New England Interstate
Water Pollution Control
Commission
Boott Mills South
1OO Foot of John Street
Lowell, Massachusetts
01853-1124
Bulletin 36
November
2OOO
LUST.
A Report On Federal & State Programs To Control Leaking Underground Storage Tanks
THi FUTURE IS iO/fl//Jff
ran MB BITE mR m
by Ann Banner Carpenter
a
ff:"
1 egulators, can you provide prompt answers
1 to public information questions and at the
1 same time be off doing a site inspection? Can
9you easily and quickly access site-specific data
from all across the regulatory landscape from one
source? Can you enter site data onto your database
from the field? If you can't, why not? The technol-
ogy is here, it's getting more user-friendly and
affordable every day, and it can make your workload
a whole heck-of-a-lot more manageable.
One-stop database shopping is already up and
running in some states and gearing up in others.
Geographic Information Systems (GISs) are being
used by regulators in many states as an important
tool in the day-to-day business of environmental
management and decision making. Mapping soft-
ware can help them analyze data, while the Internet
allows them to speed it to those who need the infor-
mation and who would otherwise be calling on the phone. Don't
have good location data? Well, all is not lost—there are several
ways to get it, including highly accurate Global Positioning
Systems (GPSs).
For regulators, the marriage of CIS, GPS, and Internet
access can be a dream come true. For users, it's as easy as shop-
ping on-line—and you don't even need a credit card.
Why Do It?
Ignorance is not bliss. GIS brings to light spatial relation-
ships that may otherwise be unknown because informa-
tion is scattered throughout various regulatory
programs. Regulatory entities often lack the wherewithal
to efficiently and effectively share information on gaso-
line releases and water supply contamination. This lack
of communication can jeopardize water quality.
• continued, on page 2
atea imMnf Am-'' ''Vyfif public
But the Truth" Book Review
•tical Issues for MTBE and Related Oxygenates
^Getting Results, PFP Style
Tank-nically Speaking: Quest for the Perfectly Reliable LLD_
's RP100 for UST System Installation Revised 1
-Oa^eisj>f.F^e^T^EJ|eleas^s-r^antaJ|!arg,^
Neyif OUST Director/Initiatives _
Alaska Tank Explosion Linked to Polyester Coveralls
HQ Update
-------
LUSTLine Bulletin 36
• The Future Is Coming...
continued from page 1
Easy access to environmental
information is important for stake-
holders, such as water suppliers,
responsible parties, consultants, and
the general public. When regulatory
personnel and stakeholders can eas-
ily access the same data, the result
should be less duplication of effort,
better communication, and more
informed decision making. And the
environment is a winner, too.
When a release has been identi-
fied, UST regulators can use CIS
mapping information to identify
such decision-making criteria as
proximity to wellhead protection
areas, drinking water wells and
aquifers, and other receptors. Water
suppliers, land-use planners, and
developers can use GIS to locate
potential sources of contamination
(e.g., UST facilities, dry cleaners,
landfills) near drinking water
sources. Realtors, developers, and
bankers can use GIS to identify,
LUSTLine
Ellen "Frye, Editor
; Rldd Pappo, Layout
f Majxfil Mpreau, Technical Advisor
* i Patricia Ellis, Ph.D., Technical Advisor
Bpnald Poltak, NEIWPCC Executive Director
'; ' i Lyim'DePont, EPA Project Officer
i i Lela Bijou, OUST Liaison
''LUSTLine is a product of the New England
Interstate Water Pollution Control Commis-
Ejivirpnrnental Protection Agency.
11": j ," ' ;i: ' • ii1 : .I '' •"' f
LUSTLyie is issued as a communication ;'
service for the Subtitle IRCRA
Hazardous & Solid Waste Amendments ''
rule promulgation process. !
1" LUSTLine is produced to promote j
information exchange on UST/LUST issues.,
Jhe opinions and information stated herein j
fare those of the authors and do not neces- 4
i sarily reflect the opinions of NEIWPCC. *,
This publication maybe copied.
! Please give credit to NEIWPCC
!!!: NEIWPCC was established by an Act of
I Congress in 1947 and remains the oldest
,{[ agency in the Northeast United States
i'ConCCrnCd with coordination of the multi-
*' ' - -' m&l ia environmental activities
of the states of Connecticut, Maine,
Massachusetts, New Hampshire,
• New York, Rhode Island, ana Vermont.
' ' '• • NEIWPCC
Boott Mills South, 100 Foot of lohn Street
Lowell, MA 01852-1124
Telephone: (978)323-7929
Fax:(978)323-7919
lustline@neLwp_c&_Qrg
i<2» LUSTLIna Is printed on Recycled Paper
among other things, potential envi-
ronmental risks.
California's GIS was developed
in response to a legislative mandate.
With the closure of several Santa
Monica drinking water wells contam-
inated with MTBE in 1997, the state
legislature passed two bills that
charged the State Water Resources
Control Board (SWRCB) with the task
of assessing the feasibility and appro-
priateness of establishing a statewide
environmental database and GIS
mapping system, beginning by con-
ducting pilot projects in the Santa
Monica and Santa Clara Valley areas.
i GIS brings to light spatial
| relationships that may otherwise be
~i unknown because information is
scattered throughout various
ill It " ml, '-' i r . . - c- '- I
regulatory programs.
In response, SWRCB and
Lawrence Livermore National Labo-
ratory (LLNL) staff members worked
together to develop GeoTracker, a
GIS that provides on-line access to
environmental data. GeoTracker is
the interface to the state data ware-
house, the Geographic Environmen-
tal Information and Management
System (GEIMS).
GeoTracker was developed to
provide timely information to coor-
dinate and support state agencies in
protecting public drinking water
sources from motor fuels contamina-
tion. The SWRCB has already real-
ized those benefits and more from
having the system up and running
and on-line (http: / / www.geot-
racker.ecointeractive.com /). The
entire state is now included, and the
agency is planning to add active fuel
sites and other data layers, such as
recharge areas.
At the regional level, EPA
Region 3 is working with its states to
develop a pilot system to obtain loca-
tion data and map wells and regu-
lated facilities, with the goal of GIS
analysis and Internet access. Like
California, Region 3 will begin with
two counties. The region is also con-
sidering developing a risk-indexing
tool to aid in planning.
If your UST/LUST program
hasn't already done so, there are sev-
eral good reasons to start planning
for GIS mapping and linking up with
the Internet:
• Better Manage Site
Cleanup Objectives
When you must respond quickly,
there is no time for an archaeological
dig in the file room. With GIS, you
can view the area where an UST
release has occurred and identify
potential receptors. The more data
available to you, the better informed
your decision and the more timely
your response.
GIS can help you prioritize sites
based on risk. "To help us with our
risk evaluation," explains Art
Shrader, with the South Carolina
UST program, "we put all our GPS
data into a common database, shook
it up twice, and asked the system to
tell us which of our LUST facilities
were within 1,000 feet of a public
water supply. We used that informa-
tion to prioritize sites that had the
highest potential risk so that we
could begin assessments on those
sites first."
Beginning in 1997, the South Car-
olina UST program identified 300
sites that met the high-risk criterion.
Since then, about every six months,
when the analysis is run again, an
additional 10 to 20 sites are added to
the assessment priority list.
• Streamline Permitting,
Enforcement, and Reporting
The Internet can streamline the per-
mit renewal and compliance process
by making it easy to enter, transmit,
and retrieve data. For example, on-
line UST permit forms and reporting
forms are available through Geo-
Tracker. These well-designed forms
help improve the accuracy of infor-
mation collected. Pull-down lists,
check boxes, and select boxes give the
user specific choices so that data are
standardized. Error-checking soft-
ware further quality-controls the val-
ues of information when it is entered
on the Internet forms to assure data
quality. When entering an address,
for example, the system will prompt
the user if the ZIP code is not consis-
tent with the city. Now, one person
can conveniently enter information
and transmit it for storage and use by
others. After the permit is approved,
the data moves into the GEIMS data-
base. Handwriting and in-box delays
are no longer an issue. Costs are
reduced and accuracy is increased.
-------
LUSTLine Bulletin 36
OeoTracker can generate reports for LUFT sites and nearby drinking water wells (e.g., within 1/2 mifei.
JBJ^Ii«!tt3LypJ.gl!g-?
^^ffiTjlEl: ''' 'I'l g'l'^ '',: Vj^t V^^^
EfBLJIFT Sites
njlActive Fuel Sites
leaking Underground Fuel Tank Report for;
Unocal *4918^S4aiiifebraCoiurty)
r Roads
r Roads
l Major Rtls
IBiSurfacB Water
-^•Watersheds
jfeGWlBasins
~i*ConfIning Layers
Leaking Underground Euel Tank Report for;
Unocal #4918 (Santa Clara County)
895 N Sail Antonio Rd
Los Altos, CA 94022
siimated tn he Wigmi 1Q Me:|iteiailea\vdl itjigl
eating Underground Inel Tank Report for!
Unocal J-I91S
-------
LUSTLine Bulletin 36
m The Future Is Coming...
continued from page 3
Vermont's Agency for Natural
Resources distributes public data via
its CIS data warehouse, the Vermont
Center for Geographic Information
(VCGI). Users can select an area on a
state map, zoom in or out, turn avail-
able data layers on or off, and click on
a well or facility symbol to find its
name, address, and more specific
information. The site has links to
important regulatory information
and notices from the water supply
and waste management divisions.
There's a link for water supply opera-
tors and links to contact staff.
The Vermont UST program Web
site provides a listing of upgraded
USTs for purveyors of fuel (required
for delivery of product), a report on
the Petroleum Cleanup Fund, and a
grant application package for
removal of farm and residential
USTs. The screens are appealing and
easy to use. Visit the site at http://
www.anr.state.vt.us / gismaps /.
Another site to visit that is ori-
ented toward public use is
Delaware's Environmental Naviga-
tor. Contaminated site information
includes general site discovery infor-
mation, a site history summary, cont-
aminants, proposed or final plans,
site status, and deed restrictions. This
site is located at http://sirb.awm.
dnrec.state.de.us /.
These Web sites give new mean-
ing to the word "accessibility," and
they suggest an openness and will-
ingness to serve the constituency.
• Facilitate Land Use
Planning, Source Water
Protection, and More
With a wealth of environmental, land
use, and historic information at your
fingertips, there is no excuse for mak-
ing uninformed land use decisions.
CIS information can be a tool for
evaluating risk to public health and
safety and risk to ground and surface
water resources. Accurate knowledge
of existing and potential source loca-
tions can help determine appropriate
setback distances for a multitude of
competing uses, such as buildings,
wells, water supplies, and parking
lots.
• Visualize and Publicize
A picture, or a map, is worth a thou-
sand words. GIS provides you with
an incomparable visual tool, in addi-
tion to its spatial analytic ability.
USGS topographic maps (at various
scales) are used as background to
help you become oriented. Digital air
photos will become available on Geo-
Tracker to enhance site location and
analysis. Want to add the dimension
of time? An additional feature of
GeoTracker is its ability to generate
water quality graphs on-line titiat plot
contaminant values across time, so
users can visualize trends.
All of the Internet mapping sites
we visited promise expansion of ser-
vices in the future, and the mapping
industry is busily developing new
applications to expand GIS capabili-
ties.
The Internet can streamline the
; permit renewal and compliance
process by making it easy to enter,
I •! i r ' i - • .
transmit, and retrieve data.
J
Test Drive a Site Today
Now that you are aware of the bene-
fits, find out how easy it is to use
interactive mapping. Take, for exam-
ple, GeoTracker. (See a demo of all
site features at http: / /www.geot-
racker.ecointeractive.com / gdemo /.)
There are various ways to locate a
site. You can use a business name, an
exact or partial address, or a case ID.
All sites in an area can be found by
entering a particular street, city, or
county. When information is really
sketchy, use a wildcard (*) to let the
computer do the work. Enter North*,
if you don't know whether North is a
street or an avenue. The computer
will display all possibilities.
Once on the map screen, you can
select various data layers or
"themes," and use "tools" to zoom in
or out, or identify a feature, such as a
leaking underground fuel tank
(LUFT) site or a well, selected on the
map. You can also select a well and
search for LUFT sites within 1,000 feet
or 2,640 feet (one-half mile), or vice
versa. This area is called a "buffer"
and is used for proximity analyses.
To get more information about a
particular well at a desired location,
first use the identify function and
then click on the well name. You will
discover links to other screens that
contain several layers of informa-
tion-—well description, public water
system information, number of LUFT
sites estimated to be within one-half
mile. You can delve deeper for loca-
tional information, including lati-
tude/longitude, the way in which the
data were obtained, estimated level
of error, nearest physical address,
elevation of a well, or public water
supply information, including sys-
tem name, class of system, number of
connections, and population served.
For specific LUFT sites, you can
access regulatory history, locational
information, more detailed site and
leak information, and number of
public wells estimated to be within a
specified distance of the site.
Events That Have Enhanced
This Information Revolution
Aside from the obvious advances in
technology and software, a few key
events have contributed to the avail-
ability, accuracy, and usability of GIS
in the environmental field:
• The development of interactive
mapping, an extension of com-
monly used GIS software to
display GIS products over the
Internet. Quality GIS software
is expensive, and this system
provides low-cost access to
valuable geographic informa-
tion. The only requirement is
Internet access and a compati-
ble browser. The user can view
and query GIS data using a
map interface.
• Recognizing that GPS has
become "a global utility" for
navigation, communication,
and emergency response, the
U.S. government abandoned
the intentional degradation of
signals (for security purposes)
in May 2000. GPS users can
now pinpoint locations with
much greater accuracy. This
ability narrows an area down
from the size of a football field
to that of a tennis court.
• The Internet is growing in leaps
and bounds. According to the
Internet Society, use of the
Internet has been doubling
annually since 1988. With an
estimated 150 million users in
1999 worldwide, Internet usage
is expected to reach 300 million
by the end of the year 2000.
-------
LUSTLine Bulletin 36
(Setting Started
There are several ways to gather sup-
port to develop GIS and interactive
mapping. In Idaho, initial work was
under way when a new manager
came on board and wanted the appli-
cation to be up and running in a
month. The focus on this effort
speeded the delivery of new equip-
ment, which made it happen soon
after.
In California, the application was
developed by popular demand.
Stakeholders were brought together
and given a sense of ownership in the
development and advancement of an
Internet mapping application. They
found that while it may help to have
a visionary at the top, not to mention
the interest of the governor and a leg-
islative mandate, the process worked
from the bottom up, rather than from
the top down. Members of various
regulatory agencies, water districts,
and the petroleum industry came
together and were able to visualize
the benefits to their respective orga-
nizations. Ideally, user groups should
remain involved and as broad as pos-
sible. Conference calls are still held
twice each month to address Geo-
Tracker issues.
It helps to form partnerships
with other state and federal agencies
to identify the benefits of sharing
information. Establish common
"themes" and data elements. Every-
one should be reading from the same
script and speaking the same lan-
guage.
How Far Out in Left Field
Are You?
If you plan to get started on GIS,
there is no time like the present to
begin the process of improving loca-
tion data. The first step is often to
"clean" your data. For example,
before you verify the location of
wells, give each well a unique identi-
fication name or number. Older
water systems may have a confusing
array of names, or a single name
applied to multiple wells. In Califor-
nia, a delay in issuance of state well
names left many wells officially
"unnamed."
The biggest barrier to performing
spatial analysis is obtaining accurate
locations. Starting with what you
already have, plot your data and find
out where each location is. If the site
appears to be out in the ocean or in a
different state, you know if s wrong.
You can also run a sample analysis.
LLNL evaluated the location of
drinking water wells in California. To
do this, the lab acquired highly accu-
rate locations for over 1,000 wells and
compared them with locations in the
state database—26 percent of the sites
were within 1,000 feet of the actual
location and 40 percent were within a
half a mile.
Improving locations for LUST
and UST facilities can be relatively
easy. Permitted facilities have
addresses, but water wells generally
do not. Thus different approaches are
needed. Geocoding software is avail-
able that can match up street
addresses with latitudes and longi-
tudes.
|E// helps to form partnerships with
j^1—"" *" n r "
Bother state and federal agencies to
I identify the benefits of sharing
'T*^ ^ JJ.*»r _ \
information. Everyone should he
rom the same script and
'$.-, speaking the same language.
1
I
In California, LLNL found that
84 percent of commercial facility
addresses produced reasonably accu-
rate latitudes and longitudes with an
estimated median error of 396 feet.
According to Dr. Anne Happel of
LLNL, the batch approach and
address matching are most cost- and
time-efficient for regulated facilities.
