EPA-AA-SDSB-85-6

                   Technical Report
    Refueling Emissions from Uncontrolled Vehicles
                          By

                     Dale Rothman

                          and

                    Robert Johnson
                        NOTICE

Technical Reports  do not necessarily  represent  final  EPA
decisions  or  positions.    They  are  intended  to  present
technical  analysis   of  issues  using   data  which   are
currently available.   The purpose  in  the release  of  such
reports  is  to   facilitate  the  exchange   of  technical
information  and   to .  inform   the  public   of  technical
developments which may form  the basis  for  a  final  EPA
decision, position or regulatory action.
       Standards Development and Support Branch
         Emission Control Technology Division
               Office of Mobile Sources
             Office  of  Air  and Radiation
        U.  S. Environmental Protection Agency

-------
                       Table of Contents

                                                        Page No

I.    Background                                             1

II.   Parameters Affecting Refueling Emissions               i
     A.     Differences Between  Vehicle Tank
           Temperature and Dispensed Fuel Temperature
     B.     Dispensed Fuel Temperature
     C.     Fuel Volatility (Reid Vapor Pressure)
     D.     Vehicle Differences
     E.     Other Factors

III.  Baseline Refueling Test Program                        8
     A.     Vehicles and Test Conditions
     B.     Test Procedure
           1.     Overview
           2.     Effects  of Vehicle Preconditioning
     C.     Test Results
           1.     Summary
           2.     Parameter Effects
                 a.     Differences Between Vehicle
                       Tank Temperature and Dispensed
                       Fuel Temperature
                 b.     Dispensed Fuel Temperature
                 c.     Fuel Volatility (Reid Vapor
                       Pressure)
                 d.     Other Parameters
           3.     Differences in Vehicles and Vehicle
                 Configuration
           4.     Prediction Equation
                 a.     Fitted Model
                       1.   Coefficients
                       2.   Variabilty
                 b.     Comparisons to Results from
                       Other Studies

IV.   Calculation of Nationwide  Emission Factors            38
     A.     Introduction
     B.     Description of Refueling Emission Factors
     C.     Calculation of Displacement Emission Factors
           l.     Methodology
           2.     Sources  of Data
           3.    .Air Quality and Health Effects Scenarios
           4.     Consumption Weighting Calculation
           5.     Emission Rates
           6.     Effects  of Fuel Weathering
     D.     Conclusions

-------
I.    Background

     The Environmental  Protection Agency  is  currently in  the
process of  developing and  evaluating  a Federal  test  procedure
for the measurement  of  vehicle refueling  emissions.   Refueling
emissions are  of direct  concern due  to their  benzene content
and the potential health effects of exposure  to  gasoline vapors
in general.  Also,  they contribute to ozone  formation,  and are
of particular concern in  areas  which  currently do not  meet  the
National Ambient Air Quality Standards (NAAQS) for ozone.

     This  report describes  EPA's baseline  program to  measure
refueling   emissions  from   uncontrolled  vehicles,    and   to
investigate  the  sensitivity  of  these  emissions  to  various
parameters.   An emission factor equation based upon  the various
parameters  will  be  developed  that  will  be  used  in  making
comparisons  with  the  results  of  other  refueling  emissions
studies.   It  will  then be  used to  estimate emission  factors
under a range of conditions.

II.  Parameters Affecting Refueling Emissions

     As was described by Hochhauser and  Campion,  the generation
of   refueling   emissions  is   a  complex  process   "involving
non-equilibrium,  unsteady   state   interphase  heat   and   mass
transfer in a  system where  the mode of  contact  between gas  and
liquid  cannot  be easily defined  or modeled."[1]   It has  been
shown,  however,  that   fairly  good  estimates  of  refueling
emissions can be obtained from empirical equations based upon a
few,   easy  to   determine  parameters.[1,2,3]   Those  parameters
that  appear to explain  the  most  variability  are:    1)   the
difference between  the  temperature  of   the  dispensed  fuel  and
the tank fuel,  2) the temperature of the dispensed fuel, and 3)
the fuel volatility.  Differences  in  the physical configuration
of  vehicles'   fuel   tanks  and  fill  necks   can  also  affect
refueling emissions,  but this  is a  variable that  can not  be
easily  quantified.   A  more  complete  description  of  each  of
these and other  parameters considered  is given in the following
sections.

     A.    Differences  Between  Vehicle  Tank  Temperature  and
           Dispensed Fuel Temperature

     A  major  factor in determining  the  level  of  refueling
emissions  is  AT,  the difference  between  the  temperature  of
the   fuel   in   the   vehicle  tank   and  the   dispensed   fuel
temperature..  The addition of fuel that  is warmer  than the  fuel
in the vehicle  tank in  turn  warms  the  tank  fuel  and  vapor
space, resulting in the  vaporization of  additional  gasoline  and
expansion of  the  vapor mixture.   This condition  is known  as

-------
                               -2-
vapor growth. On the  other  hand, addition  of colder fuel  to a
warmer tank  cools  the fuel  in the tank  and some of the vapor
present is  condensed  into liquid.  This  condition  is  known as
vapor shrinkage.   When  both fuels  are  at the  same  temperature
(AT  = 0)  neither  vapor  growth  no vapor  shrinkage occurs  and
the  volumetric  refueling losses  are  almost  identical  to  the
amount of  vapor displaced by the incoming gasoline.

     Nearly every previous study dealing  with vehicle  refueling
emissions   has  recognized the  importance  of the  relationship
between AT  and total refueling  emissions.[1,2,4,5,6,7,8,9,10,
11,12]  This  effect  is  generally  expressed as changes  in  the
ratio of either the volume  of vapor displaced,  or  grams of HC
emitted,  to  the gallons  of  fuel  dispensed.  In  all cases, an
inverse  relationship between  AT  and  volumetric  refueling
emissions  has  been seen  as  is  illustrated  in  Figure  1  taken
from  a study by the  Stanford Research Institute.   In  general,
the  same  result  holds  for  the  mass  of  refueling emissions.
However,  due to  the changing constituents  of  the vapor,  at
larger negative values of  AT a  positive  relationship  between
AT   and  mass   emissions  results.[1,2,3,6,9]    This   "turning
over"  effect is  shown in  Figure  2,  also  taken  from the  SRI
study.

     B.    Dispensed Fuel Temperature

     The  temperature  of  the  dispensed fuel (TD) can  exert  a
distinct  impact upon  refueling emissions,  separate from its use
in the determination  of  AT.   It has  long  been known that  the
amount  of   vaporization   of  gasoline   varies   directly  with
temperature.   This   is   the   reason  that   mixture   enrichment
devices are  required for  cold  starting.    All  other  factors
being  equal,  emissions  would therefore   be lower at  colder
dispensed  temperatures,  since less fuel  would be vaporized.*

     Several of the previous  studies have considered the effect
of dispensed  temperature  upon refueling  emissions.[1,2,3,4,6]
In  several  of  these,   the  value  of   AT  is  not  separately
computed  and  controlled,  so  it  is  difficult  to  separate  the
distinct  effects  that  the  dispensed  fuel  temperature  has ^on
refueling    emissions   from   its   role    in   vapor  growth   or
shrinkage.    Figures  3   and  4,  however,   show  the effect  of
dispensed  fuel  temperature,  when AT is  also accounted for, as
seen in two of the previous studies.[1,6]
     *It is interesting to  note  that this will not be  the  case
     for  the  temperature  of the   tank  fuel  (TT).    Lowering
     TT  at  a  constant  TD  will create  lower  values  of  AT,
     resulting in vapor growth and increased emissions.

-------
                                       -3-
  1.8


    x
  1.6



  1.4



  1.2



  1.0


9
I"


  aa



  0.4



  0.2
  -0.2
  -0.4
  -0.6
cn
DISPENSED FUEL
 •  S *0on.
 O 10 gtftam
 a tS
 Q 20
                                      INITIAL FUIL: 9 GALLONS
                                  IStTTEMBER 23 TO OCTOBER 12 DATA)
    -40      -30      -20      -10       0       10       20       30      40
               INITIAL TANK TEMPERATURE MINUS DISPENSED FUEL TEMPERATURE — °F
         FIGURE 1   VAPOR-LIQUID VOLUME RATIO VERSUS INITIAL AT
         Source: Stanford  Research  Institute  (9)
                                                                             so

-------
                                   -4-
-40
                                                  DISPENSED PUCt
                                                   •   S Gittom
                                                   O  tO GaHon*
                                                   A  tS Gillont
                                                   Q  ]Q GaUon
                                              INITIAL PUCt: S QAttONS
                                        10
                                                20
                                                        30
-30      -20      -10       0
   INITIAL TANK TEMPERATURE MINUS DISPENSED FUEL TEMPERATURE —
                                                                40
                                                                        50
                                                                 SA-3730-ISB
   FIGURE  2    GRAMS VAPOR PER GALLON  FUEL VERSUS INITIAL AT

   Source:  Stanford Research  Institute  (9)

-------
                                    -5-
            EFFECT OF  DISPENSED  TEMPERATURE
        7.0 --
                               FIGURE 3
    Source:  EPA Report 75-Gas-6 Part  II, August  15,1975, as  cited
    by Scott Environmental (6)
         a.o


         a.o


         7.O


         6.0


         a.o


         4.O
         2.O -



         1.O -
         OJO
           a.o
 8.O
1O.O
12JO
14JO
•Q - OI
    REID \>*3>OR R?£33LRE (FSI)
—4-—map. TEMP.=ac
                               FIGURE 4
       Source: EXXON Research & Engineering (5)

-------
     C.    Fuel Volatility (Reid Vapor Pressure)

     RVP is  a  measure  of "front-end" volatility, or the ease of
vaporization  of gasoline  at  100°F;  the  higher  the  RVP  the
greater the vaporization potential.  Refueling events occur near
this  temperature,   therefore  refueling   emission  rates  should
vary with  RVP; all other  factors being  equal,  higher  RVP fuel
yielding higher emissions.

     In many of the previous studies, the RVP of the fuel,  its
effect being recognized, was held constant.  Other studies have
attempted to explore the relationship between  RVP  and refueling
emissions  in  a quantitative  fashion.[1,2,4,6,9]  Figures  5  and
6  show characteristic  increases  in refueling  emissions  at  a
higher  RVP.[1,6]    This relationship is  also  noted in  other
studies.

     A  few   studies   have   also   considered   the  effect   of
dispensing a fuel  of one RVP into a  tank with residual  fuel of
a  different,   lower  RVP.[3,9]   The  general   result  is  larger
vapor  growth as a result of  the dispensed fuel vaporizing to
increase  the  hydrocarbon   concentration  in  the  tank  to  the
higher vapor pressure of the dispensed fuel.

     D.    Vehicle Differences

     As with  any  type  of  emissions, there will be  differences
in results  from different vehicles.   In  the  case of  refueling
emissions,   these  differences  are  primarily  related  to  the
vehicle's fuel tank system.

     Fuel tanks vary  in size,  shape, position of the  fillpipe
(i.e., rear  fill  or side fill),  fill  neck  design,  and  internal
baffling.    Differences   in  the   areas   of  the   evaporative
surfaces,  effective height  of  the fillpipe over the evaporative
surface and  turbulent  interactions  between the entering  fuel
and  existing  vapors are  among the  most  likely causes  for  the
differences  that  are observed  in  refueling  emissions  between
vehicles.

     Few of the previous studies  on vehicle refueling, emissions
have  specifically  addressed  the  issue of  differences  between
vehicles in.terms  of  emissions.   This is due predominantly to
the  fact  that most  of these  studies were  concerned with  the
efficiency of various control  strategies,  and  the percentage of
vapor  recovered was  of more  interest  than  total  emissions.
Still, Scott Environmental  Technology noted the strong vehicle
effect on  refueling emissions  as  evidenced  by the  increased
variability  in their  results  as larger  numbers of  different
type vehicles  were used.[6]   Stanford Research Institute  has

-------
                              -7-
              EFFECT OF FUEL VOLA.TILITY
      .O
     0.0
      -3O.OO
          -1OJOO
     1OJOO
       30.00
                         ft
                           AT
              - Q - PvT»= 1 2.O     -- 4 -- FVF=8 JS
                            FIGURE 5

Source:  EPA  Report  75-Gas-6 Part II,  August  15,1975,  as cited
by Scott Environmental  (6)
 u
 «
 I
     O.O
       so .00
         so .00
7OJOO
9O.OO
-a—FVV=:
                        --4---!?•.»= 1O
                           FIGURE 6
   Source: EXXON Research & Engineering (5)

-------
                               -8-
noted not  only changes in emissions  between vehicles,  but also
a change- in the shape of the  regression  line relating refueling
losses  to  AT.[9]    Also,  Exxon  Research  and Engineering  has
found average  losses at the  same test conditions  ranging from
4.5 to  5.4  gin/gal  depending  on the vehicle. [13]  Thus,  although
the previous studies have  not specifically  addressed the issue
of differences between vehicles, its effect has been noted.

     E.     Other Factors

     Several  other  factors   that  may   have  an  effect  upon
refueling emissions  have  been  considered in  previous  studies.
Among these are:  fill  rate, [1,9] amount  of  residual  fuel  in
the tank,[3,9] total amount  of fill,[1,3,9] position of  nozzle
in  the  fill-neck,[9]  and   ambient   temperature.[3,8,9]   The
magnitude of  these effects is  much less than that for  any  of
the  factors described  previously.   Therefore,  this study  has
been  designed  primarily  to  determine  the   effects   of  AT,
dispensed  temperature,  and  fuel volatility;  and  any  insights
that can  be obtained about  these other  effects  or differences
between vehicles, will be secondary.

III. Baseline Refueling Test  Program

     A.     Vehicles and Test  Conditions

     Eight  vehicles  in  all   have been tested in  the  baseline
program.  These consist of six  light-duty gasoline  vehicles  and
two  light-duty  gasoline  trucks.  The   tank   sizes  vary  from
vehicle to  vehicle,  as  do  the configurations  of the tanks  and
their internal baffling.   A  listing of the vehicles is  given in
Table 1.

     The  majority  of  the testing was  performed  on the  1983
Cutlass  Supreme,  as it  was  the  first vehicle  tested.   The
matrix  of   parameter  conditions  under which   the  Cutlass  was
tested  is  shown  in  Table   2,  along with  similar  but  less
extensive matrices for  the  1984  Escort  and the 1983  Reliant.
The  testing   of  these  vehicles  at the  various  parameter
conditions  allows   for  a  more  complete   comparison   of  the
differences  in  refueling  emissions  between   vehicles.    Of
particular  interest  here   is  the   difference  in  refueling
emissions between  side-fill and rear-fill vehicles.  Of all  the
vehicles tested, only  the 1983  Cutlass  Supreme  is  a rear-fill
vehicle,  and  the  future  fleet is  expected to be  dominated  by
side-fill vehicles.

     The remaining vehicles were tested primarily at  one  set  of
parameter   conditions.    By   testing  several  vehicles,   an
indication  of  the  range  of  refueling  emission rates   can  be
obtained.

-------
                       -9-
                        Table  1

                    Vehicles Tested
Year   Make/Model
Tank Vol. (gal)   Comments
1983  Olds. Cutlass Supreme    18.1
1983  Buick Skylark            14.5
1984  Chevrolet Celebrity      16.4
1984  Ford Escort              13.0
1983  LOT Crown Victoria       18.0
1983  Plymouth Reliant         13.0
1979  Dodge Truck W150         18.0
1979 Chevrolet 3/4 Ton
   Pickup                      19.6
                  Rear fill

                  Fuel Injected

                  Vertical Tank

-------
                            -10-
  Fuel
 9.0 RVP

11.9 RVP

10.0 RVP
12.6 RVP
                       Table 2

           Refueling Emissions Test Matrix

               1983 Oldsmobile Cutlass
Dispensed Temperature

       82°F
       92°F
       82°F
       92°F
       92°F
       82°F
 Tank Temperature

80, 92, 100, 120°F
80, 92, 100, 120°F
80, 92, 100, 120"F
80, 92, 100, 120°F
80, 92, 100°F
80, 92, 100°F
  Fuel
9.0 RVP

11.9 RVP
     1984 Ford Escort

Dispensed Temperature

       80°F
       92°F
       66°F
       80°F
       92°F
 Tank Temperature

82, 92, 100°F
92°F
72°F
82, 92, 100°F
92°F
  Fuel
9.0 RVP
11.9 RVP
  Fuel
11.9 RVP
  1983 Plymouth Reliant

Dispensed Temperature     Tank Temperature
       66°F
       66°F
       808F
       92°F

    Remaining Vehicles

Dispensed Temperature

   80, 92°F
  72°F
  72°F
  82, 92, 100°F
  92°F
 Tank Temperature
  92°F

-------
                              -11-
     B.    Test Procedure

     1.    Overview

     The refueling emissions  tests  were performed in the manner
outlined  in  Table 3.   The  test  vehicle  was  pushed into  the
Sealed  Housing  for  Evaporative  Determination  (SHED)  "cold",
i.e.,  at  ambient temperature.   At this  point there was a  10
percent fill  in  the  tank.   The  vehicle's  fuel  tank was  then
heated, either  by a  single  or  dual heating  blankets,   to  the
desired temperature,  inside the open SHED.*   The  purge  fan  was
operating inside  the SHED  during  this time. The  dispensed  fuel
had generally been heated to its desired temperature previously.

