Amoco-U.S. EPA  Pollution Prevention  Project
     Yorktown, Virginia.  Project Summary
     Amoco Corp.,  Chicago,  IL




     Prepared for:

     Environmental  Protection Agency, Washington,  DC




     Jun 92
U.S. Department of Commerce
National Techmca! Information Service

-------
             REPORT  DOCUMENTATION  PAGE
                                                                        totm
                                                                        OMB NO. 0704-0188
*»9 Dw'df n *0' thu (Oiif
 mamumirtg ihr d«i«
       PB92-228527
                                                               bu/df A
                      of iMo'mifiO" •* ntimjip0 to i*f*gf i *o«f &*' 'f^DO^v*. including thf ti**f 'o/ ffvi
                     fd. *nd co*D"«> OQ2. jnd 10 I^f 0"«f 0* M*n«
-------
Amoco-U.S. EPA
Pollution  Prevention Project
Yorktown, Virginia
Project Summary
     REPRODUCED BY
     U.S. DEPARTMENT OF COMMERCE
     NATIONAL TECHNICAL INFORMATION SERVICE
     SPRINGFIELD. VA. 22161

-------
AMOCO/USEPA POLLUTION PREVENTION PROJECT


             Project Summary
              January,  1992
          (Revised June, 1992)

-------
                        Table of Contorts

                                                             Page

List of Figures                                                ii
List of Tables                                                 iv
Foreword                                                       vi
Executive Summary                                             vii

1.0  Project Summary                             .             1-1

     1.1  Project Goals                                       1-1
     1.2  Project Organization,  Staffing and Budget           1-2
     1.3  Lessons and Results                                 1-3
     1.4  Recommendations                                    1-15

2.0  Refinery Release Inventory                               2-1

     2.1  Sampling Program Results                            2-1
     2.2  Source Identifications                              2-7

3.0  options Identification and Analysis                      3-1

     3.1  Identification                                      3-1
     3.2  Option Characteristics                              3-1
     3.3  Analyses and Impacts                                3-3
     3.4  Risk Assessment                                     3-8
     3.5  Public Perceptions                                 3-11

4.0  Ranking Method and Results                               4-1

     4.1  Option Characteristics                              4-1
     4.2  Single Criterion Rankings                           4-2
     4.3  Multiple Criteria Ranking                           4-5
     4.4  Project Options and Regulatory Requirements         4-9
     4.5  Summary of Ranking Results                         4-10

5.0  Obstacles and Incentives                                 5-1

     5.1  Reduce Barge Loading Emissions                      5-6
     5.2  Install Secondary Seals on Storage Tanks            5-8
     5.3  Upgrade Slowdown Stacks                             5-9
     5.4  Reduce Soil Intrusion into Drainage System         5-11
     5.5  Institute a LDAR Program                           5-12
     5.6  General Observations                               5-14

6.0  References                                               6-1

                        List of  Appendices
A.   Yorktown Site Background                                 A-l
B.   Project Limitations and Exclusions                       B-l
C.   Project Documentation                                    C-l
D.   potential Emergency and Upset Refinery Incidents         D-l
E.   Common Abbreviations                                     E-l

-------
                         List of Figures

Section 2.0

2.1  Simplified Flow Diagram with Release Sources

2.2  Pollution Prevention Sampling Program

2.3  Pollutant Generation Within the Yorktown Refinery

2.4  Pollutant Transfers, Recycling, and Treatment within the
     Yorktown Refinery

2.5  Total Releases Entering the Environment from the Yorktown
     Refinery

2.6  Total Airborne Emissions - Yorktown Refinery

2.7  1989 TRI Compared to Measured Emissions-

2.8  Selected Airborne Hydrocarbon Sources:   Estimated vs.
     Measured Values

2.9  Yorktown Refinery VOC Air Emission Sources

2.10 Major Sources of Refinery Solids

2.11 Solids Management

2.12 Solids Management by Type (1990)

2.13 Phenol, Ammonia, and Oil & Grease Sources

Section 3.0

3.1  Simplified Flow Diagram with Release Sources and Control
     Options

3.2  Benzene Annual Average Concentrations

3.3  Source Culpability for Benzene

3.4  Histogram of Benzene Emissions with and without Marine
     Loading Controls

3.5  Benzene Annual Average Concentrations with Barge Loading
     Control Option

3.6  Benzene Annual Average Concentrations with Secondary Seals
     on Gasoline Storage Tanks

3.7  Benzene Annual Average Concentrations with Multiple Control
     Options
                                ii

-------
Section 3.0 - Cont'fl.

3.8  Benzene Annual Average Concentrations with API Separator and
     Sewer Emissions Controls


3.9  Projected Rates of Return—Pollution Prevention Projects—
     Yorktown

Section 4.0

4.1  Hierarchy and Criteria Weights Used for Ranking

4.2  Comparison of Criteria Weights - EPA, Virginia and Amoco

Appendix A

A-l  Yorktown Refinery Location

A-2  Yorktown Refinery Block Flow Diagram

A-3  Air Emission Sources In Yorktown Area

A-4  Number of Refineries by Capacity Groups

A-5  Number of Refineries by State

A-6  Refinery Age Distribution
                               111

-------
                          List of Tables

                                                             Page

Section 1.0

1.1  Project Components                                      1-20

1.2  Project Participants                                    1-21

1.3  Comparison of Different Environmental Management
     Options for the Yorktown Refinery.                      1-23

Section 2.0

2.1  Comparison of Annual Average Predicted Impacts          2-10
     to Typical Measured Concentrations for Different
     Types of Environments for BTEX Chemicals

2.2  Reconciliation of 1989 TRI Report and Pollution         2-11
     Prevention Inventory Yorktown Refinery

2.3  Techniques Used to Determine Airborne Emissions         2-12

Section 3.0

3.1  Pollution Prevention Options from the Williamsburg      3-14
     Workshop

3.2  Selected Pollution Prevention Engineering Projects      3-16

3.3  Release Reductions from Pollution Prevention Projects   3-19

3.4a Financial Summary for Pollution Prevention Projects     3-20
     (15 percent discount rate)

3.4b Financial Summary for Pollution Prevention Projects     3-21
     (10 percent discount rate)

3.4c Financial Summary                                       3-22

3.5  Maximum Annual Average Predicted Concentrations for     3-23
     the Yorktown Refinery for BTEX Chemicals

3.6  Maximum Annual Average Benzene Concentrations and       3-24
     Benzene Exposure Associated with Various Pollution
     Prevention Options

3.7  Copper Concentrations in Yorktown Refinery Effluents    3-25
     (concentrations in ug/1)

Section 4.0

4.1  Amoco/EPA Pollution Prevention Option Characteristics   4-12


                                iv

-------
                                                             Page

4.2a. Rankings Based on Release Reduction and Exposure        4-13
     Reduction

4.2b Rankings Based on Annualized Cost and Net Cash Flow     4-14

4.3  Analytical Hierarchy Process (AHP) Ranking Using        4-15
     Workgroup Weights

4.4  Comparison of Analytical Hierarchy Process Rankings     4-16
     Using Amoco and EPA/VA Criteria Weights

4.5  Regulatory Requirements Options                         4-17

4.6a Cost-Effective Release Reduction Ranking                4-18

4.6b Comparisons of Cost-Effectiveness and Net Cash          4-19
     Flow Rankings

4.7  Cost-Effective Benzene Exposure Reduction Ranking       4-20

4.8  Option Scores by Ranking Technique                      4-21

Section 5.0

5.1  Summary of Obstacles and Incentives for Five            5-15
     Pollution Prevention Options

Appendix A

A-l  Yorktown Refinery Process Unit Capacities                A-9

A-2  Summary of Ozone Monitoring Data in the Vicinity        A-10
     of the Yorktown Refinery

A-3  Amoco Yorktown Refinery, VPDES Permit Parameters        A-ll

Appendix B


B-l  Comparison of Maximum Predicted Impacts of Criteria     B-10
     Pollutants to National Ambient Air Quality Standards

-------
                             Foreword

This volume provides a summary of work completed during a

voluntary, joint AMOCO/USEPA Pollution Prevention Project

undertaken at Amoco Oil Company's Yorktown, Virginia Refinery.

Overall goals of the Project were to (1) inventory releases of

all pollutants to the environment from the Refinery; (2)  develop,

evaluate and rank process, maintenance and operating options that

reduce these releases; and (3) identify barriers and incentives

to implementing the alternatives identified.



Special thanks are due to the AMOCO/USEPA workgroup who provided

Project oversight and direction during this two-year, $2.3

million effort.  In addition, more than 200 people, from 35

organizations participated at various times in this unique

Project.  Their enthusiasm and contributions are obvious from the

wealth of ideas developed, considered and analyzed.  Their

assistance supports a central belief of this Project:  that

developing effective solutions to complex environmental

management problems will take the best efforts of the many

'partners' in our society.  We extend a personal thanks to all

participants.
Howard Klee, Jr.                             Mahesh Podar
Amoco Corporation                            USEPA
                                VI

-------
             Amoco/USEPA Pollution  Prevention Project

                        EXECUTIVE SUMMARY

In late 1989, Amoco Corporation and the United States
Environmental Protection Agency began a voluntary, joint project
to study pollution prevention opportunities at an industrial
facility.   The Amoco/EPA Workgroup, composed of EPA, Amoco and
Commonwealth of Virginia staff, agreed to use Amoco Oil Company's
refinery at Yorktown, Virginia, to conduct a multi-media
assessment of releases to the environment, then to develop and
evaluate options to reduce these releases.  The Workgroup
identified five tasks for this study:

l.   Inventory refinery releases to the environment to define
     their chemical type, quantity, source, and medium of
     release.

2.   Develop options to reduce selected releases identified.

3.   Rank and prioritize the options based on a variety of
     criteria and perspectives.

4.   Identify and evaluate factors such as technical,
     legislative, regulatory, institutional, permitting, and
     economic, that impede or encourage pollution prevention.

5.   Enhance participants' knowledge of refinery and regulatory
     systems.

            Project  Organization/ Staffing/  and  Budget

Workgroup:  Monthly Workgroup meetings provided Project
oversight, a forum for presentations on different Project
components, and an opportunity for informal discussion of
differing viewpoints about environmental management.  Although
attendance varied, each meeting included representatives from
various EPA offices, the Commonwealth of Virginia, and Amoco.

Peer Review;  At the Workgroup's request, EPA arranged for
Resources for the Future to assemble a group of outside
scientific and technical experts.  This Peer Review Group
provided evaluation and advice on the Project workplan, sampling,
analysis results, and conclusions.   Members of this group were
paid a small honoraria for their participation by EPA.

Workshop;  A special Workshop, held during March 24-27, 1991 in
Williamsburg, Virginia, reviewed sampling data and identified
reduction options and ranking criteria.  More than 120 people
from diverse backgrounds--EPA, Amoco, Virginia,  academia and
public interest groups—attended the Workshop.


                               vii

-------
Participants;  More than 200 people, 35 organizations,  and many
disciplines were involved in this Project.   This reflected a
central belief of this Project that solving difficult
environmental problems must draw on many of society's "partners."

Cost;  Total cost for this Project was approximately $2.3
million.  Amoco Oil Company provided 70 percent of the funding
and EPA the remainder.

                       Lessons and Results

Refinery Release Inventory

A.   Existing estimates of environmental releases were not
     adequate for making a chemical-specific, multi-media,
     facility-wide assessment of the Refinery.

B.   Pollutant releases to the environment are lower than
     pollutant generation and internal transfers.

C.   The Toxic Release Inventory database does not adequately
     characterize releases from this Refinery.

D.   Site specific features, uncovered determined during the
     facility-wide assessment, affect releases and release
     management options.

Reducing Releases

A.   A workshop approach, drawing on a diverse group representing
     government, industry, academic, environmental, and public
     interests, developed a wide range of release reduction
     options in a multi-media context more quickly than either
     EPA or industry alone would do.

B.   Pollutant release management frequently involves the
     transfer or conversion of pollutants from one form or medium
     to another.

C.   Source reduction is not necessarily practical for all
     release management options, despite its cost effectiveness.

D.   In this Refinery, few release reductions pay for themselves.

Choosing Alternatives

A.   Ranking the options showed that better environmental results
     can be obtained more cost-effectively.  At this facility,
     about 95 percent of the release reductions required by
     regulatory and statutory programs can be achieved for 20-25
     percent of today's cost for these programs.
                               Vlll

-------
B.   All participants agreed on which options were the most
     effective and which were least, regardless of their ranking
     criteria or institutional viewpoints.

Obstacles and Incentives

A.   EPA may not have the statutory flexibility to simply set an
     emissions reduction "target" without prescribing how this
     target should or could be met.  If a target involves
     releases in multiple media, current administrative
     procedures discourage a coordinated approach, including the
     analysis of risks, benefits, and costs of managing residual
     pollutants in different media.

