United States
Environmental Protection
Agency
Office of Pesticides
and Toxic Substances
Washington, DC 20460
EPA 560/5-86-013
May 1986

Toxic Substances
Underground  Motor Fuel
Storage Tanks:
A National Survey
VOL.  II.  APPENDICES

-------
                         UNDERGROUND MOTOR
                        FUEL STORAGE TANKS:

                         A NATIONAL SURVEY

                            Appendices
                           Prepared by:
Westat, Inc.
1650 Research Boulevard
Rockville, MD 20850

Battelle Columbus  Division
Washington Operations
2030 M Street, N.W.
Washington, D.C. 20036
Midwest. Research  Institute
425 Volker Boulevard
Kansas City, MO 64110

Washington Consulting Group
1625 Eye Street,  N.W.
Suite 214
Washington, D.C.  20006
                             for the:

                   Exposure Evaluation  Division
                    Office of Toxic Substances
            Office of  Pesticides and Toxic Substances
               U.S.  Environmental Protection  Agency
                      Washington, D.C.  20460
                                        U.S. Environmental Protection Agency
                                        Region V,  Library
                                        230 South Dearborn Street    '
                                        Chicago, Illinois  60604
                            May 1,  1986

-------
U,S. Environmental Protection Agency
                                     DISCLAIMER
         This  report  was prepared  under  contract to  an  agency of  the
         United  States Government.   Neither the United  States  Government
         nor  any of its  employees,  contractors, subcontractors,  or  their
         employees  makes any  warranty,  expressed  or  implied,  or assumes
         any  legal  liability or responsibility  for any  third party's use
         of  or  the results  of  such  use of  any  information,  apparatus,
         product, or process disclosed in this  report,  or represents that
         its  use by such third party would not infringe on privately owned
         rights.

-------
                             CONTENTS
Appendix                                                   Page
   A      SAMPLE DESIGN AND ESTIMATION OF WEIGHTS
          AND VARIANCES	   A-l

   B      SURVEY PROCEDURES AND ELIGIBILITY AND
          RESPONSE RATES 	  B-l

   C      DEVELOPMENT OF A TANK TEST METHOD 	  C-l

   D      TANK TESTING DATA REDUCTION AND STATISTICAL
          ANALYSIS LEADING TO LEAK STATUS DETERMINATION  ..  D-l

   E      INVENTORY RECONCILIATION METHODS 	  E-l

   F      DATA COLLECTION FORMS AND MATERIALS  	  F-l

   G      NATIONAL UNDERGROUND STORAGE TANK SURVEY
          NATIONAL SAMPLE OF FARMS 	  G-l

   H      ENVIRONMENTAL DATA COVERAGE 	  H-l

   I      MULTIVARIATE ANALYSIS 	  1-1

-------
                            APPENDIX A

                   SAMPLE  DESIGN AND ESTIMATION
                     OF WEIGHTS AND VARIANCES
I.   TARGET UNIVERSE. OVERVIEW OF SAMPLE DESIGN

     The target universe, or population of interest, for the
Survey of Underground Storage Tanks consisted of all underground
tanks which store motor fuel prior to dispensing it for use as
fuel, with exceptions as noted below.  For example, in the retail
gasoline sector, this includes all underground tanks at service
stations but excludes large holding tanks at a distributor.  In
sampling, we used a tank establishment, that is, a location with
eligible tanks, as the sample unit.  Once a given establishment
was sampled, all its tanks were in the sample for the initial
data collection phase.  For the physical tank testing stage, a
subsample of the sampled establishments was drawn, and all tanks
at the subsampled establishments were physically tested.  For
purposes of list building, the target universe of establishments
was defined as a number of segments, with certain exclusions as
noted.  The following types of establishments were identified as
potentially having underground motor fuel storage tanks:

     o    Gasoline service stations;
     o    Other establishments almost certain to have underground
          storage tanks, including:
               Transit and transportation fleets (such as taxi,
               trucking, and bus companies; auto and truck rental
               companies; railroads; and auto and truck dealers);
               Marinas;
                              A-l

-------
               Airports and other air transportation related
               industries;  and
               Golf courses and country clubs;
     o    Government fleet service pumps,  including:
               Federal;
               State;
               Local — county, city, etc.; and
               Military;
     o    Large companies with 20 or more employees in other (non
          fuel-related) industries which have private fleet
          service pumps; and
     o    Farms with underground motor fuel storage tanks.

     Underground tanks containing motor fuels maintained by
private homeowners and tanks for private fleets maintained by
companies with fewer than 20 employees were excluded from the
scope of this survey.  They were not estimated to account for a
large number of underground storage tanks.  In addition, the cost
necessary to screen out businesses and residences with no
underground tanks was judged to be too great in comparison with
the anticipated low addition to the total universe from these
establishments.
     A.   Overview of Sample Design

     The sample of establishments was drawn using a multi-stage
cluster design.  The continental U.S. was divided into six
regions of interest.  The sample was drawn to provide estimates
both at the national and regional levels.  The first stage of
sampling was Primary Sampling Units  (PSUs) consisting of counties
or groups of contiguous counties with designated minimum
                              A-2

-------
estimated numbers of underground tank establishments.  The sample
of PSUs was allocated to the regions and drawn within region
proportionally to their total estimated number of underground
tank establishments.  Thirty-four PSUs were drawn.

     Within each selected PSU, three establishment frames were
developed:
     o    Fuel tank establishments - consisting of gas stations,
          establishments in other fuel-related Standard
          Industrial Classification (SIC) groups, and government
          tank locations;
     o    Large establishments - consisting of all businesses
          with 20 or more employees not already listed as fuel
          tank establishments; and
     o    Farms - consisting of all farms.

     A national sample was drawn from each frame.  For large
establishments and for farms, 600 establishments were selected
from each frame.  For the fuel tank establishments, a national
sample size of 1,618 was allocated to the regions, and six
regional samples were drawn.  In each case, the establishment
sample was drawn taking account of the PSU probabilities of
selection in such a way that the establishment samples were self-
weighting, nationally for the large establishments and farms, and
by region for the fuel tank establishments.

     Once the three samples were drawn, the large establishment
and farm samples were telephone screened for the presence of
underground tanks.  All large establishments and farms which have
underground fuel storage tanks became part of the field sample,
as did cases which could not be resolved over the telephone.  No
substitutions were made for large establishments or farms with no
underground fuel storage tanks.  The fuel establishment tank
sample consisted of establishments which were thought likely to
                              A-3

-------
have underground fuel storage tanks.   Initial field work showed
that this list actually produced about a 50 percent survey
eligibility rate; that is,  about half the sampled establishments
sampled were still in business and had underground motor fuel
storage tanks.  Although lower than anticipated,  this eligibility
rate indicates that the coverage of the target universe by the
selected SICs was probably quite good.  In order to attain our
target sample size of 800 eligible establishments, the initial
sample sizes per region were doubled for the fuel establishment
segment, for a total sample draw of 1,618 cases.
     B.   Definition of Regions; PSU Sample Design

     Table A-l lists the regions, giving the states included in
each.  They are shown on a map in Figure A-l.  The PSU frame was
developed for the entire continental U.S. as detailed in the
following paragraphs.

     For each county, the following counts were developed:

     o    Number of gas stations based on the 1981 County
          Business Patterns data (count for SIC 5541);
     o    Additional estimated number of gas stations allocated
          to counties within states on a population basis to
          bring the state totals up to the estimate provided by
          Versar to the EPA; and
     o    Total number of establishments in the selected other
          SICs (list in Table A-2)  as given by the County
          Business Patterns data.

These three counts were summed for each county to form the
estimated number of fuel tank establishments for the county.

     The counties were grouped into initial PSUs by using the
Westat Master PSU Frame developed on a population basis, which
                              A-4

-------
Table A-l.
Six regions for National Survey of Underground Fuel
Storage Tanks
1 — Northeast

     Maine
     New Hampshire
     Vermont
     Connecticut
     Massachusetts
     Rhode Island
     New York
     New Jersey
     Pensylvania
     Maryland
     Delaware
     Virginia
     West Virginia
     Washington, D.

2 — Southeast

     Kentucky
     Tennessee
     Arkansas
     Louisiana
     Mississippi
     Alabama
     Georgia
     North Carolina
     South Carolina
     Florida
                     3  —  Midwest

                          Wisconsin
                          Minnesota
                          Iowa
                          Missouri
                          Illinois
                          Indiana
                          Ohio
                          Michigan

                     4  —  Central

                          North Dakota
                          South Dakota
                          Nebraska
                          Kansas
                          Oklahoma
                          Texas

                     5  —  Mountain

                          Montana
                          Wyoming
                          Idaho
                          Nevada
                          Utah
                          Colorado
                          Arizona
                          New Mexico

                     6  —  Pacific

                          Washington
                          Oregon
                          California
                               A-5

-------
A-6

-------
Table A-2.  Selected SIC codes for fuel tank establishment frame
SIC  code

4010

4110+
 4121 +

 4131 +

 4140+


 4151

 4170


 4210+

 4231+

 4469A

 4511

 4521 +


 4582A

 4582B+

 4583

 5511 +

 5521 +

 5541+

 7512 +

 7513 +

 7519+

 7992 +

 7997B+
                 Description

Railroads, switching and terminal companies

Local and suburban passenger transportation
companies (includes airport transportation,
ambulance and limousine services)

Taxicab companies

Intercity highway transportation services

Passenger transportation charter services
(includes bus charter, rentals and tours)

School bus companies

Passenger transportation terminal and service
facilities

Trucking companies

Motor freight terminals

Marinas

Air transportation, certificated carriers

Aircraft charter, rental and leasing —
non-certificated carriers

Airports

Aircraft maintenance services

Airport terminal services

Auto and truck dealers (new and used)

Used car dealers

Gasoline service stations

Passenger car rental and leasing agencies

Truck rental and leasing agencies

Utility and house trailer rental agencies

Public golf courses

Golf and country clubs

             A-7

-------
follows the PSUs used by the Census Bureau in designing the
Current Population Survey.  This initial list of PSUs was
transformed to a final list by splitting PSUs which had large
total counts into smaller sets of counties and combining PSUs
with insufficient counts, resulting in a set of PSUs which were
as small as possible while still containing a minimum number of
fuel tank establishments.

     Once the PSUs were defined, the sample of PSUs was drawn as
follows.  For each region, a target number of PSUs was
established.  This was six PSUs per region, except in Region 5
(Mountain) where four PSUs were drawn.  Within each region, the
PSUs were sorted by an urban versus rural designation, then by
state, and finally by size (total estimated number of fuel tank
establishments).  The sample of PSUs was then drawn within each
region on a probability proportional to size basis.
     C.   Tank Establishment Frames Within PSUs; Sample of
          Establishments

     Once the thirty-four PSUs were selected, three establishment
frames were built for each PSU.  A sample was drawn from each
frame, and eligible establishments in the three samples formed
the sample of establishments.

     The first frame was the fuel tank establishment frame.  It
consisted of establishments considered to be extremely likely to
have underground fuel storage tanks.  The frame was constructed
from several sources.  A list of business establishments with one
of the target SICs (refer to Table A-2) in the selected counties
was purchased from National Business Lists (NBL).  This was
supplemented by any establishments found to have one of the
selected SICs in the large establishments list  (see below).
                              A-8

-------
Lists of Federal, state, and local government establishments in
the sampled counties with underground fuel storage tanks were
developed by extensive telephone contacts with government
officials.  In addition, a list of military establishments with
underground fuel storage tanks was provided by the military to
EPA.  These lists were keypunched and added to the fuel tank
establishment frame.

     The sample of fuel tank establishments consisted of 1,618
establishments in the country (in order to achieve a target of
800 eligible establishments).  Based on the regional totals of
number of such establishments developed in the PSU frame-building
effort, the total sample size was allocated to the six regions.
Within each region, the establishments were sorted by PSU and
SIC, and a self-weighting sample was drawn.  Since the PSUs were
sampled proportionately to the estimated number of
establishments, this resulted in an approximately equal number of
establishments per PSU within each region.  There was not a
precisely equal number per PSU for two reasons:  the PSUs were
sampled based on CBP counts and the establishments were sampled
based on actual frame counts; and the PSU sample measure of size
did not include an estimate for number of government
establishments.

     The second frame to be developed was the large
establishments frame.  This frame consisted of a list of business
establishments in the sampled counties with 20 or more employees
purchased from Dun's Marketing Identifiers (DMI).   The
establishments on this list with the fuel tank SICs (Table A-2)
were clerically compared with NBL lists, county by county, to
eliminate duplication between the two frames.  Duplicates were
deleted from the DMI list, and any establishment on the DMI list
with one of these SICs not found on the NBL list were moved to
                              A-9

-------
the NBL list.  The resulting DMI list was the frame for large
establishments not in fuel tank SICs.

     The sample of large establishments was drawn by first
sorting the frame by region, PSU, and number of employees.  Then
a self-weighting sample of 600 establishments was drawn across
the whole country.  These establishments were contacted by
telephone to determine whether they had underground fuel storage
tanks.  Those that did were part of the sample for initial data
collection; no substitution was made for establishments with no
tanks.

     The third frame was farms.  This was constructed by
obtaining a list of all farms in the sampled counties from the
U.S. Department of Agriculture, through EPA.  The list included
crop acreage for each farm.  Any establishment on the DMI list
with an agricultural SIC code was deleted from the DMI list and
added to the farm frame if it did not already appear there.

     The farm frame was sorted by region, PSU, and acreage.  A
national self-weighting sample of 600 farms was selected.  These
were screened by telephone to determine the presence of
underground tanks.  As with large establishments, no substitution
was made for farms with no tanks.
II.  PRIMARY SAMPLE UNIT fPSU) SAMPLE

     This subsection discusses the first stage sample of Primary
Sampling Units (PSUs).   Appendix H discusses the sample of farms
from PSU selection though the final sample of farms.  Thus, this
subsection and the following ones concentrate on the fuel
establishments and large establishments, although some data on
farms are included for completeness.
                              A-10

-------
     This subsection begins with a statistical description of the
six survey regions based on data gathered in the construction of
the Primary Sampling Unit (PSU) frame.  It goes on to describe
the PSU sampling process and concludes with a discussion of the
sample of PSUs drawn.
     A.   Survey Regions

     The six survey regions are defined in A-I, above, which
includes a list of states in each region (Table A-l) and a map of
the regions (Figure A-l).  Here we describe the regions
statistically in terms of characteristics important to the
present study.  Table A-3 gives several characteristics by
region, both the amounts and the percent distributions.

     The number of states and counties in each region is simply
based on the definitions of the regions.  The number of states
ranged from three states in the Pacific Region (Region 6) to 14
states in the Northeast Region (Region 1).   Alaska and Hawaii are
not included,  and the District of Columbia is counted as a state,
making the total 49.  In these 49 states there are 3,111
counties.  The number per region ranges from a low of 133, again
in the Pacific Region, to a high of 874 in the Southeast Region
(Region 2).

     The first step in constructing the PSU frame was to define
PSUs, a process described in Subsection A-I.  These consist of
counties or groups of counties with a minimum estimated number of
fuel establishments.  The minimum was set separately for each
region based on the expected number of establishments to be
sampled per PSU in each region.  The resulting PSU definition
groups the 3,111 counties into 1,362 PSUs.   The number per region
                              A-ll

-------
CM ,„
 »— -^•**
I. 1> cr Ou ,-.
0> o> t g CX

Z *- 2 CM *~
f~-
«. *'"X
O*  in
F~ t
o
o en -o o ^T*
*   — 11
l^S ^5
J2 ,w m <-• CO
*- **- o ON
o i ^~
L_ *r** w
o> ^  .i>
log
2 1
^^S§
~ yt *^J
>> c
0) O
& 'bO
3 0)
CO M

ON \O
O ^ ON ^
CM* O CD ^
CM •— T CM
CM -^ m x-.


to ^ r- ^
oq 6S to 6?
in G o C

'-"CM oo" CM
co ^- r- ^~-

CM S o S
oo ;_ ^r ^
tO *°~* ^"" ^~*

\o ^^ ^o
\D CM ON tO
TJ- >- m — .
O CO
o r^ ^^
^ £» \o ^5
fy^ ^^ \O *-Q
tO ^*^ vO ^~
CM T v-


,_ , 	
—"CM in CM


C7< OO
— T
CM fO


\O T
to r-
T 00

•^- o

m 01
in 
r-- ^-<


^•^
(~v- fi£
co" ^

o ^
to" CM


^-^
~~. "T
r-- CM
CM ^-


sg
in •—
to ^-i
ON
^1 fes
oo in
T —



ON
tO CM
in •—


to
to
to


CO
to

CO


in
CM


o
in
\Q

^O

1
rr |
o

r- o
r-- to
5~8 sg


? ~ vO -^
"~ 6? ^"6^
ON in r*- 12
CM^- COC

oo in
CM — CO Q
? w 2 1
— *"" to ^->

ro to _.
i^- ^ ^ S
CM 6? °° CM
in 12 to" —
*~ ^^

in c-i
in ^ j ^
co" 12 £ S
fO T
*~i e>? °i &?
in O\ CD —
in CM — —
00 "— 10 ^-<


to ro ^
f*> ^ t ^
•-12 o —
— to •—


r- vo
— CO


o ro
oo to
CM —

co to

c 0
3 c
in g vo «
£

,..
oo
in"
ro
ci
CM

to
IO_
to"
CO
in
00
vO
OO
. in"
o
to
0
CM
CM
'

CO
in
\q
to
ON
o
to
o
CM

ON
c-i

CO
r-
in
CM

01



r—
^
to"
ON

t-l
Rl
4J
O
H






T^"
CO
ON
^™
_^
Q.
UJ
o
-t^a
-^>
0
0.
I_
Q.
<_
0

*5
on
Q)
>
^\







*—
CO
ON
^"
of
c
O.
m
o>
c
'in
^y
CO
^
c
3
O
O
*^%.
CM


en
CO
§
a
5
m

^•3
~~
CJ
u_
•o
i
•
E
s
_0
"3
.0
ff)
*^5
^_,
s
l_
0>
Q.
X"S
T
A-12

-------
ranges from a low of 86, again in the Pacific Region, to a high
of 348 in the Southeast Region.

     Two further statistics help set the stage for the survey in
describing the regions:  the number and percent of 1980
population in each region; and the square miles and percent of
continental land area in each region.  In terms of population,
Regions 1-3 (the eastern block of regions) contain 27, 20 and 24
percent of the population, respectively, for a total of 71
percent of the population.  Regions 4-6 have 10, 5 and 14 percent
of the population, respectively.  For land area the situation is
reversed, though not as dramatic.  Regions 1-3 contain 39 percent
of the land area, while Regions 4-6 contain 61 percent.

     The next three statistics form the basis of the PSU
selection.  The number of gas stations was estimated per state by
Versar.   The distribution by region ranged from 5 percent in the
Mountain Region  (Region 5) to 31 percent in the Southeast Region
(Region 2).  Regions 1-3 contain an estimated 73 percent of the
gas stations.   The number of establishments with a Standard
Industrial Classification (SIC) code among those selected as
likely to have underground motor fuel storage tanks (see list in
Table A-2) was found as counted in the 1981 County Business
Patterns data.2  Seventy-three percent of these other fuel
establishments are in Regions 1-3.  The percent by region ranges
1Leaking Underground Storage Tanks Containing Engine Fuels.
 draft, March 1984, prepared by Versar, Inc.  The gas station
 estimates were based on figures given in the 1983 Petroleum
 Marketing News Fact Book and include all retail outlets for
 branded gasoline.
2At the time of PSU sample selection, the 1982 CBP data were not
 yet available.  They became available in time to use for final
 weights, as discussed in Subsection A-V.
                              A-13

-------
from a low, again in the Mountain Region,  of 5 percent to a high
of 31 percent in the Northeast Region.  These two figures (gas
stations and other fuel establishments)  are summed to form the
sampling measure of size.  The distribution of gas stations and
other fuel establishments follows that of the population.

     Although the PSUs were sampled based on the number of fuel
establishments, a sample of large establishments (with 20 or more
employees) and of farms was also to be drawn from the sample
PSUs.  The region statistics show that large establishments
follow the same general pattern as population and fuel
establishments:  5 percent are found in the Mountain Region and
27 percent in the Northeast Region; Regions 1-3 contain 69
percent of the large establishments as reported by the 1981
County Business Patterns data.  Farms are found mostly in Regions
2-4, which have 78 percent of farms as shown in the 1982 Census
of Agriculture.  Looking at the East versus West breakdown we
have been considering, the Eastern regions (Regions 1-3) contain
67 percent of the farms.

     In Table A-4 some of these statistics are shown for the
urban/rural breakdown.  Each PSU is designated as urban or rural
according to whether it is part of a Statistical Metropolitan
Area or not.  The majority of PSUs and constituent counties are
designed as rural (65 percent of PSUs, 77 percent of counties),
but the majority of the fuel establishments plus gas stations are
found in urban PSUs (69 percent).  The large establishments are
even more concentrated in urban PSUs, with 85 percent found there.
     B.   Sampled PSUs

     The sample of PSUs was drawn as stated in Section A-I, using
the number of fuel establishments as a sampling measure of size.
                              A-14

-------
  Table A-4.   Summary of PSU sampling frame,  urban versus  rural  PSUs
              (percent distributions  in parentheses)
Urban/
Rural
Urban
Rural
Continental
Total
Number
of
counties
722
2,389
3,111
Number
of
PSU's
482
- 880
1,36Z
Sampling
measure
of size (1 )
212,164
(69%)
93,704
(31%)
305,868
Large
establishments
(>20 empl. )
1981 CBP (2)
479,461
(85%)
103,852
(15%)
583,313
(1) Number of gas stations (Versar) plus other fuel-related
    establishments (CBP)

(2) County Business Patterns data
                                     A-15

-------
Thirty-four PSUs were drawn — six from each region, except
Region 5 where four were drawn.  Tables A-5 and A-6 give
estimates of frame counts that would result from weighting the
PSU sample data by inverse of the sampling probability.  This
gives an indication of how closely the sample reflects the frame
from which it was drawn.  Not surprisingly, the sampling measure
of size (number of fuel establishments) tracks the population
very closely, with the same percent distribution by region and
only one percentage point different for the urban/rural
breakdown.  The large establishment counts are reproduced fairly
well by the weighted sample.  The percent distribution by region
is within one or two percentage points of the population
distribution, but the urban/rural breakdown is not as close.
While 85 percent of large establishments were in the urban PSUs
nationally, in the weighted sample PSUs, 79 percent are in the
urban PSUs.

     Tables A-7 and A-8 give unweighted counts for the sampled
PSUs.  In Table A-7, we see that the 34 PSUs are composed of 76
counties.  The number of fuel establishments plus gas stations as
estimated from the Versar and CBP sources for the sampled PSUs is
27,753, and the estimated number of large establishments is
74,768.  Table A-8 shows that 11 of the 34 PSUs are rural, with
36 of the 76 counties.  The rural PSUs tend to have more counties
in order to contain the minimum number of fuel establishments.
The vast majority of both fuel and large establishments in the
sampled PSUs are in the urban PSUs (95 and 98 percent,
respectively).  In the sample, one county, Los Angeles, was large
enough to be self-representing.  This PSU accounts for the large
unweighted counts for Region 6 (Pacific) throughout the tables.

     Overall, the PSU universe appears to be well reflected in
the sample of PSUs.  Figure A-2 shows the location of the sampled
PSUs to indicate their geographic representation, as well.  The
                               A-16

-------
 Table A-5.  Weighted data from sampled PSUs, region summary
             (percent distributions in parentheses
Survey
region
1
Northeast
2
Southeast
3
Midwest
4
Central
5
Mountain
6
Pacific
r* — i : — .—
Total
Number
of
counties
561
635
912
1,660
344
114

4,227
Number
of
PSU's
210
. 341
328
327
120
73

1,399
Sampling
measure
of size (1 )
81,364
(27%)
78,974
(26%)
63,139
(21%)
36,374
(12%)
14,030
(5%)
31,988
(10%)

305,868
Large
establishments
(>20 empl.)
1981 CBP (2)
148,906
(25%)
123,360
(21%)
135,842
(23%)
57,475
(10%)
29,440
(5%)
89,358
(15%)

584,381
(1) Gas stations plus other fuel establishments

(2) County Business Patterns data, 1981

(3)  Percentages may not add to 100 due to rounding,
                                    A-17

-------
 fable A-6.  Weighted data from sampled PSUs,  urban versus rural summary
             (percent distribution in parentheses)
Urban/
Rural
Urban
Rural
Total
Number
of
counties
613
3,614
4,227
Number
of
PSU's
364
1,036
1,399
Sampling
measure
of size (1 )
207,558
(68%)
98,309
(32%)
305,867
Large
establishments
(>20 empl. )
1981 CBP (2)
462,468
(79%)
121,913
(21%)
584,381
(1) Gas stations plus other fuel-related establishments
(2) County Business Patterns data
                                         A-18

-------
 Table A-7.   Unweighted PSU sample data, region summary
Survey
region
1
Northeast
2
Southeast
3
Midwest
4
Central
5
Mountain
6
Pacific
Total
Number
of
counties

13

12

14

19

10

8
76
Number
of
PSU's

6

6

6

6

4

6
34
Sampling
measure
of size ( 1 )

5,453

3,321

2,317

5,074

1,144

10,444
27,753
Large
establishments
(>20 empl.)
1981 CBP (2)

9,051

5,888

6,555

12,573

3,058

37,643
74,768
(1) Gas stations plus other fuel-related establishments
(2) County Business Patterns data
                                    A-19

-------
 Table A-8.  Unweighted PSU sample data, urban versus rural summary
Urban/
Rural
Urban
Rural
P.-inHtiuni j]
Total 3 1
Number
of
counties
40
36
76
Number
of
PSU1 s
23
11
34
Sampling
measure
of size (1 )
26,627
1,126
27,753
Large
establishments
(>20 empl. )
1981 CBP (2)
73,305
1,463
74,768
(1) Gas stations plus other fuel-related establishments




(2) County Business Patterns data
                                     A-20

-------
   I1  '-. T ;T~V'J- <:T!" ' ^\ I
I .s, [ Ij U j !! t-~~tJv i '!!;r-A ! A : i ?;
   if
     i L
FOw|.. X H«
f&m  Va<
              • ?'.'
               S(
   s /Ijtw^lli^-iupl: -ti^lfe^S
   ^g|t ^.] ^^^^^^^^


   ! 'Ssillfff IIMllilsii£w^!=
                   (0

                   o
                   n)
                   o
                   o
                  £3
                  I/I
                   I
                  CM
                   I
A-21

-------
establishment sampling frame construction and establishment
sample draw are described in the next section.
III.  ESTABLISHMENT SAMPLE

     Once the 34 PSUs were drawn, lists of all establishments in
the three sampling sectors were constructed for the 76 counties
which comprise the 34 PSUs.  These lists are known as sampling
frames.  The initial sample of 2,818 establishments was drawn
from these frames and screened for eligibility.  Since so little
was known initially about what type of establishment would have
underground motor fuel storage tanks, the eligibility rates
themselves were an early finding of the survey.  The 896 eligible
establishments form the final sample for the survey.  This
process is described in detail below for the fuel establishment
and large establishment sectors  (which account for 2,218 initial
sample cases and 876 eligible cases).  Appendix H reviews the
process for the farm sector (600 initial cases and 20 eligible
cases).
     A.   Sample Frames for Fuel-Related Establishments and
          Large Establishments

     The sample frames were constructed as described in Section
A-I, above.  For the fuel-related establishments, several methods
of list-building were combined to result in a single list.  A
list of government agencies with eligible tanks was developed for
each PSU by a telephone search.  Federal, state and local
government officials were contacted to generate lists of all such
civilian agencies, and a list of military establishments with
eligible tanks in the sampled counties was provided to EPA by the
Department of Defense (DoD).  A list of the fuel-related business
                              A-22

-------
establishments (gas stations and other industries, see list in
Table A-2) was purchased from National Business Lists (NBL) and
supplemented by any additional establishments with one of the
selected SICs that appeared on the purchased DMI list of large
establishments.  The constructed government and military lists
were appended to the purchased establishment list to form the
fuel establishment sampling frame.

     The large establishment sampling frame was purchased from
Dun and Bradstreet's list of business establishments, the Dunn's
Market Identifiers (DMI).  This list source is more expensive
than NBL but was required since it contains the number of
employees for each establishment, which NBL does not.  A list of
all establishments in the sampled counties with 20 or more
employees was purchased.  The establishments on this list with
any of the fuel-related SIC codes were selected from the large
establishment frame and printed out.  They were clerically
compared with the fuel establishment frame, county by county, and
any such establishment not already on the fuel establishment
frame was added to it.

     Table A-9 shows the resulting frame counts by survey region
for these two frames.  The counts show fairly good (by no means
perfect) agreement with the counts in Table A-7, based on CBP and
Versar data.  For large establishments not in fuel-related
industries, the frame count is about 10 percent lower than the
CBP count.  Region 6 (Pacific) shows a higher percent deficit,
about 15 percent, and also the bulk of the amount, 5,000 cases.
For the fuel establishment sample, the total measure of size in
Table A-7 (27,753 establishments) does not include any allowance
for government and military cases, of which there were 3,139 on
the frame.  Subtracting these from the frame total leaves 30,583
establishments, or about 10 percent more than the sampling
measure of size.   Table A-10 shows the frame counts broken down
                              A-23

-------
Table A-9.  Number of establishments on the frames
            for 34 sampled PSUs (unweighted), by
            survey region

Survey
region
1
Northeast
2
Southeast
3
Midwest
4
Central
5
Mountain
6
Pacific
Total
Fuel
establ ishment
frame count

5,403

3,023

3,355

6,027

1,650

14,264
33,722
Large, non-fuel
establ ishment
(> 20 employees)
frame count

8,472

4,811

6,193

13,227

2,698

32,677
68,078
                       A-24

-------
Table A-10,
Number of establishments on the
frames for sampled PSUs (unweighted),
by urban versus rural
    Type
   of PSU
      Fuel
  establishment
   frame count
Large, non-fuel
 establishment
(> 20 employees)
  frame count
    Urban


    Rural



    Total
      33,208


       1,723



      34,931
     66,935



      1,143




     68,078
                       A-25

-------
by urban versus rural PSUs, which agrees well with the breakdown
found in Table A-8.
     B.   Establishment Sample Draw

     As described in Section A-I, above, the fuel establishment
and large establishment samples were drawn separately.

     For the large establishments, a single national self
weighting sample of 600 establishments was drawn.  The frame was
sorted by PSU and by number of employees within PSU.  Each case
was given a measure of size in inverse proportion to the sampling
probability of the PSU it was in.  A systematic sample (based on
a random start) of 600 establishments was drawn using probability
proportional to this measure of size.

     The fuel establishment sample was drawn one region at a time
so that sampling could begin before all frames were completed.
The target number of 800 eligible establishments was allocated to
the six survey regions based on their sampling measure of size.
Based on early results for eligibility rates of government and
gas station establishments, and based on the relative proportion
of the frame in each region that fell into these two categories,
the target number of eligibles was inflated to an allocated
initial sample size for each region. The net result was an
approximate doubling of the sample size.  The detailed figures
appear in Table A-ll.
                              A-26

-------
Table A-ll.  Target sample size, by region, for
             fuel establishment sample
Survey
region
1
Northeast
2
Southeast
3
Midwest
4
Central
5
Mountain
6
Pacific
Total
Target number
of eligible
establ ishment s

213

206

165

95

37

84
800
Allocated
size for
sample draw

449

415

325

194

75

160
1,618
                      A-27

-------
     C.   Eligibility Rates for Fuel and Large Establishment
          Sample

     Once the samples were drawn, they were screened for
eligibility.  Table A-12 shows the initial sample draw and number
of eligible cases, by region, for both samples.  There were
several possible reasons for a sampled establishment being ruled
out of the scope of the survey.  Some establishments were found
to be not actually located in the sampled county (48 cases for
these two samples), out of business (85 cases), or ineligible for
other similar reasons (22 cases).  Six were duplicates of another
sampled listing.  Of establishments found to be in the survey
counties and in business, 97 had only abandoned tanks and 1,084
had no underground storage tanks, or stored only non-motor fuel
substances, leaving 876 eligible establishments.

     Table A-13 shows weighted eligibility rates by type of
establishment for the survey regions and overall.  It shows that
about 80 percent of sampled gas stations were survey-eligible.
Ineligible gas stations were generally out of business.  Eighty
percent of government and military were eligible.  Some had been
mistakenly included on the frame.  Ineligible government cases
were generally out of area or storing non-motor fuel substances.
The other fuel-related industries category shows about one-
quarter eligible.  Here, the out of business rates were lower
than for gas stations, and most ineligible cases had abandoned
tanks or no tanks.  For large establishments the overall
eligibility rate was 13 percent.  Almost all of the ineligibles
in this sample were establishments which simply had no tanks.

     These varying eligibility rates show that although
underground motor fuel storage tanks are concentrated in certain
industries, they occur in establishments in a broad range of
industries.
                              A-28

-------
Table A-12.   Sample eligibility,  by region, unweighted counts of sampled cases

Survey
region
1
Northeast
2
Southeast
3
Midwest
4
Central
5
Mountain
6
Pacific
Total
Fuel establishments

Total
sample draw

447

413

324

193

75

160
1,612
Number of
eligible
establishments

225

197

161

92

42

83
800
Large establishments

Total
sample draw

158

116

142

68

29

87
600
Number of
eligible
establishments

21

18

13

7

4

13
76
        cases were  drawn, but  6 were found to be duplicates during the
  screening process.
                                     A-29

-------
 0)
 w
 0)

4-1
 o

 0)
 ex
 ;>-,
4-1

T)
 C
 cd
 bO
 0>
 W)
• r-l
 a)

 4-1
 0)
 o
 CO
 o-
 w
 0)
•r-l
 bO
•H
T)
 CU
4J
•r-l
 ID
•8
H



























4J
C

tn Xl CD
Cd rH
4J p,
w S
cu cd
CO



4-1
C >->
CU M
a *d cd ^^
2 C 4-> dP
r-l Cd -1-1 	
CU 1-1

0 's




-o
0) W
• 4-> 0)
cd -H
S-i r-l J-i
CU CU 4-1 ^~x
xi s-i w <*>
4J i 3 ^
O r-l T)
(U C
w
C
o
•H
cd s~*
4-) dP
to —
w
cd
O


K^ C
cu o

M 'bO
3 cu
cyj S-i






CO ^O ON ^^ "^ ^^^
r-H i-l iH r-l i-l











CO i — \ F^l  r— i ro r*^ c^
CM r-l CM CM CM CO





rO i — 1 CO O"\ <}" xO
oo oo oo r- co oo




4-> 4-> C
W W 4J i— 1 •!-! O
cd cd c/i cd cd »r-i
CU CU CU S-4 4J 4-1
i— 1 _C CM f! CO S ^ 4J in C *O »i-l
4J 4J *6 C 3 O
j-i 3 *i~i cu o cd
o o a o a PLI






CO
1-1











vj"
m












o
co














-------
IV.  SUBSAMPLE OF ESTABLISHMENTS FOR TANK TIGHTNESS TESTS

     The eligible sampled establishments had approximately 2,000
underground motor fuel storage tanks or manifold systems.  A
subsample was drawn for physical tank testing.  For the survey at
large, the target number of tank tests was 500.  Fifty were set
aside for farms (during the planning stage, it was not known how
many farm tanks would be found), leaving 450 tank tests for the
subsample of fuel-related and large establishments.

     At the time the subsample was drawn, it was assumed that a
manifolded system of two or more tanks connected by piping would
always be physically tested as one unit and therefore would count
as one test.  During the process of doing the testing it was
found that,  in fact, some systems were relatively simple to break
apart for testing, and this was done where possible.  In this
section, we count tanks or manifolded systems; but in the
sections reporting on tightness tests, the counts of individuals
tanks or of separate tests are generally given.

     Table A-14 shows the allocation of the 450 tank tests by
survey region.  This allocation is the estimated number of tanks
or tank systems to be tested for each category; some variation
occurred in the final sample since establishments rather than
tanks were the sampling unit.  For the farms, the number of tank
tests depended on what was found during the interviewing and tank
test scheduling.

     The allocation was made as follows.  Of the 450 tank tests,
40 were allocated to Region 5 to assure a minimum sample size for
that region.  The remaining 410 tank tests were allocated to
Survey Regions 1-4 and 6 in approximately the same proportion as
                              A-31

-------
T!
 0)

•r-l
43

 o
 O

 W
 4-)
 C
 (1)
 nj
 4J
 0)
 bO
•d
 C
 0)

4-1
 W)
 C
•H
 4J
 M
 0)
 4J
 0)
 C
 4J
4:
 ao
•H
 4-)
 tti
 4-1

 5-1
 O
4-1

 to
 4J
 C
 0)
 E
45
 to
 4J
 to
 bO
 C
 to
43
43
 cfl
H









bO
C
•r-l
4-1
CO
CO
CO
01
C

45
bO
•r-l
4J

^
C
cd
4J

}_i
o
4-1
01
r-l
g4
cd
CO
43
rj
CO

















tt)
r-l CO
43 4-1
•H £
bO CU
0> CO
•H
4-1 r-l
0 43
cd
4-1 4-1
W CO
•r-l Ct)
I-J








r-l T! CO
CO CU 4-1
6 r-l (3
4-4  B B
W cfl 43
>-i >i w w
0) B) 43 -H
43 3 r-l
E r« CO 43
3 G cd
^ cd 4-1 4-1
-U cd CO
CU



CO
4-1
C "O
4-i cu a)
o S r-i
45 CX
r-l CO E
a) .H cd
43 r— 1 CO
E 41 43
3 cd £3
IS 4J CO
w
CU



r-l
CO
^-1 3
0) ^.H .H
^ W rH r-H
6 ti) ^o
3 jM cri
Z C 4J 4J
nj td co
4-1 CU



CO
4-1 ^**
4-1 C 4-1 bO
o cu cu o C
r-l E "-I
S-l 43 45 cu r-l
cu -r4 w E ca.
43 bOv-4 -H B
E -H r-l 4J Cd
S r-l 43 CO
la cu cd 4-1 43
4-1 cfl 3
CO v-* CO
CU

r*") M
CU O
> -H
W bO
3 0)
CO r-l







CM r- 1 O CM CO VO
i-l <-l CO LTi  T3 C5 3 O
5-1 ^ -r-l CU O Cd
O O a O 3d PM
a co





o
LO
>^-













00
CT\
rH















0
LO
>J-











r-l
i— 1
i— 1
~
CM












r-
00







rH
cd
4J
O
H


 B
 a)
4-1
 (A
1
 4-1

X)
 0)
T>
 •r-l




 1
  W)
  C
 •r-l
 4J
  to   bO
  a)   C
 4-1  -H
     4-1
  tO   OT
  CO   0)
  0)  4J
  C
 4-1  r-l
 45   «
  bO  O
 •H  •*-!
 4->   W
  o   a.
 4-1
      j-i
  CO   O
 4-1  4-1
  C
  CU  T3
  G   a)
 •H   S-l
 r-l   R)
 42   CX
 Cfl   0)
 4-1   W
 to
 a)   a)
      s-i
 4-i   a)
 o   >

 a;   to
 r-l   6
 a,  a)
 E  4->
 cfl   en
 to   >,
 43   CO

 CO  45
      O
 a)   d

 •g   w
      a>
 bO  S
 C   O
 •H  CO
 s
 CO
 S-l    •
 T)  4-1
     •H
 •d   C
 a   =s
 cd
      a)
 bO  C
 C   O
 •r-l
 4-1   CO
 crj   crj
 O
 O  T3
 r-l   0)
 r-l   4-1
 CTJ   C

 C   O
 M   O
                                                                                   A-32

-------
the fuel establishment sample allocation.  Allocating the sample
in advance permitted us to draw the sample on a region by region
basis as the final eligibility results came in from the field
interview phase of the survey.

     For each region, a sampling frame was created, containing
eligible fuel and large establishments at which tanks were found
(including establishments that refused to be interviewed) .  The
frame construction waited until all cases had reached a final
status and preferably had a known number of tanks or manifolded
systems.  The frame contained the establishment ID, the number of
tanks or manifolded systems, and the establishment sampling
weight.  This list was then sorted by number of tanks, then by
PSU (from ID) , and then by fuel establishment versus large
establishment (also part of ID) .   The weights were cumulated down
the entire list.  The number of facilities to select, M-s , was
based on the allocated number of tanks, N j , and the weighted
average number of tanks per establishment, T-: , as shown in the
following equation:
= N
                j/Tj
     The sampling interval, SIj ,  was the grand total of the
weights divided by M-: (M^ was not rounded) .   The sample was drawn
in systematic fashion, beginning with a random start between 0
and SIj .   The establishments selected in each survey region have
a total number of tanks or manifolded systems close to Nj (see
Table A-14) .   Within each survey region, all underground fuel
storage tanks or manifolded systems have an equal probability of
selection for physical tightness testing.
                              A-33

-------
V.   FINAL SAMPLE WEIGHTS
     A.   Questionnaire Weights for Business and Government
          Establishments
          1.   Other Fuel-Related SICs (Other Than Gas Stations)

     The final questionnaire weights for establishments sampled
with fuel-related SICs other than gas stations were based on a
ratio adjustment of the initial sample weights for all such
screened establishments to 1982 CBP counts of these SICS,
followed by a nonresponse adjustment among the eligible other
fuel-related establishments to account for the few
nonrespondents.   The adjustments were made by survey region.  The
ratio adjustment served to put the initial sample on a known
basis, the number of establishments with one of the fuel-related
SICs in each region.  Then the eligible cases weight up to an
estimate of the number of such establishments with eligible
tanks, by region.  The nonresponse adjustment assures that the
weighted results based on questionnaires received will equal the
estimates based on screening results.
          2.   Gas Stations (SIC 5541)

     The gas stations were weighted in the same way as other
fuel-related SICs.  First, the initial sample was ratio-adjusted
by region to CBP totals for gas stations.  The eligible cases
then weight up to an estimate of the number of gas stations with
eligible tanks, by region.  A nonresponse adjustment again
assures that the weighted results based on questionnaires
received will equal the estimates based on screening.
                              A-34

-------
          3.   Other Industries (Establishments With 20 or
               More Employees)

     The sample sector of establishments with 20 or more
employees in industries not otherwise sampled (the large
establishments) was weighted the same way as the gas stations and
other fuel-related industries.  The CBP totals of establishments
of this size in all but the selected fuel-related SICs (which
include SIC 5541, gas stations) were used for a region by region
ratio adjustment of the initial sample.  The weighted eligible
large establishments then estimate the number of such
establishments with eligible tanks in the country, by region.
Since all eligible large establishments participated in the
interview phase of the survey, no nonresponse adjustment was
needed.

     Table A-15 shows the totals based on 1982 County Business
Patterns data which were used as the fixed totals the initial
sample weights were adjusted to sum to.
          4.   Government Agencies

     No national statistics are currently available to estimate
the number of individual government agencies with underground
motor fuel storage tanks, which is the universe our frame was
built to cover.  Therefore, no ratio adjustments can be made.
Nonresponse adjustments were made to account for the small amount
of nonresponse.
                              A-35

-------
Table A-15.  Known totals from 1982 County Business Patterns data
             base used for ratio adjustment

Survey
region
1
Northeast

2
Southeast
3
Midwest
4
Central
5
Mountain
6
Pacific
Total
Type of establishment

Gas station
(SIC = 5541)

28,212


22,623

27,551

12,473

6,100

13,840
110,799

Other selected
fuel-related
industries

42,173


29,825

37,391

17,786

7,881

18,565
153,621
Large
establ ishments
(> 20 employees)
not in selected
industries

158,320


109,137

131,769

67,150

30,129

84,998
581,503
                                 A-36

-------
     B.   Physical Test Result Weights for Business and
          Government Establishments

     After calculating final questionnaire weights for all
responding establishments as described above, the sampling
weights for establishments chosen for physical testing were
adjusted to sum to the estimated totals for four establishment
types  (government, gas station, other fuel-related, and other
industry) by region.  This adjustment was made by an iterative
rating procedure in which the weights were adjusted first to
regional totals, then to establishment type totals, then
readjusted to regional totals, and so forth, until no further
adjustment was needed.  This took five and a half iterations to
achieve.

     A final adjustment was made for tank test result weights.
If all selected tanks had been tested, the weight for an
individual tank or tank system test would be equal to the
establishment physical test weight.  However, some tanks were not
tested.  Thus a "tank nonresponse" adjustment was made to the
tank/tank system weights to account for the untested tanks.  A
single tank counted once (added its weight) in the count of tanks
selected and once in the count of tanks selected.  A manifolded
tank system which was not tested counted once for each tank in
the count of tanks selected.  A manifolded tank system which was
broken apart and tested as separate tanks also counted once for
each tank in each count.  A manifolded tank system which was
tested as one system counted once for each tank in the count of
tanks selected and once for each tank in the count of tanks
tested.  The ratio of the weighted count of tanks selected to the
weighted count of tanks tested was used to form the final
adjustment to tank weights.  This was done over the sample as a
whole rather than by region.
                              A-37

-------
     C.   Farm Questionnaire and Physical Test Weights

     Due to the distribution of farms within the survey regions
(both overall and in our sample) and the low yield of eligible
farms from the screening, for weighting and any regional analysis
purposes the survey regions have been consolidated into three
areas for farms (see Appendix H).   These are:

     o    East (combines Northeast and Southeast Survey Regions);
     o    Midwest; and
     o    West of the Mississippi (combines Central, Mountain and
          Pacific Survey Regions.

Total counts of farms for these areas were obtained from the 1982
Census of Agriculture and used to form ratio adjustments for
eligible farms.  Due to one refusal among farms, a nonresponse
adjustment was also made.

     Since so few farm tanks were tightness tested, no weighted
estimates will be presented for that data, and hence final
weights were not prepared for physical test results for farm
tanks.
VI.  VARIANCE ESTIMATION
     A.   Jackknife Approach to Variance Estimation

     In a complex survey such as this one, it is difficult or
impossible to estimate the variance of survey estimates directly
from algebraic formulas.  An alternative approach often used, and
                              A-38

-------
adopted for this survey, is the so-called jackknife method of
variance estimation through replication.  The idea behind the
method is to draw a collection of subsets of the sample, called
replicates, and use the subsets to form national estimates of the
statistic whose sampling variance is being estimated.  The
variability among these estimates is used to estimate the
sampling variance of the estimate based on the full sample.  [See
Sampling Techniques. 3rd Edition, W.B. Cochran, J. Wiley & Sons,
1977 for a brief discussion of the principles of this method.]
     B.   Replicate Formation

     To form the replicates, the sampled PSUs were paired and one
PSU dropped from each pair in turn.  Since there were 34 PSUs,
there were 17 pairs and 17 replicates.  The pairs were formed as
follows.  Thirty-four PSUs were drawn in six survey regions.
Except for one certainty PSU in Region 6, they were paired into
strata in straightforward fashion — PSU 1 with 2, PSU 3 with 4,
and so on.  Region 6 required some special consideration.  The
sample in the region consisted of PSUs 29 through 34, with PSU 31
being a certainty PSU.  PSUs 29 and 30 were paired.
Establishments in PSU 31 were separated into "odds" and "evens"
and these sets were treated as a pair of PSUs.  This left PSUs
32, 33, and 34 to consider.  These three PSUs were grouped into
one stratum; PSU 33 was randomly paired with 32, giving the
paired PSUs 3/4's their initial weight; and PSU 34 was given
3/2's its initial weight.   Then either the singleton or the
paired PSUs are randomly selected to be dropped for one
replicate.

     The resulting strata and random selection of which PSU to
drop from each stratum, in turn, to form a replicate (17
replicates in all)  are shown in Table A-16.
                              A-39

-------
Table A-16.  Definition of strata and replicates for
             jackknife estimation of variance
Stratum
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
PSU 1
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31, odds
(32 & 33) (3/4's)
PSU 2
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
3 1 , evens
34 (3/2's)
PSU
to drop
1
4
6
8
9
12
14
16
17
19
22
23
26
27
29
31, odds
(32 & 33)
                            A-40

-------
     C.   Jackknife Replicate Weights and Variance Estimates

     Seventeen replicates were formed by dropping a randomly
selected PSU from each stratum, in turn.  Weights were calculated
for each replicate as follows.  As an initial sampling weight for
the replicate, establishments at the selected PSU of a pair were
assigned twice their initial weight, while establishments in the
dropped PSU were assigned zero.  Establishments in all other PSUs
kept their initial sampling weight.  Then the ratio adjustment to
CBP totals by industry type and region and the nonresponse
adjustment by the same categories were done as described above
for the full sample final weights.  For tank test replicate
weights, the subsampled establishments in the replicate had their
weights adjusted by raking to the replicate total by region and
industry type, and replicate tank test nonresponse adjustments
were made.  Repeating all steps of final weight adjustment in
calculating the replicate weights ensures that the variance
estimates will reflect the impact of weight adjustments on the
variance.

     Subscripting the 17 replicates by r = 1 ..., 17, the
                                 A
variance of a national estimate, X, of a statistic X is given by:
                   ^2   17  A      A 2
                   C  — ^  / V    _ "V\
                   sx - z= (X(r)   x>

      A
where X/r\ is the estimate based on the r— replicate.  The
flexibility of this method of variance estimation can be realized
by noting that the statistic X could be not only a total (such as
number of establishments with tanks) or a proportion (percent of
                              A-41

-------
all tanks that leak) but any statistic that can be estimated from
the full sample and from each replicate in turn.
                               A-42

-------
                            APPENDIX  B
       SURVEY  PROCEDURES  AND ELIGIBILITY  AND RESPONSE  RATES
I.   IN-PERSON INTERVIEW PRETEST

     In July and October of 1984, survey packages (including
introductory letter, questionnaire, general instruction booklets,
and inventory forms) were mailed to a pretest group of 10
establishments which were previously determined to have
underground storage tanks in use.  They were selected through
liaison with local government and military officials rather than
by random sampling or from developed survey listings.  The July
pretest group consisted of seven "fuel-related" establishments
and the October group included three military installations.
Using government-operated establishments in the pretest allowed
us to prepare for problems not normally encountered in non-
government situations.  The purpose of the pretest was to
evaluate the format and wordings of the questions in the
interview for clarity and administerability; to determine the
length of administration time for the interview; and to assess
specific and overall response to the flow of the interview and
individual items in the interview.   In addition, several on-site
procedures were tested including meter testing, tank sticking,
site diagraming and soil sampling.   Several revisions to
materials and adjustments to on-site procedures were made prior
to the field period.  No results from the pretest are included in
the final estimates of the survey.
                              B-l

-------
II.  WESTAT TELEPHONE PRESCREENING AND LIST CONSTRUCTION
     PROCEDURES

     Since lists of establishments with underground motor fuel
storage tanks do not exist, it was necessary to develop
establishment frame lists for each of the 34 PSUs.  As described
in detail in Appendix A, the universe of all establishments with
underground motor fuel storage tanks was divided into three
segments:  the fuel-related establishments, large establishments
(with more than 20 employees), and farms.  Lists of fuel-related
establishments, large establishments and farms were purchased or
obtained for the 34 PSUs in the survey.  Since a list of
government establishments and locations was not available, a
telephone list construction procedure  (described in
Section B-II.B below) was used to construct government tank
establishments lists in the 34 PSUs.  In the 34 PSUs a sample of
1,618 fuel-related establishments (including government and
military establishments), 600 large establishments, and 600 farms
was drawn to be surveyed.  Since eligibility rates were expected
to be low (less than 50%) telephone screening procedures were
implemented using the Westat Research Telephone Center in order
to determine which farms and large establishments were "eligible"
(had underground storage tanks) for the survey.   (Fuel
establishments, including government and military establishments
were screened in the field.)
     A.   Government Tank Establishment List Construction

     Because there is no central listing source for government
establishments with underground motor fuel storage tanks,
federal, state, county, and city lists were developed using
extensive telephone research.  Initial contacts with officials at
                               B-2

-------
different government levels (i.e., state and county Fire
Marshall's Office, Public Works Department, local Police
Department) provided the telephone interviewer with the location
of underground storage tanks or referrals to other contacts who
could furnish information on underground storage tank locations.
Hard-copy listings were accepted by mail if the data was too
extensive to be given over the phone.  After all leads were
exhausted, using a minimum number of calls, and the lists were
determined to be complete.  They were then added in as part of
the fuel-establishment sample frame.
     B.   Farm and Large Establishment Screening

     Using the farm and large establishment sample lists,
telephone interviewers contacted the owner or operator of the
establishment and asked whether the farm or business had any
underground storage tanks in use to store motor fuel.  For those
establishments which were eligible, a contact name was obtained
to assist the field interviewer.  All establishments that could
not be located by phone (19%) or refused the screening interview
(1%), were included in the field screening efforts.  All but two
percent of the farms and large establishment that could not be
located or screened by telephone were located and screened in the
field screening effort.
III. UST SURVEY MAILOUT

     The mailout for the UST Survey began on November 26, 1984
with survey Region 6 (West coast) and continued in phases working
through Region 4 (Southwest), then Region 2 (Southeast), Region 1
                              B-3

-------
(Northeast) and Region 3 (central U.S.)-   (See survey region map
in Figure B-l.)  The last phase of the mailout was completed on
May 3, 1985, with packages being sent to Region 5 (Midwest).
Survey packages were sent certified mail to a sample of 1,965
establishments.  Included in this sample were those farms and
large establishments which could not be located through the
initial Westat telephone screening.  Survey packages were mailed
according to the schedule of the field interviewers, so that the
respondents received the survey materials approximately two weeks
prior to the interviewer's arrival at the site.  The purpose of
the survey mailout package was to allow the respondent time to
prepare for the in-person interview.

     Because the packages were sent certified mail, the date the
package was received and the name of the recipient was available
for the interviewer.  The field interviewer used this information
to trace those establishments which could not be located by
phone.  Each day, certified mail receipt cards returned were
keyed into an automated receipt control system (discussed in
Section 5-V.B).  For survey packages returned by the post office
to Westat, a log was kept indicating establishment identification
numbers and reason for return.  This information was passed on to
the interviewer, who then took responsibility for getting the
survey materials to the respondent.  Eleven percent of the
packages were returned by the post office, and less than one
percent were refused.  However, field interviewers were able to
contact nearly all of the establishments for which the package
was returned.
                               B-4

-------
                                 I

                                 ffl
                                 3
                                 tn
                                 •H
B-5

-------
A.   UST Survey Package



     Every establishment received the same package of survey

materials, which were labeled with the establishment survey I.D.

number, establishment name, and address.  The package consisted

of the following items, which are included as Exhibits in

Appendix F:


     o    Open Letter to Owners and Managers of Underground Motor
          Fuel Storage Tanks — An introductory letter that
          informed respondents of the need and purpose of the
          survey;

     o    "Certification Statement for Establishments without
          Tanks" — A labeled form for the respondent to sign and
          return to Westat if there were no underground motor
          fuel storage tanks located at the establishment;

     o    "Reporting Responsibilities of Tank Owners and
          Operators" letter — A one-page information sheet
          quoting the amended RCRA regulation that requires
          respondents to participate in the study;

     o    General Instruction Booklet — A booklet describing
          procedures for completing the questionnaire and
          inventory forms.  A "Request for Confidential Treatment
          of Business Information" form was included in the
          instruction booklet for the respondent to sign if
          necessary;

     o    Underground Storage Tank Survey Establishments
          Operator's Questionnaire — One labeled copy was
          included to be reviewed by the respondent prior to the
          in-person interview;

     o    Inventory Sheet  for Tanks with Metered Dispensing Pumps
          and Dispenser Meter Recording Sheet — Six labeled
          copies were included in the package so that the
          respondent could begin to keep inventory prior to the
          interview;

     o    Manifolded Tank  System Recording Sheet — One labeled
          copy was included in the package; and

     o    Inventory Sheet  for Tanks without Metered Disposal
          Pumps — One labeled copy was included in the package.
                               B-6

-------
     A toll-free Westat "hot line" number was included in the
introductory letter as well as in the General Instruction Booklet
to provide survey assistance for the respondent.
IV.  FIELD PROCEDURES

     Fieldwork for the UST Survey began December 2, 1984.  A
staff of seven field interviewers was trained to collect data
from the sampled establishments.  Between one and three
interviewers were assigned to cover a PSU depending on the
numbers of establishments sampled per PSU.  The interviewer's
tasks in each PSU included eliminating ineligible establishments
using field screening techniques, and scheduling and conducting
on-site interviews.  These procedures are discussed below in
Section B-IV.A and B-IV.B.  On the average, work in each PSU was
completed in 15 days.  The field phase of the UST survey
concluded on June 29, 1985.  However, data collection efforts
through the mail and by telephone for incomplete cases continued
until November 18, 1985.
     A.   Field Screening

     An interviewer's assignment list for a PSU consisted of a
call record folder for each establishment to be screened and
interviewed.  (See Appendix F for a copy of the UST call record
folder).   These lists included the farm and large establishments
which could not be located through the Westat Telephone Research
Center screening procedure.  As a part of the initial
appointment-making telephone call or visit, the interviewer
determined whether the establishment did indeed have underground
                              B-7

-------
motor fuel storage tanks on site.  In most cases, the interviewer
was able to determine whether or not the establishment was
eligible through an initial phone contact.  Where phone contact
was not possible, the interviewer traveled directly to the
establishment site to speak with the respondent.  Once
eligibility for an establishment was determined, the interviewer
then scheduled appointments for the in-person interview.  Those
establishments that sent the signed "Certification for
Establishments without Tanks" prior to the beginning of fieldwork
in a PSU were taken off the interviewer's assignment lists, and
were not field-screened.
          1.   Statistics on Eligible Establishments

     Table B-l shows the number of establishments which were
sampled, screened, and eligible for the UST Survey.
Approximately three percent of all farms and 13 percent of all
large establishments sampled were eligible for the survey (had
underground motor fuel storage tanks).  Almost 50 percent of all
fuel-related establishments sampled were eligible.  Reasons for
ineligibility are discussed in Section 5-IV.A.2.
                               B-8

-------
Table B-l.  Number of sampled, contacted, and survey-eligible
            establishments, by sample stratum



Number sampled
Number contacted
Number of establish-
ments contacted
that have tanks
("eligibles")


Farms
5981
596



20
(3.4%)

Large
establish-
ments
600
596



76
(12.8%)
Fuel-
related
establish-
ments
1,6122
1,608



800
(49.8%)


Total
2,810
2,800



896
1600 farms were sampled.  Two farms were found to be duplicates
 in the telephone pre-screening.
21,618 fuel establishments were sampled.  Six were found to be
 duplicates in the field screening.
          2.   Statistics on Ineligible Establishments

     When a sampled establishment was determined to be ineligible
for the survey the interviewer assigned an appropriate status
code on the establishment's call record, and notified the Westat
field director.  Table B-2 contains the reasons for ineligibility
and their frequency of occurrence by type of establishment.  The
majority of establishments were found to be ineligible because
they had no tanks.  Approximately 95 percent of all ineligible
farms and large establishments fall under this category.  Among
the fuel-related establishments ineligible, 73 percent had no
underground storage tanks.  All establishments in Regions 1
through 5 found to have no underground motor fuel storage tanks
through field screening procedures were instructed to sign and
                              B-9

-------
 Ul
4J
 C
 0)
 Ul
-H
rH
X)
 rd
4J
 Ul
 (1)

 (U
 en
-H
rH
 0)
 C
-H



 O

 Ul
 o
•rH
4J
 Ul
•rH
-P
 fd
•P
 fN

 CQ

 (U
 rH
 X(
 rd
 EH















rH
C)
rj
[j_|




(1)
tn
f-i
fd
hH














1
X!
fd
•P
Ul
H



1
X!
fd
•P
Ul
w




rH
fd
•P
o
EH




Ul
-P
C
0)
£2
X!
Ul
•H
rH
H
Ul
-P
a
a

c*
Ul
-H
rH
H
Ul




O
O
CO
..
(M






CO
O
VO
>.
rH





VO
cn
in





vo
cn
in







Ul
±>
C
0)
g
X!
Ul
-H
rH
XI
fd
-P
ui -d
a) a)
•P

0
in







0\<>
f^
CO





o\o
t-v
cn



I
Ul -H
-P rH
C H
0)
m -d
0 C

.
o o
55 C
Q

o\<>
CO
*
in


0
H
H

~
•^f
«
H
H
•^.^

CM
cn

0\°
o
.
H


in

0\<>
CO
•
CM
*~*

CO

X!
4J
•H
[3

Ul
4J
C
0)
g
r]
Ul Ul
•H ^
rH C
X( fd
rd 4J
4J
ui -d

^
o

Ul
4->
c
0)
g
X!
Ul
-H
rH
x>
ra

U]
0)

M-i
0

.
o
55
fe

oV>
O
•
in


vo
cn



o\o
CO

cn
**^

in
r-

^.
cn
.
H


O
H
~
O
•
CM
^^

H















Ul
Ul
0)
c
-H
Ul

XI

q_l
o


o\°
CO
•
CM


"*
in



o\o
rH

in
**^>

H
•sj

0\°
co
.
rH


o-

^
0
•
rH
•*-*

VO


4J
^
O

Ul
4J
0)

r\
Ul
•H
rH
X!
fd
4J
Ul
(U

M-i £D
0 W

.
O <*H
55 0
O

^^.
0\°
cn

•^^

vo
H



o\°
CO

rH
s«^

rH
H


o\°
VO
.
V 	

CO


o\o
co
•
v_ ^

CM


4J
J!3
O

Ul 0)
•p >
C M
Q) 3
e ui

Ul (1)
•H X!
rH 4J
X!