First, the addresses are standardized
and verified using U.S. Postal Service
software, then they are matched up
to locations using commercially
available geocoding services. "If it is
a facility owned by a corporate entity,
make sure you have the address of
the facility with the tanks, and not the
corporate office," cautions Happel.
When LLNL compared verified
well location data for over 1,200 wells
with locations in the state database,
the median error was estimated to be
2,251 feet. Obviously, when a buffer
or radius of 1,000 feet is used to
assess vulnerability, a high level of
accuracy is needed for well locations.
Accurate well locations are more dif-
ficult to obtain, but with GPS technol-
ogy improving, and becoming less
expensive and more available, help is
on the way. When data are directly
downloaded from a GPS unit into the
system, it allows for "single entry,
multiple use." It saves time and
reduces errors. Old-fashioned map-
ping onto assessor parcel maps and
digitizing is also a possibility. Cali-
fornia's DHS is committed to getting
location information to within a 25-
meter accuracy by 2003.
Security Issues
Of course, with thousands of wells
and facilities and their related data,
you can never guarantee 100 percent
accuracy and availability. All Internet
mapping sites include disclaimers
concerning the accuracy of data. A
disclaimer describes the purpose of
the application—for example, to pro-
vide a visual display of statewide or
local data from various sources. The
disclaimer directs the-user to the state
agency staff to be sure of obtaining
complete, accurate, and up-to-date
information.
The disclaimer also points out
particular requirements such as
screen resolution that are necessary
to properly view the features and
indicates that the sites are dynamic
and subject to change as more maps
are created. Some maps or databases
may be temporarily unavailable
because of updating. For the applica-
tions that allow the user to input
data, such as GeoTracker, passwords
are issued to authorized users.
Interim Steps—Your
Warm-Up Act
A number of states have plans in the
works to develop or improve their
GIS and Internet access. Virginia
plans Internet access to its geodata,
and at present has put in place an
interim step. It is a "very friendly"
GIS station located at each of the six
regional offices and headquarters
that is available to the public. The
geodata contains all 15,000 petroleum
release sites plus historical data.
Users can select counties, "pan"
around for a specific site, zoom in or
out, or identify sites within a given
radius. They can also look for a spe-
cific petroleum complaint or a site
name and address, and determine
whether the complaint is "opened or
dosed." For more detailed information,
• continued on page 10
-------
LUSTLinc Bulletin 36
SiygeKafes;
The Subsurface Fate of Ethanol
A Look at the Emerging Oxygenate
Alternative to MTBE
Susan E. Powers and David Rice
Jn response to the widespread contamination caused by MTBE-blended reformulated gasoline (RFC), legislative initiatives in
several states and at the federal level have phased out, or are trying to phase out, the use of MTBE as a gasoline oxygenate.
Ethanol is currently the most likely gasoline oxygenate alternative to MTBE. This potential for increased use ofethanol has
been most loidely acknowledged by California. The California Executive Order requiring the phase-out of MTBE also required that
an analysis of the fate, transport, and health risks associated with the use ofethanol as a gasoline oxygenate be conducted. It is clear
that California and many other states now recognize the need to understand the environmental fate of gasoline oxygenates before
any policy decisions are made regarding their widespread adoption.
The material included in this article begins to summarize findings of the study completed for California. The full report is avail-
able at httir./fwwio-erd.llnl.vov/ethanol/. The primary physical and biological properties ofethanol that have implications for ground-
water contamination are identified in this article. Future articles will focus on the uncertainties in our understanding and research
required to make sound policy decisions. (For an overview on ethanol, see LUSTLine #32, June 1999, "With the Possible Phase-Out
of MTBE, Wliat Do We Knozu About Ethanol?" by Bruce Bauman.)
Ethanol Use in Gasoline
Ethanol is currently used in oxy-
genated gasoline, albeit not as widely
as MTBE. Meeting the federal oxygen
requirement would call for 8 percent
(by volume) ethanol for oxyfuel and
6 percent for RFC. However, because
of a 54 cents per gallon of ethanol
used federal subsidy, the blending of
ethanol at 10 percent with gasoline is
popular. Several states provide addi-
tional subsidies for ethanol produced
and used in their own states. (See
sidebar on page 9.)
In Nebraska, 21 percent of all
motor fuel sold contains 10 percent
ethanol. At present, 60 percent of
gasoline sold in Illinois, and 90 per-
cent of gasoline sold in the Chicago
area, contains 10 percent ethanol.
Throughout the country, U.S. con-
sumers use more than 56 million
cubic meters (15 billion gallons) of
ethanol-blended gasoline each year.
The ethanol used for fuel is made
primarily from grains or other
renewable agricultural and forestry
feedstocks. One advantage of ethanol
is that it can be made from liquid or
solid waste, such as wood by-prod-
ucts, or agricultural waste, such as
rice straw. The ethanol used for fuel
is a high-octane, water-free alcohol
produced from the fermentation of
sugar or converted starch.
Unlike most gasoline hydrocar-
bons, ethanol loves water. These two
fluids are completely miscible. This
property has important implications,
both for the manner in which we use
ethanol-blended RFC and its envi-
ronmental impacts. For example, it is
difficult to distribute ethanol-
blended RFC because of the propen-
sity for water to absorb into the
gasoline. Gasoline distribution termi-
nals receive gasoline and ethanol sep-
arately; the two are then mixed as
they are pumped into the tanker
truck for delivery to a gasoline sta-
tion (Figure 1). With current ethanol
production capabilities, most of the
ethanol used would be produced in
Ethanol Bulk
Storage
Gasoline Bulk
^torage _ _^.
The life cycle of ethanol-RFG has some components that are significantly
different than other gasoline formulations. j;
-------
LUSTLine Bulletin 36
the Midwest and shipped by rail or
marine cargo and then by rail or
truck delivery to a final destination
terminal.
Given the nature of the ethanol
life cycle, the most likely spill scenar-
ios associated with the use of ethanol
as a primary fuel oxygenate would
involve leaks of ethanol-blended
RFG from tanker accidents or USTs
or spills of neat ethanol at distribu-
tion terminals.
CH3-O-C-CH3
CH3
Ethanol
CH3-CH2-OH
Hexane
CH3-(CH2)4-CH3
Molecular structure df gasoline
oxygenates and hexane, a representative
component of gasoline. The electrons
(") on the oxygen atoms make the
MTBE and ethanol molecules polar and,
therefore, much more water-loving
(hydrophilic) than hexane or other
gasoline hydrocarbons.
Potential Impacts of Ethanol
on Groundwater Quality
Many of the chemical properties and,
therefore, the environmental trans-
port properties of ethanol are similar
to those of MTBE. The chemical
structures of these two molecules can
help us understand their environ-
mental fate (Figure 2). The oxygen
atom in both MTBE and ethanol
makes these molecules more polar
than other petroleum hydrocarbons.
This polarity is the reason that they
"love" water, a property described
by the term "hydrophilic." Thus both
MTBE and ethanol have a relatively
high solubility in water and high
mobility in the subsurface relative to
more hydrophobic (water-hating)
gasoline constituents such as hexane.
The key difference in the envi-
ronmental fate of these two oxy-
genates is caused by the tert-butyl
group on the MTBE molecule. This
branched structure makes biodegra-
dation of MTBE very difficult. Thus,
while the ethanol molecule can be
degraded and naturally attenuated in
the subsurface, the MTBE molecule is
not effectively attenuated, allowing it
to travel significant distances from a
spill site. The net effect of these prop-
erties results in very different envi-
ronmental impacts associated with
these two gasoline oxygenates:
• MTBE deleteriously affects
groundwater quality for ex-
tended periods.
• Ethanol is not expected to be a
significant groundwater conta-
minant for extended periods.
Although we do not expect
ethanol to contaminate groundwater
as much as MTBE, it is possible that
its presence in gasoline and ground-
water near a spill site will affect
groundwater concentrations of other
constituents from the gasoline—for
example, benzene, a known carcino-
gen.
When considering the ultimate
risk of any contaminant in the sub-
surface, we are most interested in
potential groundwater concentra-
tions at some receptor point down-
gradient of a spill site (Figure 3).
Numerous processes can affect these
concentrations following the spill of a
petroleum product. The presence of a
hydrophilic compound in the gaso-
line adds additional processes we
have not had to consider previously.
Research conducted so far has identi-
fied the following important issues:
• Sufficient amounts of ethanol
can decrease the interracial ten-
sion of the gasoline, potentially
inducing greater lateral spread-
ing of the gasoline within the
capillary fringe.
• The presence of ethanol in
water can create a cosolvent
effect, increasing concentra-
tions of other contaminants.
• All of the oxygen (and other
electron acceptors, such as
nitrate, iron, and sulfate) and
nutrients needed for the
biodegradation of benzene can
be consumed as ethanol is
biodegraded.
Insufficient work has been com-
pleted to date to allow us to under-
stand the net effect of the first issue.
Both of the other two processes may
result in an increase in the concentra-
tion of hydrophobic compounds,
such as benzene, and an increase in
the distance these compounds would
travel from a spill site before attenu-
ating processes reduce their concen-
trations. Note that there are no
known field studies of the behavior
of ethanol and BTEX (benzene,
toluene, ethyl benzene and xylene)
from an UST release. Efforts are
under way to identify sites.where
ethanol-blended gasolines have been
used and presumably been released
from an UST.
• continued on page 8
Flow through unsaturated zone
reading as a gasoline pool
ort and biodegradation
(affected by rapid ethanol biodegradation)
Predicting groundwater quality and health risks at a downgradient receptor location requires
that the net effects of several fate and transport processes be understood. Ethanol in gasoline
can potentially increase concentrations of constituents dissolving from the gasoline source
and decrease rates of BTEX biodegradation.
-------
LUSTLine Bulletin 36
I Ethanol/rom page 7
Cosolvency Issues
When mixed with water in the labo-
ratory, ethanol quickly and com-
pletely transfers from ethanol-
blended RFG into the aqueous phase.
Depending on the volume fraction of
ethanol in the gasoline and the rela-
tive volumes of gasoline and water
that are mixed, it is possible that the
resulting aqueous-phase concentra-
tions of ethanol will be high enough
to increase aqueous-phase concentra-
tions of other hydrophobic com-
pounds such as benzene.
The addition of ethanol to gaso-
line affects these concentrations by
the "cosolvent effect." Cosolvency
describes the reduction of the polar-
ity of the aqueous phase when high
concentrations of organic com-
pounds, such as alcohols, are present.
Essentially, the ethanol molecules
add organic material to the aqueous
phase, making it more attractive to
other organic molecules.
Figure 4 illustrates the approxi-
mate logarithmic increase in BTEX
concentrations with increasing
ethanol concentrations. It has been
predicted that the volume fraction of
the dissolved ethanol in groundwater
systems will be less than or equal to
15 percent (i.e., 150,000 mg/L). At
these relatively low ethanol volume
fractions, BTEX concentrations in the
aqueous phase near a gasoline spill
are predicted to increase by approxi-
mately 20 to 50 percent.
The smallest percentage increase
(smallest slope) was observed for
benzene, the least hydrophobic of the
BTEX compounds. Therefore, it is
unlikely that cosolvent-related
increases in BTEX concentrations will
be significant at the field scale follow-
ing spills of ethanol-blended RFC.
Spills of neat ethanol at-a bulk termi-
nal, however, could result in very
high ethanol concentrations in a
localized area. This problem could
cause a much more significant—pos-
sibly an order of magnitude—
increase in BTEX concentrations if the
soil was previously contaminated
with a petroleum product. Field stud-
ies are in progress that should help
clarify our understanding of cosolu-
bility issues.
Biodegradation Issues
Ethanol can be degraded either aero-
bically (in the presence of oxygen) or
anaerobically (in the absence of oxy-
gen) at faster rates than can other
gasoline constituents (e.g., benzene,
MTBE). In laboratory studies with
~ 10,000
1,000
CD
u
d
u
-------
LUSTLine Bulletin 36
How Much Ethanol Is in That Gallon of RFG?
Did you know that no real formal definition exists for how much
ethanol you need in gasoline to call it gasohol? It's frequently
assumed to be 10 percent, but there is no legal definition. How
much of an oxygenate might we expect in a gallon of gasoline? To meet the
2.0 and 2.7 weight percent oxygen requirements of reformulated gasoline
(RFG) and oxygenated gasolines (oxyfuel), respectively, requires 10.8 and
14.8 volume percent MTBE. Technically, the use of ethanol in RFG requires
only 5.7 percent (7.8 percent for oxyfuel), inasmuch as ethanol has a higher
oxygen content than MTBE. In the real world, ethanol is almost always
blended into gasoline to 10 percent volume. The reason? MONEY!
There is a federal tax break for the use of ethanol in gasoline—54 cents
per gallon of ethanol used. The ethanol tax incentive has three tiers reflecting
ethanol/gasoline blends at volumes of 10 percent, 7.7 percent, and
5.7 percent.
The Specifics
The 18.4 cents per gallon federal excise tax on gasoline is used to fund the
Federal Highway Trust Fund, the primary source of federal dollars used for
road-building projects. The ethanol tax incentive is highest when ethanol is
blended at a 10 percent level. When it is blended into gasoline at that level,
each gallon of gasoline receives a 5.40 cents exemption from the federal
excise tax. (Each gallon of gasoline contains 0.1 gallon of ethanol, so the tax
exemption is 0.1 x 54 cents, or 5.4 cents.) For ethanol blended into gasoline
at 7.7 percent, each gallon of gasoline receives a 4.16 cents exemption. (In
this case, each gallon of gasoline contains 0.077 gallon of ethanol, so the
exemption is 0.077 x 54 cents, or 4.16 cents.) For 5.7 percent ethanol/gaso-
line blends, the exemption is 3.08 cents. (You do the math for this one.)
The actual economic calculus that a gasoline blender would use is fairly
complex. To grossly oversimplify things, it is usually most profitable to use
the 10 percent blend. It basically means that you can sell your gasoline at
the pump for the same price as the nongasohol station across the street but
get a 5.4 cents per gallon "rebate" from the federal government. That usu-
ally makes more "cents" than blending at 7.7 percent and getting only a 4.16
cent rebate.
A quick check of data from a recent national gasoline survey confirms
this assumption. Nationally, the average ethanol concentration in alcohol-
blended fuels during the summer of 1999 and the winter of 1999-2000 was
about 10.1 percent, with a minimum value of 9.5 percent and a maximum of
11.0 percent.
Several states provide additional subsidies for ethanol produced and
used in their own states. The take-home message here is that in the majority
of cases, if a gasoline contains ethanol, be it an RFG or oxyfuel, it is probably
present at about 10 percent by volume. •
revealed that they indeed know that
they have had spills of gasohol, but
they cannot be tracked because data-
bases archiving spill histories gener-
ally do not identify the type of
gasoline. The lack of any regulations
requiring groundwater to be tested
for ethanol content also contributes to
the scarcity of data. The paucity of
historical data confounds efforts to
understand and predict the effects of
ethanol on groundwater quality.
We should learn from lessons
associated with MTBE—namely, that
the ubiquity of MTBE in the environ-
ment was not understood until we
started to look for it. Analytical tech-
niques for assessing ethanol concen-
trations are now available. It is time
to start adding this analyte to routine
monitoring at gasoline-impacted
sites, especially in the Midwest, Cali-
fornia, and other locations where
ethanol is already in use.
At this point, no extensive mod-
eling studies are available to predict
the overall fate of ethanol and BTEX
in an aquifer following a spill of
ethanol-blended RFG. Various
researchers have conducted model-
ing studies but always with limiting
assumptions about the significance of
cosolvency or biodegradation mecha-
nisms. For example, many of the
models assume that BTEX biodegra-
dation does not occur in areas where
ethanol is present at concentrations
above some threshold value. Regard-
less of the assumptions employed,
the conclusions drawn from the vari-
ety of modeling studies'suggest that
benzene is likely to travel farther
from ethanol-blended RFG release
sites. Predictions generally show that
benzene plumes from ethanol-
blended gasoline could be from 20 to
150 percent longer than those from
nonoxygenated gasoline.
As states and the federal govern-
ment ponder the increase in use of
ethanol as an oxygenate and a bio-
mass fuel, the potential environmen-
tal benefits and costs associated with
this oxygenate must be weighed and
compared with other economic and
social implications. (Ideally, the poli-
tics won't overshadow the science.)