B.   Legislative and regulatory programs do not provide
     implementation schedules compatible with design,
     engineering, and construction timeframes.  Consequently,
     short-term "fixes" are used at the expense of more effective
     solutions.

C.   Congress, EPA, and much of industry have developed a mind-
     set and comfort level with conunand-and-control,  end-of-pipe
     treatment approaches based .on twenty years of experience.
     Many of today's problems could benefit from a different
     approach.

D.   Accounting systems in industry and government do not track
     environmental costs well, nor measure environmental health.
     Responsibility for pollutant generation and accountability
     for environmental protection are difficult to monitor.
                         Recommendations

 l.   Explore Opportunities to Produce Better Environmental
      Results More Cost-effectively.

 2.   Improve Environmental Release Data Collection/ Analysis
      and Management.

 3.   Provide Incentives for Conducting Facility-wide
      Assessments/ and Developing multi-media Release Reduction
      Strategies.  Such Strategies must Consider the Multi-
      Media Consequences of Environmental Management Decisions.

 4.   Encourage Additional Public/Private Partnerships on
      Environmental Management.

 5.   EPA Should Conduct Research on the Potential Health and
      Ecological Effects of VOCs and Reformulated Gasolines.
                                IX

-------
                             SUMMARY

1.1  Project Goals

In late 1989, Amoco Corporation (Amoco)  and the United States
Environmental Protection Agency (EPA)  began a voluntary,  joint
project to study pollution prevention opportunities at an
industrial facility.  The Amoco/EPA workgroup (Workgroup),
composed of EPA, Amoco, and Commonwealth of Virginia staff,
agreed to use Amoco Oil Company's refinery at Yorktown,  Virginia
(the Refinery), to conduct a multi-media assessment of releases
to the environment, then to develop and evaluate options to
reduce these releases.  The Workgroup identified five tasks for
this study:

    1.  Inventory refinery releases to the environment to define
        their chemical type, quantity, source, and medium of
        release.

    2.  Develop options to reduce selected releases identified.

    3.  Rank and prioritize the options using a variety of
        criteria and perspectives.

    4.  Identify and evaluate factors such as technical,
        legislative, regulatory, institutional, permitting, and
        economic, that impede or invite pollution prevention.

    5.  Enhance participants' knowledge of refinery and
        regulatory systems.

Figure 3.1 shows a schematic diagram of the Refinery, potential
release sources, and a number of pollution prevention options
identified in this Project.  Table 3,2 describes specific options
to reduce releases.  At the time this Project began, pollution
prevention was a concept predicated on reducing or eliminating
releases of materials  into the environment rather than managing
the releases later.  The Workgroup adopted this general concept
and agreed to consider all opportunities—source reduction,
recycling, treatment,  and environmentally sound disposal—as
potential choices in pollution management.  Since then, Congress,
in the Pollution Prevention Act of 1990, and other organizations,
have put greater emphasis on source reduction as the primary, if
not the exclusive, means to accomplish pollution prevention.

A central goal of this Project was to identify criteria and
develop a ranking system for prioritizing environmental
management opportunities that recognized a variety of factors
including release reduction, technical feasibility, cost,
environmental impact,  human health risk, and risk reduction
potential.  Due to the inherent uncertainties in risk
assessments, the Project focused on relative changes in risk

                               1-1

-------
compared to current levels, rather than establishing absolute
risk levels.  Because of difficulties in quantifying changes in
ecological impact from airborne emissions, changes in relative
risk were based primarily on human health effects indicated by
changes in exposure to benzene.  The risk assessment did not
include a quantitative analysis of VOCs due to limited
information on their health effects.

This project focused on pollution and potential risks posed by
normal operation of the Refinery and chronic exposure to its
releases into the environment.  Minimizing emergency and upset
events is a top priority of Amoco's facility managers.  Such
events can have catastrophic results.  However, they were not
studied in this project because: (a) prevention and control of
such events involves significantly different skills, technical
resources, and analyses than controlling releases from day-to-day
operations (AIChE, 1985); (b)  the number, type, and frequency of
incidents at Yorktown is very low; and (c) data regarding the
type of release, and relevant meteorology during the release are
not available for analysis.  Appendix D describes potential
emergency and upset events that might occur at a petroleum
refinery and the general preventative measures used to minimize
their severity and the likelihood of their occurrence.

1.2  Project Organization, Staffing.and Budget

Project Content:  The Pollution Prevention Project has many
components.  Each component defines and addresses an issue
associated with pollution prevention and facility management
choices.  These include pollutant source identification,
sampling, exposure modeling, risk assessment, etc.  Table 1.1
provides a complete list of the components in this Project.  The
Project workplan outlined the purpose and content for most of
these components  (Amoco/EPA, 1990).

Exclusions/Limitations:  A number of areas specifically excluded
or limited in this Project are described in Appendix B.  Some are
listed below:

•   Limited sampling time and data provided a "snapshot" of
    releases rather than measured annual values.

•   Very few generally accepted methodologies exist for the
    sampling used to obtain a site-wide release inventory,
    particularly for measuring air emissions.  Both EPA and Amoco
    concerns about specific sampling issues are highlighted in
    Appendix B and discussed in more detail in Air Quality Data,
    Volume II (Amoco/EPA, 1992 b).

•   The Project considered available technologies rather than
    exploring innovative techniques for reducing releases.


                               1-2

-------
•   Chemical changes of airborne pollutants were not evaluated.

•   Data and analysis focused on the Yorktown Refinery.   Site-
    specific features of this facility and its emissions may not
    apply to other refineries.  Broader regional concerns were
    not evaluated.

•   The forthcoming human health risk assessment focuses on
    potential cancer risks associated with benzene exposure
    outside the facility fenceline.

Peer Review;  At the Workgroup's request, Resources for the
Future organized a group of outside scientific and technical
experts.  This Peer Review Group provided evaluation and advice
on the Project workplan, sampling, analytical results, and
conclusions.  Members of this group were j>aid a small honoraria
for their participation and reimbursed for travel expenses to
Washington by EPA.  A report summarizing their comments is
included as part of the documentation for this Project.   Appendix
C lists all Project documentation.

Workgroup;  Monthly Workgroup meetings provided Project
oversight, a forum for presentations on different Project
components, and an opportunity for informal discussion of
differing viewpoints about environmental management.  Although
attendance varied, each meeting included representatives from
various EPA offices, the Commonwealth of Virginia, and Amoco.

Workshop;  A special Workshop, held during March 24-27,  1991, in
Williamsburg, Virginia, reviewed sampling data and identified
reduction options and ranking criteria.  More than 120 people
from diverse backgrounds—EPA, Amoco, Virginia, academia and
public interest groups—attended the Workshop.  The Workshop
sessions resulted in suggestions that further refined and
directed Project activities (Amoco/EPA, 1991a).

Participants;  More than 200 people, 35 organizations, and many
disciplines have been involved in this Project.  Table 1.2 lists
the various participating organizations.

Cost;  Total cost for this Project was approximately $2.3
million.  Amoco Oil Company provided 70 percent of the funding
and EPA the remainder.

1.3  Lessons and Results

     1.3.1  Refinery Release Inventory

A.  Existing estimates of environmental releases were not
    adequate for making a chemical-specific, multi-media,
    facility-wide assessment.


                               1-3

-------
The Yorktown Refinery had good information about the quantity of
material released to the York River from NPDES Permit monitoring
requirements, and for solid wastes as a result of internal
programs and participation in recent American Petroleum Institute
surveys  (API, 1991b).  These releases, however, made up only 11
percent of the total releases from the facility.  Available data
did not include adequate chemical-specific characterization of
the water discharge or solid waste streams.

The Refinery (and other refineries as well) could not easily
identify specific airborne hydrocarbon compounds released or the
quantity released because:

     (a)    Refineries typically do not manufacture products with
            specific chemical compositions, and therefore do not
            routinely measure chemical compositions of their
            products or emissions.  Rather, refinery products
            have specific properties such as octane, freeze
            point, and sulfur content.  Crude oil, the raw
            material used to make these products, contains
            thousands of distinct chemicals that are never fully
            separated during the manufacturing processes.
            Airborne releases from this kind of facility are
            similarly complex.

     (b)    Most hydrocarbons are released through a large number
            of widely distributed sources  (valves, flanges, pump
            seals and tank vents).  Even a small refinery may
            have more than 10,000 potentially different sources.
            Direct measurement of each of these sources is not
            practical.

     (c)    The quantities released through any single source are
            extremely small—on the order of pounds per
            year—dilute and difficult to measure.  In addition,
            some large sources that emit pollutants in the amount
            of tons per year are difficult to measure and
            quantify.  Total hydrocarbons released from Yorktown
            Refinery from all sources were approximately 0.3
            weight percent of the total crude oil processed.
            Therefore, they would not be detected through normal
            mass balances and materials accounting  (NRC, 1990).

Thus, collecting detailed, chemical specific release information
used to characterize the Refinery was expensive and time
consuming.  This Project developed a sampling and monitoring
program that included about 1,000 samples  (see  Figure 2.2).  Each
sample was analyzed for 15-20 chemicals. The sampling program
took about 12 months to complete  at a cost of about $1 million.
Even with this time and dollar commitment, only selected sources
were sampled.  The  final release  inventory was  assembled using  a


                               1-4

-------
combination of sampling, measurements,  dispersion modeling,  and
estimates based on emission factors.

Because this sampling program was a first of its kind effort,  its
scope was intentionally broad.  Subsequent analysis showed that
not all of the information obtained was necessary to identify
significant sources and potential reduction options.  For the
Yorktown Refinery (and the petroleum refining industry overall),
more general information, such as source specific VOC emissions,
is adequate to identify many of the pollution prevention projects
developed in this study.  Total VOC emissions are a good
indicator of overall emissions and can be used for tracking
emissions reduction progress.

B.  A substantial portion of pollution generated at this refinery
    is not released to the environment.

The release inventory process allowed a comparison of pollutant
generation, on-site management and ultimate releases to the
environment.  The Refinery generates about 27,500 tons/year of
pollutants.  As a result of site hydrogeology, on-site wastewater
treatment, and solid waste recycling practices, about 12,000 tons
are recovered, treated or recycled and do not leave the Refinery
site.  Of the remaining 15,500 tons about 90 percent are released
to the air.

Figure 2.4 illustrates the transfers which take place between
generation and ultimate release.  Figure 2.5 characterizes
pollutants released from the Refinery.  This site-wide analysis of
pollutant generation and release characteristics allowed the
Workgroup to focus much of the remaining Project resources on the
largest releases—airborne emissions.

Modeling studies indicated relatively little naturally occurring
transfer of hydrocarbon emissions from air into other media
(Cohen and Allen, 1991) .  Most hydrocarbons are not very water
soluble, and so are not easily removed from the air by rainfall.
Section 2.0 includes a more detailed discussion of the potential
for transfer to other media.  Although the fate of criteria
airborne pollutants (like NOX and S02)  was not studied in this
Project, they are known to be scavenged by rainfall and can
contribute to nitrogen  loads and pH changes in lakes and soil
(See Appendix B). Measurements and modeling results showed small
transfers from some surface water ponds to groundwater.
Groundwater also enters the wastewater treatment system through
the underground sewers, resulting in a net groundwater inflow.

Transfers of pollutants between media do occur, particularly as a
result of pollution management activities.  Over 370 tons/year of
hydrocarbons initially present in wastewater streams are
volatilized into air from the water collection system. More than


                               1-5

-------
2,000 tons/year of biosolids are produced by treating wastewater
in the Refinery's activated sludge system.


C.  The TRI database does not adequately characterize releases
    from this Refinery.

Title III of SARA, Emergency Planning and Community Right-to-Know
Act, created the Toxic Release Inventory  (TRI)  in 1986.   Title
III requires regulated facilities in SIC Code 20-39 to submit
annual release data on more than 300 chemicals manufactured,
produced or otherwise used in quantities exceeding certain
threshold values.  Releases to all media must be reported.  The
TRI is one way of focusing corporate attention on release
reduction opportunities.

TRI reports are based on either emission estimates, direct
measurements or a combination of both methods.   Each facility is
responsible for the accuracy of the data reported.  Industrial
facilities frequently file amendments to TRI reports to reflect
improvements in the accuracy of the estimation and measurement
techniques.

The TRI database has become the de facto national release
inventory.  The quality and utility of data reported can vary
widely.  At a plant that uses a single solvent to wash
manufactured parts, and that purchases extra solvent every year
to make up for evaporative losses, the quantity of solvent
emissions is well known and tracked through monthly purchasing
records.  A TRI report which included this solvent and plant
should be quite accurate.  However, at the Refinery, the TRI does
not report total facility emissions because:

e   The TRI is based on estimates rather than measurements.
    Estimating accuracy varies widely.  During the measurement
    portion of this Project, several new sources were identified
    whose significance had been previously underestimated.  One
    source was identified which had been overestimated.   Figure
    2.7 summarizes the results of this analysis.