4J 0
Ul
(1) 0)
a
 
-------
return a statement certifying their establishment has no tanks
(See Appendix F).   Of the 745 establishments in the survey
Regions 1-5 with no underground motor fuel storage tanks,
82 percent returned the "No Tank" form.   For establishments in
Region 6 with no underground tanks, the interviewer went directly
to the site, observed there were no tanks, and picked up the
signed form from the respondent.  This was a quality control
measure to check the accuracy of the certification.

     Establishments which had abandoned tanks, were out of
business, out of PSU, or out of the scope of the survey accounted
for about 15 percent of the ineligible establishments.

     It should be noted that if an establishment moved from the
site sampled to a different location within the PSU, the
establishment was considered eligible and the interviewer
followed the establishment to the new location to conduct the
interview.  Also,  if the owner of the establishment had sold the
business, the current owner/operator was interviewed.
     B.   On-Site Procedures

     Once at the establishment the interviewer had several types
of data to collect.  On-site procedures included an in-person
interview using the EPA Underground Storage Tank Survey
Questionnaire, a discussion on keeping inventory records,
checking the accuracy of the fuel dispenser meters, making fill-
pipe and drop-tube measurements, preparing or obtaining a site
sketch map, and locating the establishment on topographical maps,
The respondent was to gather the necessary data prior to the
                              B-ll

-------
interview to prepare for the on-site visit as instructed in the
survey package.
          1.   The Call Record Folder

     All information and associated material gathered from the
on-site visit of each establishment were kept in an individually
labeled call record folder (Appendix F) for that establishment.
The call record folder became the case jacket for the
establishment and was preprinted with forms for address and name
updating interview status reporting, contact and call recording,
interview procedures guidelines, and an interviewer debriefing
form.  All materials, such as questionnaires and inventory
information, collected at an establishment were labeled with the
establishment identification number and filed in the
establishment's call record folder.

     For each PSU worked, the interviewer received a package of
pre-labeled call record folders, each call record folder
representing a sampled establishment.  The label placed on the
outside of each call record folder contained the establishment
name, survey I.D. number, mailing address, tank location address,
contact name, contact telephone number, and the county and state
the establishment was located in.  Below this label, in the Label
Verification area, the interviewer noted any changes in the
original information on the label.  These changes were entered
into the automated receipt control system described in
Section B-V.B.  Also on the front of the call record folder, the
interviewer indicated the completion status of each on-site
procedure.  Printed inside each folder was a script the
interviewer followed which led him/her through the interview.
Also printed inside the folder were a  set of debriefing questions
                              B-12

-------
which asked how willing and prepared the respondent was for the
on-site visit.  A record of all calls to the establishment or the
respondent was kept on the back of the folder.  Each call record
folder had additional survey identification labels stapled inside
to be used for labeling any materials or records received during
the interview.
          2.   The Questionnaire

     The questionnaire body is divided into eight sections, with
each section focusing on a particular topic or concern.
               o    Section A;  Establishment Descriptive
                    Information

     Section A has two purposes.  The first purpose of the
section was to describe the type of establishment that was being
interviewed.  (Question Al was an industrial classification, for
example.)  The second purpose of the section was to find and
"screen out" any remaining "out-of-scope" cases.  Question Al had
a screening-out route for bulk fuel plants and private
residences, for example.  (Private residences were completely out
of scope.  Bulk fuel plants were only in scope if they had motor
fuel storage that was non-bulk, for use directly by motor
vehicles.  Private residences and bulk fuel plants were asked to
call the Westat home office for instructions on how to proceed.)

     Question A6 was another screening question.  Naturally,
given the nature of the survey, establishments that did not have
                              B-13

-------
underground motor fuel storage tanks were not to be interviewed.
(Also, in Question A6 any underground storage tanks that were
permanently out of service or that were used only to store non-
motor fuels such as chemicals of heating oil were excluded.)
Question All was used as a lead-in to the Tank Description Sheet
(which is described below) and also asked the respondent to
provide or draw a map of the establishment.  The primary purpose
of the map was to help the field interviewer establish the
location and linkages between the tanks, pumps, and meters at the
establishment.  The tank testing crews also used the map to help
identify the tanks to be tested, as well as to correctly number
the tanks on their data forms.
               o    The Tank Description Sheet

     The Tank Description Sheet is a two-page sheet containing
specific questions about each tank at the establishment.  A total
of 44 items about each tank include questions on the amount of
fuel held in the tank, the materials of construction, year of
installation, safety features, leak history, etc.

     Tank Description Sheet information is used in conjunction
with tank test results in order to learn more about the factors
and features of tanks that are associated with leaking.  The
information from the Tank Description Sheets was also used by the
tank testing crews.  For these reasons it was of great importance
that the tank identification number of the Tank Description Sheet
and the tank identification number on the map and the inventory
were all the same.
                               B-14

-------
               o    Section B;   Operating Practices

     The particular focus of Section B is on the establishment's
typical inventory record-keeping and inventory management
practices.  The interest here is in the establishment-kept
records, in factors associated with the accuracy of those
records, and in the kinds of records that were kept.
               o    Section C;  Operating History

     This section contains questions that fill in the
establishment's past tank history.  The Tank Description Sheets
provide basic historical information about the tanks currently in
use.  In Section C information is obtained on tanks that have
been replaced, removed without being replaced, or abandoned in
place, and in the number, the date and the reason for each of
these three actions.
               o    Section D;  Permits and Licenses

     Section D comprises two questions about permits and licenses
a respondent has to install and operate his tank.
               o    Section E:  Installation

     Section E is a short series of questions about the methods
by which the tanks were installed at the establishment.
                              B-15

-------
               o    Section F;   Protection

     Section F asks about the types of leak-protection,
corrosion-protection, and leak-detection devices that have been
installed at the establishment,  and the kinds of operating and
maintenance practices for the devices.
               o    Section G;  Information Needs and
                    Availability

     Section G focuses on the kinds of information and training
relating to tank operating and monitoring that were available to
the respondent.  Also included were questions which asked the
respondent about types of liability insurance held by the
establishment to cover sudden and non-sudden spills (and leaks)
of motor fuel.

     Interview responses varied depending on how knowledgeable
the respondent was and how willing he/she was to participate.
Often, it was necessary for the interviewer to speak with more
than one respondent to get enough information to complete the
questionnaire.  In some instances, the interviewer was unable to
get any information from the on-site respondents at all.
Operators of establishments owned by multi-establishment
corporate structures occasionally referred the questionnaire to
their home office, which was always off-site and generally
outside the PSU where the interviewer was located.  In these
cases, followup calls from Westat were made to obtain the
                               B-16

-------
completed questionnaire.  Interview response rates are discussed
in Section B-V.F.
          3.    Reviewing the Inventory Sheets

     After completing the interview with the respondent,  the next
step for the interviewer was to review the inventory forms.
Included in the survey package the respondent received were four
kinds of motor fuel inventory sheets,  a schematic diagram of the
seven most common tank and dispenser hookup systems (in the
General Instruction Booklet), and an Inventory Recording Table
(in the General Instruction Booklet) to help him choose the
correct inventory sheets to use for his establishment.  The
respondent should have started keeping inventory on these forms
prior to the interview.  Because of the complexity of the data
being gathered, the interviewer was instructed to always review
the inventory sheets with the person responsible for keeping
them.  This was not always the same respondent who answered the
questionnaire.  Depending on the size and type of establishments,
several people were sometimes involved in keeping the inventory
records.  It was the interviewer's job to make sure the
respondent understood the inventory process and was filling the
forms out correctly.  If the respondent chose to provide 30 days
of previously collected inventory, the interviewer reviewed the
data carefully and made sure all the necessary information was
provided (or that the respondent knew what information was
necessary if previously collected inventory was to be mailed in
from another location, for example, a home office where all
records were kept).

     Before reviewing the inventory forms, the interviewer had to
verify that the tanks and meters were numbered the same on the
                              B-17

-------
map drawn by the respondent in the questionnaire,  in the Tank
Description Sheets, and on all inventory forms.  It was very
important to make sure these numbers corresponded in order to
link data from inventory forms and tank tests to the
questionnaire data.  The interviewer used the Tank to Dispenser
Meter Fuel Line Connections Sheet (Appendix F) to cross-check the
linkage system.  This was done at the actual physical location of
the tanks, where tank and meter numbers were positively
identified.
     After the inventory review, respondents were told that
someone would be contacting them within the next two weeks to
check on the status of the inventory forms.  They were given a
toll free 800 number to call if they had any problems or
questions with the inventory recording procedures.  The
interviewer also gave the respondent a postage-paid pre-addressed
envelope for returning the completed forms.  Inventory response
rates are discussed in Section B-II.F.
          4.   Checking Meter Accuracy

     Once all tank and meter numbers were verified and inventory
sheets reviewed, the interviewer checked the accuracy of all
dispenser meters using a five-gallon certified meter calibration
can.  For each meter tested, a calibration  (adjustment) ratio was
recorded on the appropriate inventory form.  Using this ratio,
the inventory records were adjusted by computer to account for
the meter error.
                             B-18

-------
     The accuracy testing procedure was the same procedure used
by agencies that certify meter accuracy.  The interviewer first
pumped approximately one gallon of fuel into the can to wet the
inside.  This reduced the surface tension inside the can and
allowed for a more accurate measurement.  After wetting the can,
the fuel was returned back into the appropriate tank and the
meter reset to the zero (0.0) reading.  Next, the interviewer
pumped five gallons of fuel into the test can and read the level
of fuel according to the measuring gauge on the front of the can.
The can was used to measure error in liters or gallons.  A
"calibration ratio," which equaled the gauge reading divided by
the amount pumped into the can, was recorded for each meter
tested.  The ratio was recorded in "cubic inches" (in ) if the
fuel was dispensed in gallons or in "milliliters" (ml) if the
fuel was dispensed in liters.  A negative (-) or positive (+)
sign was always recorded with the ratio, to indicate whether the
pump was dispensing less or more than the amount indicated by the
meter.

     After recording the calibration ratio,  the interviewer
returned the fuel to the tank from which it came.  The
calibration of all meters associated with the same tank were
checked before going to the next.  If the respondent had already
started keeping inventory records, the amount of fuel returned to
the tank was recorded as a "delivery" on the inventory sheet, in
order to balance with the meter readings in the inventory
records.
          5.   Measuring the Fill Pipe/Drop Tube

     The next procedure after checking meter accuracy was to
measure the diameter of the tank fill pipe.  The interviewer also
                              B-19

-------
had to determine whether or not a drop tube was present inside
the fill pipe and, if present,  whether the drop tube was
permanent or removable.  This was done for each underground
storage tank and the data recorded on the Site Observations
Recording Sheet (Appendix F).  This information was collected by
the interviewer to help prepare the MRI crew for tank tightness
testing.  Certain factors, such as the size of the fill pipe or
the presence of a permanent drop tube hindered or prevented a
tank test.  Knowing this beforehand, the crew was prepared to
solve the problem once on site for the test.
          6.   Map Reading

     The interviewer was provided with topographical maps of each
PSU, which were included in the package with the call record
folders for establishments to be interviewed.  These are U.S.
Geological Survey maps and are graphic representations of
selected man-made and natural features of the earth's surface
plotted to definite scales.  Such maps record physical
characteristics of the terrain as determined by precise
engineering surveys and assessments.  Using a standard symbol
guide to help read the maps, the interviewer located the tanks on
the map, circled the location, and identified it using the survey
I.D. number for that establishment.  The interviewer returned the
unused maps to EPA.  The maps with tanks located on them were
returned to Westat, where they were reviewed to make sure all
establishments for that PSU were mapped, copied, then sent to
EPA.  Using the precise longitude and latitude of the tanks  from
the map, soil characteristics and other physical characteristics
of the site could be matched to the tanks specific for that
location.  There were fewer than 20 sites for which USGS
topographic maps could not be obtained, and these were covered to
                              B-20

-------
the extent possible by local street or road maps.  The data
obtained through the map linkage are discussed in Appendix H.
     C.   Interviewer Evaluation

     Immediately after leaving the site, the interviewer
completed the debriefing questions printed inside the call
record.  These eight questions were used to evaluate the overall
character of the interview and the cooperation and knowledge of
the respondents.  Table B-3 shows the debriefing statistics for
the 890 establishments surveyed.
                               B-21

-------
Table B-3.   Debriefing statistics
                                                         Percent
Percent of respondents who had
  questionnaire completed prior to interview               28%

Percent of respondents who had
inventory sheets started                                   12%

Percent of respondents who had problems
  or errors in completed parts of inventory                31%

Percent of respondents who
  understood inventory process                             98%

Percent of respondents who understood
  most/all questions in questionnaire                      99%

Percent of respondents who were cooperative                94%

Percent of respondents who were hostile                     3%

Percent of respondents who were guessing
  a lot in answering interviewer's questions                4%

Percent of establishments where it was
  necessary  to talk to more than one
  person to obtain all required information                29%
     Less than one-third of the respondents had prepared for the
on-site interview by completing the questionnaire prior to the
interviewer's arrival on site.  Only 12 percent had started

keeping inventory records prior to the interview.  Of those

respondents who had started keeping inventory records, the

interviewers found that 31 percent had errors in the completed

parts of the inventory.  Almost 100 percent of the respondents

understood the inventory process and the questions in the

questionnaire.  In approximately 30 percent of all cases it was

necessary to talk to more than one respondent to obtain all
                              B-22

-------
required information.  Even though most respondents were
unprepared for the survey, 94 percent were willing to cooperate.

     After completing the debriefing questions, the interviewer
made necessary name and address changes to the label in the Label
Verification section of the call record.  If it was necessary to
talk to more than one respondent, a contact name and phone number
for each respondent interviewed was written on the front of the
call record.  The interviewer then assigned a questionnaire
completion status for the case and circled the appropriate
completion status codes for the inventory record keeping, the
meter accuracy test, the site mapping, the debriefing, and the
confidentiality request form.  After checking to make sure that
all materials in the call record were properly labeled and
editing the questionnaire for completeness, the interviewer
returned the completed case to Westat, where it was reviewed and
entered into the receipt control system (discussed in Section
B-V) .
     D.   Refusals

     Each sampled establishment received a survey package
containing a copy of the Resource Conservation and Recovery Act
(RCRA) amendments to Section 9005(a) stating that the
responsibility of the tank owners and/or operators to furnish
information for the UST Survey.  Nevertheless, a small number of
respondents still refused to participate.  When an interviewer
encountered a refusal to participate either over the phone or in
person, he/she told the respondent that the EPA legal office
would be informed of the refusal.  The interviewer then contacted
the Westat field director immediately.  The field director
notified EPA's Office of Enforcement and Compliance Monitoring of
                              B-23

-------
the refusal by phone and by letter.   In most cases,  the
respondent agreed to participate after a phone call  from an EPA
attorney.  In other cases, a warning letter from the Waste
Enforcement Division was sent to the respondent when a phone call
did not result in cooperation.

     Some respondents refused to participate in any  part of the
interview, while others only refused to keep the inventory
records.  The number of interview and inventory final refusals is
shown in Table B-4, lines F and J respectively.  Overall, less
than one percent of respondents refused to complete  the interview
and less than two percent refused to complete the inventory
recordkeeping.

     When a respondent who had initially refused the interview
decided to participate (either as a result of a phone call or
enforcement letter) the Westat field director was notified.  If
the field interviewer was still on site in that PSU, an interview
was set up with the respondent.  If the interviewer had already
left the PSU, the person assigned to "clean-up" (see
Section B-IV.E) these special cases made the appointment and
completed the interview.
     E.   Interview "Clean-Up"

     It was necessary to use a "clean-up" interviewer who
followed behind the field teams, to handle special circumstances
when all on-site procedures could not be completed during the
                              B-24

-------
 01
 o
-H
-P
 01
•H
4J
 (d
-P
 01

 01
 d
-P
 (d
-P
 01
 o
-p
 C
 0)
TJ
 (3
 «d
•rl
-H
 0)
-p

•H

TJ
H
 0)
-H
Pn
OQ

 a)
H
43
 a)
EH

H
(0
-P
O
EH
01
1 -P
43 C
H (d Q)
0) -P g
d W 43
Pn W 01
-H
H
01
1 -P
0) 43 C
CT> (d Q)
M-P g
. >.
og CNJ

CN
og co o in
HO O CTi
vo vo co r^
«. -.
H H


O VO VO VD
o CTI r^ r^
vo in
H
CO VD O CTl
CTl CTl Og H
in m





•O 0)
0) rH
•P 43
O -H
^
n vo
.
CTI
CTl


0\°
•^ m
•
CTl
CTl


oY>
0 0
O
H
o\o
in H
CTl





J5
0)
•H 0)
> H
M 43
(d tP Q) -H
C P Q) H
O 01 t/1 0)
0 Q) C
H O  0) -P
Q) -H M C
g H (U
43 0) IS 0
MS 0) M
C 01 -H
•H H H

in t-»



01
0)
'd
1
^ o a) 0)
> d -P H
QJ -H 0) 43
-rJ s^.^ ^j .j
> ft IT
Q) 0) O H
4-> W O Q)
d d
•H O H i ft O
(DM M
 01 -H
d H H
•H flj 0)
01
TJ d <«-i
d) IH o
d i 
n CTI VD
•
H
o\°
o •<* n
n

W
Q)
^- 01
>i d Q)
MO H
0 ft 43
-P W 43 -H
d Q) O W
Q) M -H 01
t> 43 O
d >, is ft
•H M g
0 M-rl
-d -P o

O Q) M
O > 01 (d
a) d -P
M -H d 01
Q) -P
o -P g d
-P C 43 0)
Q) 01 g
-H «" -H QJO-H MM O Ti CUM O M T! d-HQ)
•OH > ft >Q) -P d ft -P 0) 0) O1 H M
0)4301 M— 0 M ftO d  O 0)
(d 01
M M M 43 M d
0) 0) 0) Q) O
43 43 43 -P 43 D
g g g (0 g 01
d d d 43 d ID
£5 J2 g -P S Pn
< « O Q W
43 0) —
IS -P
d a)
01 -H -p
•p  Q) 0)
d -P -P
01 -H Q) (d
-P rH M
d M-I ft
Q) O g <1)
T3 O 01
d M o d
o o o
ft 43 43 O
01 g -p W
0) d O Q)
M S 43 K
M H
43 0) —•

d 0)
01 -H .p
-P (d
fH <| I t.
Q) O
T) H
d M (d
O Q) W
ft 43 d
01 g -t
43 0 0 M
IS O
jj t [ i i
01 QJ a) d
•P 43 43 0)
d g g >
a) d d d

>-3 S
0)
d
o
43
ft
0)
0)
-P

H
d
01 Q)
g Q)
(d o

-------
     time the interview team was working in a particular PSU.

Some of these special circumstances included the following:


     o    The respondent most knowledgeable of the underground
          storage tanks was unavailable during the time the
          original interviewer was in the PSU.

     o    The respondent had refused to participate and decided
          to participate after the original interviewer left the
          PSU.

     o    The business was closed due to seasonal operation when
          interviews were being conducted in the PSU.

     o    The establishment was remodeling its underground
          storage tank systems and could not provide all
          necessary data at the time interviews were being
          conducted.

     o    A calibration check could not be done due to adverse
          weather conditions or seasonal operation of the
          establishment.

     o    An establishment could not be located by the original
          interviewer.
     Work done by the "clean-up" interviewer accounted for five

percent of all completed interviews.
     F.   Field Interview Data Collection Statistics



     Table B-4 contains data collection statistics for the field

interview portion of the survey.  It covers statistics on

interview and inventory response and refusal rates.
                              B-26

-------
          1.   Interview Response Rate

     The interview response rate for this mandatory survey is
nearly 100 percent overall, as well as for each sample segment.
Out of 2,800 establishments contacted, 896 had underground motor
fuel storage tanks, and were therefore eligible for the survey.
Of those, 890 or 99.3 percent completed interviews.  The highest
response rate among the sample segments was among the large
establishments, where 100 percent of the eligible establishments
provided interview data.
          2.   Inventory Response Rate

     Nearly 78 percent of the eligible establishments have
furnished complete or partial inventory data.  Even this low
response rate was achieved only after extensive edit and followup
efforts by Westat's survey staff.  Sixteen percent of the
eligible establishments have not yet provided inventory records.
It was impossible for 4.5 percent of the eligible establishments
to keep inventory records.  These reasons are discussed below in
Section B-IV.F.3.
          3.   Problems Preventing Inventory Record Keeping

     Of the 896 eligible establishments, 40 were unable to
provide inventory records for any of their tanks using the
designated record keeping procedures.  The reasons are listed
below.
                              B-27

-------
          No conversion chart — Twelve establishments were
          unable to obtain conversion charts needed to convert
          inches to gallons for their tanks because they did not
          know the dimensions of the tanks or the company which
          installed them.

          Bent fill pipes  — Nine establishments were unable to
          stick their tanks because the fill pipes were installed
          with a bend to prevent pilferage.

          Facility closed  — Seven establishments have closed
          down since the time of the interview.

          Tanks abandoned/removed — Five establishments have
          either removed or abandoned their tanks since the time
          of the interview.

          No inactive period — Inventory analysis procedures for
          tanks without meters to record the total product
          dispensed consists of an analysis of volume measurement
          changes for inactive periods.  Four establishments,
          which have tanks without meters, dispense fuel 24 hours
          a day, so there  is no inactive period to analyze.

          No way to record deliveries — Two establishments
          pumped fuel at irregular intervals from an above-ground
          tank into the underground storage tanks with no means
          of measuring the amount pumped into the tanks.

          No key to tank — The locked tank of one establishment
          was inaccessible due to delay caused by probation of
          the estate of the tank operator, who died with the only
          key in his possession.
V.   DATA PREPARATION


     Data preparation for the UST Survey began with a development

phase involving questionnaire layout and code manual design.

Inventory recording forms were developed by EPA.  The coding

format, however, was designed by Westat.  Operational phases

included document handling (including receipt control),
coding/editing, data entry, and machine editing.  Location coding

from the topographic maps is discussed in Appendix H.
                               B-28

-------
     A.   Questionnaire and Code Manual Design

     The questionnaire layout was designed for ease of data
preparation/data processing, as well as for ease of respondent
understanding and recording.  Many items were designed as
"preceded" questions, that asked the respondent to answer by
circling a code to indicate his/her response.  This eliminated
the need for a coder to translate check-marks or other non-code
symbols into coded answers.  Computer field positions were
printed in the questionnaire for most data items.  These field
positions were useful as reference locations for coders, machine
editors, and data entry staff.

     A detailed code specification manual using an automated code
book formatting program for the UST Survey Questionnaire was
developed.  This manual described the data to be encoded from the
questionnaire, item by item.  Figure B-2 lists the item
characteristics by which each data item was described in the code
manual.  Figure B-3 is an example data item description from the
Underground Storage Tank Survey Establishment Operators
Questionnaire.
                              B-2 9

-------
          Item characteristics described  in code manuals
a.   Field  position and record  number

b.   Item name (the name by which the item was called in all
     computer programs and other documentation)

c.   Quotation of the item from the questionnaire

d.   List of all code values and their definitions

e.   List of reasons for legitimate item nonresponse (the
     "inapplicable" definition)

f.   List of all missing value  codes

g.   Flags  indicating logical relationships between the item and
     subsequent items.


                 Figure B-2.  Item characteristics
                     described in code manuals
 F2A       020-022      EBEtUEGCI-CE.KISEECIIQtj

                     ++           * INAPPLICABLE, COOED 2r 8, CR 9 IN F1»,
                                   CCL 016, REC 09;  OR CODED 1, 8 OR 9 IK
                                   BOXF2, COL 19, 
-------
     B.   Inventory Data Editing

     Inventory forms were designed to include "worksheets" for
respondents to record individual meter and manifolded tank
readings, and then to record the sums of the individual readings.
Both the individual readings and the summary readings were edited
and key punched.  The raw data collected from the inventory
recording process was entered or key punched (Section B-V.D)
directly from the edited inventory forms.  A code manual and
editing instructions detailing the layout and the valid code
ranges for the inventory forms was prepared to assist the editors
and the data entry operators.
     C.   Receipt Control

     All returns were tracked by Westat's automated receipt
control system.  Each day, documents received, including
certified mail cards and "No Tank" Certification statements, were
keyed into the system.  All documents from an establishment were
linked by a survey identification number specific to that
establishment  (discussed in Section B-V.B.l).  Using this I.D.
system, returns were tracked by type of document, and reports on
the survey status and on an individual establishment status were
produced.

     For each document received, the date of receipt, a status
code and "batch" number (Section B-V.B.2) were entered into the
receipt control system using the procedure specific for that
document.  In addition, any name or address changes from the call
record were also entered upon receipt of a questionnaire from the
field.
                              B-31

-------
          1.   Survey Identification Number

     The survey I.D. number is a ten-digit number which shows the
sampling frame from which the establishment was selected, the PSU
in which it is located, and a sequential number.  The survey I.D.
uniquely identifies the establishment within the survey and links
all documents and data records for the establishment.
          2.    Questionnaire and Inventory "Batching"

     Questionnaires and inventory forms were "batched" into
groups of 10 documents for coding, editing, and filing purposes.
Each batch was given a number, which was written at the top of
the Batch Control Sheet (Figure B-4), as well as on the
questionnaire or inventory form.  Questionnaires and inventories
were batched separately.  Listed on the Batch Control Sheet were
the survey I.D. numbers of all the questionnaires (or
inventories)  and their statuses for that specific batch.
Questionnaire and inventories remained in their batches until
they were coded and sent to data entry.  If they were removed
from the batch for any reason, the date, person taking the
document, and reason were noted on the front of the Batch Control
Sheet.  A copy of each Batch Control Sheet was kept in a log for
quality control purposes for both questionnaire and inventory
batches.
                             B-32

-------
BATCH CONTROL SHEET

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.

ID LABEL











CODER:
DATE:


ri<
]
STATUS










CHECK OUT
TO/ON










BATCH
I I I I I


DATE
RETURNED










VERIFIED
BY











VERIFIER:
DATE:
% VERIFIED:

jure B-4
B-33

-------
     D.   Coding/Editing

     A staff of six coder-editors was trained to code the
questionnaire and inventory.   The initial training session
covered procedural matters as well as specific coding of the UST
Survey Operators Questionnaire and the four types of inventory
recording forms.  It included an item-by-item discussion of the
coding of the documents, practice coding examples, and group
review of the coding of practice examples.  Training materials
included code manuals, practice inventory and questionnaire
examples, and a marked-up version of the questionnaire that
linked the questionnaire to the code manual and the general
coding instructions.

     Coders were trained to edit questionnaire responses and
inventory records for consistency and completeness as they were
coding them.  Coders flagged any problems they discovered during
coding, and referred the problem documents to the coding
supervisors.  Some problems required the development of new codes
— such as when different units of measure than those specified
in the questionnaire were specified for quantity questions.
Other problems required that the respondent be called to verify a.
response or provide missing information  (a process called "data
retrieval").  In some instances, decisions could be made based on
the evidence available, by the Project Officer or by other EPA
staff.  Decisions, both general and case-specific, were recorded
in a Decision Log for future reference.

     All coding was 100 percent sight verified by a senior coder
or the coding supervisor prior to being sent for data entry.
                              B-34

-------
     E.   Data Entry

     Questionnaire and inventory data were entered ("key
punched") by highly trained data entry operators, using a key-to-
tape entry system.  This key-to-tape system is computer-driven
and provides a formatted entry keying program that minimizes many
types of data entry errors.  All data entry was 100 percent key
verified by a different operator from the entry operator.

     The questionnaire booklets and inventory records were sent
to data entry in "key batches."  Lists of the survey I.D. numbers
associated with each key batch of inventory records were made and
put into a Key Batch Control Log.  All questionnaires sent to
data entry were checked off against a list of completed
interviews, which was generated by the receipt control system.
This enabled the coders to make sure that all questionnaires were
keyed.
     F.   Machine Editing

     Machine editing is a means of data quality control that uses
a computer program to test item ranges, skip patterns, and
logical consistencies in a data file.  Such a machine edit
program was prepared for the questionnaire and for the inventory
forms.

     Machine editors were selected from among the trained coders
available from the coding staff.  The training consisted of
procedural instructions, and a walk-through using an example edit
problem.
                               B-35

-------
     The machine edit programs provided a list of test errors for
each edited case, as well as a listing of each case in error.
Each of the errors was checked, and often the hard copy of the
case was reviewed.  Updates to the data files were written on
update sheets, key-entered and run against the data file to
produce a new master file.  Then the edit cycle was rerun to make
sure that the update corrections had been made correctly.
Because of the complexity of some of the data files (particularly
inventory data files), it was necessary to rerun edit cycles
several times:  updates to some fields tended to unexpectedly
impact consistencies with other fields.

     After the final machine edit cycles, frequency distributions
for all items of the data files were reviewed by supervisors to
spot problems not captured by the machine edit programs.
VI.  DATA RETRIEVAL

     Data retrieval is the term used to refer to recontacting
respondents for the purpose of verifying or clarifying responses
to completed questionnaires for interviews.  For this study, it
was necessary to recontact respondents for problems found in the
inventory records as well as questionnaires.  These questionnaire
and inventory data retrieval procedures are discussed separately
below in Section B-VI.F.l and B-VI.F.2.  Part of the coding staff
was trained for the data retrieval process.
                              B-36

-------
     A.   Questionnaire Data Retrieval

     Recontact of respondents for questionnaire problems
generally took the form of a telephone call, though occasionally
it was necessary to mail a list of questions to a respondent.
Approximately 60 percent of all respondents were recontacted for
questionnaire data retrieval.
     B.   Inventory Data Retrieval

     Because of the complexity of the inventory record-keeping
procedure, each respondent received a "prompt" call by Westat
approximately two weeks after the field interviewer left the
site.  The purpose of the call, which was made by a staff member
trained for inventory data retrieval, was to inquire about he
status of the inventory and when the records would be completed.
The prompt caller also assisted the respondent with any questions
or problems that may have occurred about keeping the inventory.

     A large proportion of the inventory records received from
the respondents contained errors or inconsistencies ranging from
minor to major.  When these problems were spotted by coder-
editors or coder-verifiers, the inventory form was flagged for
inventory data retrieval.  The inventory data retrieval process
began with a phone call to the respondent with a discussion of
the problem. Some problems were resolved on the telephone.
Often, an explanatory letter and copies of the returned inventory
with problem areas marked were sent to the respondent.  The
respondent then sent corrected inventory records back.  It was
sometimes necessary to send multiple letters explaining the
problem before usable data was returned.  Of the 697 inventory
                             B-37

-------
responses received to date, approximately 85 percent needed data
retrieval, four percent of which needed multiple data retrieval
efforts.  At the writing of this report, there are still
establishments which have not yet responded to the data retrieval
efforts.  They account for 25 percent of all cases needing data
retrieval.
     C.   Followup of Inventory Nonrespondents

     After multiple prompt calls were made to inventory
nonrespondents, EPA sent a formal warning letter (Figure B-5) and
Status Report form (Figure B-6) to respondents who were
delinquent in returning inventory records.  Of the 300 letters
sent, 25 percent did not respond and two percent refused.

     As a result of all data retrieval efforts made by Westat and
EPA, 78 percent of all establishments have sent in inventory
records, but approximately 50 percent of all inventory records
received are complete enough for inventory reconciliation
analysis.  Of the 896 eligible respondents, 16 percent have not
yet returned inventory records.
                              B-38

-------
            UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                        WASHINGTON, D.C. 20460
CERTIFIED MAIL
RETURN RECEIPT  REQUESTED
                                                          OFFICE or;'
                                                   PESTICIDES AND TOXIC'SUBSTANCES
Dear

     The Environmental  Protection  Agency  (EPA)  has  been  informed
by Westat, the Agency's contractor for  the  National Survey of
Underground Motor Fuel  Storage Tanks, that  as of  July 31,  1985
the 30 days of motor fuel  inventory data  you are  required  to
provide EPA had not been received.

     As was explained in the  survey instructions  mailed  earlier,
Congress passed and President Reagan  signed  into  law in  1984,
amendments to the Resource Conservation and  Recovery Act (42
U.S.C., Sec. 6901) that require  EPA to  conduct  this study.  This
law also requires that  you, as an  owner or  operator of an
underground motor fuel  storage tank,  provide EPA  with the
information requested in this survey.

     I wish to stress that the evaluation of inventory data is an
essential part of this  National  study,  and  EPA  is requiring this
information from all establishments selected for  the survey.
Failure to comply with  this requirement may  result  in an
enforcement action.

     Enclosed is a form for reporting the status  of your 30-day
inventory data collection.  We ask that you  complete and return
the form within 24 hours of receipt to  verify that  you are
complying with this requirement.   Simply  check  and  complete the
correct inventory status block,  sign  and  date the form,  and nail
it in the enclosed self-return envelope.
     Thank you for your cooperation.

                                  Sincerely
                                  Martin P. Halp&r,  Director
                                  Exposure Evaluation  Division
                            Figure B-5

                                ri-39

-------
       UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
               MOTOR  FUEL  INVENTORY  STATUS  REPORT
         Please complete this form and mail in self-return
         envelope wihtin 24 hours.
         I have completed and mailed my 30-day inventory data to
         Westat.

         I am still collecting my 30-day inventory data and v/ill
         mail it to Westat by 	,
                              (date)

         I have not yet begun my 30-day inventory data collection
         but will do so immediately and mail it to Westat
         by 	,
            (date)

         I need further instructions to complete and submit my
         30-day .inventory data collection.*

         Other situation (please describe).	
                              (Signature)
                                  (Date)
*?hor.e toll free (800) 638-8985
                            Figure B-6
                                B-40

-------
                           APPENDIX C

                DEVELOPMENT OF A TANK TEST METHOD
     This appendix is a summary of the report,  "Development of a
Tank Test Method for a National Survey of Underground Storage
Tanks."  The work was conducted under EPA Contract No. 68-02-
3938, Work Assignment No. 25.^

     The appendix first summarizes the search for a suitable
tightness testing method and the reasons for the final selection.
Then the field procedures developed in the pilot test are
described.  A more detailed description of the field tightness
test plan may be found in the test and analysis plan.2
I.   SELECTING A METHOD

     In preparation for the field tightness testing, MRI first
searched for a suitable test method.  Their objectives were to
evaluate potential test methods to be used for the national
survey, to conduct a pilot survey using the test method selected,
and to develop a test plan for the national survey.  The research
was conducted in five stages.  The first stage consisted of a
1|fDevelopment of a Tank Test Method for a National Survey of
 Underground Storage Tanks," H.K. Wilcox, J.D. Flora, C.L. Haile,
 M.J. Gabriel, and J.W. Maresca, April 1986.
2"Test and Analysis Plan for the Tank Testing Program of the
 National Survey of Underground Storage Tanks," H.K. Wilcox, J.W.
 Maresca, Jr., J.D. Flora, C.L. Haile, June 10, 1985.

-------
review of current methodology for detecting leaks in underground
tanks.  Second, field observations were made of several methods
in use.  Third, of the several methods observed, five were
selected to be evaluated by conducting tests of these methods on
a single tank system at a closed service station.  Three of these
five methods were selected for further evaluation in the fourth
stage by testing tank systems at four military installations and
at an operating service station.  In the final stage, the method
chosen for use in the national survey program was tested in a
pilot study of 17 tank systems.
II.  GENERAL METHOD SELECTION CRITERIA

     The main criteria used to select a method for the national
program were:
     1.   Quantitative measurements were desired.  However, this
          did not preclude consideration of other approaches.
     2.   A detection level of 0.05 gal/h as established by the
          National Fire Protection Association, Inc., was taken
          as the target detection limit.
     3.   Minimal disruption to the station operation was
          considered to be important.
     4.   The method and equipment had to be rugged for use on
          the national survey.
     5.   The test should be applicable in a wide variety of tank
          system configurations.
     6.   The method should allow a reliable assessment of
          accuracy, precision, and sensitivity.
     7.   Costs for testing and data analysis had to be within
          the available budget.
     8.   Sufficient equipment and manpower to conduct the
          national survey were required.
                              C-2

-------
     The scope of the method selection research and pilot study
did not permit exhaustive method evaluation of all available test
methods in order to select a procedure with optimum
characteristics for all criteria.  Hence, some compromise was
necessary to proceed expeditiously with the survey.
III.  PRELIMINARY REVIEW AND TESTING

     The methods reviewed in the first stage are shown in Table
1. Those for which further evaluations were conducted are also
indicated.  The methods were classified into groups according to
their measurement characteristics.

     Five methods were selected for further testing at a closed
service station in Kansas City.  Brief descriptions of each are
provided below.  A more complete review of tank testing methods
can be found in EPA's report.3
          The ARCO method utilizes a photo optical sensor to
          monitor the level of a partially filled tank.  If the
          test conditions are set up properly, the device is self
          compensating for temperature changes.  Only the portion
          of the tank containing the product is tested.
          The Certi-Tec method uses pressure transducers which
          are located just below the surface of the liquid to
          measure level changes.  Seven thermistors are used to
          measure temperature at various levels in the tank
          during testing.  The tank is overfilled during the test
          by adding an extension to the fill pipe.  Both the tank
          and lines are tested at the same time.
3"Underground Tank Leak Detection Methods:  A State of the Art
 Review,11 EPA/600/2-86/001, January 1986.
                              C-3

-------
                  Table 1.   Leak Detection Methods  Reviewed
Detection method
Liter-
ature
review
Field
site
visits
Prelim-
i nary
testing
Devel-
opment
study
Pilot
study
Volumetric
  ARCO tank test                X
  Certi-Tec test                X
  Ethyl Tank Sentry             X
  Ezy-Chek                      X
  Heath Petro-Tite tank         X
    and line testing system
  Hydrostatic (standpipe)       X
    testing
  Lasar interferometry          X
  Leak Lokator test             X
  Mooney tank leak detector     X
  Pald-2 leak detector          X
  Pneumatic testing (air        X
    test method)
          X
          X
           X
           X
Non volumetric

  Dye method
  Vacutect method
  Helium leak detection
    method
  Tracer Research

Inventory monitoring

  Manual methods
  Automated
 X
 X
 X
 X
 X
External monitoring

 Pollulert
 Remote  infrared sensing
 Ground  water and soil
   core  samples
 Underground radar
X
X
X
X

X
                                      C-4

-------
          Leak Lokator uses a buoyancy probe to monitor level
          with a single thermistor located at the midpoint of the
          tank.  The method can be used to test a partially
          filled tank (with lower sensitivity)  or an overfilled
          tank.  Either the tank or the tank and lines can be
          tested.

          The Petro-Tite method monitors level visually in an
          extended fill pipe.  The product level is returned to
          the reference level at 15 minute intervals during the
          test.  The product is stirred continuously during the
          test to achieve a uniform temperature.  Temperature is
          monitored with a single thermistor located at the inlet
          to the pump near the top of the tank.  The tank and
          lines are all tested at the same time.

          The Varian helium leak detection method, a
          nonvolumetric method, is based on the detection of
          helium outside a tank which has been slightly
          pressurized with helium.  The tank should be empty
          during the test if the entire tank is to be tested.  It
          is also helpful to drill a number of small holes in the
          surface above the tank to assist in the location of the
          leak.  Pressure can be monitored simultaneously to
          provide a quantitative estimate of the leak rate.  The
          lines are also tested at the same time.
     A.   Experimental Procedures


     Each method was tested over a 2- to 3-day period.   A leak
simulation system was designed and fabricated by MRI and used to
draw product from the tank at a known rate.   The precision of the
leak simulator was at least an order of magnitude better than
that of the test methods.  In testing the tank, the objective of
each test group was to estimate different simulated leak rates.
The leak rates measured by each method were compared with the
rates used in the simulation.


     The data from the quantitative tank tests were analyzed to
determine the precision and accuracy of the tests.  For these
analyses the accuracy of the test (or bias)  was estimated by the
                              C-5

-------
mean of the (signed) differences between the leak rate reported
and the leak rate simulated.  A paired t-test was used to test
the hypothesis that the method was unbiased; that is,  that the
mean signed difference was 0.  A linear regression of the
reported leak rate on the simulated leak rate was calculated.  An
ideal regression equation in a tight tank would be y = 0 + l.Ox.
The scatter of the data about the regression line (correlation
coefficient, R) was used as an estimate of the precision of the
method.  The bias and precision were combined to obtain an
estimate of the root mean squared (RMS) error.
     B.   Results

     A summary of the statistical analysis for the quantitative
methods as a group is presented in Table 2.
          1.   ARCO Underground

     The ARCO method was used for 15 different simulated leak
rates, including one zero rate.  An average difference of 0.01
gal/h was observed between the rates reported by ARCO and those
calculated by MRI.  This estimated bias in the results was not
significantly different from 0 (t = 0.21, 14 degrees of freedom).
The intercept did not differ significantly from 0 and the slope
did not differ significantly from 1.  The R for the regression
was 94.3 percent, indicating that most of the variability of the
data was explained by the regression.  The RMS error estimated
for the method under the conditions of the Kansas City test was
0.05 gal/h.  The tests averaged just under an hour  (55.7 min) in
length.  In order to reduce the variability estimated with the
method, either repeated determinations or a longer test time
would be needed.
                               C-6

-------
     Table 2.   Summary of Statistical  Analyses  of Quantitative  Methods
                        Tested at Kansas  City Site
Method
ARCO
Certi-Tec
Leak
Lokator
Petro-Tite
na
15
12
22
18
Bias
0.01
-0.25
-0.01
0.05
Intercept
0.005
-0.30
-0.01
0.06
Slope
0.95
0.71
0.94
1.05
Standard
error
0.049
0.166
0.020
0.101
RMS
0.050
0.302
0.021
0.113
R2
94.3%
38.9%
98.9%
75.9%
n = number of simulated leaks.
                                    C-7

-------
          2.   Certi-Tec Method

     The Certi-Tec method was used for 12 simulated leak rates,
of which two were set at 0 and so represented the normal
condition of a tank test.  The leak rates reported by the Certi-
Tec method took slightly over an hour (average 64.3 min) for each
rate.  The estimated bias in the results (difference between the
reported rate and the simulated leak rate)  averaged -0.25 gal/h.
This bias was quite large and was significantly different from 0
(t = -5.23, 11 degrees of freedom).  The intercept differs from 0
at the 5 percent significance level and the slope differs from 1
at the 5 percent significance level as well.   The standard error
of the regression was 0.167 gal/h.  The R of the regression was
only 38.9 percent, indicating that slightly less than 40 percent
of the variability in the reported leak rates was explained by
the simulated leak rates used in the test.

     Thus this method, as implemented during this test, appears
to have substantial bias and relatively low precision.  Even
though taking several repeated determinations of the leak rates
and averaging them would reduce the random error, the bias would
remain a problem.
          3.   Leak Lokator Method

     The Leak Lokator method was used on 22 tests simulating leak
rates.  Of these, three were zero simulated leak rates and so
represented tests of the tank without any simulated leak.  Three
simulated leaks into the tank were also used.  Using the method,
the average reported leak rate was 10.8 min.
                               C-i

-------
     The bias in the determinations was estimated to be -0.005
gal/h, which was not significantly different from 0 (t = -0.23,
21 degrees of freedom).   Although the estimated slope and
intercept agree closely with the ideal, both differed from the
ideal values significantly at the 5 percent level although not at
the 2 percent level.

     These data showed a very small scatter about the regression
line, resulting in small estimated values for the standard error
of the slope, intercept, and regression.  These small standard
errors led to the borderline significances of the difference
between the regression parameters and their theoretical values.
In light of the nonsignificance of the other test for the bias
and the small magnitude of both the intercept (-0.012 gal/h) and
the bias (-0.005 gal/h), these are probably not of major
importance.
          4.  Petro-Tite Method

     The Petro-Tite method was tested under 18 simulated leak
rates, of which three were zero rates, corresponding to a tight
tank situation.  While the usual Petro-Tite test takes an average
of four leak rates each reported over a 15-min period, only five
of these determinations were based on an hour's data.  The
remaining leak rates reported were each based on a 30-min test.

     From all the tests, the bias was estimated at -0.05 gal/h
but was smaller (0.040 gal/h) when restricted to the hour-long
tests.  The bias from the complete set of tests is significantly
different from 0 at the 5 percent level but not at the 1 percent
level.  If attention is restricted to the 1-h tests, the  bias is
not significantly different from 0.  The intercept is not
significantly different from 0, suggesting that the bias is not
                              C-9

-------
statistically significant.   The slope does not differ
significantly from the ideal or theoretical value of 1 at the 5
percent significance level.   The R for the regression was 75.9
percent and the standard error of the regression was 0.101.  This
standard error is interpreted as the precision of a single leak
rate determination.  It should be noted that the normal test with
four 15-min rate determinations should be somewhat more precise
than what was reported, and that precision could be improved
further by testing for a longer period of time and averaging more
individual leak rates reported.
          5.  Helium Detection Method

     Two tests were conducted using the helium detection method.
In the first test the tank was tested in its original state.
Several large leaks were discovered during the first day's
testing, which were repaired. The next day's test revealed
substantial reduction in helium loss.

     While some helium was detected around the tank, the amounts
were generally very small and could have come from pipe fittings
or the tank bungs.  Low levels were, however, encountered in one
area.  The concrete was removed for inspection purposes to see if
a line was located in that area. None was found, but helium
levels in the excavation were moderate.

     The basic problem encountered using the helium detection
method is that helium can escape in measurable quantities through
threaded connections which have been poorly coated with sealer.
Gasoline will not normally pass through these poorly sealed
connections at measurable rates under normal operating
conditions.  This can lead to results which are hard to
interpret.  In addition, no quantitative results can be produced.
                              C-10

-------
     C.  Conclusions

     As a result of the preliminary testing in Kansas City, the
ARCO, Leak Lokator and Petro-Tite methods were selected for
further evaluation. The helium method was dropped because of the
decision that a quantitative method presented a better option for
the national survey.  The Certi-Tec method was dropped because of
the prototype state of development and its relatively lower
performance.
IV.  DEVELOPMENT STUDY TESTING

     A.  Experimental Procedures

     Five facilities were selected by the EPA for tank testing.
A total of 13 tanks were tested.  The initial plan was for each
tank to be tested by all three methods.  Difficulties in
scheduling and plumbing problems at some sites, however,
precluded a complete round of testing.

     Two types of tests were conducted at each sites:  baseline
tests which were conducted in the same manner as if no
evaluations were being conducted, and leak simulation tests which
consisted of measuring leaks under a  variety of simulated leak
rates (usually four).   The process was nearly identical to that
described for the preliminary testing.

     Three sets of data from the development study were analyzed:
baseline test data; leak simulations; and time series analysis of
the ambient volume fluctuations after the simulated leaks were
removed.
                              C-ll

-------
     The baseline data for each method was tabulated and compared
for each tank where more than one method was used to test the
same tank.  Where differing conclusions regarding the tightness
of the tank were obtained, the data and conditions of the test
were further examined in an effort to resolve the conflict.

     The data from the leak simulations were analyzed by fitting
a linear regression to the data from each tank and method by
regressing the reported leak rate on the simulated leak rate.
The intercept of this regression represents an estimate of the
leak rate of the tank or tanks system when there is no simulated
leak.  The difference between the intercept of the regression
line and the test result from the baseline test provides an
estimate of bias or accuracy of the test.  The variability of the
data about the regression line provides an estimate of the
precision of the test.  Combining these two measures yields an
estimate of the mean square (or root mean square error)
associated with the testing method.

     The third analysis consisted of a time series analysis of
the ambient volume fluctuations after the simulated leaks were
removed.
     B.  Results

          1.  ARCO Method

     The ARCO method was used to test seven tanks during the
development study.  Of these seven tanks, one tank had only the
baseline test run, one tank test resulted in the baseline test
and one simulated leak rate, and the other five tank tests all
had the baseline leak rate and several simulated leak rate tests,
                               C-12

-------
     The baseline test results for ARCO are summarized in
Table 3.  The ARCO result disagreed with the conclusion for three
tanks.  However, it must be noted that the ARCO system tested
tanks approximately 75 percent full, under no additional head
pressure.  Thus, the ARCO system provides a test most
representative of the usual operating conditions of the tank.  If
a tank system has a hole in or near the top or fill pipe, or if
there is a leak in the lines, this would not result in product
leaking under normal operating conditions.  While it may be
unlikely that all of the leaks encountered during the study are
in the top of the tank, it is a possible explanation.

     A summary of the results from the leak simulation tests
using the ARCO method are summarized in Table 4.  By this method
of testing, none of the tanks tested were reported to be leaking.
However, other test methods gave different results for some
tanks.

     The data indicate, however, that the ARCO test method
performed well at the Damneck and Pitstop North test locations.
If a single data point that appears to be an outlier is removed,
the method also does reasonably well at the Langley facility.
One of the sites (Scott Tank 18) showed essentially no regression
of the reported leak rates on the simulated leak rates.  This is
disturbing because for that test the method could not quantify
leak rates under the simulation.  One other test, at Fort Lewis,
gave a slope substantially different from 1, which indicates that
an (unknown) interfering factor is present.

     The ARCO method gave a precise determination of a leak rate
under some operating conditions.  In other cases, it failed to
give valid results for reasons that were not understood.  In
other cases, it failed to give valid results for reasons that
                              C-13

-------
        Table 3.  Summary of Baseline Results and Tank Tests Attempted
Facility and
tank
Damneck
Pits top
1 (south)
2 (north)
Scott
1 (17)
2 (18)
Ft. Lewis
1 (8C25
north)
2 (8C25
south)
3 (4194)
ARCO
+0.02 Ca
+0.02C C
0.0 C
Out of time
+0.02C C
-0.04 C
0.0C C
~
Leak Lokator
-0.077 Nb
-0.741 N
(Poor sensi-
tivity)
-0.012 C
-0.299 N
-0.178 N
Problem,
possibly mani-
folded
Leak about
gasket-could
not test
-0.027 C
-0.172 N
Petro-Tite
+0.003
-2.892
-0.05
+0.004
-0.812
—
-0.342
-3.0
C
N
C
C
N

N
N
MRI
conclusion
Tight
Leak
Tight
Tight6
Leak6
Tight
Leak
Leak
  4 (10E10)

Langley
(Poor sensi-
tivity)
-0.191d  N
-0.024   C
Tight
1 (HS tank 3)
2 (HS tank 5)
3 (MoGas)
4 (Golf
course)
Physical problem
with tank
-0.03 C
•*•"
-0.448 N
-3.0d N
—
•••
Leak
Leak
Tight
-2.540 N Leak
^Certifiable.
 Noncertifiable.
^Test OK, but leak (possibly in upper part or piping) not found.
 Test appeared OK, but data are inconsistent.
Interactive effects  between Tanks  17 and 18 were observed by Leak Lokator
 -  (negative sigh)  indicates leak out.
    indicates testing was  not conducted at that tank by the test company
   indicated.
                                       C-14

-------
        Table 4.   Results of Leak Simulation Tests Using ARCO  Method
Tank
Damneck
Pitstop south
north
Scott 18
Fort Lewis
8C25 south3
8C25 north
Langley
MoGas
Baseline
rate
0.02
0.02
0.0
0.02

0.0
-0.04

-0.03h
-0.03°
Intercept
-0.023
-
-0.092
-0.145

-0.005
-0.094

-0.336
-0.027
Bias
-0.003
-
-0.092
-0.165

-0.005
-0.054

-0.306
0.003
Slope
1.049
-
0.809
-0.044

1.140
0.493

0.419
1.167
SE
0.022
-
0.041
0.099

0.047

0.367
0.118
RMS
0.023
-
0.101
0.192

0.072

0.478
0.118
Negative = Leak out
Positive = Leak in
Bias = Intercept - base

?Two points only.
 Outlier removed.
                                     C-15

-------
were not understood.  It can detect inflow or outflow,  but would
be defeated if the water table were at a level that approximately
balances the hydrostatic pressure of the product.   It is also
subject to interference from wind and is sensitive to vibration.
It has the advantage of not requiring an overfilled tank, but
this is counterbalanced by the disadvantage of not being able to
detect potential leaks in the upper quarter of the tank.

     The ARCO method was not recommended for use on the national
survey program for several reasons.  The primary reason was the
decision to test the entire tank.  Secondary reasons were the
sensitivity of the method to interference from vibration and the
relatively high frequency of tests that did not adequately
quantify the simulated leak rates.
          2.  Leak Lokator Method

     The Leak Lokator method was used to test 10 tanks during the
development study.  Of these, two tanks had only baseline tests
and no simulated leak tests conducted.  The baseline test results
are summarized in Table 5.   The Leak Lokator test conclusions
agreed with MRI's conclusion in 6 of the 10 tank tests.  Of the
other four, the Leak Lokator test failed to certify three tanks
that had been concluded to be tight based on data from all test
methods and certified one tank that had been determined to be
leaking.

     A summary of the results from the leak simulation tests
using the Leak Lokator method is presented in Table 5.  The RMS
errors ranged from about 0.02 gal/h to 0.44 gal/h.  The standard
errors ranged from 0.015 to 0.304.  Among the tanks judged to be
tight, the standard errors ranged from 0.015 to 0.165 and the RMS
error ranged from 0.021 to 0.437.  The large values for the upper
                               C-16

-------
    Table 5.   Results of Leak Simulation Tests Using Leak Lokator Method
Tank
Damneck
(@ 120")
Pitstop south
north
Scott 17
18°
Fort Lewis
8C25 south
4194
10E10 NTCa
TCb
Lahgl ey
HS 3
HS 5
Baseline rate
-0.0775 @ 125
(+0.008 @ 118)
-0.524
-0.012
-0.299
-0.178

-0.027
-0.171
-0.191
-0.191

-0.448
-3 or more
Intercept
-0.0825
(-0.005)
-
-0.026
-0.366
-

-0.010
-0.159
-0.596
0.069

-0.641
0.126
Bias
-0.005
(-0.13)
-
-0.014
-0.067
-

0.017
0.013
0.405
0.260

-0.193
0.126
Slope
0.786
-
0.879
0.839
-

0.734
0.749
0.541
0.835

-1.78
2.43
SE
0.025
0.209
0.015
0.048
0.047

0.097
0.026
0.165
0.098

0.048
0.304
RMS
0.0255
(0.028)
-
0.021
0.082
-

0.099
0.029
0.437
0.278

0.199
0.329
Negative = Leak out
Positive = Leak in
Bias = Intercept of their (adjusted for base) regression
Intercept = Bias plus base

. NTC - not temperature corrected.
 TC - temperature corrected by Leak Lokator.
«
 Leak Lokator observed interactive  effects between Tanks #17 and #18 during
 the testing of #18.   The reasons for  this are not understood.
                                     C-17

-------
end of the range are from a test that had problems.   If that data
point is excluded, the upper end of the ranges becomes 0.048 and
0.082.  With the ability of Leak Lokator to obtain multiple leak
rate determinations fairly rapidly (about one every 10 to 15
min), one could presumably reduce these error estimates by making
several leak rate determinations at a tank and averaging them.

     The Leak Lokator method gave valid estimates of leak rates
in most cases.  The variability of a single leak rate measurement
tends to be somewhat large relative to a 0.05 gal/h criterion,
but the ability of the system to obtain leak rate determinations
in about 10 min once the test is running would allow multiple
determinations and averaging to reduce this variability.  The
method has the advantage that its level monitoring system can be
used at any desired level (head pressure).  Thus, if line leaks
are a problem, the testing could, in principle, be conducted
using a level below the piping to determine the location of the
leak.

     The hydrostatic pressure from a water table could pose a
problem for this test.  Testing did not appear to be standardized
to any specific product level.  Since the leak rate through a
given aperture would change with head pressure, testing different
tanks at different levels makes leak rate determinations
difficult to compare and quantify.
          3.  Petro-Tite Method

     The Petro-Tite method was used to test nine tanks during the
development study.  The locations of these tank systems and
reported leak rates were given in Table 3.  Three of the systems
tested had leak rates so large (in excess of 5 gal/h) that
simulation of additional leak rates on the order of 0.2 gal/h was
                              C-18

-------
not feasible.  Simulated leak rate testing was performed on five
tank systems.

     The baseline tests conducted by Petro-Tite agree with the
conclusions reached by MRI based on analysis of all the data.  It
should be noted that in some cases (e.g.  Ft. Lewis #1) where
other testers experienced difficulties, Petro-Tite would have
also had difficulty.

     A summary of the results from the leak simulation tests
using the Petro-Tite method is presented in Table 6.  The RMS
errors ranged from 0.036 to 0.193 for tanks judged to be tight.
The 0.193 is rather large, but that tank posed special problems,
leading to the conclusion that the 0.193 is not representative.
Error estimates on tanks judged to be leaking were larger,
ranging up to 0.24 gal/h.  Larger errors are to be expected for
systems with large leaks because large leaks make it difficult to
maintain product level and so therefore to obtain an accurate
volume.  However, the errors remained acceptably low relative to
the associated leak rates.

     As a result of the more detailed analysis of Petro-Tite
data, several suggestions for improving the errors involved in
the Petro-Tite method were developed.  None of these involve
significant procedural changes.  Improved algorithms could likely
result in better test results.

     The Petro-Tite method seems capable of identifying and
successfully dealing with many types of interferences in tank
testing. Although there are situations that can lead to invalid
test results, for the tanks tested in this study all tests but
one were believed to be valid.  However,  difficulties were
encountered that increased the error associated with the
estimated leak rates beyond that which is desirable.  In

                              C-19

-------
     Table 6.   Results  of  Leak Simulation Tests Using Petro-Tite Method
Tank
Damneck
Pitstop south
north
Scott 17
18
Fort Lewis
8C25 south
4194
10E10
Lang ley
golf course
Baseline
rate
+0.003
-2.89
+0.050
+0.004
-0.812

-0.342
-3.0
-0.024
-2.54
Intercept Bias
-0.009 -0.012
-
+0.069 +0.019
+0.002 -0.002
-0.774 0.038

(Could not
-0.038 -0.014
(Could not
Slope
1.01
-
1.26
1.075
0.608

fill tank)
1.50
keep filled)
SE
0.052
0.240
0.078
0.036
0.109

0.107
0.193
-
RMS
0.054
-
0.075
0.036
0.115

0.193
-
Negative = Leak out
Positive = Leak in
Bias = Intercept - base
                                     C-20

-------
difficult cases, the error rates were such that one could not
reliably detect leak rates as small as 0.05 gal/h.  Most of the
situations with large error estimates were cases where a
substantial leak was present, and hence the loss in precision did
not interfere with the detection of the leak.
          4.   Time Series Analysis of Ambient Noise Data

     Because the data obtained from ARCO was not sufficient, time
series analyses were performed only on the Leak Lokator and
Petro-Tite data.
               a.   Description of Ambient Noise Analysis

     The second analytical approach was to remove the simulated
leaks from the data to produce volume, temperature, and
temperature compensated volume time series that were longer than
normally used during a tank test.  These data were analyzed to
determine whether the results obtained during a standard tank
test period (i.e., a baseline test) were consistent with longer
test times and to determine whether the temperature-estimated
volume changes required for compensation adequately accounted for
the total volume changes in a non-leaking tank.
                    Petro-Tite Method

     Continuous time series of the change in volume and the
change in temperature (converted to volume using the product
volume and the coefficient of thermal expansion) for an entire
day of Petro-Tite testing were generated from the data collected
every 15 min by subtracting the simulated leak volume from the

                              C-21

-------
measured volume. The volume change used for this 15-min interval
was an average of the volume changes observed before and after
this period.  Cumulative time series of volume, temperature, and
temperature-compensated volume were then generated for analysis.
The temperature-compensated time series were generated by
subtracting the temperature (expressed in volume)  from the
measured volume on a point-by-point basis.  This is the same
method used by Petro-Tite.  A least squares line was then fit to
each of the three time series to estimate the mean rate of change
of volume, temperature, and temperature-compensated volume.  The
temperature-compensated volume was compared to the baseline test
results.  The standard Petro-Tite data analysis method was used
to estimate the temperature-compensated volume rate for the
baseline tests  (i.e., sum of the temperature-compensated volume
computed for four 15-min periods).
                    Leak Lokator Method

     Time series of the cumulative volume and cumulative
temperature were generated for each simulated leak sequence of
the Leak Lokator data.  Each time series ranged from a total of
40 min to over 100 min and included four to nine of the standard
Leak Lokator volume rate measure periods.  The simulated leak
rate was subtracted from the uncompensated volume rate
measurements made by Leak Lokator and converted to volume using
the reported measurement time.  These volume measurements were
then summed to obtain the cumulative volume time series.  The
mean volume rate for each simulated leak sequence was taken from
the Leak Lokator data sheets.  A continuous time series of
temperature was generated each day of testing from annotated
readings of temperature made every 5 to 10 min and placed on the
strip chart of temperature by Leak Lokator personnel.  Those
sections of the temperature time series which bracketed the
                               C-22

-------
volume data for each simulated leak sequence were used in the
analysis to compensate for temperature.  The temperature data was
converted to a volume time series and a least squares line was
fit to the data to estimate the average rate of change of volume
caused by the rate of change of temperature over an hour.  A mean
temperature-compensated volume rate was then computed for each
simulated leak period by subtracting the mean rate of change of
temperature from the mean rate of change of volume and compared
to the results from the baseline test and the other simulated
leak test sequences.
               b.   Petro-Tite Ambient Noise Analysis Results

          A summary of the mean and 95 percent confidence
intervals on the mean volume rate, temperature rate, and
temperature-compensated volume rate estimated from the long
Petro-Tite time series is presented Table 7.  The rates were
obtained by fitting a least squares line to each time series.
The confidence intervals are based on the standard deviation of
the ordinate about the regression line.  The site, tank number,
number of 15 min data points, and the test result using Petro-
Tite 's 0.05 gal/h detection criterion are also given.  For
comparison, the baseline test result is added to the table.
Agreement between the baseline test results and the long time
series results is good, except for Pitstop Tank No. 2.  The time
series from the Fort Lewis Tank No. 4 indicate that a potential
leak began several hours after the test had begun.