Conclusions drawn based on the lit-
erature review completed for Califor-
nia suggest that the effects on
groundwater resources associated
with the use of ethanol will be less
severe and more manageable than
those associated with the use of
MTBE.H
Susan Powers is with the Depart-
ment of Civil and Environmental
Engineering at Clarkson University in
Potsdam, NY. She can be reached at
sep@clarkson.edu. David Rice is with
the Lawrence Livermore National Lab-
oratory, Environmental Protection
Department, in Livermore, CA. He can
be reached at rice4@llnl.gov.
This article was prepared from the
review completed for the California
Environmental Policy Council by
Lawrence Livermore National Labora-
. tory under the auspices of the U.S.
Department of Energy (Contract W-
7405-Eng-48). Much of the work cited
in this report was conducted with addi-
tional financial support from the U.S.
EPA Science to Achieve Results
(STAR) program in the National Cen-
ter for Environmental Research and
Quality Assurance (NCERQA)
(grant number R821114).
-------
LUSTLinc Bulletin 36
DTHER OXYGENATES
In NEIWPCC's MTBE survey, states were asked if they analyze for any of the following oxy-
genates: ethanol, TBA, TAME, ETBE, DIPE, or any others. The overwhelming majority of states
Indicated that they do not analyze for any of these substances. Of the states that do:
• Two states analyze for ethanol "most of the time" and eight states "occasionally."
South Carolina and Nevada indicated that they have sites where it has been detected.
• Six states analyze for TBA most of the time, and 15 do occasionally. Six of these states
have multiple sites where TBA has been detected.
• Five states analyze for TAME most of the time, and 12 do occasionally. Four indicated
that they have a few to several sites where TAME has been detected.
• Four states analyze for ETBE most of the time, and 10 do occasionally. Three states
have sites where ETBE has been detected; Iowa had 33.
• Six states analyze for DIPE most of the time, and 9 do occasionally. Four have multiple
sites where it has been detected.
• Missouri monitors for EDB occasionally and South Carolina monitors for ETBA, TAA,
and TBF occasionally.
Only five states monitor for ethanol; two others say they do sometimes. Seven states
indicated that they have ethanol-contaminated LUST sites. Mot surprisingly, 36 states indi-
cated that they did not know whether they have ethanol-contaminated LUST sites. Kansas has
identified 70 sites that have sold a 10 percent ethanol/gasoline mix. The state is in the process
of analyzing these sites. •
For Moiie Information About Ethanol...
Blue Ribbon Panel on Oxygenates in Gasoline. Achieving clean air and dean water:
the report of the Blue Ribbon Panel on Oxygenates in Gasoline; EPA 420-R-99-
021; U.S. Government Printing Office: Washington DQ 1999
(http: / / www.epa.gov / otaq/ consumer / fuels / oxypanel / r99021 .pdf ).
Corseuil, H. X., Hunt, C, dos Santos Ferreira, R., and Alvarez, P. J. J. 1998. The influ-
ence of the gasoline oxygenate ethanol on aerobic and anaerobic BTX biodegra-
dation. Wat. Res. 32 (7), 2065.
Corseuil, H. X., and Alvarez, P. J. J. 1996. Natural bioremediation perspective for
BTX contaminated groundwater in Brazil. Water Sci. & Technol. 35, 9-16.
Heermann, S. E., and Powers, S. E. 1998. Modeling the partitioning of BTEX in
water-reformulated gasoline systems containing ethanol. J. Contain. Hydrol.
34(4), 315.
Hunt, C. S., Alvarez, P. J. J., dos Santos Ferreira, R., and Corseuil, H. X. 1997. Effect
of athanol on aerobic BTEX degradation. In: B. C. Alleman and A. L. Leeson
(eds.), In situ and onsite bioremediation, 4 (1). Batelle Press, Columbus, OH, pp.
49-54.
Rice, D. W. (and others). Health and environmental assessment of the use of ethanol
as a fuel oxygenate. Volume IV — potential ground and surface water impacts.
Published by the state of California, UCRL-AR-135949, 1999 (http://www-
http:/ / www.afdc.doe.gov/. Alternative fuels data center (USDOE). Biodiesel /
ethanol general information. U.S. government fleet information. Alternative fuel
vehicles.
hjtp:/ / www.cleanfuels.net/. Oxygenated Fuels Association.
http://www.ethanolrfa.org/. Renewable Fuels Association.
http: / /_w_ww.greenfuels.org / index.htm. Canadian Green Fuels Association.
hjtp:/ / www.ethanol-gec.org/. Governor's Ethanol Coalition.
http://5jayw.ethanol.org/main.html. American coalition for ethanol. Information
on ethanol, MTBE, publications, press releases, and more. Also links to other
organizations.
• The Future Is Coming...
continued from page 5
such as the extent of contamination
or health risks, they are encouraged
to check the site characterization
reports, which are available to the
public at specified regional offices. To
protect the privacy of responsible
parties, financial information related
to the ability to pay for cleanup is
kept separately.
New Hampshire uses a similar
approach. The Site Remediation Pro-
grams for the Waste Management
Division sponsor a GIS terminal at
their headquarters office for the pub-
lic, consultants, and other parties to
identify potential and existing conta-
mination sources and water sources.
They charge small fees to print maps
and reports. Reports can be saved to
a disk, and a feature to save maps
will be incorporated later. In the
meantime, source and receptor data-
bases with addresses are available
over the Internet on a system called
"Onestop" (http: / / www.des.state.
nh.us/onestop/). That's the warm-
up act until a New Hampshire Inter-
net GIS site is on a Web site near
you—soon.
It Just Makes Sense
The reason for getting your GIS pro-
gram up and running on the Internet
is not because everyone else doing it,
but because spatial analysis can
improve your agency's ability to pro-
tect the environment through better
information and sharing of that infor-
mation. The public can get better
(graphical) answers to questions than
they can over the phone, and you can
spend your time doing site inspec-
tions and catching the leak before it
reaches the well.
In the next issue of LUSTLine,
we'll continue our technology theme
with a story about a new integrated
UST inspection system, developed by
EPA Region 2 and New York, that
uses GPS to enter information into
the database in the field...and much
more. Stay tuned. •
Ann Carpenter is a former EPA
employee who now teaches geography
at a community college in Massachu-
setts and writes on a free-lance basis.
10
-------
LUSTLine Bulletin 36
"Nothing But the Trutfi" Uses
MTBE as Its Weapon of Choice
by Patricia Ellis
&
list came the Woburn case,
followed by the book A Civil
Action, followed by the movie
starring John Travolta. Next came
Erin Brockovich, starring Julia
Roberts, based on a case involving
Pacific Gas and Electric in Hinkley,
California, where hexavalent
chromium had been used as an
antif ouling agent in cooling towers
at a compressor station. With
MTBE playing such a major role in
my life these days, I'd been trying
to decide which of the major
MTBE impacts or which of the
many MTBE legal cases in the
works would make a good book or
movie. Now mystery writer John
Lescroart has presented us
with
Oh, wait—that book is fiction.
Nobody really dumped MTBE in a
reservoir as an act of eco-terrorism.
So far, the impacts to water
resources in real life have all been
"accidental"—accidental tank
leaks, accidental pipeline leaks,
accidental car accidents, and other
assorted accidental spills. The
gasoline leaks that cost Santa Mon-
ica more than half of its water sup-
ply, the problems that caused the
shut down 13 of 35 wells in South
Lake Tahoe, or the spill from the
Explorer pipeline that took out one
of Dallas's water supply reservoirs
for months? These incidents
weren't fiction like Lescroart's
book.
Nothing But the Truth is mys-
tery writer John Lescroart's
eleventh novel, including two that
have made the New York Times
best-seller list. The plot involves
petroleum, MTBE, evil efhanol lob-
byists, gubernatorial politics,
romantic tangles (sorry, no steamy
sex scenes), grand juries, and
lawyers, including an ambitious
prosecutor out to make a name for
himself. What more could you
want?
The book involves the death of
Bree Beaumont, a scientist who had
believed that MTBE was wonderful
for improving air quality and had
been a consultant for the Western
States Petroleum Association and
later for Calco Oil. Shortly before her
death, she had changed sides in the
volatile wars over the gasoline addi-
tive, after coming to believe that
MTBE was a cancer-causing additive
that was seeping into California's
groundwater in alarming amounts.
In the story, MTBE was big busi-
ness for the oil companies—they
were making $3 billion per year on it.
On the other hand, the ethanol indus-
try profits were a mere $45 million in
the United States just before Califor-
nia announced its MTBE ban. Now
the ethanol market was expected to
multiply exponentially. Stakes were
high.
During pre-election speeches,
gubernatorial candidate Kerry calls
for an immediate moratorium on
MTBE use. "There is no reason to tol-
erate even for one more moment this
dangerous and insoluble toxin in our
gasoline where there is an environ-
mentally safe and effective substitute
so readily available, and by this I
mean ethanol."
We need to point out at this junc-
ture that MTBE is not "insoluble." If
it were, MTBE would hardly be a
groundwater issue. While it isn't
completely miscible (as is ethanol), it
does have a rather high solubility
(45,000 ppb) compared with benzene
(1,707 ppb).
Back to the story—candidate
Kerry's opponent, speaking about the
act of eco-terrorism, says "If s not the
MTBE that has caused this terrible
crisis any more than it is guns that
kill people. People kill people, and
people—criminals—have poisoned
the San Francisco water supply.
Gasoline without additives would
have produced the same effect, and
"Nothing But the Truth" by John Lescroart,
Delacorte Press, January 2000.
no one is talking about making
gasoline illegal."
Well I can't think how many
times I heard during the Blue Rib-
bon Panel hearings, "It's not the
MTBE, if s the tanks. If 'you people'
would keep the tanks from leaking,
the MTBE wouldn't be a problem!"
A visit to Bree's office sounded
like a visit to my office—"propa-
ganda by the armload on...every
imaginable side of the additive
issue. Legislative reports, news
clippings, executive summaries
from various thinktanks, media
alerts. MTBE, ethanol, reformu-
lated gasoline." Did this get Bree
killed?
As I read the book— with the
idea of writing a review of it—I was
"sticky-noting" the pages where
MTBE was mentioned. Lescroart
had a pretty good handle on the
issues and facts involving MTBE,
putting the proper words into the
mouths of each of his characters.
Now, who can we cast as... •
Pat Ellis is a hydrologist with the
Delaware DNREC LIST Branch and
was a member ofEPA's Blue
Ribbon Panel.
11
-------
LUSTUne Bulletin 36
Diving Plumes and Vertical
Migration at Petroleum
Hydrocarbon Release Sites
by James W. Weaver and John T. Wilson
Petroleum hydrocarbons are
mostly less dense than water.
So they should float or at least
hang around the water table, right?
Not so fast. There are some fairly
common situations where we would
expect a plume of petroleum hydro-
carbons to move vertically into the
aquifer—as a result of water table
drawdown associated with pumping
from water supply wells, smearing
of contaminants due to water table
fluctuation and site investigation
activities, and movement of water
through preferential flow paths in
heterogeneous environments. But
even in the absence of any of these
circumstances, a plume may still
move downward, or "dive," into an
aquifer.
This diving situation occurs
when groundwater recharge enters
the top of a shallow water table
aquifer. Once in the aquifer, this
water begins to move in the direction
of groundwater flow. Because the
recharge water is entering the aquifer
from above, it can push contaminant
plumes downward. The amount that
a plume "dives" depends on the
amount of recharge water entering
the system and the relative contribu-
tion this additional water makes to
flow in the aquifer.
We expect that such diving
plume scenarios may occur in the
wetter parts of the country, but even
in dry climates, recharge-driven div-
ing can occur because of irrigation,
leaking water or sewer pipes, or
recharge from ephemeral surface
water features. In either case, plume
diving depends on the localized pat-
tern of recharge, the flow rate in the
aquifer, and the distribution of conta-
minants—as shown in the following
East Patchogue, New York, example.
East Patchogue, New York
A gasoline release at an East
Patchogue, New York, UST facility
_
created large BTEX and
MTBE plumes. The plumes
were detected because a
private water supply well,
located 4,000 feet down-
gradient from the source,
was in their path. The
well screen was about 50
feet below the water
table, where much of the
MTBE mass was located.
The site investigation started at this
point and went upgradient to iden-
tify the source.
Because of the importance of the
aquifer for drinking water supply,
New York undertook an extensive
investigation of the site, including
vertical characterization of the
plumes. Multilevel samplers with 6-
inch screens at 5-foot intervals were
used. A resulting vertical section
through the plume showed that
BTEX and MTBE tended to dive into
the aquifer with distance from the
source. (See Figure 1.) It was further
noted that a significant amount of
diving occurred as the BTEX plumes
passed under a gravel pit.
By studying the well
logs and performing a
detailed hydraulic charac-
terization of the aquifer with
a borehole flowmeter, investi-
gators ruled out vertical migration
controlled by stratigraphy, because
the hydraulic conductivities varied
by less than a factor of 2 over the
aquifer. This left recharge as the most
likely explanation for the plume div-
ing. The model described in the side-
bar on page 14 was used to simulate
the site and provided additional evi-
dence that recharge was the cause of
the diving.
This Patchogue example sheds
light not only on how recharge
pushes the plume downward, but
also on what happens when water
discharges from aquifers. Where
water comes up at discharge points,
so will the contaminants—along
streams, rivers, lakes, or the ocean.
40 -
0 -
S -40 H
CO
UJ
-80 -
-120
Gravel Pit
Total Xylenes
6,000
-I 1 1
4,000 2,000
Distance (ft)
A vertical cross section through the MTBE, benzene, and total xylene plumes at East
Patchogue, New York. The gasoline source is located at the right-hand edge of the sections, \
and flow is to the left. Each of the plumes dives into the aquifer with transport in the aquifer.
-------
LUSTLine Bulletin 36
The ocean is the expected destination
of the MTBE plume at East
Patchogue, where the groundwater
flow system discharges into Great
South Bay, adjacent to the southern
shore of Long Island. The groundwa-
ter and contaminants move upward
as they approach the discharge point
at the bottom of the bay.
Consequences of Missing
the Dive
What about the consequences of a
diving plume, or more to the point,
the consequences of missing a diving
plume? We averaged the East
Patchogue data set to show how the
plume would appear if sampled only
from long-screened wells. The data
were averaged over the top 10 feet of
the aquifer to simulate 20-foot well
screens—10 feet in and 10 feet out of
the aquifer.
The graphs in Figure 2 show two
sets of concentrations plotted along
the length of the plumes. The first
data set (circles) shows the maximum
concentrations from the multilevel
samplers. This set represents the
maximum concentration measured in
each sampler, regardless of depth, at
each location along the plume. It is
intended to be a reference to show
the extent of contamination on the x-
y plot.
The second concentrations (dia-
monds) are the values for the simu-
lated 10-foot screens. For these, the
MTBE concentrations all fall below
New York State's threshold of 10
/ig/L. With only these data we
would have concluded that there was
no MTBE plume at this site. The max-
imum concentrations, however, indi-
cate a significant MTBE plume in the
downgradient portion of the aquifer.
The simulated long-screened
data show that the benzene plume
appears to be shortened to about one-
third its actual length. This effect
occurs because plume diving pushes
the benzene plume out of the bottom
of the sampling network. Along the
way, the concentrations appear to
decrease, because clean and contami-
nated water mix in the well. This
mixing results in diluted samples and
lower concentrations.
Interestingly, the long-screened
data also show that the total xylenes
and benzene plumes appear to be the
same length. Here, because of sorp-
Maximum
Average of Top 10 Feel
- 6,000
- 4,000
-2,000
0
- 8,000
- 2,000
-1,000
p 20,000
-16,000
- 12,000
8,000
-4.000
5,000 4,000 3,000 2,000 1,000
Distance from Source (ft)
I Consequences of sampling only the top 10-feet of the aquifer at East Patchogue. The MTBE \
plume disappears (top), the benzene plume is shortened by two-thirds, and the total xylenes
plume appears at the same length, because it does not reach the gravel pit where diving dom-
| inates the contaminant distribution.
tion, the total xylenes did not travel
far enough in the aquifer to drop out
of the monitoring network, and there
was no apparent shortening of the
plume. Nevertheless, these two cont-
aminant distributions hint at a sam-
pling problem. Our expectation is
that there should be separation of
benzene and total xylenes caused by
sorption. In this case, the expected
chromatographic separation has been
negated by the monitoring network.
Rethinking Our Assumptions
Are plumes longer than we think
they are? The short answer is yes!
The reason we think that they are
shorter is that most LUST site moni-
toring well networks do not ade-
quately delineate contaminant
plumes in three dimensions.