     At the end  of the heating phase, the heating blankets were
unplugged, the fuel nozzle was inserted into the  fuel neck,  the
mixing  fans   were   started,  and  the  SHED   was  sealed.    A
background reading was then taken inside the  sealed SHED before
the  refueling  began.    The  refueling  was then  performed  by
turning on the fuel cart from outside the  SHED.   The refueling
ended  at  a  95  percent  fill,  either  when  the  nozzle   had
automatically  shut  off,  or  when  it  was  shut  off  manually
outside the  SHED.  The  first  method  led to  some problems with
spillage,  so in the later testing only manual shutoff was used.

     Temperature  sensors were  located  at  three   points  inside
the tank, in the fuel  cart, and at various  other  places in  the
testing setup,  and values were  recorded approximately every  two
minutes  throughout  the test.   The  measurements of  refueling
emissions  were  made  with  the  use   of   a  Flame  lonization
Detection device.

     2.    Effects of  Vehicle Preconditioning

     As was  described  in  the previous  section  detailing  the
test conditions,  the vehicle's fuel tanks were heated either  by
single  or  dual  blankets.   The  earliest   tests  on  the  1983
Cutlass were  the only  ones  where  a single  blanket was used.
Heating of the fuel  tank  was done with  the general constraint
that vapor and liquid  temperatures  not  be allowed to differ  by
more  than  6°F.   Under  these  conditions,  the time  required  to
heat  the  tank to the  desired temperature  was  in many cases
excessive,  i.e.,  an  average  of  three  hours  to  reach  a tank
temperature of  100°F.    Also,  since  the  tank  was only  heated
from below,  temperature stratification occurred inside the tank,
     In a  few of the  tests  the vehicle  was driven  on a  road
     circuit  before  the refueling,  heating the  vehicle's  fuel
     tank in the process.

-------
                              -12-



                            Table 3

                         Test Sequence

1.      Drain and refuel tank to 10  percent of fuel capacity.
2.      Push vehicle into shed.
3.      Connect heat blankets and thermocouples.
4.      Heat vehicle tank to desired temperature.
5.      Insert fuel nozzle.
6.      Close shed and start mixing  fans.
7.      Take initial sample reading  (using FID).
8.      Refuel tank to 95 percent of fuel  capacity.
9.      Check for spills and nozzle  shutoff.
10.      Take final sample reading (using FID).
11.      Disconnect heat blankets and thermocouples.
12.      Remove vehicle from shed.

-------
                              -13-


with   the . vapor   temperature   lagging  behind   the   liquid
temperature by  several  degrees.  At  the time of  refueling,  it
was  unclear  as  to  whether  an  equilibrium  existed  in  the
vehicle's fuel tank.

     The use of dual  heating blankets (the second blanket being
used to heat the  top  of the fuel tank)  drastically reduced the
heating   time  required,    and   alleviated   the   problem   of
temperature stratification in the fuel tank.   This approach was
therefore adopted as the standard tank heating procedure.

     Figures  7,  8,  and  9   show  the  results  of  the  refueling
tests  for various  methods  of  vehicle  preconditioning.   These
data are  solely from tests  on the 1983 Cutlass,  at  a constant
RVP and dispensed  temperature in each case.   Figure  7  gives  a
comparison  of  the  results  when  the  vehicle was heated by  a
single blanket, heated  by dual blankets,  or  driven  on a  road
circuit.   Figure  8 shows further  results  when the vehicle  was
heated by dual  blankets compared  to  when  it was driven on  a
road circuit.   Finally,  Figure 9 shows  the results of tests in
which the vapor and fuel temperatures  inside  the  vehicle's fuel
tank were permitted to differ markedly.

     Several conclusions  can be drawn from  these  test results.
Refueling emissions were lower by  approximately  0.7 gm/gal when
under  single  blanket  heating  versus  dual  blanket  heating.
Large  differences   in   the  fuel   tank   liquid  and   vapor
temperatures  may   affect   refueling   emissions,   with   lower
emissions resulting when the vapor temperature lags  behind the
liquid temperature.   The results  from  the  road  circuit  tests
appear to fit better with the single blanket  tests, but half of
these also fit with the dual blanket tests fairly well.**

     It  is  unclear  how  much  of  the  difference between  the
single and  dual  blanket test  results  can be  explained by  the
temperature stratification in  the  vehicle's  fuel  tank, and what
must be  explained  by other  factors.   Because of its  heating
time  advantages,  and  the  question of  tank  equilibrium,  dual
blanket heating was used in the  actual baseline  testing.   The
results  from the  tests  on the  vehicle prepared on  the  road
circuit,   which  are taken  to  represent  a real-life  situation,
suggest  that  the  dual  blanket  heating  procedure  may  yield
slightly  conservative,  but  generally  accurate,  estimates  of
refueling emissions in real life situations.
**   When  90  percent confidence  intervals  for the  regression
     lines are used.

-------
                -14-
 VEHICLE PRECONDITIONING EFFECTS
       R V P = 9 . E DI5P.TEMP.=32
  7.80
J


5
v/
Ul
LJ
Ul
in
a
j
, BB
  3.00
    SINGLE BLRNKET

    DURL BLRNKET

    RQRD PREP
   -I 5.B   -S.B
              S.B
IS.B
25. B
      AT  CDEEREE5  F)
      R V P = I B .' 0  DI5P.TEMP.=32
  9.00
/°\

—1 a. SB
n
a
>  a. 00
in
y T.BB
Ul
m..«
a
j
  B.BB
                . DURL  BLRNKET

                . RDRD  PREP
   -IS.B -10.0 -S.B   B.B  S.B   IB.B  15.B
      AT   CDEEHEE5  F>
            FIGURES 7 and 8

-------
LRREERVlpNJERNgRLD
                                          FF
         12.0
i
in
                     x  V.T .-L.T .
                     A  V . T . «• L . T
                     . . V . T . »L . T
                                      2S.0
              AT
            CDEGREES   F
                         FIGURE 9

-------
                              -16-
     C.    Test Results

     1.    Summary

     It has  been found that  the refueling emissions,  in grams
of HC per gallon of fuel dispensed, can  be estimated accurately
by  a  multiple   linear   regression  model  relating  refueling
emissions to  the difference  between  vehicle tank  temperature
and  dispensed   fuel   temperature  (AT),   the  dispensed  fuel
temperature,  and  the  fuel volatility.   The effects  of  vehicle
configuration   were   explored   and   found   to   be  of   some
significance.*   The   following  general   conclusions   can  be
reached from the results.

     a)    lower  tank  temperatures,  relative to  the dispensed
           fuel temperature,  yield higher emissions,

     b)    higher dispensed temperatures yield higher emissions,

     c)    higher RVP  dispensed  fuel yields  higher  emissions;
           and,

     d)    vehicle configuration  can  have  a  significant impact
           on refueling emissions.

A more detailed look at each of these factors follows.

     2.    Parameter Effects

     a.    Differences  Between  Vehicle  Tank  Temperature  and
           Dispensed Fuel  Temperature

     Due to  the phenomena of  vapor shrinkage and vapor growth
the difference  between the tank  temperature  and  the dispensed
fuel  temperature  has  a   significant  impact  upon  refueling
emissions.    This  difference   is  defined  herein  as  TT  -  TD
and  will be referred  to  as  AT.   The  tank temperature,  TT,
is measured as the liquid temperature in the vehicle fuel tank.

     Figure  10  shows  a  plot  of  refueling   emissions  against
AT.   The  general  trend  of  higher  emissions  at " lower  AT
values, representing more vapor  growth,  is apparent even  when
other factors such as  dispensed temperature and fuel volatility
are not considered.  Very few tests were  run  at  negative values
of  AT,  and  none  below  AT=  -12,  so  the  turnover  in  the
     The results  of  all of  the valid tests  performed  on  each
     vehicle  are  summarized in  Appendix A.  This  includes the
     special  tests  run on  the  Cutlass  in  addition  to  the
     primary baseline testing.

-------
     13.500   +
                                      REFUELING LOSSES  vs.  TTAMK-TD,sPBNsED
     12.000
CP
r^W

 1 O
  .J
  O
  55
fc,
     10.500
     9.0000
     7.5000   *
                  *
             +    *
     6.0000
     4.5000   +
     3.0000   +
      .5000   »
*          **   •
       *   *
    *   ***         *
           *
      ******               *
            2   «
        *  *
           ***      •
*      **  *2       * *
      *  *       *   *
       *               2
      *  *3 *      *   * *
       * »***       *  * *
          2 •
   **      **         2 *2*»
       2 *  *•«     * *        *
    *      *2**      *   ****    •»
      *               **2  2         *
        * * *      **    **         *  *
          2**2   *   **    *      *
           ***3*«        ***
                **»2*****             *
             *         * *            2
              *      *   *          »*         *
              *                            *  *
              *
                                            *
                                     *           *
     0.
           -14.000             .00000             14.000             28.000             42.000
                    -7.0000             7.0000             21.000             35.000             49.000
                                                    FIGURE 10

-------
                              -18-
relationship  between  AT and  mass emissions noted  earlier from
other studies is not seen here.

     b.    Dispensed Fuel Temperature

     Figures 11 and 12 illustrate the  effect  that the dispensed
fuel  temperature  can  have  upon refueling  emissions.   These
plots  are  separated  by fuel  volatility,  and  the  refueling
emissions  values  are plotted against  AT  in  order that  the
effects  of  dispensed temperature can be  separated  from  these
other parameters.   The  values plotted  at  each  AT  are the mean
responses,  along with the  standard deviation  of  the observed
test  results  (where  applicable)  for  a range  of  ATs centered
at that  point.   A smaller  standard  deviation at a  given  point
in these plots will not necessarily mean  a more precise point,
as the same  number  of tests  were not  performed  at  each point.
They  are presented here solely  to  give  an  indication of  the
variation in the test results.

     The  figures  indicate  in general  that  higher  dispensed
temperatures  will yield higher  refueling emissions.   This  is
especially  true  at  values  of  AT  around  OF0  where  a  10°F
change  in TD  produces  a   1  gm/gal change  in  emissions,  with
the  effect  being  less  notable at  higher values of  AT,  where
all values tend  to  converge.   These results are consistent with
those seen in previous studies.

     c.    Fuel Volatility  (Reid Vapor Pressure)

     Figures  13   and  14  illustrate the  effect  that the  fuel
volatility  can  have  upon  refueling  emissions.   The  form  of
these  plots  is  the same as  in  those  used  to  illustrate  the
effect of dispensed  fuel temperature,  only here the dispensed
fuel  temperature  is  held  constant  as  opposed  to  the  fuel
volatility being fixed previously.

     These figures  give  a  clear  indication  that a  higher fuel
volatility,  denoted  by a  higher  RVP,   will   yield  higher
refueling emissions,  as  was seen  in other studies.  As with the
dispensed fuel  temperature, this  effect  is  more noticeable  at
low values of AT.

     Severa'l  tests   were   run  where  a  fuel   with  a  lower
volatility than  the dispensed fuel  was placed  in the vehicle's
fuel tank.   This represents the case in which  a vehicle's fuel
weathers  and loses  some of its  volatility between refuelings.
The  expected result,  as described  in  the SRI  study,  is higher
emissions resulting from vaporization  of  the dispensed  fuel  to
increase  the hydrocarbon concentration in the  fuel  tank to the
higher vapor pressure at the  dispensed fuel.   The  results from
these tests  are  shown in Figure 14  as the three single points*,
and reaffirm the results in the SRI  study.[2]

-------
                   -19-
         DISPENSED TEMPERATURE EFFECTS
             RVP =  3.0  P£
7 ,00
.  D I SP  TEMP . =32 F
.  DI5P.TEMP.=32 F
           a.a
                   IB.a
                          2B.B    3B.B    HB.B
       AT   CDEGREE5   F
              RVP
  a  PS
                            DISP.TEMP.-82 F
                            DISP.TEMP.-92 F
            U  0
                          2B.B
                                 30.0    HB.B
       AT   CDEGREE-5   F)
            FIGURES 11 and 12

-------
                         -20-
              FUEL VOLATILITY EFFECTS
     DISPENSED  TEMPERRTUPE  B 2  F
  3 . ea
                             .   RVP =12  B  PS I
    .-ia.0    a . a
i0.a    20.a    30.a
                                            MH.B
        AT   CDEGREE5   F)
     DISPENSED  TEMPERRTURE  32  F
  t i .a
 • IB. a
\
  s.a
in 6.B
u
LJIS.B
ui
QH.a
J
  3. B
        »  RVP- 3.0  PSI
        .  RVP*IB.B  PSI
        4  RVP-ir.a  RSI
        .  RVP-I I .3*IB.B  PS I
    -IB.B    B.B
IB.B    2B.B    3B.B
        AT   CDEBREE5   F
              FIGURES 13 and 14

-------
                              -21-
     d.    Other Parameters

     It  has  been  suggested  that  other  parameters,  such  as
ambient  temperature  and  fuel  dispensing  rate  may  influence
refueling  emissions.   In  this  particular  testing program,  it
appears that  the time required to heat  the vehicle's fuel tank
and  the   dispensing  rate  of   the   fuel  may   be  of   some
significance.    The  design of  this program,  however, has made
any   significant    analysis   of   these   effects   virtually
impossible.   Thus,  although the  presence of these  effects are
recognized,   they  cannot  be   determined  here.    Also,   in
comparison  to  the  effects  due  to  AT,  the  dispensed  fuel
temperature,   fuel    volatility,   and   vehicle   configuration
(discussed  in  the  next  section),  these other  effects are  of
much lesser significance.

     3.    Differences in Vehicles and Vehicle Configuration

     For most of the vehicles  tested,  there  is  insufficient
data  to   do   independent parameter   analyses.    Therefore,   a
multiple linear  regression has been  fit using all of  the  data,
and the residuals,  the  actual  values  minus the values predicted
by the regression equation, have been  examined..   The different
patterns in the residuals from vehicle to vehicle  can  give  an
indication  of  the  vehicle effects.    Figures  15-22  show  the
residuals plotted against the  predicted values  for each vehicle
individually,  all plotted at the same scale.

     Of particular  interest  in  these  residual  plots are  the
residuals  associated with the LTD Crown Victoria,  the  Escort,
and the  Reliant.   The  residuals  associated  with  the LTD  are
quite  extreme,   higher  than those associated with  any of  the
other vehicles  aside from a  few tests  on the   1983  Cutlass.
This may  be  due to  the  LDT's  unique  fuel tank  configuration;
its height  dimension being larger than its width,  with almost
no  drop  in the fill neck.  This configuration  is  atypical  of
the automotive fleet, and the results from  the tests  on  the LTD
would  skew the  prediction equation derived  from the multiple
linear  regression  model.   Thus,   although the   LTD  shows  the
potential  range in  refueling  emission rates,  its  test  results
have not been used in formulating  the  prediction  equation  to  be
used here.
     The  results  are  also  listed-  in  Appendix  A  under  the
     heading Fuel Weathering Tests.

-------
           1983 Olds. Cutlass Supreme
                                            -22-
        1984 Ford Escort

CO
en

I
(Q
on
co
3
to
3

2
(A
0)
              Predicted  (gm/gal)
             Predicted  (gm/gal)
            1983 LTO Crown Victoria
                                                    '  1983 Plymouth Reliant
 (Q
 O) •-'
 X

 en i m«
ra
o»
O)
  m

  I -
  (A
  0
(0
3

-------
                                             -23-
              1983 Buick Skylark
                   1984 Chev. Celebrity
CO
O)  "
I.
         n  ••-••
         o»
                                              0>
(0
Cft
O
         CO
         3
          0»
          O
          05 -.»»?»•
             Predicted  (gm/gal)
                      Predicted  (gm/gal)
              1979 Dodge Truck W150
                   1979 Chev. 3/4 ton Pickup
 co ••—
 o>

V-
         a
         O)  '•"•'

         E
         O)  I 9M4
 CO
 3
         (0
         3
 (A
         (0
         4)
         a
              Predicted  (gm/gal)
                      Predicted  (gm/gal)
                  FIGURES 19-22
RESIDUAL PLOTS BY VEHICLE

-------
                              -24-
     The residual  plots  associated with  the  Escort and Reliant
both  show  a distinct  pattern of  underestimation  of  refueling
losses  (positive  residuals)  at   low  predicted  values,  and
overestimation of higher predicted values.  This  would indicate
that  the  fitted  model   here,   dominated  by  data  from  the
rear-filled  1983  Cutlass,  may  not  be  the  most  accurate  for
other vehicles, particularly side-filled vehicles  which  have a
large vertical drop in the fill neck.