     The time series of volume, temperature, and temperature-
compensated volume were generated by removing the simulated leaks
from the Petro-Tite volume time series.  The time series are 3 to
6 times longer than the standard 1 h Petro-Tite test.  The first
hour of each time series contains the baseline data.  Several
                              C-23

-------





























in
•r-
r—
ro
c
*-
O

^J
i-
ro
E
E
3
CO


.
^^

OI
2}
HJ
r-






































U"j
4-*
r— •
3
(/>
o>
S-

•p
Ol

OI
c
in
Ol
in
ro
CO



1 C
OI O
s- •!-
3 «P
•P ro
ro 
S- C
01 a/









1
oi -o
i. 01
3 -P
-(-> ro
ro m
s- c
O> O)
Q. O.

Ol O
(— U












0>
4_>
ro •• — •>
s- x:
^v^
O» r—
E ro
3 cn
Q. Q-r-^
E E
Ol O
H- u





01
S-
3
•4_)
ro
s-
01
OL
E
0)
r—









































o
>






a.
+j
ro ' — -
i- -C.
^\
O) r-
s ro
^ cn
r— N^
O
>





ai
+J
ro s~~*°
i_ _j .
\%
Ol i—
E ra
3 a
(— V^/
O
^>



























U1
+J
-p I—
in 3
O> in
r- 
s- x:
^\
O) i—
E ro
3 O>

o
>

4_)
r—
3
QJ
£-

+„>
in
cu
H-

t— *
o

£js£
LO
cn






IX




HH
O

>^
in
cn






IX



i— •^
C«3

o^
Lf)
CT>






IX





2:



•/*
c
ro
1


LZ
o
•r~
+J
ro
u
o
r:\





4^
P7
CD
•I"
'"




ro
O
0
«
O









4-*
c~
O^
H^


OO
O
o

o



rH
CM
o
,
0
1



00
o
o

o



^J-
l£>
0
.
O



00
0
o
.
o



00
«g-
0

o



0
C\J




r~H






^£
u
0)
c
£
ro
Q
3
o
cn
c
• f—
x/
ro
Ol
	 i



CM
^«
oo
O

o




+J
3
O
en
c
•r—
VX
ro
Ol
	 i


o
in
o

o



rH
r*-^
oo
.
O
1



00
oo
o

o



^T
00
0
.
o



tl£j
(Y)
o
.
o



t "-
CO
CXJ
,
o
1


J^





CM



0
• f—
s
Ol
— 1

[ ^
S-
o
u_




^J
x:
cn
•r—
r—




*3"
CM
O

O
1



O 4->
c: 3
•r- O
cn cn
c c.
•p *r~ *r~
_tZ -^ XX
cn ro ro
•r- a> 01
i— —i —i


f^* P"*** if)
CM rH OO
O rH «3-

O 0 0



CM •* 00
«*• r^- cn
O 0 O

0 O 0
1 1



m ro o
OJ OO ^f
O rH O

O O O



P** rH t*O
rH in O
000
» * •
0 O O
1 1


CM ^n r^*-
CM 
•t —
3
OJ
	 1

+J
S-
o
u.




•P
x:
cn
*r—
r—





m
0
.
o
i




c
•*—
Ol
c
•f-
sx
ro
at


rH
rH
O
.
0



OO
OO
rH
.
O




cn
o
o

o



00
rH
O
.
O



o
rH
CD
,
0



rH
in
rH
,
0



cn
rH




CM






r*
o
^>
in

•r-
CL.




^_>
x:
cn
•1—
t—




«^
o
0
.
o









) >
x:
cn



LO
0
o

0



10
o
o

o
1



^~
rH
O

O



0
cn
rH
.
o



/••"x
rH
O

O



^-
00
rH
.
0



«=*•
CM




rH




aa

^r^

+j
4J
O
U

3
O
cn
c
•r—
'w/
ro
ai
— i




CM
on
oo
,
CM
l



+j
3
O
cn
c.
•r—
^*
ro
O)
— i


cn
00
o
.
0



OO
r^.
p^x
,
CM
1



rH
lO
O

O



cn

CM

o



cn
CO
rH

O



OO
Ol
*^-
.
CM
1


CO





rH






a.
o
j ^
in
-P
• i —
a.
3
O
cn
c
>i«
x/
ro
o>





CM
rH
00

O
1



+j
3
O
cn
c
•f»
v/
ro
a>


00
CM
o

0



CM
CM

.
o
1



IT)
0
O

o'



cn
o
o
,
0



«^-
Csi
0

o



*3~
rH
r***»
.
o



vO
rH




^•>j




CO
u_

-P
0
o




































,
o
i.
01
N

O
S-
tt-.

1 ^
c
ro
u
•r-
4-
•p-
C
cn
•f*
U)
>>
P—
fO
U
*r—
^J
IO
•r-
^j
m
4^
(/>

4-^
0
c
. l
-r-*
in . •-
0) -P >>
4-^ t/) r™—
a> c
ai -P o
x:
-p ai 3:
x: o
S- -P r—
o t-
x: o H-
o
l-O >^Z!
CM C
• IT) O
CM . ••-
OO +J
•P O
u) +J Ol
J_ in S-
•r- ro T-
U 	 Id
(0 XJ O
C-24

-------
observations about the strengths and weaknesses of the method can
be made from the data.

     First, the time series for Damneck Tank No. 1 and for Scott
Air Force Base Tank No. 1, tanks declared to be tight,
illustrates the high correlation between the low frequency trends
of the temperature and volume data required for temperature
compensation.  This suggests that the method of temperature
compensation, circulation of the product and measurement of the
rate of change of temperature with one temperature sensor, is a
reasonable approach.

     Second, negative, high-frequency correlations were observed
between the temperature and temperature-compensated volume rate
time series for some of the tests.  This suggests that the method
is overcompensating for temperature effects.  These high-
frequency temperature fluctuations are probably caused by
inadequate resolution of the Petro-Tite temperature sensor.  This
increase in the high-frequency fluctuations in the temperature-
compensated volume data can be a problem if the test time is too
short.

     Third, inspection of the temperature-compensated volume rate
time series for each test suggests that a one-hour test is too
short to reliably detect small leaks.  Within a test,
fluctuations with period of 30 to 90 min are observed which are
sufficiently different from the low frequency trend exhibited by
the entire time series.

     Fourth, the time series for the tests conducted on Fort
Lewis Tank No. 2, Scott Air Force Base Tank No. 2, and Pitstop
Tank No. 1 indicate that the tanks are leaking.  The measured
temperature changes are too small to account for measured volume
changes.
                               C-25

-------
               c.   Leak Lokator Ambient Noise Analysis Results

     A summary of the mean and 95 percent confidence intervals on
the mean volume rate, temperature rate, and temperature-
compensated volume is presented in Table 8.  The site,  tank
number, duration of the test sequence, the number of Leak Lokator
volume rate measurementss in the test sequence, and the test
result based on Leak Lokator's 0.05 gal/hr criterion are also
given.  For comparison, the baseline test results are also shown.
Several observations about the data presented in Table 8 are
noteworthy.  First, the test sequence results for each tank
tested are internally inconsistent.  The results from five of the
six tanks tested could be declared tight or leaking depending on
which data sequence was used.   The results of the other tank test
(Ft. Lewis, Tank #3) indicate that the tank is leaking but cannot
determine whether the flow is into or out of the tank.   Second,
temperature, volume, and temperature-compensated volume rate data
exhibit a large range of variability compared to 0.05 gal/hr.
The high variability in the temperature compensated volume rate
suggests that the test time is too short and a single thermistor
is not adequate for measuring the mean temperature change in the
tank.  These conclusions are based on the raw Leak Lokator data
and an analysis similar to that used by Leak Lokator except  (1)
an average of four to nine standard Leak Lokator volume rate
measurements were used instead of one and  (2) the average rate of
changes of temperature over one hour was determined by fitting a
least squares line to 5 to 10 temperature values over the hour
instead of the two end points.  The uncertainty in the Leak
Lokator temperature-compensated volume rate results presented in
Table 8 is about a factor of five smaller than the uncertainty of
a single 10 min volume rate measurement and a two-point
temperature rate measurement.
                              C-26

-------














in
•r-
in
>,

"TO
c


~a

.c
Ol
•r—
QJ
:s

c
•r—
E
O
rH
1

S_
O
4->
m
_*:
0
_i

j^
Ifl
*U
QJ
	 1
l»-
0
>>
s_
re
E
E
3
i/>

CO

01

J2
 O 4->
S- T- It} ,— X ^5
3 •*-> s- .C in
+j ro \ 01
to in o> •—
s- c e to
O) QJ 3 Ol
D- Q-r— \^
E £ 0
QJ O >
1— U |X




t — t
OJ QJ O
S- 4J
3 rO ^— s &§
-P s- _c in
fO \ 01
i- O> r—
o> e 
1 X




H*H

CO *-^ %%
J- jz in
\ 01
(U i—
E 03
3 0)
r— *^**
O
1 NX
I *>


zz


1 — , — \
rt3 QJ C
+-» £ "-
o -r- e
r- 4-> ^— '
Ol
u
4-> C
ul O)
(U 3
i— o-
o>
in
-^
c
m
r-

c
o
• r-
+J
ItJ
u
o
— 1

-(->
3 C
O >r-

Ol O)
c c
"r— -#-> -i-i 'r— 4-^
^t. jc _c -i<: -C
10 O> Ol n3 Cin
O> -i- -r- CD -r-
-J t— H- — i t—
oo co t^ in
O rH CM pH
oooo
« • • »
oooo



o P-. *j- m
^- CO 00 O
oooo

oooo
1 1



r-- rH in rH
0 i-H 0 rH
oooo
• * * •
oooo



rH CM rH CO
rH in «cf  4->
3 3
O O

Ol O)
C C
i i - . i i
•*-* *r- •*— 4-*
j: ^; ^. jr
Ol fO no O)
•r- O) ID -r-
1— -J -J I—
i-H 0 CM
l£> O CO
O rH 0
* • •
OOO



CO O 00
Ol •=*• rH
rH i-H O
. • •
OOO
1 1



Ol UD l£>
en 01 CM
ooo
• * •
ooo



in O> r-H
rH CTi O
CM o o
• • •
ooo



^ CO O
rH CM CM
OOO
• * *
000



r-. r- «=*• 01
CM rH  -,-+->
3 ^, 3 C
O O O -r-

Ol O) O) O)
c c c c
12 .5 _i<; ^£
IO (D 03 rO
QJ QJ 0) QJ
_ 1 -J — I —I
CO CM VD
rH CO ^t
000
• • •
000



co oo CM
r-^ co i£>
rH O O
• * •
ooo
1 1



<5t O rH
o I-H «r
ooo
• • •
ooo



CO rH *t-
«3- O O
O O rH
• • *
OOO
1 1


CM i-H O
rH CO CM
OOO
• • •
OOO



CM O 01 CM
r-. co oo 
s-
o
U-
+J 4-> 4-1 +_) 4-) 4J +J
3 CC 333333
O-r-T- OOOOOO

O) O)O> Ol Cn CD O) O) Ol
C CC CCCCCC
^L jc ^. ^. ^:j=£:^:^ ^. ^ ^ ^ ^.
nj o> fO 10 O) 01 o> O) it} TO 1x3 co ro nj
 *r~ *t~ *r— *r- ID O) QJ Q) Q> C)
_l r- _ J _ 1 r- 1— t— 1— — J -J_J_J_I_I
VDOI.O inoioim orH^o
U3 r- CO CO r^ 1^ «*• 00 rH CM «^-
ooo oooo oooo
• «• •••• !•••
ooo oooo oooo



OOOOrH r>.COCOOO rHCOCO«^-
Kl-COlD r-HOOUD OOlCOrH
OrHrH OOOrH OrHCMCM
*tl •••• •••*
OOO OOOO OOOO
III 1 1 1 1



rH«^-o cooooom Lnoo«3-«*-
•^•i-Hi-H COr^t^^- P-.OCMrH
ooo oooo oooo
• •• •••• ••••
ooo oooo oooo



rHrHCO OltOf^iH CMCMOOOO
mcoo r^omcM ujcooo
CMOO OOOC-J CMOOO
• 4* •*•* >«•*
ooo oooo oooo
III II


cMoo«3- oimr^-in cnoomi^-
miDco oooo CMOOCO
ooo oooo oooo
• •• 4*»* t««*
OOO OOOO OOOO



rHCOr^CO CMU3CO«v|-CO OlcrjinrHvD
oioinm rHoiminLfi OICMCM^O
i-HCMrHrH OOOOO CMCOCMCMCM
OOOO OOOOO OOOOO
II 1 1 1 1 1 1

i£>m r--*+-i£>in in UD m in



VOrHrH 00<-fv- in«tl£><£>
m'«a-«4- ^f«4-«i-in m-^-inco

QJ aj a>
c c c
•r— »r- 'I—
r— i-HCMPT r— i-H CM f) -5J- i— rH CM CO «*•
aj QJ 
(ft {& fO
co oa oa

 -!->
+J in •»->
s- *-> o
o •>- u
u- a. i/)
C-27

-------
     The time series plots of temperature (converted to volume)
and uncompensated volume were generated for each of the 21
sequences of Leak Lokator data.   These cumulative time series
plots illustrate the reasons for the inconsistent test results
and the high variability.  The volume and temperature time
series, and the least squares line fit to the temperature data
are presented in the report, "Development of a Tank Method for a
National Survey of Underground Storage Tanks."

     Some difficulty is evident in using a two-point analysis
approach.  Depending on which two points are taken, a positive,
nearly zero, or negative slope can be determined because of the
large fluctuations in temperature.
     C.   Recommendations for the National Survey Testing

     The findings of the development study have resulted in
several recommendations concerning the method of tank testing to
be used in the national survey program.  These recommendations
are summarized below.
          The tank testing method should include putting a head
          of pressure on the tank.  There are two reasons for
          this.  First, proper compensation for water table
          effects are necessary if the proper conclusion is to be
          reached under high water table conditions.  Second,
          this process enhances the flow of product through small
          holes, making them more likely to be detected,
          particularly if they are near the top of the tank.
4See Footnote 1.

                               C-28

-------
          The tank test method should provide freg"<»™+-
          temperature measurements with a precise thermistor and
          adequate temperature compensation.   The product should
          be circulated or mixed during the test.  Adequate
          temperature compensation is a key to successful
          interpretation of tank test data.  Such data must
          consist of accurate temperature measurements at
          frequent intervals.  The judgment to mix is a choice of
          techniques which is associated with the better
          performance achieved by the single thermistor approach
          used by Petro-Tite over the single thermistor approach
          used by Leak Lokator.

          Data on temperature and level changes must be collected
          frequently.  This is necessary to minimize aliasing of
          the high frequency fluctuations  (out of the signal
          band) into the lower frequencies (in the signal band).
          This conclusion is based large on data analysis
          performed by Vista Research, Inc.

          Data collection must continue for an adequate period of
          time so that sufficient data for a precise analysis can
          be provided.  A minimum of 4 to 6 hours with frequent
          temperature and tank level change intervals is needed.
          While a test length of 4 to 6 hours with frequent
          temperature and level readings is desirable, the
          practical considerations of cost and disruption to an
          establishment are also factors.
          The test method must incorporate an adequate
          statistical analysis of the data to draw supportable
          conclusions about the leak rate.  None of the
          techniques were found to collect either sufficient test
          data or to provide adequate analysis algorithms.
          Improved analysis protocols will be required.
5"Analysis of the Pilot Study Tank Test Data," Vista Research,
 Inc.,  July 1985.


                              C-29

-------
V.   PILOT STUDY

     A.  Objectives

     The results from the earlier stages led to the
recommendation that a test using modified Petro-Tite equipment
and procedures be adopted for the national survey.  The major
objective of this final stage was to modify and evaluate the
performance of the Petro-Tite method as it was to be used on the
national survey.  This process included:

     o    Determining the best sampling interval for collecting
          the data; that is, the time interval at which product
          in the standpipe should be re-leveled and data readings
          made ;
     o    Determining the best length of the test;
     o    Developing and testing the analysis algorithm;
     o    Implementing the procedures operationally in the field
          to identify operating difficulties and correct them;
     o    Field testing the entire survey data collection effort
          including scheduling, data collection, and analysis;
     o    Estimating the detection performance of the method; and
     o    Finalizing the test protocol.
     B.   Overview

     A sample of 25 tanks was selected from two primary sampling
units  (PSUs) on the west coast for use in the pilot study.  The
owners and operators of these tanks were contacted to arrange for
the tanks to be tested and to schedule the tests.  Timing of the
contacts and arrangements for fuel delivery, payments, and
scheduling presented difficulties. Recommendations for mitigating
                               :-30

-------
these on the national survey were developed.  Notifying owners
earlier of the test and giving a longer lead time to arrange and
schedule the tests were found to be necessary to expedite
testing.

     Data were collected at three different time intervals and
for three different total time periods.  The resulting data were
analyzed by various methods to select the most practical and
effective data collection interval and test length.  A standard
data analysis protocol was developed for use when no testing or
data problems are identified.  Data management procedures for the
national survey were developed which included the use of on-site
computers to collect data.  Data and test review procedures were
developed to check each tank test for validity and to ensure that
the standard analysis was adequate.  A simplified analysis that
can be used in the field to visually inspect the data and
identify potential testing problems was developed and
implemented.  The tank test data were analyzed and a data report
prepared and submitted to EPA.
     C.   Data Collection

     Data identifying the tank, size, location,  product, etc.,
were entered onto the top of a spreadsheet data file utilizing a
portable computer.  Then test data are entered as each data point
becomes available.  This provided a preliminary analysis and
estimated volume change rate that can be obtained on the scene.
     D.   Data Analysis

     The data from the pilot study tank tests were analyzed with
two objectives.  One was to determine the best sampling interval,

                              C-31

-------
and the second was to determine the best total test duration.
Sampling intervals of 1, 5,  10, and 15 minutes were considered.
Data collection at 1-min intervals was found to be impractical
for the large scale survey.   Both the 5- and 10-minute intervals
provided improvements in the precision of the test data,  but the
5-minute interval resulted in better precision.  Thus, data
collection at 5-minute intervals was selected as the standard.
This analysis is presented in detail in Vista Research, Inc.'s
report.6

     Selection of the total time of the test was not so clear-
cut.  Longer test times were desirable from a data quality
standpoint, but practical limitations were also considered.  A
compromise of 2 hours of data at the low level was selected as
providing sufficient data while still proving to be practical for
the field data collection.

     The test protocol used the same equipment as for a standard
Petro-Tite test.  There were no changes in the test procedures
except for the sample interval and length of the test.

     The analysis algorithm was modified to include smoothing of
the temperature data before applying the temperature correction.
A regression line was then fitted to the corrected data to obtain
the leak rate.

      Seventeen tanks were tested in the pilot study.  A summary
of the test results is presented in Table 9.

     A family of performance curves was generated for the large
and small tanks to estimate detection performance for a given
leak rate as a function of probability of detection, probability

6Ibid.

                               C-32

-------
            
 in
+J
 Ul
 
            4-> -C
            
            1 ^-^
            o
         •o
         J-l CNI
         td ^
         'Q
         C
M r-l
o  n)
S-i  60
V-i — '
w
            O)
            •P x-s
         in J^  to
         >> to  en
         oo a* ^^^
                O
                C
                O
                C
                        cnOOOOOrHOi-Hr-lr-lrHOOOOCOOO
                                                             O£O.
                        <_J O
                                     C_> O
                                                      000
                                                                    o
                                                                           o o
            OOLOOOOOCMOi-IOOOOOOOO

            C)C3rHC)dc)C3C3OodoC>OC)OO
            OO^-rHrHCSJUDi—(OOi—ICVJCOrOOOi—I
            ooooooooooooooooo
            OOPOOO«3Ot—lCNOLnOOO«S-i—lLf)i—li—I
            OOCOe\IOOOJOr-fOOC3OOOOO

            ddrHooddoooooooodo
                III      I  +      I  J-  I             III
                           i—I  CM CO
                                                rH 
                                                                                            TJ  U
                                                                                            i.  C
                                                                                            m  o
                                                                                            •o  u
                                                                                            c  c
                                                                                            03 HH
                                                                                            4->
                                                                                            in ^-s
                                                                                               I— I
                                                                                            < ^
                                                                                            Q.
                                                                                   OJ
                                                                                >»•—
                                                                               JD -Q
                                                                                   ro
                                                                                OJ -r-
                                                                                IT3 -(->
                                                                               •r- S_
                                                                               M- Ol i—
                                                                               •r- (J       10
                                                                                J- +->   O)
                                                                                                          E a»  s-
                                                                                                          3 S-  Q.
                                                                                                      m  s -a -o
                                                                                                      (U  Ol O)  Ol
                                                                                                      r—  s- -a -a
                                                                                                      C Q. 03  m
                                                                                                      zs     aj  ai
                                                                                                          it  —J —i
                                                                                                      II
                                                                                                         —i u   n
                                                                                                      —i -z.
                                                                                                      z -^ a: a-
                                                                C-33

-------
of false alarm, and test time.   Detection performance for 0.05
gal/h leaks was unacceptable.   A test period of 1 hour or less is
too short to achieve reasonable detection performance.  For the
small tanks, test times of 1,  2, and 3 hours result in the
detection of 0.10, 0.075, and 0.05 gallon per hour leak rates
with a PD = 95 percent and a PFA < 5 percent.  For the large
tanks, test times of 1 and 2 hours result in the detection of
0.25 gal/h leak rates with a PD = 95 percent and a PFA = 2
percent and 5 percent, respectively.
     Of the 17 tanks tested, one resulted in a clearly invalid
test. One test was problematical, but the system is probably
tight.  Three tanks appear to have significant leaks, and the
remainder appear to be tight.  Due to the fact that the Petro-
Tite method places a higher head pressure on the tank than is
found in normal operation, the reported rates are overestimates
of product loss or leakage in operation.

     Since the pilot study data available for analysis was
somewhat limited, the determination of the detection limit of the
Petro-Tite method could not be established as well as hoped.
Further data from the national survey will need to be examined.
VI.  RECOMMENDATIONS FOR NATIONAL SURVEY

     The recommendations for the national survey are:


     1.   Use a modified Petro-Tite test method;
     2.   Data should be collected at 5-minute intervals for
          2 hours at each tank; and

                               C-34

-------
     3.    Data analysis should use improved algorithms to fit
          data which exhibit curvilinearity in the test results.
     The final proposed equipment configurations and data
collection, environmental measurement, and data analysis
procedures which resulted from the development and pilot studies
were specified in a separate document.7   The actual procedures
and methods which were followed in the field are documented in
Sections 6 and 7 and Appendix D of this report.
7"National Survey of Underground Storage Tanks:  Draft Test and
 Analysis Plan," Midwest Research Institute,  June 10, 1985.
                              C-35

-------
                           APPENDIX D

           TANK TESTING DATA REDUCTION AND STATISTICAL
          ANALYSIS LEADING TO LEAK STATUS DETERMINATION
I.   INTRODUCTION

     This appendix contains additional detail and in some cases a
more technical presentation of topics covered in Sections 7 and 8
of the report.  Parts II and III of this appendix provide further
details on the tightness test raw data and the initial reduction
steps which produced the basic volume change rate estimates and
the estimated within-test standard errors for these estimated
rates.  Part IV provides further detail on the retest results,
which is summarized in Part V of Section 7.  Part V of this
appendix provides a more technical description of the estimation
of total test variance than is given in Section 8.  Part VI
provides a more technical description of the leak status
determination rule than appears in Part III of Section 8, and
Part VII gives more details on the adjustment to test pressure
than appears in Part II of Section 8.
I.   DATA COLLECTION AND MANAGEMENT
     A.   Data Collected

     The tank testing data collected consist of several data
elements.  A sample of a typical Petro-Tite data sheet is dis-
played as Figure D-l.  Identifying information about the site,
tank system, and product were determined and entered as header
                               D-l

-------
   R"	
    '   '      "*""^*~°~**°
         IS. TANK TO TEST
                                                                                        8/T/35
              if-  ^i r\Vo
                       * Ao X
                                                  (J
                                       l DfTIHMIIIIIM TANK CAPAOTY-
                                                                       •»&>     JX
                                                                       «—.      ^
                                                                                n
                                                                                u
        17. nu-ur FOB TEST
                     Xj
                     'H
                      • w la


                   AND Amu IACH COMFAHTMIHT OfW OK IACH MITIMf 0 OIUVIMY OUANTITT
         II. SPECIAL CONDITIONS AM PROCEDURES TO TEST THIS TANK  ° 7
                                                                                 VAPOfl RECOVERY SYSTEM
         II. TANK MEASUREMENTS FOR
           TSTT ASSEMBLY
        20. EXTENSION HOSE SETTING
                                            21. TEMPERATURE/VOLUME FACTOR (•! TO TEST THIS TANK

                                            K ToMv MMMTI I j CdMifi I ]	- f  ~umi m !•*	• F »«• >imnil m In
                                           22.
                                           IS.
                                                 II CflOTf* tmt (Ml
                   .
             lw«lh tt faM If
                                                                              CUT:
     ?""C Pr
 3&Q
6*5
      J*.
                                                           -.035
      Go«v>A
                                                                   (nlQ
                                                                                      -.COS
                                                UftO
                                                            -.015
                                                            -.CHS
                                                                              V.OCO
33*2!
am»,	
cvm [Dr^. v> \^-  N-^.LoAw.v^
                                                            -.Q\Q
                                                                         VO
                                X
c»n_L
                                                             fx-n
                                                                              H.QQQ
                                                                                      \.OOQ
                                                                                             ^CQCii
                                                                                             -gci(u_
                                                                         VCl
                                                                              v.COO
                                                     .V1Q
                                                                              t.COQ
                                                                                              .00 <^
                                                                                              Jiai.
                                us
                                              .-310
                                                                         vQ
                                                                                             -.cot
i£sa
                                    0
                                                           v.non
                                                                              t.OQQ
                                                                                      X.OOQ
                                                                              V.OCCl
                                                                                      Y.OOO
                                 \le
                                                                          VO
                                                                              tCDCv
                                                                                     iii'iCL
                                                                         iiL
OvQT L
                                               ."an?;
cmai
                                                            V.OCTi
 3U1J
                                                           iiXiCL
                                                                              •v.eoo
                                                                         ±0.
     T.'XQ
                                                                                     i^secL
                                                                                             — ocn_
                                                                         VO
                                                                         VQ
                                                                              Y.OCO
                                                                                      V.CCd
                                                            V.OiCK
                                                                          VO
                                                                              V.CCC)
                                                                              \.CCQ
                                                                              .ir.qy.
                                                                                      v.cco
                                                                                             j^ai-
                                                      TXSO
                                                            v.c.co
                                                                          vO
                                                                              Y.CXTi
                                                                                      •vccn
 Q\51I
                                                            v.CCO
JtL
V.QDk
                                                            •V.CCO
                                                                         VO
                                                                               V.OCD
                                                                                      v.oco
oatfi
                                                            v.oco
                                                                         vO
      v.cco
                                                                                      Y.QCO
                                                                              v.ooo
                  Figure  D-l.
                                       Petro-Tite data  sheet

                                         D-2

-------
information.  Additional data needed to set up the test were
recorded.  These included the diameter of the tank, the depth
from grade to the bottom and top of the tank, and the depth of
the water table.  An initial thermistor reading was taken and the
internal check of the thermistor unit was performed.  The spe-
cific gravity of the product was measured and used to determine
the coefficient of expansion.  The tank volume was determined.
Presence of water in the tank was checked.  If water was present,
the volume of water in the tank was calculated and subtracted
from the tank volume to determine the volume of product.  A final
adjustment to product volume was to add the volume in the test
equipment (usually 2 to 3 gallons).

     After the preliminary data had been entered in the header,
the actual test data were taken and entered.  The time of reading
was entered.  The reference level was noted.  The volume in the
graduated cylinder before releveling was found and entered.
After releveling, the volume in the graduated cylinder was found
and entered as "volume after."  The fuel temperature in terms of
the digit reading on the thermistor unit was found and entered.
The actual test data used to calculate leak rates consist of the
time, the volumes before and after, the temperature, the tank
product volume, the digits per degree Farenheit, and the
coefficient of expansion.
     B.   Data Management

     The test data collected as described above were recorded on
a Petro-Tite data sheet by the test crew.  The MRI technician at
the site keyed these data into a Lotus 123 worksheet file that
had been configured to receive the data and perform preliminary
calculations.  An example printout of the data portion of this
file is shown in Figure D-2.   The MRI technician entered the
                               D-3

-------
19-Aug-85
       Page 1
       Survey ID
 Tank  Test Firm DBL CHK
 2 of  Test Crew
     2 MRI Crew  STEVE
  Fuel Type UNLEADED   Date
                     AUG 7 1985
  Tank Vol
7 API Dens
  Exp Coef
      3010 T digits      16676
      58.6 T digits/F      322
0.00060366 Leak Rate     0.001
           Std. Err  0.0058976
Ti
Hr
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
me L
Min (
17
22
27
32
37
42
47
52
57
2
7
12
17
22
27
32
37
42
47
52
57
2
7
12
17
evel
div)
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
V Before
(gal)
N/A
0.270
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.275
0.275
0.275
0.275
0.275
0.275
0.275
0.28
C.28
0.28
0.28
0.28
0.28
0.28
0.28
V After F
(gal) (
N/A
0.270
0.27
0.27
0.27
0.27
0.27
0.27
0.27
0.275
0.275
0.275
0.275
0.275
0.275
0.275
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0.28
0/29
uel Temp
digits)
16669
16670
16670
16670
16670
16670
16670
16670
16670
16670
16670
16671
16671
16671
16671
16671
16671
15671
16671
16671
16672
16672
16672
16672
16673
Tcorr dV L
(gal)
N/A
-0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.005
0.000
-0.006
0.000
0.000
0.000
0.000
0.005
0.000
0.000
0.000
-0.006
0.000
0.000
0.000
0.004
eak Rate
(gal/h)
N/A
-0.068
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.060
0.000
-0.068
0.000
0.000
0.000
0.000
0.060
0.000
0.000
0.000
-0.068
0.000
0.000
0.000
0.052
                  Figure D-2.  LOTUS data sheet
                                 D-4

-------
header data including the date, test crew,  testing company, MRI
person, and the time, level, volumes before and after, and fuel
temperature (digits).  The program calculated the leak rate,
standard error, and other intermediate values.

     After the data were entered into the computer on site, they
were stored on a diskette.  In order to facilitate expeditious
data analysis, the data were transmitted to MRI via telephone
using a modem.  The diskettes containing the data files were
shipped to MRI on a weekly basis.  The original Petro-Tite data
sheets were also shipped to MRI.

     Upon receipt of the electronically transmitted data files,
they were printed and the volume trends plotted.  Figure D-3
shows an example of such a plot.  The calculations of the leak
rate and standard error were checked.  Any unusual features of
the data such as outliers or curvilinearity were noted.  The
computer file was archived as received and the hard copy was
placed in an archive file.  A copy of the computer file was
placed in a working directory.

     When the disk containing the data file was received, the
disk file and the telephone file were compared using the IBM DOS
utility file compare program to determine whether the data
transfer was complete and accurate.  If the files were found to
differ, a new hard copy of the data and graph were printed.

     Upon receipt of the Petro-Tite data sheets, the printed data
from the computer file were checked against the raw data sheets.
Any discrepancies were corrected in the computer file.  If the
final file differed, another hard copy of the data and graph was
printed.  The final form of the computer file was archived.
                              D-5

-------
                                              •D
                                              -n

                                              i-n
1    1    1    1     1    1    1    1    1    1    1    1
                                                           1
OOO
OOOO
                                OOOO
                    auun|o/\p
                                              1
                                                                CM
                                                                 (O
                                                                 OJ
                                                                 (D
                                                                  0

                                                                 O
                                                                 (N

                                                                 O
q  q  q
odd
 I    I    I
                                                                        Q.
                                                                        E
                                                                       .   2o
                                                                      T-   £.0
                                                                           Q)
                                                                      Q)
                                                                      OT

                                                                 00   JO

                                                                 d   u
                                                                             0)

                                                                             a
                                                                             E
                                                                             a>
o


o
o
c
D

D
                                                                                 a)
                                                                                rH
                                                                                •H
                                                                     (0

                                                                     (0
                                                                     Q

                                                                     (0

                                                                     -P
                                                                     o
                                                                                Cn
                                                                                C
                                                                                •H
                                                                                w
                                                                                     (0
                                                                                     -P
                                                                                     R)
                                                                                     •a
                                                                               -P
                                                                                o
                                                                                i
                                                                               Q

                                                                               a)
                                                                               PH
                                 D-6

-------
     After the data had been checked against the original sheet,
the final data analysis was done for the tank or system.   When
the analysis was completed, a final copy of the data was  printed,
incorporating any special analysis with the final leak rate and
standard error estimates.  The final computer file was archived.
III.  DATA REDUCTION ANALYSIS METHODS
     A.  Statistical Methods Considered and Choice

     Several methods of statistical analysis of the tightness
test data were considered for use on the national survey.   This
section presents a discussion of the advantages and disadvantages
of each and gives the reasons for the selection of those used.

     The test method produced a volume change measurement at 5
minute intervals.  This change was measured directly by bringing
the standpipe to a reference level and collecting the product
recovered or measuring the additional product needed.  The other
measurement recorded at 5 minute intervals was a temperature
measurement.  This measurement was taken by means of a thermistor
probe and box.  To make this reading, a resistance bridge was
balanced by means of a dial.  The instrument reading was con-
verted to a temperature by means of the calibration chart for the
instrument.  The readings—after conversion to temperature—were
the temperature of the product in the tank at 5 minute intervals.
The temperature record of the product as measured by the thermis-
tor must be converted to an equivalent volume change using the
volume of the tank and the thermal coefficient of expansion.  One
essential difference between the volume and temperature readings
should be noted.  The temperature was recorded as a cumulative
                              D-7

-------
reading—the tank temperature—while the volumes were recorded as
differences.

     In order to make the temperature and volume data comparable,
they must be put in the same form.   Either both must be changes
or both must be cumulative.  Several approaches can be used for
the analysis.  The standard Petro-Tite approach to the analysis
of the data is to take differences  in the temperature readings.
The time interval used by Petro-Tite is 15 minutes rather than
the 5 minute intervals selected for the national survey testing.
After taking differences in the temperature readings, the change
in temperature is multiplied by the volume of the tank and the
thermal coefficient of expansion for the product to produce a
volume change due to temperature.  This is subtracted from the
observed volume change at each point.  The resulting differences
are temperature-adjusted volume changes.  The standard Petro-Tite
analysis adds up four of these 15 minute readings to obtain the
hourly leak rate that they report.   An advantage of this method
is its simplicity.  A disadvantage is that no estimate of
variability is provided.  An additional disadvantage is that four
15 minute data points do not provide sufficient data to ensure
that the test is valid.

     A similar approach could be followed for analysis of the
survey data.  Consecutive temperatures could be differenced to
obtain temperature changes for each 5 minute interval.  This
would provide a set of observed volume changes and temperature
changes.  The temperature changes would be converted to volume
changes by use of the coefficient of expansion.  At this point
two different approaches to the analysis could be used.

     One approach is to regard the observed volume changes and
the temperature volume changes as a paired sample.   In this
analysis, one would calculate differences in each pair.  These

-------
differences would be averaged to obtain an estimated leak rate.
The variability of the differences would be used to obtain an
estimate of the variability measured by the standard deviation.
The variation of the mean would be estimated by the standard
error of the differences (the standard deviation divided by the
square root of the number of terms in the average).  This would
result in n-1 degrees of freedom for the standard error.  Both
the mean and standard error (or standard deviation) would be
rescaled to an hourly leak rate.

     There are a number of advantages to this approach.  It is
directly comparable to the standard Petro-Tite tests.  It is
relatively simple and should be easily understood.   It does pro-
vide an estimate of variation.  If the volume change and tempera-
ture changes are dependent, it accounts for this by pairing the
data.  In addition, if the differences were less variable than
the original data, it would provide a more precise estimate than
other approaches.  A disadvantage is that if the data are not
dependent, it sacrifices degrees of freedom unnecessarily. In
addition, if pairing does not reduce variability, then this anal-
ysis would lose precision.

     A slightly different approach is to regard the volume data
and the temperature-volume data as two samples rather than as a
paired sample.  With this approach, the mean volume change would
be calculated as would the mean temperature-volume change.  The
difference in these two means would be calculated.   This would
result in the same estimate of the leak rate or volume change as
with the paired data.  However, there would be a difference in
the estimation of the variability.  Each set of data—volume and
temperature-volume—would have its variability estimated
separately by the sample variance.  If it were assumed that these
variance estimates were estimating the same quantity, a pooled
variance estimate could be calculated from these two.  This would
                              D-9

-------
have a total of 2n-2 degrees of freedom,  where n is the number of
data points of each type.   This approach has an advantage if
there is no inherent dependence in the two types of readings.   It
also is advantageous if pairing does not reduce variability
enough to offset the loss  in the number of degrees of freedom.

     If it were concluded  that the variation of the two types of
data is different, then the sample variances should not be
pooled.  In this case, the variance of the difference in sample
means would be the sum of  the two variances of the means (the
variance of the mean is the sample variance divided by n).   The
assumption would be that n is large enough so that the sample
mean would be approximately normally distributed.  After the
variance of the difference in the means is calculated, the square
root of this number would  be taken.  Finally, the estimated leak
rate and the standard error of it would be rescaled to an hourly
leak rate as before.  Thus, while the estimate of the leak rate
would be the same, the estimate of the variability would differ.
This approach has the same advantages of the previous approach.
The essential difference is in the calculation of the variabil-
ity.  The choice between these two approaches should be based on
whether the assumption that the temperature-related volume
changes and the observed volume changes have the same variability
is valid.  Consideration of the precision of the two measuring
instruments and of the rounding errors involved in the two
measuring processes suggests that the temperature-related volume
changes and the observed volume changes do not have the same
variance in general.  Consequently, this latter approach would be
preferred.

     The result that the variability in the temperature-related
volume data is larger than the variability in the observed volume
changes suggests that it may be advantageous to smooth the
temperature data before adjusting the observed volumes for
                              D-10

-------
temperature.  Basically, this approach would use some degrees of
freedom to smooth the temperature data by fitting a curve of some
sort to them prior to making the volume adjustments for tem-
perature.  It would use the fitted curve in the adjustments in
order to reduce variability.

     Since the temperature data as recorded represent the
temperature of the tank over time, one approach is to fit a curve
to these temperatures and use the expected or predicted values
from the fitted curve for adjustment.   In the typical test, the
temperature increased smoothly in a nearly linear fashion over
the period  (about 2 hours) of the test.  In this case, a linear
regression through the origin (or the starting temperature)
provides an adequate smoothing.   The predicted values from the
regression can be used to adjust the volume changes.  In some
cases, the temperature displayed a curvilinear form so that the
straight line fit was inadequate.  In these cases, adding a
quadratic term to the regression provided a satisfactory fit.
Occasionally, the temperature was not monotonic or displayed some
other unusual behavior.  In this event, moving averages were used
to smooth the temperature-related volumes prior to adjusting the
volumes.

     The advantage of smoothing is that it may reduce the
variability of the estimate and so improve the precision of the
test.  A disadvantage is that it is somewhat more complicated
than a linear or quadratic fit.   An additional potential
disadvantage is that it may require a different form of analysis
to be used depending on the temperature data. On the other hand,
any method of analysis should allow for diagnostics to ensure
that the data from the test meet the assumptions adequately.  It
should be anticipated that some tests will give data that do not
meet the standard assumptions.  Such tests will either be judged
invalid or will require specialized analysis.
                             D-ll

-------
     A rather different approach can be taken by cumulating the
volume differences.  This would provide two sets of cumulative
data (one for volume, one for temperature-related volume)  that
can be viewed as time series.  With this approach, time series
models could be fit to both series.  A transfer function could be
used to relate the two series and form a third series of the
temperature-adjusted volumes.  The estimate of the temperature-
adjusted volume change rate could be made from the parameters of
the time series model of the derived series.  This approach would
have an advantage if the volume measurements and temperature
measurements showed common forms of serial correlation that would
leak to a particular form for a time series model in the majority
of cases.  There are some disadvantages of this approach.   One is
that a large number of data points is required in order to fit
the time series models and have a sufficient number of degrees of
freedom.  A second is that the analysis is much more complicated
and time consuming.  A third is that the analysis must estimate
the appropriate model form for each series.  The major drawback
is that time series analysis requires more data than was
available from the tests in the national survey.

     A spectral analysis of the data from a long test during the
pilot study led to the conclusion that for test times exceeding
one hour, a sophisticated time series algorithm was not
necessary.
     B.   Standard Analysis

     As a result of the considerations of the types of analyses
available and the advantages and disadvantages of each, a
standard analysis was designed.  For the standard analysis, the
temperature-related volume change and the observed volume change
                             D-I2

-------
were both expressed in cumulative form,  beginning at zero for the
start of the low level (4-psig)  test.   A straight line through
the origin was fit to the temperature-volume data by least
squares.  The predicted values of this line were calculated and
used as a smoothed temperature correction.   The data were plotted
and inspected visually for outliers or deviations of the tempera-
ture data from linearity.  Any questionable data were checked in
detail or considered for special analysis.

     If no problems with the data were found, the predicted
values from the smoothed temperature line were used as the tem-
perature correction.  This smoothed temperature correction was
subtracted from the observed volume data for each time point.
The resulting differences were divided by the time interval to
obtain a series of volume change rates expressed in gallons per
hour, typically based on a 5 minute interval.  The arithmetic
mean of these rates was calculated and used as the estimate of
the leak rate.  The standard error of this mean was calculated
and presented as the standard error of the estimate.  In the
variance computation, n-1 was used as the divisor, where n is the
number of terms in the mean. The result was divided by n to form
the variance of the mean.  The square root of this is the within-
test standard error reported before adjusting for between-test
variation.  (See Section D.V, below, for discussion of total
variance.)

     The question of the appropriate number of degrees of freedom
was considered.  It was possible that the terms in the mean might
be correlated, implying that the actual degrees of freedom would
be less than n-l.  Spot checks of the serial correlation of the
terms showed generally no significant (at the 10% level) correla-
tions.  For a few data sets some of the lag correlations were
significant.  However, this occurred in only about 20% of the
data sets.  Those where one or more significant correlations were
                              D-13

-------
found showed no consistent pattern of which serial correlations
were significant.  Consequently,  this was interpreted as being
likely to be due to chance.  No adjustment of the degrees of
freedom is thought necessary.
     C.   Special Analyses

     A number of data set features called for a different or more
detailed analysis than that described above.   The most obvious
case was that of a manifolded tank system.   Within the set of
manifolded systems, a slightly different analysis was needed for
different numbers of tanks, and a different analysis was needed
for systems tested together as opposed to those with tanks tested
separately.

     Manifolded tanks that were separated and tested separately
provided two or more individual tank tests.  As individual tank
tests, these were subjected to the standard analysis (or special
analysis if needed).  This provided volume change rate estimates
and standard errors for each tank (and its associated lines).
These needed to be combined to estimate a system volume change
rate.  In the descriptive data presented in the first part of
Section 9, the individual test results for tanks in a manifolded
system were used separately when available.  The multivariate
analyses were restricted to single-tank systems.  Thus, creating
system volume change rates was done for completeness in the
deliverable data file.  This was done by summing the two
estimates of volume change rates.  The variability of this
combined rate was estimated by taking the variances of the
individual volume change rates and adding these. Taking the
square root of this gave the standard error of the combined rate.
This extends to any number of tanks in a manifolded system tested
separately.
                             D-14

-------
     Manifold tanks tested together provided slightly different
data.  A single standpipe (or two connected by a siphon) was
used.  A single volume change was recorded for the system every 5
minutes.  However, each tank had a circulation pump and the
associated thermistor unit to measure temperature.  In general,
each tank could have a different volume, although the usual case
was for tanks of the same volume to be manifolded.

     A temperature-related volume change was calculated for each
tank.  These were summed.  The result represented the total
temperature-related volume change.  This was used as the tempera-
ture effect.  It was smoothed as before with a least squares line
through the origin, and the temperature adjusted volume change
rates calculated as before.

     A number of other special cases were found and were dealt
with on an individual basis.  Occasionally apparent outliers were
found.  These were checked against the raw data and the test log
to see if there was any physical reason for them.  A few tests
had thermistor boxes fail during the test for some reason (rain,
FM interference).  These generally gave temperature data that
appeared as outliers.  When outliers were found and a physical
reason identified, the aberrant data were removed from the
analysis.  This generally required smoothing over the missing
data by interpolation.  If errors were identified, they were cor-
rected and the analysis redone.

     The typical data showed a monotonically increasing tem-
perature, generally linear.  A smaller proportion of the data
sets showed linearly decreasing temperature.  Some data sets
showed evidence of temperature increase that was curvilinear.  If
this curvilinearity appeared or was suspected, a test for curvi-
linearity was done by fitting both a linear and quadratic to the
                              D-15

-------
temperature data by least squares (through the origin).   If the
quadratic improved the fit significantly,  the curvilinear fit
(using both linear and quadratic terms)  was used for smoothing.

     A few cases were found where both temperature and volume
were not only non-linear, but also non-monotonic.  Provided that
they showed the same pattern, analysis proceeded.  In this event,
a five point moving mean was used to smooth the temperature data.
Equal weights were used.  This resulted in the loss of four data
points; two at the start and two at the end of the test.  The
moving mean smoothed temperature volumes were subtracted from the
volume changes to obtain temperature-corrected volumes.  These
were divided by the time intervals and expressed as gallons per
hour.  The arithmetic mean and standard error of these
temperature corrected volume rates were calculated and used as
the estimates of the volume change rate and its standard error,
respectively.

     Some tests showed volume change rates that were initially
increasing rapidly and curvilinear, while the temperature changes
were quite linear.  The volumes typically increased rapidly for
the first few times, then slowed.  This was interpreted as
relaxation of tank deformation.  The apparent relaxation appeared
to follow an exponential curve and to approach the temperature
change rate as an asymptote.  However, the constant of this
asymptote differed by tank.  The rate of relaxation may be
related to the nature of the soil in backfill and water
conditions.  When this was identified, the initial points
exhibiting this relaxation of the tank deformation were deleted
before analysis.
                               D-16

-------
     D.   Criteria for Invalid Data

     A few of the data sets from the tank tests were judged
invalid based on the analysis of the data.  This occurred quite
infrequently.

     There were a number of criteria for declaring a data set to
be invalid.  The most common was that the data showed a volume
increase even after adjusting for temperature.  Since the test
method places pressure on the tank, a volume increase cannot
occur from inflow of water.  Data that showed volume increases
after temperature adjustment that exceeded levels that could be
reasonably attributed to the variability of the measurement proc-
ess were judged to be invalid tests.  The reason for this is that
such an apparent volume increase with no explanation could be
eclipsing a small actual volume loss or leak.  Generally any tank
that showed a volume gain rate of more than 0.1 gallons per hour
after temperature adjustment was judged to be an invalid test.
The most likely explanation for such tests is that those tanks
had trapped vapor pockets.

     A variety of other data features led to the conclusion that
the test was invalid.  Some of these may also have been caused by
trapped vapor.  A few instances were found where the temperature
as recorded fluctuated erratically during the test while the
volume measurements were relatively stable.  If the temperature
data were so erratic as to preclude a temperature adjustment,
then the test was declared to be invalid.  One or two tests
showed both temperature and volume measurements that were erratic
and did not appear to track together.  These tests were also
judged invalid.  Such behavior may have been caused by incomplete
tank deformation, followed by relaxation, combined with mixing
problems.  No valid volume change rate could be estimated.
                               D-17

-------
IV.  RETEST RESULTS

     Three types of retests were conducted as part of the
national survey of underground storage tanks.  One was a back to
back retest, conducted immediately after the original test used
to estimate the leak rate.  The second was a leak simulation test
also conducted immediately after the original test.  The third
type was a complete retest conducted on a different day and
generally by a different crew.  Each of these types estimates a
different source of variation possible in the tank tests.  A
tabulation of all of the retests appears as Table D-l.  (Note
that a negative volume change is a leak, while a positive volume
change represents net inflow.  In the body of the report, leaks
are reported without minus signs.)  The simulated leak retests
are tabulated in Table D-2.  A table summarizing the estimates of
bias (accuracy) and standard deviation  (precision) based on each
type of test is presented as Table D-3.  It should be noted that
the three types of retests estimate different sources of
variation and so are not directly comparable to each other.
     A.  Leak Simulations

     The leak simulation tests were conducted after the original
test was concluded.  Generally they were only conducted when the
original test indicated that the tank was tight or had a small
estimated volume change.  The volume rate used for leak
simulation was on the order of 0.1 gallons per hour, so a large
volume change would overwhelm it.

     The purpose of the leak simulation tests was to document
that the testing method could detect leaks of known size in tanks
that appeared to be tight.  In addition, use of the leak
                             D-18

-------
Table D-l.  Retest Data Summary
Survey ID
N02784A
N131078A
N171261A
N21581B
N281389B
L01034A
L01036B
L01037A
L01037B
L02068A
G03018A
G03018B
L03095A
L03095B
G06013A
G06013B
G06028A
G06028B
G07010A
G07010B
G10020T1
G10020T2
N141107A
N141107B
N151141A
N151141C
G16005AR
G16005BR
L16394A
L16394B
N171261G
N181323C
N181323D
N181326B
G19068A
G19068B
G19068C
G19101A
N19525A1
N34128A
N34128B
Volume
2007
3985
3979
3973
11988
1039
2005
4013
4013
3989
3010
3010
6049
6048
6018
6018
2964
2964
277
566
10155
10155
1035
1033
10576
21154
1003
295
1023
1039
576
1005
1005
1033
1005
4032
1038
566
8060
6262
8000
Fueltype
DIESEL
UNLEADED
GASOHOL
PRE UNLD
REGULAR
UNLEADED
DIESEL
SUP UNLD
REGULAR
REGULAR
DIESEL #1
REGULAR
DIESEL
DIESEL
REGULAR
DIESEL
REGULAR
DIESEL
DIESEL
REGULAR
REGULAR
UNLEADED
GASOHOL
DIESEL
DIESEL
DIESEL #1
UNLEADED
REGULAR
REGULAR
UNLEADED
DIESEL
UNLEADED
REGULAR
UNLEADED
REGULAR
DIESEL
UNLEADED
DIESEL
UNLEADED
SUP UNLD
UNLEADED
Type
BTB
BTB
BTB
BTB
BTB
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
Initial
Date
0730
0822
0804
0731
0806
0709
0712
0724
0724
0809
0731
0731
0802
0802
0724
0724
0721
0721
0628
0628
0625
0626
0817
0817
0817
0817
0722
0722
0728
0728
0804
0721
0721
0722
0715
0715
0715
0712
0710
0617
0615
oc
Date
0731
0822
0804
0801
0807
0812
0826
0828
0828
0810
0827
0827
0826
0826
0828
0828
0829
0829
0826
0826
0816
0816
0831
0831
0824
0824
0828
0828
0831
0831
0810
0825
0825
0829
0828
0828
0828
0827
0822
0828
0828
Initial
Rate
-.015
-.102
.049
-.822
-.025
.013
-.055
.019
.036
.039
-.194
.060
-.036
-.032
-.153
-.089
.053
-.708
-.007
-.005
1.189
.584
.006
-.327
-.621
-.129
-.006
.046
-.021
.018
-.014
.025
.034
-.076
-.614
.070
-.068
-.078
.032
-.034
-.080
SE
.007
.018
.019
.038
.019
.014
.049
.013
.016
.014
.01
.009
.011
.013
.018
.011
.016
.018
.054
.027
.322
.028
.007
.010
.023
.008
.021
.015
.012
.012
.007
.04
.013
.018
.014
.008
.014
.02
.05
.009
.03
Retest
Rate
-.009
-.079
.040
-1.315
-.032
-.005
-.009
-.028
.017
-.019
-.226
-.005
-.117
-.047
-.097
-.325
.049
-.613
-.001
-.017
.175
.109
-.013
-.377
-.411
-.009
.025
.022
-.030
.011
-.010
-.013
-.015
-.032
-.559
-.002
-.076
-.100
.044
.044
-.053
SE
.005
.013
.010
.059
.020
.009
.008
.022
.012
.012
.005
.009
.006
.007
.016
.017
.008
.015
.009
.012
.022
.018
.010
.027
.015
.008
.011
.011
.008
.014
.004
.008
.008
.007
.018
.011
.011
.008
.030
.003
.014
                D-19

-------
 d
 3
W

x
 (0
•o
 0)
-p
 A
(0
(M
 I
Q

 0)
rH
43
 (0
EH
          Q

           S

          CO
          OrHrHOOrHOOrH


                     IOOIIIIIOIIII
rHOOQOOrHrHOOOOOiH


 I   I   I   I   I  I   I   I   I   I   I   I   I





OOOOOOOOOOOOO


ooooooooooooo




OO(N^«OOrHOOOOOO

 I   I   I  rH  I  I   I   I   I   I   I   I   I
          I




OCNOOOOOOOOOOO


ooooooooooooo


^? ^? ^^ tO '"^  p™4 ^J ^5 ^^ ^^J ^rt i^ ^^
^3* OQ ^^ 00 ^i  ^1 CO ^5 Gr* CO ^5 f™H r*^
oom^oooor-toooo

Ol   I   IOOI   I   I  O  I  O O
                                      cd        W
                                         u w
                    O in rH O
                                                   O rH rH
                   rHrHtnmOONNMrHlNrHrH

                   ooooooooooooo
                          03 PQ
                                      CQ
                                  
-------
1































































10
-p
rH
3
10
Q)
rH

-P
10
d>
-P
CD



.
l
a

0)
rH
43
id
EH
C
id TJ
0) CD rH ^-»
g rH O 43
id rH a
•P 3.rH fr
O O1 Q) ^
O (0
Q£




•0 J3
rH O
«J-H ^
•C -P 43
c id a
(d "H C^
-p > ^
co ct>
•o






•0 CM
C CD H — .
Id rH O 43
a) id rH a
S 3 rH O1
~ &Q)~
to




0)
OCM
C *'™1*
id 43

M CP
id ^
^





2;



,-»
M
rj
O
0) 43
O
(3 rH
C CD 0)
id rH a
0) 0)
g — '
r»
in
CM
o
•
o






Tj«
[•»
O
O
o
o



^-^
H
^^
VO
VO
o
o
o
•
o




H
rH










H
CD
00
O
O
•
O
1





G
0
-H
-P
id
rH
i
-H
(0

,5^
id
3


o>
m
CM
o
o





^—^
rH
*~*
H
f)
CM
O
•
0






[«»
in
o
o
o
o



,—.
H
V-"
n
in
o
0
o
•
o




^t
rH










o\
CM
VO
O
O
•
0









^
o
id
0
•P

,S({
o
(X)


m
o
in
o
o





^.
CM
N^
^f
O
in
o
•
o






in
m
CM
O
O
O



,— ^
CM
*~*
ff
in
CM
o
o
•
o




^*
m










t^
c^
CM
O
O
•
O












to

(0
a)

0)


0) Q) rH
43 0 3
-P-H O
10 -P
C
0) -H •O -P tQ Q) Id 10 -p S a) CD TJ M rH O CD 0) a 43 0) O CD •P 43 C rH 4J id v -H W tO 10 rH'H •P 3 id 10 C > CD CD-H O -P g >iC 0) C id rH rH Id -H id M X-H M Id O 4J 3 > (d -H rH 43 (3 at) •H CD O CD -P 4J CD O Id 43 C g X-P ld-H 0 -H -P Id T) rH tQ 43 C id CD id > *O tP c in -P c id CD 10 -H -P CD -O C id .P C O rH I O •H CQ, .p -O -H 10 id CD 43 CD rH > +J rH 3 M -H rH e a) u o •H 10 O to 43 ty O C CD X -H 43 id TJ >iB 0) CD rH rH O rH (3 CD • 0) 0) TJ 10 43 rH C T3 4-> CD 3 O ^lt B_j rH CD 0) rH 43 CD 0) -P 0 43 - id TJ 10 -H CD -P >H 4J to id id 0) > g -P -H 0) rH 4J M id 10 -P CD CD O 4J JJ 10 0) -H rH 0) ft 43 0) g -P 0 0 C O CD id 0-H M-rH rH o > id fep > CM D-21
-------
simulation allows for an estimate of the accuracy of the test as
well as its precision.  The accuracy refers to the ability of the
test to measure a known volume change,  while the precision of the
test refers to its ability to reproduce measured rates.

     Thirteen leak simulation tests were conducted.  Two of these
were conducted on tanks that had estimated volume rates that
indicated that the tanks were probably leaking.  These tests were
excluded from the analysis because variability is known to
increase for leaking tanks.  The results from all of the leak
simulation tests are tabulated in Table D-2.  Using the leak
simulation results from the tanks with small estimated volume
changes (less than 0.1 gallons per hour in absolute value) gave
the following results.

     Three rates were calculated from leak simulations.  The
first was a baseline rate for the tank.  This was estimated dur-
ing the regular tank test.  While the leak simulation was con-
ducted, a measured rate was estimated.   This is the rate observed
by the testing method during leak simulation.  It is presumed to
be composed of the tank rate plus the simulated rate.  The simu-
lated rate is calculated by collecting product drawn from the
tank at a constant rate, weighing it on a triple beam balance,
and converting the weight to volume at the temperature of the
product in the tank.  The difference between the observed rate
during the simulation and the baseline rate provides an estimate
of the simulated rate.  The difference between this and the
actual simulated rate can be used to assess the accuracy of the
test.

     The average difference between the measured rate  and the
simulated rate was -0.00891 gallons per hour, based on the 11
leak simulations where the tank was not estimated to be leaking.
If the other two simulations are included, this mean difference
                              D-22

-------
increases to -0.0184 gallons per hour.  The difference between
the measured rate and the simulated rate is interpreted as an
estimate of bias.  The variance of the differences about their
mean provides an estimate of twice the within-test precision plus
any variance due to taking successive 2 hour test periods.
Taking half the variance of differences estimates the variance
itself.  The estimate was 0.00066 gallons per hour squared for
the 11 tests.  (It was larger, 0.00291 gallons per hour, if all
13 tests were used.)  A mean squared error (MSB) can be
calculated to incorporate both types of error—accuracy and
precision.  The mean squared error is the sum of the bias squared
plus the within-test variance.  In this case it was 0.00o74
gallons per hour squared (or 0.00325 gallons per hour squared for
all 13 tests).