"Conventional" monitoring
wells are primarily designed to moni-
tor for the presence of free product
floating on the water table. To accom-
plish this task, conventional wells are
constructed with relatively long
screens that bisect the water table.
This approach is meant to allow for
seasonal fluctuations in water table
elevation in the hope that the screen
will extend below the lowest low-
water elevation and just above the
highest high-water elevation.
Also, many monitoring networks
consist of relatively few wells, most
of which are located on the LUST site
property. We've seen that such net-
works are not well suited for deter-
mining the true extent of a plume,
nor can they provide accurate infor-
mation about the vertical distribution
of either contaminants or hydraulic
conductivity. The lack of such data is
a critical limitation for performing a
quantitative risk assessment.
Groundwater samples drawn
from these conventional wells repre-
sent composite samples. Because
they mix waters of varying true con-
centrations, they are diluted and give
a falsely low impression of the sever-
ity of contamination.
So how do we interpret concen-
trations of contaminants that are
below state or federal action levels?
Here, an old dictum applies: The
absence of evidence is not evidence
of absence.
It may be that, sometimes, low
concentrations are just that. But we
need assurance that the wells have
been located such that they actually
• continued on page 14
13
-------
LUSTLitie Bulletin 36
m Diving Plumes from page 13
sample the plume. When we sample
from the wrong place, as our
Patchogue example shows, we may
well think that concentrations are
lower than they actually are. As a
consequence, we incorrectly believe
that plumes are shorter than they are
actually, even to the point of appar-
ent nonexistent.
With the prevalence of biodegra-
dation of BTEX (and some emerging
news concerning MTBE biodegrada-
tion), we may be too quick to
attribute short plumes to natural
attenuation, rather than to the true
cause—a sampling error.
An Approach to Assessing
Plume Diving
Is there a universal prescription for a
practical assessment of plume diving
and vertical migration? We're afraid
not. The site assessment process
involves putting together the pieces
of the puzzle to delineate the extent
of contamination. One part of that
puzzle is a determination of the verti-
cal contaminant distribution. Because
of the potentially detrimental conse-
quences of missing a diving plume,
the site investigation should be
designed to ensure that a diving
plume doesn't extend out of the
range of the bottom of the monitoring
network. Site-specific factors, such as
geology, hydrology, land use, and
site geochemistry, provide the evi-
dence for plume diving. The follow-
ing factors should be evaluated in
planning a site investigation:
• Geology and the Sampling
Network
What land form contains the plume?
Is it a flood plain, delta, or coastal
plain? Do drilling logs indicate that
there are discrete zones that yield
plentiful water and other zones that
do not?
Core logs give information
needed to define the stratigraphy,
including the geologic units, their
consistency, and their orientation.
Has a cone penetrometer or borehole
flowmeter test been performed?
Have the monitoring wells been
tested to determine hydraulic con-
ductivity? Have they been tested to
determine whether they are screened
in intervals known to yield water?
What are the properties of the aquifer
14
The OnSite Plume Diving Calculator
Calibrated numerical groundwater flow models, such as MOD-
FLOW, can be used to show how much recharge-driven diving
might occur in an aquifer. Inasmuch as these models are not
applied at most LUST sites, we'd like to suggest that you try some sim-
pler alternatives.
From our experience at the East Patchogue site, a simple simulation
model was developed to estimate the prospects for recharge-driven
plume diving. The model is a part of EPA's on-line tools for site assess-
ment called OnSite. Use of the tools requires only a standard browser and
Internet access. The tools are available at http://www.epa.gov/aihens/
software / training / WebCourse / part-two / onsite.
The plume diving model allows an aquifer to be split into segments,
each with its own hydraulic conductivity, recharge rate, and length. The
upgradient and downgradient heads are specified in the aquifer, as is a
starting point that represents the source and a well location. Given these
specified aquifer parameters, an estimate is given for how deep the top of
the plume goes below the water table at the specified well location.
The software has been used on several Long Island sites and found to
match the observed plumes. The model, however, is based on a simple
one-dimensional conceptualization, and it won't be appropriate for all
sites. It does, however, give an idea of the prospects for recharge-driven
plume diving.
As the site investigation moves away from the source, the model can
be used to predict plume diving. Sampling of the aquifer can then show if
the predictions were correct. More to the point, sampling can show if the
plume is diving, and the model results give a guide for determining the
vertical extent of contamination. •
with regard to depth, thickness, and
hydraulic conductivity? Does the
well network characterize the aquifer
in two or three dimensions?
Clearly, vertical delineation of a
contaminant plume requires three-
dimensional characterization, which
may include the use of permanently
installed multilevel wells or tempo-
rary push points. Because of the
expense of installing and sampling
from multilevel wells, temporary
push points can be used to reduce the
cost of vertical delineation.
Using this push technique, loca-
tions can be sampled without
installing permanent wells when the
vertical location of the plume is not
known. Permanent monitoring wells
can be installed after the plume has
been located. Determining the hori-
zontal extent of contamination may
not necessarily be a simple task. Push
technologies can be of great benefit
here as well.
Answers to the questions posed
above help define the stratigraphy,
which serves as the geologic control
on the contaminant distribution. In
many flood plains, for example, the
surface material is primarily heavy
silts and clays that have been
deposited during historical flood
events. Beneath these surface silts
and clays are sand and gravel
deposits associated with previous
meanders of the river. The water
table is frequently in the surface silt
and clay. So the materials with a
capacity to carry groundwater (sand
and gravel) and transport a plume
occur at the bottom of the sequence
of deposited sediment.
The Elizabeth City, North Car-
olina, "Old Fuel Farm Site" illustrates
the effects of recharge, stratigraphy,
and sampling. It is described in a
report, Natural Attenuation of MTBE
in the Subsurface Under Methanogenic
Conditions, available from EPA at
http: / / www.epa.gov / ada / pubs /
reports.html.
• Hydrology and Land Use
From a topographic map, what do
elevations of areas such streams and
lakes indicate about the groundwater
flow system? How much annual rain-
fall occurs? What recharge estimates
have been developed for the area or
are commonly used? What are the
land use patterns?
-------
LUSTLine Bulletin 36
For example, tKe flow-system at
East Patchogue is determined by the
regional flow on Long Island, where
water generally flows from the center
of the island in the north to the Great
South Bay in the south. An average of
44 inches of rain falls each year, and
the United States Geological Survey
(USGS) estimates that the average
recharge is about half of the rainfall.
The UST facility property is
paved and adjacent to a highway.
Downgradient, the land uses include
light commercial, playing fields, a
gravel pit, and medium-density resi-
dential areas. Each of these land uses
influences the pattern of recharge,
from very low recharge where the
surface is paved to very high recharge
in the gravel pit.
From this information, we could
have suspected that the contaminants
were likely to travel toward the bay.
If the plumes moved away from the
service station property and out from
under the paved area, there would be
a good chance for diving behavior,
particularly if the plumes reached the
gravel pit—as indeed they did.
At other sites, unlined drainage
ditches, leaking water mains and
sewer pipes, irrigation, and the flow
pattern in the aquifer can determine
the vertical distribution of contami-
nants.
Thus, where recharge is likely to
be the plume diving instigator, the
amount of water that infiltrates the
area above the plume, and the
amount that this recharge contributes
to flow in the aquifer, determines
where and how much diving will
take place.
• Geochemistry
Simple geochemical tests can be used
to spot a plume that is diving because
of clean water recharge. In general,
uncontaminated recharge water at
the top of an aquifer will have oxy-
gen concentrations that exceed
1 mg/L, iron concentrations that are
less than 0.5 mg/L, and methane con-
centrations that are less than 0.1
mg/L. Groundwater that has been
contaminated with petroleum hydro-
carbons will generally contain oxy-
gen concentrations that are less than
0.5 mg/L and may contain concentra-
tions of iron and methane that are
greater than 1 mg/L. If the ground-
water is sampled with a bailer, the
sample is usually contaminated with
atmospheric oxygen during sam-
pling, and the rule of thumb for oxy-
gen should not be applied.
In general, clean recharge water
will have low dissolved organic car-
bon (DOC), usually less than 1.0 to 2.0
mg/L. The plume will usually have
elevated DOC, often exceeding 10
mg/L.
Putting the Pieces Together
The stratigraphy of an area provides
the first indication that plume diving
should be considered. Are the conta-
minants contained in dipping strata?
If so, off-site migration is likely to be
controlled by the stratigraphy. Dip-
ping or not, the plume direction will
be dictated by the flow that water
takes through the geologic structure.
The groundwater flow rate and
an estimate of the petroleum release
date provide clues about travel time
to various downgradient locations. If
the rate is low enough, the plume
may never reach that gravel pit or
unlined ditch that is waiting to drag
it to the depths of the aquifer.
So, before taking the site investi-
gation off-site, can an estimate of
plume diving be made? In simple
aquifers, the OnSite plume diving
calculator can be used to estimate
diving at a specific location. (See
sidebar on page 14.) Subsequent sam-
pling with a direct push probe can
provide confirmation (or not) of the
location of the plume, both vertical
and horizontal, before a commitment
to permanent monitoring wells is
made.
From our work on sites with div-
ing plumes, it's clear that the
prospects for plume diving need to
be factored into site investigations.
This information can be used to
determine whether diving is likely to
occur in the downgradient plume. If
diving is a possibility, then the sam-
pling design must be such that
plumes are fully characterized
through the design of the monitoring
network.
By the way, plume diving is not a
new concept. It was evident in data
collected from the first Borden
Aquifer dispersion experiment con-
ducted in the 1980s (MacKay et al.,
1986, A natural gradient experiment
on solute transport in a sand aquifer,
Water Resources Research, 22(13)
2017-2029). It was also observed in
data from the extensive USGS Cape
Cod field study (LeBlanc et al., 1991,
Large-scale natural gradient tracer
test in sand and gravel, Cape Cod,
Massachusetts, Water Resources
Research 27(5), 895-910). •
Acknowledgment
The U.S. Environmental Protection Agency
through its Office of Research and Develop-
ment conducted the research described in this
article. It has been subjected to Agency review
and approved for publication. loseph Haas of
the New York State Department of Environ-
mental Conservation provided the East
Patchogue, New York, data set.
Jim Weaver is with the Ecosystems
Research Division ofEPA's National
Exposure Research Laboratory, Office
of Research and Development (ORD),
in Athens, Georgia. He can be reached
at weaver.iim@epa.yov.
John Wilson is with the Subsurface
Protection and Remediation Division
ofEPA's National Risk Management
Research Laboratory, ORD, in Ada,
Oklahoma. He can be reached at
Wilson.iohnt@epa.vov.
DIVING PLUMES
in NEIWPCC's MTBE survey, when
asked if they investigate MTBE plumes
differently from BTEX plumes because
of the potential for diving plumes, 4
states answered "yes" and 15
answered "sometimes." When asked if
they require three-dimensional char-
acterization of MTBE plumes, 14 of the
19 states that answered "yes" to the
previous question answered that they
do "occasionally," 3 answered "most
of the time," and 1 said "always."
Delaware indicated that the answer
depended on the project officer
whether it was "occasionally" or
"most of the time." Montana com-
mented that if a vertical gradient is
apparent, nested wells will be required
to verify whether a diving plume
exists.
When asked if they are taking any
extra steps to make sure MTBE is not
migrating beyond standard monitor-
ing parameters, 19 states answered
"yes." When asked what kinds of
steps, most said that they are using
multilevel wells, nested wells, deeper
wells, and/or more wells located far-
ther downgradient from the source. •
15
-------
iiie Bulletin 36
Oxygenates
Analytical Issues for MTBE and
Related Oxygenate Compounds
by Deana M. Crumbling and Barry Lesnik
Questions have been raised about which analytical methods for MTBE and
related analytes are appropriate within the context of state and federal LUST
programs. To help answer some of these questions, we've prepared the fol-
lowing overview of the current status of MTBE analysis from the perspective and
experience of EPA's Office of Solid Waste (OSW) Methods Team, the group responsi-
ble for developing and maintaining the SW-846 Methods manual, and EPA's Technol-
ogy Innovation Office (TIO).
Because MTBE is not currently a RCRA-regulated analyte, it has not been validated
in any SW-846 method at this time. Neither Method 8021 nor Method 8260 has been validated
for MTBE. As it stands, analyses of a few of the oxygenated analytes that are of more
recent interest to LUST program personnel [e.g., tert-buryl alcohol (TEA) and ethanol]
have already been validated and published in SW-846 methods several years ago.
"Approved Methods"
Confusion often arises in the search
for an "approved method." A com-
mon misconception is that when a
method is published in SW-846 it
becomes an "approved method" and
is, therefore, required across the
board. This is not true. In fact, any reli-
able method may be used, whether it
is published in an EPA methods man-
ual or suggested as an alternative
method. (Note: Some state UST pro-
grams require the use of "EPA-
approved methods.") Any method
used must be able to determine the
analytes of concern in the matrix of
concern at the action level of concern.
SW-846 Methods
Requirements for using specific
"EPA-approved methods" in the con-
text of waste programs are discour-
aged. The use of prescriptive
analytical methods is counterproduc-
tive to the generation of reliable data,
because samples encountered in
waste programs are too varied and
complex for any single method to
work for all samples all the time. For
this reason, the SW-846 manual uses
a performance-based approach to
analytical methods.
SW-846 is intended to provide
general guidance, not prescriptive
requirements. There are no "refer-
ence methods" in SW-846, in the con-
text that the term is used in the Office
of Water Programs. Part of the mis-
understanding regarding analytical
method requirements stems from the
fact that EPA water programs do
require "EPA-approved methods"
when implementing the Safe Drink-
ing Water Act (SDWA) and the Clean
Water Act (CWA). However, meeting
SDWA and CWA requirements is not
usually the driver for projects within
waste programs, and "EPA-
approved methods" are not required.
Thus there will never be a pre-
scriptive method for MTBE within
OSW programs, even when MTBE is
included as a target analyte in pub-
lished SW-846 methods. However,
the growing interest in oxygenates
indicates that an SW-846 method that
addresses MTBE would be highly
valuable to the UST/LUST commu-
nity.
By the time they are published,
SW-846 methods have undergone
thorough evaluation and peer review
to (1) determine the level of method
performance that can be expected
under "typical" conditions, and (2)
identify what interferences might
compromise method performance
and what to do when it happens. We
believe that some of the existing SW-
846 methods are appropriate for the
sample preparation (Method 5031)
and determination (Methods 8015
and 8260) of MTBE and other related
target analytes in aqueous matrices.
However, it would take a "demon-
stration of applicability" to prove it.
In the Meantime
Until MTBE analysis is validated in
an SW-846 method, which methods
could be used for MTBE and related
analytes? The simple answer is that
"any method that can be demon-
strated to measure the constituent of
concern, in the matrix of concern, at
the level of concern, and at the degree
of accuracy as identified as necessary
to address the site decision" can be
used. (See http:// duin.org / down-
load /char/articled^, pdf. page 2.) Of
course, demonstrating that a method
is working as expected on a variety of
real-world sample types takes time
and technical expertise, and that
means that the answer may not be so
simple. With this caveat in mind, lef s
look at oxygenate analysis in relation
to existing SW-846 methods.
Method 8015
The chemical and physical properties
of the target analytes and the poten-
tial for interferences in the samples
submitted for oxygenate analyses
must be considered when selecting
potential sample preparation and
determinative methods. SW-846
Method 8015, "Nonhalogenated
Organic Compounds Using
GC/FID," was developed for the
analysis of oxygenates and is
expected to be applicable to MTBE
and other related compounds. (Note:
All SW-846 methods may be accessed
on-line at http:// www.epa.eov/
SW-846/main.htaO
Method 8015 is a determinative
method [gas chromatography/flame
ionization detector (GC/FID)] only.
As such, it is merely part of the
16
-------
LUSTLine Bulletin 36
"analytical metKod" picture. A
"determinative method" applies to
the analytical instrumentation used
to generate the analytical result. A
sample preparation (or sample intro-
duction) method is needed to get the
target analytes from the sample
matrix into the analytical instrumen-
tation.
An appropriate sample prepara-
tion method must be applied to the
original sample (such as water or
soil) so that the analytes can be trans-
ferred from the matrix onto the GC
column. If the sample preparative
method is not appropriate for the
analyte, transfer of the analytes from
the original sample into the instru-
ment may be incomplete or unpre-
dictable, and the.final results may be
erroneous due to low recovery, no
matter how good the instrumental
determinative method is.