«»   The measurement phase of this Project revealed substantially
    higher TRI reportable emissions from the blowdown stacks than
    had been estimated previously.  On the other hand,
    measurements revealed that emissions from wastewater sources
    had been overestimated.  Amoco has filed an amendment to its
    past TRI reports for Yorktown to reflect new data.  Figure
    2.7 compares the starting TRI data with results obtained from
    the Project.

•   The TRI focuses on specific chemicals which account for only
    a portion of the total emissions.  In the Refinery's case,
    the TRI report covers only 9 percent of the total

                               1-6

-------
    hydrocarbons released, and only 2.4 percent of the total
    releases to all media.  Criteria pollutants—CO,  NOX,  S02,
    and PM-10—are not reportable in the TRI.

•   Some activities and emissions are exc.luded by EPA from record
    keeping requirements, such as emissions from barge loading.
    At this facility, barge loading operations account for about
    20 percent of the total benzene emissions (See Figure 3.4).

Finally, TRI provides an approximate inventory of selected
materials released to the environment.  TRI data by itself does
not allow for meaningful risk evaluation or comparisons on a
facility basis, because it does not define the facility's
relationship to nearby populations and ecosystems.

D.  Site specific features determined during the facility-vide
    assessment, affect releases and release management options.

National programs, by design, address overall problems in
specific media.  But these programs seldom consider site-specific
differences in developing standards.  Other refineries, and
indeed other industrial facilities, can use the general sampling
approach developed here to obtain the facility-wide release
inventory.  However, each site will exhibit unique geophysical
and process characteristics. Each assessment plan must include
these site-specific characteristics in its design and focus.  As
an example, the Yorktown Refinery does not have a hydrofluoric
acid (HF) alkylation unit and HF was not measured.  HF can pose a
significant health risk if managed improperly, and may need to be
tracked at facilities that use it.

Groundwater;  As a result of a clay soil layer, unique
hydrogeology, the placement of the underground drainage system
relative to the water table, and local climate, groundwater
movement at this site is minimal.  In fact, the underground
drainage system is acting as a groundwater collection unit,
sending groundwater to the Refinery's wastewater treatment plant.
Thus, groundwater at this site is not leaving the property.
Furthermore, sampling showed surprisingly low levels of
groundwater contamination, compared to other refineries (LA
Times, 1988).

Marine Loading Emissions;  Yorktown Refinery uses marine
transportation for receiving all crude oil and shipping more than
80 percent of  its products.  Estimated releases from product
loading operations are 784 tons/year of VOCs.  Computer modeling
analysis showed this source had the greatest impact on exposure
of nearby residences to Refinery hydrocarbon emissions.
Therefore, it  would be useful to include marine loading emissions
in this facility's environmental management plans.  Many other
refineries rely more on pipeline, rail and truck shipments to


                               1-7

-------
handle crude and products, and would thus not expect to find the
same potential impact from marine operations.

Airshed Status;  As discussed in Appendix A, the Refinery is
located in an airshed classified as an attainment area for all
criteria pollutants including ozone.  Therefore, relatively few
hydrocarbon emission controls have been required or installed at
this facility.  The sampling program and release reduction
options focused on hydrocarbon releases.  Many other refineries
in ozone non-attainment areas have already installed extensive
hydrocarbon emission controls.  Consequently, other facilities
may have a significantly lower percentage of hydrocarbon
emissions.  Similarly, NOX, CO, PM-10 and SO2 emissions have been
more tightly controlled in some other airsheds  (such as the Los
Angeles basin) which do not meet NAAQS for these pollutants.

    1.3.2  Release Reduction Options

A.  A workshop approach/ drawing on a diverse group representing
    government/ industry, academic/ environmental and public
    interests developed a wide range of release reduction options
    in a multi-media context more quickly than EPA or industry
    alone would do.

The release inventory described in 1.3.1 above, served as the
basis for identifying ways to reduce releases.  A 3-day
brainstorming Workshop, held in Williamsburg, Virginia generated
more than 50 potential release reduction options for the
Refinery.  These ranged from producing a single grade of gasoline
to specific technical options for particular equipment or
processes.  Table 3.1 lists all options identified.

The Workgroup subsequently narrowed this list to 12 options for
more careful, quantitative analysis.  This winnowing process
considered only those options that were technically feasible now,
offered potentially large release reductions, addressed different
environmental media, and posed no process or worker safety
problems.  Projects designed to comply with several current or
anticipated regulations were also included.  Table 3.2 lists
engineering projects included for further analysis.

The Workshop also addressed screening criteria  to help prioritize
the options, potential barriers and incentives  for
implementation, and permitting concerns.  The diverse viewpoints
brought to all these discussions helped guide subsequent Project
activities. These views reinforced the Workgroup's desire to
consider broader issues such as multi-media release management
consequences, future liability impacts, etc.  The Workshop was
able to consider these issues more comprehensively than either
government or industry alone would normally do.
                               1-8

-------
B.  Release management frequently involves the transfer or
    conversion of pollutants from one form or medium to another.

It is not at all unusual for pollutants to be converted and
transferred from one form or media to another as part of a
pollution control practice.  For example,  scrubbers used to
remove acidic pollutants from many electric utility stacks
generate large volumes of calcium sulfate sludge (EPRI, 1983)
which must also be managed.  For options developed at the
Yorktown Refinery:

•   Modifications of the underground drainage system and process
    water treatment plant  (required under the Benzene Waste
    Operations NESHAP; Federal Register, 1990) will improve
    process water treatment and reduce air emissions, but produce
    more solid waste such as biosolids and fully spent activated
    carbon.

•   The Refinery has limited sludge processing capacity.  K«>eping
    soils out of sewers would reduce the amount of sludge in the
    API Separator and thus allow for more on-site management of
    other solid wastes, reducing offsite disposal.

•   Installing an electrostatic precipitator would reduce FCU
    particulate (PM-10) emissions (catalyst fines), but transfer
    the additional collected particulates to land disposal.

•   Burning hydrocarbons that cannot be economically recovered
    generates other criteria pollutants which may also need to be
    managed.

None of these transfers or transformations are bad, in and of
themselves.  The Project simply pointed out the need to
recognize, plan, and manage these changes at an early stage of
the release management cycle.

C.  Source reduction options were more cost-effective than most
    treatment and disposal alternatives.  Nevertheless, source
    reduction alone was not adequate to achieve all the desired
    or legally required release reductions.

The Workgroup agreed to consider the waste management
hierarchy—source reduction, recycling, treatment, and safe
disposal—as the basis for developing release reduction options.
Technologies identified and analyzed fit into this hierarchy.
Time and budget constraints limited technology choices to
conventional, proven solutions rather than exploring innovative
alternatives.

However, less than half the options identified qualified as
"source reduction."  Had the options been limited to only source
reduction, the scope of potential opportunities for reducing

                               1-9

-------
releases and improving environmental quality would have been
unnecessarily restricted.

If all source reduction options identified in this Project were
implemented, benzene and total hydrocarbon emissions would be
reduced by about 25 percent and 16 percent,  respectively.   The
Workgroup concluded that a cost-effective strategy for the
Refinery would have to include a mix of source reduction,
recycling, treatment and disposal options.

Of the source reduction options considered,  most appear to be
significantly lower cost than recycling, treatment, and disposal.
Source reduction options considered have had an average cost of
$650/ton of pollutant recovered.  The remaining seven options
analyzed had an average cost of $3,200/ton,  nearly 5 times
higher.  The cost-effectiveness of individual options varied form
a low of $190/ton for secondary seals on gasoline storage tanks
to a high of $128,000/ton for the treatment plant upgrade.

D.  While release reductions do not always pay for themselves,
    some environmental improvements can be made at a net cost
    savings to the Refinery.

The Refinery is relatively efficient in managing materials.  An
ongoing weight-loss management program to capture lost material
has been in place at all Amoco refineries for a number of years.
Approximately 99.7 percent of the incoming crude is converted to
useful products and refinery fuel.  The hydrocarbon release
reduction options identified in this Project dealt with the
remaining 0.3 percent.

Despite the relative efficiency of the Refinery, two source
reduction options—seals on gasoline tanks and & leak detection
and repair program—have net cost savings and a positive rate of
return.  Amoco did not know this before this Project.  On the
other hand, some of the source reduction options and all
treatment options were not economic investments for the Refinery.
For example, fitting all fixed roof storage tanks with secondary
seals would result in much higher cost for relatively little
additional reduction in hydrocarbon emissions compared to fitting
only gasoline storage tanks.  Treatment options generally require
significant capital outlays with no return in the form of
recaptured or improved product.  Technology options with positive
rates of return are shown in Figure 3.9..  Options that have
negative return are not shown.

    1.3.3  Choosing Alternatives

A.  Ranking the options showed that better environmental results
    can be obtained more cost-effectively.
                               1-10

-------
Compliance with current and anticipated regulations requires
controls for eight sources types, reducing airborne hydrocarbon
releases by 7,300 tons/year at an average cost of $2,400/ton.
The Refinery could reduce about 7,100 tons of airborne
hydrocarbons each year (or about 97 percent)  by controlling six
sources at about 25 percent of the cost.  This cost-effectiveness
comparison does not account for possible benefits to other media.

If allowed to address both hydrocarbons and listed hazardous
waste, the Refinery could reduce about 7,500 tons per year at an
average cost of about $500/ton using its choice of sources and
techniques.  Table 1.3 provides a more detailed comparison of
different Release Management Strategies, results and costs.

These results are all the more significant because the options
evaluated were neither selected nor developed ahead of time with
a target reduction goal in mind. Nor did the selection process
have a goal of meeting regulatory requirements in some
alternative fashion.  This suggests that even more impressive
results might be achieved, if that were the focal point at the
beginning.

B.  All participants agreed on which options were the most
    effective and which were least/ regardless of their ranking
    criteria or institutional viewpoints.

The Project used a multi-dimensional prioritizing process  (the
Analytical Hierarchy Process, AHP) in which weights were
developed for all criteria used to rank alternatives.  These
criteria included cost, release reduction, timeliness and changes
in benzene exposure, among others.  The process allowed the
Workgroup to assess the significance of and interactions between
criteria—how changes in one criterion affect other criteria and
total rankings.

All options were considered legally acceptable, and no specific
regulatory requirements were imposed on the decision making
process.  Although different organizations brought different
perspectives to the discussions, each organization reached the
same conclusions about which options would be most effective and
which were least.  The driving forces in this prioritization were
cost and relative risk reduction, as measured by benzene
exposure.  A variety of sensitivity studies confirmed this
initial set of preferences.

Amoco ranked control of marine loading losses as the most
effective—though not the lowest cost—option.  A second tier of
options included installing secondary seals on tanks, instituting
a leak detection and repair program, and upgrading blowdown
stacks.  All four were also viewed as reasonably effective
pollution prevention projects.  In total these four projects
would prevent or capture almost 6,900 tons of releases annually

                               l-ll

-------
at a cost of about $510/ton.  EPA and Virginia selected the same
five options, in this hypothetical case with no specific
regulatory requirements.  See Items 4 and 5 in Table 1.3.

    1.3.4   Obstacles and Incentives to Implementing Pollution
            Prevention

After identifying several alternative environmental management
options, it is reasonable to ask why these options are not being
implemented.  What can be done to encourage their use?  The
following discussion summarizes the general findings based on an
assessment of potential obstacles and incentives for implementing
five highly ranked options.  For more details, see Section 5.0.

A.  EPA does not have an explicit policy goal and may not have
    the statutory authority to simply set a release reduction
    "target" without prescribing how this target should or could
    be met.  When the target involves releases in multiple media,
    current administrative procedures discourage a coordinated
    approach/ including evaluating risks/ costs and benefits of
    managing residual pollutants in different media.

Requirements under many statutes and regulations prescribe how
release reductions should be achieved, sometimes in terms of
which technology should be used, often in terms of which specific
sources should be controlled.  For example, the Benzene Waste
Operations NESHAP focuses on a specific emissions source to a
single medium—benzene emissions from wastewater.  The rule
requires control of benzene emissions from this single source.

Data from this refinery indicated that wastewater is a small
contributor to total benzene releases.  Amoco and EPA disagree
about some of the specific measurements and results.  These are
discussed }'.- ..citail in Air Quality Data, Volume II  (Amoco/EPA
1992b).

A number of pollution prevention approaches developed in this
Project are more effective in controlling benzene emissions, and
less costly to implement than the benzene NESHAP.  Other
refineries might find other sources that present more cost-
effective control opportunities.  Focusing on individual sources,
rather than on desired overall "performance," limits the ability
to achieve the most cost-effective control.