     A  further  comparison can  be made  by fitting a  multiple
linear  regression model,  based upon  the same  parameters  as
discussed  before,  for  the  Cutlass  test  results  and for  the
Escort test results.  These  are  the only two vehicles that have
large   enough  data   bases   to  make   reasonable   parameter
estimates.   The resulting fitted models are as follows:

Cutlass:  Refueling Loss (gm/gal) =
   -5.584 - 0.114[AT(°F)] + 0.0857[TD(°F)] + 0.520[RVP(psi)]
           R2 = 0.856

Escort:   Refueling Loss (gm/gal)  =
    -6.687 - 0.039[AT(°F)] + 0.081[TD(°F)] + 0.545[RVP(psi)]
           R2 = 0.912

The  resulting  equation  indicates  that the primary differences
in refueling emission between these vehicles  lies in the  amount
of  vapor  shrinkage  or  vapor  growth  that  occurs during  the
refueling event.  These processes would naturally be related to
vehicle configuration,  so this result appears reasonable.

     A  few  more insights  can  be gained  from an  examination of
the  residual  plots.   Aside  from the  Celebrity whose  residuals
are  consistently negative, although not  as extreme as those for
the  LTD, the residuals for  all  vehicles generally fall  within
+ 1  gm/gal.  This  includes the test results from the two light-
duty trucks which agree well with the  prediction  eguation, even
though it is based primarily upon automobile tests.

     In  summary,   it  is  clear  that differences  do  exist from
vehicle to  vehicle.   Nevertheless,  an eguation based  upon  all
of the  data,  except  the LTD tests for reasons  as noted before,
appears to work well on average.

     4.     Prediction Eolation

     a.     Fitted Model

     A  primary goal  of  this  study .has been  to develop  an
emission factor  equation based  upon the  parameters that  affect
refueling   emissions.   This  has been  achieved  by  fitting  a

-------
                              -25-
multiple  linear  regression model  with  the  data from  seven of
the eight vehicles tested.*

     The prediction  equation  developed  for  refueling emissions
from an uncontrolled vehicle is given as follows:

     Refueling Loss (gin/gal) =
  -5.909 - 0.0949[AT(°F>] + 0.0884[TD(°F)] + 0.485[RVP(psi)]
           R2 = 0.786
           MSE = 0.732

     This  equation  will be used to  estimate emission  factors
under  a  range of  conditions,   and  also  will  be compared with
results  from other  refueling  emission  studies.   The range of
conditions over which actual tests were made is given below:

     To = 66-68°F; RVP = 9.0-11.9 psi;  AT = 0 to 10°F
     To - 78-85°F; RVP = 9.0-12.6 psi;  AT = -2 to 40F°
     To = 88-95°F; RVP = 9.0-11.9 psi;  AT = -12 to 32F°

     1.    Coefficients

     Each of the  parameters included in the regression model is
statistically  significant   at   a confidence  of  99.9  percent,
i.e.,  there  is  less  than a 0.1  percent  probability that any of
the  three  parameters has  no   effect  upon refueling emissions.
The  magnitude of  the effects  due to each  parameter is  given by
the  associated coefficient  in  the regression equation.   A 10F°
increase  in  AT   will   lower  refueling  emissions  by nearly  1
gm/gal;  a   10°F   increase  in   TD   will   increase  refueling
emissions  nearly  1  gm/gal;  and  a  1 psi  increase  in RVP will
increase refueling emissions nearly 0.5 gm/gal.

     There was some  consideration as  to whether a . linear model
is sufficient to  explain the  data over  the  range  of conditions
where  the  regression equation is applicable.   A  look   at  the
residuals  (actual  gm/gal minus  predicted gm/gal)  can  give an
indication  as  to  whether  the  assumption  of  linearity  is
appropriate.  Figures  23-26 show the  residuals  plotted against
the  predicted  values and  against each  independent  parameter.
The  residual scatter  in  these plots  appears  random,   and no
systematic trends  are evident,  which would indicate significant
nonlinearity.  Also, several other forms of  the regression were
considered  in  which   interaction   and  nonlinear   terms  were
included.   These   are   presented in  Table  4,   along with  the
associated  R2 .    The  R2  value  is   a  measure  of  a  model's


*    Does not  include  testing on  the  LTD or the  special tests
     on the 1983 Cutlass (single blanket, etc.)

-------
                             -26-
                             Table 4

          Alternative Formulations of Regression Model


Form of Model _ _ R2

Gm/Gal =
1) exp [ao+ai(TT)+a2(RVP)+a3(TD)(TT)+
     ay(TD)(RVP)]                                          .805
2) exp [a0+a.l(TT
     a,(RVP)]                  .                            .823

3) a0-l-al(TD)+a2(T0)(TT) +
            T|)                                            .613
4) a0+ai(AT)+a2(AT2)+a,(TD)+
     ay(RVP)                                               .790

5) a0+al(AT)+a2(TD)+
     a3(RVP)                                               .786

-------
2.5000   +
2.0000
 \.5000
 1.0000
-1.0000   *
 -t.5000   +
                           RESIDUAL  VALUES vs. PREDICTED VALUES

                           Fitted Model    AM Vehicles
                                                                                                                   I
                                                                                                                   to
 -2.0000  *
                         3.7778
2.0000           .a.ma    . Kt.M    5.5556       4
        2.88B9            4.6667
                              PREDICTED (gm/gal)
   	*	*	
        9.1111
8.2222            10.000

-------
 2.5000
 2.0000   *
                              RESIDUAL VALUES VS.  TTANK-TOI8PKN8BD
 1.5000
                                      *        *
                              «»* *       *   *
                              *                *
                                      *
 1 .0000
                               *
                              * 2
\*  .50000
i
+~r

g

M
«J
 .00000
-.50000   +
-1.0000
                           «             *   »        »   *
                    *      *          *      •           *
                               2  *     -      *        »   *
                              ***       *   «•         »  » »
                            •         •                   *
                               '*       ,  «•   * *
                    *            «       * *   »  2    *
                              «*        *    *
                                          *
                                  * »    •    »  *  »        *
                            *     *    2  *   •*        *
                            » ««  «          •  »
                             *»» «    *    *       *
                              ***       2 *         *
                               * *  *
                                  *  *2    *
                               •*«2    *
                           »»            *•
                             *   *
                              *    •
                                                                                                                             00
                                                                                                                              I
-1.5000   +
                            *   *
-2.0000
                                                                     -*-»•"
           14 000              .00000             14.000             28.000
                   -7.0000             7.0000             21.000             35.000
                                                                                42.000
                                                                                         49.000.
                                           TTANK~TD i s FENS BD  (°F)

-------
                                      RESIDUAL VALUES  vs.  DISPENSED TEMPERATURE
    2.5000   +
                                                                                  * *
    2.0000   +

                                                                                    *




    1.5000   *                                               *                   ,    ,.  .

                                                                               *
                                                                *                *

                                                                                 * * *

    I.OOpO   +
                                                             *****
                                                           *
            +                                             * * * 2  *                  •
^"*                                                         »» »                »
*rt                                                         ,        *              . 2   •
0*   50000   *                                              22*              ** 3 •
N  '                                                      *         *                  *
                                                             ***    .   *         *
             .                                            ««*«•             »**2                                         ',
~                                                           2 •                                                               £
J                                         '	          *      	   •   	                         I
S  .00000   ^	•	**	—	    ill                - •  -
J3                             *                        2   *     *            *     . •
O                                *                      * **   *  2                  *
*Tj          <.                   *                           **•*                 •
M                                                   ',**••*•
g                                                         • • 3 *      • • '          *

    -.50000   *                                              ^ •                   ',.*.'
                                                                                *    * *  *
            +                                            * *                     • *  •
                                                             *                  *•  •  3    •
                                                        •              *         *     2*

    -1.0000   *                                                                         *  **
                                                           2                          *
                                                                                   *

            +

                                                             *

    -1.5000  *                             '                                      *        .
    -2.0000   *                                                            j   	__+	+	+	+

           60.000  *          6B.OOO            76.000            84.000            92.000
                    64.000            72.000            80.000            88.000,            96.000.


                                                     ID TEMPERATURE  (°F)

-------
2.5000  +
      RESIDUAL VALUES vs.  FUEL VOLATILITY


                                          2
2.0000


1.5000 +


+


1.0000 *


X-s t
l-»
(0
\ .50000 *
§


a
3 .00000 < 	
Q
l-»
09


-.bOOOO +


•f


-1.0000 *


*
-1.5000 *
•fr
-2.UOGU »




•
*



2

4
*
* . *
4 2
22
3 *

3
3
4
2
.4
2
•
2
5
3
3 •

•
*




*
2 •
5 •
2
*
»
2
3

3
*
3
4
4
4 •
2 «
4
5 •
2

4
5
2
5
5
*
3
2

5
2
3
2

2

l*
+ + Jt > — -*-—*-—+ 	 *
      8.5000
               9.0000
9.5000            10.500            11.500
         10.000            11.000            12.000
                                                                                                                   I
                                                                                                                   U)
                                                                                                                   o
                                                                                                                   I
                                                                          12.500
                                                                                  13.000
                                    REID VAPOR PRESSURE  (psi)

-------
                              -31-
ability  to.  predict  trends  that  are  presented  in the  data.
These results show that the linear model  is  sufficient  to model
the  baseline refueling  data,   and its  simpler  form makes  it
easier to interpret.

     2.    Variability

     There is a large amount of variability  in  the results from
the  refueling  emission  tests;  values  ranged  from under  3.0
gin/gal  to  over  11.0  gm/gal.    Nearly   80  percent  of  this
variation  is  explained  by  the  three  parameters:   TD,  AT,
and  RVP,  as  indicated  by the  R2 value associated  with  the
regression  model.   However,   a   fair   amount   of  variability
remains unexplained as shown by the mean  squared error  value of
0.557 gm/gal.

     Much  of   the  remaining   variability  is  due   to   the
differences in vehicles  as discussed  before, but  other  factors
are  also  involved.   Several parts  of the  test  procedure  are
subject to certain degrees of  error,  and can therefore  lead to
test variability.  The first of  these  involves the  heating of
the fuel tank  in  the vehicle as mentioned before.   What effect
the  heating  rate may  have  is  unclear.   The  same can  also be
stated for the dispensing rate of the fuel.  The heating of  the
dispensed fuel  in the  fuel cart also varies somewhat, and could
very possibly slightly affect the RVP of the fuel.

     All  of  the  above  effects  are   generally  negligible,
however, in  comparison to the  effects  caused by  even  a small
fuel spill or  spitback  at the end  of a refueling.  Tests  in
which  spills  estimated  at  over  1/2 of  a cup  occurred  were
generally  voided,  except  where  no  significant  effect -was
noted.  However,  even  a  spill as  small as  1  1/2  tablespoon
could generate  a  one gram per  gallon increase  in emissions if
it were to  completely evaporate.*  The concern  over spills  was
large enough to change the test procedure used  in this  program
to call  for  manual  shutoff of  the  dispensed fuel.  Also,  the
effect  is   large  enough  to   warrant  being   considered   in
determining the total  emission factor.   This will be discussed
more in the sections  dealing with the refueling emission factor.

     b.    Comparisons to Results from Other Studies

     The prediction  equation derived here can  be:   1)  used to
make comparisons  with  the results measured  from other  studies,
and  2)  compared to prediction  equations  derived  elsewhere.   A
brief summary  of  the results from other  studies,  and  how they
compare to those  in  this study, is given in Table  5.   Overall
the results of this study are in good agreement with past work.
*  Using 10 grams/tablespoon and assuming a 15 gallon refueling.

-------
                              -32-


                              Table 5

             Summary of Results from Previous Studies
Year

1975
Study

Scott
T0(°F)

55-70
  AT(°F)  RVP(psi)

-30 to 30 8.8, 12.0
Comments &
to Current
Re I ati
Study
                                                              on
1975
Scott    80-85
(station)
        -20 to 30 8.0-8.8
1975
CAPE9
30-90   -40 to 40 7-13
  Laboratory
  study;(lab)similar
     effects   seen
  for    RVP,    To,
  and           AT;
  resulting  gm/gal
  0-0.5      gm/gaI
  I owe r

  Field study of 4
  Stage          11
  Recovery  Systems
  on vehicles  at a
  service   station
  ;   some  base!ine
  testing     done;
  mean  values  from
  baseline  results
  agree  we 11  wi th
  predictions  from
  this study.

  Performed      in
  SCOTT   mini-shed
  on        vehicle
  tanks;  no  tests
  of       positive
  ATs   at   higher
  T0s;    can   only
  compare      with
  current  study  at
  around    AT=OF°;
  agreement      is
  good    in    this
  range.

-------
                             -33-

                        Table 5 (cont'd)

            Summary of Results from Previous Studies


                                             Comments & Relation
        Study    T0(°F)      AT(°F)     RVP(psi)    to   Current


1975    SRI      68-85  -30 to 45 6.9-8.6(10.6) Tests     at     a
                                               station    on    a
                                               vehicle       fuel
                                               tank;
                                               comprehensive
                                               tests  looking  at
                                               many   independent
                                               variables     that
                                               affect   refuel ing
                                               emissions;    good
                                               general  agreement
                                               with       current
                                               study;    slightly
                                               low   at   low  ATs
                                               and    higher   at
                                               high ATs.

1976    ER&E    10-100 -20 to 10 7-13          Fuel  tank   tested
                                               in      control led
                                               environment;
                                               vapors   collected
                                               in   Tedlar   bag;
                                               yields     results
                                               1-1.5        gm/gal
                                               lower  at  0°  AT,
                                               nearly   equal   at
                                               10° AT.

1976    Union   75-85  -20 to 30 8.8-9.0       Tested       random
                                               vehicles    at    a
                                               refuel ing
                                               station;  29 used
                                               for       base Iine
                                               results;
                                               estimates        1
                                               gm/ga I   I ower   at
                                               0°             AT,
                                               approximately
                                               equal at AT  10°

-------
                             -34-

                         Table 5 (cont'd)

             Summary of Results from Previous Studies
Year

1978
Study   T0(°F)    AT(°F)

ER&E     85      -1
                  RVP(psi)

                     9.1
       Comments &
       to Current
         Re I at
         Study
                                                              on
         Looking        at
         efficiency  of  an
         onboard   control
         system    on    3
         vehicles.
         Baseline
         est imates   .7-1.6
         gm/gal       lower
         than    predicted
         by current study.
1978
Mob i I
82-85
8.2-12.0 Tests
         single
          on     a
         veh i cIe.
Good    agreement
with  predictions
from      current
study.

-------
                              -35-
     In order to estimate  the results from other  studies  using
the prediction  equation,  information is  required on the  test
conditions:  fuel  RVP,   tank  temperature,  and  dispensed  fuel
temperature.  Also  the  refueling losses need  to  be reported in
total  grams per gallon  of refill,  or  in a  form that can  be
readily converted  to this  form.  Only  a  few of  the  previous
studies met all of these criteria.

     In their  tests  involving   Stage  II  vapor  recovery  vapor
balance   systems   at   a   retail   gasoline   station,   Scott
Environmental    arrived    at    estimates    for    uncontrolled
emissions.[6]  Their  study involved two phases of testing:  the
first on  thirty  control  vehicles at  a service station  and  the
second  on  random   vehicles.   For  each  phase,   two  series  of
vehicles  were  tested.   Average  RVP  and  dispensed temperature
are only  provided  for  the two  series  in the second phase  of
testing.  These two series  yielded  baseline emissions of  5.505
and 5.593  gm/gal.   The  average  for  the other factors  are also
given for these two series;  RVP = 8.0286  and  8.6440,  dispensed
temperature  =  81.265  and  81.0196  °F,  and tank  temperature  =
81.867  and  82.0796  respectively.   Using  these  conditions  and
the regression  equation derived in  this  report,  estimates  of
5.109 and 5.352 gm/gal  are obtained.   These are  slightly  lower
than  obtained  by  Scott,  but still  well  within  the range  of
uncertainty in the data.[6]

     In a  study  done by  the  Mobil  Research and Development
Corporation  in  1978 a series  of refueling emission  tests  were
run on  a  1978  Pontiac Sunbird.   During these  tests the  vehicle
was equipped with  an onboard control  system, so  the  total  HC
emissions  given is  the  sum of the HC  collected in the  canister
and the refueling  emissions measured  in the  SHED.  These tests
were performed  in  a  SHED,  in the  same general  manner as  the
test  in  this  study, with the  only  exception  being that  the
vehicle was preconditioned by driving and not just  heating  of
the fuel tank.[13]

     The  Sunbird   was   tested   at   the  following  conditions:
dispensed  temperature =  82-85°F,   AT =  OF0, and RVP  ranging
from  8.2  to  12.01bs.   The  resulting losses, along with  the
estimates    from .this  study  are  given  below.    The  equation
derived in  this  report  is  not strictly applicable  at RVP levels
under  9.0,  but  the  estimates  are  given here  regardless.   Mean
estimates  from this study  are generally on the high  end of  the
ranges  given by  Mobil,  but also note the  variability  in  their
test results.