Section 1.1 of Method 8015 cov-
ers some of the sample preparative
methods that have been shown to be
applicable for a variety of oxy-
genated compounds. Note that Sec-
tion 1.1 shows that the purge-
and-trap technique is rarely success-
ful for highly water-soluble oxy-
genates. Purge-and-trap works best
for analytes that are both volatile and
relatively insoluble in water (e.g., BTEX
compounds).
MTBE, however, is more soluble
in water than BTEX compounds, and
this characteristic decreases its purg-
ing efficiency relative to those com-
pounds, creating the possibility that
interferences in complex sample
types could render purge-and-trap
analyses susceptible to imprecision
and poor method sensitivity due to
unpredictable sample-specific purg-
ing efficiencies. This generalization is
even more true for oxygenated ana-
lytes that are more water-soluble
than MTBE is.
As with any preparative or deter-
minative method, evaluation of sam-
ple-specific characteristics in relation
to expected method performance (to
meet project-specific needs) for spe-
cific analytes is required to determine
whether purge-and-trap or some
other sample preparation method
can consistently provide the expected
data quality.
Section 1.1 of Method 8015 rec-
ommends that samples to be ana-
lyzed for highly water-soluble
oxygenated organic compounds
[such as tert-butyl alcohol (TBA) and
ethanol] be prepared using direct
injection or azeotropic distillation
(Method 5031). Direct injection alone
(into a GC/FID) has been shown to
achieve detection limits in the range
of 400-500 ppb for TBA and ethanol.
Azeotropic distillation tech-
niques can be used to concentrate
samples for these alcohol analytes.
The azeotropic distillation sample
preparation/concentration technique
has been shown to produce detection
limits in the vicinity of 10 ppb when
the concentrated samples are ana-
lyzed by GC/FID. Vacuum distilla-
tion (SW-846 Method 5032) and static
headspace (SW-846 Method 5021)
could also be considered as poten-
tially viable sample preparative
methods. Each of these preparative
methods will have its advantages
and drawbacks. Additional develop-
ment work for both preparative and
determinative methods will be
required to validate routinely applic-
able methods across the range of oxy-
genate compounds, sample types,
and detection limits that are now of
interest.
g; The selection of any analytical
^method must always consider the
I™*—
-ultimate use of the data. TJie use of
^project-specific systematic planning
i£- can ensure that data collection
IN. -
^ methods are cost-effectively
matched to the project's
decision-making needs.
i
I
Method 8021
SW-846 Method 8021—Aromatic and
Halogenated Volatiles by GC Using
Photoionization (PID) and/or Elec-
troconductivity (E1CD) Detectors—or
a similar method that relies on a pho-
toionization detector, is not recom-
mended as a determinative method for
MTBE and its associated oxygenates.
PID is most sensitive to compounds
that contain double bonds (which is
why this method is a good determi-
native technique for BTEX com-
pounds).
MTBE and related compounds,
however, do not contain double
bonds. Although the PID analysis
will respond to the oxygen atom in
these compounds, the response is
weaker than the response for BTEX
compounds and, therefore, may be
subject to interference and false posi-
tives when real-world samples con-
tain significant amounts of other
contaminants such as petroleum
hydrocarbons. (See pages 9 and 15 of
"An Evaluation of MTBE Impacts to
California Groundwater Resources,"
available at http:/ /www-erd.llnl.
gov / mtbe / pdf / mtbe.pdf.) The E1CD
detector of Method 8021 works only
for compounds containing halogen
atoms, and MTBE does not possess
this characteristic either.
Method 826O
GC with a mass spectrometer (MS)
detector is also appropriate as a
determinative method for oxy-
genates, as long as a sample prepara-
tive method appropriate to the
sample has been used. MS offers the
advantage of unambiguous identifi-
cation of target compounds. It is a
good idea to keep a few things in
mind if a GC-MS method for MTBE
and other oxygenates is discussed in
terms of SW-846 Method 8260:
• Method 8260 has not been vali-
dated by EPA for use with
MTBE.
• Method 8260 is a GC-MS deter-
minative method only—the
sample preparation method is
separate. Purge-and-trap is not
specified by Method 8260 and,
in fact, is not recommended for
the few alcohol analytes that
have been validated in Method
8260 (e.g., ethanol and TBA).
(See the Appropriate Prepara-
tion Technique table in Section
1.1 and Section 1.2.)
• Improved performance of
Method 8260 for MTBE and
other oxygenates can be
expected if instrument operating
conditions are modified to
accommodate that particular
analyte group (rather than try-
ing to generalize operating con-
ditions to accommodate the
entire range of 100-plus vali-
dated analytes in the Method
8260 list).
I continued on page 18
-------
LUSTLine Bulletin 36
• Analytical Issues for MTBE
from page 17
Field Methods
In addition to their analysis by GC-
MS in the fixed laboratory, MTBE,
ethyl t-butyl ether (ETBE), and TEA
have been successfully analyzed in
the field by using a field-portable
GC/MS and heated (at 60°C) static
(i.e., equilibrium) headspace. Method
performance information (provided
by Field-Portable Analytical, Inc.)
shows detection limits in the range of
4-5 ppb when the MS is operated in
full-scan mode. When operated in
selective ion monitoring (SIM) mode,
detection limits down to 0.2 ppb are
possible with full positive identifica-
tion of the analytes. As samples are
analyzed in the field at the time of
collection, issues regarding sample
preservation are avoided.
Depending on the nature of the
project, field analysis can signifi-
cantly decrease costs by supporting
real-time decision making according
to an Expedited Site Assessment
approach. (See EPA 510-B-97-001,
"Expedited Site Assessment for
Underground Storage Tank Sites: A
Guide for Regulators," available on
OUST's Web site at http: / /www.epa.
gov/swerustl /pubs/index.htm#sam.')
The Decision-Making Factor
Above all, the selection of any analyt-
ical method must always consider die
ultimate use of the data. Data for risk
assessment purposes typically need
lower detection/quantitation limits
than when data are used to delineate
a plume or to place monitoring wells.
The use of project-specific systematic
planning can ensure that data collec-
tion methods are cost-effectively
matched to the project's decision-
making needs.
The flexibility inherent in SW-846
methods permits "mixing and match-
ing" of sample preparation and
determinative methods so that the
needed method sensitivity and accu-
racy can be achieved. As long as data
are of known quality, and that qual-
ity has been matched to the decision-
making needs of the project, any
reliable method can be used.
Although we have discussed in
general terms the various analytical
method options for oxygenates that
might be explored, a more specific
18
answer to the question, "What
method should be used for MTBE
and/or oxygenates for this particular
project?", first requires that the pro-
ject manager clearly specify the
intended use of the data. When this
use is known, the required detection
limits can be determined, the desired
turnaround time for the data results
can be estimated, and the most cost-
effective option for generating the
data (i.e., the sampling program and
the analytical methods) can be
deduced.
For More Information
Information about SW-846 and the
selection of analytical methods can be
found through the Clean-Up Infor-
mation Web page of the Technology
Innovation Office at http: / / cluin.
org/charl.htm and the OSW Meth-
ods Team home page at http: / /
www.epa.gov / SW-846 /. More infor-
mation about how site characteriza-
tion and cleanup can be made more
cost-effective can be found at
http://www.clu-in.org/products/
failsafe.htm. •
Deana Crumbling is an analytical
chemist with EPA's Technology Inno-
vation Office, and works on improving
the accessibility and application of site
characterization tools to waste site
cleanups. She can be reached at
Crumbling.Deana@epamail.epa.gov.
Barry Lesnik is the Organic Methods
Manager with EPA's Office of Solid
Waste Methods Team and is responsi-
ble for the organic methods included in
SW-846 and regulatory issues dealing
with chemical analysis. He can be
reached at lesnik.barn/@eva. mv.
Results of NEIWPCC
Survey of State
Experiences with MTBE
to Be Posted on Web
In August, the New England
Interstate Water Pollution Control
Commission (NEIWPCC) con-
ducted a survey of all 50 state
LUST programs to ascertain their
experiences with monitoring for
and cleaning up MTBE releases
from USTs. The survey, funded
by the EPA Office of Under-
ground Storage Tanks, was
undertaken to provide the states
with a better picture of how
each state program is currently
dealing with MTBE and other
oxygenates. This very com-
prehensive survey consists of
34 questions and numerous
subquestions. NEIWPCC received
responses from all 50 states.
After sending the compiled
results to all states for a final
review, NEIWPCC plans to post
the results with an executive
summary on its LUSTLine Web
site (lustline@neiwpcc.org^ on
December 15, 2000. NEIWPCC
will encourage states to update
their information periodically and
will present the information in the
next issue of LUSTLine and at the
national UST/LUST conference in
Albuquerque, New Mexico, in
March. You will also find issue-
specific summaries in this issue of
LUSTLine with relevant articles.
Look for the "What Our Survey
Shows" boxes. •
toEWTEeHNICAL BULLETINS EBQM ;APL
The American Petroleum Institute (API) has recently completed several technical bul-
letins that address aspects of LUST remediation. These bulletins can be downloaded
from http://www.api.ehs/saresbul.
• Dissolution of MTBE from a Residually Trapped Gasoline Source—September
2000, Bulletin No. 13.
• No-Purge Sampling: An Approach for Long-Jem Monitoring—October 2000,
Bulletin No. 12.
» Strategies for Characterizing Subsurface Releases of Gasoline Containing MTBE—
August 2000, Bulletin No. 11.
• Simulation of Transport of Methyl Tert-Butyl Ether (MTBE) to Groundwater from
Small-Volume Releases of Gasoline in the Vadose Zone—June 2000, Bulletin No.
10.
• Non-aqueous Phase Liquid (NAPL) Mobility Limits in Soil—June 2000, Bulletin No. 9.
-------
LUSTLine Bulletin 36
Getting Results, PFP Style
ray for performance (PFP) is"acommon-seme~appwach to LUST~sitej:leanup. Payments are made as con tmnmatwn levels go
' P^-— ^^" *x a?**~ 2 IB*. ] Sfr~" 8 ^—"~ * A,
"n and. cleanup goals are achieved and maimamed. The price, interim payment milestones, contamination-level goals, and
'mejimitfor reachingthe goals are all firmly fixed at the beginning ofthe cleanup ana
-„•„ ,.~j £,,7T., "£scflpg clauses" are tgrjjtten into the contract tha±.can release \
not changed thereafter. Contaminant
reduciiornsjns-nsured fan
faulty sitelcharacterization
WttWn the PFP frame
the lowest-price bid submit
yrk
•ed in i
new release).
there are significant vjjjyations on how stales pricjfyheir PFP c\
in open competition among qualified contractors or by negotf
South Carolina, and Florida^-all PFP pioneers—havejjgen pricing theif
|P - J J6fe— & JT iipBs,
negotiate^ price with the cleanup contractor. South Carolina conducts o
der. FlorMa has experimented-withboth negotiation and bidding and h,
one totaWprice. And predictions that lowest-bid contracjjfs would product
Theijeadership of these states and the success ofth~&r_P_FP cleanups
cleanups using
'Pftitive bidding a
also awarded "bun
'.oddy work have n
§»p a wealth ofexp>
contractor if thejiggd arises (eg, a
;r
>e set by award to
*•* _,r ,*
ing their^unJPFP initiatives. These states arefindingmjat their PFP cleanups are typically lest
tive, conclude in the expected time, and spur innovationjn cleanup technology and managemei
Thefollowing thr'ee articles shed some light on the Oklahoma, South Carolina, and Flo
progressjjsupportfor developing a state PFP program y
bailable through OUST/EPA Regwnl
ips. Prices m(
"onjvith the contractor Oklahoma,
rent methods. Oklahoma state staff
the cleanup to the lowest bid-
" of multiple,PFJLcleanup sites for
wed true.
that others can adapt in develop-
"Sxpensive and environmentally effec-
iy contractors.
fa PFP programs. PFP is a work in
A t
I Office representatives.
After Some PFP Growing Pains
Oklahomans Realize PFP Benefits
by Richard McKay
iy state starting a performance approach to remedi-
ation will go through some growing pains. For the
smoothest transition possible, those involved must accept
that remediation programs will undergo a paradigm
change. Leaders, technical groups, and accounting staff
must think outside their traditional time-and-materials
program models.
The growing pains can be intensified if the regulatory
and fund groups work independently. In our experience,
the ideal arrangement is to have a technical staff with both
regulatory and fund authority. When one department is
managing the technical and financial aspects of a case, a
consistent message is projected to consultants and tank
owners, and the potential for having a case slip between
the cracks is reduced. If this setup is not available, a spirit
of cooperation between all parties is required.
Regulators, fund groups, and consultants are all
entrusted with the responsibility to protect human health,
safety, and the environment and to ensure that cleanup
funds are spent effectively. These purposes will be under-
mined if the parties involved do not consciously work
together toward a common goal.
Statutory Roadblocks
Although statutes may not specifically provide or allow
the authority for a state agency to enter into performance-
based cleanup contracts, this omission may not necessarily
be an obstacle. Recognizing that the concept of perfor-
mance-based cleanup contracts made a lot of sense, the
Oklahoma Petroleum Storage Tank Division (PSTD)
implemented a pay-for-performance (PFP) program in
1996 on a voluntary basis. The necessary forms were cre-
ated, several contracts were signed, and the remediation
systems were installed and implemented.
Despite challenges to our statutory authority, the per-
formance-based reimbursement program prevailed. There
were some parties, however, who felt specific authority
was needed. Thus, with the full support of the state's
petroleum marketers, statutes were passed in 1998, man-
dating that all work be preapproved and empowering the
authority to enter into preapproved purchase orders and
performance contracts. The rules were revised to make the
preapproval process mandatory. Now all site work must
be preapproved and most site remediation is performance
based.
Benefits of Pay for Performance
Oklahoma's PFP program provides benefits for the envi-
ronment, for fund protection and management, and for
claim processing. These benefits include the following:
• Consultants now install better-designed remedia-
tion systems. Thus our most difficult sites are being
cleaned up, and all site cleanups are progressing
faster.
• continued on page 20
19
-------
LUSTLinc Bulletin 36
m Results, PFP Style from page 19
• Consultants must guarantee results, and no pay-
ments are made until incremental goals are attained.
• The flow of fund money is manageable, because
remediation costs are fixed and controlled through
negotiation and the use of TankRACER software.
We are thus able to encumber preapproved site
remediation monies in predictable amounts.
• The time it takes to pay claims is shorter, and there
are minimal disputes over reimbursements and dis-
allowances. Disputes over reasonable prices are
eliminated.
• Minimal claim support documentation is required,
minimal erroneous or questionable documents are
received, and payment for work that is not per-
formed is eliminated.
• Collaboration between the tank owner, the consul-
tant, and the state has improved so that tank owners
are more likely to view the agency as an advocate
than as a headache.
Oklahoma's PFP program has shifted the consultants'
focus from keeping cases open on a time-and-materials
basis, with little incentive to close a case, to achieving
results to make money. As a consequence, the rate at
which groundwater benzene concentrations are reduced
has changed from a small, slow decrease over several
years to a large decrease within a few months.
For example, in cases where contaminant reduction
milestones have been achieved, on average, the 25 percent
milestone has been achieved in 6 months, 50 percent in 8
months, 75 percent in 11 months, and 100 percent in 16
months from baseline concentrations measured prior to
system start-up. In each case, the consultant signed a per-
formance contract guaranteeing results in three to five
years from system start-up, and the existing remediation
system was replaced by an entirely new system. None of
the previous systems had been able to maintain contami-
nation levels below site-specific cleanup levels, and most
showed very little progress.
Under PFP, the consultant guarantees that the soil and
groundwater readings in the remediation area will be
below cleanup levels for all chemicals of concern (COCs)
before the system can be turned off, and the readings must
remain at or below site cleanup levels for six months
before the final contract payment is made. When we con-
vert a time-and-materials site to PFP, contamination levels
typically drop suddenly, rebound somewhat, and then
continue to decrease.
Changes on the Run
The PSTD has gone through several episodes of growing
pains since implementing its program in 1996. Our experi-
ence with writing performance contracts has helped us
dose a number of loopholes. For example, system design
was initially not specifically itemized as part of the final
cost. One consultant contested this policy, so we changed
our guidelines.
We have received many ideas for program revisions
from consultants—in the spirit of cooperation—to
improve the contract, rather than take advantage of an
20
omission. By keeping an open mind throughout this
process, our agency has had the opportunity to learn from
its mistakes, as well as from people outside the agency,
such as consultants and their attorneys.