RCRA requires application of the Best Demonstrated Available
Technology  (BOAT) to a hazardous waste before it can be disposed.
BOAT standards are typically based on a destruction technology
rather than on methods at the higher end of the pollution
prevention hierarchy.

One proposal now before Congress  (S. 1081) to reauthorize the
Clean Water Act would amend 304(b) of the Act and require EPA to

                               1-12

-------
promulgate effluent guidelines which reflect the application of
best available control technology (BAT) for all categories of
pollutants.  This Congressional proposal, which does not reflect
the Administration's position, could limit the Agency's ability
to set environmental protection priorities.

B.  Legislative and regulatory programs do not provide
    implementation schedules compatible with design, engineering,
    and construction timeframes.

Most regulatory and statutory programs require compliance within
six months to at most three years after promulgation of a final
rule.  In some cases, compliance requirements do not consider
normal maintenance schedules and economic penalties associated
with facility-wide shutdowns.  Consequently, short-term "fixes"
which can meet legal deadlines, are used at the expense of more
cost- and environmentally effective, long-term solutions.

A typical refinery project for processing oil using established
technology and design procedures, normally takes 2-3 years from
initial design to startup, assuming there is agreement on what to
build, no unusual equipment delivery problems, no additional
safety considerations, and no prolonged startup difficulties.
Many projects take longer when regulatory applicability, scope or
design criteria are unclear, or new technologies are involved.

For example, the benzene NESHAP rule discussed above was
promulgated in March 1990 (under the 1977 Clean Air Act
Amendments).  Statutory language required compliance with the
regulations within two years.  In this case, significant
differences in interpretation between EPA and the regulated
community took more than one year to resolve and to clarify the
regulatory requirements.  An acceptable understanding is a
prerequisite to engineering and construction.  It was physically
impossible to design, engineer, procure, construct, and start up
the required control within the remaining one year compliance
time frame.

C.  Congress, EPA and much of industry have become used to
    command-and-control, end-of-pipe treatment approaches based
    on twenty years of experience.  These well established
    problem solving approaches are difficult to change.

In the 1970's, environmental regulations successfully helped
reduce point source emissions to air and water.  End of pipe
treatment was successful partly because many industrial firms and
permitting authorities had little experience dealing with these
problems, and found the specification of technical solutions
offered a "road-map" for how to proceed along an uncharted
course.  These requirements also provided a relatively "level
playing field" for US industry.  Many of today's problems are


                               1-13

-------
sufficiently different than those of the early 1970's that they
can benefit from alternative approaches.

D.  The short time taken by the Virginia Air Pollution Control
    Board to issue or modify air permits is not a deterrent to
    installing technologies to reduce airborne emission at this
    site.

Most of the technical options would reduce air releases at the
Refinery.  However, obtaining permits to install most of these
technologies would probably not be a problem since the Virginia
Air Pollution Control Board is estimated to take about six months
to issue a permit  (Virginia is a delegated state for issuing air
permits).

However, information generated through a facility-wide multi-
media assessment is a necessary first step to not only developing
a strategy to reduce these releases, but also to exploring such
implementation options as integrated permits.

E.  Inadequate accounting for both the benefits and costs of
    environmental regulations is an obstacle to developing a more
    efficient environmental management system.  Responsibility
    for pollutant generation and accountability for environmental
    protection are difficult to quantify.

At many industrial plants, such as Amoco's, waste management
costs are frequently charged to a central environmental
management division rather than to the operating unit that
generates the waste.  Remediation costs for clean-up of
contaminated soil, for example, are frequently charged against
another cost center, rather than to the generator of the
contamination.  This separation between release generation and
costs is a disincentive to manage releases more effectively.

Few EPA accounting systems measure direct benefits of the
Agency's activities, such as improved ecological health,
biodiversity, reduced risk to human populations, etc.  Rather,
accomplishments are usually measured in terms of activities such
as permits written, amount of fines collected, or number of
enforcement actions pursued.   (GAO, 1991)  The lack of direct
connection between Agency activities and environmental results
reduces accountability for program costs and benefits.  Without
adequate measurement systems, it is difficult to tell when
environmental management practices actually improve the
environment.

        1.3.5.  Education/Communications/WorXing Relationships

This Project enhanced knowledge of both government and industry,
and generated information that EPA and Amoco can use.
                               1-14

-------
The study provided an opportunity to educate individuals within
EPA and Amoco.  Based on plant visits and information exchanges,
EPA personnel better understand how a refinery works, the
complexities of the refining processes, and the difficulties in
obtaining reliable environmental release data.  This improved
understanding will be useful as the Agency considers future data
needs for regulatory development and permits.

Similarly, Amoco personnel better understand how EPA develops
regulations, the type of information needed, and the Agency's
operating constraints. This will be useful for Amoco in
interacting with EPA and other government agencies.

The detailed release information developed in this Project could
be useful to all three media offices:  air, water, and solid
waste.

•   The Office of Air and Radiation may be able to use air
    monitoring and modeling information for developing MACT
    standards and improving emission factors.

•   The Office of Solid Waste should be able to use sampling and
    monitoring information for characterizing RCRA Subtitle D
    wastes and management practices.

»   The Office of Water should be able to use wastewater sampling
    information to evaluate Petroleum Refining effluent
    guidelines, and the biomarkers research results in evaluating
    aquatic health measurement tools.

The working relationships between various EPA offices, State and
Amoco personnel were quite fragile when the Project began.
Individuals brought their institutional viewpoints to initial
discussions.  By agreeing at the beginning of the Project that we
may not necessarily agree with all findings and conclusions,
people showed a willingness to discuss issues and focus on data
and factual information.  Many of the perceived and real
differences in views were more easily dealt with in a factual
setting.

1.4  Recommendations
                              %
     1.4.1     Explore Opportunities to Produce Better
               Environmental Results More Cost-effectively.

Data from this study show that the Refinery can meet a release
reduction goal more cost-effectively than by meeting reductions
prescribed by current regulatory or legislative requirements.

For example, the ranking analysis shows that given the
opportunity the Refinery could remove about 97 percent of tons of
airborne hydrocarbons at about 25 percent of the cost of reducing

                               1-15

-------
them under current and anticipated regulations.   The cost-
effectiveness of the flexible option is about $6QO/ton compared
with the cost-effectiveness of $2,400/ton for regulatory
requirements.

EPA might evaluate options for setting a goal or target for
reducing multi-media releases from a facility, and then allow the
facility to develop an alternative compliance strategy to meet
the goal.  This alternative strategy would allow the facility to
meet the goal at a lower cost, include interim milestones, and be
enforceable.  This strategy would also make appropriate
information available to ensure that the reduction targets will
be met.

This strategy might also include commitments to other
environmental improvements such as cogeneration, additional
reductions in releases, wetlands restoration, wildlife habitat
enhancement, creation of new wetlands, controls on nonpoint
sources of pollution, improved environmental data collection and
research.  The cost savings realized from meeting requirements
under a more flexible approach make it possible to realize
additional environmental benefits which are presently foregone
because of the high costs of many regulatory programs.

     1.4.2     Improve Environmental Release Data Collection,
               Analysis and Management.

Data from this study show that an emissions inventory could be
improved by measuring releases and developing new emission
factors.  For example, the emissions inventory at the beginning
of the project did not account for all potential releases to the
environment.  Some releases were excluded because the Agency has
excluded them from reporting  (e.g., barge loading operations);
some releases were not included because the sources and the
amount were thought by Amoco to be insignificant (e.g., blowdown
stacks); some emissions were overestimated (e.g., API Separator);
and some releases were underestimated  (e.g., coker pond).
Jointly established sampling and analysis protocols could help
improve data quality, so that reported values more accurately
portray facility releases.

Data currently collected in response to regulatory or permitting
requirements could be evaluated to determine how its utility and
quality might be improved.  For example, TRI data quality and
utility could be improved by:

•    Providing more inclusive estimates of facility-wide releases
     to all media.  The Project found the exclusion of marine
     loading operations from TRI reporting requirements conveyed
     an inaccurate picture of total facility releases.
                               1-16

-------
•    Reporting groups of chemicals, rather than individual
     species, especially if these chemicals have similar
     structural, physical and toxicological properties.
     Requiring reporting of all VOCs for refineries,  rather than
     specific compounds like xylene (and its individual isomers),
     would provide a meaningful measure of refinery releases.
     That is because xylene poses approximately the same risks
     and has physical characteristics similar to the hundred of
     undifferentiated VOC compounds not covered in TRI.  For a
     refinery, where a complex mixture of chemicals are released
     from most sources, tracking many separate chemicals does not
     make good use of technical, laboratory, and environmental
     management resources.

•    Reporting other selected chemicals of concern for
     demonstrated human health or ecological impact separately.
     At a refinery, chemicals such as butadiene, benzene, and
     nickel may be good indicators of risk/release potential and
     management practices.  Other industrial sectors would need
     to track different specific chemicals.

•    Improving emission factors for estimating releases based
     upon information developed in thit project, and additional
     work by EPA/industry task groups that could focus on the
     different data collection needs of discrete industry
     sectors.

The Project had great difficulty collecting and verifying
environmental release data from the site.  Emissions from these
sources are complex and measurement techniques are rudimentary.
Many emission measurements varied with time.  For example, the
Coker pond emissions varied by a factor of three within a few
hours.  Better sampling and analysis methods and statistical
tools are needed to analyze variability.  Research is also needed
to develop methods that can verify release inventories within
reasonable confidence limits, accounting for specific differences
in emissions factors.
     1.4.3     Provide Incentives for Conducting Facility-vide
               Assessments, and Developing multi-media Release
               Reduction Strategies.  Such Strategies Should
               Consider Multi-Media Consecruences of Environmental
               Management Decisions.

This Project demonstrates that more cost-effective environmental
protection programs can be designed by allowing companies to
consider site specific factors and focus on results.

A detailed facility-wide, multi-media assessment identified the
most significant medium  (air) and releases sources, both in terms
of quantity and impact on the surrounding area.  Specific

                               1-17

-------
technology options were then developed to deal with these
sources.  The significance of sources identified in this Project
were not initially known or apparent to the participants.
Proposed solutions could not have been developed in the absence
of data which identified their importance.

For example, hydrocarbon emissions from barge loading operations
(784 tons annually) and blowdown stacks (5,200 tons annually)  are
significant.  However, the Refinery did not know this prior to
this Project, nor did the existing regulations require the
collection of this data.  Thus, it did not develop control
options to reduce these emissions.

Several technologies considered for reducing releases, transfer
pollutants from one medium to another or convert pollutants to
different forms.  Since human health and environmental
consequences vary from one medium to another, viewing a release
problem in the context of net environmental effects is essential
to developing more sound solutions.

The current institutional framework and procedures for developing
regulations do not include multi-media assessments and analysis.
Current practices should be reviewed to determine how they could
be modified to use information from such assessments.  An
integrated pollution prevention and management strategy would
facilitate development of release management options that produce
better environmental results.  (EPA/SAB, 1990a; EPA/SAB, 1990b;
OMB, 1991)

At present, industry has little incentive to conduct such
assessments because it does not have an opportunity to implement
their findings.


     1.4.4     Encourage Additional Public/Private Partnerships
               on Environmental Management.

The Yorktown experience demonstrates the opportunities and
pitfalls that can occur when government and industry work
together.  The opportunities are significant.  The pitfalls are
worth overcoming.  All organizations—EPA, Virginia and Amoco—
sought to develop and test innovative environmental management
approaches that, unlike most traditional "command and control"
approaches, consider risk reduction, address multi-media
concerns, maximize environmental benefits, encourage efficient
use of resources, and promote facility-specific implementation
choices.  While it will take time and patience to overcome
decades of distrust, such joint government/industry efforts can
result in more cost-effective environmental protection by
providing the opportunity to share different viewpoints and
skills.
                               1-18

-------
In this study, for example, EPA brought expertise on the type of
information needed to develop regulations,  and their operating
constraints, while Amoco brought an understanding of refinery
operations and economics.  By helping to educate each other and
develop a mutual understanding of issues and technology, Amoco,
EPA and the Commonwealth of Virginia together agreed on the most
significant emissions from the Refinery and the most promising
approaches to reducing them.

Public/Private partnerships could also be used to leverage Agency
resources for providing improved data needed to develop
regulations.  This Project illustrates a possible approach to
collecting data, assessing technologies and characterizing a
facility within an industry that took less time and Agency
resources but relied more on private support.

     1.4.5     Conduct Research on the Potential Health and
               Ecological Effects of VOCs.

The Refinery is a major source of the area's VOC emissions.
However, information on the potential adverse health effects of
VOC emissions is rather limited (Graham, 1991) .  Research is
needed to better characterize health and ecological effects of
VOCs that can be used in conducting risk assessments.  This study
could also build on efforts currently underway at the American
Petroleum Institute, and the Chemical Industry Institute of
Toxicology (CUT) and others.