     The bias is clearly not significant, in that it does not
differ significantly from zero (t = -0.347, 10 degrees of
freedom).  As a result, the variance and the mean squared error
are nearly identical.  A measure of variation often used is the
standard deviation  (or root mean squared error if bias is
present), which is the square root of the variance (or MSB).
This measure has the advantage that its units are the same as the
measurement, gallons per hour.  The standard deviation was
estimated to be 0.0257 gallons per hour for these data.
     B.  Back to Back Retests

     Back to back retests were conducted on a total of 18 tanks,
which includes the 13 tanks with leak simulations.  Five tanks
had back to back retests without leak simulation.  The purpose of
the back to back retests was to estimate the stability of the
test method.  That is, to ensure that the volume change estimate
                              D-23

-------
did not differ markedly if based on the succeeding 2 hours after
the test.

     As with all of these tests, variability is expected to be
larger if the initial leak rate or volume change is larger.  For
this reason, the results of the back to back retests are pre-
sented primarily for those tests with volume change rates less
than 0.1 gallons per hour in absolute value.  Retest results for
tanks with larger volume rates were more variable but generally
consistent.

     The average difference between the original and retest for
the 14 tests with small volume changes was 0.00629 gallons per
hour.  The estimate of within-test plus change over 2 hour
periods variance was 0.00053 gallons per hour squared, giving a
mean squared error of 0.00057 gallons per hour squared.  The
corresponding standard deviation was 0.0231 gallons per hour and
the root mean squared error estimate was 0.0239 gallons per hour.
The mean difference was not significantly different from zero
(t = 0.272, 13 df).

     If all 18 back to back retests are used, the estimates are
slightly larger.  The mean difference was -0.0134 gallons per
hour, with the variance and MSB being 0.00893 and 0.00910 gallons
per hour squared, respectively.  The mean difference did not
differ significantly from zero  (t = -0.14, 17 df).
     C.  Complete Retests

     The complete retests consist of revisits to the site on a
different day.  Typically this includes a different crew and
involves rescheduling and refilling the tank.  The complete
retests incorporate all of the features of a tank test and so
                              D-24

-------
include all the sources of error including potential difference
from crew to crew and differences due to weather conditions,
nearby traffic flow, day of the week, etc.  In addition,  there is
a possibility that the tank is different at the time of the
retest.  In fact, two of the retests originally scheduled were
cancelled when it was found that the tanks had been repaired
between the initial test and the scheduled retest.  In addition,
two retests were performed and it was then discovered that the
tanks had been repaired between the initial test and retest.
These data are also not included, as they would measure an
additional source of variation which is not of interest,  i.e.,
repair.  Two other retests were performed on tanks that were
initially determined to have large vapor pockets.  These two
tanks were retested later and on retesting were again found to
have large vapor pockets.  The results of the test and retest for
these tanks with vapor problems agreed qualitatively; however,
the numerical agreement was not close.  The reason for this may
be that the vapor pocket trapped in the tank was of different
size.  There were also different ambient conditions that would
affect the vapor differently.  For these reasons, the vapor
retests were not included in the estimate of the variance from
the retests.

     The mean difference from the subset of 34 good complete
retests was 0.00297 gallons per hour.  For complete retests,  the
variance of the differences between initial and retest rates
estimates twice the total variance; that is, the within-test plus
between-test components.  We report here the corresponding
estimated total variance.  The estimated total variance was
0.00254 gallons per hour squared, giving a mean squared error of
0.00255 gallons per hour squared.  If attention is restricted to
initial tests with estimated volume change rates of less than 0.2
gallons per hour in absolute value, the results change slightly.
For this set of 30 retests, the mean difference was 0.0137
                              D-25

-------
gallons per hour, while the variance was 0.00181 gallons per hour
squared.  This resulted in a mean squared error of 0.00200
gallons per hour squared.   Neither mean difference is
significantly different from zero (t = 0.059,  33 df,  t = 0.322,
29 df,  respectively).   The cases with larger volume change rates
were somewhat more variable, however.

     As noted above, there were two retests of tanks that had
vapor problems.  The initial test results showed volume increases
of 1.189 gallons per hour and 0.584 gallons per hour,
respectively, based on very short test times.   The retests based
on longer times gave volume increases of 0.175 gallons per hour
and 0.109 gallons per hour, respectively, with again the
conclusion of a trapped vapor pocket.  Both of these retests
agreed on the presence of vapor.  The difference in apparent
volume increase rates may be due to a number of factors.  The
initial test was terminated quite early.  The early termination
may have led to variable results.  The size of the vapor pocket
may have differed between the initial and retest.  The changes in
conditions—temperature, barometric pressure—that affect the
vapor pocket may have differed.  All of these could lead to the
observed differences in apparent volume increase rates.  However,
the consistency of the test and retest in identifying the tank as
having a problem with trapped vapor suggest that the test method
is consistent in identifying problem tanks.

     There were two tanks that were retested after the tank was
repaired.  One of these had an initial leak rate estimated to be
-0.057 gallons per hour with a standard error of 0.004 gallons
per hour.  The rate estimated on the retest was -0.017 gallons
per with a standard error of 0.0094 gallons per hour.  Although
the tank was considered to be leaking by the NFPA Standard 329
and the owner took corrective action, the volume change rate
estimated initially was fairly small.  The second tank had an
                              D-26

-------
initial leak rate estimated as -0.137 gallons per hour with a
standard error of 0.009 gallons per hour.   On the retest,  the
estimated volume change was -0.132 gallons per hour with a
standard error of 0.007 gallons per hour.   Little change was
observed.  However, on the retest, the testing company certified
the tank as tight based on the last hour of data, where they
estimated a rate of -0.044 gallons per hour.  The data from this
test showed little difference from the initial test.  Except for
the known fact that some repairs were done to the tank, there
would be no reason to exclude it from the retest data.  Even the
former retest would not be viewed as suspect from the change in
estimated leak rates.

     The retest data analysis showed no evidence of bias in the
test methods.  Both the back to back retest and the leak sim-
ulations estimated within-test (plus variation from one 2 hour
period to the next) standard deviations on the order of 0.025
gallons per hour.  The complete retest data gave a total standard
deviation estimate of 0.05 gallons per hour.
V.   ESTIMATION OF TOTAL VARIANCE

     The various types of retests offered not only a means of
estimating both within- and between-test variation,  but also
evidence that the between-test variation is sizeable compared to
the observed variance of a single test result.   In order to use a
statistical hypothesis testing approach to determine whether the
observed leak rate in a given test is evidence of a leak rather
than due to measurement fluctuation, the total variance must be
estimated for each test.  This was done by estimating the
between-test variation from all the data taken together and
adding this to the estimate of within-test variance generated by
the data from each test.  The within-test standard error was
                             D-27

-------
squared, the overall between-test variance added, and the square
root of the sum was taken as the estimate of total standard error
used in the leak status decision process.

     Two sources of information were used to estimate the
between-test variance.  The two sources agreed fairly well, which
served as a validity check on the results.  The two estimates
were then averaged (using relative weights based on the number of
cases each estimate was based on) to form the needed estimate of
between-test variance.  Table D-4 summarizes this process.

     The complete retests provided one data base from which to
estimate between-test variance.  For a retested tank i, let k
index the test (1 or 2) and j index the 5-minute volume change
measurement for a given test.  Then a given 5-minute volume
change measurement, X-j, can be written:
     xikj = Li + dik + eikj                    [Equation D-l]

where

     l>j_    =   tank i's true leak rate under test conditions;
     d^k   =   random measurement error of LJ due to differences
               from one test occasion to another; and
     e^kj  =   random measurement error of the individual
               5-minute volume change measurement for this test.

Since the various quality assurance double-testing methods showed
no evidence of bias, it is reasonable to assume that

          E(eikj) = 0
          E(dik)  = 0.
                             D-28

-------

-p
o
c

en
0)
-P
fO
rH
Q)
tO
A
0

Q)
g
j3
<-\
O
£>

T5
Q)
>
Q)
01
XJ
O

c
0

TS
Q)
U)
fO
X*
— '
Q)
O
C
to
•H
^_l
to
r> '—•
0)
-p M
01 0
Q) Ifl
-P 01
1 0)
C rH
o) a
0)
5 -p
-P 0)
0) 0)
XJ -P
o o
(H
01
Q) T!
-P Q)
fO -P
B 01
•rH £
-P -1-1
CO T!
H tO


.
•3*
1
Q

0)
rH
X!
10
£-)

O
0) 0)
1 \ 1 ^—
fO fO 'O »^H
B -P 5H a
•H O fC &>
-P -P T3 N —
0) C
W tO
-P



4-)
01
T! Q) QJ
Q) -P UCM
-P 1 C —
to c to x!
E o)-rH a
•H o) fn tn
-P 5 rO -^
03 -P >
W 0)
XJ




-P
T3 01
0) 0) 0)
-P -P OtN
fO 1 C ^
B c to x;
•H -H -H fl.
4J X! rH &1
01 -p <0 — •
H -H >
•^




0)
o
c
T3 fO
Q) -H
-P rH ^^*
tO tO X!
B > a
•H &i
±> <-\-^
01 tO
H -P
O
-P



^








0)
01
to
CQ

tO
-P
fO
Q






^.
O
in
o
.
0








fSJ
f\j
(N
O
O
.
O









CO
CO
o
o
o
.
o










^
in
CM
O
O
.
o





^
CO







-P

Q)
-P
0)
}_l

Q)
-P
0)
rH
a
g
o
u

^
rH
in
o
.
o








CO
CTi
rH
O
O
.
O









CO
t>
o
o
o
.
o










f^
VO
OJ
o
o
.
o





CO
CO
H






X! 0) XJ
-P B C a
•H d 0) cr>
^ rH Q)
O |5 (N
01 > -P •
-P 0) 0
01 T3 XJ
Q) 0) Ti
-P rH Q) C
J3 fji fO
rX 01 q
c to to o
fO Q) XI •
&H B o o

<*
H
in
o
.
0








o^
cr*
rH
o
o
.
o









in
VO
o
o
o
.
o










*3*
VD
CN
O
O
.
O





r.
VO
rH




0)
-P
tO
n
•r-\
-P
01
Q)

T3
Q)
C
-H
O
B
O
u
D-29

-------
We also assume that




          E(di:Ldi2) = 0,


          E(ei:Lei2) = 0,
and that the dj^ and e^-i each have a constant variance,  denoted

as
             = between-test variance
and
          a2 = within-test variance
           w
where the mean of the e^j is taken over all measurements for the


k-th test of the i-th tank, usually 24.




     Starting with Equation D-l, an estimate of total  variance


can be based on the two estimated  leak rates,  Xj^ (the initial

rate) and x^2  (the retest rate) as follows:
          E(Xil  - xi2)


          = E(dn +  ei;L  -  di2  -  ei2)2



          - E(dn -  di2)2  +  E(ei]L  -  ei2)2




because the d^k  and  e^k  are independent.   This,  in turn, equals:
                 2«2.
                   w
                              D-30

-------
Thus


                               -  xi2)2)

and
                      n    2
          E((l/((2)-(n)))i:  E  S?  )  =
                                     w
where
 ik
                                iic " xikj)
     Therefore, letting
           b,retest
                                          - xi2)
    n   2  o
    Z   ZsJ

    L-l k=l 1]
                                                   [Equation D-2]
we have
E
-------
     Tests on tanks which can be assumed not to be leaking
provide a second estimate of 6"2.  Here,  the true leak rate is
zero, and we have

          xij = di + eij                       [Equation D-3]

with assumptions on d^ and e.^ as stated above.  (We suppress k
since only one test was done on these tanks.)  In this case we
have
and
                              C n	    n  ^ /
                                                  [Equation D-4]
Clearly
          E(S2            ) = o~2 .
            v b, tight tanks'    b
Defining tanks which can be assumed not to be leaking requires
some decision-making.  By limiting this group to tanks with
measured average volume change between 0.0 and 0.2 gallons per
hour, the tanks which may be leaking  (negative measured volume
change) are eliminated as are the test results which are likely
due to vapor pockets (high positive measured inflow) .

     The results of applying Equation D-3 to the 34 retests and
Equation D-4 to the 133 measured volume changes between 0.0 and
0.2 gallons per hour are shown in Table D-4.  It can be seen that
the two approaches yield similar estimates and in particular
indicate the importance of the between-test component of the
                              D-32

-------
total variation in x^.  It should be noted that these figures are
all as measured,  and not as adjusted for test pressure.   The
adjustment deflates the measured leak rate by about half (the
factors range from 0.395 to .608), but is applicable only to
actual leaks, since it adjusts the rate from test pressure to an
assumed operating pressure.

     To get one estimate of between-test variance to use in
adjusting within-test standard error up to total standard error,
the two estimates described above were averaged with relative
weights based on the number of cases each was based on:
     34/167 (0.00222) + 133/167 (0.00193) = 0.00199

Thus, to estimate the total standard error for a given observed
leak rate, 0.00199 was added to the reported (within-test)
standard error squared, and the square root taken.  This total
standard error was used in the statistical hypothesis test method
for determining leak status described in Section 8 of this
report.
VI.  DETERMINATION OF LEAK STATUS

     The physical tightness test for each tank system provided an
unbiased estimate of volume change rate and an estimate of the
within-test variability of that rate.  The complete retest data
provided an estimate of the between-test variability of the
measured rates.  However, the test itself did not provide a
definitive leak status determination, that is, an unequivocal
"yes" or "no" to the questions "Is this tank tight?" or "Is this
tank leaking?"  In order to estimate the number of tanks in the
country that are leaking and to look at the subset of leaking
                              D-33

-------
tanks to investigate factors associated with leaking,  such a
determination must be made (or the test result ruled
inconclusive) for each tested tank system.   Two approaches were
considered for making this determination:   a cut-off rule,
comparing the observed volume change rate to a pre-determined
cut-off; or declaring a system leaking or not by a hypothesis
testing approach.  The latter approach was chosen for the study
determination of leak status.  Two drawbacks of the cut-off
approach were that there was no scientific basis for establishing
a specific level for the cut-off at the time of the survey, and
that it did not take into account the differences in precision
achieved by the individual tests.

     The null hypothesis to be tested in determining leak status
is:

               H0  : Li = 0

where Li is the true leak rate of the tank.  The alternative is

               HA  : Li > 0.

As shown in Part V, above, we model the test result, "x^, as
having a total variance composed of a within-test and between-
test component.  This total variance is estimated as
                2 = S2.+ sj
                t    wi   b
where the first term is the within-test variance measured from
the i-th tank test data and the second term was estimated as
described above (Part V).  The test statistic is therefore
                              D-34

-------
and is compared to one-tailed tables of the Normal distribution
to determine whether HO can be rejected at a certain level of
significance.  If HQ is rejected,  we say the tank system is
judged to be leaking.

     Several significance levels were examined,  as was the trade-
off between significance and power.  The power was estimated for
a specific leak rate after adjusting the leak rates and their
associated standard errors for test pressure (see Part VII,
below, for this adjustment procedure).   A significance level of

-------
logically consistent to adjust test volume change rates for pres-
sure in the event that the system was judged to be tight.

     The adjustments are based on Bernoulli's law.  More
specifically, adjustments are based on Torricelli's form of the
Bernoulli equation.  In order for the adjustments to be reason-
able, the assumptions for these physical laws must hold.  It
should be noted that the assumptions for Torricelli's and
Bernoulli's law assume that the flow is through an orifice with
neither resistance nor turbulence.  In practice, this is not the
case.  While the flow rate will be generally small enough so that
the assumption of a turbulence is reasonable, and so that the
head change is slow enough to be neglected, in most cases, leaks
will probably be through corroded sections and will be into soil
which may present some resistance.  The effect of resistance
would be to lower the flow rate.  However, how much the flow rate
would be lowered under the different pressures is not known.
Consequently, the effect of violation of these assumptions on the
adjustment to the leak rates is not known.  It is assumed to be
negligible.  There are some other, implicit assumptions.  These
include that the orifice is constant, that the temperature and
density do not change, and that the product is not viscous.

     Torricelli's form of Bernoulli's law can be used to
calculate adjustments to the flow rates.  In order to do this,
several assumptions must be made.  The set of assumptions used in
these calculations is detailed below.  A step by step procedure
for the adjustments is given first.  These are the adjustments to
be made in the ideal situation where the tank system leak was
quantifiable and a valid line test with quantifiable leak rate
was done.  In our data base, among tank systems judged to be
leaking with quantifiable leak rates, only 39 percent had valid
line test leak rates.  Since the majority of cases had no valid
line data and the separate analysis described in Section 8 of the
                              D-36

-------
report showed that line leaks accounted for a very small
proportion of system leaks when they were done,  leak status and
leak rate as reported in the Major Findings are based on measured
tank system leak rates adjusted directly to operating conditions,
without adjusting for line test results.  We present the line
test adjustment procedure since it was used for the analysis in
Part V of Section 8 and for future use in analyzing data
collected in the national survey.
     A.   Adjusting the Line Leak Rate to the System Leak Rate

     Since the line test is conducted at higher pressure than the
system test, the leak rates estimated from the line test are not
directly comparable to those estimated from the system test.
This adjustment accounts for the difference in pressure and
adjusts the line test rates to be comparable with the system test
rates.  These adjustments are calculated differently for pressure
systems and suction systems and for gasoline and diesel fuels.

     The assumptions made for this adjustment are the following.
These are in addition to the assumptions needed for the use of
Bernoulli's equation to adjust the flow rates.
          The orifice where the leak (if any) occurs is where the
          line joins the top of the tank.
          The tank is assumed to be buried to a depth of 3 feet
          to the top of the tank.
          The water table is assumed below the bottom of the
          tank.
          Three tank diameters are assumed: 48 inches, 64 inches,
          and 96 inches.
                              D-37

-------
     Table D-5 gives the adjustment factors to adjust the rates
estimated from the line test to the conditions assumed for the
system test.  The factors as presented are multiplicative.  To
convert a rate estimated from the line test to the equivalent
system rate, multiply the estimated line rate by the factor in
the table.

     The difference by type of delivery system results from the
fact that the line test is conducted at 15 PSIG for suction lines
and at 50 PSIG for pressure lines.
     B.   Subtracting Line Rates From System Rates When Valid
          Line Results are Present

       After adjusting the line test results by the factors in
Table D-5, the line test rates would be comparable to the system
test results.  The line test rates could be subtracted to obtain
an approximate tank rate.  This is the rate for the tank system
excluding delivery lines, but still including any other plumbing
such as fill pipes, vent pipes, etc.

     If a system has more than one delivery line, each line test
rate would be adjusted, then all line test rates subtracted from
the system rate.  For the tank systems for which the line was
found to be untestable, the line rate cannot be separated from
the system rate.
                               D-38

-------
Table D-5.  Adjustment factors for line test rates
        Tank  diameter             Suction       Pressure
          48  inches                0.431         0.236
     (0  -  1,000  gallons)


          64  inches                0.395         0.216
   (1,101  - 7,000  gallons)


          96  inches                0.317         0.174
   (7,001  - 15,000 gallons)
                            D-39

-------
     C.   Adjusting the Tank Rate (or System Rate)  to Assumed
          Operating Rate

     Since the test is conducted at elevated pressure, flow rates
through any orifices will be larger under the test conditions
than they would be under actual tank operation.   The magnitude of
the difference depends on a large number of variables.  In
particular, flow rates would vary by location of the hole in the
tank (distance from the bottom), amount of fuel  in the tank, and
pressure of a water table part way up on the tank.   The
adjustment factors would also vary with diameter of the tank.
Since diesel tanks were tested at the same pressure (hence at a
lower head-distance) as gasoline tanks, the adjustment also
varies with fuel type because of the density difference.

     The standard assumptions for calculating the adjustment
factors presented in Table D-6 are as follows.  These are in
addition to the basic assumptions of Bernoulli's law.

     o    The water table is assumed to be below the bottom of
          the tank.
     o    The tank is assumed to be buried to the depth of 3 feet
          from grade to top of tank.
     o    Three tank diameters are assumed  (48,  64, and 96
          inches).
     o    The average operating level of the tank is assumed to
          be half full.
     o    The orifice or hole  is assumed to be in the bottom of
          the tank.

     Table D-6 then gives adjustment factors to adjust the
estimated tank system leak rate to the assumed standard set of
operating conditions.  The factors should be multiplied by the
leak rate estimated under the  system test to obtain the adjusted
                               D-40

-------
Table D-6.  Adjustment factors for tank (system) rates*
                                  Adjustment factor
     Tank diameter              Gasoline       Diesel
        48 inches                0.395         0.430
   (0 - 1,000 gallons)


        64 inches                0.456         0.496
 (1,101 - 7,000 gallons)


        96 inches                0.558         0.608
 (7,001 - 15,000 gallons)
*If a standard height had been used for both fuels, the
 gasoline column would apply to both.
                         D-41

-------
leak rate.  Note that this adjustment can be done to the system
test leak rate, or to the leak rate remaining after any relevant
line leak rates have been adjusted to test conditions and
subtracted off.

     Multiplying the rates estimated under the system test by the
adjustment factors given in Table D-6 will give adjusted rates
for the assumed standard set of operating conditions described in
the assumptions above.
                               D-42

-------
                            APPENDIX E

                 INVENTORY  RECONCILIATION METHODS


I.   EPA INVENTORY RECONCILIATION METHOD

     EPA has developed a simple method1 for monitoring
underground motor fuel storage tank inventory records to detect a
systematic deficit which may be attributable to a leak.  The
method is based on counts of the number of daily underages found
in the inventory record and is simple enough to be implemented by
a tank operator without excessive calculation or burdensome
record-keeping.  As originally formulated,  the method is intended
for application as the "first line of defense against leaks" in
an on-going monitoring program.  Thus,  the approach is sequential
in nature and involves making a decision on the presence or
absence of an inventory deficit at the end of each 30-business-
day period, based on a comparison between the cumulative count of
daily underages and certain statistically-derived "action
numbers"1.  A cumulative number of underages in excess of the
appropriate action number was to be interpreted as evidence of a
deficit.  The statistical model and calculations underlying the
method were detailed in the report from Battelle Columbus
Laboratories to EPA2.  The basic method required modification for
application to the inventory data collected in the survey because
each sampled facility provided only a single, one-time record of
^U.S. EPA, Office of Toxic Substances, "More About Leaking
 Underground Storage Tanks:  A Background Booklet for the
 Chemical Advisory," (October 1984).
2David C. Cox, "Performance of the Chemical Advisory Inventory
 Analysis Method Under Various Scenarios," Report from Battelle
 Columbus Laboratories to EPA under contract No. 68-01-6721
 (April 1984).
                              E-l

-------
30 days' inventory for analysis.  The purpose of this section is
to describe the statistical model on which the modified EPA
method is based.

     The decision rule for the proposed method will be defined by
considering a well-run station where the only sources of
discrepancy in the inventory records are (i) a daily leak of
magnitude L and (ii)  unavoidable random error in the daily stick
measurement of the tank.  Successive daily errors are assumed
independent and identically normally distributed with mean zero;
this assumption is supported by the research of Warren Rogers3'4.
Hence, we can write:
               xi = xi + ei

where X^ is the ith daily stick measurement, Xj is the true
quantity of gasoline in the tank at the close of the i**n day, and
          ^
e^~ N(0, Q  ) is the stick measurement error.  Now consider a
period of n days, assuming for simplicity that the station is
open every day.  The process of balancing inventory at the end of
each day, as described in the literature5 and assuming that there
is no metering error at the pump6, leads to a set of daily
variances (discrepancies),

          d  = -L + e  - e_, i = 1, ... n.
3"Inventory Reconciliation system," Warren Rogers Associates.
4Warren Rogers, personal communication.
5American Petroleum Institute:  "Recording Practices for Bulk
 Liquid Stock Control at Retail Outlets,"  (1977).
6Metering error, if present, can be estimated and removed  from
 the record, see American Petroleum Institute,  "Recommended
 Practice for Bulk Liquid Stock Control at Retail Outlets,"
 (1977).
                               E-2

-------
Let N be the total  number of negative daily variances,

          N =  #{i|l < i < n, d-L < o}.

     Clearly,  large values of N suggest that there is a leak,
i.e., L > 0.   The exact probability distribution  of N is,  in
general,  very difficult to derive.  However, of  the special case
of no leak, i.e., L = 0, the calculation has been carried out7.
Table E-l shows  the distribution for the case n = 30 of most
interest.  In  general,  we must rely on a normal approximation to
the distribution.   This is derived as follows.  We first find the
mean E(N) and  variance V(N) as follows.  Define:

         P - Pr(d1 <  0)
         P1 = Pr(d. < 0, d.+1 < 0)
         Ij « 1 ,  if  d. < 0
             0 ,  else
                        2
Then £(1^) = p, E(Iflj) * p  If |j-1| > 1  (because then 1^,  I. are indepen-
dent), EUjIj+j) «  pr  Thus
         E(N) = E(i|1 I.)  = np.  Also

         E(N2) - l(ft l\  + 2.Z. 1^
              = np + Ztn-Dpj + [n(n-l) - 2 (n-l)]p2
Therefore
         V(N) = E(N2) -(E(N)f                                     [1]
             « np(l-p) - 2(n-l)(p2 - P])
             s a(L)2
VWarren Rogers,  "The Exact Null Distribution of the Number of
 Negative  Daily  Variances," Report from Warren Rogers Associates
 to EPA,  (September 1984).
                               E-3

-------
Table E-l.
Probability distribution of the number of negative
daily variances, N, for the no-leak case, based on
30-day inventory
No. Of
negative variances
< 10
11
12
13
14
15
16
17
18
19
> 20
Probability of
occurrence
0.0024
0.0121
0.0456
0.1161
0.2022
0.2432
0.2022
0.1161
0.0456
0.0121
0.0024
                               E-4

-------
We approximate N by a normal distribution with mean np +0.5 and
             ^
variance a(L) .  The mean is taken as np +0.5 to provide an
approximate continuity correction for use in the upper tail of
the distribution, in which our greatest interest lies.

     To check the accuracy of the approximation, consider the
case L = O.  Then,
               Pi = pr n)
(exact)
0.6216
0.3784
0.1762
0.0601
0.0145
0.0024
Pr(N > n)
(approximate)
0.6217
0.3783
0.1788
0.0630
0.0162
0.0029
                              E-5

-------
     Clearly the approximation is sufficiently accurate over the
range of n reported.  For L / 0,  the exact distribution of N has
not been derived.  We will rely on the normal approximation in
such cases.  The mean and standard deviation of the approximating
distribution have been calculated and are shown in Table E-3.
Table E-3.
Mean and standard deviation of normal approximation
to the distribution of N, the number of negative
daily variances, for various values of the daily
leak rate L, for a 30-day inventory
L
(gallons)
2
3
4
5
6
7
8
9
10
Mean
16.46
16.93
17.41
17.88
18.34
18.81
19.27
19.72
20.16
Standard
deviation
1.636
1.641
1.647
1.654
1.665
1.678
1.684
1.699
1.707
     The final feature for which we must account before we can
determine the decision rule is round-off error.  In practice,
inventory values are typically reported to the nearest gallon so
that an exact inventory balance, i.e., a zero variance, can occur
due to round-off.  This is fairly common in actual inventory
data.  We will assume that a zero variance is reported if the
actual variance is less than 0.5 gallons in absolute value.
Thus, a negative variance is reported only if the actual variance
is less than -0.5 gallons.  Let N* be the number of negative
variances actually reported and assume a/T =25 gallons.   Then
the distribution of N* should be approximated by a normal
distribution with mean and standard deviation shown in Table E-4.
                               E-6

-------
Table E-4.
Mean and standard deviation of normal approximation
to the distribution of N*, the number of negative
daily variances accounting for round-off error, for
various values of the leak rate L, for a 30-day
inventory
L
(gallons)
0
1
2
3
4
5
6
7
8
9
10
Mean
15.26
15.74
16.22
16.69
17.17
17.64
18.11
18.58
19.04
19.49
19.94
Standard
deviation
1.633
1.634
1.635
1.638
1.644
1.650
1.660
1.672
1.681
1.687
1.703
     Now suppose we have 30 days' inventory and there is no leak.

Using the approximating distribution from Table E-4 the number of

daily variances observed should have the distribution shown in

Table E-5.
Table E-5.
Probability distribution of the number of negative
daily variances, N*, observed when no leak is present
n = number of
negative






variances
15
16
17
18
19
20

(Pr(N* >
0.564
0.326
0.142
0.047
0.011
0.002

n)






                              E-7

-------
Thus, if we make 18 or more negatives our criterion for deciding
that a deficit is present, there is approximately a five percent
false-positive rate.  That is, a tank with no leak and no source
of error in inventory other than random measurement error due to
sticking has approximately a five percent chance of being
erroneously classified as a leaker.  Note that false-positives
due to other factors such as theft are not accounted for here.
The detection capability of this version of the EPA inventory
analysis method can now be calculated using the values given in
Table E-4.  Results are shown in Table E-6.
Table E-6.
Probability of detection of leaks of various sizes
using the modified EPA inventory method based on
3 0 days'  data
Actual leak
Gallons/day
1
2
3
4
5
6
7
8
9
10
Gallons/hour
.04
.08
.12
.17
.21
.25
.29
.33
.37
.42

Detection
probability
0.08
0.14
0.21
0.31
0.41
0.53
0.64
0.73
0.81
0.87
Thus, leaks of at least nine gallons per day or more have better
than 80 percent chance of detection.  It should be noted that the
detection capability of the simple inventory method based on only
30 days' data would be expected to be poor.  The method was
designed, as explained previously, for use as a tool for on-going
monitoring programs.
                              E-3

-------
II.  WARREN ROGERS ASSOCIATES' INVENTORY RECONCILIATION METHOD

     Warren Rogers Associates (WRA) has developed a computerized
system for analyzing daily inventory data from underground
storage tanks in order to identify leaks8.  The details of the
method are proprietary.  This section provides a brief
description of publicly-available information on the model and
should not be interpreted as an evaluation or endorsement by EPA.

     The WRA system was developed in response to the perceived
inadequacy of conventional, routine inventory accounting in
detecting small or moderate leaks.  Typically, such leaks are
masked in the data by a variety of errors.  For example, a single
delivery error of 300 gallons could mask a 10 gallon-per-day leak
based on 30 days' inventory.   The purpose of the model is to
isolate, identify, and quantify these errors.

     Errors accounted for include:

          Delivery errors;
          Unexplained additions;
          Pump meter error;
          Temperature effects;
          Stick error; and
          Tank or line leaks.

Occasionally, other, rarer, errors will appear, e.g., use of an
incorrect tank conversion chart, or theft.  The data required by
the model include only daily stick readings, deliveries, and
sales.
8Warren Rogers Associates, Inc., "Inventory Reconciliation
 System," (undated).
                              E-9

-------
   The  basis  for  the model  is that the major errors and
discrepancies in the inventory data are very distinct in their
characteristics and thus in the way they contribute to the total
record.  Thus, for example, an unrecorded over-delivery or an
unrecorded removal will cause a permanent shift in the record
which remains as a fixed component in all future observations.
This effect can be estimated and removed from consideration when
evaluating the possibility of a continuing day-to-day trend
indicative of a leak.   By contrast,  a large stick error caused by
a mistake in reading the stick or conversion chart will typically
cause a large discrepancy in that day's inventory which will be
followed the next day by a discrepancy of similar size in the
opposite direction.  The two discrepancies will tend to cancel
out in the cumulative inventory record.  The "signature" of a
pump meter error is different:  such an error will induce day-to-
day errors of constant sign proportional to the through-put of
the tank.

     WRA's report to clients includes a record of day-to-day
variances and the cumulative variance between book inventory and
stick measurement for the period.  It also provides:

          Over-  or under-deliveries by date of occurrence and
          amount.  That is, the discrepancy between the amount of
          product actually delivered as opposed to the amount
          reported;
          Unexplained one-time gains or losses also by date and
          amount;
          Meter  errors at the pump;
          Trends which are indicative of either a tank or line
          leak;  and
          Effects of possible disparities between the ambient air
          temperature and underground temperature.
                              E-10

-------
As a special contribution to this study, WRA also provided a
"data quality code" based on professional interpretation and
experience.  The data quality code is explained in Table E-7.  A
sample WRA inventory report is shown in Figure E-l.  Based on a
discussion with the developers of the WRA model, the false-
positive rate is five percent, comparable to the modified EPA
method.
Table E-7.  WRA data quality code
          Category
          Definition
              1
              2

              3
              4
Confident of the result
The trend could have been
delivery-induced
The trend is noisy but believable
No confidence in the trend due
to the data
Data is questionable and requires
further investigation.
III. ENTROPY LIMITED INVENTORY RECONCILIATION SYSTEM

     Entropy Limited has developed the Precision Tank Inventory
Control (PTIC) system9.  The analysis is based on principles
similar to the WRA system and accounts for the same types of
errors and discrepancies.  Entropy appears to consider thermal
effects and vapor losses more comprehensively than does WRA.
However, additional input data to the system is required for
these analyses.
9Entropy Limited, "Precision Tank Inventory Control,"  (1984)
                             E-ll

-------
        CO
a.
o
z
OOUIO    CO

h-ZCOf««  CO




O    T  Z
O  00  >
  O) •    QC
       .  O
LUUJOO  Z I
OCOa.<4»  01 I
a  3—  > I
UJ
OC
           _>
           —•

           <

           Ji I\!  I -^ |   fSJ •$• U"N ^ <\l "•<        I
           <      I      I  I  I  I  I  I           I  I  I  I  I  I  -• I -i-i-i — —i-i-i
           >                                             I    I  I  I  I  |  I  I


           s:




           LU

           T
           <
           M4***«**««*«*****»**«««t«t»«*««*
           o£ tn o r*« if\ i*~ \f\ r** eo O —* -o ir» so oo o •*• o —' * ^ o Is* ?w     i       77        ^     T i  i   ^   •  i   ?T   i   *^V
                                                                              LU
           >                                                                  O
           -4                                                                  Z


           O          '                                                        rX


           Sj-^ 00 O «O'J'00 M f* <*> GO «^ I/MA-a" O ^-VO r-O-^ «O I*-C7< «M »O iTv m 
                                                                              1-4
                                                                              H-
                                                                              <
                                                                              _J

                                                                              s.
                                                            ,^^ir\rgco^—<   3
           «-»O»^Of*«oornt*>^i^O>p">o^>cooiA>-*OpgfM^oi^Ofs)>J'^-<(M'aeo   o

           cO                                                                  Q
                                                                              a


           cO                                                                  LU
           oi                                                                  a.

           0:000000000000000000000000000000   u.
           uj       o        m       <-<         o         •*  . .,      o


           -J                                                                  Z
    O     OI                                                                  01
    o     o

3  O
OI  -*        ttt»t»t«»»t«»»«»«»»»»»»»»»»»»t

           oi u> »o «M •* o *• 
                          -  -•-•                        •^mcor«»'Ococo^o
                                                                                                           OJ
                                                                                                                   at
                                                                                                                   a.
                                                                                                                   ai
                                                                                                                   CC
                                                                                                                       o
                                                                                                                       2:
                                                                                                                       <
                                                                                                          UJ
                                                                                                          Z

                                                                                                          U.
                                                                                                          O

                                                                                                          oc
                                                                                                          UJ
                                                                                                          03
                                                                                    O)  X
                                                                                    >  I-
                                                                                    Ul
                                                                                    a
                                                                                                                               O
                                                                                                                               O4
                                                                                                0
                                                                                                •p
                                                                                                c

                                                                                                I
                                                                                                H
Q)
M
3
                  .42
                   •««•   UJ
>

o
                  •-O   (rt
                  zo
                  
-------
                                  LU
OOLUO    00

t—zoor-  uo
<>  oo  >-
  LU »    CC
          o
          H-
LULUOO  Z
'JQ3tX^  LU
O  3—  >
aiALU    Z
  OZ    —

LU

ac

3
%-. *
O I





>.
•
3£ «h-^f
»•« Q— •
Q LU Oj 
Z LU Q«-l
LU f— ^ LT>
oC rO O O> ^
H* • <-« • O
00 K- ^0 "^ X
>. I 00 K H»
_J 00 Z^
1-1 LU a
< -j LU s:
o  _J
«4









_J
<
wt
H>
Z
LU
aj
LU
u,
a.
b«|
Q

LU
ae
^
H"
<5
a:
UJ
a
5!
LU
b^
^^







TREND

»0 LL.
>0 O
•
rsi LU
1 LJ
Z
^
^j
<-«
LU
^4
Z
s_D
i_i
oo









                                                                                             Z  O
                    LU
                        LU
                                                          E-I3

-------
     The PTIC system reports its leak findings as an estimated
leak rate, in gallons per day,  and as a "probability of leak"
(see the sample inventory report in Figure E-2).   According to
the model's developers, the probability of leak is based on a
Bayesian-type analysis which accounts for various factors
including the quality of the inventory data.   Details are
proprietary.  Typically, the decision rule is phrased in terms of
the leakage probability as follows:

          Leak probability                 Decision
               < 10%                    Tank is tight
             10% - 50%                  Inconclusive
               > 50%                    Tank is leaking

The 50 percent cutoff point corresponds to a false-positive rate
of approximately two percent.  To obtain a more typical five
percent false-positive rate, a cutoff of 30 percent leak
probability should be used to decide that the tank is leaking.
                               E-14

-------
                            939
II
••*
x
|

s.
>«
g*
1! II
»ei
:£
dz
s_
1-
c
1 t»
•>V NM
25
E°
Sr rj
Sr.;
£
Ml
II
l«

«-
u •
«g
££
»*
N u
"1
* X
MM
55
*v *•
Sf-



UN
;j
s

M
_

rl"
'X
" £
W
11



I/,

S. 5
P ^ 2
'-* J- '-5
•» \* ""
i" 7J 2 '~
ci — s 3
0 =S - K
y i! «:
t»: ,> M u
*H*1
IJSSS
P
Z CS O O
1^
>
55
EC
t_
<
C
r-
S
M4
-J
—
i
i
c
ff.
1
S
S
x
2

ss
n^
C'
^
KI
,
t^*
H-*
^
t-l
P
tr
»
SC
^«

fe
dk,
^5
^
1^
T
u.
y
to

a
z


2-®?';*C>JX*CW?l*'~NIB9v~f1!:>5"':''B;*~'*'~*'L":*®'fflt<*1!'
S *«' ,«I '^ e- *» •£• — ~^ j c^ ~ ?• ?i — c ?• c- rt •• c? •* c: f- 2 — d ^ •* '-" t' 3
3 i — — i N — i~ « *• 3 K — ~ (• l>- S « W O CS ^ C S i f ' !- i» C l»
y ' -,_»c -.', ," , 7 -^.- i i i K i - i i i N i i i « i
— 1 1 1 II 1


>SO®OS>C«Cs>COC®C*ttCSOOOffOO'JO©SCC*
.•^^,^.«,— ,^ — »,<.,—, ^., — j.^cCCOSC'C'OSC'CC'CCS;®
.^ --=,-• ^ o « O
J5 -M .•«•-»«<»*


i;c«s2:®s-. c 5>cocc>©c®e;*s>**®5>c»eo»oco'eo
^ -i j .^ -* ,-! .. «,' ,>'. ~ i« -^ ->' -^ -j — — c — rj -i .- — »i ^ i^" N! — ^ -* !_•? o
», s, f ••! -- i S- i i .-. -; *5 * ;; i r- ?i -i —• '•"• 13 i^ *f -; K :i S s^ fi " ?C c. s>
;• r: ;-i i^ t> e; i» i^ ;: — fr i* ?i « f ! f < «>• !- w ^ v — *•**« n f J c N j. 3
X - - - -

£ 3 S •* 3 SS * » « S * 9- W Ift — -• " N N L-5 l» rt f '. >* N «> W — O » N »
' ^ tw O * M J *• — ?-'*MCrjCJ"C'J' — — N — C — '*^1«*>Cfl"}'5>O
E * •* s: K « w ^ f j N » «f n r> c c ~ t- •• > N c ; K :: i; «; f i * — * r? cj
• •—..-^.AlAlM^ •Mwn^«^AlAiv4 VHVMW^^lw^'lwvww^AJN ^tlMMvM
.^ p« w« ^ pj g>| ^ "• •• ^ ^^ , 4 i,^ ^^ »« •« **fcv »f
X
a
£^p§^i§|8§§f§§P§§fip§pSSp§
INri^wNoaid^N^ttoNa^^^gSS^S^g^-^"*10
NNN\\\\\\\\X\V.X\NNX\N.S.XXV.X\\\\\
--«e9*»*NOo.g;;fi»*w*Na9:e-«»*«v9Na»«-



O.N
B ^
^ w
K-
S


> o
J»
&c a>
— L-5
IA


[£ o
<^i
K TS
8*
K

at«
Jk •
i ""
£2
— 1A
u
N4
e
b^
> O
y rt
I
•^
P



5° 4ii
C5 I1-*
* *
w w
n t*"

Nrt
«
u II

s. r-
-; -3
^ ^
y^ ^>
S A
^\ ^
+ 1
— •"
^ V ^

r **
SKM
^SS
*•" < ^1
=: aa
pi
S S!
t. M S
gig
pi
S
VFER INFLOW
EC
c;
!?"*
<
X
x
•^
•^
_,
S
<
>
£
*,
c-
«I

U,
^ ,
H
£3
A''
U**

^
o
O.
i
•j
rr ff,
^ *
^ ^i
X u
> >
aa








f — S b» Lt
;> — ^ •* r>
?•- a C M »•
C' G C- — —
. . . -l .
C •* •*•• !- O
n c M i





n

a
n U
X
~ < II
Pi
M -ay
^ SM?
1" .si
§oaai
o . < —
u<:x> i,
> » > <
aaaa-j
Entropy Limited Inventory Report

•
CS
1
W
(1)
M
3
hn
O1
•H
PM

















E-15

-------
                                                                                              L'C'z
                                    f a
                                    B t« — O
                             W

                             c:
                               — ,s> r-
                                d ~*

               EL s g zr a
                     a
                     z
                     *;

                     e
                                    — — — C ® LT SS C> <


                                    V f^> **1 I** -*, >• *». —. ;

                                    fc is " e 1 ~ F- " '
                                                            §
  c? N


  K O
  S\i N
  N rt

   N
      :~ .& a a a
      >>;>>>

      O OCOO

      >;>>->•>.


      E2 « 2 3
                                                            9

                                                            9
                                       = D
  7) 'J3
 :.- >•
^aa
* tf 3S
t-Ka

Sig
c5=>
  z z
                 >
               s. a a a a
               n n ii n n
                z
                -*

                c.
               IT •* \O N ••   —
               t- sc w o 9-   y
                  — N —   —  a a
        2

        3
        3
   O   SS
        3    s
a a  =:
     — W C5 Ov W O
     j  — « ff,) »
                E
                            sr U
                7  S3
             — in ift L-S LO «  a
                            •JO
           icss  o  -;
       \ \ \ \ \
       NN& « O *
     a  — « n
       \ \ \ \ \       „
       ----CJ  -  -<
a  ~ =
a  a a
        a    a
                         I
                                               o
                                              ii ^
               . _ _  ;32  -
          -I    -. ~ fr  ^, \ ^. 'i
          ^    _;j:a    z-3-
          ft    <• < w  &- — <
          i,    £2 ^p    i^  — —   ni;ic

          >.  ?^^^2^.1pg

          §  ^g2'^S33-s

          S  S-^^-2S§3
                                                         w
                                                        si
                                                        s. 'J
                                                        O'JO
                                                        a -«

                                                       . >«M
_K<


!£g
                                                            a
                             -aJ-H>SOS:3  -
                                               a 75
                                  O
                                                    - J
                                    :: a
                                    29
                                                     E-16

-------
        APPENDIX F






DATA COLLECTION FORMS AND MATERIALS
             F-l

-------
                           APPENDIX F



                        TABLE OF CONTENTS
                                                             Page
Open Letter to Owners and Managers of Underground
  Motor Fuel Storage Tanks; W.D. Ruckelshaus, United
  States Environmental Protection Agency, October 15,
  1984	   F-ld

General Instructions:  U.S. Environmental Protection
  Agency Underground Storage Tank Survey	   F-5

     Instructions for Completing the Inventory Sheet
       for Tanks with Metered Dispensing Pumps	   F-24

     Instructions for Completing Dispenser Meter
       Recording Sheet	   F-27

     Instructions for Completing the Manifolded Tank
       System Recording Sheet	   F-31

     Inventory Sheet for Tanks Without Metered Dispensing
       Pumps	   F-40

Reporting Responsibilities of Tank Owners and Operators...   F-43

Certification Statement for Establishments Without Tanks..   F-44

Establishment Operator's Questionnaire:  United States
  Environmental Protection Agency Underground Storage
  Tank Survey	   F-45

     Tank to Dispenser Meter Fuel Line Connections	   F-76

     Site Observations Recording Sheet	   F-77

Call Record/Case Folder:  United States Environmental
  Protection Agency Underground Storage Tank Survey	   F-79

Instructions:   Preparation for Tank Testing	   F-83

     Environmental Conditions Data Sheet	   F-84

     Temperature Profile Data Sheet	   F-85
                               F-3

-------
                      APPENDIX F

             TABLE OF CONTENTS (Continued)


                                                        Page

Site Diagram and Detail Diagram Sheet	   F-86

Picture Description Data Sheet	   F-87

Critical Features Data Sheet	   F-88

Edit Checklist	   F-89

Simulated Leak Test Volume Data	   F-90

Simulated Leak Data Form:  Leaking Underground
  Storage Tank	   F-91

Petro-Tite Data Sheet	   F-93

Line Test Data Sheet	   F-94
                           F-4

-------
                        WASHINGTON D.C  20460
            OCT15 iy«4
              OPEN' LETTER TO OWNERS AND MANAGERS OF
               UNDERGROUND MOTOR FUEL STORAGE TANKS
    The Environmental Protection Agency  (EPA)  is conducting  a
national survey to learn more about the  problem of  leaking
underground motor fuel storage  tanks and piping.   The  purposes  of
the study are to 'find out how widespread the  leakage problem is,
and to coilecc information on factors  that cause tanKs to  leak.
The study will help the Agency  assess  the impact of  leaking  tanks
on the economy and the environment, and  the  need for Federal
regulations to prevent leaking  tanks.

    I am writing to personally  ask for your  participation  in this
vital project, the results of which could have a major impact as
to how we deal with this potential environmental threat.

    Let me assure you that EPA  is not  conducting this  survey to
locate owners of leaking tanks  to take legal action  against
them.  To do so would defeat the purpose of  the survey.  In  the
case of leaking tanks, however, EPA will request that  the  owner
report any leak to the proper local authority  and  take corrective
action such as tank repair, replacement  or removal  from use.

    In order to conduct this study, EPA  has  selected a random
sample of about 1,000 establishments nationwide including  f ar-ns ,
gasoline service stations, transportation-related  businesses,
businesses with private gas pumps, and government  facilities.
The sample of 1,000 estaolishments was selected to  represent as
many types of underground storage tank facilities  as possible in
order to develop national estimates" of leakage on a  scientific
basis.  Your establishment is one of the 1,000 selected to
participate in this important study.
                                 F-5

-------
    Within the next 2 weeks, an interviewer from Westat, Inc., a
private contractor conducting the Survey for EPA, will be
contacting you to schedule an appointment for an interview with
you at your place of business.  A copy of the interview form is
enclosed.  We would appreciate it if you x^ould take the time to
fill out the questionnaire before the interviewer arrives, but do
not mail the questionnaire back to EPA.   The interviewer will
review your answers with you during the visit.

    In addition to the interview, the interviewer will be making
a sketch map of your facility layout, and will want to know where
each of your tanks is located.  It would be helpful if you have a
map of your tank and dispenser layout ready to show the
interviewer.

    As part of the survey,' we will be asking you to provide
product inventory records for a 30-day operating period, so it is
necessary that we know the accuracy of your pump readings.  If
the calibration of your pump  (or dispenser) meters has not been
checked and certified within  the past three months, the
interviewer will need to check the meter calibration with a
certified 5-gallon metering can.

    Your inventory data for each tank system will be analyzed by
computer to identify and explain any shortages or overages.
Results of the analyses will be provided to you at no cost and
will be confidential if you so request.  Later, we will want to
conduct professional tightness tests on some fraction of the
tanks inventoried in the survey.  All tests will be provided free
to the participant, and, if requested, results will be treated as
confidential by the Agency.

    The enclosed booklet of General Instructions will provide you
with definitions of key terms, answers to questions you might
have about the survey, and directions on completing the
questionnaire and providing inventory information.  If you have
any further questions about this questionnaire, or need any other
assistance, please call Westat at the toll-free survey assistance
number 800/638-8985, and ask  for the EPA Specialist.
                                 F-6

-------
    You may claim confidentiality for all or any part of your
response under 40 CFR Part 2.  You should do this when you
provide the information to the interviewer.  A confidentiality
request form is included in the instructions booklet.

    Although EPA is conducting the survey.under Federal
authority, we are seeking your full and active participation on a
cooperative basis.  I hope we can count on your help.

Enclosures

                          Sincerely,
                          William D. Ruckelsnaus
                                   F-7

-------
U.S. ENVIRONMENTAL  PROTECTION AGENCY
  UNDERGROUND  STORAGE  TANK SURVEY
       GENERAL  INSTRUCTIONS
                 Prepared by:

         WESTAT
         An Employee-Owned Research Corporation
               ci-iBivci • Rock./.He rv/ID SOBSO • .3CM S51 -1 5OO
                         F-9

-------
   GENERAL INSTRUCTIONS  FOR  COMPLETING THE
    ESTABLISHMENT OPERATOR'S  QUESTIONNAIRE
   PLEASE READ THE FOLLOWING INSTRUCTIONS  BEFORE  YOU  BEGIN TO
   FILL OUT THE ENCLOSED QUESTIONNAIRE.   IF  YOU SHOULD NEED
   FURTHER ASSISTANCE,  CALL WESTAT AT THE  TOLL FREE SURVEY
   ASSISTANCE NUMBER, (800) 638-8985,  AND  ASK FOR THE EPA
   SURVEY SPECIALIST.
                     PURPOSE OF THE SURVEY
     The Environmental Protection Agency (EPA)  is  conducting
this study to learn more about the problem of  leakage  in  under-
ground storage tanks.   The purposes of  this study  are  to  find
out how widespread the leakage problem  is,  and  to  collect infor-
mation on factors that cause tanks to leak. The study will help
the Agency assess the  impact of leaking tanks  on the economy and
the environment, and the need for Federal  regulations  to  prevent
leaking tanks.
                HOW ESTABLISHMENTS WERE SELECTED
     Establishments were selected to participate  in  this  survey
from a preliminary listing of facilities  that  are likely  to have
underground storage tanks.  This  list was compiled by  EPA from a
variety of sources, including government  agencies, federal program
rosters, and private and telephone directories.   Your  facility
was not purposely chosen from this listing,  but sampled on a
probability basis using scientific random selection  procedures.
The purpose of the probability selection  procedures  is to obtain
a broad representation of kinds of establishments with underground
motor fuel storage tanks.

     If your company operates more than one  establishment that
has underground motor fuel storage tanks, the  establishment you
are to respond for can be identified by the  facility's name and
address on the questionnaire label.   If the  questionnaire label
does not provide you with enough  information to know which estab-
lishment to respond for, please call the  EPA Survey  Specialist
at the toll free hot line number,  (800) 638-8985.
                                    F-ll

-------
               HOW THIS  SURVEY WILL BE CONDUCTED
     Within the next two weeks, an interviewer from Westat, Inc.
will be contacting you to arrange an appointment for an in-person
interview with you at the establishment location.  (Westat, Inc.
is a survey research company that is assisting the EPA in con-
ducting the Underground Storage Tank Survey.)  Enclosed with
this instruction booklet is a copy of the questionnaire, so that
you will know what questions the interviewer will ask.  In order
to answer some of the questions, you may need to consult your
records, so you should prepare your answers to the interview
before the interviewer calls.  Since the interviewer will record
your answers in a separate copy of the interview, the enclosed
copy is yours to keep.
                           AUTHORITY
     This survey is being conducted under authority of Sections
9005 and 9009 of the Resource Conservation and Recovery Act
(RCRA), as amended by the Hazardous and Solid Waste Amendments
of 1984.  Subsections (a) and (b) of Section 9005 detail EPA's
authority for conducting the survey and the conditions under
which EPA will treat information provided by owners and opera-
tors as confidential business information (see CONFIDENTIALITY).
Section 9009 details EPA's responsibilities in conducting studies
of underground storage tanks.
                          REIMBURSEMENT
     Section 9009(f) specifies that owners or operators of under-
ground storage tanks shall be provided "fair and equitable reim-
bursement" for "costs, including the loss of business opportunity,
due to closure or interruption of operation of an underground
storage tank solely for the purpose of conducting studies author-
ized by this Section."  Under Section 9009(f)(2), claims for
reimbursement must be "filed with the Administrator [of EPA] not
later than 90 days after the closure or interruption which gives
rise to the claim."
                           F-12

-------
                         CONFIDENTIALITY
     Section 9005(b) of RCRA, as ammended requires EPA to make
survey information available to the public upon request, unless
you have requested that the information be treated as confiden-
tial business information under 40 CFR, Part 2 and Section 1905
of Title 18 of the United States Code.  As explained in the
Administrator's open letter, you can request that all of the
information you provide be treated as confidential business
information, or that certain items be treated as such.  Informa-
tion that has been determined by EPA to be confidential business
information cannot be made available to the public by EPA, but
can be made available to authorized officers, employees and
representatives of EPA, and to the Congress, if requested.

     Although EPA is conducting this survey under Federal
authority, we are seeking your participation on a cooperative
basis.  Be assured that the contractor and staff conducting the
survey are pledged not to disclose the name or address of any
participant.  The contractor provides survey data to EPA identi-
fied only by a participant code number.  Only if an establishment
refuses to participate will the name and address be given to
EPA.  Should this occur, the Agency may be required to take
legal steps to obtain data necessary to the survey.  However, we
would use legal action as a last resort and would strive to
avoid its use.

     If you want to request that some or all of the information
you provide will be treated as confidential business information,
please read and complete the "Request for Confidential Treatment
of Business Information" form enclosed with this package.  You
should give the completed, signed request form to the interviewer
at the time of the interview.
                                 F-13

-------
        REQUEST FOR  CONFIDENTIAL TREATMENT

                OF BUSINESS INFORMATION
     I hereby request that information I have provided to the
Environmental Protection Agency in response to (certain/all)
the questions in the "Underground Storage Tank Establishment
Operator's Questionnaire" or the "Inventory Record Form"  be
treated as confidential business information under 40 CFR
Part 2, and Section 1905 of Title 18 of the United 'States
Code.
     LIST THE QUESTION NUMBERS OF THE RESPONSES FOR WHICH YOU
ARE REQUESTING CONFIDENTIAL TREATMENT:     	
                     PLEASE PRINT  OR  TYPE:
ESTABLISHMENT NAME:
MAILING ADDRESS:
                                     Street
                      City               State            Zip


TELEPHONE:
                                                      Extension
ESTABLISHMENT OWNER/
  OPERATOR:
                        (Print or type)          (Signature)
DATE :                	/	/	
                     Month    Day    Year


                            F-15

-------
                   DEFINITION  OF TERMS
Cathodic Protection - Used to reduce or eliminate corrosion
     of a metallic structure which is in contact with corrosive
     soil by applying an electric current to the structure
     which is greater in strength and opposite in direction to
     the current that is causing corrosion.

     Passive (galvanic) Cathodic Protection  - The required
          current is generated by the corrosion of sacrificial
          anodes, such as Magnesium or Zinc, which are attached
          to the surface of the protected material (tank or
          pipe) in the soil.

     Impressed Current Cathodic Protection - The required
          current is provided by an external source and is
          passed through the system using non-sacrificial
          anodes, such as Carbon or Platinum, which are buried
          in the ground.


Continuous Electronic Monitoring System - This system could
     include the following:

     •    thermal conductivity sensors;
     •    electrical resistivity sensors;
     •    gas detector; and
     •    interstitial monitoring in double-walled tanks.
Establishment - The term establishment is used to mean a
     commercial or non-commercial location that is used for
     any purpose other than just a residence.  That is, any
     location that is used for a nonresidential purpose (even
     if it is also used as a residence) is considered to be an
     establishment.  Examples of establishments include gaso-
     line service stations, farms, schools, factories, fire
     stations, highway maintenance facilities, parks, stores,
     offices, delivery services, military installations,
     airports, etc.  (If you believe that your facility does
     not fit the definition of an establishment, please call
     the toll free survey assistance number,  (800) 638-8985,
     and explain your situation to the EPA Survey Specialist.)
                                F-16

-------
External Corrosion Protection System - This system could include
      the following special equipment or materials:

           cathodic protection;
           electric isolotion;
           polyethylene wrappings;
           coatings; and
           paints.


Inventory Reconciliation - The balancing of "book" inventories
      against observed inventories (meter/dipstick readings).


Manway - A means of entrance into an underground storage tank
      allowing internal inspection.


Motor Fuel - Any substance that is used to power a motorized
      vehicle (such as an automobile, boat, airplane, truck,
      etc.).  For example, motor fuels such as:

      •    leaded gasoline;
      •    unleaded gasoline;
      •    diesel fuel;
      •    aviation gas;
      •    jet fuel; and
      •    gasohol.


Pressure Pump Delivery System (also called submerged pump delivery
      system) - This system works on the principle of positive
      pressure to push the liquid from a low point to a high
      point using a submerged pump (coupled with an electric
      motor) mounted inside the tank.


Remote Gauge - A measuring device that indicates the quantity of
      fuel stored in a tank on an external scale or dial.


Secondary Containment - A secondary enclosure or barrier intended
      to contain any spills or leakage from the primary storage
      tank or from pumps, piping and other equipment.  These may
      include:

      •    concrete vaults or basins;
      •    plastic or clay lined basins;
      •    soil sealants (soil cement or bentonites); or
      .    double-walled tanks or pipes.
                              F-17

-------
Siphon Pump Delivery System (also called suction pump delivery
      system) - This system works by drawing liquid from a low
      point because of a vacuum at a high point, using a suction
      pump.  This pump is located at grade (i.e., ground level),
      either directly above the storage tank or, as in the case
      of some dispensing operations, at some distance from the
      storage tank (at the pump islands).


Underground Storage Tank - A large vessel or container-placed
      beneath the surface of the earth used for storing and
      handling of liquids (such as petroleum products) or waste
      materials (such as used or waste oil).


Used or Waste Oil - Oils (whether used or unused) that are no
      longer fit for their intended use because of contamination
      or degradation.  These oils include, but are not limited to

      •    automotive engine oils;
      •    gear lubricants;
      •    diesel engine oils;
      •    railway diesel oils;
      •    oil storage and treatment residuals (such as bottoms);
      •    hydraulic oils;
      •    metal working oils;
      •    transformer oils; and
      •    oils contaminated with water.
Water Finding Paste - A paste applied to the bottom of the
      dipstick which changes color when it comes in contact with
      water.
Water Table - The upper limit of the portion of the ground
      (soil) wholly saturated with water.
                                 F-18

-------
       ORGANIZATION  OF THE QUESTIONNAIRE
The Establishment Owners/Operators Questionnaire is  designed  to
obtain data on your establishment's underground fuel and waste
oil storage operation,  including such items as tank  design,
operating and installation characteristics, tank corrosion pro-
tection and tank leakage monitoring.  The questionnaire  is
divided into seven sections,  as follows:


A.   Screening Information

     This section of eleven questions asks for information about
     the establishment  itself,  including  questions about the
     type of establishment, the owner and operator of the estab-
     lishment, and the  number of tanks at the establishment.
     Question A.11 provides instructions  for completing  Tank
     Description Sheets for the establishment.

          Tank Description Sheets - A Tank Description Sheet
          must be completed for each underground tank.   Ques-
          tions asked will include information on specific tank
          characteristics, such as reported age, size and typi-
          cal fill volume, manufacturer,  installer,  materials of
          construction, inspections or leak tests, and other
          design characteristics.


B.   Operating Practices

     This section asks  questions about practices such as taking
     tank inventories using a dipstick, checking and recording
     dispenser meter readings,  inventory  procedures  after a
     delivery and inventory reconciliation or "balancing"
     between stick readings,  dispenser meter readings, and
     delivery records.


C.   Operating History

     In this section you will be asked about any tanks that have
     been replaced, removed without being replaced,  or abandoned
     in place, and in what year and why this occurred.