Through our experience, we've incorporated many
important defining points into our performance contract,
including the following:
• Items that the contract price includes or excludes;
• The remediation system warranty area;
• Fair and reasonable payment terms;
• Which party takes responsibility for damages
caused by the tank owner or his or her employee;
• The situations that will allow the contract to be
renegotiated (i.e., secondary release, continuing
release, or migration of a plume onto the site);
• A provision that ensures continual system opera-
tion;
• Appropriate penalties if a consultant abandons
remediation activities prior to termination of the
contract;
• Points at which to take baseline samples;
• Lab analyses that should be run, schedules for sam-
pling wells, and conditions under which the consul-
tant will be able to change labs during the course of
the contract;
• Ways that reduction payments are related to BTEX
concentrations and the method of calculation;
• The method for measuring free-product reduction;
and
• A sampling protocol to qualify for reduction pay-
ments, reserve the agency's right to verify all sam-
pling data, identify key monitoring wells and
compliance monitoring wells, and determine a rea-
sonable period to monitor for rebound once all
wells are below cleanup levels.
The term of a PFP contract varies based on site-spe-
cific conditions, the chosen remediation technique, and
the operating history of similar techniques. For example,
after writing several contracts, we found that in clay-rich
soils, the time it takes to achieve the final 25 percent
reduction can be longer than the time it takes to attain the
75 percent reduction milestone. To compensate for this
slowdown, many of these performance systems have been
enhanced by localized dig and haul operations, additional
remediation wells to increase well density, the introduc-
tion of nutrients to increase bioremediation, or the intro-
duction of oxygen-releasing materials. The state also
allows a PFP system to be modified from the original
scope of work, provided that modifications are performed
within the terms of the contract and at no additional cost.
Without the ability to make these modifications, the
consultant risks leaving up to 40 percent of the perfor-
mance contract on the table. Keep in mind, there are also
sites where the 100 percent milestone has been achieved
six months after start-up, leaving the consultant with 2 to
21/2 years of operation and maintenance money as pure
profit. These cases create an established history from
which the agency can learn what a reasonable remedia-
tion time frame should be and apply that lesson to future
-------
LUSTLine Bulletin 36
contracts. Although system performance varies, the con-
sultant must ultimately achieve the final goal in a reason-
able time frame and deliver a site that is ready to be
monitored for closure.
Negotiating a Fixed Cleanup Price
One of the primary objectives of PFP cleanup is to achieve
results at a reasonable price. Cleanup prices can be set
through negotiations and/or bidding. Since the program's
inception, Oklahoma has used a customized price build-
up computer program, TankRACER, to determine a rea-
sonable price. Detailed printouts from TankRACER are
used as support documents for negotiating a final contract
price with the consultant. Through this negotiated proce-
dure, we have saved a total of $875,000 over the consul-
tants' original proposals, which can then be used for
characterization and restoration work on other sites. Typi-
cally, the TankRACER price varies by only 4 percent, on
average, from the final contract price, and assures all par-
ties that the final negotiated price is reasonable.
Reasonable Cleanup Goals and Price
Cleanup goals have a direct effect on the performance
remediation price and are commonly based on a category
system, a maximum contaminant level (MCL), or a tiered
risk assessment. Oklahoma changed from a category sys-
tem to a tiered approach in 1996, allowing more reason-
able and achievable site-specific cleanup goals that are
protective of human health and can be attained at a rea-
sonable price. Based on this tiered approach, the consul-
tant guarantees that soil and groundwater will be remedi-
ated to site-specific cleanup goals at a negotiated price that
includes all remediation costs. Today the average perfor-
mance site remediation price using air sparge and soil
vapor extraction techniques is $498,000 for a 26,500 yd3
plume or $18.80 per yd3.
Had we utilized a risk-based program to determine
reasonable cleanup levels and instituted a performance
program from the inception of our tank program, we esti-
mate we could have saved as much as $6.48 million on just
41 sites that were changed from time and materials to per-
formance. This savings assumes that each site moved from
site assessment directly into PFP remediation. These cases
represent a small portion of the sites that require remedia-
tion. The economic consequences of not instituting pro-
grams to determine site-specific cleanup goals and a
reasonable site remediation price could be substantial.
Since these changes were instituted, we have been
able to prioritize each site, use better budget controls, and
move the worst sites more quickly toward implementing
corrective action. In addition, the consultants are now
more inclined to develop and use remediation techniques
that are faster, more efficient, and more cost-effective. •
Richard McKay is Supervisor of Special Projects, Programs,
and Operations at the Oklahoma Corporation Commission,
Petroleum Storage Tank Division. He canbe reached at
r.mckay@occ.state.ok.us.
South Carolina
The Light at the End of the Invoices
PFP Bids Well in South Carolina
T
by Arthur Shrader
JL he South Carolina Bureau of UST Management imple-
mented a pay-for-performance (PFP) remediation pro-
gram for state-funded cleanups in 1997. Our program
goals were to encourage cleanup contractors to be more
efficient and effective, achieve cleanup goals at a reason-
able price, and simplify the invoicing process. In an effort
to streamline cumbersome time-and-materials procedures,
we implemented PFP, using the competitive bidding
process. As a result of this move, we're seeing a nice,
bright light at the end of our invoices.
Here's how PFP works in South Carolina. Our depart-
ment solicits bids for proposed projects in the state gov-
ernment biweekly publication, South Carolina Business
Opportunities. Prior to the advertisement, department staff
members prepare specification packages that contain an
assortment of information necessary to assist interested
contractors in preparing their bids—stated cleanup goals
(based on site assessment activities and current levels of
chemicals of concern in key monitoring wells), site maps,
summarized technical data, and other relevant informa-
tion. Contractors are also encouraged to review the entire
project file located at the agency's Freedom of Information
office. Inasmuch as South Carolina certifies UST rehabili-
tation contractors, a bid bond is not required. Any UST-
certified contractor is welcome to submit a cleanup
proposal.
Contractors that wish to respond to the solicitation
submit a proposal that specifies a cleanup method or com-
bination of methods, an estimated time for completion,
and the total cost. UST program staff members evaluate
the proposal to determine whether the proposed technol-
ogy is feasible, the estimated time is protective of recep-
tors, and the total cost is reasonable, based on the costs of
similar cleanups. If more than one proposal meets all of
these parameters, the contractor offering the lowest bid is
selected.
When the contract is awarded, the selected contractor
submits a detailed corrective action plan along with a per-
formance bond or irrevocable letter of credit equal to the
amount of the award to guarantee that the project will be
completed successfully . The department approves the
plan and notifies the public of the proposed corrective
action before work begins.
• continued on page 22
-------
LUSTLine Bulletin 36
• PFP in South Carolina from page 21
As the project progresses, the contractor is paid a per-
centage of the total cost as agreed cleanup milestones are
achieved. Both program staff members and the cleanup
contractor take duplicate water samples for analysis by
separate laboratories to verify progress. When the contrac-
tor reports that cleanup goals have been reached, monitor-
ing wells are installed at random locations selected by staff
members to verify mat the total affected area has been suc-
cessfully rehabilitated.
Results Keep Getting Better
Since 1997, six cleanups using this process have been com-
pleted. All were completed within the schedule outlined
in the proposal. Of the ongoing 180 PFP cleanups, 57 per-
cent have reached 75 percent of the cleanup goals, 35 per-
cent are in post-startup and are achieving goals, and 8
percent are in the corrective action plan development
stage.
At their own costs, contractors routinely install addi-
tional treatment points or excavate additional soils in the
main source area after implementation of the initial correc-
tive activities to accelerate the cleanup. The use of more
durable equipment (to eliminate downtime and to reuse
the equipment at the next job) is also quite common.
Table 1. Examples of decreased prices over time at South Carolina
PFP sites.
Free Product
& Dissolved
Free Product
Only
1997
$275,000
$180,000
1998
$180,000
$100,000
1999
$133,216
$30,000
2000
$117,000
$29,500
As contractors become more familiar with the PFP
process, bid amounts for cleaning up similar size ground-
water plumes have been further reduced. (See Table 1.)
Based on available data, our cleanup costs are more
directly attributable to plume size than to other factors,
such as geology or levels of mass reduction. The cleanup
of a larger plume requires more treatment points and a
greater overall effort than a smaller plume does. MTBE
plumes are more costly to clean up, because they are typi-
cally larger than BTEX plumes.
De-crazyfication
As we've eased into PFP, we've found that voluminous
invoices depicting time-and-materials charges are a thing
of the past. Invoices that are received from contractors
that have achieved a cleanup milestone consist of a single
page indicating the percentage of the bid price that is due.
The quarterly monitoring report documents the amount
of reduction, and the split-sample laboratory data verify
progress for the selected monitoring points. The payment
approval process is typically completed within two days
(the invoice is approved for payment or returned until
progress is documented).
PFP focuses the contractor, the regulator, and the
fund administrator on environmental results. Although
our PFP program is still in its infancy, we have already
been witness to more timely and efficient cleanups, lower
costs, and the opportunity for our staff members to spend
their time more appropriately overseeing cleanup activi-
ties, not reviewing invoices that resemble novels. •
Arthur Shrader is the Director of the Assessment and Correc-
tive Action Division of the South Carolina Department of
Health and Environmental Control, Bureau of Underground
Storage Tank Management. Art can be reached at
shradeaa@columb26.dhec.state.sc.us.
Florida
*—U**K !••
\ \ For a Third of the Price, Who'd Be Without It!
j \- \ Competitive Bidding; for Petroleum
xl Cleanup in Florida
I
by Brian Dougherty
t was one of those projects you don't really want. You
don't think it will work, but you need to do your best to
make it work. In 1996, the Florida legislature mandated
that the Department of Environmental Protection (DEP)
initiate a pilot program on bidding for petroleum site
cleanup services. We went along because we had to, not
because we thought it was a good idea. Now, four years
later, it seems like maybe bidding is a good idea. Our
experiment with competitive bidding for petroleum-cont-
aminated site cleanup services has demonstrated not only
22~ ~~
that it works, but that it gets the job done for one-third of
the price of preapproved cleanup work. (That's one-third
of the price, not one-third less).
To date, the department has accepted six bids for
petroleum cleanup services: five for site assessment work
and one for PFP cleanups. The results are easy to see. (See
Table 2.) Florida has saved an estimated $815,671 (64%) by
bidding the work rather than doing it under our preap-
proval program.
-------
LUSTLine Bulletin 36
1 Table 2. Summary at petroleum cleanup program bids. \
Type of Work
Assessment
Assessment
Assessment
Assessment
Pay-for-Performance Cleanup
Assessment
Total
No. of
Sites
11
10
7
10
3
26
No. of
Responses
24
27
34
16
11
11
Cost for Requested Services
(Total for all sites)
Bid Award
$ 22,500
$ 32,751
$ 24,900
$ 34,075
$227,550
$ 123,768
$ 465,544
Preapproval
Estimate
$ 41,800
$ 95,000
$ 108,260
$ 95,509
$ 665,500
$ 275,146
$1,281,215
Cost Difference Between
Preapproval and Bid
Difference
Preapproval-Bid
$ 19,300
$ 62,249
$ 83,360
$ 61,434
$ 437,950
$151,378
$815,671
Percent
Reduction
46%
66%
77%
64%
66%
55%
64%
What About Quality?
An obvious question at this point is, What about the qual-
ity of the work? Do you still get good work at such bargain
prices? The answer is yes. The quality of the work per-
formed under the bidding pilot has been as good as typi-
cal work under preapproval. That is not to say that there
have not been any problems with the work. We did have
to provide a few reminders that the scope of work was a
bid specification that had to be performed exactly and
completely. But all of the problems were resolved satisfac-
torily.
The Advantages of Bidding PFP
Bidding a PFP cleanup has distinct advantages over nego-
tiating the same cleanup. The one-time bid to establish the
price for the cleanup rules out any need to negotiate the
price. When you bid the cleanup, the burden for determin-
ing the best, most efficient strategy for the cleanup is
placed squarely on the consultant. The consultant must
price the job as competitively as he or she can.
Bidding the cleanup avoids the common negotiating
pitfalls associated with estimating the cost of the treatment
technology, estimating total cleanup time, and estimating
the monitoring time. In a negotiated PFP, these estimates
tend toward the high side to provide as much contingency
as possible. This tendency is not necessarily bad, and the
need to cover contingencies is real, but it can be difficult
and time-consuming to whittle the contingency down to a
level that is acceptable to both parties.
Getting the Specs Right
Bidding the cleanup work did initially require consider-
ably more administrative overhead than it would have
under conventional preapproval. Much of this additional
effort reflected our "learning curve" in developing a pre-
cise specification and learning the administrative proce-
dures. Most of this effort is now behind us.
The single most time-consuming aspect of preparing
the invitation to bid is the development of the exact bid
specification. As is common for bid work, the specification
must be exact and precise. In this regard, we found that
the assessment bids were more difficult to spec than the
PFP bids. The PFP bids were easier to prepare because it is
fairly straightforward to write a bid specification for a
completed cleanup.
The assessment bids were more time-consuming to
prepare because the exact locations and depths of wells
had to be specified. Then, because it is nearly impossible
to predict exactly how much work will be required to
complete the assessment, the specification development
effort had to be repeated for each subsequent scope of
work.
After our first couple of bids, we found a way to elim-
inate the need for follow-up bid specifications. We did so
by developing a fee schedule that sets forth a precisely
defined minimum scope of work that will be awarded. If
the minimum amount of work is insufficient to complete
the assessment, then we can award the additional work
necessary to complete the assessment, using the fee sched-
ule. This approach eliminates the additional overhead of
having to prepare and offer a new bid. Bidding the
cleanup work will always require a more formalized
approach than preapproval, but the net overhead ends up
being about the same.
Two Thumbs Up
Bidding petroleum cleanup services has yielded tremen-
dous cost savings with no decrease in the quality of work.
The administrative overhead was a burden at first, but we
have already made many improvements to our internal
processes to reduce that burden. The fee schedule
approach will further reduce the administrative overhead
for assessment sites. This reduction in administrative
overhead, coupled with the dramatically lower prices,
suggests strongly that bidding petroleum cleanup services
is an unbeatable way to manage Florida's petroleum
cleanup program. •
Brian Dougherty is an Environmental Administrator with
the Florida Department of Environmental Protection. He can
be reached at Brian.Doughertu@dep.state.fl.us.
23
-------
LUSTLine Bulletin 36
nically Speaking
by Marcel Moreau
• Marcel Moreau is a nationally -;
SKagnized petroleum storage specialist _ |:
e.folumn, Tank-nically Speaking, \
|J is a regular feature o/LUSTLme. As ;:,
pi always, we welcome your comments and j
|s- questions. If there are technical issues \
f that you would like to have Marcel :\
- - -- discuss, let him know at '.•;•
'marcel.moreau@juno.com. -• ]
— -- ' -- --
The Quest for the Perfectly Reliable LLD, or
S/i0ttW Electronic Line Leak detectors Hatfe an Annual test
of Operation?
M & yfirst encounter with the electronic line leak detector (ELLD) "test of operation" issue came a few years ago during a
J\/m compliance inspection. The recordkeeping at the facility was pretty good, but there was no record of an annual test ofoper-
X F P ation of the ELLD. The maintenance person said that he had checked with the manufacturer to obtain test procedures and
had been told that the device did not need to be tested.
At the time, this statement seemed to me to be a bit presumptuous on the part of the manufacturer. Nevertheless, the rules did say
that test procedures were to be performed "in accordance with the'manufacturer's requirements," so the manufacturer did seem to
have some ground to stand on.
I have since heard the question, "Do electronic line leak detectors need to be tested?", many times from inspectors and havefol-
Imved discussions concerning the issue with Internet interest groups. There are two main schools of thought on the issue:
• The "proof is in the pudding" school. This view holds that, "The rule says a test should be done, and there is only one true test
of operation and that is to see if the device can actually find a leak"—a view held primarily by regulators.
• The "father knows best" school. This view holds that, "I build these things and I know how they work. These devices are pretty
smart, can tell when they are not working right, and don't need any additional testing"—a view held primarily by some man-
ufacturers. This vieio is also popular with UST owners who have invested in ELLDs, in part, to avoid the cost of annual test-
ing of mechanical devices.
Although I believed the points made by both sides had some validity, my own tendency has been to lean toward the regulatory
vitno of "the proof is in the pudding." Having done a little more research into the matter, however, I am beginning to lean toward the
"father knows best" school.