EPA should also undertake research to develop indicators that
measure impacts on the ecosystem of multi-media releases from
industrial facilities.  This Project looked at several biomarkers
that show promise as indicators in aguatic environments.  Limited
information and methods' for assessing ecological risk limits the
ability to conduct comprehensive risk assessments, and measure
changes in environmental quality.
                               1-19

-------
        Table 1.1

   Project Components
Biomarkers
Chemical Fate and Transport
Communications
Cost Estimation
Decision Making Methodology
Engineering
Environmental Impact
Exposure Modeling
Facilities Management
Group Dynamics
Meteorology
Public Perceptions
Regulatory/Legislative Policy
Risk Assessment
Sampling
Source Identification
           1-20

-------
                            Table 1.2

    Participants  in the AMOCO/EPA Pollution Prevention Project

U.S. Environmental Protection Agency

Office of Air and Radiation
Office of Solid Waste and Emergency Response
National Advisory Council on Environmental Policy
  and Technology
Office of Research and Development
Office of Policy, Planning and Evaluation
Office of Water
Office of Pesticides and Toxic Substances
Office of Air Quality Planning and Standards
Region III

Amoco Corporation

Environmental Affairs and Safety
Public and Government Affairs
Art Services
Analytical Services
Groundwater Management Services

Amoco Oil Company

Refining and Transportation Engineering
Research and Development
Yorktown Refinery
Whiting Refinery

Commonwealth of Virginia

State Water Control Board
Department of Waste Management
Department of Air Pollution Control

Academic Institutions

Virginia Institute of Marine Science, College of William and Mary
University of California at Los Angeles
University of Michigan
                               1-21

-------
                      Table 1.2 (Continued)

Consultants

ICF/Clement International
Research+able
ENSECO Laboratories
Radian Corporation
Linnhoff-March
York Laboratories
Murry/Trettel Consulting Meteorologists
Industrial Marine Service/ Inc.
James R. Reed and Associates
Industrial Economics, Incorporated
Abt Associates
Resources for the Future

Peer Review Committee Members

Dr. Clifford S. Russell, Vanderbilt Institute for Public Policy
Studies (Chair)
Ms. Jolene Chinchilli, Chesapeake Bay Foundation
Mr. A. Ray Dudley, Array Enterprises, Inc.
Dr. John R. Ehrenfeld, MIT Center for Technology, Policy and
Industrial Development
Dr. John D. Graham, Harvard School of Public Health
Dr. Robert J. Huggett, Virginia Institute of Marine Science
Ms. Frances H. Irwin, Conservation Foundation
Dr. Joseph F. Malina, Jr., University of Texas
Dr. John J. McKetta, University of Texas
Mr. David R. Patrick, Clement International Corporation
Dr. James G. Quinn, University of Rhode Island
Dr. Mitchell J. Small, Carnegie Mellon University

Amoco/USEPA Workgroup Members

John Atcheson                        Mark Klan
David Berg                           Howard Klee
Doug Blewitt                         Donna Kraisinger
Walter Brodtman                      Jim  Lounsbury
Kirt Cox                             Keith Mason
Catherine Crane                      Richard Olin
Jim Cummings-Saxton                  Pat  Pesacreta
Christine E. DeLuca                  Mahesh Podar
Dan Fort                             Alex Ross
Deborah Gillette                     Manik Roy
Madeline Grulich                     Marv Rubin
Deborah Hanlon                       Dale Ruhter
Janice Johnson                       Debora Sparks
Mark Joyce                           Mary Spearman
Sharon Keneally-Baxter               Pat  Woodson
                              1-22

-------
                                                                               Table 1.3

                                          Cooper!son of Different Environmental Managencnt Option* for the Yorktotn Refinery
Selection Criteria for Release Reduction Projects
1. Current and Expected Regulatory Requirement*
(Table 4.5) (Note 3)
2. Cost-Effective Release Reduction (Table 4.6)
3. Cost- Effective Benzene Exposure Reduction
(Table 4.7)
4. Multiple Criteria (Table 4.4) (Note 4)
4a. Work Croup (Top 4)
4b. Amoco (Top 4)
4c. EPA/Vlrglnla (Top 4)
5. Most Favored--All Rankings, All E valuators
(Table 4.8)
Ho. of
Project*
8
6
6
4
4
4
4
Project ft
(Note 5)
4. 5c. 7A, 7B.
7C 8, 9. 11b
lib. 5c. 7A.
6. 8. 9
5c. 9. lib.
7A, 4. 1
9. 5c. 7A. 11c
9. 5c. 7A. lie
9, 5e, 7A, 11c
9, 5a or 5b,
7A, 11b or 11c
Material
Released
(Note 1)
VOC/HC
VOC/HC
Listed
HU
VOC
VOC
VOC
VOC
VOC
Total
Release
Reduction
Tons/Yr.
7.300
7.500
7,100
6.900
Capital
Cost
S)M
53.6
10.7
13.2
Four optl
effect!
10.2 .
Annual
Cost
SW
17.5
3.8
4.2
ons Mere cons
ve in dlfferer
3.5
Benzene
Exposure
Reduction, X
99
87
90
stently selected
it ranking exercl
87
Average
Cost
$/Ton
2,400
510
590
•s dost
ses.
;sio
Motes;
1.   VOC
     HC
          • Volatile Organic Compounds
          • Liquid Hydrocarbons
Listed HU • Solid, Hazardous Waste
2.   Values are rounded.  See tables 4.1 through 4.7 for
     details.

3.   Regulatory and Statutory Programs considered Include
     Benzene NESHAP, Ozone non-attainment, likely Clean
     Air Act requirements under MACT and HON rules.

4.   Multiple criteria Included release reduction
     potential, benzene exposure reduction potential,
     cost, Impact on liability, transfer-ability to other
     facilities, status In pollution prevention hierarchy.
     etc.  C«e Section 4.0 for discussion.

5.   See Table 3.2 for Project descriptions.
                                                                                 1-23

-------
2.0  REFINERY RELEASE INVENTORY

Amoco's Yorktown Refinery is a 35-year old,  53,000 barrel per day
facility that manufactures gasoline, heating oil,  LPG,  sulfur,
and coke.  It is located on the York River in Virginia,  near the
Chesapeake Bay.  The Refinery was used for this study primarily
because of it's proximity to Washington, D.C.  During the course
of this program, over one-hundred EPA and Virginia regulatory
personnel were able to visit the facility for a first-hand view
of refinery operations and practices.  Appendix A provides a site
description, facility details, and background information on
emissions, permits, and administrative procedures in the
Commonwealth of Virginia.  Figure 2.1 provides a schematic flow
chart of the Refinery and potential releases to different
environmental media.

Pollution prevention concepts generally start with source
reduction as a key management option.  But source reduction and
the benefits therefrom, can only occur once specific sources are
known.   The environmental sampling program helped identify
specific release sources in the Refinery that could later be
targeted for potential process or operational changes to reduce
releases and their environmental impact.

Information on all release sources was not available at the
beginning of this Project.  Complex industrial sources,  like the
Refinery, contain hundreds, sometimes thousands, of potential
release points.  It is technically difficult and impractical to
monitor and measure each of these points.

Most refinery measurements have focused on compliance with
end-of-pipe pollution control requirements such as their NPDES
water discharge permit or RCRA groundwater monitoring plan.
Permit monitoring requirements demand certain information for
wastes that may leave the Refinery.  Monitoring resources are
typically allocated to meet permit requirements rather than at
the point of generation. As a result, a comprehensive multi-media
release inventory based on individual process unit contributions
was not available.

2.1 Sampling Program Results

To overcome deficiencies in existing data, a multi-media sample
collection and analysis approach was used.  This program differed
from many existing regulatory programs which dictate the kinds of
compounds or mixtures to be reported, as well as their units of
measure—"oil and grease" in the NPDES water permit program, for
example.  Rather, each medium was sampled for selected chemicals
such as benzene, toluene, ethylbenzene, and xylene (BTEX), as
well as for particular chemical species expected to be present in
specific media such a metals, and polynuclear aromatics.  One
goal was to identify specific chemicals present in all media,

                               2-1

-------
both within the Refinery and entering the environment beyond the
Refinery property line.

As in any experimental measurement program,  uncertainties remain
in the values measured.  Environmental emission measurements are
difficult to make and most have no well defined,  accepted
scientific protocols.  In many cases, measurements made by one
technique were cross-checked against measurements or estimates
made using a different technique.  Individual reference documents
discuss the uncertainties associated with sampling in each
medium.  It is worth noting that some measurements represent the
first attempt to determine emissions from some specific sources.
Appendix B provides more details about the sampling program and
its limitations.

About 1,000 separate samples were collected during this program.
They provide the first major database showing all releases from a
single facility into all environmental media at one point in
time.  Figure 2.2 shows the sample distribution by media,
excluding duplicates and field blanks required for QA/QC
purposes.

The probable accuracy of most measurements is + 100 tons/year.
However, to avoid losing information in roundoff, a higher level
of accuracy of ± 1 ton/year is used throughout the report.  Major
conclusions from the sampling program include:

Distribution of Releases Within the Refinery;  The Refinery
generates about 27,504 tons/year of materials that reach all
environmental media—air, surface water, groundwater and land.
Figure 2.3 summarizes the generation of pollutants, prior to any
internal recycling, transfer or disposal.  Airborne emissions
account for 48 percent of the total; solid waste 29 percent; and
surface water 14 percent.  In addition, biosolids from wastewater
treatment account for 9 percent of solid wastes.   However, a
combination of natural conditions and pollution management
systems designed for materials recovery and compliance with
existing regulations and permits, substantially shifts the
quantity and distribution of the pollutants between media.  This
is discussed below.

Internal Refinery Management;  Figure 2.4 shows how the releases
described above are managed in the Refinery.  About 44 percent
(12,124 tons/year) of the material generated is handled on-site
through treatment, recovery and recycle, and does not leave the
facility.

    Air.  An additional 378 tons/year of hydrocarbons evaporate
from Refinery drainage and wastewater treatment systems, mixing
with other airborne hydrocarbons.  A total of 13,609 tons/year of
airborne material leaves the Refinery.


                               2-2

-------
    Water.  Wastewater flows to the API Separator where an
average of 2,690 tons/year of oil (about 50 barrels per day)  is
recovered and recycled to the Refinery for processing.   A small
amount of groundwater is also recovered by the drainage system
and mixed with process wastewater.  The activated sludge
wastewater treatment system generates 2,420 tons/year of
biosolids which are ultimately recycled on-site with other solid
and hazardous wastes.  Treated effluent discharged to the York
River contains about 46 tons/year of suspended solids and other
material.  This discharge is normally about 10% of the amount
permitted under the Refinery's NPDES permit, although under
infrequent upset conditions, the discharge may approach permit
values.

    Solid and Hazardous Wastes.  Spent caustic (3,788 tons/year)
is sent off-site for recovery of remaining caustic value and
naphthenic acids.  Most catalysts are recycled for recovery of
additional activity or metals.  Spent cracking catalyst is sent
to Amoco's Whiting, Indiana Refinery (613 tons/year) for use as
equilibrium catalyst.  Spent Ultrafonning catalyst is returned to
metals reclaimers to recover platinum for reuse in new catalyst.
Spent desulfurization catalyst and polymer catalyst are non-
hazardous and are buried in an on-site landfill.  Sludges
collected from the API Separator are a listed hazardous waste
under RCRA regulations.  They are combined with other solid
wastes such as biosolids from the wastewater treatment plant and
recycled to the Refinery's coker  (4,398 tons/year).  In the
coker, hydrocarbons are converted into saleable products, while
water is recovered for treatment.

Most remaining solid waste is construction debris and
contaminated soils which are landfilled on- and off-site.  A
total of 1,725 tons/year of solid waste were landfilled in 1990.
Potential off-site landfills are screened for management
practices, compliance records, and other factors by Amoco
personnel prior to approval for use by any Amoco facilities.

Distribution of Releases Leaving the Refinery;  Nearly 89 percent
(13,609 tons) of the releases to the environment from the
Refinery are airborne.  Figure 2.5 shows total releases to all
environmental media leaving the Refinery.  The high percentage of
airborne emissions leaving the facility focused both the sampling
program and subsequent identification of the projects on this
medium.  As noted below, groundwater contamination under the
facility is small, and is not moving off-property.  Consequently,
no contaminated groundwater is reaching the external environment
from the Refinery.

Types of Airborne Emissions;  Figure 2.6 shows the division of
airborne emissions between criteria pollutants—SO2, NOX, CO and
particulates  (PM-10)—and VOC hydrocarbons.  Criteria pollutants
result primarily from combustion or other stacks.  Nearly 60

                               2-3

-------
percent of air emissions are hydrocarbons (VOCs).   The sampling
program focused on VOCs and their sources because  much less was
known about these sources.