D.   Permits and Licenses

     This is a short section  about any special permits or licenses
     needed for tank installation or storage of flammable materials


                             F-19

-------
E".   Installation

     Section E includes overall questions about how the tank
     was installed.
     Protection

     This section asks questions about any protection systems
     in use against external corrosion, and any monitoring
     systems used to detect tank leakage.
     Information Needs

     Section G is about the kinds of information and services
     relating to tank monitoring that are currently available
     to you.
                                 F-20

-------
                USE  OF THE QUESTIONNAIRE
          The questionnaire has been  designed to minimize  the
effort required for it's completion.   "Skip patterns" have been
incorporated  to enable respondents to by-pass sections of the
questionnaire which are not relevant  to them.  The following
section describes how you are  to  complete the questionnaire in
preparation for the call from  a Westat interviewer.
                       EXAMPLES OF QUESTIONS


          Most of  the questionnaire  items are straightforward
and require  only the circling  of  the correct code(s)  or the
completion of short answers  on the lines which  are provided.
The following examples illustrate the use  of other question
formats  found throughout the questionnaire.


          Example  A

          Some questions require  that you indicate  a  distance or
frequency and also circle  the  correct unit of measurement or
time as  indicated in the sample questions  below.   Different
units have been  specified  for  your convenience.   Please do not
neglect  to circle a unit code  (as shown) or to  write  in an
appropriate  unit of measurement or time.   This  question, as with
all questions,  includes  its  own instructions printed  in capital
letters  and  enclosed in  brackets.
E4.   What is the shortest distance between any of your tanks and any neighboring underground
     tank or other solid underground structure (such as a basement wall, sewer, or utility
     vault;?  [ENTER DISTANCE AND CIRCLE UNIT CODE]


                                   SHORTEST DISTANCE FROM        /
                                    UNDERGROUND STRUCTURE: 	t/7            /23-2£
                                   INCHES	01
                                                                        , 29-3C
                                                                 v—'
                                   OTHER [SPECIFY];
                               F-21

-------
F2.   How often do you inspect your external corrosion protection system?  [ENTER
     FREQUENCY AND CIRCLE UNIT CODE]
      IF YOU NEVER INSPECT THE EXTERNAL CORROSION PROTECTION SYSTEM, CHECK HERE
      AND SKIP TO F3.
                                        FREQUENCY OF INSPECTION:
                                        [CIRCLE ONE]:
                                         PER DAY	01
                                         PER WEEK	02
                                         PER MONTH	(JT,
                                         PER YEAR	04
                                         OTHER [SPECIFY]:
                                                                             05
                                                                                    719
                                                                                    720-22
                                                                                    723-24
            Example B
                                                                         "no"
            Other questions  require that you code a  "yes or
answer  for  each  category listed,  as  indicated in  the sample
question below.   The "Other [SPECIFY]"  line  enables  you  to enter
an answer  not  covered  by the preprinted response  categories.
                   cn yf t*ie following fael tvpes «ere atorea in 5113
                 :an* luring the past 12 moncns' [CIRCLE JNE CODE fOR
                    Jnieaaea  lasnune
                    Die-re, 'ue: .  . .
                    Otner UPEC'.r*;:
                                                  / 58-55
                                         F-22

-------
           Example C

           When a series of  similar questions  apply  consistently
to a  given category,  they have been  formatted into  tables  or
grids to facilitate  the administration of the questions.   Notice
also  that Question C6b requests that all applicable response
categories be  circled, not  just the  most prominent  one, as indi-
cated in the sample  question  below.
  C6.
Please answer the following questions about each tank that has been removed without
being replaced.  [SPACE HAS BEEN PROVIDED FOR UP TO FOUR TANKS. IF MORE THAN FOUR
TANKS HAVE BEEN REMOVED WITHOUT BEING REPLACED, WRITE THE ANSWERS FOR THE ADDITIONAL
TANKS ON A PLAIN SHEET OF PAPER]

C6a. In what year was the
(first/second/third)
tank removed?


C6b. Why was tne tanK
removed?
[CIRCLE ALL THAT
APPLY FOR EACH TANK]
a. Because it

was leaking?. . . .
b. Because other tanks
were being removed
at that time? . . .
c. Because it was no
longer needed/ in
use7 	
d . Or for some other
reason [SPECIFY]: .


First Tank

—1 1 1
(year)
720-23






01

^-.
M32J


03

04
(specify)
724-33
Second Tank

~74
(year)
734-37




^^^

(oy


02


03

04
(specify)
738-47
Third Tank

81
(year)
748-51




^_^^

\^y


02

/" — N
( 03 )
vy
04
(specify)
752-61
Fourth Tank


(year)
762-65






01


02


03

04
(specify)
766-74
                                F-23

-------
                             SKIP  INSTRUCTIONS


            Skip instructions  indicate the next question  to be
 answered.   They  s.ave  time  by allowing  you  to ignore  irrelevant
 questions.   The  following  is an  example of a skip instruction
 attached  to an answer category.


 81.   Do you (or another establishment employee) inventory the contents of your tank(s) by
      neasuring the depth of the contents with a dipstick? [CIRCLE ONLY ONE CODE]

                                       YES [GO ON TO B2]	1       ...
                                                                                 / 10
                                        NO  [SKIP TO 85]
            Skip instructions  are sometimes not attached  to an
 answer but  are enclosed in  a box,  as  shown  below.
814.  Ho* often is the accuracy of your dispenser meters checked?  [CIRCLE ONLY ONE CCOE]
IF THE ACCURACY OF YOUR DISPENSER METERS IS NEVER CHECKED,  CHECK HERE
AND SKIP TO 816.
                                                                                732
                                      DAILY	01
                                      WEEKLY	02
                                      EVERY TWO WEEKS	03
                                      MONTHLY	C4     /33-34
                                      ANNUALLY	05
                                      OTHER [SPECIFY]: 	  06
                                      F-24

-------
         THE  MOTOR  FUEL INVENTORY SHEETS
     Enclosed in the survey package  are  four  kinds  of  sheets  for
keeping daily motor fuel inventory records.   The  type  of  tank
and dispenser systems you operate will determine  which inventory
sheet(s) you will need to use.   You  may  need  only one  kind of
sheet or as many as three kinds.

     In Figure 1 on the following page,  you will  find  schematic
diagrams of the seven most common tank and dispenser hookup
systems currently in use.  These seven hookup systems  are listed
in Table 1, below.   Use the diagrams in  Figure 1  to determine
which tank and dispenser hookup system(s) you have.  Then use
Table 1 to determine which kind(s) of sheet(s) you  should use
for inventory recording.
  Table 1. Inventory Recording
Possible tank/Dispenser Meter/
Dispenser Hookups
Single tank, unmetered
Single tank with single
dispensing meter
Single tank with multiple
dispensing meters
Custom Blending:
2 tanks, 2 dispensing meters,
1 dispenser
Custom Blending:
2 tanks, multiple dispensing
meters and dispensers
Manifolded Tanks:
Multiple interconnected tanks,
multiple dispensing meters
Manifolded Tanks, Custom
Blending
Appropriate Inventory Review Forms
Inventory
Sheet
for Tanks
without
Metered
Dispensing
Pumps
X






Inventory
Sheet for
Tanks with
Metered
Dispensing
Pumps

X
X
X
X
X
X
Dispensing
Meter
Recording
Sheet

X
X
X
X
X
X
Manifolded
Tank System
Recording
Sheet





X
X
                             F-25

-------
                                 Figure 1
                   Schematic Diagrams of Possible Tank/
                     Dispensing Meter/Dispenser Hookup
 Single tank.
                Single tank.
Single tank.
Custom Blending:
 unmetered'  single dispensing  multiple dispensing  2 tanks, 2 dispensing
                                      meters         meters,  1 dispenser

     Custom Blending:
2 tanks, multiple dispensing
   meters and dispensers
          Manifolded Tanks:
    multiple interconnected tanks,
     multiple dispensing meters
                                  M1]   IM21  IM3J   IM4
                   Manifolded Tanks, Custom Blending
                                  F-26

-------
     Regardless of which inventory sheets you use, you will need
to provide 30 complete inventory readings for each of your tanks.
It is preferable that each of these readings represents one
operating day.  Many tanks and tank systems are inactive (not
used) for certain days during the week.  If your tank(s) are
inactive on a particular day, you can use the inactive day as an
inventory day only if you take and record dipstick readings for
the tank(s) for that day.  (You cannot carry down the closing
stick readings from the previous day. )  You must provide actual
stick readings (or remote gauge readings, if available) for each
of the 30 inventory days.  If your dispensers are metered, you
must also provide meter readings for each of the 30 inventory
days.  If you do not have complete inventory information for a
day, do not use that day as an inventory day.

     Instructions for using each of the four kinds of Motor Fuel
Inventory Sheets, along with example copies of the Sheets, are
provided on the following pages of this booklet.  After you have
used Figure 1 and Table 1 to determine which inventory sheets
you will be using, please read the instructions on how to complete
the sheets.
          If you have any questions about:

          •   Which sheets you should use for your tanks;
          •   How to complete the sheets; or
          •   Any recording problems you may have;

          please call Westat at the toll-free survey
          assistance number, (800) 638-8985, and ask for
          the EPA Survey Specialist.
                              F-27

-------
                 INSTRUCTIONS  FOR COMPLETING THE
     INVENTORY  SHEET  FOR TANKS WITH METERED DISPENSING  PUMPS
     The Inventory Sheet for Tanks with Metered Dispensing Pumps
is used for any individual tank or system of connected tanks
(i.e., manifolded tanks) that has one or more metered dispensing
pumps.  The sheet is used to record daily physical inventory
measurements (stick readings and deliveries) and volume of fuel
pumped from the tank, as calculated from dispensing meter read-
ings.  You will need one Inventory Sheet for Tanks With Metered
Dispensing Pumps for each tank (or system of tanks) that has
metered dispensers.

     You should fill out one line of the Inventory Sheet for
each day that inventory readings are taken.  (Days for which
inventory readings are not taken should not be entered on the
sheet.)

     •    In Column 1, enter the date of the reading (day and
          month).

     •    In Column 2, enter the opening dipstick reading, in
          gallons.  (On days 2 through 30, opening dipstick
          reading will be the same as the closing stick reading
          of the line above.)

     •    In Column 3, enter the day's deliveries to the tank,
          in gallons.

     •    In Column 4, enter the sum of Columns 2 and 3.  (This
          is your "opening physical inventory."

     •    In Column 5, enter your closing dipstick reading to
          the nearest quarter inch.

     •    In Colunn 6, enter your closing dipstick reading,
          converted to gallons (using your conversion chart for
          this tank).

     •    In Column 7, subtract the amount in Column 6  (your
          closing stick inventory) from the amount in Column 4
          (your opening physical inventory) and write the
          remainder in Column 7.  This column represents the
          quantity gone from the tank, according to your physi-
          cal inventory records.
                               F-28

-------
     .    In Column 8, enter the "meter sales" (the number of
          gallons pumped from the tank according to your meter
          readings).  You must record the actual meter readings
          and calculate the meter sales on a Dispenser Meter
          Recording Sheet.  Column 8 of the Inventory Sheet
          should equal Line I of the Dispenser Meter Recording
          Sheet for the same date.

     The Inventory Sheet for Tanks With Metered Dispensing Pumps
is printed as a four-page booklet along with a Dispenser Meter
Recording Sheet.  (The dispenser Meter Recording Sheet is the
last three pages of the booklet.)  Six copies of the Inventory
and Dispenser Sheet booklet are included in the survey package.
If there are more than six tanks with metered dispensers at your
establishment, please photocopy as many additional sheets as are
required.
                              F-29

-------
   INVENTORY  SHEET  FOR  TANKS WITH METERED DISPENSING PUMPS
                                                                        Tank Number:
                                                                        Dispenser  Meter  Numbers:
                      (Name  of  Facility)
                       (Street  Address)
                                                      Type  of Fuel:

                                                      Size  of Tank:
      (City/Town)
          (State)
                   (Zip)
                                                                        Year  Installed:
                                             Dipstick*  Inventory
       Column 1
 Column 2
                                Column  3
                                             Column  4
                             Column 5   Column  6
                                      Column 7
                                       Column 8
 Day
         Date
 Opening
Dipstick*
Inventory
(gallons)
 Deliveries

(in gallons)
[Column 2]
   plus
[Column 3]
 Closing
Dipstick*
Inventory
(inches)
 Closing
Dipstick*
Inventory
(gallons)
Gone from
   Tank:
[Column 4]
  minus
[Column 6]
Meter sales**
  (gallons)
 (from meter
   sheet)
  3  I
 10
 11   I
 12
 13
 15  I
 16
 17  I
 18
 19
 20
 21
 22
 23
 25
 26
 27
 28  j
 30
          I
              I
            I
           I
 *If tank has remote gauge,  check here  	|  and use remote gauge readings instead of stick readings.
"Transferred from Line I  of Dispenser Meter Recording Sheet.
                                                F-30

-------
                   INSTRUCTIONS  FOR  COMPLETING
                 DISPENSER  METER RECORDING  SHEET
     The Dispenser Meter Recording Sheet is used to record daily
meter readings and to calculate volume of fuel pumped for all
dispenser meters connected to an individual tank or system of
tanks.  One 30-day set of Meter Sheets is kept for each individ-
ual tank or connected tank system.  On each day of inventory
readings, record each meter's closing reading (in gallons) on
Line G ("Today's Closing Meter").  Record "Yesterday's Closing
Meter" on Line H.  (For Day 2 through Day 30, "Yesterday's
Closing Meter" will be the same as Line G ["Today's Closing
Meter"] from the day before.)

     The gallons of fuel dispensed daily through a given meter
is calculated by subtracting "Yesterday's Closing Reading" (Line
H) for that meter from its "Today's Closing Reading" (Line G).
Enter the difference between the two readings in Line I for each
meter.  This is the number of gallons dispensed (pumped) through
that meter during that day.  After you have entered the gallons
dispensed by each meter in Line I, add up Line I for all meters
and enter that figure in the column marked "Line I Totals."  The
"Line I Total" figure is the daily "gallons dispensed" for all
meters.  The "Line I Total" must also be recorded for the same
day in Column 8 of the Inventory Sheet for this tank.

     Dispenser Meter Recording Sheets are printed in a four-page
booklet, along with an Inventory Sheet. Six copies of this book-
let are included in the survey package.  Please photocopy extra
copies if needed.
                             F-31

-------
*
» n
8 ^
_j
u
4)
| £
It)
0}
0. =sb
S

1
t*
1 £


. (^
• £
1s
1
01
tt ^
s
1


— ' CM
a; *»
r
i
I — ' *~
c «&
! I


I
i
'£.
C
,=
s
1
EJ
0,
or

I

a,
*j
a





























u
4)
1
Ol
X
C
>.

2
CJ





-























u
u
£
c
03
c
,-H
LJ
n
!

CO
01
X

-



























s~*
X
1
co
c
c.
CO
i

a
CJ
"*































u
£
Ol
c
CD
O
CO
X


u





























L.
01
I
c
ta
o
t-
n
a


=

CM




























X
43
1!
n
c
a
to
0)
c
o

a
N-































01
£
?
n
o
n
X


CJ





























u
u
£
g
CO
o
u
CO
el
•s
u

CO
X

-



























^_^
X
CJ
•s
m
c
c
a
n
n
c
o

CJ
•^































t-
0)
?
(0
o
0)
X

2
"



.

























u
01
£
|
9)
O
U
n
CO
•o

CO
01
X

-



























_^
-I.
I
c
c
c
n
01
c
0

CO
*H































u
01
£
•-H
CO
o
I—I
n
X

•c
o
CJ






























01
V
a
c
83
O
U
n
n
u
01

to
01
=

*




























X
t
s
Ol
a
(C
m
O

3
^































u
CT(
C
0)
o
CO
X

T3
O
t—
"





























fc-
U
0)
.5
CO
o
LJ
CO
a
•o
01

n
01
X

-



























_
X
1
CJ
•c
0)
01
c
1
03
C
o

•— 1
0)
43
I-H































«
I
H
1(0
o
03
X


"





























u
01
01
1
in
o
o
n
a
•c
01

CO
01
X

-




























X
u
•a
V
£
tDl
n
c
o


OH































u
0)
?
I'**
0}
O
0]
x


43





























k4
01
01
.s1
03
O
U
O
CD
4)

0)
X

CD



























^_^
il
1
43
"S
S

X

„



























^_^
X
1
43
"8
n
c
a
a
•f-4
co
§

3
•^































V
c
c
c
n
X


u






























01
£
c
c
u
CO
a
•c
01

CO
X

0



























^.^
X
1
•o
0)
CO
c
CO
o

3
**


                                                                           ^  01
                                                                           a  o>
                                                                           w iS
                                                                           0  x
                                                                           u)  «-
                                                                           i-  O
                                                                           s -s
                                                                           £-0
                                                                           ^ 00
                                                                           .2  c
                                                                           JJ  O
                                                                           •o "
                                                                               c
                                                                            L.  0)
                                                                            O  I-
F-32

-------
£
*
* en
CO
0) •*-*
C 0
-J
t,
CD
j_i CC
C *fc
Z
u
CU
U *fe
z
u
CU
z:
0)
+j J"\
CU •»
U
a
1 *



u
w -^
| 5. «!=
*•

1_
.i_> f^f
U =ft£

£
i -
c. %
z
1
1






c
CJ
=

g.
C

"E
c
C3


(H
J
Z

cc




cc
—







































£

g


o
X


a
0

LJ









































4,

CU


c

o
LJ

-w
.£
L.

*-
a;

=





*-
*~




































i

LD
•O

en
c
01
c.
CO
in
c
o

cc

^™





















...















—







CU
z

a
c


o
n


CU
o

J










































*J


~J
c
M
0
LJ

_W
CO
1-1
CU

CU

*





(NJ
«—




































1

CJ
•a
CU
c


en
CO
c
c

CD

^""













































cu
r

C"
C

(0
o
CO


a
o

CJ









































CU

CO


c
CO
0
LJ

cn
CO
•a
u,
eu

CU

=





^
*~






































LJ
-n
CU
c

c,
cn
CO
c
c

CD

*"H













































CU
OJ

tJ
C

cn
c
n


cc
0

CJ








































!_
0




c
en
O
CJ

1C
re
u.
CU

(U

X





,
^
c
u_
C
(C
E
CD
L.
O

*
                                                                                   F-33

-------
     =
i
a
    S

—   £


S   S
ir   —
    IT

z   f
C=   u-
    cr
C£   ^j



I   £
cr   •-

cTi   z


?   S
r;
&
-
• CO
1 s
_J
tri
V
.LJ CO
u
0)
V *fc
r
u
V
^ Nfl
u
V
^J .A
V

1

1 u
•s *
1

cc
JSJ
1-
.w •-
z *
1

! ™
c
5
! .c
CO
•5

c
u
ft:
CK
£
*
1

[•••
'e


a
"^


























U
u

cr
c
X
c

CJ

CO
>.


































u
V
f
?
CO
o
en

X
0
u
Cl
«
01

z





*-
fsl

















	 I






	 ,
z
1
^
c
u
a
Cl

0

o:
0
a

—

































u
1
I
CO
c

u

CO
X
•c
o

CJ




1


























t.
u
£
a
c
o
u
if)

>

*D
L.
o
s

X





INI

























	 ,
X
C3
•o
CD
C
a
en

2

en
c
c
«

-^

































j.
1
a
c

u

.w
•o
o

CJ































u
u
£
o
c
CO
o
LJ

X

•o
u
01
CO

s





F*
(N
























,_^
X
CJ

CO
c
41
C.
X

Q

O
ce

—

































01

I
to
o

u

GO
X
0

CJ































u
0)
V
X
51
0)
o
tJ
10

>
a
u
«

X





-a-
tN
























^^^
X
1
CJ3

cn
c
0)
a
CO

a

tn
§

CN
























^-^
X
1
CJ
7,
CO
I
09

o

1
a

—

































01
01
01
c
CO
0

o

a
X
•o
o

u































b*
01
1
o
c
0}
o
CJ
0)

X
(0
-a
L4
n

X





xO
CN
























^^
X
CJ
•a
0)
a.

^

CO
§
^-1
eg

~

































u
£
C1
c
CO
o

u

CO
X
•§

CJ































u,
0)
1
o
c
CJ
0)

X

T3
»

X





f^.
CN
























^^,
Z^l
f
^
S
s
CO

a

CO
c
o

CJ
—

































01

1
o

u

."
•§

CJ































u
01
£
o
cn
o
u
to

X
a
•a
01
cn
01

z





CO

























^^
z
1
u
•o
01
CO
1
CO

a

c
o
"a

—

































u
1
o
c
n
3

U

CO
X
0

CJ































bl
01
I
CT
c
to
c
u
eo

>.
re
u
0)

=





5\
fN
























^_^.
X
1
•o
01
spens

^

CO
c
o
a

—

































01

cn
c

u

CO
•§

CJ

































1
1
o
u
CO

V
re
u
a;
n
OJ

X





w
"














1
I
I


i
1




. 	 ^
X
'i
CJ
7
w
1
CO

o

cn
c
o
03

—







                                                                                                                                                      a,
                                                                                                                                                      c.
                                                                                                                                                     c.
                                                                                                                                                   0)  l-
                                                                                                                                                   (-  C
                                                                                                                                                   0) —


                                                                                                                                                  1  !
                                                                                                                                                  *J  C


                                                                                                                                                  *^ cc
                                                                                                                                                   =  g
                                                                                                                                                   eo  ce
                                                                                                                                                      c
                                                                                                                                                   I-  CC
                                                                          F-34

-------
                 INSTRUCTIONS  FOR COMPLETING  THE
             MANIFOLDED TANK SYSTEM RECORDING SHEET
     The Manifolded Tank System Recording Sheet is an eight-page
booklet that is used whenever two or more tanks are connected by
piping to make a multiple or manifolded tank system.  One Mani-
folded Tank System Recording Sheet booklet is to be used for
each manifolded tank system that is to be inventoried.

     The purpose of the Manifolded Tank System Recording Sheet
is to provide a convenient way to keep individual daily stick
and delivery records for each tank in the system.  At the end of
each day, you should add up and record each line of inventory
measurements (Lines A through F) for all tanks in the manifolded
system.  These daily totals are entered in the "Tank System
Totals" column of the Manifolded Tank System Recording Sheet,
and then transferred to the appropriate columns of the Inventory
Sheet for the tank systems.  The "Transfer to Inventory Sheet"
column on the righthand side of the sheet indicates that Inven-
tory Sheet column number to which the total should be transferred,
You must also complete a Dispenser Meter Recording  Sheet for
the tank system.
                              F-35

-------
•o
 CO
-a
 CO


JZ
 o

 X
 u
 CO
 C
 CT
               CO
               a.
C-l t-i
CO O -J
c 4j co .c
o > en
c~ c
a o
.* CO -H
C JJ CO
CO CO 4-J
>— >•, O
l/l I—
c cc
CO «k
c r-
CD 4k
C ^
CO 4k
a «k
c 
-------
to

CD
5
O
£
t- u
CD O -*-i

u c
S (0
J* CO — c
C J-> O
CO CO 4->
t— >» o
to »•:
"c eo
C P>
a "fc
^
a *
H-
a *h
H-
a «fc
H-

a «h
C (N
0 •»
c «-
a *h



en
C
o

1
a
CJ

c


CO
a
c
•H
•a
b
o
CJ
CO
er


a
jj
i




cu
X
a
X
OJ
Q
CN
O
CJ



















LJ
CU
C
CO


CO
T3

CU

a


^^

^^
CO

CO
c


y
r)


uO
a
c
c

3
<




KV
O






































CO

03
a



cu
•t-i
CO
.^
F-H
Q
ca




*
O


























^-v
CD

•f




•^
C
a

c


i~l
u



U-
o
a
1 1
o
CJ



.
^
o

































,^— y
CO
cu
£.
U
C
H


V
O
• H

to
c
c
CO
o
CJ
a




*
o



































f—^
CO

CO
01


V
U


to
Cl
c
CO
o
CJ
uJ




r^
a






















LJ

i

CJ



CO

o
o


^f
c
a



u


cu
c-
CJ
^H
cu
u.
L.




M
a




















LJ
cu
c
CO

>
a
•o
u
cu

CO
cu



^^y
en

CO
o


V
u


in
c
c
c
cu
a
o
*




~
o






































en

o
o
N.^

CO
CO
u
CO
.^
*~4
c*
ca




<±
o


























^— ^
ca






Y
c
at

c



CO
3
L.

0
CO
1 >
o
CJ



V,
^
o

































s~^
CO
CO

o
c


.V
a


tn
c
c
CO
o
c_
a




vO
O



































^— y
a

CO
a


Ly
o


to
c
c
CO
o
CJ
LJ




i-
o






















LJ

1

O



en

03
a1


V
C
a

E
0

u.

cu
c
8
-H
CU
3
L.
L-




CN
O




















UJ
CO
c
en

x
CO
-a

CO

05
CO



f-*.
CO

CO
C"


V
o


tn
c
c
c
0
c§
<




«
o






































CO

a
o
s_y

CO
CU
^_,
cu
.^
p«(
ICO
Q
22




•3-
O


























^— ,
ca






^
c
a

c
H


cu



0
CO
) .
o
CJ



r-
lA
O

































^~^
0)
cu
jZ
o
c


V
O


tn
01
c
n
0
CJ
C




*
o



































^-^
CO

a
en


^
CJ


tn
c
c
CO
a
CJ
LJ




r>
o






















C



CJ



en

CO
a


^
c
CO

e
o
(^
u-

cu
c
o
CJ
^
CO
'-
u.




CNI
o




















UJ
CO
c
en

>>
£
•a

CO

CO
CO



, — .
CO

o
a-


-^
o


tn
Cl
c
CO

«




KN
O






































CO

CO
o
*_^

CO
CO
CO
^
^^
c
cc




*
o


























^-^
m






^
c
CO
t—
c



cu
3
L-

0
CO
) 1
c
CJ



CO
*
o

































_^~,
CO
CO
^~
o
c


^
c.


e/)
01
a
0
CJ
a




vO
o



































^— >
CO

a
C1


•^
o


tn
a
c
B
0
CJ
Lu




r-
o






















LJ

i

o



CO

cc
o


V

CO
K—

e
o



CO
c
o
CJ
_,
3J
ut
L.




                                                                                                       CO
                                                                                                       >s
                                                                                                       tn
                                                 F-37

-------
a
u
to



 41 .C
CD > 0!
1- C
^ H- 1
S CD
•2 * •£
c *j eg
O CO 4J
t— X O
in H?
"c co
ce 4k
1 =
eg 4k
eg 4k
"c -a
eg 4k
K-
CO 4k
eg «
a H
01
o

CD
O
U
O





O
u




























o
1

^
a
tn
gl

co
o
-H
U
UJ





(°


















UJ
1

u


CO

a
o

K-

1
u
u-
41
IQ
C_3

,^
41
3
U.
U.





cS
















UJ
41
C
•H
CO
^,
CO

41

co
41


'o
'o
O

^
U
Ul
o
c

c
41
<§
«t





3




























"co

co
O

CD
41
bi
41

.^

m





o
u





















ai

+

^
•^
^
s
c

rH
4)
D
U.
C

CD
£
U


r—


cS


























^^
o
41
J=
U
C
•*4

_^
U
•H
.U
O
c

CO
o
u
o





o
u




























CO
s

Jt
u
•H
01
g

CO
o
u
UJ





°


















UJ
1

u


CO
^H
s

1

1

01
o

^-4
s
u.
^





o
u
















UJ
2
•H
CO
^,
a

41
.u
CO

^

CO
CD
o

^
.Jj
01
o

c
41
*





o
u




























CO

co
o

CD
4)
41

>H
^
CD





(3





















CO



^
^-^

c
CD
C


41
3
U.
O

CD
O
H-
u


CM


O


























^^
CO
41
1
• «H

^
•^
01
g

CO
a
u
a





O
u




























co
13
o

^
u
01
gl
H
CO
o
u
UJ





o
u


















UJ
1

u


CD

CD
O

1

§
£



^4
U.
U.



                                                                                                                 CO
                                                                                                                 >,
                                                                                                                in
                                                                                                                 I
                                                                                                                 §

                                                                                                                 0}
                                                                                                                 o
                                                                                                                 u
                                                                                                                 co
                                                                                                                 u
                                                                                                                 0)


                                                                                                                 .u
                                                                                                                 CO


                                                                                                                 CD


                                                                                                                 O






                                                                                                                 §
                                                                                                                 J^

                                                                                                                 n
                                                                                                                 c
                                                                                                                 CD
                                                                                                                 a
                                                                                                                 u
                                                       F-38

-------
UJ
UJ
o
u
a
IS
b b
U. O  01
b C
>— NH
§03
-1
C 4-1 03
00 00 JJ
1- X 0
in H-
!*
c r»
09 «h
H-
!-
|c
09 Id
H-
03 «h
H-
C CM
03 «h
09 «k
H-




09
C
O


cS

c
•H

1
o

cc

OJ




«j
03
a
X
s


O
CJ














UJ
a
c
09


03
•a
b
_Jj
09
1)

^.^ ,
"o?
"o
01

^
o
••4
tn

o
c
c
0)
5
•*






o
CJ




























09

03
O


03
•H
b
i
CD






0
CJ





















^— V

+

•X


09
C
• H
_
4)
3
U.
14-
O
03
"o
CJ



5


5



























^-v
09
•H

JC
U

in

E
§
a
o






^j




























03
C

_^
U

*

o
c
93
O
u
U4






O
U

















UJ

1

CJ

^^
00

03
a


§
b
u_
OJ

u
1
u.
u.






o
CJ














UJ
a
c
03


03


£
09
1)

^^
03
O

_^
u
•H
Ul

O
c
£
a
o
*






cS




























03

03
O


03

b
I
m






c3





















^— s
CO


^


s
c

^
0}
3
U.
It.
o
03
O
CJ



2


o
u



























^— ^
inches


o
• H
in

g
09
O
a






o
u




























'a
03
c


o

in

a
c
09
O
CJ
UJ






o
CJ

















UJ

1

u

^
09

03
O
1

|
b
U.
01
c

3
U.
U.






o
CJ














UJ
OJ
c
09

X
09

b
0)
03


^^,
09
03
C


U
•H

                                                                                                                           C
O

a


3
*••*
O
U
                                                                                                                           03
                                                                                                                           O
                                                                                                                           09

                                                                                                                           03
                                                                                                                          jj
                                                                                                                           a
                                                                                                                           00
                                                                                                                           X
                                                                                                                          en
                                                          F-39

-------
in

CJ
a
ce
a
CJ
in
>-
in
Q
LJ
Q
^
a
u u
0) O 4->
**- 4-1 Oi
c ±J o j;
CO > 17?
£ £
E CO
-X CO —i
C JJ CO
09 09 .tJ
1— >< o
in >—
C CO
09 4k
C r-
09 4k
C SO
H-
C WX
0) 4k
c -a-
09 4k
09 4k
C CM
09 4k
C «-
CO 4k




CO
C
o

^
CO
CJ

c


CO
g
H
u
o
0
CO
cc

u
CO

I



V
eg
c
>,
CO
a

CM

O
CJ



















UJ
CO
c
CO


CO
•o
L.
CO

CO
HI

^x

CO
"3
D

_^
U

in

a
c
c
0
a.
o
*





r*>

0
CJ




































^
CO

0
c


CO

u.
a
•H
2
CO





-»

o
CJ


























^-^
ca

^





c
CO
h—
C


CO
D
u.
u.
o
CO
o
CJ



£

lA

o
CJ

































, — ,
CO
CO
CJ
c


_^
CJ

in

o
c
CO
o
CJ
a





\43

O
CJ



































CO
eg
O

^
o

in

31
c
CO
o
CJ
UJ





p~

o
CJ






















UJ



CJ


f~^
CO

CO
c

^
c
a
h-

§
U

CO
c
c
CJ
__,
«
u.
u.





CM

0
CJ



















UJ
CO
c
CO

>•
CO
•a


. i
CO
CO



CO
a
o

.X
u

1/5

cn
c
c
CO
a.
a
*





r*\

O




































^^^
X

Q
cr
'^^

CO
CO

u
CO
.^
CO
CO





^

8


























J^^.
cn

+





c
at
*"
c


CO

u.

o
"a
o
t—
CJ



CO

lA

o
CJ

































, — ,
CO
CO
c
c


_^
u
• H
J^
in

01
c
CO
o
•-H
CJ
Q





SO

0
O



































CO
rH
09
D

^
U
-H
ft

O
c
CO
0
G
UJ





p^

o
CJ






















UJ

t

CJ


f~^
CO

ca
o

^
c
09


e
a
u

CO
c
3
^4
U.
u.





CM

O
CJ



















UJ
CO
c
••H
ca


CO
•a

CO

CO
CO



CO
CO
a

_^
O
• H
ft

C
c
—
3
*





Ks

O
CJ




































,^
CO

09
a
**^

CO
CO
• r-l
b.
CO
•H
&
ca





-^

0
CJ


























^-^
ca

^





c
a

c

_
CO

u-

o
CO
"o
H-
U



5

lA

O
CJ

































^-^
CO
CO
c.
CJ
c


_^
0
-H
ft

O1
C
CO
o
CJ
Q





^o

a
CJ



































CO
09
0

_^
0

in

ol
c
CO
o
CJ
UJ





r^

o
CJ






















UJ

1

CJ


f — ^
CO

CO
c

_^
c
CO


o
u

CO
c
3

CO
ut
u.





CM

o
CJ



















UJ
o
c
CO


a

tj
V

CO
CO



CO
CO


^
u
H
in

Ol
c
c
CO
a.
a
•*





K\

O
CJ




































^. ^
CO

CO
Cl


09
CO

CO
H
&
CD





•3-

a
o


























^— *.
ca

^





c
a
•~
c

^-i
j)

u.
u-
o
09
O
CJ



o
CM

lA

O
CJ

































^~,
CO
CO
?
•H

_^
CJ
• H
in

o
c
09
O
CJ
O





SO

o
u



































"oO
CO
O

_^
u
•1-4
in

o
c
co
o
CJ
UJ





p*

0
o






















UJ

1

o


^~*
o

09
o

_^
a
^_

R
u
It.
CO
c
cS
rH
09
U.




                                                                                                                                                                                             ca

                                                                                                                                                                                             01

                                                                                                                                                                                             J?


                                                                                                                                                                                             ^
                                                                                                                                                                                             a
                                                                                                                                                                                             U
                                                                                                                                                                                             tl
                                                                                                                                                                                             O
                                                                                                                                                                                             u
 OJ

JZ
                                                                                                                                                                                             a
                                                                                                                                                                                             u
j=
4-1

 o
OJ

 09

"3

"o
>—


 01

 CO

in


^
 09

-------
s
u
ce
Ul
>—
tn
>-
tn
u u
0) 0 +>
<*- .U  CD ,C
> 0?
U C
h- t-H
§00
c - "3
eg co jj
>- >. o
l«
03 •»
|s
|S
CD •»
H*

CO «k
eg «
1 =




a
0
03
O

CM

O
u












Ul
CD
C
03


03
•o
Ui
U
CO
01
9-


"oj

a
o

JX
o
1/5

O
c
1
o
*




1^

o
u



























CO

03
57
(0
01

U
01
I
CO




^

o
o


















CO
•f



O
C
I
o
<




K\

o
u



























03
PH
J
03
01


I
CD




^

O
U


















CO


^



c
CO
t—
c

1
u_

o
03
ts
U


CM

IA

O
u























^^
00
1)
y
•H

U
to

o
c
03
0
U
a




*o

3

























03

03
O

U
.u
to

g
CO
o
o
Ul




p*.

o
u














Ul

1

u

CO

03
o

_^
Q
h-

1
It-
CD
c
cS
•~4
u.
u.



                                                                                                                                                                     CO
                                                                                                                                                                     ^
                                                                                                                                                                     to
                                                                                                                                                                     c
                                                                                                                                                                     IB
                                                                                                                                                                      CD
                                                                                                                                                                      CO
                                                                                                                                                                      o
                                                                                                                                                                      u
                                                                                                                                                                      4)

                                                                                                                                                                     JT
                                                                                                                                                                      C
                                                                                                                                                                      O
O
u
                                                                                                                                                                      CD
                                                                                                                                                                      a
                                                                                                                                                                      0)

                                                                                                                                                                     JZ
                                                                                                                                                                      CO

                                                                                                                                                                     "3
                                                                                                                                                                     ^J
                                                                                                                                                                      o
                                                                              F-41

-------
s
LjJ
cr
O
U
a

5
(4 U
O o 4J
<«- 4J CO
C 4-1 CD jC
2 > (75
e a
JJ CO rH
eg a 4J
1— >. O
in >::
Is
eg 4k
h-
CO 4k
CO .
CM

o
u












LJ
CO

a
T3
CU
4J
CO
cu

^s


CO
a
c
*
•H

Ul

q
c

c
1
*




«

0
u


























co
eg
^**
CO
CO

u
cu

.^

en






cS
















a

•f


^^


£
co
c


u.

u-
o

a
"e
u



CM
,A

H
CO

>.
CO

co

ca
c



,—*
cc
a
01
j,

S'


01
c

c
1
<




^

o
u


























CO
a

a
Cv

i

.«.
&
CD




^

O
U
















£

^


x^, ^


c
CO
c


u.

u-
o

s
4-1
c
u



vO
lA

O
u






















ff^
a
cc
?

^

j_l
t^

?

a
c
a




vO

cS
























s-^
. CO
"3
o
Jt

L
Ifl

?

a
c
LJ




p^

cS













LJ

1
U

^~*
cc

CO
c


CO
H-
§


ct

cS

^
K
U.
U.




CM

°












UJ
ca

>.
a
•o
cu
4J
CO
cu



s~*
CO
CD
IT
*

. i
^

cr
c

c
I
*




^

o
u


























CO
CO
01

CO
cu

cu

• H
&
en




_^

a
u
















en

+


*^,


c
CO
c

3
U.

U-
0

CO
a
u



PS.
CM
^

O
u






















*-**.
a
cu
u
c

•g

4J
tn

CJ
c

u
c
CJ
a




M5

o
u
























^— V
co
"3
CJ
X
•tH
. i
W

c?
•H
CO
a
u
LJ




^

O
U













LJ

1
U

y~x
CO

eg
c


1
t—
i
u_

cu

1Q
fj

^^
U.
U.




CM

O
u












s
•H
CO

>.
CO
•u
bt
u

a
cc



s~*
CO
CT
^

. 1
w

a

c
cc
•*




^

o
u


























CO
a

a
cc

cu
3

3
en






cS
















in

•f


^^

Jlf
c
a
c


u.

b.
o

a
"c
H-
U



co
CM
kA

«s






















^Si
tn
cc
£
c,
c

^

J_
w

CT

a
c
c
o




vfl

O
u
























^•s
CO
"3
01
^

i i
W

c?

cc
c
c
UJ




r-

cS













LJ

1
U

^•v
cr

a
c


1

§
it.

c
c
cS

^^
a
u.
u.




                                                                                                                                   CO
                                                                                                                                   X

                                                                                                                                   cu
                                                                                                                                   cu
                                                                                                                                   O
                                                                                                                                   4J
                                                                                                                                   O
                                                                                                                                   cu
                                                                                                                                   o
                                                                                                                                   u
                                                                                                                                   c
                                                                                                                                   o
                                                                                                                                   •o
                                                                                                                                   cu
cu
                                                                                                                                   CO
                                                                                                                                   ^H
                                                                                                                                   co
                                                                                                                                   4-1
                                                                                                                                   o
                                                                                                                                   (4
                                                                                                                                   CU
                                                                                                                                   '4.

                                                                                                                                   2
                                                                                                                                   CO
                                                                  F-42

-------
ce

X
o

a
U.
>—t
X.
U b
O O 4J
!»- jj a
C JJ O £
co > o!
b C
H- M
8 CO
.* a -i
C JJ CO
co co *j
^ x, o
in >Z
' l«
JC
c r«
(D «k
h-
"c »o
co •»
^
1*
^
C rdinqs in Gallons
3
OJ
cc
b
03
Jet
2


0)
4-1
O
Q
X
eg
a
CM
O
u









gals) (Yesterday's Line E)
.*
u
•M
^>
en
o
c
c
<§
<


K\
cS









IB
CO
o
CO
®
•H
U
?
•1-1
^
5
00


-*
o
u









a
+
-
'to
*H
CO
a
j<
u
•*H
^1
to
g
&
a
f


1*1
cS









00
o
a
0
u
•»4
b
01
>*H
^H
<§
CD



-------
                  INSTRUCTIONS FOR COMPLETING
   INVENTORY SHEET FOR TANKS WITHOUT METERED DISPENSING PUMPS
     The Inventory Sheet for Tanks Without Metered Dispensing
Pumps is the only sheet to be used with tanks having unmetered
dispensing pumps.  Without metered dispensing pumps, it is
difficult to use inventory records to monitor for fuel losses,
because the quantities of fuel being pumped from that tank are
unknown.  As a result, inventory calculations must be based on
stick readings alone.  You will need an accurate dipstick and
the correct inches-to-gallons conversion chart for your unmet-
ered tank.

     You will need to make a series of 30 opening and closing
dipstick readings of your unmetered tank.  There should be one
or more days between each of the 30 readings.  Figure 2, below,
shows two plans for taking the 30 readings.
                            Figure 2

          Inventory Readings Plans for Unmetered Tanks
PLAN A:   Immediately before each withdrawal or delivery of
          fuel, enter the date and opening stick readings for
          the tank on the inventory sheet.  Immediately after
          the withdrawal or delivery make and record the closing
          stick reading on the inventory sheet.  Deliveries
          should be entered from the delivery receipt you re-
          ceive from the fuel truck driver.  (All deliveries
          will be made when the facility is "open," since the
          delivery will be occur between the opening and closing
          stick readings.)
PLAN B:   At the beginning of each operating day (before any
          withdrawals of fuel) record the date and opening stick
          reading for the day.  At the end of the day (after all
          withdrawals of fuel) record the closing stick readings.
          If a delivery occurs while "closed" (after the closing
          dipstick reading was taken) record the quantity deliv-
          ered (from the receipt) on the line for the following
          day and circle the code (2) for "closed."  If a deliv-
          ery occurs while your facility is open, record the
          quantity delivered on the line for the day the delivery
          occurred and circle the code (1) for "open."  NOTE:
          It is not necessary to have withdrawals or deliveries
          during an operating day in order to fill in an inventory
          line, as long as you make and record both opening and
          closing stick readings.

                                F-44

-------
     If your tank is used very infrequently (once a day or less)
you may wish to follow Plan A.  Plan A requires that you record
dipstick readings on the tank each time you use the tank.  If
the tank is used more than once a day, you should follow Plan B.
Plan B requires that you record dipstick readings at the opening
and closing of each operating (business) day.

     The step-by-step instructions for recording inventory on
the "Inventory Sheet for Tanks Without Metered Dispensing Pumps"
are:
                                 s
     •    In Column 1, record the date that the inventory reading
          will be made (day and month).

     •    In Column 2, record the opening dipstick reading, in
          inches (to the nearest quarter inch).

     •    In Column 3, record the opening dipstick reading, in
          gallons (as calculated from your inches-to-gallons
          conversion chart for this tank).

     •    In Column 4, record the closing dipstick reading, in
          inches (to the nearest quarter inch).

     •    In Column 5, record the closing dipstick reading, in
          gallons (as calculated from your inches-to-gallons
          conversion chart for this tank).

     •    In Column 6, record the amount delivered to the tank
          since your closing reading on the line above.  (The
          "Gallons Delivered" should be taken  from the receipt
          provided by the fuel delivery truck  driver.)

     •    Finally, in Column 7,  please indicate whether the fuel
          delivery was made before  the opening stick reading on
          this line (i.e., when the facility was closed) or
          during the time between the opening  and closing stick
          readings (i.e., when the  facility was open).
                              F-45

-------
INVENTORY SHEET FOR  TANKS WITHOUT  METERED  DISPENSING  PUMPS
                    (Name of Facility)
                                                     Tank  Number: 	

                                                     Type  of  Fuel: _

                                                     Size  of  Tank:	

                                                     Year  Installed:
    (City/Town)
                     (Street  Address)
         (State)
(Zip)
Dipstick* I
Column 1 Column 2 Column 3
Opening Opening
Dipstick* Dipstick*
Reading Reading
Day Date (inches) (gallons)
1 [
2
3
4 I
5
6 I
7 [
8
9
10
11
12
13
14
15 I
16
17
18
19
20 I
21 I
22
23
24 I
25
26
27
28
29
30
nventory
Column 4 Column 5 Column 6 Column 7
Has delivery
Closing Closing made while
Dipstick* Dipstick* Deliveries open or closed?
Inventory Inventory (Circle one)
(inches) (gallons) (in gallons) Open Closed
I 1 2
1 2
1 2
I 1 2
I -12
I 112
1 2
1 2
I 1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
I 1 2
I 1 2
1 2
1 2
1 2
1 2
1 2
1 2
1 2
    »If tank has
     readings.
remote gauge, check here |	|  and use remote gauge readings instead of stick
                                             F-46

-------
                REPORTING  RESPONSIBILITIES

           OF TANK OWNERS  AND OPERATORS
     On November 8, 1984,  President Reagan signed the  Hazardous
and Solid Waste Amendments of 1984, amending the Resource
Conservation and Recovery  Act (RCRA).   Section 9005(a)  of  RCRA,
as amended, states:

     "FURNISHING INFORMATION—For the  purposes of developing  or
     assisting in the development of any regulation, conducting
     any study, or enforcing the provisions of this subtitle
     [Subtitle I of Title  VI, 'Regulation of Underground Storage
     Tanks], any owner or  operator of  an underground storage
     tank . .  . shall upon request of  any officer,  employee or
     representative of the Environmental Protection Agency duly
     designated by the Administrator,  .  . . furnish information
     relating  to such tanks, their associated equipment,  their
     contents, conduct monitoring or testing, and permit such
     officer at all reasonable times to  have access to,  and copy
     all records relative  to such tanks  [underline  added for
     emphasis].For the purposes of developing or  assisting  in
     the development of any regulation,  conducting  any study, or
     enforcing the provisions of this  subtitle, such officers,
     employees or representatives are  authorized —

          "(1) to enter at reasonable  times any establishment or
     other place where an  underground  storage tank  is  located;

          "(2) to inspect  and obtain samples from any  person  of
     any regulated substance contained in such tank; and

          "(3) to conduct  monitoring or  testing of  the tanks,
     associated equipment, contents, or  surrounding soils, air,
     surface water or ground water.

     Each such inspection  shall  be commenced and completed with
     reasonable promptness.

     Section 9006, "FEDERAL ENFORCEMENT," gives EPA the authority
to issue compliance orders and to commence civil actions for
noncompliance  with the requirements of Section 9005.   Section
90006(a)(3)  authorizes EPA to seek civil penalties  for violation
of such an order, not to exceed  $25,000  per day of  continued
noncompliance.
                            F-47

-------
               UNITED STATES EN V iRONMENTAL PROTECTION AGENC1

                           WASHINGTON, D. C. Z^-i-'.C
Dear Establishment Owner/Operator:

     If there are no underground motor fuel storage tanks located at
your establishment, please sign the certification statement below
indicating this and return in the postage paid envelope provided.  If
there are abandoned or out of service underground motor fuel storage
tanks at this establishment you should not sign this statement.  If
the interviewer calls after you have mailed the signed statement,
inform him/her that you have done so.
                                            Sincerely,
                                            Martin P. HaXper, Director
                                            Exposure  Evaluation  Division
Establishment Name:
Establishment Address:

Establishment Telephone:
                      CERTIFICATION STATEMENT FOR
                     ESTABLISHMENTS WITHOUT TANKS
         THE OWNER OR THE OPERATOR OF THE FACILITY, OR
         HIS AUTHORIZED REPRESENTATIVE, SHOULD SIGN
         AND DATE THE CERTIFICATION WHERE INDICATED.
         THE PRINTED OR TYPED NAME OF THE PERSON SIGN-
         THE CERTIFICATION SHOULD ALSO BE INCLUDED
         WHERE INDICATED.
CERTITIFICATION:

     I hereby certify that there are no underground motor fuel  storage
tanks at the establishment at the above address.  I am aware  that
there are significant penalties for submitting false information,
including the possibility of a fine.
Print or type name
Signature
       F-48
Date signed

-------
                                       OMB No.: 2070-0037

                                       Expires: December 31, 1985
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

            UNDERGROUND STORAGE TANK SURVEY

          ESTABLISHMENT OPERATOR'S QUESTIONNAIRE
                         Conducted by:
                  WESTAT
                  An Employee-Owned Research Corporation
                     ea'TfB
-------
A1.
                             A.   SCREENING INFORMATION

What type of establishment is this?  [CIRCLE ONLY ONE CODE]
                                                                                                      01
           a.  FARM OR RANCH	01                                    /60-61
           b.  GASOLINE SERVICE STATION.

               [PLEASE SELECT ONE OF THE FOLLOWING SUBCATEGORIES]:
                                                                                                    762-63
               b1.  FULL SERVICE STATION (WHERE MOTOR
                    VEHICLE REPAIR WORK IS DONE)	02
               b2.  LARGE,  HIGH VOLUME STATION	03
               b3.  CONVENIENCE STORE	04
               bit.  SELF SERVICE GASOLINE STATION	05
               b5.  OTHER [PLEASE DESCRIBE] 	

                    	  06                                    /64-65

           c.  MILITARY FACILITY 	  07
           d.  FEDERAL AGENCY OR OFFICE	08
           e.  STATE AGENCY OR OFFICE	09
           f.  LOCAL GOVERNMENT AGENCY OR OFFICE 	  10
           g.  MARINA	11
           h.  TAXI SERVICE OR COMPANY	12
           i.  BUS FLEET FACILITY	13
           j.  TRUCK FLEET  FACILITY	14
           k.  AIRPORT OR AIRFIELD	15
           1.  RAILROAD DEPOT	16
           m.  OTHER BUSINESS [PLEASE SPECIFY YOUR
               ESTABLISHMENT'S PRIMARY PRODUCT OR                                                   /66-67
               SERVICE]: 	  17


                                                                      PLEASE DO NOT COMPLETE THE
           n.  BULK FUEL PLANT OR TERMINAL	18 —*• REST OF THIS QUESTIONNAIRE!
           o.  PRIVATE RESIDENCE THAT IS NOT ASSOCIATED               PLEASE CALL WESTAT AT THE
               WITH A FARM OR RANCH	19 —*•  800-638-8985 (TOLL FREE NUMBER)
           p.  OTHER [SPECIFY]: 	  20      AND ASK FOR THE "EPA SPECIALIST."
                                             BOX A1

      IF A1 = MILITARY, FEDERAL, STATE OR LOCAL AGENCY (CODES 07, 08, 09 OR 10), CHECK

      HERE     AND SKIP TO A6.  OTHERWISE, GO ON TO A2.
                                                                                                     768-69
                                                                                              770
A2.
Is this establishment owned and/or operated by a major petroleum company?  [CIRCLE ONLY
ONE CODE]

                                  YES	
                                  NO	
                                                                                                     771
                                           F-50

-------
 A3.    What  is  the  name  and  address  of the  owner  of this establishment?

            Owner's Name    	

            Owner's Address
A4.    What  is the  name  and  address  of  the  operator of this establishment?

            Operator's Name:     	
            Operator's Address:
A3.    What  is the motor  fuel that  is  stored  at  this establishment  used for:   retail sales, whole-
       sale  sales, or for use by the establishment  itself?   [CIRCLE ONE CODE  FOR EACH ITEM]

                                                                                       YES   NO

                                         a. RETAIL  SALES	1     2      /72
                                         b. WHOLESALE SALES	1     2      /73
                                         c. USE  BY  THIS  ESTABLISHMENT	1     2      774
                                         d. OTHER [SPECIFY]:	  1     2      /75
                                                                                                     X76-77
A6.   Does this establishment have any underground  storage  tanks  that  are  used to store
      motor fuel?  [CIRCLE ONLY ONE CODE]                                                            /78
                                        YES  [GO  ON  TO  A7].  .   1
                                        NO	2
PLEASE DO NOT COMPLETE THE REST OF
THIS QUESTIONNAIRE!  PLEASE CALL
WESTAT AT 800-638-8985 (TOLL FREE
NUMBER) AND ASK FOR THE "EPA
SPECIALIST."
A7.   How many underground storage tanks currently in use  are  used  to  store  motor fuels?

                                                NUMBER OF  TANKS: 	       /79-81


A8.   Does this establishment have any underground storage tanks that  are  used  to
      store used or waste oil?  [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO  A9]	1
                                                NO [SKIP TO A11]	2
                                                      F-51

-------
A9.   How many underground storage tanks currently in use are used to store used or waste oil?

           NUMBER OF USED OR WASTE OIL UNDERGROUND TANKS: 	         783-85


A10.  What (is/are) the capacity/ies of your used or waste oil tank(s)?  [ENTER CAPACITIES IN
      GALLONS]

           a.  Capacity of used or waste oil tank #1		 gallons     /86-91

           b.  Capacity of used or waste oil tank #2		 gallons     /92-97

           c.  Capacity of used or waste oil tank #3		 gallons     /98-103
                                           F-52

-------
A11.  Please fill out one Tank Description Sheet for each underground storage tank that this
      facility uses to store motor fuel.  There are six (6) Tank Description Sheets bound into
      this booklet.  If there are more than six underground storage tanks at this establishment,
      either photocopy as many additional sheets as are required to describe all the tanks, or
      write the answers to the questions for each extra tank on a plain sheet of paper.

      TANK DESCRIPTION SHEET INSTRUCTIONS

      1.   Use the space on the next page to draw a map of the underground tank area at your
          establishment.   On the map,  show the location of each tank, the pumps/dispensers for
          each tank,  and  any buildings and features associated with the tanks (such as a garage,
          driveway, or wall).   See the example map below showing a gasoline service station with
          three tanks and two  pump islands.

      2.   Assign a number to each underground  storage tank at this establishment,  and write
          that number on  the tank in your map.   (See example below.)   Also write the tank number
          in the upper lefthand  corner of the  Tank  Description Sheet  for that tank.

      3.   It is only  necessary to fill out Tank Description Sheets for tanks that  are on site
          at this establishment.   Do not  fill  out  Tank Description Sheets for any  tanks that
          this establishment may  use,  own or maintain off site.

      4.   If another  establishment  uses or maintains an underground storage tank on your
          establishment's  site/property,  you should complete a Tank Description  Sheet for  that
          tank and include it  on  your  map.

      5.   Large establishments with  more  than one tank area  may  find  it  easier to  draw
          individual  maps of each tank area, rather  than  drawing one  large  map.

                                                 Map
                                                    T--53

-------
                                                                                                                     021
                                                    TANK DESCRIPTION SHEET
T1.
15.
To.
18.
What is trie capacity of this tank7  .That is, what is
the maximum numoer of gallons of fuel it can hold9)
[ENTER CAPACITY  IS GALLONS]

     TANK DESIGN CAPACITY:
                                     	 gallons
                                                        1o-2l
      What is the average amount of fuel in this tank just
      before a aeliverv7  .'That is, what is the low ooint of
      the product level9;  [ENTER QUANTITY I\ GALLONS]
           AVERAGE CONTENTS BEFORE DELIVERY:
                                                      gallons
                                                        22-21
      What is the average amount of fuel delivered to this
      tank''  [ENTER QUANTITY IN GALLONS]

           AVERAGE SIZE Of DELIVERY:
                                         	 gallons
                                                        28-33
      What is the maximum number of gallons of fuel that has
      ever been stored in this tank7  *That is, how full have
      vou actually filled it7!  [ENTER QUANTITY IN GALLONS]
           LARGEST QUANTITY HELD IN TANK:
                                                      gallons
                                                        '34-39
      In what year was this tank installed1'

           YEAR OF INSTALLATION:
Was this tank new or used wnen it was installed''
[CIRCLE ONLY ONE CODE]

     NEW [SKIP TO T8]	1
     USED [GO ON TO T7]	2
     DON'T KNOW [SKIP TO T8] . .  .   8
                                                 /44
How old was this tank when it was installed7

     AGE IN YEARS: 	
                                                 ,'45-47
Is this tank scheduled for replacement or repair
within the next 12 months7  [CIRCLE ONLY ONE CODE]
           YES
           NO.
                                                       748
                                                       ,'49
T10.
T11.
      In what year was this tank last repaired7

           YEAR LAST REPAIRED:
      What types of repairs were done to this tank9

      REPAIRS:
                                                       ,'50-53
                                                       754-55
                                                       756-57
T12.  Which of the following fuel types were stored in this
      tank during the past 12 months7  [CIRCLE ONE CODE FOR
      EACH FUEL TYPE]
      b.
      c.
      d.
      e.
      f.
     Leaded gasoline .
     Unleaded gasoline
     Diesel fuel .  . .
     Aviation fuel  . .
     Gasohol 	
     Other  [SPECIFY]:
YES

 1
 1
 1
 1
 1
 1
T13.   Does this tank have a pump?

           YES [GO ON TO T14]	1
           NO [SKIP TO T18]	2
NO

 2
 2
 2
 2
 2
 2
                                                       /5B-65
                                                       766
                                                                                                          TANK
                                                                                                          NUMBER:
                       T14.  How many pumps are connectea  to tnis tank7

                                  NUMBER OF PUM"S 	
                                                                               D/-6S
                       T13.  Does this tank have a "suction" or a "submerged"  \ores-
                             sure  pump deliverv system7   [CIRCLE ONLY ONE CODE]

                                  SUCTION	01
                                  SUBMERGED	02
                                  OTHER [SPECIFY:: 	  03
                                                            T16.
                                                            T17.
                                                            T18.
                                                                              /70-71
                             How many dispensers '.nozzles! are connected to this tank7

                                  NUMBEK OF NOZZLES: 	
                                                                               72-74
                             Do the proouct dispensers (nozzles) for this tank nave
                             meters to measure the total quantity of product that nas
                             been pumped from the tank7  [CIRCLE ONLY ONE CODE]
                                                                       YES
                                                                       NO.
                                                                                                            1
                                                                                                                         ,'75
                                                                  Is this tank attached to another tank by pipes or lines7
                                                                  [CIRCLE ONLY ONE CODE]
                                  YES [PLEASE SPECIFY THE
                                    TANK NUMBERS) OF
                                    THE CONNECTED TANK(S)]
                                                                                                                         776
                                                                             NO.
                                                                  T19.
                                                                              /77-7S
                             How is this tank situated in relation to the water
                             table7  Is it:  [CIRCLE ONLY ONE CODE]
                                  Completely above the water table. .
                                  Partially above and partially
                                   below the water table	
                                  Or, is the top surface of the tank
                                   completely below the water table .
                                  Other [SPECIFY]: 	
                                                               01

                                                               02

                                                               03
                                                               04
T9.   Has this tank ever been repaired7  [CIRCLE ONLY ONE CODE]   T20.

           YES [GO ON TO T10]	1
           NO [SKIP TO T12]	2
           DON'T KNOW [SKIP TO T12], . .  8
                                                            T21.
                                                                  T22.
                                                            T23.
T24.
                                                                              779-80
                             Does this tank have a manway or other means of being
                             entered for internal inspection7  [CIRCLE ONLY ONE CODE]

                                  YES [GO ON TO T21]	1
                               .   NO [SKIP TO T23]	2
                                  DON'T KNOW [SKIP TO T23]. .  .  8
                                                                              781
                             Has the interior of the tank ever been inspected7
                             [CIRCLE ONLY ONE CODE]

                                  YES [GO ON TO T22]	1
                                  NO [SKIP TO T23]	2
                                  DON'T KNOW [SKIP TO T23]. .  .  8
                                                                              782
                             When was the most recent internal inspection of this
                             tan1-.7

                                  MOST RECENT INSPECTION: 	
                                                                              783-86
                             Has the tank ever been tested for leaks after it Has
                             placed in service7  [CIRCLE ONLY ONE COOT]
                                                                             YES [GO ON TO T24]	1
                                                                             NO [SKIP TO 126]	2
                                                                             DON'T KNOW [SKIP TO T26]. . .  8
                                                       787
What test method was used to test the tank?  (Please
give the brand name of the test, if known, and describe
the test procedure.  If more than one method was used,
describe all methods used.)
                                                                        METHOO(S):
                                                                                                                   788-89
                                                                                                                        790-91
                                                                  F-55

-------
                                              TANK  DESCRIPTION SHEET  (Continued)
T25.   In what year  was the tank  last  tested  by  this/these
      methods?