The Electronic LLD and the
Testing Issue
For the first 30 years after its intro-
duction in the mid-1950s, the LLD
remained an entirely mechanical
device (see "Of Blabbermouths and
Tattletales," LUSTLine #29). Since the
implementation of the federal rules,
however, a number of manufacturers
have developed LLDs that are con-
siderably more sophisticated than the
original mechanical models and rely
on electronic components to do their
job. Although the mechanical devices
(MLLDs) are still the most common
type in service, the ELLDs are mak-
ing headway in the marketplace.
Annual testing of MLLDs has
been a requirement of the fire codes
since long before the federal rule. The
24
federal UST rule (and most state UST
rules) have adopted this requirement
as well. Typical language states that
"an annual test of the operation of the
leak detector must be conducted in
accordance with the manufacturer's
requirements" (40 CFR 280.44(a)).
The testing of MLLDs is fairly
straightforward. Because all of the
working parts are concealed and the
MLLD is self-contained, there is no
way to test it other than to generate a
leak and see if the MLLD responds.
The typical test procedure involves
connecting a testing device into the
piping system at the crash valve at
the base of the dispenser. The testing
device typically includes pressure
gauges and a small valve that can be
carefully adjusted to allow three gal-
lons per hour (gph) of product to leak
out of the piping and into a suitable
container.
The "test of operation" issue,
however, becomes more complex
with ELLDs. These devices are usu-
ally capable of conducting more
accurate 0.2 or 0.1 gph tests, in addi-
tion to the 3 gph test. Because the fed-
eral definition of a line leak detector
is written as a performance standard
(detecting 3 gph leaks at 10 psi in one
hour), the annual test of operation of
LLDs applies only the 3 gph function
of ELLDs. There is no requirement in
the federal rules to evaluate the abil-
ity of the ELLD to detect leaks of 0.2
or 0.1 gph on an annual basis.
-------
LUSTLine Bulletin 36
The Question Please...
The debate concerning ELLD test
procedures boils down to this point:
many regulators want to continue the
tradition of testing operation by gen-
erating leaks and seeing if they are
detected; some manufacturers insist
that their ELLDs are completely self-
testing and need no additional evalu-
ation.
Note that not all manufacturers
claim that their ELLDs are self-test-
ing. In fact, some state that the test of
operation should consist of generat-
ing a leak and verifying that it is
detected.
To understand the bases for the
opposing opinions, we need to
understand a little more about the
operating principles of ELLDs and
the types of "self-testing" they are
capable of conducting.
Types of ELLDs and How
They Work
There are two basic types of ELLDs:
flow-based and pressure-based. Both
types attempt to evaluate the
integrity of the piping immediately
after each customer has finished dis-
pensing product. The test may
require from less than a minute to as
long as 10 minutes to complete. If
another customer arrives and turns
on the pump, the test is aborted and
restarted when this customer is fin-
ished dispensing.
In general, both types of ELLDs
have the ability to turn the pump on
and off and to communicate in the
form of displays and/or printers.
They also have some computational
and/or logic circuitry that can deter-
mine if a piping run is tight and eval-
uate, to some degree, how well the
ELLD itself is functioning.
Are ELLD self-tests sufficient to
meet the regulatory standard of
"annual test of operation...
Jojiducted in accordance with the
^manufacturer's requirements" or
not? Well, it depends...
Pressure-Based ELLDs
Pressure-based ELLDs are the most
common type of ELLD. These ELLDs
monitor the pressure in the line after
the pump has been turned off. A
check valve in the system is used to
maintain some pressure in
the piping. A leak in the
piping will reduce the
amount of liquid in the
pipe and produce a loss
of pressure in the piping
that can be measured.
A pressure trans-
ducer—a device that con-
verts changes in pressure to
changes in voltage—is
installed in the piping to
detect pressure changes. The
bigger the leak in the pipe, the
faster the pressure in the pipe will
drop. If the pressure drops more
than a certain amount in a set interval
of time, then a failed test results.
Brands of pressure-based ELLDs dif-
fer principally in how much pressure
is held in the pipe at the beginning of
the test, the length of time during
which the pressure is monitored, and
the number of times the test is
repeated before a leak is declared.
The common use of flexible pip-
ing in today's UST systems has pre-
sented a bit of a challenge to
pressure-based ELLDs. In a rigid pip-
ing system, very small losses of liq-
uid will produce fairly large pressure
drops, because the volume of the pip-
ing is relatively constant over the
operating range of pressures that
submersible pumps produce.
In a flexible piping system, how-
ever, the range of pressures normally
encountered (0 to 30 psi) produces
relatively large changes in the vol-
ume of the piping system. Like a bal-
loon (though to a much lesser
degree), the flexible piping expands
as pressure increases and contracts as
pressure is reduced. The contraction
of the piping has the effect of main-
taining some pressure in the piping
as liquid leaks out, thereby prolong-
ing the time required for the pressure
in the pipe to drop a given amount.
In a rigid piping system, the
pressure drop in the piping due to a 3
gph leak is quite rapid. In a flexible
piping system, the pressure drops at
a much more leisurely pace. Because
pressure-based ELLDs monitor pres-
sure change over time to determine
whether a leak is present, the device
must use a longer test interval to
detect a 3 gph leak in flexible piping
than in rigid piping.
Pressure-based ELLDs must be
programmed at installation so that
the length of the test interval is
adjusted for the flexibility and length
of the piping system in which it is
installed. In some models, this infor-
mation must be entered into the
device manually. In other models, a 3
gph leak is created in the piping at
the time of installation, and the
device is operated in a "learn" mode,
whereby a series of tests are run to
empirically determine the length of
the test interval, based on the pres-
sure decay curve that is actually pre-
sent.
How Pressure-Based ELLDs
Test Themselves
There is no question that pressure-
based ELLDs can conduct a certain
amount of self-testing. Because the
device controller can operate the
pump on its own, the controller
knows that when the pump is off, the
pressure should be at the approxi-
mate holding pressure of the check
valve; when the pump is on, the pres-
sure should be at the operating pres-
sure of the pump. If the measured
pressure is outside of these ranges,
then the controller knows that some-
• continued on page 26
~ 25
-------
LUSTLine Bulletin 36
• Tank-nically Speaking
from page 25
thing is not right and a warning can
be activated.
The warnings will not only deter-
mine whether the transducer is mal-
functioning, but may also identify
other system problems [e.g., running
out of product (pump on, but line
pressure too low), or a defective
pump control that keeps the pump
motor running all the time (pump
supposedly off, but line pressure too
high)].
The comparison of expected ver-
sus actual pressure readings is typi-
cally conducted as part of the
protocol for the 0.2 or 0.1 gph tests.
The 0.2 gph tests are initiated when-
ever the system has not pumped
product for some period of time (typi-
cally about a half hour) and in all but
the most active 24-hour facilities are
usually conducted on a nightly basis.
Thus the operating condition of the
transducer is typically evaluated on a
nightly basis, and the successful com-
pletion of a 0.2 gph test is a reason-
ably reliable indicator that the ELLD
is functioning properly.
Although I have not investigated
all available brands of ELLDs, I
expect that there is significant varia-
tion in the sophistication of the self-
testing that is conducted by the
different models. In addition, many of
these devices have evolved over time
so that earlier software versions may
not self-check to the same level as
later versions.
To guard against improper pro-
gramming, some ELLD models
establish their default piping (the
type of piping assumed to be present
unless the installer reprograms the
ELLD) as a fairly long run of the most
flexible piping type. This assumption
lengthens the test interval signifi-
cantly, which is likely to result in fre-
quent "false alarms" if the piping is,
in fact, a more rigid variety. Little can
be done to thwart the person who
might intentionally program the
ELLD for operation in a rigid piping
system (when the piping is actually a
flexible variety) to reduce this "false
alarm" rate.
Flow-Based ELLDs
Flow-based ELLDs typically work by
keeping the pump motor operating
after the customer has hung up the
__
nozzle. This procedure maintains the
piping at operating pressure. The
ELLD controller then closes an isola-
tion valve at the pump end of the
line. The closing of this valve sepa-
rates the piping into what I will call
the "pump side," which is very short
and extends from the pump motor to
the isolation valve (in some cases the
valve is inserted in the pump mani-
fold in the traditional LLD port), and
the "dispenser side," which contains
the bulk of the piping and extends
from the isolation valve at the pump
to the dispenser.
After the isolation valve is
My gut instinct is that the number of
instances where a leak was missed
by a malfunctioning ELLD will run a
distant third to the instances where
! a warning light resulted in a service
call or was completely ignored.
i
closed, the pump side and the dis-
penser side remain open to one
another via a small passageway in
which a flow-sensing device is
installed. The pump motor continues
to run during the test period to main-
tain a constant (operating) pressure
on the pump side of the isolation
valve. In a tight piping system, the
dispenser side of the isolation valve
will maintain the original (operating)
pressure, and there will be no flow
through the small passageway,
because the pressures on both the
pump and dispenser sides of the pip-
ing will remain equal.
If the dispenser side of the piping
has a leak, however, the pressure will
drop on the dispenser side of the iso-
lation valve. Liquid will now flow
through the flow-sensing pathway,
because the pressure is greater on the
pump side than the pressure on the
dispensing side of the piping. This
flow rate is measured. If it exceeds
the threshold set for the device, a
failed test is declared.
In flow-based ELLDs, the pres-
sure in the entire piping run during
the test period is maintained at a con-
stant level, because any product
leaked from the dispenser side of the
piping will be replaced with product
from the pump side. Because the test
pressure is constant, there is no need
to take into consideration variations
in leak rate due to pressure changes
in the pipe. The ELLD test protocol is
the same whether the device is
installed in rigid or flexible piping.
How Flow-Based ELLDs Test
Themselves
Flow-based ELLDs are capable of
some fairly rigorous self-tests. In
some devices, after the completion of
a 3 gph test, a small bypass valve on
the dispenser side of the isolation
valve is opened. This valve generates
a calibrated 3 gph leak of product
from the dispenser side back into the
tank. The ELLD then checks whether
it can correctly measure this simu-
lated leak with the flow sensor. If it
can, then everything is fine; if it can't,
then the device communicates a
warning to the operator that the sys-
tem is not operating properly.
Another flow-sensing brand sim-
ply keeps the flow-sensing pathway
open while the pump is dispensing
fuel and checks whether the flow
sensor registers flow. While this
approach is not as quantitative as the
approach described previously, this
particular flowmeter has no moving
parts, so calibration is not a big issue.
And the Answer Is...
So, are ELLD self-tests sufficient to
meet the regulatory standard of
"annual test of operation—conducted
in accordance with the manufac-
turer's requirements" or not? Well, it
depends...
For Pressure-Based ELLDs
For pressure-based ELLDs, I think
the answer is a little murky. Some
devices seem to offer a reasonably
comprehensive test of operation. The
main omission is that the ability to
detect a 3 gph leak in a specific pip-
ing run is not determined. But the
EPA interpretation of the LLD test
requirement is that a specific leak
rate does not need to be detected
(http: / / www.epa.gov / swerustl /
compend/rd.htm, question 16). Thus
this omission does not seem signifi-
cant according to EPA's reading of
the rules.
I suspect that some other brands
and older models of pressure-based
ELLDs probably fall short of a thor-
ough self-test. I can think of two
-------
LUSTLine Bulletin 36
items that would serve as helpful
compliance evaluation tools for both
inspectors and storage system own-
ers. One would be a list of ELLD
devices that includes the manufac-
turer's official recommendations for
the "annual test of operation" for that
device. This information would be
useful to compliance inspectors who
need to know whether documenta-
tion of a field test must be produced
or whether the manufacturer believes
that the ELLD's internal testing is
sufficient to meet the regulatory
requirements.
The second item would be an
independent review of each ELLD
model that a manufacturer claims is
"self-testing" to evaluate whether
that claim seems reasonable or spe-
cious. These items are beyond the
scope of this article. If there is enough
interest, however, perhaps EPA
could fund such a review, or the
National Work Group on Leak Detec-
tion Evaluations might consider
undertaking such a review.
For Flow-Based ELLDs
Flow-based ELLDs that generate
quantitative leaks and determine
whether they are correctly detected
should meet most everyone's defini-
tion of a self-test. Flow-based ELLDs
that use a nonquantitative flow test
seem to provide a reasonable test of
operation. All of the flow-based
ELLDs of which I am aware fall into
one of these two categories.
For Those Who Are Still
Dissatisfied
My reading of the regulations is that
the privilege of deciding what a "test
of operation" is rests with the manu-
facturer of the device. The disagree-
ment arises because of the varying
definitions of "operation." My dictio-
nary says that "operation" means
"the quality or state of being func-
tional," and that "functional" means
"performing or able to perform its
regular function." These definitions
leave lots of room for manufacturers
and regulators to have differing opin-
ions about what is meant by "test of
operation."
To tilt the definition in the regu-
latory direction would require a spe-
cific definition of "test of operation."
Such a definition might read, ""test of
operation' shall mean a procedure to
confirm that a LLD will detect a leak
of 3 gph by simulating a 3 gph leak
and verifying that the LLD responds
by shutting off the flow, restricting
the flow, or sounding an alarm. Man-
ufacturer's recommendations should
be followed when conducting the test
of operation." In the absence of a reg-
ulatory definition for "test of opera-
tion," however, the federal
regulations and the dictionary give
the manufacturer of the LLD wide
latitude in setting its own definition.
My Two Cents
My own thinking has evolved such
that I would rather have a well-engi-
neered, self-testing device that evalu-
ates itself on a daily (or nearly daily)
basis than a non-self-testing device
that is evaluated by a person of
uncertain competency on an annual
basis.
I've been somewhat reassured in
researching this article that the most
popular ELLDs do a fair amount of
self-checking that will realistically tell
the facility operator (if he or she is in
a mood to listen) whether the ELLD
is functional. While the self-checking
may not take into account all possible
contingencies, there seem to be clear
benefits to automatic self-testing ver-
sus an annual test of operation con-
ducted by a fallible human.
Either way, there are no guaran-
Either way, there are no guaran-
tees that every leak will be detected in
a timely fashion. As former OUST
employee David Wiley has been
known to say, "We should not let the
perfect become the enemy of the
good."
It would be helpful to have some
real-world data that would reveal the
number/percentage of the following
events:
• The number of piping leaks
that have occurred where the
ELLD did not detect the leak;
• The number of times service
personnel have responded to
customer reports of ELLD
warning lights or messages;
and
• The number of times service
personnel responding to other
problems have discovered
ELLDs in warning mode.
My gut instinct (whaf s yours?) is
that the number of instances where a
leak was missed by a malfunctioning
ELLD will run a distant third to the
instances where a warning light
resulted in a service call or was com-
pletely ignored.
Maybe what is really needed is
an annual test to verify that facility
operators understand their leak
detection equipment. •
ELLDs - Tips for Inspectors
If an internal diagnostics system detects a problem, some ELLDs will
not proceed with the more sensitive 0.2 gph test. If an ELLD has not
completed a 0.2 gph test in the last week, it may indicate that die device
is not working correctly (unless the facility is extraordinarily busy).
Read alarm messages carefully (this task may require consulting an
owner's manual or technical manual for the device) to understand
what the ELLD is telling you (and the owner).
For pressure-based ELLDs, consult the programming manual for the
device to find out how to verify that the type and length of piping that
have been programmed into the ELLD are consistent with the piping
actually present at the facility.
Get a copy of the California State Water Resources Control Board's
(SWRCB's) new booklet, Understanding Line Leak Detector Systems. The
booklet discusses the technological principles of LLDs and provides an
overview of installation, inspection, maintenance, and special features
of these devices. While you're at it, check out SWRCB's Understanding
Automatic Tank Gauging Systems booklet. To obtain copies of either
booklet, fax your request to the SWRCB UST Program at'(916) 341-
5707, call (916) 341-5775, or visit its Web site at http:/ /www.swrcb.
ca.eov / cwphome / ust. •
27
-------
LUSTtitie Bulletin 36
by Robert N. Renkes, Executive Vice President, Petroleum Equipment Institute
PEI'S RECOMMENDED PRACTICE FOR UST SYSTEM INSTALLATION (RP100) REVISED
The federal underground storage tank regula-
tions (40 CFR 280.20) require that all USTs and
piping subject to these rules be properly
installed in accordance with a code of practice devel-
oped by a nationally recognized association or inde-
pendent testing laboratory and in accordance with the
manufacturer's instructions.
Of the three recommended practices mentioned in
40 CFR 280, only one, PEI's Recommended Practices for
Installation of Underground Liquid Storage Systems (FBI
RP100-2000), has been updated and revised about
every three years since it was first issued in 1985.