-------
                              -36-


     RVP   .  tt Tests     Total HC (qm/gal)     This Study*

   8.2-8.5      8            3.72-6.82            5.52
   8.6-9.0     10            4.2-5.60             5.74
  10.3-10.6     3            5.9-7.1              6.54
  11.2-11.4     3            5.5-7.0              6,95
  11.8-12.0     3            7.0-7.2              7.24

     The Stanford  Research Institute study  involved  tests on a
26-gallon General Motors and a 26-gallon Ford  fuel  tank used in
1973-4  vehicles,  at  a  service  station.   The results  from the
tests on the  GM  tank at a fill  rate  of 5.3 gallons  per minute
and  a  fill   of  20  gallons  are given below  along  with this
study's  estimates  at  the given  conditions.[9]   These  tests
represent  those  most similar  to this  study's  testing.   Where
the  conditions  fall  within  the  ranges  for  which  the  EPA
equation is applicable,  the  agreement is good, generally within
.50 gm/gal.   Only  when  the equation is extrapolated  far beyond
'its  applicable range is  there a significant  disagreement with
the SRI results,  illustrating the dangers of such extrapolation.


 AT     To.         RVP
-17     79.5       8.5
  0     80         8.5
  1     80         8.5
  2     78         8.5
 27     76         8.5
 35     79         8.5
 41.5   79.5       8.5
 43     77         8.5

     Exxon Research  and Engineering performed a series of tests
on 3 vehicles in 1978,  in order to determine  the  efficiency of
an onboard control  system.[13]   The vehicles tested  were:   a
1978 Caprice,  a  1978 Pinto,  and a  1978 Chevette,  and  the test
conditions  were:   AT  =  -1F°,  TD  =  85°  F,  and  RVP  =  9.1
psi.    The  resulting averages  for  each  vehicle,   along  with
predicted  values from this study are given below.

                              GM/GAL
         Vehicle           ER&E[13]   This Study

         Caprice            4.9         6.1
         Pinto              4.5     •    6.1
         Chevette           5.4         6.1
GM/GAL
SRI[2]
5.13
5.03
4.82
4.52
3.07
2.09
1.26
1.42
This Study
6.85
5.28
5.19
4.92
2.37
1.88
1.30
0.94
Difference
1.72
.25
.37
.40
-.70
-.21
.04
-.48
     Using 83.5°F and mid-range of RVP interval.

-------
                              -37-
     In  this  case,  the  results  from  this   study  appear  to
significantly  overestimate the  Exxon  results,  especially for
the Pinto.   The testing on these vehicles  was  performed with a
prototype  refueling canister  on each  vehicle,  and the  tests
were part  of a  larger  test sequence including measurements of
evaporative  and  exhaust   emissions.    These   differences  and
consideration  of the  fact that  a  comparison  is being  made
between  individual   vehicles   and  one  case  and  a  population
average  in  the other case  can explain some of the discrepency
in  the  results.   Also,  these  vehicles  are   older  and  of  a
different fuel tank design than those tested here.

     Four prediction  equations that  consider  factors other than
AT have  been found  in  the relevant  literature.   One  of  these
has a  correlation coefficient, r, of only  0.25 associated with
it, so it  has  not been  included  in the  analysis  here. [4]   The
remaining  equations,  and  associated parameter  regions  where
they are applicable are given below.

CAPE9(EPA)[3]

gm/gal = exp[-0.091703 + 0.0011521(RVP)(TD) - 0.0012605(TT)
                    + 0.054094(RVP)  + 0.00010725(TD)(TT)]
         R2 = 0.945
         RVP =  7 psi
         RVP = 10 psi
         RVP = 13 psi
SE - 5.6%
TD = 50 to 90 °F
TD = 40 to 80
TD = 30 to 70
op

op
TT = 50 to 90 °F
TT = 40 to 80 °F
TT = 30 to 70 °F
Exxon Research and Engineering Co.[l]
gm/gal = exp [-1.23 + 0.0185(TD) + 0.00170(TT) + 0.118(RVP>]

         R2 = 0.951     SE = 12.4%
         RVP = 7 to 13 psi
         To = 10 to 100 °F
         TT = 30 to 90 °F
         AT > -20 °F

Union Oil[2].
gm/gal = -15.178+0.1503(TD)+0.002523(TD)(TT)-0.0000002099(T0)2(TT)2

         R2 = 0.5740     SE = 0.3873 gm/gal
         RVP = 8.8 to 9.0 psi
         TD = 75 to 85 °F
         TT = 70 to 115 °F

     The equations  determined  in  the CAPE-9,  Exxon  and  Union
studies are  based upon 140, 43, and  29  tests,  respectively.[1,
2,3]   The  testing done  in  the CAPE-9  and  Exxon studies  was

-------
                              -38-


performed  on  a vehicle  tank  in  a laboratory  setting.[1,3]
Union performed its  testing on  vehicles refueling at  a retail
gasoline station.[2]

     As is  readily apparent, the form of  these three equations
differ among themselves,  and from the equation derived  in this
study.  This makes a direct comparison  of  the results  somewhat
difficult.  Figure 27, however, shows plots  of refueling losses
versus  for  each  equation  and  the  Stanford  results,  at  a
dispensed temperature of 79°F and an RVP of  8.5 psi.  An RVP of
8.5 psi is  slightly  out of the applicable  ranges  for the Union
results and those  derived here,  but the figure is  still useful
for comparison.  Figure 28  shows  a further comparison  of  this
study's results to earlier work  by EPA as  cited  in the  Scott
study.[6]   This  is shown  in a separate figure as the conditions
are slightly different from those show for the other studies.

     Figures 27 and 28 show very  good general  agreement between
the results from the  CAPE 9, Stanford,  the earlier work by EPA,
and  this   study,  over  their  applicable  ranges  of  AT.   The
studies by  Union  and Exxon yield  somewhat lower  estimates of
refueling losses at  negative values  of AT,  but  their  results
are  not  radically  different.    All  in all,  considering  the
differences  in  testing apparatus and procedures,  the  results
from  the ,various  studies tend to  confirm  each  other  and  the
results derived here.

     Considering the  results  from  these  studies,   it  appears
that  the  prediction  equation  derived in this  report generally
provides  reasonably  accurate estimates  of  refueling emissions
based upon the  given  parameters  within  its applicable parameter
ranges.   Therefore, there  should be no  problem in using  it to
determine  average  emission factors  and  to determine  control
system designs  and efficiencies.

IV.  Calculation of Nationwide Emission Factors

     A.     Introduction

     Analysis of the baseline test data  has  yielded an equation
that  can  -be used  to  calculate emission  factors  representative
of   various   AT,    TD,   and   RVP   conditions   within   the
approximate  limits   of  the  values  for   the  original  test
parameters  (see section   III  C  4).   Given  this  ability  to
determine  emission  rates  for  different  conditions,   it  then
becomes   necessary   to   determine   the  most   representative
conditions  in  order   to  calculate  a  refueling emission factor
that  will  accurately  reflect   national   uncontrolled  in-use
emissions  levels.    Because  the  conditions   that   determine

-------
                               COMPARISON of EPA RESULTS to OTHER STUDIES
I
en
co
I
       •r,
      i/,-
      o
              .n
            6.0 -
            5.0 -
            4:0 -
            3,0 -
                                     RVP = 8.5 psi DISP.TEMP. = 79  F
            2
       .0 -
            1 .0 -
            0.0
       -40.00
                             -20.00
-o— EPA84     	+---SRI
   0.00           20.00

AT (DEGREES F)
                                      • — o— CAPE 9
                -A	ER&E
   40.00


	x— UNION
                                             FIGURE 27

-------
                          COMPARISON of EPA RESULTS to SCOTT STUDY
ta
w
(/)
o
      7.0
      6.0 -1
      5.0 H
      4.0 -I
3.0 -I
      2.0
      1.0-1
                               RVP = 8,8  psi DISP,TEMP=85 F
      0.0
         30.00
                         -10.00
10.00
                    EPA84
                               AT  (DEGREES F) ilJirhlTAI
                           .^4---SCOTT ENVIRONMENTAL
                                                                                          o
                                                                                          I
30.00
                                           FIGURE 28

-------
                              -41-
emission  rates  (basically  temperature  and volatility  of  the
fuel)  vary from  region  to  region  and  from  season  to  season
within  a  given  region,  it  will  be  necessary  to  identify
regional and seasonal temperatures  and fuel characteristics and
then to  apply  the appropriate  averaging to determine national
emission  factors.   It  is also important  to examine seasonal
emission factors to ensure that summer  and  winter  emissions are
not significantly different from the annual average value.

     There are  two basic uses  of  a refueling emission  factor:
(1)  to calculate  air  quality effects  and   (2)  to  determine
health risk due to exposure to the pollutant  in question.   The
air quality effect of VOC from refueling emissions  consists  of
the  role  these  emissions  play  in  ozone  formation.    Ozone
formation  tends  to be  a seasonal phenomenon,  with most  NAAQS
violations occurring  during  the spring  and summer months,  i.e,
May through September.  The emission factor  used in air  quality
calculations   should   therefore   appropriately  reflect   the
conditions that are found during the ozone season.

     In addition  to  their role as  ozone precursors,  refueling
emissions may also have environmental  health effects.   Benzene,
a  known  human  carcinogen,  is  present  in  small  amounts  in
gasoline.  In addition,  recent  studies  have indicated that other
species  of VOC contained in  refueling emissions  are possible
carcinogens.[15]   Although  the effect on  humans  is  not  fully
known, refueling  emissions  may pose  a  health risk to  service
station employees, self-service gasoline customers,  and persons
residing  near   service  stations.    Because   such  exposure  risk
represents  a  year-round  problem,  the  emission  factor used  in
determining  health "  risk   should   represent   average   annual
conditions,   although   if   there   are  significant   seasonal
variations the  additional risk posed by  these variations  would
have to be evaluated.

     The  remainder  of this report will  begin  with a discussion
of the total emission factor,  which includes both spillage and
displacement  losses.   An appropriate  spillage  emission  factor
will be selected.  The  process for developing  the displacement
emission factor will  then be described in detail, including the
methodology  used,  sources  of  data,   selection  of- seasonal
scenarios  for  air quality  and  environmental  health effects,
determination of  representative temperature and  fuel volatility
parameters  and  emission  factors  for  these  scenarios.    The
question  of  seasonal differences  in  the emission  factor  will
also be  addressed.  Finally,  a representative national emission
factor for refueling will be presented.

-------
                              -42-
     B.    Description of Refueling Emission Factors

     There are  two types  of  refueling  losses that  comprise a
total  refueling  emission factor.  These  are  spillage of liquid
gasoline  during  the  course  of  the  refueling  operation  and
displacement  losses,  or  the  vapor  that  is forced out  the
fillpipe  during  refueling.   Displacement  losses  occur  during
every  refueling  operation, while spillage or "spitback"  is a
more infrequent occurrence.

     1.    Spillage Losses

     A varying portion of the total  refueling  loss results from
the  spillage  of  liquid  gasoline during  the  refueling process.
The  amount  of such  spillage  can vary from  a  few  drops  on the
side of  the car  or pavement as the  fueling  nozzle is withdrawn
from the  fillpipe  to a cup or more  spurting out  on  the ground
as  a  result  of  "spitback"  due  to  poor  fillneck design  or  a
malfunctioning fuel  nozzle.   Probably the  majority  of  spills
are  less extreme,  coming about as  a  result of  motorists  or
service station attendants attempting  to "top off"  the vehicle
tank  by  restarting  the  nozzle after   automatic  shutoff  has
occurred.  Such spills are normally  not large, on  the order  of
a tablespoon  or so.   A spill  of one tablespoon leaves a 9 to 10
inch diameter  circular spot on the service station pavement and
results  in  emissions of  about  10 grams.   Thus  on a  10 gallon
fill,  the spillage  would equal  about one gram  per  gallon  of
fuel  dispensed.    Larger  spills  such  as  those  accompanying
spitbacks  or  nozzle malfunctions  can   lead  to  significantly
higher emissions.  A one-half cup spill  for the  same 10 gallon
fill  leads  to emissions  of   about  8  grams  per  gallon.   Thus,
overall,  spills are of concern.

     Of  course  not   every  fillup,  or  even  every attempt  at
"topping off"  results  in  a  fuel  spill and different  amounts  of
fuel  are  spilled  each  time.   Unfortunately  very  few data are
available  regarding  either  the  quantity   or   the  estimated
frequency of  fuel  spills,  and there  is considerable variance in
the existing estimates.  EPA's emission  factor document (AP-42)
presents  a  value  of about  0.30 grams  per gallon based  on a
comprehensive  study conducted by  Scott Research  Laboratories  in
the  early  1970's.[16]    However,  an  in-depth  review of  this
study  reveals  the  authors  belief   that   the-  spillage  rate
estimates  should  be  viewed  as  minimum values,  rather  than
averages, due to  the presence of observers,  the  technique used
to  estimate  spill  amounts,  and  the  fact that  the stations
studied were primarily full serve rather than self serve.[17]
However,   another   EPA  contractor report  cited  an estimate  of
1.36 grams per gallon, and a  ,1980 Calfornia study  conducted  by

-------
                              -43-
the  South . Coast  Air   Quality  Management  District  provided
information which  indicated an average  spillage rate  of  about
0.80  grams per  gallon  for  uncontrolled nozzles.[18,19].   It
should be noted that the latter study  included  the brief period
of  fuel  shortages  in  1979,  which  may  have  encouraged  an
abnormal  amount  of   "topping off"  of  fuel  tanks  and  hence
slightly  higher  than normal  spillage.   The  wide  variation  in
the  available  data on  spillage  rates (more than a  factor  of
four among  the three studies) is of some concern.   While  there
appears  to be  good  reason  for  the  variation,   the  data  is
inadequate  to   allow  determination  of   a   revised  emission
factor.   In the  absence  of  more definitive information  on this
topic,  the .30  grams/gallon  rate contained  in AP-42  seems  to
represent the best available  estimate  of the  spillage emission
factor,  so this value  will  be  used in  this analysis.    The
remainder  of  the  discussion  will   focus on  the  displacement
emission  factor,  but it   should be  noted  that   the  emission
factor for spillage must be added to  the displacement emission
factor in order to  arrive at a total refueling emission factor.

     2.    Displacement Losses

     As discussed  earlier,  there are  three primary factors and
several   secondary  factors   that  determine   the  displacement
emission  rate  for  refueling  operations.   The  primary factors
are  (1)  the  dispensed  temperature  (TD)  of  the  gasoline  (2)
the Reid  Vapor Pressure (RVP)  of the gasoline  and  (3) AT,  or
the difference between the  temperature of the residual gasoline
in the  vehicle tank  (TT) and the dispensed  temperature of the
gasoline  used  to  refill the  tank  (ie.  TT-TD).   To  develop
emission  factors for  refueling operations,  it will be necessary
to look at  these parameters  on a seasonal and  national basis.
The  most   significant  of the secondary  factors are  fuel tank
configuration  differences,   the  effects  of  which  have  been
described in sections  II and III  above,  and differences between
the  RVP of the  dispensed  fuel  and the residual  fuel in  the
tank, due to weathering of  the tank fuel.

     C.    Calculation of Displacement Emission Factors

     This section will derive nationwide average values for the
three  major  determinants,  AT, TD,  and  RVP,   from  which  the
uncontrolled   displacement    emission   factors   for   several
scenarios will then be calculated.

     l.    Methodology

     Overall,  the methodology used in  this process is to weight
the  available  regional  temperature and RVP  data by regional
highway fuel  consumption to  determine average  national values

-------
                              -44--
for  the  appropriate  time  periods  for   each  scenario  to  be
evaluated.   These average  values  can then  be  used with  the
multiple  linear   regression  equation  developed   earlier   to
calculate representative  emission factors.  The  available  data
indicate  that  there  is  a  considerable  amount of  regional  and
seasonal  variation  in the  the  temperature and RVP parameters,
making  such  a weighting  process  necessary.   Also,  the  fuel
consumption pattern  is  far  from  uniform  throughout  the U.  S.
and  the  available data  are  not  all  aggregated  at the  same
levels.   Fuel  consumption  and  RVP  data  are  available  on  a
monthly   state-by-state   basis  while  AT  and  T0   data   are
available only on a  monthly  regional  basis.   The methodology
used to  aggregate  and weight these parameters will be discussed
following  a  brief description of  the  sources  of  the  fuel
temperature, RVP and fuel consumption data that were used.