           YEAR LAST TESTED:  	
                                                       ;16-19
T26.   Of "hat material is this tank constructed''   [CIRCLE ONLY
      ONE CODE]

           FIBERGLASS-REINFORCED PLASTIC  .   01
           STEEL	02
           OTHER [SPECIFY]: 	  03
                                                       /20-21
T27.   Is the inside of this tank lined'  [CIRCLE ONLY  ONE CODE]

           YES [GO ON TO T28]	1
           NO [SKIP TO T30]	2
           DON'T KNOW [SKIP TO T30].  .  .   8
                                                       /22
T28.   In what year was the lining installed9

           YEAR LINED: 	
                                                       X23-26
T29.   Of what material is the liner constructed9  [CIRCLE
      ONLY ONE CODE]

           EPOXY-BASED RESINS	01
           FIBERGLASS REINFORCED PLASTIC  ...   02
           ISOPHTHALIC POLYESTER-BASED RESINS.   03
           POLYURETHANE-8ASED RESINS 	   04
           OTHER [SPECIFY]: 	  Q5
                                                       727-28
T30.  Is the outside of this tank coated?  [CIRCLE ONLY ONE
      CODE]

           YES [GO ON TO TJ1]	1
           NO [SKIP TO T32]	2
           DON'T KNOW [SKIP TO T32]. .  .   8
                                                       ,'29
T31.  Of what material is the coating constructed?  [CIRCLE
      ONL1 ONE CODE]

           FIBERGLASS/EPOXY	31
           4SPHALTIC MATERIAL	02
           URETHANE	03
           COAL TAR EPOXY	04
           OTHER [SPECIFY]: __	  05
                                                       /30-31
T32.  Is there secondary containment for this tank?  [CIRCLE
      uNLY ONE CODE]

           YES [GO ON TO T33]	1
           NO [SKIP TO T34]	2
           DON'T KNOW [SKIP TO T34]	8
 33.  Is this secondary containment a:

           concrete basin7	
           clastic-lined earth basin9 	
                                                       .'32
           clay-lined basin7	
           double-wall tank9	
           or something else [SPECIFY]:
01
02
03
04
05
                                                       ,'33-34
T34.   Is there secondary containment for any equipment that is
       attached to this tank (such as pipes, pumps, valves,
       etc.)9  [CIRCLE ONLY ONE CODE]

           YES [GO ON TO T35]	1
           NO [SKIP TO T36]	2
           DON'T KNOW [SKIP TO T36]	8
 T35.   Is this secondary containment a:

           concrete basin9	01
           plastic-lined earth basin9 	  02
           clav-lined basin9	03
           double-wall piping9	04
           or something else [SPECIFY]: 	  05
                                                       /35
                                                                                                          TANK
                                                                                                          NUMBER:
                    T36.   What  is  the name of the company that installed the
                          tank9
                               INSTALLER:
                    T37.   Is there a paved surface over tne tank9

                               YES [GO  ON  TO T38]	    1
                               NO  [SKIP TO T40]	2

                    T38.   Is this  pavement:

                               asphalt9	01
                               concrete?	02
                               gravel?	03
                               other [SPECIFY]:  	  04
                    T39.
                    T41.
                          How  thick  is the pavement?

                               THICKNESS:
                                   [CIRCLE ONE]:
                                    INCHES	01
                                    FEET	02
                                    OTHER  [SPECIFY]: 	
                                                                03
                    T40.   What 13 the distance from the surface to
                          the top of the tank9

                               DISTANCE TO SURFACE:
                                   [CIRCLE ONE]:
                                    INCHES	01
                                    FEET	02
                                    OTHER [SPECIFY]:
                                                                03
                          Does this tank have any of the following kinds
                          of protection against corrosion?  [CIRCLE ONLY
                          ONE CODE FOR EACH ITEM]
                                                               Yes  No
                          a.    Passive cathodic protection
                               (using sacrificial  anodes)9 .  .
                          b.    Cathodic protection using
                               impressed current9	
                          c.    Other [SPECIFY]: 	
T42.   Has this tank ever been founa to be leaking9
      [CIRCLE  ONLY  ONE CODE]

           YES [GO  ON TO T43]	1
           NO  [SKIP TO T44]	2
           DON'T  KNOW [SKIP  TO T44] ...  8

T43.   How was  the leak detected and/or verified?
      [CIRCLE  ALL THAT APPLY]

           INVENTORY RECONCILIATION ...  01
           ENVIRONMENTAL MONITORING ...  02
           FACILITY INSPECTION	03
           TANK TESTING	04
           OTHER  [SPECIFY]: 	
                                                               05
                    T44.  Have the lines 'piping)  for tnis tank ever
                          been 'ound to be leaking?  [CIRCLE ONLY
                          ONE CODE]
                               'iES.
                               SO .
                               DON"
                                                                           ,38-39
                                                       .,40
                                                       /41-42


                                                       ,''43-45




                                                       /46-47







                                                       /48-50




                                                       /51-52
                                                       /53

                                                       /54
                                                       /55
                                                                           ,'56-57
                                                                           ,•58
                                                                           •59-60
                                                                           .61-62
                                                                           ,63-64
                                                                           ,65-66
                                                                           ,'67-68
                                                       F-56

-------
                                                                                                                     G2!
                                                    TANK DESCRIPTION SHEET
                                                                                                          TASK
                                                                                                          NUMBER:
11.
T2.
T3.
T4.
T5.
T6.
T7.
T8.
T9.
T10.
Til.
What is the caoacit\ of tnis tank7   Tnat is. what is
tne maximum number of gallons of fuel it can hold'''
[ESTER CAPACITY IN GALLONS]

     TANK DESIGN CAPACITY:
           T14.  How many pumps are connected to this tanK7

                           rt or PUMP5 _
                               	 gallons
                                                " '16-21
What is the average amount of fuel in this tank just
before a delivers"1   That is, what is the low point of
the product level''   [ENTER QUANTITY IN GALLONS]
                                                                                             _
                                                                                                                         ,67-o9
                                                                  T15.   Does this tank nave a "suction" cr a "submeroed" .pres-
                                                                        sure;  pump deliverv svsten;'1  [LiRCLE ONLY ONE CODE]
                                                                             SuCTION 	
                                                                             SUBMERGED .  .  .  .
                                                                             CTHER [SPECIFY j:
                                                     C1
                                                     02
                                                     03
           AVERAGE CONTENTS BEFORE DELIVER*:
                                                      oallons
                                                      " 22-27
      What is the average amount of fuel delivered to this
      tank7  [ENTER QUANTITY IN GALLONS]

           AVERAGE SIZE OF DELIVERY:
                                                                                                                   •70-71
                                                            T16.  How many dispensers ,nozzles, are connected to this tank7
                               	 gallons
                                                  28-33
What is the maximum number of gallons of fuel that has
ever been stored in this tank7  (That is, now full have
vou actually filled it7)  [ENTER QUANT I TV IN GALLONS]
                                                                             NUMBER OF NGZZLES:
                                                                  T17.
                                                                   72-7«
                 .0 the product dispensers  nozzles'1 for tnis tank have
                 meters to measure tne total quantity of product that nas
                 been pumped from the tank7  [CIRCLE ONLV ONE CODE]
           LARGEST QUANTITY  HELD IN TANK:
                                                      gal Ions
                                                        34-39
                                                                       YES
                                                                       NO.
      In what year was this tank installed7

           YEAR OF INSTALLATION:
                                                                                                                         775
                                                            T18.  Is this tank attached to another tank by pipes or lines7
                                                                  [CIRCLE ONLY ONE CODE]
                                                 /40-43
Was this tank new or used when it was installed7
[CIRCLE ONLY ONE CODE]

     NEW [SKIP TO T8]	1
     USED [GO ON TO T7]	2
     DON'T KNOW [SKIP TO T8] . .  .   8
                                                 744
How old was this tank when it was installed7

     AGE IN YEARS:  	
                                                 745-47
Is this tank scheduled for replacement or repair
within the next 12  months7  [CIRCLE ONLY ONE CODE]
                                                                             YES [PLEASE SPECIFY  THE
                                                                               TANK NUMBERvS)  QF
                                                                               THE CONNECTED TANK(S)]
                                                                                                                         '16
                                                                             NO.
                                                                                                                   /77-7S
                                                            T19.  How is this tank situated in relation to the water
                                                                  table7  Is it:  [CIRCLE ONLY ONE CODE]
           YES
           NO.
                                                       748
                      Completely above the water table. ...
                      Partially above and partially
                       below tne water table .........
                      Or, is the top surface of the tank
                       complete!)  below the water table ...
                      Other [SPECIFY]: _
                                                                                                                01

                                                                                                                02

                                                                                                                03
                                                                                                                         779-80
      Has this tank ever been repaired7  [CIRCLE ONLY ONE CODE]   T20.   Does this tank have a manway  or other means of being
                                                                        entered for internal inspection7  [CIRCLE ONLY ONE CODE]
           YES [GO ON TO T10]	1
                                                                             YES [GO ON TO T21]	1
                                                                             NO [SKIP TO T23]	2
     NO [SKIP TO T12]	2
     DON'T KNOW [SKIP TO T12]. .  .  8
                                                       749
      In what year was this tank last repaired7

           YEAR LAST REPAIRED:
      What types of repairs were done to this tank7

      REPAIRS:
                                                       750-53
                                                       754-55
                                                                       DON'T KNOW [SKIP TO T23]. .  .   8

                                                            T21.   Has the interior of the tank ever been inspected7
                                                                  [CIRCLE ONLY  ONE CODE]

                                                                       YES [GO ON TO T22]	1
                                                                       NO [SKIP TO T23]	2
                                                                       DON'T KNOW [SKIP TO T23]. .  .   8
                                                                                                                         781
T12.
                                                 756-57
Which of the following fuel tvpes were stored in this
tank during the past 12 months?  [CIRCLE ONE CODE FOR
EACH FUEL TYPE]
                                                                  T22.
                                                                  782
                 When was the most recent internal inspection of this
                 tank?

                      MOST RECENT INSPECTION:
                                                                                                                         783-86
                                                                  T23.   Has the tank  ever  been tested for leaks after  it  was
                                                                        placed in service7  [CIRCLE  ONLY  ONE  CODE]
      a.    Leaded gasoline .
      b.    Unleaded gasoline
      c.    Diesel fuel .  . .
      d.    Aviation fuel  . .
      e.    Gasohol 	
      f.    Other  [SPECIFY]:
                                     YES

                                      1
                                      1
                                      1
                                      1
                                      1
                                      1
T13.   Does this tank have a pump7

           YES [GO ON TO T14]	1
           NO [SKIP TO T1B]	2
NO

 2
 2
 2
 2
 2
 2
                                                       758-65
                                                       766
                                                                       YES [GO ON TO T24]	1
                                                                       NO [SKIP TO T26]	2
                                                                       DON'T  KNOW [SKIP  TO T26].  .  .   8
                                                                                                                   787
                                                            T24,   What  test method was used to test  the tank?   (Please
                                                                  give  the brand name of the test,  if known,  and describe
                                                                  the test procedure.   If more than  one method was used*
                                                                  describe all  methods used.)

                                                                  METHOD(S):  	             	
                                                                                                                   788-89
                                                                                                                         790-91
                                                                 F-57

-------
                                              TANK  DESCRIPTION  SHEET  (Continued)
T25.
T26.
T27.
T28.
T29.
T30.
T31.
132.
In what year was the tank last tested by this/these
methods7

     YEAR LAST TESTED: 	___
                                                 /16-19
Of what material is this tank constructed9  [CIRCLE ONLY
ONE CODE]

     FIBERGLASS-REINFORCED PLASTIC .   01
     STEEL	02
     OTHER [SPECIFY]:  	  03
                                                 ,20-21
Is the inside of this tank lined1  [CIRCLE ONLY ONE CODE]

     YES [GO ON TO T28]	1
     NO [SKIP TO T30]	2
     DON'T KNOW [SKIP TO T30]. .  .  8
                                                 722
In what year was the lining installed9

     YEAR LINED: 	
                                                 ,'23-26
Of what material is the liner constructed?  [CIRCLE
ONLY ONE CODE]

     EPOXY-BASED RESINS	01
     FIBERGLASS REINFORCED PLASTIC ...  02
     ISOPHTHALIC POLYESTER-BASED RESINS.  03
     POLYURETHANE-BASED RESINS 	  04
     OTHER [SPECIFY]: 	  05
                                                 /27-2S
Is the outside of this tank coated7  [CIRCLE ONLY ONE
CODE]

     YES [GO ON TO T31]	1
     NO [SKIP TO T32]	2
     DON'T KNOW [SKIP TO T32]. . .  B
                                                 /29
Of what material is the coating constructed?  [CIRCLE
ONLY ONE CODE]

     FIBERGLASS EPOXY	01
     ASPHALTIC MATERIAL	02
     URETHANE	33
     COAL TAR EPOXY	04
     OTHER [SPECIFY]: 	  05
                                                 ,'30-31
Is there secondary containment for this tank7  [CIRCLE
ONLY ONE CODE]

     YES [GO ON TO T33]	  1
     \0 [SKIP TO T34]	2
     DON'T KNOW [SKIP TO T34]	3
                                                       '32
 33.  Is tnis secondary containment a:

           concrete basin7	01
           plastic-lined earth basin7 	   02
           clay-lined basin?	C3
           double-wall tank"1	04
           or something else [SPECIFY]:  	  05
                                                       , 33-34
T34.  Is there secondary containment for any equipment that is
      attacned to this tank -such as pipes,  pumps, valves,
      etc.)7  [CIRCLE UNLY ONE CODE]

           VES [GO ON TO T35]	1
           NO [SKIP TO T36]	2
           DON'T KNOW [SKIr" TO T36]	3
T35.  Is this secondary containment a:

           concrete basin7	01
           plastic-lined earth basin7 	   02
           clay-lined basin7	03
           douole-wall piping7	04
           or something else [SPECIFY]:  	  05
                                                       •35
                                                                                                          TANK
                                                                                                          NUMBER:
                                                                     T36.   What  is  the  name of the company  that  installed  the
                                                                           tank7
                                                                     T41.
                                                                                INSTALLER:
T37.   Is there a paved surface over the tank7

           YES [GO ON TO T38]	    1
           NO [SKIP 10 T40J	2

T38.   Is this pavement:

           asphalt7	01
           concrete7	02
           gravel7	03
           other [SPECIFY]:  	  04
T39.   How thick  is tne pavement7

           THICKNESS:
               [CIRCLE ONE]:
                INCHES	01
                FEET	02
                OTHER [SPECIFY]:  	

                	  03

T40.   What is the distance from the surface to
      the top of the tank9
                                                                                DISTANCE  TO SURFACE:
               [CIRCLE ONE]:
                INCHES	01
                FEET	02
                OTHER [SPECIFY]:  	

                	  03

      Does this tank have any of the following  kinds
      of protection against corrosion?  [CIRCLE ONLY
      ONE CODE FOR EACH ITEM]

                                           Yes   No
      a.   Passive cathodic protection
           fusing sacrificial anodes)? .  *   1     2
      b.   Cathodic protection using
           impressed current9	1     2
      c.   Other [SPECIFY]: 	  1     2
T42.   Has this tank ever been found to Be leaking7
      [CIRCLE ONLY  ONE CODE]

           YES [GO  ON ro T43]	1
           NO [SKIP TO T44]	2
           DON'T  KNOW [SKIP  10 T44] ...  8

T43.   how was the leak detected and/or verified9
      [CIRCLE ALL THAT APPLY]

           INVENTORY RECONCILIATION ...  01
           ENVIRONMENTAL MONITORING ...  02
           FACILITY INSPECTION	03
           TANK TESTING	04
           OTHER  [SPECIFY;: 	

                                           05
                                                               T44.  Have the lines  piping,1  for this tank ever
                                                                     been found to be lea«ing7
                                                                     ONE CODE]
                                                                                                [CIRCLE ONLY
                                                                          NO ....
                                                                          DON'T KNOW
                                                                                                                            /38-39
                                                                                                                             40
                                                                                                                            , '41-42
                                                                                                                            / 46-47
                                                                                                                            ,'48-50
                                                                                                                            751-52
                                                                                                                            ,53
                                                                                                                            • 54
                                                                                                                            '55
                                                                                                                            '56-57
                                                                                                                      ,53
                                                                                                                      ,'59-60
                                                                                                                       61-62
                                                                                                                       63-64
                                                                                                                      '65-66
                                                                                                                      -67-68
                                                       , 36-37

                                                      F-58

-------
                                                    TANK DESCRIPTION SHEET
      What is the caoacitv of this tank''   That is, what is
      the maximuiTi number of gallons of" fuel it can hold9
      [ENTER CAPACITY IN GALLONS]

           TANK DESIGN CAPACITY:
                                     	 gallons
                                                       .16-21
      What is the average amount of fuel in this tank just
      before a aeliverv9  .That is, what is the iow point of
      the product level",  [ENTER QUANTITY  IN GALLONS]
           AVERAGE CONTENTS BEFORE DELIVERY:
                                                      gallons
                                                        22-27
T3.
T4.
T5.
T6.
T7.
IB.
      What is the average amount of fuel delivered to this
      tank'  [ENTER QUANTITY IN GALLONS]

           AVERAGE SIZE OF DELIVERY:
                               	 gallons
                                                "-28-33
What is tne maximum number of gallons of fuel that has
ever been stored in this tank9  ,That is, how full have
vou actually filled it9,  [ENTER QUANTITY IN GALLONS]
     LARGEST QUANTITY HELD IN TANK:

In what year was this tank installed''

     YEAR OF INSTALLATION:
                                                      gallons
                                                        34-39
                                                 /40-43
Was this tank new or used when it was installed9
[CIRCLE ONLY ONE CODE]

     NEW [SKIP TO T8]	1
     USED [GO ON TO T7]	2
     DON'T KNOW [SKIP TO T8] . .  .  8

How old was this tank when it was installed9

     AGE IN YEARS: 	
                                                 745-47
Is this tank scheduled for replacement or repair
within the next 12 months9  [CIRCLE ONLY ONE CODE]
           YES
           NO.
                                                       749
T10.
T11.
      In what year was this tank last repaired9

           YEAR LAST REPAIRED:
      What types of repairs were done to this tank9

      REPAIRS:
                                                       750-53
                                                       754-55
T12.
                                                 756-57
Which of the following fuel types were stored in this
tank during the past 12 months9  [CIRCLE ONE CODE FOR
EACH FUEL TYPE]
           Leaded gasoline .
           Unleaded gasoline
           Diesel fuel .  .  .
           Aviation fuel  .  .
           Gasohol 	
           Other [SPECIFY]:
                                     YES

                                      1
                                      1
                                      1
                                      1
                                      1
                                      1
T13.   Does this tank have a pump9

           YES [GO ON TO T14]	1
           NO [SKIP TO T18]	2
NO

 2
 2
 2
 2
 2
 2
                                                       758-65
                                                       766
                                                                                                          TANK
                                                                                                          NUMBER:
                                                                  T14.
                                                                  T15.
                                                                  how manv pumps are connected to this tank0

                                                                       NUMBER OF PUMPS 	
                                                                                                                    '67-69
                                                                  Does trus tank have a "suction" or a "suomerged" * pres-
                                                                  sure  pump delivers svstem9  [CIRCLE ONLY ONE CODE]

                                                                       SUCTION	01
                                                                       SUBMERGED	02
                                                                       CTnER [SPECIFY]: 	  03
                                                            T16.
                                                                  T17.
                                                                  T18.
                                                                  ,'70-71
                 how many dispensers '.nozzles- are connected to this tank9

                      NUMBER OF NOZZLES: 	
                                                                  X72-74
                 Do the product dispensers ^nozzles) for this tank have
                 meters to measure the total quantity of product tnat has
                 been pumped from the tank9  [CIRCLE ONLV ONE CODE]
                      YES
                      NO.
                                                                        Is this tank attached to another tank by pipes or lines9
                                                                        [CIRCLE ONLY ONE CODE]
                                                                             YES [PLEASE SPECIFY THE
                                                                               TANK NUMBER-'S  OF
                                                                               THE CONNECTED TANKlSi]
                                                                                                                         776
                                                                             NO.
                                                                  T19.
T9.   Has this tank ever been repaired9  [CIRCLE ONLY ONE CODE]   T20.

           YES [GO ON TO T10]	1
           NO [SKIP TO T12]	2
           DON'T KNOW [SKIP TO T12).  .  .  8
                                                            T21.
                                                                  T22.
                                                                  T23.
F24.
                                                                  777-78
                 How is this tank situated in relation to the water
                 table9  Is it:  [CIRCLE ONLY ONE CODE]

                      Completely above the water table. ...  01
                      Partially above and partially
                       below the water table	02
                      Or, is the top surface of the tank
                       completely below the water table ...  03
                      Other [SPECIFY]: 	  04
                                                                  779-80
                 Does this tank have a manwav or other means of being
                 entered for internal inspection9  [CIRCLE ONLY ONE CODE]

                      YES [GO ON TO T21]	1
                      NO [SKIP TO T23]	2
                      DON'T KNOW [SKIP TO T23]. . .  8
                                                                  781
                 Has the interior of the tank ever been inspected9
                 [CIRCLE ONLY ONE CODE]

                      YES [GO ON TO T22]	1
                      NO [SKIP TO T23J	2
                      DON'T KNOW [SKIP TO T23], . .  8
                                                                  782
                 When was the most recent internal inspection of this
                 tank9

                      MOST RECENT INSPECTION: 	
                                                                  783-86
                 Has the tan* ever been tested for leaks after it was
                 placed in service9  [CIRCLE ONLY ONE CODEJ
           YES [GO ON TO T24]	1
           NO [SKIP TO T26]	2
           DON'T  KNOW [SKIP TO T26].  .  .   8
                                                       787
      What test method was used to test the tank9  (Please
      give the brand name of the test,  if known,  and describe
      the test procedure.  If more than one method was used,
      describe all methods used.)

      METHOD(S):	     	
                                                                                                                   788-89
                                                                                                                         790-91
                                                                F-59

-------
                                                                                                                   0-51
                                              TANK  DESCRIPTION  SHEET  .'Continued)
T25.   In what year was the tank last  tested  by  this/these
      methods''
           YEAR LAST  TESTED:
                                                       /16-19
T26.  Of what material  is this tank constructed''   [CIRCLE  ONLY
      ONE CODE]

           FIBERGL* o-REINFORCED PLASTIC  .   01
           STEEL	02
           OTHER [SPECIFY]:  	  03
                                                       /20-21
T27.  Is the inside of this tank lined?  [CIRCLE ONLY ONE CODE]

           YES [GO ON TO T28]	1
           NO [SKIP TO T30]	2
           DON'T KNOW [SKIP TO T30].  .  .   8
                                                       /22
T28.  In what year was the lining installed''

           YEAR LINED: 	
                                                       X23-26
T29.  Of what material is the liner constructed?  [CIRCLE
      ONLY ONE CODE]

           EPOXY-BASED RESINS	01
           FIBERGLASS REINFORCED PLASTIC  ...  02
           ISOPHTHALIC POLYESTER-BASED RESINS.  03
           POLYURETHANE-8ASED RESINS 	  04
           OTHER [SPECIFY]: 	  05
                                                       /27-28
T30.  Is the outside of this tank coated''  [CIRCLE ONLY ONE
      CODE]

           YES [GO ON TO T31]	1
           NO [SKIP TO T32]	2
           DON'T KNOW [SKIP TO T32]. .  .  8
                                                       /29
TJ1.  Of what material is the coating constructed?  [CIRCLE
      ONLY ONE CODE]

           FIBERGLASS/EPOtY	01
           ASPHALTIC MATEHIAL	02
           URETHANE	03
           COAL TAR EPOXY	04
           OTHER fSPECIFY]: 	  05
                                                       /30-31
T32.   Is there secondary containment for this tank7  [CIRCLE
       ONLY ONE CODE]

           YES [GO ON TO T33]	  1
           NO [SKIP TO T34]	2
           DON'T KNOW [SKIP TO T34]	8
 T33.   Is this secondary containment a:
                                                       ;32
           concrete basin9	
           plastic-lined earth basin7 .
           clav-lined basin9	
           double-wall tank7	
           or something else [SPECIFY]:
01
02
03
04
05
                                                       ,'33-34
 T34.   Is there secondary containment for anv equipment that is
       attached to this tank (such as pipes, pumps, valves,
       etc.)7  [CIRCLE ONLY ONE CODE]
           1ES [GO ON TO T35] . . .
           SO [SKIP TO T36] ....
           DON'T KNOW [SKIP TO T36]
 T35.   Is this secondary containment a:
           concrete basin7	
           plastic-lineo earth basin7  .
           clay-lined Dasin7	
           double-wall piping7	
           or something else [SPECIFY]:
                                                 1
01
02
03
34
05
                                                       . 35
                                                                                                          TANK
                                                                                                          NUMBER:
                    T36.   What  is  the name of the company that installed the
                          tank?
                                                                                INSTALLER:
                    T37.   Is  there  a  paved  surface over  the tank7

                               YES  [GO  ON TO T38]	    1
                               \0  [SKIP TO  T40]	2

                    T38.   Is  this  pavement:

                               asphalt7	01
                               concrete7	02
                               gravel7	03
                               other  [SPECIFY]: 	  04
                    T39.   How thick  is the pavement7

                               THICKNESS:
                                   [CIRCLE ONE]:
                                    INCHES	01
                                    FEET	02
                                    OTHER [SPECIFY]:  	
                                                                03
                    T40.   What is the distance from the surface to
                          the top of the tank?
                    T41.
                                                                                DISTANCE TO SURFACE:
               [CIRCLE ONE]:
                INCHES	01
                FEET	02
                OTHER [SPECIFY]: 	

                	  03

      Does this tank  nave any  of  the following  kinds
      of protection against corrosion?  [CIRCLE ONLY
      ONE CODE FOR EACH ITEM]

                                           Yes   No
      a.   Passive cathodic orotection
           ; using sacrificial  anodes)? .  .   1     2
      b.   Cathodic protection using
           impressed  current7	1     2
      c.   Other [SPECIFY]: 	  1     2
T42.   Has this tank ever been found to be leaking7
      [CIBCLE ONLY ONE CODE]

           YES [GO ON TO T43]	1
           NO [SKIP TO T44]	2
           DON'T  KNOW [SKIP TO T44] ...  8

T43.   How was the leak detected and/or verified?
      [CIRCLE ALL THAT APPLY]

           INVENTORY RECONCILIATION ...  01
           ENVIRONMENTAL MONITORING ...  02
           FACILITY INSPECTION	03
           TANK TESTING	04
           OTHER  [SPECIFY]: 	

                                           05
T44.  Have the lines 'piping) for this tank ever
      oeen founa to be leaking7  [CIPCLE ONLY
      ONE CODE]

           YES	1
           NO	2
           DON'T 
-------
                                                                                                                   L91I
                                                    TANK DESCRIPTION SHEET
T1.
T2.
T3.
T4.
T5.
T6.
T7.
T8.
What is the capacity of this tank?  (That is, what is
the maximum number of gallons of fuel it can hold?)
[ENTER CAPACITY IN GALLONS]
           TAN.  JESIGN CAPACITY:
                                	 gallons
                                                 /16-21
What is the average amount of fuel in this tank just
oefore a delivery9  (That is, what is the low point of
the product level?)  [ENTER QUANTITY IN GALLONS]

     AVERAGE CONTENTS BEFORE DELIVERY: 	 gallons
                                                 .'22-27
What is the average amount of fuel delivered to this
tank?  [ENTER QUANTITY IN GALLONS]
     AVERAGE SIZE OF DELIVERY: 	gallons
                                                 /28-3J
What is the maximum number of gallons of fuel that has
ever been stored in this tank?  (That is, how full have
vou actually filled it?)  [ENTER QUANTITY IN GALLONS]
           LARGEST QUANTITY HELD IN TANK:
      In what year was this tank installed?

           YEAR OF INSTALLATION:      	
                                                      gallons
                                                     '  X34-39
                                                 /40-43
Was this tank new or used when it wss installed?
[CIRCLE ONLY ONE CODE]

     NEW [SKIP TO T8]	1
     USED [GO ON TO T7]	2
     DON'T KNOW [SKIP TO T8] . .  .  8
                                                 /44
Haw old was this tank when it was installed?

     AGE IN YEARS: 	
                                                 /4S-47
Is this tank scheduled for replacement or repair
within the next 12 months?  [CIRCLE GNLY ONE CODE]
           YES
           NO.
                                          1
                                                       /48
T9.   Has this tank ever been repaired?  [CIRCLE ONLY ONE CODE]
T10.
Til.
     YES [GO ON TO T10]	1
     NO [SKIP TO T1Z]	2
     DON'T KNOW [SKIP TO T12]. .  .  8

In what year was this tank last repaired?

     >EAR LAST REPAIRED:  	
                                                       /49
      what types of repairs were done to this tank9

      REPAIRS:
                                                       /50-53
                                                       /54-5S
T12.
                                                 756-57
Which of the following fuel tvpes 
-------
                                              TANK  DESCRIPTION SHEET  'Continued
T25.
126.
      In what  year  was the tank  last  tested  by  this/these
      methods9
           YEAR LAST TESTED:
                                                 /16-19
Of what material  is this tank  constructed7  [CIRCLE  ONLY
ONE CODE]

     FIBERGLASS-REINFORCED PLASTIC .   01
     STEEL	02
     OTHER [SPECIFY]:  	  03
T27.
T29.
                                                 /20-21
Is the inside of this tank lined?  [CIRCLE ONLY ONE CODE]

     YES [GO ON TO T28]	1
     NO [SKIP TO T30]	2
     DON'T KNOW [SKIP tQ T30].  .  .   8
                                                 ,'22
In what year was the lining installed9

     YEAR LINED: 	
                                                 ,'23-26
Of what material is the liner constructed7  [CIRCLE
ONLY ONE CODE]

     EPOXY-BASED RESINS	01
     FIBERGLASS REINFORCED PLASTIC  ...   02
     ISOPHTHALIC POLYESTER-BASED RESINS.   03
     "OLYURETHANE-BASED RESINS 	   04
     OTHER [SPECIFY]:                     05
T30.
T31.
                                                 727-23
Is the outside of this tank coated7  [CIRCLE ONLY ONE
CODE]

     YES [GO CN TO T31 ]	1
     NO [SKIP TO T3Z]	2
     DON'T KNOW i SKIP TO T32). .  .  8
                                                 ,'29
Of wnat material is tne coating constructed7  [CIRCLE
ONLY ONE CODE]

     FIBERGLASS EPOXY	01
     ASPHALTIC MATERIAL	02
     URETHANE	03
     COAL TAR EPOXY	04
     OTHER [SPECIFY]:               05
                                                       ,30-31
      Is there secondary containment for this tank9  [CIRCLE
      ONLY ONE CODE]

           YES [GO ON TO T33]	  1
           SO [SKIP 10 T3«]
           DON'T KNOW [SKIP TO T34]
                                                       ,32
      Is this secondary containment a:
           concrete basin9	
           plastic-lined earth basin9 .  .   .
           clav-iined basin9	
           couDle-wall tank'	
           or something else [SPECIFY]:
                                           01
                                           02
                                           03
                                           04
                                           05
 T34.
                                                 ,'33-34
 Is  there secondary containment for any equipment that is
 attached to this tanx • such as pipes, pumps, valves,
 etc.,7  [CIRCLE ONLY ONE CODE]

      YES [GO ON I'D T35J	  1
      NO [SKIP TO '36]	2
      DON'T KNOW [SKIP TO T36]	8
                                                        '35
       Is  this secondarv containment a:

           concrete basn9	01
           clastic-lined earth basin" 	  02
           clay-lined basin9	03
           double-wall piping9	Oa
           cr something else [SPECIFY]: 	  05
                                                                                                          TANK
                                                                                                          NUMBER:
                                                               T36.   What  is  the  name of the company that installed the
                                                                     tank7
                                                                                INSTALLER:
T37.   Is there a paved surface over the tank7

           YES [GO ON TO T38]	    1
           NO [SKIP TO T40]	2

T38.   Is this pavement:

           asphalt9	01
           concrete7	02
           gravel7	03
           other [SPECIFY]:  	  04
                                                                     T39.   How thick  is the pavement7

                                                                                THICKNESS:
               [CIRCLE ONE]:
                INCHES 	
                FEEt	
                OTHER [SPECIFY]:
                                                                                                                 01
                                                                                                                 02
                                                                                                                 03
                                                                     T40.   What 15 the distance from the surface to
                                                                           the top of the tank9

                                                                                DISTANCE TO SURFACE:
               [CIRCLE ONE]:
                INCHES	01
                FEET	02
                OTHER [SPECIFY]: 	

                	  03

T41.  Does this tank have any of the following kinds
      of protection against corrosion7  [CIRCLE ONLY
      ONE CODE FOR EACH ITEM]
                                                                           a.   Passive catrtodic protection
                                                                                .using sacrificial anoces!9
                                                                           b.   Cathodic protection using
                                                                                impressed current9	
                                                                           c.   Other [SPECIFY]: 	
                                                               Ta2.  Has this tank ever been ""ound to De leaking'
                                                                     [CIRCLE ONLY ONE CODE]

                                                                          YES [GO CN TO *43]	1
                                                                          NO [SKIP TO T-i-j	2
                                                                          DON'T KNOW [SKIP TO T44J ...  8
T43.
      How was tne leak detected and/or verified7
      [CIRCLE ALL THAT APPLY]
INVENTORY RECONCILIATION .
ENVIRONMENTAL MONITORING .
FACILITY INSPECTION
TANK TESTING 	
OTHER [SPECIFY]:

. . 01
. . 02
03
. . 04

05
T44.  Have the lines .piping; for this tank ever
      ceen ""ound to be leaking9  [CIRCLE ONLY
      ONE CODE]

           YES	i
           NO	2
           DON'T KNOW	3
                                                                                                                            ,38-39
                                                                                                                            ;40
                                                                                                                            ,41-42
                                                                                                                            /48-50
                                                                                                                            ,51-52
                                                        53
                                                                                                                      •54
                                                                                                                      . 55
                                                                                                                       56-57
                                                                                                                             5B
                                                                                                                             59-60
                                                                                                                            ,bl-s2
                                                                                                                            '63-o4
                                                      F-62

-------
T2.
T3.
U.
T5.
16.
T7.
T8.
                                                                                                                    L9ii
                                                    TANK DESCRIPTION SHEET
      What is the capacity of this tank7  .That is, what is
      the maximum number of gallons of fuel it can hold''!
      [ESTER CAPACITY IN GALLONS]
     TANK DESIGN CAPACITY: 	 gallons
                                                  16-21
What is tne average amount of fuel in tnis tank just
before a delivery7  (That is, what is the low point of
the product level7}  [ENTER QUANTITY IN GALLONS]
           AVERAGE CONTENTS BEFORE DELIVERY:
                                                      gallons
                                                        22-27
      What is the average amount of fuel delivered to this
      tank7  [ENTER QUANTITY IN GALLONS]
           AVERAGE SIZE OF DELIVERY:
                                   	 gallons
                                                 .28-33
'"hat is the maximum number of gallons of fuel that has
ever been stored in this tank?  iThat is, how full nave
vou actually filled it?)  [ENTER QUANTITY IN GALLONS]
           LARGEST QUANTITY HELD IN TANK:
                                                      gallons
                                                        34-39
      In what year was this tank installed7

           fEAR OF INSTALLATION:
                                                  '40-43
Was this tank new or used when it was installed?
[CIRCLE ONLY ONE CODE]

     NEW [SKIP TO 13]	1
     USED [GO 3N T0 T7]	2
     DON'T KNOW [SKIP TO T8] .  .  .  8

How old was tnis tank when it was installed7

     AGE :N YEARS: 	
                                                 ,45-47
Is this tank scheduled for replacement or repair
within the next 12 months7  [CIRCLE ONLY ONE CODE]
                                                       /44
           VES
           NO.
                                                        49
T10.
111.
      In what vear was this tan* last repaired?

           YEAR LAST REPAIRED:      	
      What types of repairs were done to this tank7

      REPAIRS:
                                                        54-55
T12.
                                                  56-57
Which of the following fuel t-.pes were stored in *:nis
tank during tfe past 12 months7  [CIRCLE ONE CODE FOR
EACH CUEL TYPE]
      b.
      c.
      d.
      e.
     Leaded gasoline .
     Unleaded gasoline
     Diesel f'jel .  . .
     Aviation fuel  . .
     Gasohol 	
     Qtner [SPECIFY]:
YES

 1
 1
 1
NO

 2
 2
 2
                                                                                                          TANK
                                                                                                          NUMBER:
                       T14.  How many Dumps are connected to this tank7

                                  NUMBER OF PUMPS 	
                                                                              ,67-69
                       T15.  Does this tank have a "suction" or a "submerged" *pres-
                             sure) pump delivery system7  [CIRCLE ONLY ONE CODE]

                                  SUCTION	01
                                  SUBMERGED	02
                                  OTHER [SPECIFY]: 	  03
                                                            T16.
                                                                  T17.
                                                            T18.
                                                                              ,'70-71
                             How many dispensers '.nozzles) are connected to this tank7
                                                                             NUMBER OF NOZZLES:
                                                                               72-74
                             Do the product dispensers  .nozzles) for this tank have
                             meters to measure the total quantity of product that has
                             been pumped from the tank7  [CIRCLE ONLY ONE CODE]
                                                                       YES
                                                                       NO.
                                                                                                                         '75
                                                                  Is this tank attached to another tank by pipes or lines7
                                                                  [CIRCLE UNLY ONE CODE]
                                                                             YES [PLEASE SPECIFY THE
                                                                               TANK NUMBERIS) OF
                                                                               THE CONNECTED TANK(Si]
                                                                             NO.
                                                                                                                         776
                                                                                                                         /77-T
                                                                  T19.
T9.   Has this tank ever been repaired7  [CIRCLE ONLY ONE CODE]   T20.

           YES [GO ON TO T10]	1
           NO [SKIP '0 T12]	2
           DON'T KNOW [SKIP TO T12].  .  .   8
                                                            T21.
                                                                  T22.
                             How is this tank situated in relation to the water
                             table7  Is it:  [CIRCLE ONLY ONE CODE]

                                  Completely above the water table. ...  01
                                  Partially above and oartiallv
                                   below the water table	02
                                  Or, is the top surface of the tank
                                   completely below the water table ...  03
                                  Other [SPECIFY]: 	  Ou
                                                                              • 79-80
                             Does this tank 'lave a manwav or other means of being
                             entered for internal inspection7  [CIRCLE ONLY ONE CODC]

                                  YES [GO ON TO T21!	1
                                  NO [SKIP TU T23]	2
                                  DON'T KNOW [SKIP TO T23). . .  8
                                                                              '81
                             Has the interior of the tank ever been inspected7
                             [CIRCLE ONLY ONE CODE]

                                  YES [GO ON TO T22]	1
                                  NO [SKIP TO T23]	2
                                  DON'T KNOW [SKIP TO T23J. . .  8
                                                                              ,82
                             When was the most recent internal inspection of this
                             tan'<7
                                                                             MOST  HECENT  INSPECTION:
                                                                        Has the tank ever oeen tested for leaks a'ter it
                                                                        placed in service7  [CIRCLE ONLY  JNE CCDEJ
                                                                                                                          33-86
                       T24.
     YES [GO ON TO F2oJ ......  1
     NO [SKIP TO T26] .......  2
     DON'T KNOW '.SKIP TO T26]. .  .  8
                                                 •87
•hat test method was used '.o r.est the tank7   Please
give the orand name of tne test,  if Known, and descrioe
the test procedure,  if more than one method was used,
describe all methods used.,1
T13.   Does this tank nave a  pump7
           VES [GO ON  TO pa],
           '.0 [SKIP T0 T18J.  .
                                                       ,58-65
                                                                        METHOD'S,:
                                                                                                                          88-39
                                                                                                                          90-91
                                                         F-63

-------
                                              TANK  DESCRIPTION SHEET  Continues.
T25.
T26.
In wnat year was the tank last tested by this/these
methods9
           YEAR LAST  TESTED:
                                                 ,-16-19
Of wnat material is this tank constructed''  [CIRCLE ONLY
CNE CODE]

     FIBERGLASS-REINFORCED PLASTIC .   01
     STEEL	02
     OTHER [SPECIFY]:  	    03
T27.
T28.
T29.
                                                 ,20-21
Is the inside of this tank lined7  [CIRCLE ONLY ONE CODE]

     YES [GO ON TO T28]	1
     NO [SKIP TO T30]	2
     DON'T KNOW [SKIP TO T30]. .  .  8
                                                 ,'22
In wnat year was the lining installed9

     YEAR LINED: 	
                                                 '23-26
Of what material is the liner constructed?  [CIRCLE
ONLY ONE CODE]

     EPOXY-SASED RESINS	01
     FIBERGLASS REINFORCED PLASTIC ...  02
     ISOPHTHALIC POLYESTER-BASED RESINS.  03
     POLYURETHANE-BASED RESINS 	  04
     OTHER [SPECIFY]: 	  05
T30.
T31.
                                                 ,'27-23
Is the outsiae of this tank coated9  [CIRCLE ONLY ONE
CODE]

     YES [GO ON TO T31]	1
     NO [SKIP TO T32]	2
     DON'T 
-------
                                                                                                                   I  021
                                                    TANK DESCRIPTION SHEET
T1.
T2.
T3.
T4.
T5.
T6.
T7.
T8.
T10.
What is the capacity of this tank9  (That is, what is
the maximum number of gallons of fuel it can hold'7)
[ENTER CAPACITY IS GALLONS]
     TANK DESIGN CAPACITY: 	...    gallons
                                                 ,16-21
What is the average amount of fuel in this tank just
before a delivery'7  '.That is, what is the low ooint of
the product level')  [ENTER QUANTITY IN GALLONS]
           AVERAGE CONTENTS BEFORE DELIVERY:
                                                      gallons
                                                        22-27
What is the average amount of fuel delivered to tnis
tank'  [ENTER QUANTITY IN GALLONS]
           AVERAGE SIZE OF  DELIVERY:
                                   	 gallons
                                                  28-33
What is the maximum number of gallons of fuel that has
ever been stored in this tank?  vThat is, how full have
vou actually filled it')  [ENTER QUANTITY IN GALLONS]
           LARGEST QUANTITY HELD IN TANK:
                                                      gallons
                                                       /34-39
      In what year was this tank installed?

           YEAR OF INSTALLATION: ________
                                                 /40-43
Was this tank new or used when it was installed?
[CIRCLE ONLY ONE CODE]

     NEW [SKIP TO T8]	1
     USED [GO ON T0 T7]	2
     DON'T KNOW [SKIP TO T8] . .  .  3
                                                 ,'44
How old was tnis tank when it was installed'

     AGE IN YEARS: 	
                                                 ,'45-47
Is this tank scheduled for replacement or reoair
within the next 12 months'  [CIRCLE ONLY ONE CCDE]
           YES
           NO.
                                                       ,48
T9.   Has this tank ever been repaireo?  [CIRCLE ONLY QNE CODE]

           YES [GO ON TO T10]	1
           NO [SKI? TO T12]	2
           DON'T KNOW [SKIP TO T12]. .  .  8
      In what year was this tank last repaired'

           YEAR LAST REPAIRED:
                                                       /50-53
T12.
'13.
      What types of repairs were done to this tank'

      REPAIRS:
                                                 /56-57
Which of the following fuel types «ere stored in this
tank during the past 12 -nonths'  [CIRCLE CN£ CODE FOR
EACH FUEL TYPE]

                                                 NO
                                                        56-65
                                                       '66

0.

d.
P
r.

does



Diesel fuel 	


Other [SPECIFY]:

this tank nave a pump'
YES [GO CN rQ Mi], ...
NO rsi'ip ro rial 	
YES
1

. . 1
. . 1
. . 1
1

. . 1

                                                                                                          TANK
                                                                                                          NUMBER:
T14.   How many pumps are connected to this tank'

           NUMBER OF PUMPS 	
                                                       ,67-69
T15.   Does this tank have a "suction" or a "submerged" (pres-
      sure)  pump deliverv system'  [CIRCLE ONLY ONE CODE]

           SUCTION	01
           SUBMERGED	02
           OTHER [SPECIFY]: 	  03
                                                        70-71
T16.   How many dispensers (nozzles) are connected to this tank?
                                                                             NUMBER OF  NOZZLES:
                                                        72-74
T17.   Do the product dispensers (nozzles) for this tank have
      meters to measure the total quantity of product that has
      been pumped from the tank?  [CIRCLE ONLY ONE CODE]
                                                                       YES
                                                                       NO.
                                                                                                                         /75
                                                            T18.  Is this tank attached to another tank by pipes or lines'
                                                                  [CIRCLE ONLY ONE CODE]
           YES [PLEASE SPECIFY THE
             TANK NUMBER;s) OF
             THE CONNECTED TANKlS)]
                                                                             NO.
                                                                                                                         ,'76
                                                                                                                         '77-78
T19.   How is this tank situated in relation to the water
      table'  la it:   [CIRCLE ONLY ONE CODE]

           Completely above the water table. ...  01
           partially  above ana partially
            below the water table	02
           Or,  is the top surface of the tank
            completely below the water table ...  03
           Other [SPECIFY]: 	  04
                                                       ,79-80
T20.   Does this tank  have a manway or other means of being
      entered for internal inspection'  (.CIRCLE ONLY ONE CODE]

           YES  [GO ON TO T21]	;
           NO [SKI» TU T23]	2
           OCN'T KNOW [SKIP TO T23].  . .  3
                                                       /81
T21.   Has tne interior of the tank ever been inspected?
      [CIRCLE ONLY ONE CODE]

           YES  [GO ON TO T22]	1
           NO [SKIP TO T23]	2
           DON'T KNOW 'SKIP TO T23],  . .  8
                                                       /82
T22.   When was  tne most recent internal inspection of this
      tanK?
                                                                             MOST  RECENT  INSPECTION:
                                                                                                                         ,83-36
                                                                  T23.   Has  the  tank  ever  been  tested  for  leaks  after  it  was
                                                                        slaced  in  service'   [CIRCLE  ONLY ONE  CODE!

                                                                             '•'ES [GO  ON  TO  T24]	1
                                                                             NO  [SKIP TO T26]	2
                                                                             DON'T  KNOW  [SKIP TO  T26].  . .   8
                                                                                                                         37
                                                                  T24.   What  test  method "as useo to test  the  tank'    Please
                                                                        give  the brand name  o*~  the test, if  known, ana describe
                                                                        tne  test procedure.   If nore than  one  method was  used.
                                                                        describe all  methods used.

                                                                        METHODvS;:
                                                           F-65

-------
                                              TANK DESCRIPTION  SHEET  .Continued,
T25.
T26.
In what year was the tank last  tested by  this,'these
methods'"
           YEAR  LAST  TESTED:
                                                 .16-19
Of unat material is this tank constructed''  [CIRCLE ONLY
ONE CODE]

     FIBERGLASS-REINFORCED PLASTIC  .   01
     STEEL	02
     OTHER [SPECIFY]:  	  03
T27.
T28.
T29.
                                                 ,20-21
Is the inside of this tank lined"1  [CIRCLE ONLY ONE CODE]

     YES [GO ON TO T28]	1
     NO [SKIP FO T30]	2
     DON'T KNOW [SKIP TO T30).  .  .   8
                                                 /22
In what year was the lining installed''

     YEAR LINED: 	
                                                 723-26
Of -hat material is the Uner constructed'  [CIRCLE
ONLY ONE CODE]

     EPOXY-BASED RESINS	01
     FIBERGLASS REINFORCED PLASTIC  ...   02
     ISOPHTHALIC POLYESTER-BASED RESINS.   03
     POLYURETHANE-BASED RESINS 	   04
     OTHER [SPECIFY]: 	  	       05
T50.
T31.
                                                 ,77-28
Is the outside of this tank coated9  [CIRCLE ONLY ONE
CODE]

     YES [CO ON TO T31]	1
     NO [SKIP TO T32]	2
     DON'T KNOW [SKIP TO [32]. .  .  8
                                                 •29
Of what material is the coating constructed7  [CIRCLE
ONLY JNE CODE]

     FIBERGLASS LPOXY	01
     ASPHALTIC MATERIAL	02
     URETHANE	03
     COAL TAR EPQXY	04
     UTHER [SPECIFY]: 	  05
T32.
 53.
                                                 ,30-31
Is there secondary containment for this tank7  [CIRCLE
ONuY ONE CODE]

     YES [GO ON T3 T33]	  1
     NO [SKIP TO T3*]	2
     DON'T KNOW [SKIP  TO f?4 ]	8
                                                        32
      Is this secondary containment a:
           concrete basin7	
           plastic-lined earth basin7 ....
           clay-lined basin"	
           Double-wall tank7	
           or something eise [SPECIFY]:  	
                                           01
                                           02
                                           03
                                           34
                                           05
T34.
                                                  33-34
Is there secondary containment for any equipment that is
attached to this tank  .such as pipes, oumps, yalves,
etc.;7  [CIRCLE ONLY ONE CODE]
           YES [CO ON TO T35] .  .  .
           NO [SKIP TO '36] ....
           DON'T KNOW [SKIP TO T36]
                                                        35
T35.
      is this secondary containment a:

           concrete oasin7	01
           plastic-lined earth basin7 	  02
           ciay-li-'ed basin7	33
           3oubie-*all pioing7. 	  04
           or something else [SPECIFY]: 	  05
                                                                                                          TANK
                                                                                                          NUMBER:
T36.   What is the name of the company that instaliea the
      tank7
                                                                                INSTALLER:
                                                                     T37.   Is  there  a  paved  surface  over  the  tank7
                                                                                YES  [GO  ON  TO  T38]
                                                                                NO  [SKIP TO TuO]  .
                                                                     T38.   Is  this  pavement:
           aspnalt7	01
           concrete7	02
           gravel7	03
           ether [SPECIFY]:      	  04
T39.   How thick is the pavement7

           THICKNESS:
               [CIRCLE ONE]:
                INCHES 	
                FEET	
                OTHER [SPECIF^]:
                                                                                                                01
                                                                                                                02
                                                                                                                 03
                                                                     T40.   What  is  the  distance  from  the  surface to
                                                                           the  top  of the  tank7

                                                                                DISTANCE  TO  SURFACE:
               [CIRCLE ONE]:
                INCHES	01
                FEET	02
                OTHER [SPECIFY]:  	

                	  03

T41.   Does this tank have anv of the following  *inos
      of protection against corrosion7  [CIRCLE UNLY
      ONE CODE FOR EACH ITEM]
                                                                                Passive ca^hodic  protection
                                                                                .using  sacrificial  anodes'7
                                                                                Cathodic protection using
                                                                                impressed current7	
                                                                                Other [aPECIFi]:  	
T42.  Has this tank ever been found to De leaking7
      [CIRCLE ONLY ONE CODE]

           v£S [GO ON TO i43j	1
           NO [SKIP TO iitaj	2
           OON'T KNOW [SKIP TO r44J .  .  .  8
T43.
      How was the leak Detected and, or verified7
      [CIRCLE ALu InAT APPLY]
           ,'NVENTGKY RECONCILIATION .
           ENVIRONMENTAL MONITORING .
           FACILITY INSPECTION. .  . .
           TANK TESTING 	
           OT«ER [SPECIFY]: 	
                                                                                                          01
                                                                                                          02
                                                                                                          03
                                                                                                          04
                                                                                                                05
                                                               U4.  Have the lines  piping,  for this tank ever
                                                                     been ""ound to oe leaking7  [CIRCLE ONLY
                                                                     ONE CODE]
                                                                          YES. .
                                                                          NO ..
                                                                          DON'T
                                                                                                                            ,38-39
                                                                                                                            .40
                                                                                                                             41-..2
                                                                                                                            ,43-45
                                                                                                                            ,46-47
                                                                                                                            ,48-50
                                                                                                                             51-52
                                                                                                                       54
                                                                                                                       55
                                                                                                                             56-57
                                                                                                                             61-o2
                                                                                                                             63-64
                                                                                                                             fc5-b6
                                                                                                                            ,67-o8
                                                       ,36-37

                                                      F-66

-------
                                    B.  OPERATING PRACTICES
                                                                                                     ra
B1.   Do you (or another establishment employee) inventory the contents of your tank(s)  by
      measuring the depth of the contents with a dipstick?   [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO B2]	1
                                                                                                     /16
                                                NO  [SKIP TO B5]	2
B2.   How often do you inventory the tank contents?  [CIRCLE ONLY ONE CODE]

                                                TWICE DAILY	01
                                                DAILY	02
                                                WEEKLY	03
                                                EVERY TWO WEEKS	04      /17-18
                                                MONTHLY	05

                                                OTHER [SPECIFY]:	  06
B3.   Do you have a chart (or charts) that show how to convert the depth of the product in
      the tank(s) to gallons?  [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2
B4.   Are the inventory (stick) readings recorded in a log or journal or other permanent
      record such as a daily inventory report?  [CIRCLE ONLY ONE CODE]

                                                YES	1       /20
                                                NO	2
B5.   Do any of the underground motor fuel storage tanks at this establishment have remote
      gauges (either float or electronic) that show the quantity of product in the tank?  [CIRCLE
      ONLY ONE CODE]

                                                YES [GO ON TO B6]	1
                                                NO [SKIP TO B8]	2


86.   How often do you (or another establishment employee) inventory the contents of your
      tank(s) by reading the remote gauge(s)?  [CIRCLE ONLY ONE CODE]

                                                TWICE DAILY	01
                                                DAILY	02
                                                WEEKLY	03
                                                EVERY TWO WEEKS	04      /22-23
                                                MONTHLY	05
                                                OTHER [SPECIFY]: 	  06


                                               F-67

-------
B7.   Are the inventory (gauge)  readings recorded in a log or journal or other permanent record
      such as a daily inventory  report?  [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2
88.   Do the product dispensers for your tank(a) have meters that record the total
      quantity of fuel that has been pumped from the tank(a)?  [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO 89]	1        .
                                                NO  [SKIP TO 816]	2


89.   Do you (or another establishment employee) check and record the dispenser meter
      readings for the tank(s)?

                                                YES [GO ON TO B10]	1
                                                NO  [SKIP TO 812]	2


810.  How often do you check and record the dispenser meter readings?  [CIRCLE ONLY
      ONE CODE]

                                                TWICE DAILY	01
                                                DAILY	02
                                                WEEKLY	03
                                                EVERY TWO WEEKS	04      /27-2S
                                                MONTHLY	05

                                                OTHER [SPECIFY]:	  06


811.  Are the dispenser meter readings recorded in a log or journal or other permanent
      record such as a daily inventory report?  [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2       /29
812.  Do you (or another establishment employee) check the accuracy of your dispenser meters
      to make sure the meters correctly measure the amount pumped?  [CIRCLE ONLY ONE CODE]


                                                YES	1       /30
                                                NO	2
                                          F-68

-------
B13.  Does anyone other than you or another establishment employee (such as a state or
      county Weights and Measures official) check the accuracy of your dispenser meters?
      [CIRCLE ONLY ONE CODE]
                                                YES
                                                NO
314.  How often is the accuracy of your dispenser meters checked?  [CIRCLE ONLY ONE CODE]
           IF THE ACCURACY OF YOUR DISPENSER METERS IS NEVER CHECKED,  CHECK HERE
           AND SKIP TO B16.
B15.  About how often is it necessary to recalibrate (adjust the gauge of) your
      dispenser meters?  [CIRCLE ONLY ONE CODE]
B17.  Are inventory (stick or remote gauge)  readings of your tank(s) taken immediately before
      receiving a fuel delivery?  [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2
                                                F-69
732
                                                DAILY	01
                                                WEEKLY	02
                                                EVERY TWO WEEKS	03
                                                MONTHLY	04      733-34
                                                ANNUALLY	05

                                                OTHER [SPECIFY]:  	  06
                                                DAILY	01
                                                WEEKLY	02
                                                EVERY TWO WEEKS	03
                                                MONTHLY	04      735-36
                                                ANNUALLY	05
                                                OTHER [SPECIFY]: 	         06
B16.  Approximately how often do you receive deliveries to your tank(s)?

                                                FREQUENCY:  	                  737-39

                                                 [CIRCLE ONE]:
                                                 PER WEEK	01
                                                 PER MONTH	02      740-41

                                                 OTHER [SPECIFY]: 	  03

-------
B18.  Are inventory (stick or remote gauge) readings of your tank(s) taken immediately after
      receiving a fuel delivery?  [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2
819.  la the quantity delivered to each tank recorded in a log or journal or other
      permanent record such as a daily inventory report?  [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2       /44


820.  Do you reconcile your inventory (stick or remote gauge) readings with your book inventory
      (meter readings and deliveries)?

                                                YES [GO ON TO 821]	1        ,
                                                NO [SKIP TO 822]	2


821.  How often do you reconcile your tank inventory (stick or remote gauge) readings with your
      book inventory (meter readings and deliveries)?  [CIRCLE ONLY ONE CODE]

                                                DAILY	01
                                                WEEKLY	02
                                                EVERY TWO WEEKS	03
                                                MONTHLY	04      /46-47
                                                ANNUALLY	05
                                                OTHER [SPECIFY] 	  06


822.  Do you ever use water-finding paste to check the water level in the bottom of your tank(s)?
      [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO 823]	1
                                                NO [SKIP TO C1]	2


823.  How often do you use water-finding paste to check the water level in the bottom of
      your tank(s)?  [ENTER FREQUENCY AND CIRCLE UNIT CODE]

                                                FREQUENCY: 	          /49-51
                                                [CIRCLE ONE]:
                                                 PER DAY	01
                                                 PER WEEK	02
                                                 PER MONTH	03      /52-53
                                                 PER YEAR	04
                                                 OTHER [SPECIFY]:  	  05
                                          F-70

-------
                                     C.   OPERATING HISTORY
C1.   Have any tanks at  this establishment  ever been replaced?   [CIRCLE  ONLY  ONE  CODE]

                                                YES  [GO ON  TO C2]	1
                                                NO [SKIP TO C4]	2
                                                DON'T KNOW  [SKIP TO  C4]	8
C2.   How many tanks have been replaced?
                                                NUMBER REPLACED:
                                                                                              716
                                                                                              717-19
C3.
Please answer the following questions about  each tank that  has been replaced,  beginning
with the tank that was replaced  most  recently.   [SPACE HAS  BEEN PROVIDED FOR UP  TO  FOUR
TANKS.  IF MORE THAN FOUR TANKS  HAVE  BEEN REPLACED,  WRITE  THE  ANSWERS  FOR THE  ADDITIONAL
TANKS ON A PLAIN SHEET OF PAPER.]

C3a. In what year was the
(first/second/third)
tank replaced?


C3b. Why was the tank
replaced?
[CIRCLE ALL THAT
APPLY FOR EACH TANK]
a. Because it
was leaking?. . . .
b. Because other tanks
were being replaced
at that time? . . .
c. Because it was no
longer needed/in
use7 	
d. To increase storage


e . Or for some other
reason? [SPECIFY]: .




First Tank



(year)
720-23





01


02


03

04


05

(specify)
724-35

Second Tank



(year)
736-39





01


02


03

04


05

(specify)
740-51

Third Tank



(year)
752-55





01


02


03

04


05

(specify)
756-67

Fourth Tank



(year)
768-71





01


02


03

04
i

05
j
(specify)
772-83

                                                F-71

-------
C4.   Have any tanks at this establishment  ever been removed without  being replaced?  [CIRCLE
      ONLY ONE CODE]
                                                YES [GO ON TO C5]	1
                                                NO [SKIP TO C7]	2
                                                DON'T KNOW [SKIP TO C7]	8
                                                                                                    CM)
                                                                                              716
C5.   How many tanks have been removed without being replaced?

                                                NUMBER REMOVED:
                                                                                              /17-19
C6.
Please answer the following questions about each tank that has been removed without
being replaced.   [SPACE HAS BEEN PROVIDED FOR UP TO FOUR TANKS.  IF MORE THAN FOUR
TANKS HAVE BEEN  REMOVED WITHOUT BEING REPLACED,  WRITE THE ANSWERS FOR THE ADDITIONAL
TANKS ON A PLAIN SHEET OF PAPER]

C6a. In what year was the
(first/second/third)
tank removed?

C6b. Why was the tank
removed?
[CIRCLE ALL THAT
APPLY FOR EACH TANK]
a. Because it
was leaking?. . . .
b. Because other tanks
were being removed
at that time? . . .
c. Because it was no
longer needed/in
use7 	
d. Or for some other
reason [SPECIFY]: .

First Tank


(year)
720-23



01

02

03

04
(specify)
724-33
Second Tank


(year)
734-37



01

02

03

04
(specify)
738-47
Third Tank


(year)
748-51



01

02

03

04
(specify)
752-61
Fourth Tank


(year)
762-65



01

02

03

04
(specify)
766-74
C7.   Have any tanks at this establishment been abandoned in place?  ("Abandoned in place" means
      that the tank is no longer in use but has not been removed.)  [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO C8]	1
                                                NO [SKIP TO 01]	2
                                                DON'T KNOW [SKIP TO 01 ]	8
                                                                                                      071
                                                                                              716
                                           F-72

-------
C8.
How many tanks have been abandoned?
                                                NUMBER ABANDONED:
                                                                                              /17-19
C9.   Please answer the following questions about each tank that has been abandoned in place.
      [SPACE HAS BEEN PROVIDED FOR UP TO FOUR TANKS.   IF MORE THAN FOUR TANKS HAVE BEEN
      ABANDONED IN PLACE,  WRITE THE ANSWERS FOR THE ADDITIONAL TANKS ON A PLAIN SHEET OF
      PAPER]

C9a. In what year was the
(first/second/third)
tank abandoned?