Employee Exposure;  Within the Refinery fenceline, maximum
observed and/or calculated airborne concentrations of TRI
reported chemicals were below OSHA action and permissible
exposure levels for 8-hour time weighted average exposures.

Air Quality Beyond the Fenceline;  Impacts on air  quality by the
Refinery were calculated using air dispersion modeling techniques
and the emissions inventory developed during this  study.   Table
2.1 compares calculated concentrations of benzene, toluene and
ethylbenzene at several locations near the Refinery with reported
values for typical urban, rural and remote settings.  The area
has lower hydrocarbon concentrations from the Refinery than
typically found in urban air.  As discussed in Appendix A,
automobiles and an adjacent powerplant contribute  some of these
chemicals to the air.  Biogenic (natural) sources  also
contribute.  In the entire middle Atlantic region  natural sources
provide about 40 percent of airborne hydrocarbons, with a higher
percentage in more rural areas like Yorktown (Placet and Streets,
1989) .

Diversity of Emissions;  Crude oil contains thousands of
individual hydrocarbon species.  Emissions from oil refining
operations reflect this diversity.  Despite an extended sampling
and analysis program, all species emitted could not be
identified.  Selected samples were analyzed for 150 organic
compounds.  The analysis identified about 80-90 percent of the
compounds present.  The remaining 10-20 percent which were
unidentified are most probably structural isomers  of typical
volatile organic hydrocarbons found in petroleum product
mixtures.  A number of small quantity emissions of unidentified
hydrocarbons were found in most of the airborne samples.   These
compounds would be expected to exhibit similar physical and
toxicological properties to compounds identified.

Comparison with TRI Emissions;  The Refinery's 1989 TRI report
showed 371 tons of reportable chemicals released from all sources
to all media.  Based on measurements and modeling conducted for
this project, releases of TRI chemicals were 893 tons, about 2.4
times higher than reported.  This difference reflects: (a) the
identification of blowdown stacks as a significant source  (430
tons) whose contribution was previously unknown and unrecognized,
(b) the addition of marine loading losses (165 tons) which are
not reportable for TRI (Oge, 1988), and  (c)  lower emissions from
the API Separator (-86 tons).  Emissions from the  inactive
landfarm, a coker pond and sewer vents were also identified as
new sources.  Figure 2.7 illustrates these changes.  Table 2.2
provides a reconciliation of the reported TRI values and
measurements made for twelve reported chemicals.  Comparing the

                               2-4

-------
reported values and measurements on the same basis (excluding
marine loading losses and blowdovn stack emissions)  shows
measured values of 298 tons/year, versus the 371 tons/year
reported—about 20% less than reported.

Emissions Excluded from TRI;  The sampling program helped
identify and quantify other emissions excluded from TRI reporting
requirements.  Some of these are excluded because they are below
the threshold amounts that trigger reporting.  Some are excluded
because certain operations, such as barge loading, are not
considered reportable under some circumstances by EPA (Oge,
1988) .

Some chemicals are excluded because they are not listed, although
they have substantially similar physical and toxicological
properties to chemicals that are listed.  The isomers of
trimethylbenzene illustrates this last point:  1,2,4-
triraethylbenzene is a reportable chemical.  The 1,3,5- and 1,2,3-
trimethylbenzenes are not.  All three occur in crude oil and
gasoline and in emissions from a refinery.  All three isomers
have similar physical and toxicological properties.

Because of the wide diversity of emissions coming from crude oil,
the TRI report only covers about 11 percent of the total
hydrocarbon emissions leaving the facility.  The unreported
emissions belong to the general class of volatile organic carbon
compounds (VOCs) associated with petroleum products and
processing.  As noted above, some new sources were also
identified during the study.  Total Refinery airborne emissions
of all non-methane hydrocarbons—both TRI reportable and
non-reportable—are about 7,905 tons per year.

Cross-media Transport;  Cross-media transport—the natural
transfer of a pollutant from its medium of release into a
different medium—is not significant for chemicals which are only
slightly soluble in water  (Allen, Cohen, and Kaplan, 1989).
Studies performed by UCLA's National Center for Intermedia
Transport showed that most hydrocarbons released into the air do
not transfer rapidly into other media.  Therefore, examination of
air quality effects ignoring inter-media transfer is a reasonable
analytical approach.  Water soluble compounds, such as methanol
and MTBE, can transfer from air into water and soil media under
certain conditions  (Cohen, Allen, et. al. 1991).

As discussed elsewhere, a number of hydrocarbons initially
released into various water streams partially volatilize into the
air.  Several pollution control technologies—some used currently
and others proposed for future regulations—transfer pollutants
from one medium to another.  For example, the wastewater
treatment plant converts waterborne hydrocarbons into 2420
tons/year of biomass.  At the Refinery, this sludge is recycled


                               2-5

-------
to the coker where it is converted into products.  However,  not
all refineries have cokers.

Comparison vith AP-42 Emission Factors: Emission factors
established by EPA are most frequently used to estimate total
airborne emissions from different types of refinery equipment.
Most AP-42 Factors do not provide information about the
composition of these emissions, except by general class—e.g.,
VOCs, particulates, etc.  The measurement program allowed a
direct comparison between several measured or inferred emission
rates and emissions calculated using these factors.

For fugitive and tank vent emissions, the AP-42 emission factor
calculations showed good agreement with measurements.  Measured
emissions from the coker pond were about 40 percent greater than
estimated using AP-42 factors.  (Estimated emissions were
determined using the AP-42 emissions factor for an oil-water
separator.  Both the Coker pond and API Separator perform this
function.  There is only a single emission factor for different
types of uncovered separators).

Measured emissions from the API oil/water Separator were 2,100
percent lower than estimated with AP-42 factors.  A combination
of reasons can probably explain this discrepancy:  (1) the limited
database for this emission factor (no measurements), (2) improved
refinery operating practices since the original data was
collected in 1959, and  (3) improved measurement techniques during
the last thirty years (API, 1991a; API, 1990).  The EPA Office of
Air Quality Planning and Standards (OAQPS) has given this
emission factor a D rating on a scale of A (good) to E  (poor, no
data) (EPA, 1988 EPA, 1985a).  Overall measured emissions from
these four sources were about 60 percent of the amounts
estimated.  Figure 2.8 compares measured and estimated emissions
for selected sources.

Groundwater;  Subsurface contamination, detected during the
sampling period, was significantly less than that observed at
other petroleum refining facilities  (Los Angeles Times, 1988).
Contamination found appeared limited to shallow soils and/or
groundwater.  These unusually low levels of contamination are the
result of (a) natural soil conditions, (b) no significant spills,
(c) a higher than usual percentage of above grade piping, and  (d)
the underground process sewer system acting as a continuous
groundwater recovery system.  About 35,000 gallons per day of
groundwater are collected in the underground sewer system and
sent to the wastewater treatment plant.  This last point was an
unexpected finding of the study (Amoco/EPA, 1991e).

Solid Waste Management;  The Refinery generated over 10,500 tons
of solid waste and spent caustic in 1990.  However, more than 80
percent of the solid waste is recycled or treated either on- or
off-site and does not enter the environment. Remaining materials

                               2-6

-------
are disposed of in approved landfill sites.   See Figure 2.11 for
more details (Amoco/EPA, 1991d).

Solid Waste Composition:  Most of the solid wastes result from
activities associated with the Refinery process water collection
and treatment system.  Nearly 1,000 tons/year of soils enter the
drainage system where they become oil-coated sludge.

Surface Water;   The existing water treatment plant is very
effective in removing contaminants from process waters prior to
discharge.  Except MTBE, most contaminants were not detectable in
the treated effluent (Amoco/EPA,  1991c).  Samples and
biomonitoring results showed some localized sediment
contamination in the Refinery's stromwater pond.  Water
discharged from the pond is routinely monitored and continues to
meet NPDES permit reguirements.

Public Perceptions;  Quality of life issues are important to area
residents who value the historical and rural aspects of their
environment.  Most concerns focus on the  community
infrastructure stresses caused by commercial and residential
development, traffic, and limited sanitary sewer system capacity.

The Refinery has a relatively low profile within the community:
awareness is usually limited to residents of the immediate area
and triggered by odors from the facility.  The Refinery is not
viewed as a major problem, nor is there a reservoir of good will.
There are a number of military facilities in the area.
Government activities are viewed with some suspicion.  Relative
to other environmental problems,  air quality was most often cited
as the environmental medium of concern to residents. Watermen
were more concerned about water quality and preserving their
commercial fishing practice.  They attributed all pollution to
sources other than themselves  (Amoco/EPA, 1992c).

Ecological Impacts;  This Project also undertook the development
and testing of several biomarkers through research with the
Virginia Institute of Marine Science.  The goal of this work was
to develop tools which might be useful indicators of changes in
Refinery releases to the River.  This work is still in progress,
and will be reported in detail as part of the comprehensive
Project documentation  (Amoco/EPA, 1992d),

2.2 Source Identification

Air Emission Sources; Most releases from the Refinery into the
environment are into the air.  A combination of direct source
measurements, ambient air monitoring, modeling, and emission
factor calculations improved the emissions inventory and air
dispersion models.  Emissions from sewer vents, water ponds, the
inactive landfarm, and API  (oil/water) Separator were measured
directly  (Amoco/EPA 1991b).  Mass balances and inlet/outlet water

                               2-7

-------
analyses were used to determine emissions from the cooling tower
and wastewater tanks.  In general,  mass balance techniques are
not sufficiently accurate for most inventory calculations
(NRG, 1990).

Ambient monitoring both upwind and downwind of the Refinery was
used to infer information about emissions from such fugitive
sources as  leaks from process valves, flanges, pump seals, and
tank vents.  Emissions from barge and truck/rail loading
operations  were calculated using standard AP-42 emission factors
(EPA, 1988) and actual Refinery loadings for 1990. Emissions from
blowdown stacks were calculated using the AP-42 emission factor
which gives a total emission rate based on refinery throughput.
Composition measurements made at Yorktown were used to define
chemicals in the total flow.  A number of separate attempts to
measure emissions directly are discussed elsewhere (Amoco/EPA,
1992a).  Table 2.3 summarizes the different techniques used to
define the  airborne emissions.  The result of applying this
combination of tools is the composite source chart shown in
Figure 2.9.  In this chart, specific emission sources are
identified  and quantified.  This source information provides a
starting point for identifying and prioritizing pollution
prevention  opportunities.

The chart reveals a number of useful results about airborne
emissions from the facility.  First, the three process blowdown
stacks are  the largest source of airborne hydrocarbon emissions.
At the beginning of this study, these were thought to be minor
sources.  Barge loading losses represent a second major source of
emissions.  Fugitive losses from process equipment and from tank
vents are the third and fourth largest sources, respectively.
The coker blowdown pond is also a significant emissions source.
While this  was recognized prior to this study, its size and
significance in relation to other sources was unknown.  Emissions
from the coker pond do not contain many chemicals in the TRI, and
therefore not all emissions would be reported as part of the
annual TRI  filing.  The Refinery sewer system (sewer vents and
API Separator) is a relatively small contributor to emissions.

A single measurement is the basis for landfarm emission estimate.
Therefore,  this should be used with caution.

Solid Waste Sources;  Figure 2.10 summarizes solid waste sources
identified  from the sampling program.  In 1990 the Refinery
generated about 10,500 tons of solid and hazardous waste
(including  spent caustic).  Sludges accumulated in the drainage
and water treatment system account for a majority of these
solids.   Contaminated soils, tank sediments and spent catalysts
account for the remaining solids.  Spent caustic, although an
aqueous solution, is usually classified and handled as hazardous
waste for waste management purposes.  Figures 2.11 and 2.12
summarize how these wastes are managed.

                               2-8

-------
Many of these wastes are a natural byproduct of refining
operations.  At Yorktown/ crude oil, for example, contains more
than 1,000 tons of sediment from the producing formations, which
ultimately deposit in the refinery's API Separator or storage
tanks.  Local soil contributions (about 1,000 tons/year),
however, represent a large potential opportunity to reduce solid
waste generation.

Subsurface Sources;  A combination of well samples and computer
modeling helped identify potential sources to and sinks for the
subsurface aquifer.  The most significant sources of water
reaching the subsurface were natural recharge from rainfall, and
the water from the coke fines settling basin.  More important
from an environmental impact standpoint, the Refinery's
underground drainage system apparently is recovering groundwater
and adding it to the wastewater treatment plant.  Consequently,
there appears to be no movement of contaminated groundwater off
site.