     2.    Sources of Data

     a.    Fuel Temperature

     Dispensed  temperature  and  AT  data  used  for calculating
emission factors are available  from a 1975 gasoline temperature
survey conducted  for  the American  Petroleum  Institute (API) by
the Radian  Corporation. [20]   The  year 1975 is considered to be
a  representative  year  in  terms  of  temperature,  since  the
average annual ambient temperature was within one degree of  the
30  year   mean.   The  study surveyed  56  U.S.  gasoline stations
located  in  22  cities; these were  grouped into  six geographic
regions.   The  six regions  and the locations for  the stations
surveyed  are shown  in  Figure  29,  and the  monthly  AT  and TD
values from the survey are shown in Appendix B.

     Not  all  of  the  stations  reported data  for  all  months of
the  year,  resulting  in  a  few gaps  in   the  data.   The  most
serious  of  these  gaps occurred in the Pacific Northwest (region
6  in  Figure  29)  where   AT data  were reported   only for  the
month  of May.  Since the  Pacific  Northwest accounts for  only
about  3.5 percent  of  the gasoline  consumed for  highway  use in
the  U.S., it  was concluded  that  this region could be omitted
from the analysis without  seriously affecting the  accuracy of
the  results.   Alaska  and  Hawaii  were  also  omitted  from  the
study,  since  no   AT  or  TD  data  were  available  from  these
states.   Other minor gaps  of  a  month  or  so  in  AT  and TD
data,  primarily  in the  North  Central U.S.  and  the  Far  West
(regions  4  &  5),  were  filled by points  interpolated from  the
existing data.

-------
                           Survey sample stations
                                                              cn
                                                              I
29

-------
                              -46-
     b.    Nationwide Fuel Consumption

     Nationwide fuel consumption  (gasoline)  by  state  was taken
from  the  1983 version  of  the  DOT/FHwA publication entitled
Highway  Statistics  -  Table  MF-26.    This   table   contains
estimates  of  monthly  gasoline  consumption  for  each  state  and
the District of Columbia.  Table MF-26 is shown in Appendix B.

     To allow  for  further  calculations,  the monthly  state fuel
consumption figures  were summed  for  each region thus providing
monthly regional fuel consumption values.

     c.    RVP

     The   RVP   data   were  taken   from   1983   ASTM  maximum
specifications  for  the  U.S.[21]   The  maximum  specifications
were used  rather  than current  actual levels on  the  assumption
that recent  increases  in  RVP would  continue and that  by 1989
in-use RVP  levels  would be  essentially  at  the  maximum values
specified by ASTM.   This  is  already  the case in some areas of
the country.   The  RVP  values  for  each state and  month are also
shown in Appendix B.

     To get  the RVP data  on the  same level of  aggregation as
the temperature data,  the  state RVP  data was divided into  the
same  regions  as  the  temperature data and  then  consumption
weighted to get weighted RVP for each region in each month.

     3.    Air Quality and Health Effects Scenarios

     To  facilitate  assessment  of  seasonal  variation  in  the
emission rates,  five  seasonal  air guality  and  health effects
scenarios  were established  and  AT,  TD,  and RVP values  were
calculated for  each.   The first  scenario is simply  the annual
average  value  for  the   nation   or   region.    Two  additional
six-month  scenarios  were  chosen  to  represent  warm  weather
versus cold weather  conditions.   "Winter"  is comprised  of  the
months October through  March,  while  "Summer"  consists  of  the
months April through September.   Two additional  scenarios were
chosen to  represent  the months in which most ozone  violations
occur.   These   include  a  "Five  Month"   scenario  (May  through
September), and  a  "Two  Month"  scenario  for the  two  peak ozone
violation months (July and August).


     4.    Consumption Weighting Calculation

     In order  to  calculate national  average AT,  TD,  and  RVP
values for the five scenarios mentioned  above, monthly regional

-------
                              -47-
data were consumption-weighted  by  the regional fuel consumption
values  for   the months  in  question.   As  explained  earlier,
monthly  state  RVP and  fuel  consumption values  were aggregated
on  the  same monthly regional  basis  as  the  TD  and AT  data.
The  generalized equation  for  calculating  consumption  weighted
values for each scenario is as follows:

           n
           I  (ATR,M)(FCR.M)
    AT=  R,M=1	:____^_
                   FCr o t a 1


     Where R = region number from Figure 29
     M =   month   number    (of   the   seasonal   scenario,   not
           necessarily of the calendar year)
     n =   number  of  months and number of  regions evaluated in
           a given scenario
     ATR.W =     temperature  differential  (AT)  for  region  R
                 during month M
     FCR,H   =   fuel consumption for region R during month M
     FCrotai =   total national  fuel  consumption   (less  region
                 6, Alaska and Hawaii).

The  key  parameter shown  in  the  equation  above  is AT.   The
consumption weighted  values  for  the  other  two key parameters
for  any given  scenario  can  be  determined by  substituting the
appropriate monthly  TD  and  RVP  values in  the  equation  above.
This can  be done  for  each of the scenarios  mentioned above to
get the appropriate values of the key  parameters for use  in the
refueling emission equation.

     The  results  of these  weighting calculations  are  shown in
Table  6.    Regional  and  national  average  AT,   TD  and  RVP
information  is  presented  for five  scenarios:   annual  average,
summer, winter,  five  month  ozone   season  and  two  month  peak
ozone season.  Regional fuel consumption  values used in the the
weighting   calculations   and   percentages   of   total   fuel
consumption  for  each  region are   also  shown  for  comparison
purposes.   As explained above, Region  6  (Pacific Northwest) has
been omitted, as have  Alaska and Hawaii.   As would be expected,
TD,  RVP and -AT  values vary both  seasonally  and  from  region
to region  for  any given season.   Reasons for  this variability
are discussed below.

     As can be seen from the table,  dispensed fuel temperatures
vary seasonally and  from region  to  region.   This is due largely
to climatic factors  such as ambient  temperature and the  amount
of  solar  radiation.   Other  relevant  variables   include  the
volume  and depth  of  the   underground service  station  tanks,
layout  of  the  fuel piping, composition of the  surface over the
tanks and associated piping  (e.g. concrete,  asphalt, grass) and

-------
                         -48-
                Table 6




Weighted Temperature and  RVP Parameters
REGION:
SCENARIO:
Average Annual
Fuel Consumpt.
(gal x 106)
% Total
RVP (PSI)
A T (°F)
TD (°F)
Summer (Apr-Oct)
Fuel Consumpt.
(gal x 106)
% Total
RVP (PSI)
AT (°F)
TD (°F)
Winter (Oct-Mar)
Fuel Consumpt.
(gal x 106)
% Total
RVP (PSI)
A T (°F)
TD (°F)
Nat'l Avg.
96,050.4
100.0
12.6
+4.4
68.9
51,846.3
100.0
11.5
+8.8
76.2
44,204.4
100.0
13.9
-0.8
60.3
1
N.East
41,658.5
43.4
13.3
+5.7
62.3
22,815.1
44.0
12.2
+10.7
70.7
18,843.5
42.6
14.6
-0.3
52.0
2
S.East
20,381.2
21.2
12.4
+4.0
81.8
10,689.0
20.6
11.4
+6.8
86.7
9,692.1
21.9
13.4
+0.9
76.4
3
S.West
11,977.6
12.5
11.4
+3.7
70.5
6,232.4
12.0
10.1
+7.6
78.6
5,745.3
13.0
12.8
-0.4
61.8
4
N.Cent.
10,225.6
10.6
12.6
+5.5
66.2
5,690.2
11.0
11.2
+11.7
74.3
4,535.5
10.3
14.3
-2.4
56:1
5
Far W.
11,807.6
12.3
11.7
+0.1
70.5
6,419.7
12.4
10.5
+3.9
77.2
5,388.0
12.2
13.3
-4.4
62.4

-------
-49-
Scenario
Ozone - 5 Mo.
Fuel Consumpt.
(gal x 106)
% Total
RVP (PSI)
A T (°P)
TD(°F)
Ozone - 2 Mo.
Fuel Consumpt.
(gal x 106)
% Total
RVP (PSI)
A T (°F)
TD (°F)
NatT Avg.
(May-Sept)
43,995.8
100.0
11.3
49.4
78.8
(Jul-Aug)
18,664.7
100.0
10.9
49.9
82.7
1
N.fest
19,459.4
44.2
12.0
4-11.5
. 73.8
8,326.2
44.6
11.5
4-12.5
78.0
2
S.East
8,956.0
20.4
11.2
4-7.5
88.0
3,760.0
20.1
10.9
4Q.2
90.5
3
S.West
5,244.4
11.9
9.9
4-7.1
80.8
2,147.7
11.5
9.8
4-7.0
83.5
4
N.Cent.
4,869.8
11,1
10.9
4-12.1
79.0
2,103.0
11.3
10.5
4-13.3
86.5
5
Far W.
5,466.2
12.4
10.3
+5.1
79.0
2,327.8
12.5
10.0
4-3.2
83.0

-------
                              -50-
protection .from  solar radiation for  the  tank system.  Although
there  is  a  strong  correlation  between  ambient  temperature
(TA)  and dispensed temperature, variation exists due  to these
other factors.   In  general,  the Radian study for API  shows the
average  dispensed  temperature  parallels  the  average  ambient
temperature curve,  with a positive  offset (i.e., TD  is always
higher than  TA).   The  amount  of  the offset  varies seasonally
and regionally, undoubtedly due to climatic differences.

     The  RVP  values  shown  in  the  table  are  ASTM  maximum
recommended values.   Figure 30  shows the average  seasonal and
regional variation  in these ASTM  RVP values and  the resulting
national averages  for winter and  summer.  The  regions  by which
these data are  aggregated are those  shown in Figure  29.   ASTM
assigns  each  state a "volatility  class"   and  specifies maximum
recommended  monthly  RVP  limits  based  on  the  climatic  and
topographic   factors.    The   five   volatility   classes   are
designated A,  B, C,  D,  and E,  corresponding  to  maximum  RVP
limits of  9,  10,  11.5,  13.5,  and  15  psi,  respectively.   In
addition, a number  of  states  have formally adopted RVP limits
similar  to  the  ASTM   recommended   levels.[22]   Particularly
noteworthy is California, where  RVP  is limited to 9 psi during
the  months  of  the highest  ozone concentrations  in  order  to
decrease VOC emissions.

     In-use  RVP  is  essentially  determined  by  the  gasoline
refiners, subject  to state  laws and voluntary  compliance with
the ASTM recommended  limits.   RVP  varies  seasonally  as  well  as
regionally,  based  primarily on  how  the climate  and topography
of  an area  affect  vehicle  operation.    For  example,  RVP  is
higher in the winter to  assist in cold  starting but decreases
in summer to  avoid  vehicle  driveability problems such  as vapor
lock.  RVP  values  are  generally  higher  for the  northeastern
U.S.  than  the  southeastern  U.S.   In  general,   for any  given
season or  area, RVP  is higher  as ambient temperature for any
given month  decreases.    The overall  trend  in  in-use  RVP  has
been  toward  higher and higher  values  in recent years,  due  to
changes  in  vehicle  design  and gasoline  refining  practices,
leading to the conclusion that  by 1990 the  in-use values will
approximate the ASTM maximum limits.

     The  final  factor,  AT,   also  varies   seasonally,   with
positive  values  being  more  predominant  in  the  summer  and
negative values more  prevalent  in  winter.   Although  there  is  a
certain  amount  of  regional  variation,  the seasonal values are
very similar for all areas of the country.

     There is  also  a certain  amount  of diurnal  variation that
affects  AT   values,   which  explains  the   presence  of  some

-------
                SEASONAL  RVP  VALUES
                     ASTM MAXIMUM SPECIFICATIONS
Q.
I
CL
          AVG ANN



          I R1   E33 R2


          I NAT. AVG,
    SUMMER


SEASONAL SCENARIOS

     R3    SS
                          Figure 3 o

-------
                              -52-
negative  values  in  the  summer  and  positive  values  in  the
winter. ,  TD  is  more stable  than  TT  due  to the  insulating
effect of the ground in which the service  station  storage tanks
are  buried.   Since  vehicle  tank  temperatures follow  ambient
temperatures  more closely,   the  likelihood  is  strong that  TT
will  be  lower  for  those  vehicles  fueled  in  the  morning,
resulting   in  negative,   or  at   least  less  positive,    AT
values.   Conversely,   diurnal  heating  would  likely  result  in
higher  TT  values  in  the  afternoon,  resulting  in  positive
ATs.   The  distance  a  car  is  driven  before  refueling  also
affects  TT   (TT  increases  with  distance   driven,   up   to  a
point)  which  in  turn  affects    AT.    Since  these  diurnal
effects  are  recurring  and  ongoing,   one   would  expect  the
differences  between  summer and  winter  ATs  to  be  caused  by
climatic and not diurnal variations.

     5.    Emission Rates

     Given  the  weighted  regional   and  national  average  TD,
AT  and RVP  values  in Table  6 for  each of  the  five scenarios
under  consideration,   we  are  now  prepared  to  calculate  the
emission rates  for  each  of  the scenarios and to  assess  how the
variation in  the  key  parameters  affects emission  rates.   These
emission rates  for  the different  scenarios are calculated quite
simply by  substituting the TD,  AT   and  RVP  values  of Table  6
into  the  multiple  linear  regression  relationship  developed
earlier (given below)  and solving for the emission rate (ER).

         ER = -5.909 - 0.0949AT + 0.0884 TD + 0.485 RVP

The  results  of  these  calculations  for   each   of   the  five
scenarios are shown in Table  7.   This emission rate data can be
compared  regionally  within  each   scenario  and   between  the
various  scenarios  on a  seasonal  basis  for  each  region  and
nationally.

     Turning  first  to  the  regional  evaluation  within  each
scenario,  several  points  should be  noted.   First,  overall,  the
regional values for each  scenario  are relatively  uniform given
the  variation  seen  in  the key  parameters  of Table 6.   All
values  fall within + 10 percent  of  the  national average  for.
that  scenario,   with  the  exception  of  the  southeastern  U.S
(Region 2).-   In each  of  the  five  scenarios the  emission rate
expected in  the Southeast exceeds the national average for that
scenario by between  16  and 19  percent.  The higher  emission
rates  in the  southeastern US  apparently occur because of the 10
to  16F°  higher  dispensed  temperatures  encountered  there,  as
compared  to  the  national  average.   Simply by  using  the
coefficient  for  TD in the emission rate equation,  it  can  be

-------
                                      -53-
                                      Table 7
                          Displacement Emission Factors I/
                                               REGION:
SCENARIO;

Average Annual

Summer (Apr-Sep)

Winter (Oct-Mar)
Nat'l Avg.
5.9
5.6
6.2
iep) 5.6
.ug) 5.7
1
N.East
5.5
5.2
5.8
5.3
5.4
2
S.East
7.0
6.6
7.2
6.6
6.6
3
S.West
5.5
5.2
5.8
5.4
5.6
4
N.Cent.
5.5
5.0
6.2
5.2
5.6
5
Far W.
6.0
5.6
6.5
5.6
6.0
-    Displacement losses only  -  a spillage factor must be  added to derive a total
     refueling emission factor.

-------
                              -54-


determined that  the 10 to  16F°  difference in  TD  results in an
increase, of  0.9  to 1.4 g/gal  in the  emission  for  the  various
scenarios.    This  easily accounts  for the  significantly higher
emission rate in the this  region.

     Climatological   differences   offer   the   most    likely
explanation  for  the  higher  dispensed  temperatures   in  the
Southeast.    As   explained   earlier,   the   average  TD   value
generally  follows  the trend  of  the   annual  average  ambient
temperature curve for  any  given region,  but  there  is always  a
positive  offset  (i.e.,   the  TD  value  is  greater  than  the
average ambient  temperature),  probably  because  of solar  heat
gain  and  the thermal  storage  effect  of the  ground,  which  in
turn are modified  by  the  other  factors  noted  above in  section
C-4.  The  magnitude of the  offset  varies during  the course of
the year for most regions,  particularly  where the  ground may be
frozen  during  the  winter  months.    In  such  areas   TD  may
approach   the  ambient temperature  in  the  Spring, when  the
ambient  temperature  rises   relatively  guickly while  the  soil
temperature  increases  more  gradually.    The  offset   for  the
Southeast,  on the  other hand,  is relatively constant throughout
the year,  likely because over most of the area  the ground never
freezes and  because of greater  solar  gains and  higher  ambient
temperatures, particularly during the Winter months.