RP100 has been revised once again and is now
available to individuals and firms interested in the lat-
est guidance on proper techniques and methods for
installing UST systems.
The group responsible for writing the recom-
mended practice, PEI's Tank Installation Committee,
reviewed over 160 comments—with more than 40 per-
cent coming from regulators—and made many
changes throughout the document. Here are a few
examples of the revisions:
• A more extensive list of post-installation testing
recommendations is included in the reformatted
chapter on testing.
• The requirement that the bottom of the excava-
tion be covered with leveled backfill material
has been removed. Instead, the instructions read
that tanks should be installed to facilitate water
removal. In addition, a new section asks the
installer to consider the installation of a water-
gauging port at the end opposite the tank fill
pipe. (See diagram below.)
• New and stronger warnings on the use of vent
restriction devices have been incorporated into
the chapter on piping.
• A new section on fill piping has been added.
• About a dozen drawings and figures have been
revised.
• Substantial releases from secondarily contained
pressure piping systems have occurred because
of the failure of interstitial sensors or leakage of
product out of the secondary containment
before the product could reach the sensor. The
document now recommends that line leak detec-
tors be installed on all pressurized piping sys-
tems, including those with secondary
containment.
RP100-2000 supersedes the previous recom-
mended practice of the same name, published in 1997.
Copies are available for $30 (includes postage and
handling) from PEI, P.O. Box 2380, Tulsa, Oklahoma
74101-2380. Phone: (918) 494-9696. Fax: (918) 491-9895.
On-line orders can also be placed on PEI's secure
server at wTvw.pei.org / catalog. •
Piping Dimensions
Water gauging 6min.—
port
Dimensions shown are generally accepted limits. Sloping all lines to the tank facilitates future testing. (Revised diagram.)
28
-------
LUSTLine Bulletin 36
MTBE
Santa Clara Study Examines Causes of
Fuel/MTBE Releases at Active UST Sites
In May, the Santa Clara Valley
Water District released a report
titled An Evaluation of MTBE
Occurrence at Fuel Leak Sites with
Operating Gasoline USTs. The report
presents the results of a study the
District undertook because of con-
cern over a disturbing rate of occur-
rence and magnitude of MTBE
releases at facilities where gasoline
USTs are still in use, compared with
sites no longer storing or dispensing
fuel. The study was designed to eval-
uate potential causes of significant
MTBE contamination at operating
UST facilities, including evaluation
of the occurrence of undetected
releases and identification of actual
and potential weaknesses in fuel
storage, management, and delivery
operations.
To identify the most reasonable
source and cause of the MTBE conta-
mination, available MTBE data for
more than 150 LUST sites with oper-
ating USTs were reviewed. A final
study population of 16 LUST sites
with operating UST systems and high
concentrations of MTBE in ground-
water was selected. The sites were
evaluated in detail to determine
whether new releases were occurring
and not being detected by the moni-
toring systems.
The results of the 16 case studies
reveal that 13 sites are suspected of
having an undetected release from
the current UST system, and 2 sites
are suspected of having a pre-
upgrade release. Releases at two of
the sites with significant MTBE
plumes were discovered only
because of significant nuisance con-
ditions that warranted investigation.
In one case, vapors in an off-site
sewer prompted investigation of the
USTs. In the other case, detection of
MTBE in a nearby municipal well
prompted an investigation. The
release locations for these two sites
have been confirmed.
In one case, a leak in the primary
piping drained to the submersible
turbine pump but was not contained
or detected. In the other case, the
owner identified the release mecha-
nism as a vapor release associated
with the vapor recovery system, as
well as potential breaches in various
portions of the system.
A specific release mechanism
could not be definitely identified for
the other 14 sites. At two sites where
detailed testing of the secondary con-
tainment portions of the systems was
conducted, components, such as dis-
penser pans and turbine sumps, were
not liquid-tight, but a release from
these components was not con-
firmed.
An attempt was made to deter-
mine the most likely source location
and release scenario resulting in sig-
nificant MTBE concentrations at each
of the 14 sites. An evaluation of the
available data indicated the follow-
ing suspected source areas:
• Twelve of the sites appeared to
have source locations around
the tank complex. Five of these
are suspected of having releases
associated with inadequate or
absent sumps around the sub-
mersible turbine pump.
• Two sites appeared to have
problems associated with lined
trenches.
• Two sites appeared to have
source areas located around the
tanks and piping.
• Three sites appeared to have
source areas near the
piping/ dispenser.
• One site had inconclusive data.
At least 7 of the 16 sites had
received notices of violation or had
problems noted by an inspector
related to monitoring system compo-
nents. Four other sites had under-
gone third-party inspections during
which components were found to be
either not liquid-tight or not vapor-
tight.
Increased Vigilance Needed
Inadequate, or lack of, secondary
containment and improper monitor-
ing seemed to be the common theme
associated with suspected unde-
tected releases. The study concluded
that the results of this and similar
studies should be used to assess the
threat to water resources posed by
storage of fuels in USTs and the oper-
ation of fueling facilities.
Whereas most of the cases exam-
ined in this study involved double-
walled tanks and piping—14 of the
16 sites had double-walled tanks, 10
had double-walled piping, and 6 had
lined piping trenches—the study rec-
ommended that particular attention
be paid to single-walled UST systems
that have no protection against low-
volume leaks that may not be
detected by monitoring or integrity
testing. Conversely, the mere pres-
ence of a double-walled UST system
does not guarantee protection against
undetected and/or significant re-
leases.
The study also recommended
that although operating UST facili-
ties are not currently required to con-
duct environmental monitoring at
non-LUST sites, such monitoring
should be considered the only
method to confirm the presence of
MTBE in groundwater and evaluate
the threat that past and continuing
undetected releases pose to ground-
water resources.
The report is available on the
District's Web site at http://www.
scvwd.dst.ca.us / wtrqual / factmtbe.
htm. A slide presentation based on
the report is also available from the
same Web site. •
;_ For more information, contact
James S. Crowley, P.E.,
g__ Engineering Unit Manager
^Leaking Underground Storage
^Tank Oversight Program, Santa
g* Clara Valley Water District, at
(408)265-2607 (x2638).
.;„ J
-------
LUSTLine Bulletin 36
"There's Still a Lot of Work to Do"
OUST Director Announces New
UST Program Initiatives
It seems clear that EPA's new
Director of the Office of Under-
ground Storage Tanks (OUST),
Cliff Rothenstein, has a mission. He
is no stranger to the goings-on of
EPA's storage tank program.
Rothenstein assumed his position at
OUST after serving three years as
Deputy Assistant Administrator for
the Office of Solid Waste and Emer-
gency Response (OSWER), where he
worked on EPA and the Administra-
tion's positions and strategies on
Superfund, solid and hazardous
waste, and UST legislation. He was
also the agency's chief negotiator on
Superfund and solid waste legisla-
tion, directing the Administration's
congressional negotiating teams.
Since taking up up his new role at
OUST in June, Rothenstein has been
quick to get down to business. "I
think if s great," he says. "This pro-
gram presents a whole new and dif-
ferent type of challenge. There is a lot
of good work being done here. I just
want to help the program gain a little
more visibility, so people realize there
is still more work to be done. We've
done some creative things in the past,
and we'll continue to try to do that."
As Director, Rothenstein is gear-
ing up to implement four new UST
program initiatives:
• Launching numerous UST-
fields pilots to assess, clean up,
and foster redevelopment of
abandoned or closed UST sites;
• Taking steps to improve and
maintain UST system compli-
ance with spill, overfill, corro-
sion protection, and leak
detection requirements;
• Taking steps to accelerate
cleanups for the 160,000 petro-
leum releases that have already
occurred; and
• Evaluating the performance of
UST systems to determine
what's working and what, if
any, regulatory changes should
be made.
On October 23, Tim Fields,
EPA's Assistant Administrator for
30
OSWER, sent a memorandum to all
10 of EPA's Regional Division Direc-
tors for Underground Tanks dis-
cussing the four initiatives. See EPA's
Web site at www.epa.gov / oust /
and click on the "What's New" but-
ton to view the complete memo.
"Right now, we will continue to
work with the regions and states to
develop the details on these initia-
tives," says Rothenstein.
OUST will step up its USTfield
efforts by conducting a series of pilot
programs where the states and the
local communities work together to
perform site assessments of the aban-
doned tank universe in selected
cities, particularly in areas where
communities are trying to promote
redevelopment.
As part of this initiative, EPA
announced on November 2 that it
• has targeted communities in 10
states to receive $100,000 each for
assessment and cleanup of aban-
doned tanks. The 10 communities
include: Nashua, New Hampshire;
Trenton, New Jersey; Wilmington,
Delaware; Anderson, South Car-
olina; Chicago, Illinois; Kansas City,
Missouri; Albuquerque, New Mex-
ico; Salt Lake City, Utah; Oakland,
California; and Portland, Oregon.
EPA plans to select 40 more UST-
fields pilot projects in 2001.
"We will focus on filling in the
gap at Brownfield sites, where tanks
exist but Superfund money cannot
be used," says Rothenstein. "Almost
every Brownfield site has some aban-
doned tanks. If they find them, they
can't do anything about them,
because they can't use that money.
Our USTfields initiative will allow
the state and the community to actu-
ally do the assessment arid maybe
even the cleanup at those tank loca-
tions."
"As for increasing compliance,
we might look to establishing tar-
gets, whereby we try to get more and
more compliance each year, "says
Rothenstein. "We need to come up
with a good, common definition of
'operational compliance/ so that
Oust Director Cliff Rothenstein.
everybody's working off of the same
page—to the extent that we can. We
might see if there is a way to pro-
mote the idea of multisite compli-
ance agreements with owners—
maybe for cleanup, too^-so that you
basically negotiate one agreement
that applies everywhere and save
time and resources."
"I know a lot of cleanups have
been done, but there is still a bit of a
backlog," he explains. "We're look-
ing at different ideas—perhaps
establishing annual targets for clean-
ing up sites. We envision that the
regions and the states will work
together to establish these targets.
We are also looking for ways to use
innovative approaches, such as pay
for performance, to speed up
cleanups."
He says the agency will go back
and evaluate the UST requirements.
"We won't get all the answers imme-
diately, but we want to focus on
some of the critical questions so we
can at least get some preliminary
answers."
Rothenstein has a mission for the
tank program. "I think there's been a
sense that this program's been work-
ing well and nothing needs to be
done," he says. "But frankly, there
are 160,000 releases out there that
still need to be cleaned up—far more
than Superfund or RCRA combined.
There are still a lot of active tanks
that are not in compliance. There's a
whole universe of abandoned tanks
that people aren't really sure what to
do about—part of the reason for the
USTfields initiative. We've done a
great job, but there is still a lot of
work left to do." •
-------
LUSTLine Bulletin 36
Health & Safety _
Alaska Tank Explosion Linked
to Polyester Coveralls
by Ben Thomas
On May 16, 2000, a Soldotna,
Alaska, UST owner was hos-
pitalized with first- and sec-
ond-degree burns on his face after
one of three underground tanks he
was preparing for removal exploded.
The three tanks had been out of ser-
vice for some time and were mostly
empty. Anticipating the arrival of the
tank removal contractors, the tank
owner was using an electric pump to
try and remove the last bit of gasoline
from the tank when the explosion
occurred. At the time, the tank owner
was wearing an insulated jumpsuit
that was a polyester / cotton blend.
The Ignition Source
The owner had unknowingly created
what is known in the fire-fighting
business as a "fire triangle." For a fire
to occur, three critical ingredients
must be present: fuel, oxygen, and an
ignition source.
The tank already contained a suf-
ficient fuel-oxygen mixture. Because
it was mostly empty, there was plenty
of space for the gasoline at the bottom
of the tank to vaporize. When the
owner tried to remove the product,
there was probably a high level of
explosive vapors in the tank. This
environment set the stage for a dan-
gerous situation.
The owner's coveralls were 35
percent cotton and 65 percent poly-
ester—enough polyester to
create some static while the
tank owner was working.
The static buildup from the
polyester suit, plus the dry
Soldotna air, created a spark.
All the spark needed was a
fuel source and oxygen.
When the owner got near
enough to the gasoline
vapors fuming out of the fill
pipe, the static from his
^ Picture of charred pump and
baseball cap after tank explosions
sends owner to hospital in
Soldotna, Alaska
coveralls was the ignition source
needed to trigger the explosion.
Think Twice Before...
Be careful when working around gaso-
line. It may be easier than you think to
create a "fire triangle." A tank that is
mostly empty can be a greater fire haz-
ard than one that is full of gasoline.
Hire trained professionals to properly
and safely remove a UST from the
ground. And seriously consider using
cotton-only work garments when
working around gasoline. •
Ben Thomas is with the Alaska Depart-
ment of Environmental Conservation
UST program. He can be reached at
Ben Thomas@envircon.state.ak.us.
LU.S.T.LINE
Q One-year subscription. $1 8.00.
Q Federal, state, or local government. Exempt from fee. (If you wish to have LUSTLine sent to your home, please submit
your request on agency letterhead.)
Q Please take my name off your mailing list.
Q Please send me back issues of LUSTLine. Fill out name and address — no P.O. boxes. Back issues cost $3.00 per
issue or $50.00 for a complete set.
If ordering back issues, please indicate LUSTLine issue ^s _ _
Q Please send me a LUSTLine Index.
Name Company/Agency
Address _
Street City/Town State ZIP
Please enclose a check or money order (drawn on a U.S. bank) made payable to NEIWPCC.
Send to: New England Interstate Water Pollution Control Commission
Boott Mills South, 100 Foot of John Street, Lowell, MA 01852-1 124
Phone: (978) 323-7929 • Fax: (978) 323-791 9 • lustline@neiwpcc.org • www.neiwpcc.org
We welcome your comments and suggestions on any of our articles.
-------
Two New Compliance
Assistance Tools Now
Available
The EPA Office of Underground
Storage Tanks (OUST) is distribut-
ing the following two new compli-
ance assistance tools to Regional
and state UST programs to help
increase compliance levels:
• Operating and Maintaining
Underground Storage Tank Sys-
tems: Practical Help and Check-
lists (EPA 510-B-00-008) is a
50-page manual containing brief
summaries of the federal UST
requirements for operation and
maintenance (O&M), as well as
practical help that goes beyond the
requirements. Checklists prompt
the user—the UST owner or opera-
tor—to look closely at the kinds of
equipment in use and explain how
to keep that equipment working
properly over the lifetime of the
UST system. The manual provides
recordkeeping forms that also help
the UST owner and operator keep
equipment operating properly. If the
user has questions, the manual pro-
vides contact information for addi-
tional help. The O&M manual is
available in electronic form for free
viewing, printing, and downloading
at wwTV.epa.gov / swerustl / pubs /
index. htm#ommanual. The manual
will be updated periodically, and the
most current edition will always be
available at this Web site.
• Automatic Tank Gauging Sys-
tems for Release Detection: Ref-
erence Manual for Underground
Storage Tank Inspectors (EPA 510-
B-00-009) is a 140-page manual that
can help state and EPA inspectors of
USTs evaluate how well UST owners
and operators are using their auto-
matic tank gauging (ATG) systems to
comply with release detection
requirements. The manual also pro-
vides handouts that UST inspectors
can distribute to UST owners and
operators to help them understand
the proper operation and mainte-
nance of their ATG systems.
The manual contains a sum-
mary of specifications, based on
third-party evaluations, for ATG
systems that detect leaks from USTs
and their piping. Each summary
provides information on certified
detectable leak rate/threshold, test
period duration, product applica-
bility, calibration requirements,
restrictions on reports, sample
reports, and more. The ATGrnan-
ual is available in electronic form
for free viewing, printing, and
downloading at www.epa.gov/
swerustl / pubs / index.htm#auto-
matic. The manual will be updated
periodically, and the updated ver-
sion will always be available at this
Web site.
Both manuals can be adapted and
tailored by states or industry to match
precisely their specific requirements
(for assistance, contact Jay Evans at
703-603-7149 or evans.iav@eva.sov).
LU.S.T.UNE
New England Interstate Water
Pollution Control Commission
Boott Mills South
100 Foot of John Street
Lowel^ MA 01852-1124
Forwarding and return postage guaranteed.
Address correction requested.
LU.ST.UNE INDEX
Auctut ItSiDuIIotln II -June SOOO/BaUgtln 135
The updated LUSTLine
Index—the long and
action-packed story of
USTs and LUSTs—
is now available.
Call NEIWPCC
for your copy at
(978) 323-7929
------- |