C9b. Why was the tank
abandoned?
[CIRCLE ALL THAT
APPLY FOR EACH TANK]
a. Because it
was leaking ....
b. Because it was no
longer needed/in

c. Or for some other
reason [SPECIFY]: .


C9c. How was the tank
abandoned?
[DESCRIBE PROCEDURE, OR
CIRCLE ALL THAT APPLY]
a. Tank was drained. .
b. Tank was washed . .
c. Tank was cut open .
d. Tank was sand
filled 	
e. Tank was cement
filled 	
f. Other [SPECIFY]:




First Tank



(year)
720-23





01


02

03
(specify)
724-31




01
02
03

04

05


(specify)
732-43

Second Tank



(year)
744-47





01


02

03
(specify)
748-55




01
02
03

04

05


(specify)
756-67

Third Tank



(year)
768-71





01


02

03
(specify)
772-79




01
02
03

04

05


(specify)
780-91

Fourth Tank



(year)
792-95





01


02

03
(specify)
796-103




01
02
03

04

05


(specify)
7104-115

                                                F-73

-------
                                    D.  PERMITS AND LICENSES
                                                                                                     OH

01 .   Were you required to obtain a special building permit or license in order to have your
      tank(s) installed?  [CIRCLE ONLY ONE CODE]
                                                NO ....................  2
                                                DON'T KNOW ................  8
D2.   Are you required to maintain a special permit or license to store flammable or hazardous
      material at your establishment?  (Often these permits are called Hazardous Use or Hazardous
      .Materials permits, and are issued by the state, county, or local fire marshal.)  [CIRCLE
      ONLY ONE CODE]

                                                YES	1
                                                NO	2
                                                DON'T KNOW	8
                                           F-74

-------
                                        E. INSTALLATION
E1.   What type of fill was used to backfill around and over the tank(s)?  [CIRCLE ONLY ONE CODE]

                                                a. Clean sand (with no large rock)?. ...  01
                                                b. Pearock or pea gravel?	02
                                                c. Soil from the excavation?	03      /18-19

                                                d. Or some other kind of fill [SPECIFY]: .  04
E2.   (Is the tank/are any of the tanks) installed with the bottom resting on or in a concrete
      or packed earth pad?  [CIRCLE ONE CODE FOR EACH ITEM]
                                                                                     Yes   No
                                                a. A concrete pad or cradle? ....  1     2       /20
                                                b. A packed earth pad?	1     2       /21


E3.   Are any of the tanks strapped to a concrete pad?  [CIRCLE ONLY ONE CODE]

                                                YES	1       /22
                                                NO	2
                                                DON'T KNOW	8
E4.   What is the shortest distance between any of. your tanks and any neighboring underground
      tank or other solid underground structure (such as a basement wall, sewer, or utility
      vault)?  [ENTER DISTANCE AND CIRCLE UNIT CODE]

                                                SHORTEST DISTANCE FROM
                                                  UNDERGROUND STRUCTURE:  	          /23-28

                                                [CIRCLE ONE]:
                                                 INCHES	01
                                                 FEET	02      /29-30

                                                 OTHER [SPECIFY]:                           03
                                               F-75

-------
                                         F.   PROTECTION

                                                                                                    OH
F1.   Has any type of special equipment or materials been installed to prevent  external
      corrosion of the tank(s)?  [CIRCLE ONLY ONE CODE]

                                                YES [SPECIFY AND GO ON TO F2]:  	          /16

                                                  	  1
                                                NO  [SKIP TO F3]	2        /17-18
                                                DON'T KNOW [SKIP TO F3].  .	8
F2.   How often do you inspect your external  corrosion protection system?   [ENTER
      FREQUENCY AND CIRCLE UNIT CODE]
       IF YOU NEVER INSPECT THE EXTERNAL CORROSION PROTECTION SYSTEM,  CHECK HERE
       AND SKIP TO F3.
/19
                                                FREQUENCY OF INSPECTION:   	          /20-22

                                                [CIRCLE ONE]:
                                                 PER DAY	01
                                                 PER WEEK	02
                                                 PER MONTH	03       723-24
                                                 PER YEAR	04

                                                 OTHER [SPECIFY]:            	  05
F3.   Since you began using the tank(s), have you ever had the tank(s) completely drained and
      cleaned out?  [CIRCLE ONLY ONE CODE]
F4.   Does the tank system have a continuous electronic monitoring system to detect tank leakage?
      [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO F5]	1        .
                                                NO  [SKIP TO F6]	2
                                         F-76

-------
F5.   How often is the electronic monitoring system inspected for maintenance?  [CIRCLE ONLY ONE
      CODE]
                                                a. Annually?	01
                                                b. Twice a year?	02
                                                c. Three or four times a year?	03      727-28
                                                d. Or at some other interval? [SPECIFY]:
                                                                                            04
F6.   Have pressure piping (or line) leak detectors been installed at this establishment to
      detect leaks in the piping (lines)?  [CIRCLE ONLY ONE CODE]

                                                YES [GO ON TO F7]	1
                                                NO [SKIP TO G1]	2       729
                                                DON'T KNOW [SKIP TO G1]	8
F7.   How frequently are the pressure piping leak detectors tested to make sure they are operating
      correctly?
           IF THE PRESSURE PIPING LEAK DETECTORS ARE NEVER TESTED, CHECK HERE |	|
           AND SKIP TO QUESTION G1.
730
                                                FREQUENCY: 	          /31-33

                                                [CIRCLE ONE]:

                                                  PER DAY	01
                                                  PER WEEK	02
                                                  PER MONTH	03      /34-35
                                                  PER YEAR	04
                                                  OTHER [SPECIFY]: 	  05


F8.   Have the pressure piping leak detectors ever given false leak signals?  [CIRCLE ONLY
      ONE CODE]

                                                YES	1
                                                NO	2       736
                                                DON'T KNOW	8
F9.   Have the pressure piping leak detectors ever detected actual leaks in the piping system?
      [CIRCLE ONLY ONE CODE]

                                                YES	1
                                                NO	2       737
                                                DON'T KNOW	8
                                                F-77

-------
                                     G.  INFORMATION NEEDS
G1.   Have any of the companies from whom you receive your fuel products asked you to keep inven-
      tory records (dipstick readings,  meter readings and delivery records) for your tank(s)?
      [CIRCLE ONLY ONE CODE]
                                                YES.
                                                NO .
                                          716
G2.   Has anyone ever given you training or explanatory literature about any of the following
      topics?  [CIRCLE ONE CODE FOR EACH ITEM.  IF YOU HAVE RECEIVED INFORMATION OR TRAINING,
      PLEASE INDICATE FROM WHOM]
                       Type of Training
        a.  Keeping inventory records.
        b.  Doing inventory reconciliation
            calculations	
        c.  Measuring the quantity of product
            in a tank using a dipstick and
            conversion chart	
        d.  Checking pump meter accuracy.
        e.  Line leak detection and testing.
        f.  Tank or line leak prevention.
        g.  Tank tightness testing methods.
        h.  Leak monitoring methods (such as
            observation wells)	
Did you
receive?
YES   NO

 1     2
 1     2
 1     2
 1     2
 1     2
If "Yes," from whom?
                            /17-19
                                          /20-22
                                          723-25
                            726-28
                            729-31
                            732-34
                            735-37
                                          738-40
                                          F-78

-------
G3.   If you found out that (your tank/one of your tanks) was leaking, would you probably:
      [CIRCLE ONLY ONE CODE]
           a.   Replace it with another tank	01
           b.   Line it and continue to use it	02      /41-42
           c.   Abandon it in place	03
           d.   Or something else [SPECIFY]:  	  04


G4.   How much do you expect it would cost you to:
                                                a.   Replace a tank?	$	  /43-4B
                                                b.   Line a tank?	$	  /49-S4
                                                c.   Abandon a tank in place? ....    $	  /55-60

G5.   Do you have an insurance policy that covers you against damage to people or property
      caused by sudden spills of motor fuel?  [CIRCLE ONLY ONE CODE]
                                                *»	1       /
                                                NO	2       /61

G6.   Do you have an insurance policy that covers you against damage to people or property
      resulting from non-sudden spills (including leaks) of motor fuel?  [CIRCLE ONLY ONE CODE]
                                                YES	1
                                                NO	2       '"
                                               F-79

-------
                 TANK  TO  DISPENSER METER FUEL uim  CONNECTIONS

                 Instructions:   Mark  (X)  in  each  block  for which  there
                 is  fuel  line  (pipe)  connection  from  the tank to  the
                 dispenser  meter.   (If  more  tanks than  spaces, use
                 additional sheets.)
Disp.
Meter
Number
M-l
M-2
M-3
M-4
M-5
M-5
M-7
M-3
M-9
M-l-
M-ll
M-l 2
M-13
M-14
M-15
M-l 6
M-17
M-i a
M-19
M-20
Tank Mumoer and Product
T-l





















7-2 1 T-3 1 7-4 ! T-5 | T-6 1 T-7 i T-3




























































1

i










i
i









!





i








j
i
!
i


i
1

















1







i













1
Does che facility have a  leak monitoring  system  (for tanks oc piping)  that  is  not
electronic (such as ocservation wells)?
            YES 	
            NO 	

            If YES, describe
                                     F-80

-------
                     Site Observations Recording Sheet
Site Code  Label
Date

Size of fill pipe
(I.D. )
Drop Tube
(permanent or
removable)
Tank 1


Tank 2


Tank 3


Tank 4


Tank 5


Tank 6


Site Code  Label
Date

Size of fill pipe
(I.D. )
Drop Tube
(permanent or
removable)
Tank 1


Tank 2


Tank 3


Tank 4


Tank 5


Tank 6


                                    F-81

-------
                                                                     OMB Ho.:  2070-0037

                                                                     Expires:  December 31, 1985
UNITED STATES  ENVIRONMENTAL  PROTECTION AGENCY
                   UNDERGROUND STORAGE  TANK  SURVEY
            I
                                         I
  HAILING ADDRESS; Verified? ... |	|
LABEL VERIFICATION

          LOCATION ADDRESS:  Verified' ...
               (ESTABLISHMENT NAME)
                   (ADDRESS)
                                                             (ESTABLISHMENT NAME)
                                                                 (ADDRESS)
                 (CITY/STATE/ZIP)

  CONTACT NAME AND PHONE;  Verified? ... |~

  Contact Name: 	    	     	
                                                                (CITY/STATE/ZIP)
               Contact Phone:
 A.  Questionnaire Status;
 B.  Inventory Status (CIRCLE ONE)

    1 = Started
    2 = Not Started
    3 = Obtained
    4 - Refused
    5 = Other (SPECIFY) 	
 C.  Can Teat (CIRCLE ONE)

    1 = No Meters
    2 = Complete
    3 = Partial Complete
    4 * Refused
    5 = Other (SPECIFY)
          D.  Mapping (CIRCLE ONE)

             1 = Complete
             2 = Other (SPECIFY)
          E.  Debriefing (CIRCLE ONE)

             1 = Complete
             2 = Other (SPECIFY) 	
          F.  Confidentiality

             1  = Form Enclosed
             2  = Waived
             3  = Other (SPECIFY)
                                         Conducted by:
                            WESTAT
                            An Employee-Owned Research Corporation
                                               • MD POQfiO • 3CM251-1SOO
                                          F-83

-------
[IN YOUR TELEPHONE CALL TO SET UP THE APPOINTMENT,  ASK R IF HE/SHE HAS COMPLETED THE QUESTIONNAIRE FORM AND BEGUN THE
INVENTORY RECORDING.  IF NOT, ENCOURAGE THE R TO DO SO.]
INTRODUCTION!

1.  Hello, my name is (YOUR NAME), from West at.
    you about your underground storage tank(a).
    Midweat Research Inatitute.  (He/She) will be drawing a map of the tank area(a)  and taking aome pictures of the
    surface area(s) over the tank(a).
[SHOW IDENTIFICATION CARD].   I'm here to conduct  the interview with
The other member of my team is                                from
    [IF YOU HAVE ANY OBSERVERS ON THIS INTERVIEW,  INTRODUCE YOUR OBSERVERS.   OTHERWISE,  ASK IF THERE IS A PLACE  WHERE
    YOU AND THE RESPONDENT CAN SIT DOWN AND GO THROUGH THE QUESTIONNAIRE.  A SIDE-BY-SIDE SEATING ARRANGEMENT IS
    PREFERABLE, SINCE THIS ALLOWS YOU TO READ FROM THE RESPONDENT'S WORKING  COPY OF THE  QUESTIONNAIRE.   IF THE K DOES
    NOT HAVE HIS SURVEY MATERIALS IN SIGHT, SUGGEST THAT HE OBTAIN THEM — THAT THE INTERVIEW WILL BE MORE EFFICIENT,
    AND THAT YOU WILL HAVE TO RECORD CERTAIN INFORMATION ON HIS INVENTORY FORMS LATER ON.]


CONFIDENTIALITY!

2.  As was mentioned in the letter and in the General Instructions, you can  (claim/ask for) confidentiality for  all
    or part of your responses to the questionnaire.  The way you do this is  by filling out  the form that la in the
    General Instructions booklet.  Have you decided to claim confidentiality for any of  your answers7

                                       YES	 1 [OBTAIN COMPLETED CONFIDENTIALITY FORM FROM
                                                     RESPONDENT.  PUT ID STICKER ON TOP  OF  FORM.]
                                       NO	 2


[BEGIN INTERVIEW, READING QUESTIONS, ITEM-BY-ITEM.  READ THROUGH THE TANK DESCRIPTION SHEET ONLY FOR THE FIRST TANK.
FOR THE SECOND AND FOLLOWING TANKS, READ THE QUESTION NUMBERS AND/OR ABBREVIATED QUESTIONS.]
INVENTORY RECORDS:
3.  Next, we need to review your inventory records.  The General Instructions booklet discussed keeping inventory
    records for each of your tanks.  Have you started to keep these records yet7

                                       YES	 1  (GO TO 4)
                                       NO	 2  (GO TO 5)


4.  (IF YES):  That's great!  Here is a postage paid  envelope in which to send the completed inventories to Westat.
    A Westat interviewer will be calling you in a few weeks to check with you on any problems you might be having
    with the inventoriea.  While I am here,  I need to review your inventory sheets and initial them.  [IF NECESSARY,
    PROBE:  If I could take a look at them now, I would appreciate it.]  [IF THE TANK(S) DISPENSER(S) ARE METERED]:
    I will also need to record the results of my metering can tests of the dispenser meters on the inventory sheets
    for each tank.  (GO TO 8)
5.  (IF NO):  la there any reason why you have not started the inventories7

                                       YES	 1   SPECIFY:
                                       NO.
                                                                                                          (GO TO 6)
                                                   2  (GO TO 6)
6.  Will you be able to start the inventories today7
                                       YES	 1  (GO TO 7)
                                       NO	2  [PROBE FOR WHEN THEY WILL BEGIN:
                                                                                      ] (GO TO 7)
                                       REFUSED	7  (TERMINATE)
                                       IMPOSSIBLE. 9  [SPECIFY WHY:
                                                                                      ] (TERMINATE)
7.  Here la a postage paid envelope in which to aend the completed inventoriea to Westat.


8.  When we receive your inventory(ies) they will be computerized and run through a computer program that checks for
    gains and loaaes that can't be accounted for, auch as over- and under-deliveriea, theft or pilfering, or leakage.
    We will let you know the results of that computer analysis.
                                                               F-84

-------
DISPENSER METER ACCURACY CHECKS:
We have  found  in the past that a major problem  in analyzing  inventory records  is  that some dispenser meter  readings
are just  alightly  inaccurate.  Often these meter errors  show up  in the computer analysis as small  leaks.  For that
reason,  we  are checking out the accuracy of all dispenser meters, using a 5-gellon meter teating can.

Our accuracy checking procedure is  the same procedure  that  is  used by the agencies that certify meter  accuracy.  We
will not  be adjuating your meters if we find that they are misreading.  What we will do is record  the  amount of  pro-
duct pumped intgo  the can according to the dispenser meter,  and  record the amount in the metering  can  according  to
the gauge on the can.  I'll need to record this on your  inventory sheets as well  as my copy of the questionnaire.
The information will be fed into the computer program  to correct for metering  error in the results.

We will be  pumping five gallons of product into the test can from each dispenser  that has its own  meter.  We will
then be pouring the five gallona of product back into  the tank from which it was  pimped.  IF K HAS BEGUN  INVENTORY:
I will need to record the returned product as a "delivery" to the tank on your inventory sheet.  FOR THE  FIRST
MEASUREMENT:   First, I need to wet the inside of the can with  about a gallon of product, and pour  it back into its
tank.

MAKE SURE YOU  WILL BE ABLE TO RETURN THE PRODUCT TO THE  TANK BEFORE YOU BEGIN  PUMPING.  DO ALL THE METERS FOR A
TANK BEFORE MOVING ON TO THE NEXT PRODUCT TYPE.  AFTER ALL OF THE TANKS ARE DONE, WASH THE CAN OUT WITH DETERGENT
AND WATER,  AND DRY IT AS COMPLETELY AS POSSIBLE.
DEBRIEFING:

To be completed immediately after leaving the site.

D1.  Did £ have the questionnaire completed?

                                       YES	 1
                                       NO	 2


D2.  Did j^ have the inventory sheets started?
                                       YES	 1 (GO  TO D3)
                                       NO	 2 (GO  TO 04)
D3. Did fl have errors or problems in the completed parts of the  inventory7
                                       YES	 1   (DESCRIBE;
                                       NO	 2
04.  Did £ understand inventory process?
                                       YES	 1
                                       NO	 2
D5.  Did J^ understand most/all of the question^ in the questionnaire7

                                       YES	 1
                                       NO	 2
06.  Was H:                                       YES   N0_

     a.  cooperative?	   1     2
     b.  hostile?	   1     2
     c.  guessing a lot?	   1     2
     d.  Other (SPECIFY) 	

                                                   1     2
07.  Was it necessary to talk to more than one R to obtain all required information?

                                       YES	 1
                                       NO	 2
D8.  Comments:
                                                               F-85

-------
                                                                     TIME BEGAN:

                                                                     TIME ENDED:
                                                    A.M.
                                                    P.M.
                                                    A.M.
                                                    P.M.
                                RECORD OF CALLS
ATTEMPT
NUMBER




















DAY




















TIME
am
pm
am
pm
am
pm
am
pm
am
pm
am
pm
em
pm
am
pm
am
pm
am
pm
am
pm
am
pm
am
pm
am
pm
am
pm
am
pm
am
pm
em
pm
am
pm
em
pm
RESULT
CODE




















COMMENTS/PROBLEMS




















                                 RESULT  CODES
PRELIMINARY RESULT CODES

1  APPOINTMENT
2  RESPONDENT NOT AVAILABLE
3  RESPONDENT NOT LOCATED
4  RESPONDENT ILL
5  REFUSAL/BREAKOFF
6  RESPONDENT BROKE APPOINTMENT
7  LANGUAGE PROBLEM
8  OTHER
9  OUT-OF-SCOPE
FINAL - HAIL TO SUPERVISOR

11  COMPLETE
12  PARTIAL COMPLETE
13  ESTABLISHMENT CANNOT BE LOCATED
14  RESPONDENT UNAVAILABLE
15  REFUSAL/BREAKOFF
16  RESPONDENT AVOIDING INTERVIEW
17  LANGUAGE PROBLEM:   NO INTERPRETER
18  NO TANKS
19  OUT OF PSU (MOVED)
20  OUT OF BUSINESS (CLOSED)
99  OTHER (SPECIFY) 	
                                   F-86

-------
                    INSTRUCTIONS
                 PREPARATIONS FOR TANK TESTING

1.    If you are not responsible for making the following testing arrange-
     ments, please notify those who are  as soon as  possible.   Please notify
     other persons who may be involved,  including the tank owner and those
     at your firm's or regional offices.

2.    Immediately contact your fuel supplier or distributor to make
     arrangements for filling your tanks.   Explain  any tank filling
     problems to the test coordinator from Midwest  Research Institute (MRI)
     when he calls.

3.    Fill any business vehicles before the fuel drop  off.   As necessary,
     make arrangements for alternate sources of fuel  for those vehicles on
     the test day.

4.    Fuel delivery must be finished before 8:00 a.m.  of the test day.
     If the test crew has to wait for fuel drop off,  it means that  testing
     will not be finished until later that evening.
5.    Co^I«ts«2^r;aHJJ.::;<«»«Jli 3au& until the fuel  level  comes up into neck of
     the fill pipe.   Use your tank dipsticks to  determine when the tanks
     are "full":  the fuel depth,  as measured by  the  dipstick,  should equal
     the tank diameter.   (In many tanks,  you can see when the fuel reaches
     the fill pipe neck.  However,  for tanks with drop tubes,  you must use
     the 'dipstick to know when it is full . )  Testing cannot be done if the
     tanks are not completely full.

6.    Once filled, the tanks cannot be used until testing is complete.
     Make arrangements to keep the tanks out of  service.   Your business
     does not need to be closed during this time,  but the tanks must remain
     inactive.
                           FINAL CHECKLIST

           D   Notify responsible  individuals.
                D   Owner
                D   Main or regional  office
                D   others

           D   Contact supplier or  distributor

           D   Fill  business vehicles before  filling tanks

           D   Fill  tanks before 8:00 a.m. on test day

           D   Fill  tanks completely

           D   Arrange to keep  tanks  out  of service



                                  F-87

-------
                     ENVIRONMENTAL  CONDITIONS  DATA SHEET
Sit(
Dati
Tesl
Tesi


Time

2 Code Label
a
t Firm
t Team


Barometric
Pressure







Surface


Tar



Temperature
Ambient


k No.



°F
Subsurface







Comments

Climatic conditions.
                                        F-88

-------
                          TEMPERATURE PROFILE DATA
Site Code Label
Test Team        	       	   Date
Test Crew                                            Tank No.
START
TIME
END
TIME
                                  Figure  3
                                       F-89

-------
                   Site Diagram and Detail  Diagram Sheet
Site Code Label
Map
Test Firm
Tesl> Team	              Date
                         Sketch Area and Dimensions
                                    F-90

-------
                            PICTURE  DESCRIPTION



Site Code Label



Team 	         Date





Picture No.                              Description



     1





     2
     4






     5






     6






     7






     8






     9






    10






    11






    12






    13






    14






    15
                                     F-91

-------
                        01
                     c  re
                     ...  01
                     •— .c
                     t_  w
                     Ol. 01
                     3  >

                     1°
                     —  01

                     10 i—
                     O  I-

                     01  '
                     Ol  C
                     u  re
                     re t—



                     lii
                     ISI  >|
                       Ol
                       •g
                     -*  re
                  j: c  •-.
                  -j re o
                  ai-
                  01    E
4J

 Ol

 O)
   §••
00


 (O
 l/l

 O)
-P

 ro

 O)
 (O

 O
          re
          o
 >    o>  c
 u E  s.  o
 01 01  = —


->, £  =i

o "^  ot ^'1




   «-i  "I
                  ago
                  o u  E!
                  I. 01  Ci
                  o c. ad
                  o

                     —  Ol

                  O C  OJ

                  b, til —-'
                  o *
                     U  Ol
                  n, ._  -J

                  M -> — I
                  —
                     01
                     Ol 01
                     =  SJ
   o

   o

   •-1? Zj
   en      I
                  o
                     —  01
                  si i—  a
                  *j ••— .—
                   01    -*l

                   M ^  C

                  — o  re I
                O  i-
                   O O)*J
                b  irt C  U
                01  C —  3


                E  a. s.  o
                3  >/> O  U
               z — o a.

                  o u
    •g
     o
                     ou

                     01 -O
                     a. o

-------
Site Code Label



Team                                              Date
                              EDIT CHECKLIST





          n    Site code label on all  pages.









          D    Be sure all  maps are numbered sequentially.









          n    Photographs  of critical parameters.









          c    Site code labels on photographs and filed in notebook.









          a    Check to see that all  data sheets are filled our correctly.
                                     F-93

-------
                             SIMULATED LEAK TEST
                                 VOLUME DATA
     Site Code Label

     Date
     Tank No.

     Test Team

     Test Firm
                                  Volume,  g
       Time
Bottle No.
Final  Wt.
Tare Wt.
Total Wt.
Spec-! fie Gravity.
Temperature:
                                       F-94

-------
                                 SIMULATED  LEAK  DATA  FORM
                             LEAKING UNDERGROUND STORAGE TANK


Site Code Label

Test Crew	  Date  	 Tank No. 	

Test Team	

             Rotameter                   Start     End      Elapsed    Measured
Rotameter     Setting     Nominal  Rate    Time      Time     Time       Volume     Calculated
   No.          mm            gph         Clock     Hour      gms        gph          Rate
                                             F-95

-------
ol SuM>lf« Own*, or D*tl*r
]B 15. TANK TO TEST
£J Identity by (xmlion
-1
^" B>»n4 IrxJ Gtt«J*
16. CAPACITY Fr?
By moil »ccuT*t* r i
Gllloni ClDoni I 1
11 ttwr« tfoubl ** to T(u> C«p*crtv ' j^| [~|
SM Section 'DETERMINING TANK CAPACITY" j ]
17. FILL-UP FOR TEST Stick Raadlnga
to H in
Stick Watar Bottom
baton FIB-up " tnvantnrv
to* In.
FW Up. STICK BEFORE AND AFTER EACH COMPA
Oaltona
RTMENT DROP OR EACH METEflED DELIVERY QUANTITY
Stitioo Ch.rt
Tink Minuficlurtr i Chirt
Compuny Engin««ring D»1«
DlhM

Total Gallons
GaNona aa Raadlng


                                                    Product In tuft lank (up to flit pip*)
STOUOHTON. MA. 02072-098*
(617) 344-1400
1
n.
M
• 18. SPECIAL CONDITIONS AND PROCEDURES TO TEST THIS TANK VAPOR RECOVERY SYSTEM
^1 SM manual Mcliont •pp)*»W« ChMk telow and rwwrd proc*dor« in log (26) p 1 S(-g< ,
^1 n Witor hi bvth Q3 High iratar tabto In tank •zcavttlon Q Un«<«) twlng iMMd wtth LVLLT Q St«g« H
W 19. TANK MEASUREMENTS FOR
TSTT ASSEMBLY
Bottom at tank tn Qr^to*
Add 30" for 4" L .
Add 24" tor 3" L or »hr *M| , .


Total tubinf to Mumbl. Appro. InuU
20. EXTENSION HOSE SETTING
ExMnd hoM on tuction tub* 8~ or mor«
tolow 1*nk too '
*» Fill DID* »t*nd* •bow gr*d«, utt top of Ml
26
l» OF TEST FTOCEDUMS
27.
m
MM
114 Ir)





































za.
Ricorit dttuli of tttting up
•nd running titt (Ulifull
lififlth tl IIM il nitrfid )





































29.
lu*q
*.





































21 TEMPERATURE/VOLUME FACTOR {•} 1
Is Today Warmer' 1 J CoWef ' 1 J 	 * F Prodod m
22. TMrm*l-S«n«or rMdtng *rt*r dreuMton
23. otglti p»r T In range of •xpoctod crwngi
24.
25.
X
0 TEST THIS TANK
Tank 	 ' F FiH-up Product on Truck 	 __• f Expected Crtinoe | + Of - )
dVll
*•*
NtlWMt

total quantity In coafllclarrt of aipantlon for voluma changa In thla tank
tui tank (It or 17) Invotvad product par *F
* _ TKI. 1.
volume Chang* p*r *F (24) Digits p*r T In Met Votumt chtngi p*r dkgIL t«it
Rang* (23) Computa to 4 daclmal plactt lactor (a)
30. HYDROSTATIC
PRESSUflE
COKTROl
tlin^piM Ltvtl
M l*ck*t
tl
H*itf.nf





































Unit*
Khitk
AitlH^





































* mi« ituiiKatim m
«co« n m\ UL
32 f,*«.
feWtlt*
l*f*>*
R.H..I





































All*>
B.,.-,





































Pr^Kt
IbptNtl (-)
frrtirt
M»MrU(+)





































11
"" mm unn cwmsimi
R{ FUTNIil
35.
ThtrMl
UMW
hWto|





































36
CK^H
H.itaf*
lt*N -
10





































37.
C*»Mtlt>W
W - (•! -
ElfMMH *
CvMrecliM ~





































38. WET VOUIMf
CHtNGtl
EACH READING
T«p*irit*r«
AdfllNMHt
bpmnon(»}w
ConltKlion ( - f
#Mffl-*37{T)





































39
ACCUMULATID
CHANGt
M H|k LtM( racwl
IiWMMlKttH
tlluli il i«|ili
CMftp.***
mfPltrtarlt)





































                                            F-96

-------
Data Chart for Tank System Tightness Test
               petro Tite
PLEASE PRINT
1. OWNER Property [J
Tank(s) Q]
2. OPERATOR
3. REASON FOR
TEST
(Explain Fully)
4. WHO REQUESTED
TEST AND WHEN
5. WHO IS PAYING
FOR THIS TEST?
6. TANK(S) INVOLVED
7. INSTALLATION
DATA
8. UNDERGROUND
WATER
9. FILL-UP
ARRANGEMENTS
10. CONTRACTOR.
MECHANICS.
any other contractor
involved
11. OTHER
INFORMATION
OR REMARKS
12. TEST RESULTS



13. CERTIFICATION
Date
Serial No of Thermal
Sensor
TANK TESTER

Name Address Representative Telephone
Name Address Representative Telephone

Name Address
Telephone




Name Title Company or Affiliation Date
Address
Telephone


Billing Address City
Title Telephone
State Zip
Attention of Order No Other Instructions
Identify by Direction Capacity Brand/Supplier Grade



Location Cover Fills Vents
North inside driveway, Concteie, Black Top, Size, Titefill make. Drop

Depth tn thp WatPf tahlp "

Tank* to h« fillpH hr Dat« Arranged hy
Approx. Age Steel/Fiberglass



Siphones Pumps
Suction, Remote,
Which tanks7 Make it known
Is the water over the tank ?
D ^s D No

Name Telephone
Extra product to "top off" and run TSTT. How and who to provide ? Consider NO Lead.
Terminal or other contact
for notice or inquiry

Company Name Telephone







Additional information on any items above. Officials or others to be advised when testing is in progress or completed. Visitors or observers present
during test etc
Tests were made on the above tank systems in accordance with test procedures prescribe!
as detailed on attached test charts with results as follows:
Tank Identification Tight Leakage Indicated





i for petro Tit**
Date Tested





This is to certify that these tank systems were tested on the date{s) shown. Those indicated as "Tight" meet the criteria established by
the National Fire Protection Asoclatlon Pamphlet 329.
Testing Contractor or Company. By Signature
Technicians _ _ Address

-------
r-tfoublo[tyyChocM—\
1 COMPANYM^VlNCORPORATEO [
46I4 MOSKCT KANSAS CITY. Ma 64 1 30
PHOMC (0I6) 931-5032
11 IDENTIFY
EACH LINE
AS TESTED
































	 -


LIQUID VOLUMETRIC LINE LEAK TEST
1 LOCATION 	 	 	 — 	
DATE OF TEST
JOB #

2 OWNER 	 , 	 	
3 OPERATOR 	 	 —
DS""°n' T«l*phon« No
4 REASON FOR TEST 	


g TFST neoi,iF$TFn RY
• SPEC

"•'"• """"" UIJ«' N« Billing Addr.,i ' ' 	
At INSTBIJCTIONS
ANir.(S| NAMF
B IS A K-M TANK TEST TO BE | 	 | YES » MAKE AND TV
MADE WITH THIS LINE TEST? i — i PUMP OR DIS
|_J NO
12 TIME
(MILITARY)




































13 LOG OF TEST PROCEDURES,
AMBIENT TEMPERATURE.
WEATHER, ETC




































14 PRESSURE





































ORkPi




































PE OF
•ENSFRS
COVER APPROXIMATE
IB VOLUME
BURETTE




































READING




































NET
CHANGE




































16 TEST RESULTS
CONCLUSIONS. REPAIRS AND COMMENTS


©Copyright
1976 Kent-Moore Corporation
F-98

-------
11 IDENTIFY
EACH LINE
AS TESTED





















12 TIME
(MILITARY)






















WEATHER, ETC





















14 PRESSURE
BEFORE





















AFTER





















IB VOLUME
BURETTE
BEFORE





















READING
AFTER






















NET
CHANGE





















18 TEST RESULTS
CONCLUSIONS REPAIRS AND COMMENTS






































































































































































































































































































































































































































































































































































































































^


















































































































N











































































































































































































































































5^?\ Kent-Moore
yJJ TOOL DIVISION



















-































-




































-




































-



















































































































































































































































































































































—




































"
























































































































































































































































































































































































































































))























































































































































1501 SOUTH JACKSON STREET • JACKSON. MICHIGAN 49Z03
T»»fOOO« 517/784-8561
Fofm No U76-145 (J-2S515-88)
                                                                  F-99
                                                                                                                              Litho in U.S A

-------
                          APPENDIX G

             NATIONAL UNDERGROUND  STORAGE  TANK  SURVEY
                     NATIONAL SAMPLE  OF  FARMS
I.   INTRODUCTION AND SUMMARY

     The survey of underground motor fuel storage tanks is
designed to provide national estimates of the number of
underground motor fuel storage tanks at the end use point and the
number and percent of these tanks which leak.  The survey design
defined three segments of the overall target universe of
establishments with underground motor fuel storage tanks:
     o    Fuel establishments (gas stations and establishments in
          other fuel-related or fuel-using industries) which by
          the nature of their business are likely to have such
          tanks;
     o    Large establishments (20 or more employees) which by
          virtue of their size may have an underground motor fuel
          storage tank; and
     o    Farms, of which over half have motor fuel storage
          capacity, but an unknown proportion store motor fuel
          underground.

     The sample design for the survey is a two-stage cluster
design.  The first stage is survey locations, called Primary
Sampling Units (PSUs) and consisting of counties or groups of
counties.  The contiguous United States was divided into six
survey regions, based on rough similarity of soil type and
condition, as defined in Figure G-l.  Thirty-four PSUs were
drawn, six from each region, except four PSUs were drawn from
Region 5.
                              G-l

-------
Figure G-l.
Six regions for National Survey of Underground
 Storage Tanks
  1  —  Northeast

       Maine
       New Hampshire
       Vermont
       Connecticut
       Massachusetts
       Rhode  Island
       New York
       New Jersey
       Pensylvania
       Maryland
       Delaware
       Virginia
       West Virginia
       Washington, D. C,

  2  —  Southeast

       Kentucky
       Tennessee
       Arkansas
       Louisiana
       Mississippi
       Alabama
       Georgia
       North  Carolina
       South  Carolina
       Florida
                   3 -- Midwest

                        Wisconsin
                        Minnesota
                        Iowa
                        Missouri
                        Illinois
                        Indiana
                        Ohio
                        Michigan

                   4 — Central

                        North Dakota
                        South Dakota
                        Nebraska
                        Kansas
                        Oklahoma
                        Texas

                   5 -- Mountain

                        Montana
                        Wyoming
                        Idaho
                        Nevada
                        Utah
                        Colorado
                        Arizona
                        New Mexico

                   6 -- Pacific

                        Washington
                        Oregon
                        California
                                G-2

-------
     Among the three survey segments, fuel establishments and
large establishments are both concentrated in the same areas,
where the population is.  Drawing a sample of PSUs which is
optimal for both of these segments is therefore no problem,
because they occur together.  Farms, however, tend to be found in
the opposite places, those with sparse population.  So optimizing
the design for farms is in direct opposition to optimizing the
design for fuel establishments and large establishments.  Since
the fuel establishments are the major focus of the survey,
accounting for about 800 of the approximately 920 expected
establishments with underground motor fuel storage tanks, the
sample of PSUs was optimized for fuel establishments by being
drawn in proportion to the number of fuel establishments in each
PSU.  As noted above, the resulting sample of PSUs is not optimal
for studying farms.

     The second stage of sampling is the sample of establishments
within the selected PSUs.  Three sample frames (master lists)
were developed for the 34 sampled PSUs — one for fuel
establishments, one for large establishments, and one for farms.
Samples were drawn from each list:

     o    1618 fuel establishments;
     o    600 large establishments; and
     o    600 farms.

These establishments were contacted to determine whether they
were eligible for our survey; that is, whether they had
                              G-3

-------
underground motor fuel storage tanks.  The eligibility rates were
(approximately):

     o    50 percent for fuel establishments;
     o    15 percent for large establishment; and
     o    Less than 5 percent for farms.

     This appendix discusses the national farm sample of 600
farms to be screened.  Subsection II discusses the target
universe of farms and describes the farm sampling frame on a
national basis.  The 1982 Census of Agriculture conducted by the
Census Bureau is taken as the standard count of farms, and a list
developed by the Agricultural Stabilization and Conservation
Service (ASCS) of the U.S.  Department of Agriculture  (USDA)  is
the sample frame used.  For the nation, overall, this frame
offers good coverage of the farm universe.  Subsection III
reviews the survey design with reference to the farm sample and
compares Census figures with ASCS figures for the selected PSUs.
In this subsection, it is seen that the coverage of farms by the
frame is weak in some parts of the country.  Section IV concludes
the appendix with a discussion of the ratio-adjustment weighting
method proposed to minimize total sampling error in the farm
estimates.
II.  TARGET UNIVERSE OF FARMS AND SAMPLING FRAME
     A.   Two Farm Data Sources

     Two sources of information on farms were used in designing
and conducting this survey.  One is the 1982 Census of
Agriculture (the most recent) conducted by the Bureau of the
                              G-4

-------
Census.  This source is used as the most reliable source of
national statistics about farms.  The second is the "1983
Deficiency Master File" developed by the Agricultural
Stabilization and Conservation Service (ASCS) of the U.S.
Department of Agriculture (USDA),  which is used as the list, or
sampling frame, for farms.

     The Census of Agriculture is a data collection and
tabulation effort which is as inclusive as possible.  The 1982
Census lists 2,240,976 farms in the U.S.   A farm is defined by
Census as "any place from which $1000 or more of agricultural
products were sold or normally would have been sold during the
Census year."  Tables provide breakdowns of these farms by size
of farm, value of sales, type of crop, etc., both nationally, by
state and by county.  Some of these figures are reviewed later in
this section.

     What the Census of Agriculture does not provide is a list of
farms or farm operators in specific places.  Thus, for an actual
sampling frame we used the USDA/ASCS 1983 Deficiency File.  This
is a list of farms developed by the USDA containing about
1,942,000 listings  (87 percent as many as the Census total).  The
original impetus for the development of the file was to provide a
mechanism for payment distribution for the PIK (Payment-in-Kind)
program for 1983.  In 1983,  the PIK program was so popular that
USDA believes that almost everyone engaged in growing PIK program
crops (which include various cash grains and upland cotton)
applied for it, and hence is listed on the Deficiency File.
Because they saw a chance to have a near-Census of farms on a
data file, USDA made a special effort to also include listings of
farms not eligible for the PIK program.  The basic data were
gathered by the ASCS county agents.
                               G-5

-------
     The official USDA/ASCS statistics indicate that of 2,018,000
farms known to the ASCS, 1,942,000 (or 96 percent)  are listed on
the Deficiency File.  The ASCS definition of a farm is all of the
land farmed under one operation.

     Only about 57 percent of the farms listed on the Deficiency
File (1,116,000 farms)  are farms that are eligible for the PIK
program.  The remaining 43 percent of farms on the list are not
eligible for the PIK program.  Some portion of the ineligible
farms are ineligible because they were not growing PIK program
crops,  others because they did not choose to apply for the PIK
program.  Because of the 96 percent coverage of farms known to
them, ASCS believes the Deficiency File is a very complete list
of farms in the U.S.

     In exploring the universe of farms and comparing the two
data sources, we take the 1982 Census of Agriculture as the
primary source of information on the nation's farms.  Although
the ASCS total is less than the Census total, it is probable that
the ASCS list is not simply a subset of the farms counted by the
Census, but a partially overlapping list.  This is due to the
fact that the two lists are constructed by different
organizations for different purposes, are based on different
information, and have different definitions as for including and
counting specific cases.  However, we can get a summary of the
nation's farms from the Census and a rough idea of the ASCS
coverage of those farms.
     B.   Summary of the Target Universe Based on the 1982 Census
          of Agriculture

     The figures presented here are taken from Vol. 1, Part 51,
U.S. Summary and State Totals of the 1982 Census of Agriculture.
                               G-6

-------
The first table lists total numbers of farms by size and sales
categories.

     It seems likely that farms with small acreage or low sales
volume would be less likely to have underground motor fuel
storage tanks and would also be less likely to be included on the
ASCS file than large farms.  Table G-l indicates that a number of
farms are quite small, with 8 percent of farms reported having
one to nine total acres.  Also, many farms have quite low sales
figures.  Nearly one-quarter of farms reported on had less than
$2,500 in sales in 1982.

     The Census also gives figures for storage of various fuels
(although unfortunately for our survey, no question was asked as
to whether the storage was underground).  Table G-2 summarizes
the storage capacity data for 1982.

     This indicates that roughly half of all farms reported
gasoline or gasohol storage, and about 40 percent reported diesel
storage.  The overlap of the two groups is not given but is
presumably fairly high.  However,  the number of farms with
substantial storage capacity is much less — 2 percent reported
2,000 gallons or more diesel storage capacity, and 1 percent
reported that much gas storage capacity.  Taking 1,000 gallons or
more as a cutoff,  7 percent of farms reported this much gasoline
storage capacity and 8 percent reported this much diesel storage
capacity.

     In conclusion, based on the 1982 Census of Agriculture,
there were about 2.2 million farms, of which 8 percent were
smaller than 10 acres, one-quarter had less than $2,500 in sales
for the year, and perhaps 10 percent have 1,000 gallons or more
fuel storage capacity.  This last assumes a substantial overlap
between storers of gasoline and diesel fuel.  If there is little
                               G-7

-------
 Table G-i,
Farms by acreage and sales
(1982 Census of Agriculture)
  Total U.S.  Farms
                                          2,240,976
By acreage

1-9

10 or more

   10 - 49
   50 - 499
   500 - 1,999
   2,000 or more
                                            187,665

                                          2,053,311
                                  449,252
                                1,238,162
                                  301,320
                                   64,577
By sales

Less than $2,500

$2,500 or more

   $2,500 - $9,999
   $10,000 or more

      $10,000 - $99,999
      $100,000 - $499,999
      $500,000 or more
                        840,583
                        274,580
                         27,800
                                            536,327

                                          1,702,973
                                  560,010
                                1,142,963
(1,676 abnormal farms not reported by sales - institutional,
research and experimental farms,  and Indian reservations.)
                              G-8

-------
 Table G-2.  Fuel storage capacity,  1982*
             (1982 Census of Agriculture)
Farms reporting fuel expenses
Storage capacity reported, farms
1,000's gallons
Farms with storage capacity of:
1 - 499 gallons
500 - 999 gallons
1,000 - 1,999 gallons
2,000 or more gallons
Storage capacity reported as
"no", farms
Storage capacity not reported, farms
Gasoline
and Gasohol
1,123,463
583,853

616,650
352,925
136,455
17,433
451,895
422,083
Diesel
fuel
924,863
648,605

471,646
262,902
140,896
49,419
150,210
245,380
*Includes above-ground tanks and containers,  as well as under-
 ground tanks.
                              G-9

-------
overlap, as many as 15 percent of farms may have 1,000 gallons or
more motor fuel storage capacity.
     C.   Comparison of Census and Sample Frame

     The sampling frame, the ASCS 1983 Deficiency File, is
primarily a data base of farms rather than a source of
statistics.  Hence, we do not have extensive national or state
statistics on this file.  Nationally, we can compare the number
of farms from Census (2,240,976)  and the ASCS file (1,942,437),
showing that the sample frame file has 87 percent as many farms
as the Census.  (Note that these are not necessarily completely a
subset of the Census farms, as mentioned above.)

     We also can compare total cropland acreage between the two
data sources.  The Census shows 445,362,028 acres of total
cropland on 2,010,609 farms with cropland, while ASCS shows
443,850,049 acres of total cropland on its 1,942,437 farms.  The
ASCS definition of cropland is "tillable soil" — the land does
not have to have been planted, only to be suitable for planting.
The Census definition includes three categories:

     o    Harvested cropland;
     o    Cropland use only for pasture or grazing; and
     o    Other cropland.

The two definitions appear to be quite similar.

     The sample frame thus covers 99.7 percent of the total
cropland reported in the Census and has 96.6 percent as many
farms as those reporting cropland in the Census.  It appears that
farms with no cropland is an area of sparse coverage for the ASCS
                              G-10

-------
list.  The major categories of land in farms not included in
total cropland are:

     o    Pasture and rangeland other than cropland and woodland
          pastured (418,264,264 acres);
     o    Woodland (87,088,255 acres); and
     o    Land in house lots, ponds, roads, etc. (36,082,032
          acres).

So farms with pasture, rangeland or woodland and no cropland are
more likely to be in the Census but not the ASCS list.  However,
in the Census 90 percent of farms listed had cropland, so farms
with none are relatively rare.

     Other types of farms which may tend to under-represented by
the ASCS list (based on discussions with Tom Meyer of ASCS) would
be growers of fruits and vegetables.  Most farms grow more than
one crop, and so many fruit or vegetable farms may also have a
PIK-eligible crop or may be listed as an ineligible farm on the
ASCS file.  According to Census data, 69,109 (3.1%) of farms
reported vegetables harvested for sale and 123,663 (5.5%)
reported land in orchards.  On a national basis, these farms do
not represent a major portion of the target universe, although on
a regional basis their proportion varies.  These figures are
presented as a way of assessing the potential for undercoverage,
but we have no direct way of determining the ASCS coverage of
these types of farms.
III.  SAMPLE DESIGN FOR UST SURVEY. FARM SEGMENT

     In this subsection we again review the survey sample design,
emphasizing the aspects relevant to the farm sample.  The design
was a two-stage cluster design.  The contiguous U.S. was divided
                             G-ll

-------
into six survey regions,  as presented in Figure G-l shown
earlier.  The first stage of the sample was survey locations,
known as Primary Sampling Units (PSUs).  These PSUs consisted of
counties or groups of counties and were chosen by region with
probability proportional  to number of fuel establishments.  The
second stage was the within-PSU selection of farms.  Farms were
selected from a sampling  frame based on the ASCS list for the
selected counties with within-PSU probabilities determined so
that the overall probabilities of selection would be equal for
all farms.  We give more  details in the following sections.
     A.   First Stage Sample of Survey Sites (PSUs)

     The first stage in the two-stage sample design was of PSUs,
which were counties or groups of counties.  Within each region,
six PSUs (four in the Mountain Region) were selected with
probability proportional to their number of gas stations and
fuel-related establishments.  As discussed in Subsection I, this
is the optimal design for studying fuel establishments — the
main focus of the survey.

     Table G-3 shows some statistics on number of farms, by
region, based on the 1982 Census of Agriculture.  The first two
columns give the total farms in each region and the corresponding
expected sample size, by region, for an equal probability sample
of 600 farms to be screened for underground motor fuel storage
tanks.  Regions 1, 5 and 6 have expected sample sizes of less
than 100, with Regions 5 and 6 less than 50.  Next, in column 3,
we have used the inverse of the PSU probability of selection as a
PSU weight and weighted the 1982 Census of Agriculture farm
counts for the selected PSUs up to the regional level.  By
comparing these figures with column 1, we see that our sample of
PSUs has considerable variance from the actual totals.  As
                             G-12

-------
 Table G-3.  Farm summary based on 1982 Census  of  Agriculture,
             all farms
Region -1-
1 -Northeast
2-South
3-Midwest
4-Central
5-Mountain
6-Pacific
Continental
U.S. Total
Agriculture
Census count
222,099
548,926
725,699
464,680
121,777
152,630

2,235,811
Expected
farm sample2
60
147
195
125
33
41

601
Weighted count,
sampled PSU's
123,714
283,226
908,358
494,029
147,071
104,164

2,060,562
Expected
farm sample2
36
82
264
144
43
30

599
•'•Regions are defined in Figure G-l-

2These farms are to be screened for the presence of underground motor
 fuel storage tanks.
                                   G-13

-------
mentioned in Subsection II, this is due to the PSU sample
selection being based on the number of fuel establishments, a
measure inversely correlated with the number of farms.

     Finally, column 4 gives the expected sample size based on
the 1982 Agriculture Census counts for our PSUs.  Regions 5 and 6
are still very low, and Regions 1 and 2 have a lower sample size
than expected from the regional totals.
     B.   ASCS List for Selected PSUs

     The actual sample was drawn from a sample frame based on the
ASCS 1983 Deficiency File.  This file was described in Subsection
II above on a national basis.  Here, we compare the ASCS file
counts to the Census counts for our sampled PSUs and present some
relevant Census figures on a regional basis.  The actual sample
frame used was a modification of the ASCS file, which we describe
below, leading to the final sample sizes.

     In Table G-4, the Census of Agriculture counts are compared
with the ASCS file counts for the sampled PSUs on a region-by-
region basis.  The third column shows the percent coverage the
ASCS file had.  For the 76 counties in our 34 PSUs as a group,
the ASCS file had 70 percent as many listings as there were farms
counted in the Census of Agriculture.  On a region-by-region
basis there is quite a bit of variation in this coverage.  The
ASCS list has good to excellent coverage of Regions 2 through 4,
which together contain 70 percent of all farms according to the
Census; and fair to poor coverage of the rest of the country.
For Region 3, the Midwestern region, ASCS actually has more
listings — 118 percent as many as the Census.  For Region 2
(South) and 4 (Central), the ASCS had fairly good coverage — 90
percent and 79 percent as many listings, respectively, as the
                             G-14

-------
Table G-4.   Raw  farm count based on sampled PSUs  (1982 Census
             of Agriculture and 1982 ASCS Defiency File)



Region 1
1 -Northeast
2-South
3-Midwest
4-Central
5-Mountain
6-Pacific
Continental
U.S. Total
Raw counts, sampled PSD1 s
1982
Agriculture
Census
3,743
6,619
13,367
11,025
4,472
10,851

50,077
1983 ASCS
Deficiency
File
1,573
5,969
15,787
8,706
2,305
504

34,844

Percent
Coverage
ASCS File
42%
90%
118X
79%
52%
5%

70%
1-Regions are defined in Figure G-
                               G-15

-------
Census.  For Regions 5 (Mountain)  and 1 (Northeast),  the coverage
was only about half — 52 and 42 percent as many listings,
respectively, in ASCS as the Census count.   Finally,  for Region 6
(Pacific),  the coverage was very low — the ASCS list had only 5
percent as many listings as the Census for this region.

     Several attempts to understand these discrepancies have met
with limited success.  The two data sources rely on different
bases to get their lists of farms and farm operators, employ
different (and to a great extent not thoroughly documented)
definitions of "a farm" and have different basic philosophies of
the importance of complete coverage.  We were able to determine
that our ASCS list is a list with one record per farm, as defined
by the County Agent, so that the comparison in Table G-4 is the
relevant one.

     We expected that vegetable, fruit or livestock farms would
be at greater risk of under-representation on the ASCS list, so
Table G-5 presents the counts of these types of farms by region,
with the percent of all farms in the region, based on the 1982
Census.  A farm may, of course, have crops in more than one
category.  For example, a cattle ranch with pastureland would
likely also grow feed grain and be eligible for the PIK program.
Farms with land in vegetables or orchards might also have PIK-
eligible crops, or be on the ASCS File as ineligible.  The most
striking statistic in Table G-5 is that, while nationally 5.4
percent of farms have land in orchards, in Region 6  (Pacific),
33.7 percent of farms have land in orchards.  It seems quite
probable that this is a contributing factor to the severe
discrepancy between the ASCS frame and the Census in that region.
Region 1 (Northeast) has a higher rage of farms with vegetables
(7% versus 3.1%) than the national average but scarcely enough to
account for listing less than half of all farms in that region.
                              G-16

-------
 (d
m T>
    C
 C  (d
 O iH
    0)
 0)  M
-P 10
rH Id
 o
-H
    C
    id

r- «-
S 2
5 t
0
^— «*^
W
U-
T?
C CO
(O CD
<- S
§ 5
> CD
* s.
CO 4>
£ >
«- c
w **~
u_





4-f
CD
O
c_
m
^f
Q.



l_
CD
.Q
£
z
*^
c
CD

Q-
<_
CD
A

U
z
*
c
o
^>
CD
OH



M M M M M M
oo in r-. co oo r«-
co CD in r^ CD «•




r^ r** ^o w ^o CT^
CO CD Tj- TT IO r-v
co ^r •<*• — r*. co
— in to ro oj —
in in — in co r-.
— to «a- to
»C »« »« M M M
rv — to r>- to r^.
in in —  o r^ in  *— oo oo
in t> CNJ co in ro
^- O> CO Is- OO CO

in o> r> •«- oo i>
4->
11 ~ £.
1* |rls
*->•»-> f •*-' ^ —
C_ 3 -o C 2 °
o o r: CD o w
z to C o 3EI Q.
I i i i i i
*- «>j 10 ^- in CD

M
<«•
CD



^»
O>
P»-
•«
Px
  O
        10  >
                                                                                               CO
                                                                                               49
        S.   *
        c  cn
        <_  _
        o  —
                                                                                               *
                                                                                               •K
                                                                                                         CD
                                                                                                         CD
                                                                                                         O


                                                                                                         CSJ

                                                                                                           m
                                                                                                          CO
                                                                                                          a>
                                                                                                          Q.
                   in
                   to
                   CM

                   C\J
                                                                                                          CO
                                                                                                          CD

                                                                                                          "a.
                                                                                                          CL
                                                                                                          (O
                                                                                                          (O
                                                                                                          o
                   in
                    o»
                    c
                                                                                                           CO

                                                                                                           £
                    o
                    3
                    CO
                                                                                                               CO
     CD
CD   O.
*-   <0
0>   «-
  >  O>

«°   £
 CO   *i
 
            **       C
                                                                                                               tM
                               CD

                               ro

                               i"—

                               X7
                                C
                                
-------
     The basic pattern in Table G-4 is good coverage to over-
coverage in those parts of the country which contain the majority
of all farms (Regions 2, 3, and 4 contain 1,739,305 farms, or 78
percent of the total, see Table G-3),  and fair to poor coverage
in the remainder of the country.  This underlying distribution of
farms, combined with the pattern of over- and under-coverage and
the PSU selection probabilities, results in a fairly decent
national estimate of number of farms based on weighted ASCS data,
even though the regional estimates are poor.  These weighted
figures are shown in Table G-6, along with the expected sample
size based on weighed ASCS file counts.  Regions 1, 5, and 6
continue to lose sample cases due to list undercoverage of those
regions.
     D.   Sampling Frame and Actual Farm Sample

     In order to use the ASCS list as a sampling frame, two
modifications were made.  First, the list of farms was collapsed
into a list of farmers by aggregating records with the same name
and address.  We would thus be able to increase the number of
farms sampled without increasing the costs by sampling 600
operators and interviewing them regarding "any farm land you own
or operate" in the specific counties they were sampled for.  For
those few who reported underground storage tanks, we then
determined which distinct farms have such tanks and how many.
The second frame modification was due to the use of a purchased
list for the large establishment segment of the overall survey.
Any large establishments with agricultural SICs were removed from
the large establishment frame and matched against the ASCS list.
If they did not already appear on it, they were added to the
frame.
                              G-18

-------
Table G-6.   Weighted farm counts from ASCS  1983  File,  expected and
            actual sample sizes
Region *
1-Northeast
2-South
3-Midwest
4-Central
5-Mountain
6-Paciflc
Continental
U.S. Total
Weighted counts, sampled PSU's
1982
Agriculture
Census
123, 714
283,226
908,358
494,029
147,071
104,164
2,060,562
1983 ASCS
Deficiency
File
52,376
301,055
1,105,519
512,376
132,621
5,652
2,109,599
Farm sample
size expected
from ASCS file2
15
86
314
146
38
2
601
iRegions are defined in Figure G-l

2The'.e farms are to be screened for the presence of underground motor
 fuel storage tanks.
                              G-19

-------
     From the final frame of farm operators thus established, a
sample of 600 cases was drawn with within-PSU probabilities set
so that the entire sample had equal probability.  Table G-7
reviews the results of farm operators by region, column 1 shows
the distribution of farm operators by region, column 2 gives the
number of distinct farms this represents, and column 3 shows the
farm estimate based on the unadjusted sample weights.  Comparing
these estimates back to the Census totals in Table G-3, we see
that there is quite a bit of region to region variation, although
the grand total is fairly close.  This indicates that a ratio
adjustment would improve the sampling error of estimation for
this survey, which we describe in the next subsection.
IV.  STATISTICAL ADJUSTMENT OF WEIGHTS TO MINIMIZE SAMPLING
     VARIANCE

     In the previous subsection, it became apparent that the
actual sample of farms based on the ASCS list does not accurately
reflect the regional distribution of farms as measured by the
1982 Census of Agriculture.  Further, in subsection II we found
that the underground tank survey regions are very unequal in
numbers of farms.  In order that our final estimates of number
and proportion of farms with underground tanks reflect regional
variation and totals more closely, we propose a system of
adjustments to the sample weights by region.  Since some of the
six survey regions have such small sample sizes, we also propose,
for farm estimates only, consolidating the survey regions into
three areas which have about the same number of farms and which
will have over 100 sample cases each.  The proposed consolidation
is given in Table G-8, which shows the three consolidated
regions, their Census totals, the unadjusted sample estimates,
and the approximate adjustment factor to apply to the sample
weights so that our final sample estimates (of numbers of farms)
                              G-20

-------
Table G-7.  Results of farm sample draw

Region1
1-Northeast
2-South
3-Midwest
4-Central
5-Mountaln
6-Pacific
Continental
U.S. Total
Number of
farmers (operators)
sampled2
11
88
324
142
33
2

600
Number of
farms
operated
17
94
354
159
33
2

659
Weighted
number of farms
using sample weight
53,395
295,242
1,111,868
499,398
103,649
6,282

2,069,834
Regions are defined in Figure G-l.

2These farms are to be screened for the presence of underground motor
 fuel storage tanks.
                              G-21

-------










o
-H
i *
4-
<0
^
•a
C
(0
W
0)
+J
 o
a) ^j
M O
(0
T> <«-l
a)
-p -P
(0 C
Tf 0)
-H g
rH 4J
O W
u d
C-r->
O T!
U (0

CO
1
O
0)
r-l
43
(0
EH
c -a
0 (U
2 £ o^
4^ 4^ *^ r^
w w <-> 2
£E 3 «0 0
•— ^ <»— 1-1

w >
If
4J 4^
•C c>
o a>
5 *
? "







x?
4>
4->
w c
•o o
if
c *-
o
o









tf)
c
o
m
^f*

O (O
•t M
— in
Is- (SJ
r- r-




4-J
U)
*-> 0}
% I
z 1
E








+->
(A
•'-' —
 (9
w fl>
a> -C
5* „
t- o -^
o en o
Z T, >
1 S -
^ c
o^ 1
-- 10


M
— "



0>
 .,-
^ £
4J Q.
*r w
0 OT
» 2
a> w
> jF
^- *-







»
« .H
t u__
h c i:
^ « °
® 2 w
<_> c O-
1 3 T?
(O 0 C
«S H «J
in
n
«•
G-22

-------
will equal the Census totals.  The actual adjustment was made
after the field work had been completed, so that the final number
of actual farms contacted was used.  After this adjustment, the
ratio of largest to smallest weight was about 3.4 to 1, not an
excessive gap.

     In assessing the quality of the final estimates for farms,
for these three consolidated regions and nationally, we have
computed sample variances based on the final weights.  There is a
qualitative aspect to the accuracy as well, in which we
acknowledge that coverage of the far West Coast especially is
fairly low, and the estimates for the Western consolidated region
may contain some bias if these three states are strongly
different in terms of underground motor fuel storage from the
rest of the west.  However, since the West Coast accounts for
only 20 percent of farms in Survey Regions 4, 5 and 6, it would
have to be extremely different for the survey estimates of this
consolidated region to be significantly affected.
                             G-23

-------
                              APPENDIX H

                     ENVIRONMENTAL DATA COVERAGE


I.  INTRODUCTION


     Environmental data coverage by  existing data bases and literature

was explored for  geographic  locations of  the  OTS Leaking Underground

Storage Tank survey. '2'3  Data  sources were  located and subsequently

reviewed for their usefulness.  From the pertinent literature and data

sources found,  environmental data sets were  derived for survey areas

and organized within an automated data base.   Parameter  choices  were

directed toward   use  in  leak analyses and  fuel migration  modeling

studies.   The data  sets were compiled  into a Basic  Site Information

File  containing  locators,  descriptors,  and cross-reference  keys

pointing to additional  soil,  climate, and  groundwater information for

the sites  in the  survey.   Fuel component  chemical  and  physical  data

were also compiled and tabulated.
      "Literature  Searching  for Leaking Underground Storage Tank
       Project," General Software Corporation, 1985.
     2
      "Environmental Scenario Assemblage for Leaking Underground
       Storage Tanks," General  Software Corporation, 1985.