     Second,  comparing   the  seasonal   (summer   and   winter)
emission rates to  the  average  annual  rates for each region, all
the emission rates are within +10 percent of  the  average annual
value  except for the  North  Central  U.S.   In  this  case  the
seasonal variation  is on  the  order  of  12-13  percent,   due  to
slightly greater  seasonal  variation  in  the absolute values  of
the key parameters.   This  relatively  small  seasonal  variation
in  the emission factors  is  likely  due  to the  existence  of
offsetting factors in the conditions that determine both Winter
and  Summer  emission  rates.    In the Winter  months, RVP's  are
high and AT values tend  to be  more negative  than during  the
Summer.  Both of  these trends  would tend to increase emissions,
but they are offset by lower dispensed temperatures, which tend
to decrease  emissions.  Conversely, during  the  Summer  dispensed
temperatures  are  higher,  which  would  increase emissions,  but
the higher  temperatures  are offset by lower  RVP's and  positive
AT values,  both of which tend to decrease emissions.

     Third,  comparing  the  two  ozone scenarios to  the  average
annual  scenario,  the  emission  rates for  all  five regions  and
the national  average do not  vary by more than 10  percent.   In
this case  the average annual values for  each region exceed the
ozone scenario values  in a  range of 0 to about 6  percent.   So
overall  there  is  good  agreement  between  the average  annual

-------
                              -55-
emission  rates  and the  emission  rates  expected  in the  ozone
prone months.   This  is true  for all regions  and on a national
level.

     6.     Effects of Fuel Weathering

     In addition to the three primary factors  discussed above,
fuel weathering also  affects  refueling  emission  rates.   Fuel
tank weathering results  in  a  difference  in  RVP between  the
dispensed fuel  and the residual fuel in the  vehicle tank,  with
fuel in  the tank  losing  volatility due to  the  evaporation  of
lighter  ends  in  the  gasoline.   The  very  limited amount  of
baseline  testing  that  was done with  lower  RVP  fuel  in  the
vehicle  tank indicates that  an increase  in emissions, on  the
order of  a  gram per  gallon,  resulted from an RVP difference  of
approximately 1.9  psi  between the  tank  and the  dispensed  fuel
(see  Section   III.C.).    This  general  phenomenon  was   also
observed  in the SRI study.[2]  Unfortunately, neither  the  EPA
nor the SRI  data are  adequate to fully characterize  the effect
of the  RVP  difference,  although they do show  the direction and
give a rough idea of the magnitude of the change.

     In order  to be able to  include  the effect of tank  fuel
weathering  in   the emission  rate  calculation,  one  would  also
need to  know  the  average amount  of  in-use  weathering  that
occurs  between  refuelings,  in  addition  to the  effect of  the
resultant difference  in  RVP  values on  the  refueling  emission
rate.   This  includes  both  the  different  vehicle  and  fuel
effects.  Since neither  of  these variables  can  be  determined
with any  certainty at  this time, the effect  of tank weathering
has  not  been   included   in  the  emission  factor  calculation.
Although  refueling emissions  may thus be  somewhat understated,
this effect may be  partially  offset  by the  method  of  RVP
determination for  the calculation.   Use  of the ASTM maximum RVP
limits  represents  EPA's  best judgment  of  future RVP  levels.
However,  if  in-use  RVP  levels  should  be  lower  than  these
maximum values,  the  resulting decrease  in the emission factor
would tend  to  be  at  least partially offset  by  an increase  in
refueling emissions due to fuel weathering.

     D.     Conclusions

     At the beginning of this   investigation it  was felt  that  it
might be  necessary to develop both a seasonal emission factor
for  air  quality  calculations and  -an average annual  emission
factor  for   health exposure  risk  calculations.    However,  the
relative uniformity of  the seasonal emission  factors indicates
that the  average  annual  values can be  used for  both  purposes
without introducing any significant  error  into the air  quality

-------
                              -56-

                                 .1
calculations.  If only  air  quality calculations  were involved,
it might  be more  appropriate  to  use  only  a summer  emission
factor, although by  so  doing two important "by-products" of the
air  quality calculation,  the  emissions   inventory  calculation
and  the calculated  lifetime emissions  reduction per  vehicle,
would  both  be understated.   An annual  average  would  be  more
appropriate for these latter  two purposes as well as for health
exposure risk calculations.

     On the other hand,  use of  an average annual  emission  rate
for  air guality  determinations  may  theoretically overstate the
air  quality benefits   somewhat.   The  difference  between  the
summer  and  average  annual  emission  factors  is relatively small
(less  than  5  percent),  however,  and  any  differences   in  air
quality  calculations,   i.e.  SMSAs  brought  into  compliance  or
percent change in  air  quality,  would likely  disappear  in  the
roundoff of  the  EKMA model.  Thus in practical terms,  it would
likely  be  very difficult to see  any  differences  in  the  air
quality outputs  resulting  from the use  of the  average  annual
values, whereas there are real   advantages  to  its use  in terms
of   emissions   inventory,    lifetime   emissions   and   cost-
effectiveness calculations.

     Finally, it does not appear that there will be  a  need for
seasonal  emission  factors   for health   effects  purpose as  a
result  of   changes   in  the  amount   of  benzene   and   other
potentially hazardous  species  in  the  total VOC  emitted.  Such
emissions are  a  function of the  percentage  of  the  hazardous
pollutant present  in liquid  gasoline  and the same  temperature
considerations that  affect   the basic  VOC  emission rate.   In
order  to  have significant  seasonal  variation in  the emissions
of   these   hazardous   species,   then,    either   the   seasonal
percentage of  these  species  in  the  liquid gasoline  would  have
to vary significantly  or,  since  such emissions are  normally
expressed as  a percentage  of  total VOC emissions,  winter and
summer VOC emission  factors would  have to  differ significantly
from the annual average.  As  stated above, the latter condition
is not  the  case.   Correspondingly,  the  1983/84  NIPER  gasoline
surveys show  no  significant  difference  between the  winter and
summer  benzene  percentages  in  liquid  gasoline.   For  these
surveys  average  summer  and  winter  benzene  fractions   in  the
liquid  gasoline  averaged about  1.3 percent.[23,24]   Similar
data are not available  for  other potentially hazardous species,
but there is no  reason  to  believe  that  the liquid  fraction  of
these species varies regionally or seasonally.  Therefore it is
reasonable  to  conclude  that   seasonal   differences   in   the
emission factors  will not necessitate  separate emission factors
for either health effects or air quality purposes.

-------
                              -57-
     For  these  reasons,  it  was  decided  to  use  the  average
annual  displacement  value   of   5.9  grams  per  gallon for  all
calculations.  Adding 0.3 grams  per gallon  for spillage results
in  a  national  average  refueling  emission  factor of  6.2  grams
per gallon.

-------
     Appendix A





BASELINE TEST RESULTS

-------
   1983 Oldsmobile Cutlass  Supreme




Dispensed Temperature = 82°F  3 RVPs

Test
845638
845639
845637
845632
845636
845631
845628
845630
850113
850114
850115
850117
851354
851355
845950
845951
845945
845947
850057
850104
845943
845944
845946
850105
850106
845941
845942
845948
845642
845641
845627
845294
845625

RVP
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
12.6
12.6
12.6
12.6
12.6

T(8F)
1.5
0.
10.8
8.1
22.0
18.5
36.7
37.0
12.0
7.7
12.3
11.9
2.3
1.5
-1.1
-.8
11.0
10.0
10.0
11.1
17.0
21.7
20.5
17.2
18.0
34.0
39.5
38.8
-2.0
7.5
8.2
16.6
19.0

Disp.
Temp('F)
80.5
82.0
82.0
83.9
81.0
83.5
82.3
82.0
80.5
83.3
80.0
79.8
81.7
81.5
81.0
83.0
81.0
83.2
82.0
80.8
85.0
80.8
83.0
83.8
81.8
86.0
82.0
81.2
82.5
84.0
84.0
• 84.0
81.0'

Losses
(gm/gal)
5.209
5.456
5.021
5.308
4.407
4.632
3.115
2.918
4.934
4.934
4.331
5.126
5.060
5.133
7.831
8.074
6.490
6.133
6.395
5.947
4.842
4.432
5.743
5.327
5.007
3.830
3.500
3.514
8.938
7.366
6.892
6.290
6.081

Liq.
TernpCF)
82.0
82.0
92.8
92.0
103.0
102.0
119.0
119.0
92.5
91.0
92.3
91.7
84.0
83.0
79.9
82.2
92.0
93.2
92.0
91.9
102.0
102.5
103.5
101.0
99.8
120.0
121.5
120.0
80.5
91.5
92.2
100.6
100.0

Vap.
Temp(°F)
83.0
83.0
93.9
92.5
102.0
100.5
116.0
117.0
92.3
93.0
93.4
90.5
82.5
82.5
82.5
83.0
92.3
92.5
93.0
93.0
102.0
102.0
103.0
102.5
98.5
118.3
118.0
118.5
83.0
92.5
92.5
102.0
101.0


Disp. Losses
Gals (gins)
14.8
14.7
14.6
14.6
15.0
15.5
14.8
14.6
15.1
15.1
15.1
15.1
15.0
15.0
14.8
14.8
14.7
15.0
15.2
15.0
14.6
14.8
14.8
15.0
14.9
15.3
14.6
14.8
14.5
14.5
14.8
14.5
14.9
77.1
80.2
73.3
77.5
66.1
71.8
46.1
42.6
74.5
74.5
65.4
77.4
75.9
77.0
115.9
119.5
95.4
92.0
97.2
89.2
70.7
65.6
85.0
79.9
74.6
58.6
51.1
52.0
129.6
106.8
102.0
91.2
90.6
Disp.
Time
(min.)
2.52
2.53
2.58
2.67
2.63
2.70
3.00
2.58
1.97
1.98
1.98
2.02
1.98
1.97
3.30
3.42
3.25
3.33
1.98
2.00
3.35
3.13
3.33
2.05
1.97
3.50
3.27
3.23
3.62
3.58
3.83
3.68
3.38
Heat
Time
(min.)
28.00
24.00
52.00
50.00
36.00
37.00
74.00
67.00
42.00
32.00
30.00
48.00
28.00
32.00
24.00
24.00
60.00
46.00
46.00
45.00
35.00
40.00
37.00
39.00
39.00
68.00
61.00
72.00
25.00
48.00
60.00
39.00
41.00
  Dispensed Temperature = 92°F  3 RVPs
845289
845290
845280
845292
850110
850111
850112
851356
851357
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
-11.0
8.3
29.0
27.8
1.5 .
1.5
2.2
15.2
8.4
91.0
92.0
91.0
91.2
90.5
90.5
90.0
88.0
92.0
6.952
6.324
4.257
4.097
6.470
6.831
5.887
4.700
6.060
80.0
100.3
120.0
119.0
92.0
92.0
92.2
103.2
100.4
81.8
101.0
119.0
116.0
92.5
93.5
94.2
100.0
97.0
14.5
14.2
14.8
14.5
14.9
15.4
15.1
15.0
15.1
100.8
89.8
63.0
59.4
96.4
105.2
88.9
70.5
91.5
2.88
2.95
2.78
2.82
2.05
2.07
2.00
2.00
2.05
26.00
38.00
67.00
67.00
34.00
34.00
36.00
44.00
32.00

-------
845931 11.9 -1°-5
845932 11.9 -10.3
845935 11.9 2.0
845937 11.9 1-0
850054 11.9 2.5
850053 11.9 -1.0
845938 11.9 9.5
845939 11.9 10.1
845933 11.9 22.1
845936 11.9 30.6
850055 11.9 25.9
850056 11.9 28.8
845956
845957
845954
845955
845953

845107
845276
845102
845103
845109
845105
845106
845275

845286
845287
845961
845959
10.0
10.0
10.0
10.0
10.0

9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0

9.0
9.0
10.0
10.0
-11.0
-10.0
1.8
2.0
11.8

-10.8
-11.0
2.0
-1.0
-3.2
7.1
8.0
27.5

10.5
-3.5
-6.0
-3.9
92.0
91.3
90.0
91.0
90.5
93.0
92.0
91.8
92.0
90.9
93.0
91.2
91.0
90.0
90.2
90.0
90.8

90.8
90.0
90.0
92.7
93.2
91.6
91.5
92.5

92.5
92.0
92.0
91.6
10.250 81.5 81.5
11.431 81.0 83.0
8.307 92.0 91.5
7.270 92.0 93.0
10.060 93.0 90.8
9.765 92.0 92.0
8.066 101.5 102.5
7.815 101.9 103.0
4.921 114.1 109.9
3.311 121.5 119.0
4.066 118.9 117.8
4.821 120.0 117.2
8.597
8.128
6.966
7.240
6.479
Single
6.353
6.651
5.980
6.153
6.554
5.027
5.060
3.453
Road
4.815
6.060
7.985
6.514
Dispensed Fuel IOTP =
850904
850913

850885
850887
850901
850902
11.9
11.9

11.9
11.9
11.9
11.9
0.0
3.1

3.2
3.0
4.7
4.0
92.0
90.8
Large
90.8
91.0
89.0
89.0
9.566
8.731
80.0
80.0
92.0
92.0
102.6
82.3
83.0
93.0
92.3
102.0
15.2 155.8
15.3 174.9
15.0 124.6
15.2 110.5
14.9 149.9
15.3 149.4
15.2 122.6
15.1 118.0
15.1 74.3
14.8 49.0
15.2 61.8
15.1 72.8
14.9
14.9
14.8
15.0
14.6
128.1
121.1
103.1
108.6
94.6
4.52
4.53
3.98
3.77
3.08
3.28
4.22
4.22
4.05
3.60
2.55
2.32
3.48
3.43
3.33
3.42
3.33
32.00
24.00
46.00
53.00
40.00
42.00
40.00
39.00
67.00
65.00
49.00
58.00
21.00
24.00
48.00
48.00
35.00
Blanket Data
80.0
79.0
92.0
91.7
90.0
98.7
99.5
120.0
Prep Data
103.0
88.5
86.0
87.7
11.9 TanK
92.0
93.9
78.0
76.0
86.5
86.0
84.0
95.6
96.4
114.0

105.9
90.0
87.5
88.7
Fuel KVP
88.0
90.3
15.0
14.6
15.0
15.0
14.8
15.0
14.9
15.0

14.6
13.4
13.5
14.2
= 10.0
14.5
14.5
95.3
97.1
89.7
92.3
97.0
75.4
75.4
51.8

70.3
81.2
107.8
92.5

138.7
126.6.
2.85
2.83
2.93
2.97
2.93
2.93
2.80
2.87

2.60
3.15
3.22

2.27
2.25
36.00
42.00
96.00
178.00
74.00
212.00
173.00
306.00

182.00
172.00
185.00
159.00

16.00
19.00
Vapor -Liquid Temperature Differences
8.220
8.140
8.947
9.093
94.0
94.0
93.7
93.0
86.8
87.0
104.0
102.5
15.0
15.0
15.0
15.0
123.3
122.1
134.2
136.4
2.07
2.10
2.10
2.07
16.00
16.00
36.00
34.00

-------
                                  1984 Ford Escort
                        Dispensed Temperature =  80°F 2 RVPs
851161
851162
851163
851165
851164
851166
851213
851214
850304
850012
850014
850311
850314
850312
850313
850013
850303
850011
850310
850309
850307
850308
851160
851167
851168
851169
851211
846446
846447
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
11.9
9.0
9.0
9.0
9.0
11.9
11.9
3.7
4.0
9.3
12.5
21.2
22.5
10.5
20.5
1.5
12.0
12.7
2.7
11.8
11.1
13.4
3.7
3.7
13.3
4.5
12.0
22.0
23.0
20.2
8.9
4.5
4.8
6.0
2.0
.7
80.0
79.0
80.7
81.0
80.3
78.5
81.0
80.5
4.456
4.308
4.269
4.574
4.221
3.159
4.673
4.346
83.7
83.0
90.0
93.5
101.5
101.0
91.5
101.0
81.2 6.091 82.7
80.0 5.750 92.0
80.3 5.606 93.0
80.0 5.595 82.7
81.2 5.694 93.0
81.0 5.654 92.1
79.1 5.455 92.5
80.0 6.255 83.7
80.0 5.991 83.7
79.6 5.890 92.9
78.5 6.279 83.0
80.5 6.077 92.5
80.5 5.390 102.5
80.0 5.202 103.0
80.0 5.567 100.2
Dispensed Temperature
85.0
89.5
89.7
89.0
92.5
92.3
4.740
5.221
5.115
4.875
7.903
7.228
93.9
94.0
94.5
95.0
94.5
93.0
83.0
83.0
91.9
92.5
101.8
102.5
91.7
101.0
84.0
94.0
94.0
82.3
92.7
92.2
91.8
83.9
84.0
93.0
83.7
95.0
103.0
105.0
102.0
= 908F 2 RVPs
94.5
93.2
91.0
96.0
98.2
98.0
10.3
10.4
10.4
10.8
10.4
10.7
10.4
10.4
11.0
10.8
10.9
11.1
11.1
10.7
11.2
11.0
11.6
10.9
10.4
10.4
10.5
10.4
10.4
10.4
10.4
10.4
10.4
11.3
11.4
45.9
44.8
44.4
49.4
43.9
33.8
48.6
45.2
67.0
62.1
61.1
62.1
63.2
60.5
61.1
68.8
69.5
64.2
65.3
63.2
56.6
54.1
57.9
49.3
54.3
53.2
50.7
89.3
82.4
4.46
1.37
1.37
1.43
1.43
1.42
1.37
1.35
1.47
1.42
1.43
1.45
1.45
1.40
1.47
1.43
1.45
1.45
1.38
1.43
1.43
1.40
1.38
1.40
1.42
1.38
1.38
1.63
1.63
17.00
16.00
28.00
32.00
40.00
44.00
28.00
42.00
25.00
57.00
56.00
32.00
50.00
52.00
58.00
24.00
25.00
50.00
28.00
45.00
40.00
38.00
42.00
32.00
30.00
32.00
30.00
26.00
21.00
                             Dispensed Temperature = 66°F
850305
850306
11.9
11.9
7.5
5.0
66.5
'67.5
5.029
5.286
74.0
72.5
73.9
72.0
10.5
10.5
52.8
55.5
1.35
1.37
6.00
5.00