Surface Water Sources;  Identification of specific surface water
sources was complicated by an old, underground drainage system
that was not designed for access and sample collection.
Nevertheless,  samples from major arteries helped identify the
sources for primary pollutants of concern such as oil & grease,
BOD5, ammonia, total suspended solids, sulfides, metals, BTEX,
and phenol.  Figure 2.13 shows the process sources.  The crude
unit desalter and crude tank water draws have the highest
pollutant loadings.
                               2-9

-------
                                               Table  2.1
                  Comparison Of Annual Average Predicted Impacts To Typical Measured
                Concentrations For Different  Types Of Environments  For BTEX Chemicals
CHEMICAL

BENZENE
TOLUENE
ETHYLBENZENE
MAXIMUM PREDICTED
CONCENTRATION
AT THE FENCELINE1
(ug/M3)
1.3
5.4
1.6
MAXIMUM
PREDICTED
CONCENTRATION
AT THE CLOSEST
RESIDENCE
(ug/M3)
0.6
2.4
0.7
NOTE 2
TYPICAL REMOTE
CONCENTRATION
(ug/M3)
0.51
0.19
0.06
NOTE 2
TYPICAL RURAL
CONCENTRATION
(ug/M3)
1.5
1.3
0.7
NOTE 2
TYPICAL URBAN
CONCENTRATION
(ug/M3)
5.7
7.7-12.0
2.7
1Maximuia on  land.

2J.  J.  Shah  and E.  K.  Heyerdahl,  1988.
                                                 2-10

-------
                                                   Table 2.2

       Reconciliation of 1989  Toxic  Release  Inventory Report  and Pollution Prevention Inventory
                                    Yorktown Refinery (Units  are Tons/Yr)
Chemical
enzene
oluene
thylbenzene
ylenes
yclohexane
aphthalene
rimethylbenzene
thylene
ropylene
utadiene
ethanol
TBE
TOTAL
1989 TRI Report
41.0
91.5
24.5
107.0
3.8
2.8
45.5
8.5
30.5
0.036
3.3
12.1
371
B/D Stack
Additions
32.4
56.6
45.7
121.6
26.5

42.2
40.6
64.1
0.18


430
Cokar Additions
1.8
3.1
1.0
6.3


1.1





13
Barge Loading
Additions
15.0
52.8
14.0
69.2

5.1



'1*

8.8
165
Hastewater
Subtractions
-0.1
-23.2
-7.3
-33.2
0.3

-22.2





-86
Total
90.1
180.7
77.9
270.9
30.6
7.9
66.6
49.1
94.6
0.2
3.3
20.9
893
es; . •
TRI as  % of total VOC'8 - 11%
VOC'a as % of crude run * 0.3%
TRI Column excludes 0.175 tons chlorine
TRI Reports only 1,2,4-trimethylbenzene.
  Other columns report 1,2,4- as well  as 1,3,5- and 1,2,3-isomers.
Hastewater Subtractions includes the net effect of reduced emissions  from the API Separator
  and increased emissions from sewer vents.
The- totals are rounded.
                                                      2-11

-------
                                             Table  2.3
                           Technique Used to Determine Airborne Emissions
  Source Type
  Basis of Emission Estimate
                Speciated Emissions
'I SEPARATOR
,RGE LOADING

.OWDOWN STACKS
>KER POND
•OLING TOWER
ACTIVE LANDFARM
•ADING RACK

OCESS FUGITIVE

WER VENTS
NKS
Direct Measurement
AP-42

AP-42
Direct Measurement
Water Sampling, Mass Balance
Direct Measurement
AP-42

AP-42 Default Components

Direct Measurement
AP-42
Direct Measurement
Proportional to Product Compositions Loaded in
  1990
AP-42 and Direct Measurement of Composition
Direct Measurement
Direct Measurement
Direct Measurement
Proportional to Product Compositions Loaded in
   1990
Proportional to Compositions Measured for Similar
  in-stream Equipment at Another Refinery
Direct Measurement
Proportional to Product Compositions Loaded in
  1990
                                                2-12

-------
              Fig. 2-1
              Yorktown refinery
              Simplified  Flow Diagram
              with Release Sources
                                                  Crudt oil
V^—^-V
 Water
 to sewer
                                                                                                   Fugitive HC
                       Spent
                       catalyst
                       T
                    To recovery,
                    recycle, and
                    landfill
\iiJu
   till  !i«i||
     Ultratorming


        Water to sewtr
1
            Vent losses

  £  ^  Treating and blending
             tanks
                           Furnace oil
                                                                                            Slowdown
                                                                                            stacks
                                                                                To recovery,
                                                                                recycle *nd ItndtJII
  k Refinery
1    fuel gas
                                                      ' Vipor recovery
                                                           Polymerization—MTBE
                                                    T   T
                                                      ©  ©
                                                    Gasoline
                                                                                                              Treated wat
                                                                                                              to fork Kvt
                                                                                             To recycle and landfill
                                                                         Sulfur
                                                   Marin*
                                                   loading
                                                                                                       Solid
                                                                                                       Airborne emission
                                                                                                       ground water
             Petroleum coko
                                 Liquefied
                                 petroleum gas
                                 (LPGI

-------
                 Figure 2.2
Pollution Prevention Sampling Program
             (Number of Samples)
     Air 630
   Groundwater143
                                  Solid Waste 37
                                 Soils 39
                            Water110

-------
                             Figure 2.3
              Pollutant Generation  Within
                  the Yorktown Refinery
                   (Prior to Recycling, Transfer and Disposal)
     Airborne Criteria
         5704
   NOx,SO2, CO, PM-10
   Airborne Hydrocarbon " '
          7527
                             Haz. and Solid Waste
                                   8104
                            Catalysts, Sludges, Spent
                                  Caustic
                              Biosolids from Tr'mt
                                   2420
Units are Ton/year
               Waterborne Material
                    3749
            Oil, Susp. Solids, Inorganics

Total Generation =  27,504 Tons/year

-------
                                             Figure 2.4
                                             Pollutant  Transfers/Recycle/Treatment
                                             within  the Yorktown Refinery
                                             (Units are tons per year)
        Pollutant Generation
        at Refinery:  27,504
        (See Figure 2.3)
Air
»• • •

Airborne criteria: 5704
NOx, SOj, CO. PM-10
Airborne hydrocarbons:
7527
Water

Stormwater
                                    13.231
Waterborne material: 3749
Oil, susp. solids, inorganics
Ground water
Solid Waste

Hazardous and
solid waste:
8104

Catalysts, sludges,
spent caustic
                                                            ,,  .     .          .
                                                            Hydrocarbon evaporation
                                                            from water 378
                                           ^
                                                 Recovered oil:
                                                 2690
                                                             Biosolids
                                                             from
                                                             treatment:
                                                             (wet tons)
                                         Caustic to
                                         off-site recycle:
                                         37RR
                                                              Catalysts to
                                                              off-site recovery:
                                                                                                             Pollutant Releases
                                                                                                             from Refinery: 15,380
                                                                                                             (See Figure 2.5)
                                                                                                   Airborne criteria:
                                                                                                   5704
                                                                                                                      51%
                                                                                                                      •* ' /0
                                                                                                   Airborne
                                                                                                   hydrocarbons:
                                                                                                   7905
                                                                                         Settling basin !••
                                                                                                                   Stormwater to York River
                                                                                                                                to York River 46

                                                                                                                                   0.3%
                                                                                         Sludges and oils to
                                                                                         on-site recycle:


                                                                                             3r
                                                                                                                 rjn-site
                                                                                                                 landfill:


                                                                                                               *
                                                                                                                               end disposal: 1725
                                                                                                             Off-site
                                                                                                             disposal:

-------
                       Figure 2.5
           Total Releases Entering the
      Environment from Yorktown Refinery
                        Watertorne
                                   Djsposa|

                                   1725
                Airborne Criteria
                   5704
                   37.1%
                           Airborne Hydrocarbon
               Air
              13,609
              88.5%
Units are Ton/year
Total Releases = 15380 tons/year

-------
                          Figure 2.6
     Total Air Emissions - Yorktown  Refinery
                  SO2
                  3802
         NOx



        PM-10
         560
Hydrocarbons
                                                    7905
Units are Ton/year             Total = 13,609

Note: The Adjacent Virginia Power Plant Releases About 36,466 Tons / Year
    See Appendix A for More Details

-------
                   Figure 2.7
    1989 TRI  Inventory Compared to
           Measured Emissions
    Tons/year
                                       New
                                       Total
                                       893
                                       11,000

               New
              Source
              + 443
Reported
  371
                       Not
                     Reportable
                      + 165     Better
                            Measurement
                               -86
-200
               Stack  Marine  Wst Water  Measured
             & Other

-------
                     Figure 2.8
  Selected Airborne Hydrocarbon Sources
       Estimated vs. Measured Values
API Sep.
 1278
Fugitive
 796
             Tanks
              633
        Estimated
       (AP-42 Factors)
       Total = 2871
                     Coker
                      164
                             Fugitive
                              796
                   Coker
                   261
                Measured
             (Direct/Indirect Meas.)
                Total = 1751
 Units are tons/year

-------
                    Figure 2.9
             Yorktown Refinery
        VOC Air Emission Sources
 B/D Stacks
   5200
             Fugitive 796

                  Barge Loading 784

                     Land Farm 53

                      API Sep. 61
                                         Tanks 633



                                        Sewers 117
                                       Coker 261
Units are tons/year
Total = 7905

-------
      Figure 2-10.Major Sources
                       of Refinery Solids
14,000

12,000

10,000

 8,000

 6,000

 4,000

 2,000

    0
      T/Yr (Wet)
                              I^PI API Separator Sludge
                                  WWTP Sludges
                              HH Tank Sediment (H)
                              lH Tank Sediment (NH)
                              I   I Spent Catalyst
                              mi Contaminated Soils
                                  Spent Caustic
                                  Other
        1988
       1989
1990   CRUDE
 Note:
1,000 tons of sediment in crude oil is recovered as solid waste.

-------
Fig. 2-n Solids Management
 14,000
 12,000
      T/Yr (Wet)
                                      Landfill (Offsite)
                                      Landfill (Onsite)
                                      Landfarm
                                      Recycle (Offsite)
                                      Recycle (Onsite)
         1988
1989
1990

-------
Fig. 2-12  Solids  Management
             by  Type  (1990)
  5,000
  4,000
  3,000
  2,000
  1,000
    0
     T/Yr (Wet)
       Recycle
        Onsite
Recycle
 Offsite
Landfill
 Offsite
                  •1 API Sludge (Haz)
                  im WWTP Sludges
                  I  I Tank Sediment (Haz)
                  II Spent Catalyst
                     Contaminated Soils
                  l-;:---:';!;:l Spent Caustic
                     Other
                  Bi Other (Haz)

-------
                        Figure 2.13
    Phenol, Ammonia, Oil & Grease Sources
100
   % Contribution
 0
                                    <——I  Cooling Tower
                                    !•  Gas Treater/Tanks
                                    GHH  Ultraformer
                                        Combination Unit
                                        Coker
     Phenols
Ammonia
O&G

-------
3 . 0  OPTIONS IDENTIFICATION AND ANALYSES

After assembling the Refinery release inventory,  the Project
identified potential process and operating changes that might
impact these releases.

3.1  Identification

In March 1991 more than 120 representatives from EPA, Amoco, the
Commonwealth of Virginia, academic, environmental and consulting
organizations met for a 3-day brainstorming Workshop in
Williamsburg, Virginia.  The Workshop participants developed a
variety of potential release reduction options that might be used
at the facility, considered ranking criteria, permitting issues,
and obstacles and incentives for implementation.   Workshop
sessions included both a structured review of process synthesis
techniques and a more free-wheeling idea-generation/discussion
session.  Participants proposed over 50 concepts for further
consideration covering energy conservation (affecting criteria
pollutant releases), volatile hydrocarbon controls, solid waste,
groundwater, and surface water streams.  Table 3.1 lists all the
projects identified.

To meet Project schedule and budget constraints,  the Workgroup
later selected twelve projects for more detailed analysis.  The
twelve chosen were felt to:  (1) be feasible with current
technology,  (2) offer significant potential for release
reductions,  (3) have manageable (or no) impact on worker safety
concern, (4) be amenable to more quantitative analysis in the
time available, and (5) address concerns in different
environmental media.  Table 3.2 provides a brief description of
each project.  Figure 3.1 shows where each option fits into the
overall refinery flow.

3.2  Option Characteristics

Preliminary material balances and engineering designs were used
to analyze each potential option.  Some of this work was
completed specifically for this Project.  Other portions were
completed as part of environmental engineering work at Amoco for
the Refinery.  For each option the following items were
determined:

Capital Cost;  Cost estimates with a ±25 percent accuracy were
made for these scoping studies.  Additional engineering effort
would be required to prepare an estimate with a + 10 percent
accuracy typically needed for management approval.