      "Environmental  Scenarios Supporting Movement of Complex Mixtures
       to Groundwater," General Software Corporation, 1986

      "Chemical-Physical Parameters  and  Processes  Effecting  Petroleum
       Fuel Migration", General Software Corporation, 1985.
                                 H-l

-------
II.  DATA SOURCE AVAILABILITY AND COVERAGE






     In the  search  for soil,   climate,  and groundwater  information,




only major readily accessible sources were considered.   These  sources




include,   among  others,  the County  Soil  Surveys   of  the  Soil




Conservation Service,  USGS publications, and  the  NAWDEX data base.   A




summary  of the sources located and  descriptions of the information




which they contain are presented in Table  H-l.






     The County Soil Surveys  of the Soil Conservation Service  provide




the  most  complete  and  comprehensive   information  on  soil




classification.   The survey status  of the  original  76 counties in  the




LUST survey  is  provided  in Table H-2.  The SCS  Soils-5  computerized




data base  contains most of the information covered  in  the published




surveys.   There were 914  site  locations recorded,  and  of these,  over




450 were covered by modern soil surveys, but  approximately 150 of  the




latter were designated as urban land  or  mixed land  complexes and were



not fully  described.






     USGS  publications and  the NAWDEX Groundwater Site Inventory




provide  variable  coverage for groundwater  and  subsurface geologic




information.   For  areas  not covered,  regional  ranges  were  recorded




from "Ground-water regions of the United States" by R.C.  Heath  or  from




the ENVIRLOC  database  as cited in Table  H-l.   These ranges must  be




used  with caution,   however,  since  they  are  broad  geographic




approximations only.
                                 H-2

-------
     To obtain up to date,   reliable climatic information, parameters



were  requested  directly from  the  National Oceanic  and Atmospheric



Administration  (NOAA).   Currently,  NOAA  is  compiling parameter



summaries  from  approximately 3000 U.S.  Weather  stations  from their



databases for the Exposure  Evaluation Division of OTS.  Publications



summarizing portions of this data  include  the  Climatic Atlas of the



United States and the Statistical Abstract of the United States.  Soil



Surveys frequently contain brief climate summaries as  well.






III. BASIC SITE  INFORMATION FILE






     The Basic Site  Information File  was designed in support of the



Leaking Underground  Storage  Tank survey  from the  work performed in a



preliminary study  described in the  Task 8 report  of  EPA  Contract



68-02-3970.   Data, data ranges, and cross  reference  keys  covering a



variety of locator,   climate, soil,  and  groundwater information were



included  in the file to  enable  the  user  to have  a  general



understanding of site location  and conditions,  and to obtain further



information as necessary.





     The  file itself  contains four  sections:   site  location and



identification,  climate,  soil,  and  groundwater/geologic.   The



parameters in the file and  their  corresponding lengths are  shown in



Table H-3.  Tables H-4 through H-6 are  examples  from the Basic Site



Information File.
                                H-3

-------
     A.   Site Location and Identification



     The site location and identification portion includes identifiers



ranging   in  resolution  from  general region to  specific  site.   These



locators aid in  the determination of  the number of sites within a



particular  state/ county, or region, and in the  location of the actual



site on  a USGS  topographic map.






     The LUST Regions (Pacific, Mountain, Central, Midwest, Northeast,



and Southeast) are  the largest divisions contained in the  file,



dividing the United  States into  six parts for  survey purposes.   The



PSU, or  primary sampling unit, is a  further  division of the  LUST



Region which encompasses one or more counties.   There are 34  PSUs



included in the LUST Survey which cover a total  of 76 counties.






     The state  and county  PIPS   codes,  or  Federal  Information



Processing  Standards,  are  numeric  codes  for each  state and county.



The state and county codes are two and three  digits respectively, and



are sometimes  combined  into a single five  digit  identifier.   Being a



standard identifier, the  PIPS Code helps  to avoid confusion due to



spelling errors and nonuniform abbreviations.





     The USGS  Topo Quad information is provided  for easy  reliable



geographic location.   This  information  includes the name of the




topographic quadrangle on which the site may be found, the map scale



of the quadrangle,  and  the bottom right coordinates  of the map.  This



information  may  be useful in the  future for digitization of mapping



and site location.
                                 H-4

-------
     Survey sites  were usually received marked  on a  USGS  topo map.



Sometimes,  however,  sites were  marked on nonstandard or state road



maps,  or occasionally not marked  at all.   If a topo quad could be



determined for a site, this information was included in the  file,



otherwise it was omitted.






     The Soil  Survey Area  information provides the name of  the Soil



Conservation Service County  Soil  Survey  covering  the site,  the year



the  survey was published,  and  the survey area  code.  County soil



surveys cover  a  county,  group of counties,  or sections  of counties.



Sites  located in  areas with  no  current published soil survey are



labelled "Area not surveyed"  at  this point.  Sites  not marked,  or



marked on  large scale  maps  are labelled  "Site  not specifically



marked", or  with some  other pertinant descriptor.   The Survey Area



Code is obtained from section one  of the Soil Conservation Service Map



Unit Use File (MUUF).  Every current county survey has a corresponding



code,  which  is the county  PIPS code for single whole county surveys.



For  partial county  and  multi-county surveys,   codes are  600  numbers.



These codes are found by searching the MUUF for state and survey area



name,  and are  used for finding cross  reference keys to specific soil



information.






     The specific site locators are the site ID,  latitude,  longitude,



and approximate elevation.  The site ID is an alpha-numeric code  taken



from the marked  topo  maps  as received.  The number includes the PSU.




For sites with multiple tanks,  a letter is tacked  onto  the end of the
                                H-5

-------
ID  (i.e.  A,  B, C,  etc.)  identifying each tank, so  that  each tank has



its own unique record in the event  that soil conditions may differ.






     The site coordinates were determined by measuring those marked on



USGS topo maps with a gridded ruler to the nearest 1/16 inch and then



performing the necessary calculations.  The coordinates were presented



in the file in degree:minute:second format.   Sites received marked on



maps with  insufficient scale or resolution were included with general



information only  (i.e.  no  specific coordinates).   The  elevation  was



taken from the topo map.






     The Hydrologic Unit,  or HU Code, is a numeric code assigned to a



drainage basin or distinct hydrologic feature by  the  Office of Water



Data Coordination.   Although  the HU Code  is applied mainly to surface



water,  it is sometimes used to organize groundwater studies.   An



example of this  is O.K. Todd's major water resource divisions in



Ground-Water  Resources  of  the  United States.   These  major divisions



correspond to  the first  two digits of the  HU Code,  as shown in



Figure 1.  HU Codes are  available from ENVIRLOC.





     B.  Climate





     State Climatic Divisions   (SCDs) are  areas  within  states which



have similar  climates.  The National  Weather  Service  has defined 353



divisions in the  United States   which frequently  follow  county




boundaries.  These  divisions, which were retrieved from GEOCOLOGY, for



survey locations,  will  help  determine the closest applicable weather



station  from  which to take climate data.   NOAA will provide rainfall
                                 H-6

-------
statistics to the Exposure  Evaluation Division  for those stations



recording hourly  precipitation  as well  as mean temperature and



humidity by SCO.






     C.  Soil






     The  soil  information included in  the  basic  site file  provides



some parameters plus soil  type keys for obtaining additional data from



Soils-5f the soil data base of the Soil  Conservation Service.






     The Soil Map Unit is  an  alpha-numeric which is obtained  from the



Soil Conservation Service published soil surveys.   The unit  is  found



by locating the site on one of the soil maps in  the county survey,



usually by comparison with the marked  topo map.  The Soil  Map  Unit and



the Survey Area Code are then used to extract the  Soils-5 Recnumber



from  the  Map Unit Use  File (MUUF)  section  three.  The  Soils-5



Recnumber consists of the  two character  state abbreviation and a four



digit number which  together determine the record to access within the



Soils-5 data base.   The additional information include such parameters



as permeability,  pH,  percent clay, etc.  A sample of the available



data is shown in Table H-7.






     If a county or part of a county did not have  a current published



soil survey, a soil type inference was made  using  surrounding county



soil surveys,  making either an  individual  soil type inference or a



major  association  inference as  shown in  the  site  file.   Soils-5
                                H-7

-------
Recnumbers were then found as  before.  If an inference could not be



made with reasonable confidence, then no inference was made.






     Additional  information in the Basic Site File includes seasonal



high water table,  availability of C-Horizon  (subsoil)  parameters,  and



relative  corrosivity to steel and concrete, all of which could be



useful for the prediction of possible tank leaks.   The seasonal high



water table information provides  a depth range, water table type,  and



the  months  of  common occurrence.   Availability   of  C-Horizon



information is  a  yes  or no  indication of whether  county soil survey



data  include the  mineral  subsoil.   Risk of Corrosion  is  a relative



parameter (low,  moderate, high) determined primarily by  drainage class



and  texture,  total acidity,  resistivity  at field  capacity,  and



conductivity of saturated extract, as  described in part  603  of  the



National Soils  Handbook  of  the SCS.






     D.   Groundwater/Geolog i c





     R.C.  Heath divided the United  States into major groundwater



regions (referred  to in  the site  file as Heath Regions)  in his report



"Ground-water  regions of the United States".  Figures 2  and 3 show the



boundaries of the  fifteen regions.  Heath established  ranges  for



transmissivity,  hydraulic  conductivity,  and porosity for  these



groundwater regions,   which may be  used  if  actual data is  not



available.  These  ranges are very general,  however,  and  should be used



with  caution.
                                H-8

-------
     Space is provided  for the NWWA (National Water Well Association)



subregion for future  input.   The NWWA is  currently  organizing



hydrogeologic parameter  ranges for  subsets of the region of R.C.



Heath.






     A literature search  was performed  in the National  Water Well



Association bibliographic  data base to  locate articles and  studies



describing  aquifers  in  the  areas  of  interest.    Most  of the



publications  were USGS reports  which contain  good groundwater and



geologic descriptions.   These USGS  publications were used  to develop



the groundwater file  which is cross referenced in  the Basic Site File.






     Extensive searches were performed  in  the NAWDEX Ground Water Site



Inventory to obtain water and well information.   Site  Resolution



(position  with respect  to aquifer),  Well Usuage Description (domestic,



public, industrial,  etc.),  and  Depth to Groundwater  were usually



obtained from the GWSI.  Well sites within  five minutes latitude and



longitude  of  a survey site were  used to determine  the  parameters at



that site.  If no  well sites  were  within this  radius of the LUST site,



Depth to  Groundwater  was taken from ENVIRLOC  (this  appears as a



range).   The other  literature  sources  previously mentioned were



occasionally  used when  available.






     The Basic Site Information  File,  Soils-5,  the Groundwater



Information File,  and  the future NOAA weather  data,  will be  useful



tools providing reasonable environmental  scenarios to  the modeller.
                                H-9

-------
This will aid not only in modeling efforts for the LUST study, but  for



future modeling studies as well.
                                 H-10

-------
                   Table H-l.  Information Source Summary  (1 of 5)
   Source

Literature:

County Soil
Surveys USDA Soil
Conservation
Service
USGS Publications
  Water Resources
  Data
  Guidebooks for
  Fieldtrips
  Water Resources
  Bulletins
Parameters
soil type,  level,  slope,
permeability, pH,
available moisture
capacity, temperature,
precipitation, soil
texture, % fragments,
sieve analysis, liquid
limit plasiticity,
index, shrink/swell
potential, erosin
factors

depth to groundwater soil
bulk density, cation
exchange capacity, organic
content, clay content
surface water data

observation well number,
location, hydrologic unit,
groundwater level,
well characteristics,
aquifer type, groundwater
quality

thickness and
characterization of
rocks and water
bearing formations

hydrogeology of
principal aquifers,
saturated thickness
ranges, temperature,
water level,
characterization of
core samples, analysis
Geographic Coverage
  and Frequency
most US counties
 (down  to  60  inches
 only)
                                                        some counties
all US states

most US states,
site specific
US, site specific
 (usually to bedrock),
 info variable by
state

US, site specific,
 info variable by
state
                                H-ll

-------
                   Table H-l.  Information Source Summary (2 of 5)
  Source
  Water Resources
  Bulletin
  Geological
  Circulars
  Water Resources
  Investigations
  Open File
  Reports
Resources of the
United States
O.K. Todd, 1983
Premier Press
Parameters

of rock samples,
hydraulic conductivity,
specific gravity,
particle size, porosity,
water quality

hydrogeology of
prinicipal aquifers,
saturated thickness
ranges, temperature,
water level,
characterization of
core samples, analysis
of rock samples,
hydraulic condictivity,
specific gravity,
particle size, porosity,
water quality

soil chemistry,
tranmissivity,
hydraulic conductivity,
thickness, sieve
analysis, soil layers

well data, water
quality, pumping
and drawdown studies

Water level, aquifer
description
precipitation,
occurrence of
groundwater,
storage coefficient,
evapotranspiration,
base of fresh water,
potentiometric contours,
basement slope
Geographic Coverage
  and Frequency
US, site specific
info variable by
state
US, site specific,
info variable by
site
US, site specific,
info variable by
site

US, site specific,
info variable by
site

US major groundwater
regions, info available
for most regions
                                H-12

-------
                   Table H-l.  Information Source Summary  (3 of 5)
  Source

Statistical
Abstract of
the United
States, 1984
US Dept. of
Commerce,
Bureau of Census

Climatic Atlas
of the United
States, 1968
U.S. Dept of
Commerce,
Environmental
Science Services
Administration,
Environmental
Data Service

Hourly
Precipitation
Data, NOAA,
US Environmental
Data Service
(monthly
publication by
state)

Topographic Map
Series, USGS,
Reston, VA

Ground-Water
Regions of the
United States,
R.C. Heath,
USGS Geological
Survey Water-
Supply Paper
2242
Parameters

mean temperature,
precipitation, days
w/precipitation
greater than 0.1 inch,
average snowfall,
average percent sunshine,
average windspeed

temperature,
precipitation,
state climatic
divisions, humidity,
evaporation, snowfall
radiation, skycover,
wind speed
hourly precipitation
Geographic Coverage
  and Frequency

selected US cities
US  (maps)
US weather stations
elevation, coordinates
groundwater regions,
descriptions, ranges of
transmissivity,
porosity, hydraulic
conductivity, and
recharge
US, most areas
US groundwater regions
                                 H-13

-------
                   Table H-l.  Information Source Summary (4 of 5)
  Source

NOAA
(National
Oceanic and
Atmospheric
Administration)
Data Bases:

Geocology,
Oak Ridge
National
Laboratory
(contained in
GEMS)
NAWDEX
(National Water
Data Exchange)
Ground Water
Site Inventory,
USGS, Reston, VA
National Ground
Water Information
Center Data Base,
National Water
Well Association,
Worthington, OH
Parameters

temperature,
wind speed,
precipitation,
state climatic
division, sky cover,
humidity
monthly temperature
monthly evaporation

state climatic divisions
within counties

soil great groups

well description,
groundwater level,
water use,
lithology,
transmissivity,
hydraulic
conductivity,
storage coefficient,
water quality

bibliographic,
key word search
covers current
literature including
USGS publications
Geographic Coverage
  and Frequency

US weather stations,
data collected variable
by station
US state climatic
divisions
eastern US counties

US


eastern US

US site specific,
data variable
by site
global, major emphasis
in US, literature
dependent
                                 H-14

-------
                   Table H-l.  Information Source Summary  (5 of 5)
  Source

ENVIRLOC,
Soil/HU Code,
General Software
Corporation
Landover, MD
Soils-5, USDA
Soil Conservation
Service,
Washington, D.C.
Parameters

approximate depth
to groundwater
ranges, soil
parameter ranges,
Hydrologic Unit
Code, Heath
Groundwater region
number

essentially same
information and
coverage as published
surveys
Geographic Coverage
  and Frequency

continental US by
Zip code or coordinate
most US counties, info
for most counties (with
modern published surveys
only)
                                H-15

-------
                 Table H-2.  Status of County Soil Surveys (1 of 6)
  County

Arkansas:
  Garland
California:
  Alameda
  Los Angeles
  San Mateo
Colorado:
  El Paso
  Teller
Connecticut:
  Hartford

  Tolland

Florida:
  Duval
Illinois:
  DuPage
Indiana:
  Grant
Survey Name
Alameda Area (excludes
western section)

Los Angeles County, West
San Fernando Valley Area

San Mateo Area
(excludes northern
section)
El Paso County Area
(excludes northwestern
section)
Hartford County

Tolland County
City of Jacksonville,
Duval County
DuPage and Part of
Cook Counties
Grant County
Survey Status
                            Mapping in
                            progress
Complete


Complete


Complete




Complete
                            Mapping not
                            started
Out of print

Complete


Complete



Complete
Out of print
Mapping in
progress
Year
Published
1966


1979


1961
1980
1962

1966


1978



1979



1915
                                 H-16

-------
  County

Iowa:
  Pottawattamie
Table H-2.  Status of County Soil Surveys (2 of 6)


  Survey Name                 Status Survey
Kansas:
  Johnson

  Waynedotte
Kentucky:
  Bullitt
  Jefferson

  Oldham

Minnesota:
  Ramsey


Mississippi:
  Issaquena

  Warren

Missouri:

  Caldwell

  Carroll
  Chariton

  Clinton

  DeKalb
  Pottawattamie County
  Johnson County

  Leavenworth and
  Waynedotte Counties
  Jefferson County

  Oldham County
  Washington and Ramsey
  Counties
  Issaquena County

  Warren County



  Caldwell County

  Carroll County




  Chariton County

  Clinton County

  DeKalb County
                   Year
                   Published
Out of print       1914

Mapping in
progress
Complete           1979

Complete           1977



Mapping in
progress

Complete           1966

Complete           1977


Complete           1980



Complete           1961

Complete           1964



Complete           1974

Out of print       1912

Mapping in
progress

Out of print       1912

Complete           1983

Complete           1977
                                 H-17

-------
                 Table H-2.  Status of County Soil Surveys (3 of 6)
  County

  Gentry
Montana:
  Hill
  Liberty


  Toole
Nebraska:
  Arthur
  Blaine

  Custer

  Grant


  Hooker

  Logan

  Loup

  McPherson

  Thomas

New Hampshire:
  Hillsborough

  Rockingham
Survey Name
Arthur and Grant
Counties

Blaine County

Custer County

Arthur and Grant
Counties

Hooker County

Logan County

Loup County

McPherson County

Thomas County


Hillsborough County

Rockingham County
Survey Status

Mapping
complete
                            Mapping not
                            Started

                            Mapping not
                            started

                            Mapping not
                            started
Complete


Out of print

Complete

Complete


Complete

Complete

Out of print

Complete

Complete


Complete

Out of print
Year
Published
1979


1954

1982

1979


1964

1974

1937

1969

1965


1981

1959
                                 H-18

-------
  County

Iowa:
  Pottawattamie
Table H-2.  Status of County Soil Surveys (2 of 6)


  Survey Name                 Status Survey
Kansas:
  Johnson

  Waynedotte
Kentucky:
  Bullitt
  Jefferson

  Oldham

Minnesota:
  Ramsey


Mississippi:
  Issaquena

  Warren

Missouri:

  Caldwell

  Carroll
  Chariton

  Clinton

  DeKalb
  Pottawattamie County
  Johnson County

  Leavenworth and
  Waynedotte Counties
  Jefferson County

  Oldham County
  Washington and Ramsey
  Counties
  Issaquena County

  Warren County



  Caldwell County

  Carroll County




  Chariton County

  Clinton County

  DeKalb County
                   Year
                   Published
Out of print       1914

Mapping in
progress
Complete           1979

Complete           1977



Mapping in
progress

Complete           1966

Complete           1977


Complete           1980



Complete           1961

Complete           1964



Complete           1974

Out of print       1912

Mapping in
progress

Out of print       1912

Complete           1983

Complete           1977
                                 H-17

-------
                 Table H-2.  Status of County Soil Surveys (3 of 6)
  County

  Gentry
Montana:
  Hill
  Liberty


  Toole
Nebraska:
  Arthur
  Blaine

  Custer

  Grant


  Hooker

  Logan

  Loup

  McPherson

  Thomas

New Hampshire:
  Hillsborough

  Rockingham
Survey Name
Arthur and Grant
Counties

Blaine County

Custer County

Arthur and Grant
Counties

Hooker County

Logan County

Loup County

McPherson County

Thomas County


Hillsborough County

Rockingham County
Survey Status

Mapping
complete
                            Mapping not
                            Started

                            Mapping not
                            started

                            Mapping not
                            started
Complete


Out of print

Complete

Complete


Complete

Complete

Out of print

Complete

Complete


Complete

Out of print
Year
Published
1979


1954

1982

1979


1964

1974

1937

1969

1965


1981

1959
                                 H-18

-------
                 Table H-2.  Status of County Soil Surveys (4 of 6)
  County

New York:
  Albany
  Essex


  Fulton


  Hamilton


  Queens


  Rensselaer
Ohio:
  Greene

  Miami

  Montgomery

  Preble

Oregon:
  Clackamas


Rhode Island:
  Bristol

  Kent

  Washington
Survey Name
Albany County
Rensselaer County
Greene County

Miami County

Montgomery County

Preble County


Clackamas County Area



Rhode Island

Rhode Island

Rhode Island
Survey Status


Out of print

Mapping in
progress

Mapping not
started

Mapping not
started

Mapping not
started

Mapping not
started

Out of print

Mapping complete


Complete

Complete

Complete

Complete


Complete
Complete

Complete

Complete
Year
Published
1942
1937
1978

1978

1976

1969


1985



1981

1981

1981
                                  H-19

-------
                 Table H-2.  Status of County Soil Surveys (5 of 6)
County

South Carolina:
  Lexington

  Richland

Tennessee:
  Chester

  Henderson

  Madison

Texas:
  Brooks
  Collin

  Harris
Survey Name


Lexington County

Richland County




Henderson County

Madison County
Collin County

Harris County
Survey Status


Complete

Complete


Mapping complete

Complete

Complete


Mapping in
progress

Complete

Complete
Year
Published
1976

1978
1960

1978
1969

1976
  Hays

  Kenedy


  Travis

  Willacy

  Williamson

Utah:
  Salt Lake



  Tooele
Comal and Hays Counties     Complete
                   1984
Travis County

Willacy County

Williamson County
Salt Lake Area
(excluding eastern
section)
Mapping not
started

Complete

Complete

Complete


Complete
                            Mapping in
                            progress
1974

1982

1983


1974
                                 H-20

-------
                 Table H-2.  Status of County Soil Surveys (6 of 6)
  County

Washington:
  Cowlitz
  King
  Snohomish
  Wahkiakum

Wyoming:
  Campbell
Survey Name
Cowlitz Area (eastern
part excluded)

King County Area
(eastern part
excluded)

Snohomish County
Area (eastern part
excluded)
Campbell County
                   Year
Survey Status      Published
Complete           1974
Complete           1973
Complete           1983
                            Mapping complete
Out of print       1955

Mapping in
progress
  Johnson


  Sheridan
Johnson County, Southern
Part

Sheridan County
Complete           1975
Out of print       1939

Mapping in
progress
                                 H-21

-------
Table H-3.  Parameters and Record Lengths included
       in the Basic Site Information File.
  LUST Region                        30
     PSU                              2
     State FIPS                       2
     County FIPS                      3
     USGS Topo Quad                  30
     Scale                            9
     Bottom Rt Latitude               8
     Bottom Rt Longitude              9
     Soil Survey Area Name           80
     Year Published                   4
     Survey Area Code                 3
     Site ID                         11
     Latitude                         8
     Longitude                        9
     Elevation (ft)                   5
     HU Code                         10
     SCO                              3
     Weather Station                 35

  Soil Map Unit                       5
     Series                          53
     SoilsS Recnumber                 6
     Soil Inference                  14
     Inference From                  33
     Inference Associations          56
     Inference SoilsS Numbers        54
     C-Horizon Info                   3
     High Water Table                43
     Corrosivity to Steel            13
     Corrosivity to Concrete         13

  Heath Region                       27
     NWWA Subregion                  23
     GW & Geologic Description      207
     Site Resolution                 35
     Well Usage Description          35
     Depth to GW (ft)                 7
     GW Cross Reference              25
                       H-22

-------
                                       o
                                       ro
                                       i—
                                       m
                                       g
                                                                            .0
                                                                            CO
U
                                                                            §

                                                                            S
                                                                            o
                                                                            CO


                                                                            U}
2
o
o
3
^j
O
'4-1
P
•f-1
i-
<, <
•J
4<^
-H
ot
•H
•*
•V
J3
•
-I
01
i— 1
CO

gion: Central
&
^\
w
3
2: Arthur
s
z
m _,
- §I
«*> .. °
0 5P Q
QJ Qj
" H &
CO fc, EH
•• ^J
jb co



Scale: 1:62500


o
o
C)
r— 1
• •
_1 J
• H
JJ
Bottom Right La


gitude: 101:30:00
amo: Arthur and Grant Counties
c z
Bottom Right La
Soil Survey Area 1


r-.
i-H

Year Published:


CM
-3"
I-H
O
o
1
5
5
in
o
f-H
1
•H
3
i — i
o
vO
0)
Survey Area Codi
: N230000635
a
a
•H
CO
§
•H
4-1
5
co
u
is
CM
O

de: 10180014
cS

Series: Valentine Fine Sand

u
cfl
a
r-l
•H
£

i — i
o
0
U
0
in
en
i— i
1

Inference From:

Inference :
ence Associations:
i-l
•H (1)
•H <4-l
o c


05
ence SoilsS Number
iw

High Water Table: GT 5.0ft
To Concrete:

r-H
o 3J
G 1 i
CO
o o
i[ i 1 1
o ^
N -H
•H W
S <§

NWWA Subregion:

c
.-i
PL,
00
•H
X
1
»
5
a
5
m
S
U
T3
C
CO
TD
C
cfl
CO
01
c
•H
4-1
T3
(U
4-1
C3
T3
i— 1
0
!f,
O
O
C
3
1
05
01
'4-1
•H
3
cr
co
T3
C
cfl
01
C
3
Q
•H
4-)
d Geologic Descrip
§
§

T3
C
CO
Resolution: dune s
3
•H
CO

Irrigation

Usage Description:
H
iH
2

0
o
•
m
^^
, to Groundwater(ft
5
t



oss Reference: 52
S
5

                                                H-23

-------
 CJ
 e
 CO
 •o
 V.
 c
0")
43
CO
H
Midwest
on
             oo
             1 — 1

             • •

             03
             &
             M
             b


             S
             (0
             4J
st


CO
m
o
CO
CU
M
t<
£
I
f-- Sweetser
E
C
2
%
3
a
£
D
o
o
o
CO
o
<»•
1
•H
4->
Scale: 1:24000
Bottom Right La
gitude: 85:45:00
anio: area not surveyed
C 2
S (0
0)
4J !H
£, <
5 *
« 0)
4-> CO
4->
O -H
£


Year Published:
o
vO
OO
• •
4J
U-I
8
•H
4->
(0
0)
t-4
w

CM
O
CO
0
-i
6
S
in
en
r-l
•H
a
Inference From: Miami County, 1979 (103)

Soil Inference: Ba.Pw
Blount-Pewamo Association

Inference Associations:
: IL0014,MI0042
en
Inference Soils5 Number
High Water Table: 0-3. Of t, perched apparent Dec-May

C-Horizon Info: yes
s
o
T— 1
S
s
u
/^
u
p
4=
M
•H
Corrosivity to Steel: h
5
•ft
ff
in
/^
»-M
CO
r— I
CO
S-l
4-1
C
CU
u
T3
CU
4-1
cfl
•H
O
cfl
r-t
oo
C
o
z
vO
• •
g
•H
rr>
I
5
ro
*
ion: Unconsolidated sand and gravel deposits over water bearing limestone and
dolomite bedrock
jj
GW and Geologic Descrip
i-
cu
'4-i
•r*
^
o-
CD
-^
O
O
v<
•a
CD
43
"O
C
CO
•O
OJ
4-1
CO
TJ
•rH
i — 1
O
•f.
C
o
o
c
Site Resolution: over u


Well Usage Description:
0
i
CM
CO

Depth to Groundwater(ft!


GW Cross Reference: 29
                                                            H-24

-------
                                                    o
                                                    CNJ
 C
 3
 O
 flj
 >
 3
                                        C
                                        3
                                        C
 u
•H
 CO
_L

CU
(0
H
       i
Corrosivit


r-
•H
C
S-l
1
TH
CO
4-1
tn
n)
C
U
4-1
rH
O
U
•H
4-1
C
CO
iH
4-1
<
O
I— 1
§
•H
5
GJ
32
C Description: Layers of clay, sand, shells, and limestone, very shallow water table, springs
and seeps common
-H
S
1
T3
g


0)
•H
3
cr
CO
CO
•a
•H
o
rH
0
rH
rH
cfl
tn
0)
o
•H
4->
P
rH
S
s
3
-H
C/3


•H
U
tn
0)
E
o
•a
u
•H
i— 1
3
Cu
C
0
•H
4J
CO
GO
•H
S-l
•H
•H
4J
a
•H
J-l
I
1
W
D
rH
rH


dwater(ft): 23.00

B
S


CO
OJ
i
0)
J-i
O
14-1
in


                                                             H-25

-------
                          u
                          I/I
                          V.
u-i
 to
 o
co
 CO
 >
 a
 c
 o
 i-l
 o
O


c.
 01
,c
 o

 
      «
           u

           I/I
           UJ
           T
            O
            UJ
            a
                                o
                                tu
                               >
                               ^
                               ^

                               i/i
                               i
                               u

                               a

                               o
                      3

                         —    U

                         2 > a

                         *$£
                    o
                 IT a
                 >  Z
                 «  4
                 -J  ft
                                     u u>

                                     Zfe
                                     a j
                                     s m
                                     in
                                         •
                                     u z
                                     z «
in »-
u _/
u <-
« in
ft
a 1/1
lu 3
»- C
   U
O ft
Z 4
« u

V <
o u

4 >-
z  z
O  tfl

in  z
in  >
UJ  O
C  ft
z  •-
-  1
   c
o  —
U  -I

R  - a
-  < o
C  ft ft
10  I- U
   I/I
o  o o
UJ  S Z
Z  I/I <

4  U O
ft  Z Z
_J W O
u u z
> - <
   z a
 • ^
ft    a
u. m r.
u. UJ U.
O Z
   w o
IL Z U
o » «n

ui n
>-    lu
V Z ft
•• « «
m o
z j in
o    <
u »- u
   _i a
in - *
hi ut
M     •
ft 7 N-
U » Z
M O U
   ft O
>• C ft
j    u
j x a
it > i
1 1 1 t- XI IT O *•>
1 1  i «n n
i ft »- -i
j , 	
1 1 O t- 1 C IT O
I i -. - i « n «
1 1 3 » l i i i
1 o — l in o o
1 1 •- .J I nj <\j rj
1 1 _i I
r OF MATERIAL LESS
J=_EASSiMC_SiEyE_«lA_
L__lQ__i 	 4Q__i_2Qfl__J
100 9*-lOO 9O-IOO
100 90-100 70-95
9^-100 90-100 B5-100I
1 ! Z II
i u. a i
1 1 U 41 1 O O O
1 iH'jIII

| ( u - M
1 1 < L* 0 0 O
i i ff *•> ai
1 IU. A -1

U !
IUJ
1 M
1 ftj 1
1 UJ O
1 £J >-
ia r
1 01 m i
I 1 4 « «
1.1 II
l-t 44
Id 1
11/1 1 • •
II « 4 »
1 Ci ill
1 IU 1444
	 ESJIMftJ
UNIFIED
	
. . CL-ML
.-ML
.-ML. CL
1 1 L' U 0
1 1
1 1 • » •
1 1 J _l _l
i lisa
1 UJ 1
l ft i
i 5 i
i •- i
1 K 1
1 »- 1
I 4i in
1 0 1 U.
1 t - U. —
1 1 in > tn
1 Z - 1 0
1 K • 1 •* in 10
1 1 UJ I
1 l»-i 1
1 lui 1
1 > I en »
1 - 1 71 _»
1 \r l— i— |
10)11
1 ft 1 1 1
1 ft 1 J II
1 O 1 uj Ol
1 U 1 IU H
1
1
. 	 	 	 J
JIONfwiKO lOPGANICl
[Q3SI FPOD. IMATTFP 1 	
L_I_ifiBQUEi_iECIl_l_Sl
51 «L 1 1-! 1 	 B
5| *L 1 1-2 1
	 L 	 1 	 1
1 C ul I r) n m
1 ft 41 W1 « « •
1 Ul U.I 1 • • •
1 1 4)

*- U* Zl O C O
1 I f» M
1 Ul O*

u:
1 >- U
1 «- N
1 Z in iv. IM M
!3§ vvv
1 4 J
i V: 7
1 D tee*
l J •-<-<• • «
1 •" 1- It S f> *
i c c- ai i i i
I IL 1 • • •
i a i« « r-
~"1 »~i
It~ i
u u | nj c> o
1 _i < -J rj — fvi
1 CD tt ZJ • • •
1 4 4 >4 O O O
l_l U M l i l
I — zi N *> n
4 er H — -« —
1 > i_ M • • •
r: !oo°
i > *
i *
II "< 0 O O
1 4 >• ftl • • •
1 lu r- Zl rj "J M
— — -^ Z •- NJ l l l
1 ft _J Z) < « 0
1 tM — • —^ • 1 •
1 £ O H O 0 O
1 1
i_i iir o o
1 3 >• "H ^ » 1"
1 B >- m • • •
I f- 1/1 m i i i
1 1/1 z v r o tr
1 — u> O) M »n n
IO e ~i • • •
1 — ia m m
1 4 U 1 1 1 1
j_< a 1 o o c
1 Z ~ l o
| ». i |rv n «
1 _l I U<
1 • Z' •-!
1 t- t- O «
i z •• •< en
UJ O l-l U>
1 »- a m 01
i ex >i
1— — -^-H
i D a i
l> ft C!
1 5 IS 1
1— — -H— <
IIJII
1 m >- ZJ
1 ri c x i
1 UJ •- -1
la--»H
1-4 • II
ted *- zi 1 1
.-» z -< 1
1— 1 — — 1— 1
1 1 in II
1 1 in 1 1
1 1 IL II
1 1 Z 1 1
1 wl O II
1 U ft II
1 01 4 II
i en r i i
UJ 1 1
1 1 »• Zl Ol
1 l c ••« a
1 1 lu ~< Al
-J— _|~4
1 1 V, 1

i 01 z ii
1 1 0 1 1
ea ft ii
UJ < 1 1

1 UJ T II
i a.i >- ~» i
i uj a z n
1 Ul UJ -4 1
1 1 O -< 1
i— «— — i— <
ii ii
1 in 1 1
1 I X II
i »- i
!_• C 1
i er>i i i
l M. i
i i
i en c ii
i u Z ii
1 M — II
i jn ii
U-J-J
Id II
1 M Z II
1 3-1 >- -< 01
1 1 ft •-! «
1 U kJ «
1 1 1
1 lt/»
1 Zl 1
1 f 1
i a
1 Zl 1
1 — <-H
1 1 1
1 1 1
1 1
._ __ T
r%8AQ — T —
OKIOUO1J
r 	 i
1 >H
1 Ul
1 ZJ uJ
1 U> 73
i > a
1 IkJ
1 ICC
                 IU 4
                 S ft O
                 »• VS t-
   1C. Z  II  I   I
I   Ik, -  loo
I   10 -  I
                                          j b Z    ill
                                          1 IL x   Icon
                                          IO -   i
                                                                                                               n
                                                           H-26

-------
                                                    CD

                                                    •a
                                                    a
                                                    3
                                                    O
                                                    VJ
                                                    OC

                                                    .c
                                                    4-'
                                                     00
                                                     c
                                                    •H
                                                    -a
                                                     c
                                                     o
                                                     O,
                                                     CO
                                                     4)
                                                     O  CO
                                                     o  ao
                                                     C -O
                                                     3 O
                                                        H
                                                     U
                                                     •H e
                                                     so -H
                                                     o
                                                     ^H -a
                                                     o ai
                                                       o
                                                     O -H
                                                     i-l SO
                                                     CO  0)
                                                     2:  1-
                                                      0)
                                                      u

                                                      QO
H-27

-------
                                             OJ
                                            X

                                            n-l
                                             o

                                             CO
                                             c
                                             o
                                            1-t
                                             SO
                                             V
                                            to

                                            "O
                                            a
                                            3
                                            o
                                            V*
                                            00
                                           CN
                                            M
H-28

-------
                                           01
                                           1-1
                                           i-H
                                           «
                                           >

                                           —*
                                           S
                                           •H
                                           >

                                           i—I
                                           fH
                                           <
                                           c
                                           o
                                           •H
                                           oo
                                           CT3
                                           01
                                            00
                                           b-;
H-29

-------
                           APPENDIX I
                      MULTIVARIATE ANALYSIS
I.   INTRODUCTION

     While the tables presented in Section 9 provide a useful
descriptive look at leaking tanks and conditions under which
leaks occur, they do not take into account the simultaneous
effects of many variables.  To respond to this analytical need,
multivariate statistical models have been developed to examine
the relationship between leak status (1 = leak, 0 = no leak) [or
leak rate (gallons per hour)] and various explanatory variables.

     The advantage of the multivariate analysis is that it
provides a method of assessing the contribution of individual
explanatory factors, while simultaneously controlling for other
variables.  The procedures used also allow a step-wise approach
(i.e., first finding the one variable that best predicts leak
status [or leak rate], then the second best predictor, etc.) and
a test for the statistical significance of coefficients of each
variable in the model.  The results of the multivariate analysis
have been summarized in the next subsection so that the reader
may learn the outcome of the multivariate analysis without having
to go through all the mathematical details.  The technical
details on mathematical formulation can be found in later
subsections, along with the final equations for the multiple
regression and logistic regression models developed.
                              1-1

-------
II.  SUMMARY OF MULTIVARIATE ANALYSIS RESULTS

     The major results of the modeling efforts are presented
below.  The reader should also note the caveates and limitations
at the end of this summary.
     A.   Multiple Correlations

     The multiple correlation coefficients (R) from the final
regression models (which retained only variables with significant
regression coefficients — see Subsection C for confidence
levels) were about .30 for leak status and .45 for leak rate,
demonstrating low to moderate predictive ability.  This
corresponds to R2 values of aboue .08 and .20, respectively.
Since R2 can be interpreted as the fraction of variance accounted
for by the model, it is clear that the models do not account for
most of the variance in leak status and leak rate.
     B.   Predictors of Leak Status

     Based on the coefficients in the regression and/or logistic
models, the probability that a tank system leak tends to increase
for:

     o    Older tanks,
     o    Tanks with no leaded gasoline stored,
     o    Tanks with passive cathodic protection, and
     o    Tanks for which no log of deliveries is kept.
                               1-2

-------
The positive relationship between leak probability and passive
cathodic protection might seem surprising.   A possible
explanation is that passive cathodic protection tends to be used
in areas which have a history of corrosion/leak problems.
Another explanation could be that passive cathodic protection is
strongly correlated with the storage of aviation fuel and, thus,
might be a proxy for this fuel type.  (The multivariate model
equations for leak status may be found in Section III, which
follows.)
     C.   Predictors of Leak Rate

     Among leaking tank systems, the leak rate tends to be larger
for:


     o    Fiberglass tanks;
     o    Tanks not on a concrete pad;
     o    Tanks both old and steel (i.e., an interaction
          effect) ;
     o    Tanks attached to other tanks; and
     o    Tanks in establishments with operators trained to check
          for line leaks.

The above factors are not indicators of leak likelihood, but of
larger leak rates among leaking tank systems.  The last factor
may well be a case of reverse causality — i.e., where tank
systems leak heavily, operators are trained to detect line leaks
(rather than vice versa).
*More precisely, fiberglass tank systems show less increase in
 leak rate as they get older.
                             1-3

-------
     D.   Limitations and Caveats



     In addition to the comments about the limitations of the

scope of the study presented in Section 8, the following

limitations and caveats apply to the multivariate analysis:
          Only business, government and military sectors are
          included (no farms).

          Manifolded tanks that could not be separated for
          tightness tests are not included.

          Although a long list of 49 potential explanatory
          variables were included, there are other possible
          variables which were not in our data base and whose
          effects are, therefore, not accounted for.  In
          particular, soil characteristics were not available for
          analysis and use in the models.  However, backfill
          around the tank (e.g., sand/gravel) is included and may
          be more relevant.

          The multivariate analysis finds "measures of
          association" rather than causality.  Naturally, since
          the variables used were suspected of affecting leaking,
          the discovery of a statistically significant
          association tends to affirm a causal linkage.  But the
          reader is cautioned that a different covariate could be
          the real causative factor, as in all statistical
          correlation studies.   For example, the variable "age of
          tank" could represent the effects of aging, per se, or
          age of tank could be a proxy for different installation
          techniques which changed over time, or different resins
          used in the manufacture of fiberglass tanks in
          different production years.
                             1-4

-------
III. MULTIVARIATE MODEL DEVELOPMENT PROCEDURE


     A.   Overview

     Two regression models (one to predict leak status and one to
predict leak rate) were developed using the variables in Table
1-1 as candidate predictor variables.  (Table 1-1 also appears as
Table 9-31 in Section 9 of this report.)   The regression analysis
followed a number of preliminary steps before arriving at the
final models.  This included elimination of variables with too
many missing variables (X13/  X16, X18) and variables with nearly
constant values (X8, X9,  X2i/ X23).  Stepwise regression runs
were made to obtain a reduced set of variables which best
predicted leak status or leak rate.  Finally, individual
regression coefficients were examined to ensure statistical
significance.  Sample sizes are shown below for the final model.
                                      Sample
          	Model	       Size
          Leak Status Regression        327
          Leak Status Logistic          380
          Leak Rate Regression           99
                             1-5

-------
  Table  1-1.   Simple Correlation of Leak Status with Explanatory Variables
Explanatory
Variable
X1
X2
X3
X4
X5"
X6
X7
X8
X9
X10
X11
X12
X13
X15
XI 6
X17
X18
XI 9
X20
X21
X23
XZ4
X25
<26
X27
X2B
X29
X30
X31
Meaning
Gas Station
1 Underground tanks
Tank capacity
Average low fill level'3'
(Age of tank)2
Leaded gasoline
Diesel fuel
Aviation fuel
Casohol
Other
Suction pump
Depth buried
Mater level
Tank tested
Years since test
Tank material
Tank lined
Tank coated
Passive cathodic protection
Impressed current catn.
protection
Other protection
Previous tank leak
Previous line leak
Frequency of delivers
Sand fill
Gravel fill
Concrete oad
Packed earth uaa
Oist. to nearest tank or
structure
Definition
1 = Yes; 0 = No
Number at facility
Gallons
As fraction of tank capacity
in 'years;
1 - yes; 0 = No
1 = Yes; 0 = No
1 = Yes; 0 = No
1 = Yes; 0 = No
1 = Yes; 0 : No
1 = Yes; 0 s No
Inches from surface
to top of tank
Inches from surface
to water table^4'
1 if tested after placed
in service; 0 otherwise
Since most recent test
1 = steel: 0 = fiberglass
1 = Yes; 0 = No
1 = Yes; 0 r No
1 = Yes: 0 = No
1 = Yes; 0 = No
1 = yes; 0 = No
1 s Yes: 0 = No
1 = Yes: 0 = No
Number per year
1 = Yes: 0 = No
1 z Yes: 0 = No
! i Yes: 0 i No
1 = Yes; 0 : No
I feet!
Correlation'1) »ith Y1,
Leak status
(1 = Leak; 0 = No Leak)
-.08
.12
.14
-.05
.11
-.26
.24
.13
-.07
.08
.003
.10
-.15
.03
.Ou'
.02
.07
-.01
.10
0
-.08
-.05
.05
-.05
.03
.006
.07
.03
'.04
Correlation'1' with Y2,
Leak rate (gal/Hr),,
among leaking tanks'^'
-.06
.10
.34
-.07
|
-.11
-.08
.07
0
.29
-.12
-.006
-.005
.01
-.21 :
-.09
.02
-.25
.05
0
0
-.04
.23
-.003
-.10
.16
-.09
-.09
-.09
 Pearson's correlation coefficient; Kendall's Tau-6 was also calculated for all Y1 correlations and found to be the same for
 nearly every  variable.

"Using data only from individual  leaking tanks with quantifiable leaks.

 I.e., just before product  is added.

 At time of test.
                                                    1-6

-------
 Table  1-1.  Simple Correlation of Leak Status with  Explanatory Variables
              (Continued)
explanatory
Variable
X32
X33
X34
X35
X36
XT 3
XT4
XT18A
XT19
XT 20
XT 36
X85
X819
XC7
XC8
XF1A
XG20
XG2E
XG2F
XC2G
Meaninq
Interaction: age & material
Interaction: gasohol &
material
Permit to install
Permit to store
Average high fill level*6'
Average fuel delivery
Max. ever stored
Attached to other tank
Tank proximity to water
table
Manway with tank
Not self-installed
Remote gauge
Log of deliveries
Any abandoned tank'5^
t Abandoned tanks
Corrosion prevention equip./
mat.
Trained to check pump
Trained to check line leaks
Trained to cneck leak
prevention
Trained to check leak
monitoring
Definition
(X5 ) (1-X17)
X9 (1-X17)
1 = Yes; 0 s No
1 = res: 0 = No
Aa fraction of tank capacity
in gallons (to one tank)
gallons
1 : Yes; 0 : No
1 : above: 2 = ..irtiaily
above: 3 = below; 4 : other
1 = Ves; 0 = No
1 * Yes; 0 z No
1 s Yes; 0 * No
1 * Yes; 0 = No
1 = Yes; 0 = No
(coded as zero if none)
1 = Yea; 0 s No
1 = Yes; 0 : No
1 = Yes; 0 = So
1 = Yes; 0 - No
1 = Yea; 0 s No
Correlation'1) with Y1,
Leak status
(1 : Leak; 0 s No Leak)
-.03
0
.12
.02
-.06
.15
.11
.22
.13
.19
.12
-.00$
-.03
-.03
• .12
•
-.02
.14
.10
.10
.15
Correlation'1) with Y2,
Leak rate (gal/Hr),
among leaking tanks'2'
-.07
0
.17
.09
-.09
.23
.29
.24
.28
.13
.12
.05
.002
.03
-.09
-.12
.24
.18
.15
.17
5At that facility.

 I.e., Just after product is delivered.
                                        1-7

-------
     B.    Multiple Regression Models
     Two models were constructed:
     [1]  Leak Status Model:
          (among all tanks
          with tightness test)

     [2]  Leak Rate Model:
          (among leaking
          tank systems only)
                         Dependent Variable, Yl
                           1  if leak
                           0  otherwise

                         Dependent Variable, Y2
                           leak rate in gal/hr
     Both models were run using the predictor variables in Table

1-1.  The general form of the model is:
Y - b
                          b2X2
where a few of the variables were interaction terms and the b's

are regression coefficients estimated by a least-squares

procedure.  In addition, a non-linear transformation was used for

one of the X variables.  Age2 was used rather than Age because

data plots suggested a non-linear increase in the percentage of

tanks that leak as a function of age.
                             1-8

-------
     C.   Logistic Regression Model

     For the leak status model, an alternative logistic
regression model was run.  The dependent variable can be
reexpressed as an odds ratio*, in the form:
     rial locr   Probability of Leaking Tank
     L  J   y    Probability of Tight Tank
                  b0 + b^ + b2Xx + ...
     This alternative formulation of Model [1] should more nearly
satisfy the homogeneity of variance assumption for regression.

     The coefficients (b's) for the Logistic Model are estimated
by maximum-likelihood methods rather than least-squares.
IV.  FINAL MULTIVARIATE MODELS

     Using the procedures defined above, linear and logistic
regression models were developed for leak status.  For leak rate,
a separate linear regression model was developed.  The final
models appear below.
 The assumed underlying model for the logistic regression is
 Y = l/[(l + exp (-bQ-biXn - b2X2 ....)].  From this expression
 it can be shown that log [Y/(l - Y) ] - bo + b^, + b2X2 + ...
 In this equation Y is the probability that the tank system leaks
 and 1 - Y is the probability that it does not leak.
                              1-9

-------
     Leak Status  Models


     [1]    Regression Model  :

           Yl  =  .22  + .00019 Xg  -  .25  Xg  +  .0044  X12***  + .18  X2Q


                         * * * *
     [la]   Logistic  Model    :

           log  Probability  of Leak = 1>3  _  >63  x  _>0  17  x
            *  Probability  no Leak               °          x^
                                       -  .38
 All coefficients significant at the 94 percent confidence level
 or better (except coefficient of X20 at 78 percent confidence
 level).
**(Age)2 was used rather than Age because this non-linear
 transformation showed a stronger correlation with leak status.



***The regression model found a + coefficient, but the logistic
 model found a - coefficient.  This may be a case of X12's
 collinearity with other variables.  However, no strong
 collinearities were detected with XT,.  (See Tables 1-2 and 1-3
 in Section V.)  Therefore, the relationship with X12, depth tank
 is buried, is inconclusive based on this mixed result.
****A11 coefficients significant at the 94 percent confidence
 level or better.
                              1-10

-------
     [2]   Leak Rate Model*****:

           Y2 = .91 - .67 X1? - .54 X29 - .0068 X32******

                i   /T *} V     J-  *5 K V
                  . D^ Arpi QA T . ZD AG2E
     The reliability of the model was examined in several ways.
For the regression models, the multiple correlation coefficient,
R, provides some overall measure of the predictive ability of the
model.  These results are shown below.
                 Multiple Correlation
                    Coefficient, R                R2

  Equation      Unadjusted   Adjusted    Unadjusted   Adjusted
    [1]             .30       .29           .093       .081

    [2]             .50       .45            .25         .20
     All coefficients significant at the 97 percent confidence
 level or better.
******This is an interaction term which was included to capture
 the more than additive effect of age and material type together.
                            1-11

-------
     The "adjusted" values of R and R2 adjust for degrees of
freedom in the model and, therefore, provide a better estimate of
how reliably the model might predict leak status and leak rate
for other tank systems beyond the modeling data set.  The R2 term
can be interpreted as the proportion of the variance in Y that
can be explained for by the model.  Thus, the model is able to
account for less than 10 percent of the total variance in leak
status and only about 20 percent the variance in leak rate.

     The reliability of the coefficients of the X's in equations
[1], [la] and [2] were also examined to ensure that the value is
not likely to be a chance occurrence.  The probability that these
coefficients are not chance occurrences is 94 percent or more for
each of 9 of the 10 parameters in these equations.  The remaining
coefficient had a 78 percent probability of being a non-chance
occurrence (i.e., there is a very low probability of the observed
coefficient occurring if its true value were zero).  It should be
noted that these probabilities of non-chance occurrence applies
one variable at a time — i.e., with many variables tried in the
model, the probability of at least one chance selection of a
variable increases.
V.   RELATIONSHIP BETWEEN EXPLANATORY VARIABLES
      fCOLLINEARITY)

     Multicollinearity frequently exists in large data sets.
Pairwise collinearity is one sample form, and is relatively easy
to visualize.  In order to test for such "first order"
collinearity in the models, the correlations between all pairs of
independent or predictor variables  (i.e., X's) were computed.
The results shown in Table 1-2 indicate low pairwise
collinearity, except for X17  (tank material) and X32 = [(1 - tank
                             1-12

-------
Table 1-2.  Collinearity  (intercollelation) of X's in models
A.   Leak status regression and logistic models —
     Pearsons Correlation Coefficient between
     explanatory variables
1
4
X6
X12
X20
XB19
4
1




X6
-.03
1



X12
-.07
-.06
1


X20
-.08
-.12
.07
1

XB19
.10
.002
.09
-.04
1
B.   Leak rate regression model — Pearson's Correlation
     Coefficient between explanatory variables

X17
X29
X32
XT18A
XG2E
X17
1




X29
.09
1



X32
-.80
-.07
1


XT18A
.13
.38
-.10
1

XG2E
.05
.08
-.11
-.02
1
                              1-13

-------
material) x (Age)2] in the leak rate model (correlation of -.80).
The variable,  X32, is an interaction term.  The correlation of
X17 with X32 is close to the correlation of Age2 with -Age2.
Therefore, a large intercorrelation would be expected.

     Table 1-3 shows correlations between variables in the models
and variables not in the models.  (Variables with small
correlations,  less than .20, are not included.)  Any large
correlations could be considered as proxies (or substitutes) for
the model variable with which they are strongly correlated.  For
example, in the leak status model, passive cathodic protection
(X2g) is strongly correlated (correlation coefficient = .62) with
aviation fuel (X8).  Therefore, the apparent increase in the
likelihood of a leak with passive cathodic protection, might be
due, in large measure, to its relationship with aviation fuel
storage.

-------
Table 1-3.   Correlation Between Model  X's and X's  not  in the  Model
 A.
     Leak Status Model
       Model X
                                 Non Model X's
                                       Pearson's
                                      Correlation
                                      Coefficients
                                        (>  -20)
 2                o
X5, (Age of Tanks)'1

Xg, Leaded gasoline

XT2> Depth buried

Xorv, Passive cathodic
  X
   B19'
        deliveries
                          Xy  (Diesel fuel)
X2 (# Underground tanks)
Xg (Aviation fuel)
Xlg (Tank lined)
X29 (Concrete pool)
XT18A (Attached to  other  tank)
     (Manway with tank)
     (Trained to  chek line  leaks)
XQ2F (Trained in  leak protection)
XG2H (Trained in  leak monitoring)

X13 (Water level)
X-ig (Years since  test)
X3/^ (Permit to install)
^35 (Permit to store)
None

-.39

None

 .33
 .62
 .34
 .38
 .29
 .41
 .24
 .27
 .31

 .30
 .34
 .20
 .20
                                     1-15

-------
B.    Leak Rate Model
      Model X
Non Model X's
 Pearson's
Correlation
Coefficients
   (> .20)
Xiy, Tank material Xi (Gas station)
Xy (Diesel fuel)
^11 (Auction pump)
X13 (Water level)
X15 (Tank tested)
Xig (Years since test)
Xlg (Tank lined)
X19 (Tank coated)
X32 (Interaction: Age & material)
X~29> Concrete pad X2 (# Underground tanks)
X- (Average low fill level)
X-^g (Years since test)
X2Q (Passive cathodic protection)
X3Q (Packed earth pad)
X34 (Permit to install)
X3g (Average high fill level)
X.T.-J (Average fuel delivery)
XjlSA (Attached to other tank)
X~20 (Manway with tank)
•^G2H (Drained in leak monitoring)
X32, Interaction: X-Q (Suction pump)
Age & material X-j^ (Tank capacity)
X-^g (Years since test)
Xiy (Tank material)
X18 (Tank lined)
X-ig (Tank coated)
^T18A' Attached to X2 (# underground tanks)
other tank X3 (Tank capacity)
Xy (Diesel fuel)
X13 (Water level)
Xig (Years since test)
X25 (Previous line leak)
X2q (Concrete pad)
X3Q (Packed earth pad)
X34 (Permit to install)
X3r (Permit to store)
X3g (Average high fill level)
Xrp3 (Average fuel delivery)
Xj^ (Maximum ever stored)
X^2Q (Manway with tank)
Xr2n (Trained to check pump)
-.21
.22
.42
-.29
-.28
-.37
-.35
.66
-.80
.46
.24
-.48
.26
-.20
.24
.28
.20
.38
.52
.24
-.29
.26
.49
-.80
.53
-.54
.48
.22
.23
-.30
-.23
.28
.38
-.29
.29
.25
.25
.35
.33
.40
.24
                                    1-16

-------
Leak Rate Model (Continued)
      Model X
                            Non Model X's
 Pearson's
Correlation
Coefficients
   (> -20)
 X,
  G2E'
Trained to
check line
leaks
                          X7
                          X
                           8
(# underground tanks)
(Diesel fuel)
(Aviation fuel)
                          X1Q  (Other fuel)
                          ^16  (Years since test)
                          X2g  (Gravel fill)
                          X.-QQ  (Tank proximity to water  table)
                          X.r.20  (Manway with  tank)
                                (Trained to check pump)
                                (Trained in leak protection)
                          XQ2H  (Trained in leak monitoring)
      .25
     -.25
      .25
      .41
     -.44
      .21
      .39
      .22
      .40
      .89
      .68
                                     1-17

-------
 50273-101
  REPORT DOCUMENTATION
         PAGE
l._ REPORT NO.
            EPA' 560/5-86-013
3. Recipient's Accession No.
 4. Tttto end Subtitle

    Underground Motor Fuel  Storage Tanks:  A National Survey
                                                9. Report Dete
                                                     May,  1986
 7. AMI**.)  Di.tz,  St*ph«m K. ;* Flora, Jairus D., Jr.;6
           Strenio,  Judith F;* Carmen J. Vincent* «t al.
                                                S. Performing Organization R«pt. No.
 *. Performing Onjanliatton Name and AaMress

        •Weatat. Inc., 1650 Research Blvd., Rockvllle, MD 20850
        "Midwest Research Institute, 425 Volker Blvd., Kansas City, M8 (4110
        Battelle Coluabus Division, Washington Operations, 2030 H Street, M,
         Washington, D.C. 20036
        Washington Consulting Group, 1625 Eye Street, NW, Washington D.C. 20006
                                                10. Project/Task/Work Unit No.
                                                     Task 3
                                                11. Contract(C) or Grsnt(G) No.
                                                (O  EPA No. 68-02-4243
                                                ...        and
                                                (G>  EPA No. 68-02-3938
 12. Sponsoring Organization Nam* and Address
              Environmental Protection Agency
              Office of Toxic  Substances
              Exposure Evaluation Division
              401  M  Street, S.W.
 	Washington,  D.C.  20460	
                                                13. Type of Report & Period Covered
                                                    Final Report
                                                    Feb. '84 - May '86
                                                14.
 IS. Supplementary Notes
 1C. Abetract (Limit: 200 word*)

     A  nationally representative  sample of 2,812  establishments  were  interviewed  to
     determine  the  presence  of  underground motor  fuel  storage  tanks.    This  sample
     represented  establishments  in fuel-related  industries  (1,612),  large  establish-
     ments  in  all other industries (600),   and farms  (600).   A  total of 890  of  these
     establishments  were found to have  a total of 2,445  underground motor fuel  storage
     tanks.   Only  19 farms with 34  tanks were  found.   The following national estimates
     were  made:    there  are  796,000  underground motor  fuel storage  tanks  at  326,000
     establishments  in the  Unisted  States  --  158,000  of these  are  on 79,000 farms.   A
     subsample of  218 establishments  was selected  for tank tightness testing,  using a
     modification  of a  commercially available  test.    The  method over-filled  the  tank
     system into a standpipe,  and thus  detected leakage  anywhere  in the  system  of tank
     vessel,  pipes,   lines,  joints,  and  fittings.   Among  the  non-farm  establishments
     tested,  the  following  estimates  were  made:   35  percent (189,000) of tank  systems
     were  judged  to be  leaking  under test  conditions;  the  average  leak  rate  of  those
     systems  with  quantifiable leak rate, adjusted for test pressure, was 0.32  gallons
     per hour; half  the leaks among all systems judged to be leaking were 0.25  gallons
     per hour or less.
 17. Document Analysis  a. Descriptors

    Establishment characteristics,  leaking underground  tanks, survey design,  tank
    characteristics, tank establishments, tank  tightness  testing, underground  motor
    fuel storage tanks

    b. Identifiers/Open-Ended Terms
   e. COSATI Field/Group
 IS. Availability Statement
                                                          19. Security Class (This Report)
                                                            Unclassified
                                                         20. Security Class (This Pace)
                                                                11
                                                           21. No. of Pates
                                                             Vol I: 227
                                                           	Vol II: 330	
                                                           22. Price
(See ANSI-Z39.18)
                                         See Instructions on Reverse
                                                          OPTIONAL FOMM 272 (4-77)
                                                          (Formerly NTIS-3S)
                                                          Department of Commerce

-------