-------
  1983  Plymouth Reliant
851151
851150
851152
851153
851154
851155
845486
845487
845492
845493
851157
851156
851159
851246
11.9
11.9
11.9
11.9
11.9
11.9
9.0
9.0
11.9
11.9
9.0
9.0
11.9
11.9
2.S
1.7
10.0
19.3
14.1
18.0
1.5
-1.5
3.0
1.3
4.3
4.0
5.5
5.6
80.0
81.8
81.0
80.0
79.9
81.0
90.5
91.5
89.0
91.0
68.0
69.0
68.0
68.3
6.265
6.640
6.439
4.969
5.316
6.109
5.577
5.677
6.839
7.036
3.763
3.939
5.299
5.323
82.5
83.5
91.0
99.3
94.0
99.0
92.0
90.0
92.0
92.3
72.3
73.0
73.5
73.9
84.3
84.0
89.0
97.5
94.0
99.7
86.5
87.0
88.3
89.5
73.2
72.0
77.5
78.0
9.8
10.0
9.8
9.6
9.5
10.1
10.4
9.9
11.2
11.2
9.7
9.8
9.7
9.9
61.4
66 A
63.1
47.7
50.5
61.7
58.0
56.2
76.6
78.8
36.5
38.6
51.4
52.7
1.32
1.32
1.30
1.28
1.48
1.37
1.92
1.87
3.12
3.17
1.30
1.32
1.28
1.30
16.00
16.00
32.00
44.00
32.00
38.00
38.00
38.00
42.00
43.00
6.00
4.00
12.00
14.00
    1983 Buicfc SKylarfc
846410
846413
846454
846453
846412
846452
11.9
11.9
11.9
11.9
11.9
11.9
2.7
13.0
-.9
-2.0
.5
.5
90.5
85.0
94.3
93.5
92.0
92.0
7.925
6.471
7.540
7.353
7.365
7.708
93.2
98.0
93.4
91.5
92.5
92.5
92.2
91.0
91.9
90.0.
90.6
94.3
13.4
13.6
13.7
13.6
13.7
12.0
106.2
88.0
103.3
100.0
100.9
92.5
3.75
2.32
2.15
2.05
2.05
1.80
28.00
28.00
26.00
23.00
20.00
19.00
 1984 Chevrolet Celebrity
850205
850209
850208
850207
850206
850204
11.9
11.9
11.9
11.9
11.9
11.9
12.7
10.7
13.2
.7
-.5
2.6
80.0
81.0
78.8
91.8
92.5
89.4
4.632
4.596
4.706
6.333
6.876
6.080
92.7
91.7
92.0
92.5
92.0
92.0
92.2
92.0
92.1
96.0
94.5
94.0
13.6
13.6
13.6
13.8
13.7
13.8
63.0
62.5
64.0
87.4
94.2
83.9
1.78
1.78
1.85
1.95
1.95
1.93
44.00
38.00
40.00
58.00
38.00
36.00
   1983 LOT Crown Victoria
850400
850401
850402
850404
850405
11.9
11.9
11.9
11.9
11.9
11.2
11.8
11.0
1.0
1.8 .
80.8
79.2
80.5
90.8
90.0
7.448
7.200
7.558
11.166
10.500
92.0
91.0
91.5
91.8
91.8
95.0
94.0
94.2
95.4
94.2
15.4
14.5
15.4
15.7
15.6
114.7
104.4
116.4
175.3
163.8
2.02
1.90
2.05
2.30
2.25
58.00
51.00
42.00
34.00
36.00
1979 Chevrolet 3/4 Ton PicJcup
850689
850690
850691
850686
850688
11.9
11.9
11.9
11.9
11.9
14.5
13.2
10.3
-.1
2.6
81.7
82.3
82.3
91.3
89.7
6.048
5.813
5.795
7.916
7.964
96.2
95.5
92.6
91.7
92.3
92.0
93.0
93.5
93.2
93.5
16.6
16.6
.16.6
16.7
16.6
100.4
96.5
96.2
132.2
132.2
2.33
2.27
2.30
2.43
2.58
32.00
28.00
28.00
38.00
32.00

-------
850990  11.9
850991  11.9
850987  11.9
850988  11.9
                                 1979 Dodge Truck W150
12.5.
12.0
2.1
.2
• 80.0
80.0
91.7
91.8
6.593
6.456
8.984
8.950
92.5
92.0
93.8
92.0
94.0
92.7
95.0
91.8
18.2
18.0
18.3
18.0
120.0
116.2
164.4
161.1
2.47
2.38
3.30
2.78
46.00
42.00
30.00
40.00

-------
          Appendix B
Fuel Consumption Weighting Data

-------
                                 HIGHWAY  USE OF  8A80LINE BY NONTH8 -  1883
C04WUI* rot TM CAUMM MM
MOM M M*l«*I* «f NDTM-f Ml i
                                                   'IWOMMM •» «AllM*l
 TMII NT-M
MWNMt 1*14
•TATS
AlAIANA
ALASKA
ARIZONA
AkKAKSAS
CAiiFoiaiA
COIOHAOO
COIMCCTieUT
OELAWAII
oist. or coi.
FIOIIOA
CIOICIA
NAUAII
IDAHO
ULIMI*
IMIAM
IOWA
KANSAS
KIRTUCKT
IOUISIAM
MAIM
NAMIAND
NASSACNMtm
MICHICM
NUaiSOTA
NISSIMIMI
HISSOUII
NMTAM
UIIAIKA
MVAPA
MM NAMTMIM
MW JIHSSV
MW WtllCO
MM TOK
>0*TH CAMM.IIA
MWTN MKOTA
OHIO
OKLAHOMA
OMCOH
mMS«lVMIA
•NODI lit AW
SOUTH CAtOlIM
SOUTH MKOTA
TSMlMtt
TIKAt
HTAH
WflMMT
•IIICIIIIA
WAKHIMTM
WMT VIMIIIA
WISCOMU
moMiwt
TOTAL
FUCK ITA£I
TA» *ATt
o*
MCtNMI SI
II* CUTS
Tl»
•ALLO*)
U
II
II
s.s
1
It
14
II
14. 1
I.T
7.1
i.»
14. •
II
II. 1
It
1*
I*
1*
It.*
II
It
I*
»
II
It.l
It
14
II
It
It
U
•.I
It
It
15
u
u
u
i*
u. »
it
•
».»§
-
•
MWAIT
IIT.Ttl
ll.TII
llt.ttt
•l.tft
744, Til
1*7,117
•4. ft*
IT.MT
II. Ml
Ml. Ill
M*.4M
IS. II*
14. »4?
M4.«M
111. Itl
M.141
•4.1(1
tl«.i(0
1(4.111
M.I4*
llt.tIT
llt.(l(
tM.tU
111,4*1
TI.I1T
!4«.T*t
»l.(S*
41, (Tt
tt.ut
M.ltt
It*. tit
«»,»••
»l».t(»
MI. Ml
M.m
M>,»I4
I14.M?
Tt.tt*
t?t.*4»
M.4IS
t«.(44
K.TM
IM.t«4
til.***
41. M4
14. m
1 »•,!»»
IM.lt>
41. M(
IM.4M
IT.tft
• .NT .Ml
t.S4
flMMM
lt».M(
•K.ttt
ItT.Mt
••.IT*
T!T,«4S
tl.TM
M.tTI
It.tTt
11.114
1T4.*!1
Mt.*l*
tt.ltt
M.Mt
IT4.M4
Mt.tM
M.tM
(I.MI
•T.M*
IM.ttt
M.MT
IM,»4t
Ut.Ttt
tM,M4
U4.tM
M.4N
!4*.tT(
M.Mt
44. »M
ti.tr*
M.Mt
tta.iM
ta.Mt
ttl.MI
110. TM
It.UT
M*,tl(
114.114
TT.tll
ITt.ttt
It. 1(1
IM.MI
U.TI4
l»T,t4«
•M.M*
M.Mt
tt.ITt
IOT.MI
IHT.tM
4*.*M
IIT.TIt
K.I4J
(.(M.TII
t.Ti
NUtCN
Ut.MI
It.ltt
114,114
•t.tT4
t 10. 1*1
US. II*
•T.Ttt
tl.TtT
It .Ml
4tl ,MT
Ml,(l(
M.M*
11. Ml
141. ttl
1*1. ttt
ltt.lt!
M.TM
III. let
IT4.4W
IT.tIT
141, 4IT
ITt.TM
(41.144
141. Sit
M.M1
iM.tai
t».m
• 14. tit
it. (it
M.tM
IT*. 11*
M.44I
It*, Ml
III.M4
H.M4
UI.I4I
IU.M1
M,t4l
IM.tTT
M.Mt
II*. 141
M.MI
IT*. Ml
•M.MI
tl.4tl
IT. (41
1*1. tl*
IM.MT
tl.M!
IM.M4
M.IIT
T. Ml. 4*1
!.*>
MIK
I4«.»7I
11,11*
II*. Ut
tl.IM
Ml. Ml
III.***
1*1. 4M
M.MT
11.14*
4U.W
Ml. 1*4
14, M4
11.11*
IM.lll
Ml. Ml
IM.M*
M.4K
114.4*1
ITt.lll
•1T.IJ*
1*1.4*1
ITI.IU
174.1**
IM.MI
M.Mt
IM.tlt
IT.tll
U.1I4
M.M4
IT, lit
1(1.1*7
17.171
1*7.147
141. TIS
M.MI
Ml. IS*
ltl.lt*
M.MI
lll.lt*
M.MI
117. II*
14,114
1*4.7*7
(TT.(I(
M.m
it, M*
Ml ,1*1
III. Ill
((,(14
141, TI4
It. 11*
1. 1(1.474
•.IT
NAT
I**.***
II. Ml
II*. II*
It*. 17*
114.IM
III. Ml
111,71*
IT. 141
14. Ml
4*1. ttt
1 tl.TM
14. IX
IT. Ill
411.117
' tlt.tlt
114. Ml
1*1. Til
•I1I.T44
114. ttl
41,111
III. Ttl
111.141
III. Ill
II*. tl*
104. Ill
S14.II1
M.TI4
(t.Mt
41.11*
14.MT
MI.4IT
71. Tl*
471.111
141. Ml
M.MI
4*1. Ml
1*1.171
101. Ill
111.414
M.MI
IM.MI
tf.tll
lll.lt*
T«l.*l*
14.111
M.MI
117.1*1
III. 41*
It.***
ITI.ITT
11.11*
•.Mt.M*
*.ll
MM
II*. IIS
It. Ttl
II*. tit
M.4M
Ml. Ill
114. TM
III. Ill
St.TT*
14,111
Ml, III
III. tit
14.111
It.ttt
MI.ITt
III, III
IM.MI
111. Ml
III. til
IT1.ITI
M.SM
1(4. IM
ItT.ltl
llt.tM
III. 144
111. Ml
tt*.*lt
I I.MI
IT, III
41. Til
tl.M*
"IM.47I
Tl.M*
411.111
14T.1I1
It. IT)
1*1,114
lit. (It
IM.4T*
171. Ill
M.ltt
lit. IT*
ll.ttt
IM.tlt
TM.1M
(!.*!•
II. (14
111.7*7
1*7.11*
TI.*T*
1*1. IM
It. Ill
• .•M.Mt
*.M
,. . ii,.!!" IS!11 1* OM OF A SIKHS INT-SI TMCWCH NT-MI «IVIM A* AJULTIIS Of NOTOI-FML
JSSliSr1!?1;^-^!.*^,***" °" «»««• '«">" »T*W NOTC«-Hlll TA« ACIKIIS. II THIS! TMIM
S^iSTSLJ! iSlX!i! Xl!" "*»••«. u MMI TO HAKI TM MTA umrom AM CONFLIU. ISTIFMTI* «f
KSXJJSS'.SPf JSS.41" "»•* >• *M> «**«» •' TIM TIMIAl NISHUAT AMIJHSTIATION. TM
U*SiII5".5*ilS!SS?L!lf!!11 '• "*•* «•«*•««• '•«• TM MMJHSTM »ATA UCOMK u IAILI NT-I.
• - ~«. TO AMIOIINATI ACTUAL HSt. TW NMTM.T •ISTIItHTIMI MM IfTINATM S» TM FltltAl
ML*
llt.M*
1.14*
Itl.tTl
III.M4
l.ttl.tll
111, Ml
llt.M*
tl.lt*
II.1T1
411, Ml
Itl.tTt
IT.tIT
41.111
' 411. Itl
147,1*1
I1T.IT4
llt.lt*
IM.I1T
III. 411
M.M*
1*1.11*
ttl, til
MI.7M
ITI.tTT
IM.tlt
I4I.IM
M.MI
Tt.ttt
47. !!»
M.Mt
IM.MI
4*. MI
Mt.Ml
1*1,141
It. Ttl
411.117
17*. Ml
III. 117
414.114
14.71*
IM.4M
M.IM
111.11*
TIT. 171
M.IM
14.114
Itl. 11*
IM.I44
Tl.M*
IM.tTT
M.4I1
t.MI.ITI
».*T
AMMT
171.171
ll.tl*
III.1T1
lit.***
i.MT.in
llt.TM
114. M*
11.7*7
11.141
41*. TM
IM.tT*
M.tM
44. Itl
417.71*
I4I.ITI
141.14*
II*. Ml
I4*.*7»
114 .Ml
M.tM
IM.MI
117. Itl
IM.IM
IM.MI
IM.M*
144. M*
11,111
11.11*
4T.I4I
44.1*7
MI.M7
M.M*
•M.ITT
llt.TM
M.Mt
4I4.MI
ITI.14T
III. Ml
III.Mt
M.M4
141, »TI
M.IM
III. 114
III. Ill
•1.141
M.TM
. (M.IM
III. t 14
Tt.MI
IM.tt*
M.MI
• .IM.M*
*.W
MrTINM*
Itl. It*
lt.lt*
IM.tTT
M.tTI
tit .IT*
111,41*
III. ttt
IT. Ml
14, MT
171.411
111,411
tt.MI
M.MT
ttt. Itl
IM.Mt
lll.tl*
111.441
111. Ml
IM.1T*
M.TM
III. Ill
Itl. Ml
114, IM
ItT.III
M.4M
Mt.lll
M.IM
•1.174
41. IM
M.TM
MI.IM
Tl.M!
471.111
IM.MI
11. IM
111 .Ml
I44.TM
1*1.71*
Mt.Ml
M.tM
llt.M*
tt.MI
IM.Mt
Mt.M*
M.MT
It. Ill
II 1.141
IM.MI
17. IT*
174. IM
M.Mt
• .IM.MI
•.M
OCTMtl
141.111
1.41*
III. Ml
IT. 114
111,471
111,141
IM.MI
14.111
I4.(4(
4*1. 114
lit. lit
tl.IM
11. Ml
III. 114
1M.IIT
111.7*1
IM.tlt
141.4*1
177.144
4*. 4(1
II1.M4
111. 411
lit. Ill
III.Mt
11.411
III .114
11.111
ll.Mt
M.MI
I4.1M
Ml.tM
M.M*
4T1.I7I
141.11*
11.4*4
Mt.lll
IM.Mt
17.141
Ml. 171
11.41*
II*. 411
11.171
1*7.14!
•11.141
IT.tIT
1I.IIT
114.1*1
111,171
(I.II4
IM.M4
II. Ill
• .M1.IM
• .M
MWNMI
141. ttl
M.IM
II1.4IT
M.lll
•M.IM
IM.IM
M.IM
11,111
11.17*
411,111
111,111
14. IM
11.144
Ml. Ttt
It*. 11*
IM.IM
14.lt*
I1T.I4*
ITI.I4I
.M.M*
14*. IM
1*7,111
Mt.lll
144.111
M.IM
IM.MI
ll.Mt
M.M*
M.M*
M.Mt
ITI.lt*
M.M*
411. Ill
111.1*1
II .Ml
141.411
141.171
M.M4
Itl. ITT
M.M*
II*. tit
11,11*
III.Mt
Ml.tM
41, TIT
14. Tt*
ItT.Ilt
141. »TI
•4.7*1
IM.IM
It. 141
l.Mt^M
l.M
MCMMk
IM.Mt
ll.Mt
II*. M*
(1.1*4
141. Ill
M.IM
ll.tl*
14.111
11.117
441. Ill
141. Ml
14.17*
M.lll
ttl. Ml
114. Ttl
M.11T-
•I.MI
IM.MI
Itl.MI
M.MI
14*. 4M
17*. Ml
ttt.M*
Ut.MI
II. Ml
lit. 411
14.1*1
U.AI7
14.111
M.M*
ITI.M*
M.M*
IM.Mt
114,114
ll.Mt
141, M4
IM.IM
•I ,»•»
Mt.tM
tl.M*
II*. Ill
II. 41*
171.711
IM.Mt
•l.M*
It.ltt
III.MT
141.41*
M.IM
14*. IK
l*.tl*
7.IM.1TI
t.M
TOTAl
l.llt.44t
IM.Mt
l.tll.MI
I.Mt.MI
II.II4.KT
I.IM.I4I
1. ITT. Ml
MI.IM
IM.Mt
4.IM.II*
t, 141. 11*
IM.IM
411. Ml
4. Ml. TIT
I.MI. Ill
1. 114. MI
I.II4.IM
1.1(4. (IT
t, MI.IM
Mt.tM
l.M*. 144
t.MT.ITT
I.MI. II*
I.MI. 47*
1. II*. Til
1.414. Ill
411. »l*
IM.M4
414, IM
t*4.*44
•.lit. IT*
III. 114
*.III.*M
I.TM.IM
IM.MI
4,411,111
1. IT*. 441
l.l*l.*74
4, lit. 411
Ml. lit
I.4M.TT4
lit. 4*1
t.ttt.TTI
1. III. 147
III. Itl
111.114
1.471. 414
1. 117. 141
771.411
I.MT.tlt
tM.MT
M.M*. II*
IM.M
CHAHCI
FION
llll
MLIMI
1
t
tt.MT
-ll.Mt
M.MI
'•.111
114.171
-M.II4
11.140
M.TM
-T*l
1*7. 117
M.TII
TM
II. Ill
III. IT*
44.171
14, (11
• II. IT*
11.147
K.TIT
.•II
• .•41
M.lll
-11.411
•T.Mt
It, III
M.lll
11.417
-t.MI
7*1
11.141
til. Til
IM
-17.7*4
M.TII
-I. Itl
M.MI
-11.171
-TO*
II. Ill
-1.4*7
11,111
-7,111
-11.411
141.111
1.114
l.lll
11,1*1
M.MT
-14.141
11.174
-14.111
1. IM.Mt
-
ri«-
CIIT
1.1
-t.i
t.i
i.(
t.T
-1.1
I.I
1.7
-9.4
1.
1.
0.
I.
I.
1.1
I.T
l.t
1.7
*.*
1.4
1.1
I.I
-I.I
t.«
I.I
1.4
1.1
-I.I
t.t
I.I
I.I
*.l
-I.I
I.I
-1.*
*.*
•l.t
-1.1
1.*
-I.*
l.t
-1.1
-l.«
1.1
1.0
*.*
t'.t
-1.1
1.1
-4.4
l.t
-
HIMMAT AMUIimATIO* HIM WMt MT». TM tlTtltt 111 TAMI W-M Will MOT MMI WITH THOtt
1* TAW.I NT-IMA. TAIlt NT-M MOM MIIIIUAT «M »M.* Mill TMII NT-ltCA MrilCT* MAM CAllMt
•r «A*oiiM urotnt *» WMXIMII timitirroM TO TM ITATI TAIIM ACIKUS IACH MOTN. •~<-'"'
tl ttt TM IATII M OT MMMM 1, ltd M* FM IATI CHAWII WWIM THI «tM HI TMll
MT-ltTT. ITATU -WTH VAtlAlll IATH Ml AltO INMiril* III TMII NT-HIT.

-------
Listing of STATEDATA-RV at 07:49:43 on JUL 23. 1985 for CC1d=SN81
Page
$
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31'
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
ST
AH
AL
AR
AZ
CA
CO
CT
DC
OL
FL
GA
IA
ID
IL
IN
KS
KV
LA
MA
MD
ME
MI
MN
MO
MS
MT
NB
NC
NO
NH
NJ
NM
NV
NV
OH
OK
OR
PA
HI
SC
SO
TN
TX
UT
VA
VT
Ml
WN
WV
MY
HA
JAN
15.0
13.5
1S.O
13.5
13.5
15.0
15.0
15.0
15.0
13. 5
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
15.0
11.5
FEB
15.0
13.5
13.5
13. 5
13.5
15.0
15.0
15.0
15.0
13.5
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
13.5
15.0
15.0
15.0
13.5
13.5
15.0
15.0
13.5
15.0
1S.O
15.0
13.5
15.0
13.5
13.5
15.0
15.0
15.0
15.0
15.0
15.0
15.0
11 .5
MAR
15.0
13.5
13.5
11 .5
13.5
13.5
15.0
13.5
15.0
13.5
13.5
15.0
13.5
15.0
15.0
13.5
13.5
13.5
15.0
15.0
15.0
15.0
15.0
13.5
13.5
15.0
15.0
13.5
15.0
15.0
15.0
11.5
11.5
15.0
15.0
13.5
13.5
15.0
15.0
13.5
15.0
13.5
11 .5
13.5
13.5
15.0
15.0
15.0
15.0
15.0
1 1 .5
APR
15.0
11.5
11.5
10.0
11.5
11.5
13.5
13.5
13.5
11.5
11.5
13.5
13.5
13.5
13.5
11.5
13.5
11.5
13.5
13.5
13.5
13.5
13.5
13.5
11.5
13.5
13.5
13.5
13.5
13.5
13.5
10.0
11.5
13.5
13.5
11.5
13.5
13.5
13. 5
13.5
13.5
13.5
11.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
11.5
MAY
15.0
11 .5
11 .5
10.0
11 .5
11.5
13.5
11.5
13.5
11.5
11 .5
11 .5
11.5
11 .5
13.5
11.5
11 .5
1 1 .5
13.5
13.5
13.5
13.5
13.5
.5
.5
.5
.5
.5
13.5
13.5
13.5
10.0
10.0
13.5
13.5
11 .5
13.5
13.5
13.5
.5
.5
.5
0.0
.5
.5
13.5
13.5
13.5
13.5
11.5
11.5
JUN
13.5
11 .5
11 .5
9.0
10.0
10.0
11 .5
11.5
11 .5
11.5
11 .5
11 .5
10.0
11 .5
11 .5
10.0
1 1 .5
11 .5
1 1 .5
11 .5
11.5
11.5
11.5
11.5
11.5
10.0
10.0
1 1 .5
11 .5
1 1 .5
1 1 .5
9.0
10.0
1 1.5
11 .5
10.0
1 1 .5
11.5
11 .5
11.5
10.0
11.5
10.0
10.0
1 1 .5
1 1 .5
11.5
11 .5
11.5
10.0
11 .5
JUL
13.5
11 .5
10.0
9.0
10.0
10.0
11 .5
1.1.5
11 .5
11.5
11.5
1 1 .5
10.0
11.5
11 .5
10.0
11 .5
1 1 .5
11 .5
1 1 .5
11.5
11.5
11.5
10.0
11.5
10.0
10.0
11 .5
10.0
11 .5
11 .5
9.0
10.0
11.5
11.5
10.0
10.0
11 .5
1 1 .5
11.5
10.0
11.5
10.0
10.0
11.5
1 1 .5
11 .5
11 .5
11.5
10.0
11 .5
AUG
13.5
10.0
10.0
9.0
10.0
10.0
11.5
11.5
11 .5
11.5
10.0
11.5
10.0
11 .5
11 .5
10.0
1 1 .5
10.0
1 1.5
11 .5
11.5
11.5
11.5
10.0
10.0
10.0
10.0
10.0
10.0
11 .5
11.5
9.0
10.0
11.5
11.5
10.0
10.0
11 .5
11.5
10.0
10.0
10.0
10.0
10.0
.5
.5
.5
.5
.5
10.0
11.5,
SEP
15.0
11.5
11.5
9.0
10.0
10.0
11.5
11.5
11 .5
11.5
1 1 .5
11 .5
10.0
11 .5
11 .5
10.0
11 .5
11 .5
13.5
1 1 .5
13.5
13.5
11.5
11 .5
11.5
11.5
10.0
11.5
11 .5
13.5
13.5
10.0
10.0
13.5
11.5
10.0
1 1 .5
13.5
13.5
11.5
10.0
11 .5
10.0
10.0
11.5
13.5
11 .5
11.5
11.5
10.0
1 1 .5
OCT
15.0
11.5
13.5
10.0
11.5
11.5
13.5
13.5
13.5
11.5
11.5
13.5
1 1.5
13.5
13.5
11.5
13.5
1 1 .5
13.5
13.5
13.5
13.5
13.5
13.5
11 .5
13.5
11.5
13.5
13.5
13.5
13.5
11.5
10.0
13.5
13.5
11.5
13.5
13.5
13.5
13.5
11.5
13.5
11.5
11.5
13.5
13.5
13.5
13.5
13.5
1 1 .5
11 .5
NOV
15.0
13.5
13.5
11.5
13.5
13.5
15.0
15.0
15.0
13.5
13.5
15.0
13.5
13.5
15.0
13.5
15.0
13.5
15.0
15.0
15.0
15.0
15.0
13.5
13.5
15.0
13.5
13.5
15.0
15.0
15.0
13.5
11.5
15.0
15.0
13.5
13.5
15.0
15.0
13.5
13.5
13.5
13.5
13.5
15.0
15.0
15.0
15.0
15.0
13.5
1 1 .5
DEC
15.0
13.5
15.0
13.5
13.5
15.0
15.0
15.0
15.0
13.5
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
13.5
15.0
15.0
15.0
15.0
15.0
15.0
13.5
15.0
15.0
13.5
15.0
15.0
15.0
15.0
15.0
15.0
15.0
11.5

-------
Listing of OELTATMO at 07:49:15 on JUL 23. 198S for CC1d=SN81                                                              Page

                                                                          OCT<  NOV   DEC
                                                                          3.4   0.2  -5.4
                                                                          2.4  -7.8  -3.5
                                                                          2.7  -4.B  -1.9
                                                                          2.1  -3.4   6.1
            5-FW   -3.7  -3.2  -4.0  -2.9  11.9   3.7   0.0   6.3   3.9   1.1 -11.3  -5.9
1
2
3
4
5
REGN
1-NE
2-SE
3-SW
4-NC
JAN
-10.3
1.9
3.4
0.2
FEB
0.2
5.4
-8.2
-9.0
MAR
8.2
7.4
5.4
-11 .3
APR
6.3
3.4
10.3
9.3
MAV
14.5
7.4
11.6
7.6
JUN
15.6
6.1
10.0
19.3
JUL
15.9
3.5
4.9
15.5
AUG
9.1
13.0
9.1
11 .2
SEP
1 .7
7.4
-.8
6.3

-------
Listing of TOISPMO at 07:49:30 on JUL 23. 1985 for CC1d=SN81                                                               Page
1
2
3
4
6
6
7
REGN
1-NE
2-SE
3-SW
4-NC
5-FW
6-NW
JAN
43
69
54
50
54
999
FEB
45
74
57
51
57
48
MAR
48
73
61
41
62
49
APR
S3
80
67
47
67
S3
MAV
66
84
76
63
72
59
JUN
74
87
82
74
77
63
JUL
78
90
83
88
83
999
AUG
78
91
84
85
83
73
SEP
72
88
79
83
79
71
OCT
66
85
76
75
74
60
NOV
59
83
67
63
67.
49
DEC
46
73
54
52
58
42

-------
                              -58-
                           References

     1.    "An   Experimental   Study   of   Vehicle   Refueling
Emissions," Albert  M. Hochhauser  and  Raymond J.  Campion,  SAE
Paper No. 760307, Feb. 1976.

     2.    "Testing of  a  Vapor  Balance  Service  Station  Vapor
Control  System  by  the  California  ARB  Test  Procedure  — BREA
June  1976,"  Technical  Memorandum  from M.J.  Dougherty to  Mr.
Cloyd  P.  Reeg,  Research  Department,  Union  Oil  Company  of
California, August 19, 1976.

     3.    "Expansion   of   Investigation   of   Passenger   Car
Refueling  Losses,"  EPA-460/3-76-006   U.S.  EPA,   OAWM,  OMSAPC,
ECTD, September 1975.

     4.    "Healy Phase  II Vapor Recovery  System Certification
Report," Scott Environmental Technology, June 1982.

     5.    "A Service Station Test  of  a  Vapor  Balance System
for   the  Control  of   Vehicle   Refueling   Emissions,"   A.M.
Hochhauser and  L.S.  Bernstein,  Exxon  Research  and Engineering
Company, July 1, 1976.

     6.    "Evaluation of  Test Procedures  for Measuring Vehicle
Refueling Emissions,"  API Publication No. 4276, July 8,  1976.

     7*    "Service   Station  Vapor   Recovery:    Vapor  Balance
System  Performance  Diamond Bar,  California,  March 5-12,  1974,"
Atlantic  Richfield Company,  Products  Division,   Research  and
Engineering,  April 8,  1974.

     8.    "Vapor   Recovery  Nozzle    Development  and  Field
Testing,"  B.E.  Weidenaar,  H.J.   Grimes, and  R.G.  Jewell,  SAE
Paper No. 741038, October  1974.
                               /  '
     9.    "A  Study  of  Variables  that  Effect  the  Amount  of
Vapor  Emitted During  the Refueling of  Automobiles," API  Report
CEA-21, May 16,  .1975.

     10.   "Vapor  Control  Efficiency  of  Simple   Displacement
Systems  at Two Service  Stations,"  Technical  Memorandum from
M.J.  Dougherty,  Research  Department,  Union  Oil  Company  of
California, September 22,  1975.

     11.   "Air   Pollution  Emission   Test:     Emissions   from
Gasoline Marketing  Operations at  Exxon Retail Station,  Hayward,
California," U.S. EPA, OAWM, OAQPS, EMB, April 1975.

-------
                              -59-
     12.   "Air   Pollution  Emission   Test:    Emissions   from
Gasoline  Marketing  Operations  at  Standard  Oil  of  California
Retail Station, Davis,  California,"  U.S.  EPA, OAWM, OAQPS, EMB,
May 1975.

     13.   "On-Board  Control of  Vehicle  Refueling  Emissions:
Demonstration  of   Feasibility,"   API  Publication  No.  4306,
October 1978.

     14.   "Vapor   Control  Concepts,"   M.W.   Lieferman,   in
"Vehicle  Refueling  Emissions  Seminar,"  API  Publication  No.
4222,  December 1973.

     15.   "A  Chronic  Inhalation  Study  With  Unleaded  Gasoline
Vapor," Journal of  the  American College of Toxicology,  American
Petroleum Institute, 1984.

     16.   "Compilation  of  Air  Pollutant  Emission  Factors,"
(AP-42), U. S. EPA, OAWM, OAQPS, 1977.

     17.   "Investigation  of Passenger  Car  Refueling  Losses,"
APTD-1453, Scott Research Labs for EPA and CRC, September, 1972.

     18.   "Utility  of  Reactivity Criteria  in Organic  Emission
Control Strategies  for  Los  Angeles," Final Report, EPA Contract
No. 68-02-1735, December, 1975.

     19.   "Phase  II  Vapor Recovery  Evaluation  Program,"  South
Coast Air Quality Management District, c.1980.

     20.   "Summary   and   Analysis   of   Data   From   Gasoline
Temperature  Survey Conducted at  Service Stations  by  API" (API
Publication No. 4278), Radian Corporation, 1976.

     21.   "Standard  D439-83,"  Annual  Book  of_  ASTM Standards,
Part 23, American Society for Testing and Materials, 1983.

     22.   "Digest   of   State   Inspection  Laws   -  Petroleum
Products," Fourth Edition, API Publication 926, 1985.

     23.   "Motor" Gasolines,  Winter  1983-84"   (NIPER-135  PPS
84/3),  U.S.  DOE,  National  Institute  for  Petroleum  and  Energy
Research  (NIPER), June,  1984.

     24.   "Motor Gasolines, Summer  1984"  (NIPER-138 PPS  85/1),
U.S. DOE, NIPER, February,  1985.

-------