Operating and Maintenance Cost:  Costs were estimated as a
percentage, of total capital cost and project complexity.
Depending upon the project, this cost varied between 3 and 6
                               3-1

-------
3 . 0  OPTIONS IDENTIFICATION AND ANALYSES

After assembling the Refinery release inventory,  the Project
identified potential process and operating changes that might
impact these releases.

3.1  Identification

In March 1991 more than 120 representatives from EPA, Amoco, the
Commonwealth of Virginia, academic, environmental and consulting
organizations met for a 3-day brainstorming Workshop in
Williamsburg, Virginia.  The Workshop participants developed a
variety of potential release reduction options that might be used
at the facility, considered ranking criteria, permitting issues,
and obstacles and incentives for implementation.  Workshop
sessions included both a structured review of process synthesis
techniques and a more free-wheeling idea-generation/discussion
session.  Participants proposed over 50 concepts for further
consideration covering energy conservation (affecting criteria
pollutant releases), volatile hydrocarbon controls, solid waste,
groundwater, and surface water streams.  Table 3.1 lists all the
projects identified.

To meet Project schedule and budget constraints, the Workgroup
later selected twelve projects for more detailed analysis.  The
twelve chosen were felt to:  (1) be feasible with current
technology,  (2) offer significant potential for release
reductions,  (3) have manageable (or no) impact on worker safety
concern, (4) be amenable to more quantitative analysis in the
time available, and  (5) address concerns in different
environmental media.  Table 3.2 provides a brief description of
each project.  Figure 3.1 shows where each option fits into the
overall refinery flow.

3.2  Option Characteristics

Preliminary material balances and engineering designs were used
to analyze each potential option.  Some of this work was
completed specifically for this Project.  Other portions were
completed as part of environmental engineering work at Amoco for
the Refinery.  For each option the following items were
determined:

Capital Cost;  Cost  estimates with a +25 percent accuracy were
made for these scoping studies.  Additional engineering effort
would be required to prepare an estimate with a + 10 percent
accuracy typically needed for management approval.

Operating and Maintenance Cost;  Costs were estimated as a
percentage, of total  capital cost and project complexity.
Depending upon the project, this cost varied between 3 and 6


                               3-1

-------
percent of total capital.  It also includes depreciation,  taxes,
insurance and other indirect costs.

Liability Cost Rating;  Each project was evaluated qualitatively
for its potential to affect future remediation,  catastrophic and
product quality liability concerns.

Timeliness;  The number of years needed to complete each project
was estimated, subject to current equipment maintenance schedules
and operating limitations.

Transferability;  Qualitative assessment of the ability to use
the project technology within other refineries and other
industries was made.

Revenues;  Revenues were estimated for those projects where
saleable materials were recovered.  The quantity of recovered
material was equivalent to the emissions reduction.  All
recovered hydrocarbons were valued as gasoline at $0.75/gallon,
with an average density of 6.5 Ibs/gallon.

Annualized Cost;  These costs were estimated as the sum of
annualized capital costs and all variable expenses.  Future costs
were discounted at 10 percent (or 15 percent) to determine their
present value, assuming a project life of 15 years.

Pollution Prevention Mode;  One or more of the pollution
prevention modes in the pollution prevention hierarchy was
assigned based on review, discussion, and consensus among
Workgroup members.  These classifications were not obvious in
several cases and required extended debate.

Net Release Reduction;  Estimates of emissions reduction
(tons/year) vary in accuracy.  Additional emissions sampling and
more detailed engineering analysis would be needed to improve
these estimates.  Where possible, generation and transfer of
releases in other media were included in estimating the "net"
change in release.

Recovery Cost:  For liquid hydrocarbons or VOC emissions the
equivalent annual cost was divided by the net release reduction
volume to determine an average $/gallon for each option.  This
number is equivalent to the price which would have to be charged
per gallon of recovered material to recover capital, operating,
maintenance and distribution costs.

Cost-Effectiveness;  The equivalent annualized cost was divided
by the net release reduction to determine a $/ton cost-
effectiveness for all options.
                               3-2

-------
Net Present Values;  Present value of all cash flow, including
initial capital, operating expenses, taxes, depreciation,
indirect costs, revenues, etc.

Resource Utilization;  Qualitative estimates were developed for
each option's effect on raw materials and utilities requirements.

Effects on Secondary Emissions;  The impact of each project on
other emissions were judged qualitatively.  For example,
increased power requirements would normally increase emissions
from utility systems.

3.3  Analyses and Impacts

Ranking and prioritizing these options required specific,
quantitative (and sometimes qualitative data) about each choice.

    3.3.1  Release Reductions

Table 3.3 provides estimated reduction in releases  (tons/year),
type of material released, control technology efficiency,
expected time required to complete installation of the particular
control technology, and type of release management option used
for control (source reduction, recycling, etc).  The efficiency
values shown for options 5a-5e are average values for the entire
set of tanks included in each option.  They do not show the
efficiency of a particular type of seal on an individual tank.
Details of specific control efficiencies are provided in Air
Quality Data, Volume II  (Amoco/EPA, 1992b).

    3.3.2  Financial Analysis

EPA and Amoco use different methods to estimate capital costs,
rates of return, economic benefits, and cost-effectiveness.
Major differences are:

    capital cost estimates;  Amoco has developed specific
    cost/capacity correlations for commonly used equipment based
    on past experience.  Installation costs are based on industry
    or company construction standards, and site-specific labor
    rates.  EPA typically uses more general cost data available
    in the literature or provided by individual companies.

    income tax;  Amcjo's rate-of-return evaluations include
    current Federal and State income taxes, which total about 36
    percent of before-tax income.  EPA excludes income tax from
    its calculation of cost-effectiveness.

    depreciation;  EPA does not consider depreciation when
    estimating compliance costs.  Amoco includes depreciation in
    tax and cash-flow calculations.  Depreciation rates are set
    by Internal Revenue Service rules for  specific kinds of

                               3-3

-------
    equipment.  At present, the Capital Cost Recovery System
    (CCRS) is used to calculate annual depreciation over a 10-
    year period.

    interest rate;  EPA typically uses a 10 percent discount rate
    in determining the net present value of a cash flow stream.
    Amoco normally calculates the interest rate (internal rate of
    return) at which the present value of benefits is equal to
    the present value of the costs.  This may be higher or lower
    than 10 percent, depending upon individual circumstances.

Amoco's estimates for capital, operating, and maintenance costs
were used in this report.  A ten percent discount rate was used
in calculating present values, with the exception of Table 3.4a,
where a 15 percent discount rate was used.  Cost comparisons
using EPA and Amoco's approaches are discussed in more detail in
Projects, Evaluation and Ranking (Amoco/EPA, 1992f).

Table 3.4a summarizes the quantitative financial analysis
described in Section 3.2.  For each option, the Table shows the
Present Value (PV) of the capital cost, the PV of
operating/maintenance cost, the annualized cost and cost-
effectiveness in dollars/ton.  PV is a useful tool for comparing
different options that have different cash flows and different
time periods, since all cash flows, present and future, are
expressed in current dollars.

All dollar values, including the annualized cost and cost-
effectiveness, are in 1991 dollars.  Other financial information
is on PV basis that discounts future cash flows using a 15
percent discount rate.  Table 3.4b presents this same
information, except that cash flows are discounted at 10 percent.
The higher discount rate is more typical of that used at Amoco
for project evaluation purposes.  EPA typically uses the lower,
10 percent, rate to evaluate cost-effectiveness.

Table 3.4c provides a different perspective on the information
contained in Table 3.4b.  Table 3.4c shows the net present value
of all cash flows including initial capital, operating and
maintenance expenses, taxes, depreciation, indirect costs and
revenues from product recovery.  The net cost-effectiveness
column represents the net cost in $/ton for each option.  The
values shown in brackets for projects 11 and 5a indicate that a
net profit is generated for each ton recovered.  The final column
of Table 3.4c (and Figure 3.9) presents the rate of return for
each option.  Values are presented only for options that have a
positive rate of return.

    3.3.3  Off-site Impacts

The influence of different management options on the exposure of
surrounding human populations to releases from the Refinery was

                               3-4

-------
assessed using computer modeling tools.  For those options that
involved changes in releases of solid waste or surface water
discharges, adequate tools were not available.  However,  an
independent risk screening showed that neither the surface water,
nor solid waste presented a significant human exposure pathway
(see Section 3.4).  No additional modeling was done for options
that might affect groundwater, since groundwater is not a source
of drinking water near the site and is not migrating off-site.

    3.3.4  Dispersion

As noted several times in this report, the most significant
emissions were airborne.  For those options that involved a
change in emissions affecting air quality, impacts were modeled
using standard air dispersion techniques, similar to those used
to confirm the emissions inventory.  Exposure estimates were
developed for three classes of chemicals: benzene, toluene,
ethylbenzene and xylene (BTEX), other chemicals reported in the
Refinery's TRI submissions, and criteria pollutants (S02, N02,
PM-10, and CO) .  Similar modeling techniques were used for all
three classes.  A resource document detailing air modeling
methods and results discusses these techniques in detail
(Amoco/EPA, 1992b).  The discussion here focuses on the impact of
benzene emissions, since this turned out to be the chemical
species of greatest concern, relative to other releases.

For benzene (and other BTEX compounds) the emissions inventory
described in Section 2.0 was used as input to the ISCST air
dispersion model.  The model used one year of hourly
meteorological data collected at the National Weather Service
station at Norfolk, Virginia.  Norfolk is located about 50 miles
from Yorktown, and experiences a similar land-sea breeze
condition.

A mathematical receptor grid was established around the refinery
containing 859 points.  Both a fine grid (250 meter resolution)
and a coarse grid (1,000 meter resolution) were used, to give
better resolution near the Refinery where concentrations changed
more quickly.  The model calculated ground level concentrations
at each receptor point for each hour of the year as well as
annual average concentrations.  Thus, about 7.5 million computer
calculations were completed to model impacts of each chemical
studied for each different control scenario considered.

    3.3.5  Results

Current Refinery Operations

Modeling results are summarized in several ways.  First, Figure
3.2 presents isopleths of annual average benzene concentrations
overlaid on a USGS topographic map of the Refinery area.  Each
                               3-5
                                                                         s

-------
curved line represents a line of constant benzene concentration,
much as elevation lines are used on topographic maps.

Second, Table 3.5 provides maximum predicted ground-level
concentrations in micrograms per cubic meter (ug/M3) for  various
receptor locations:  within the refinery (9.2), beyond the
fenceline on the York River (6.2),  beyond the fenceline on land
(1.3), and at a nearby residence (0.6).

Third, additional dispersion modeling to identify source-receptor
relationships and define culpable sources for the receptor points
identified above.  Figure 3.3 illustrates these results.   For
example, at the point of highest concentration in the  York River,
93 percent of the impacts result from barge loading emissions.
Whereas at the nearby residence, barge loading emissions—are
responsible for 53 percent of the total calculated concentration.
Storage tank emissions and blowdown stacks account for most of
the remaining concentration.  Barge loading accounts for a
significant fraction of the total impacts at both receptor sites.

To put these values in some perspective, Table 2.1 compares
calculated concentrations with ambient air quality data obtained
from past EPA studies (Shah, 1988).  For benzene, refinery
impacts at the fenceline are similar to those observed in a rural
environment.  At the nearby residence, benzene concentrations are
similar to those observed in a remote pristine setting.
Ethylbenzene impacts were similar to benzene.  Toluene impacts
were somewhat higher, falling between typical rural and urban air
quality.  No comparable information was available for xylene.

Pollution Prevention Options

Changes in emissions resulting from simulating the different
pollution prevention options in Table 3.2 were also modelled,
using identical air dispersion techniques.  Again, benzene
emissions are discussed here, because of their potential health
impacts.

For each option considered, revised benzene emissions for the
affected source(s) were used to calculate a new emissions
inventory.  For example, recovering barge loading losses reduces
benzene emissions by 11 tons per year.  Figure 3.4 shows a
modified histogram reflecting this change.  However, the
histogram only shows the impact of reduced benzene emissions on
the total.  It is more helpful to ask how the reduced emissions
effect exposure of people outside the fenceline.  These changes
become more apparent when plotted as revised isopleths on the
same USGS map used for Figure 3.2.  Figure 3.5 shows the new
isopleths  (shown as dotted or red lines) on the original
inventory map.  Close examination shows that high benzene
concentrations around the barge loading area have disappeared.
                               3-6

-------
Furthermore, the most outlying isopleth (showing a concentration
of 0.12 ug/M3)  has  moved in towards the Refinery center.   This
indicates that the area impacted by Refinery emissions has been
reduced.

Ultimately the new concentration information can be converted
into population exposure and risk.  Each pollution prevention
option can be viewed and compared in this same way, leading to
calculation of changes in relative population risk for each
option, compared to current operations at the Refinery.  This
will be done as part of the relative risk assessment described in
section 3.4.   As an interim measure of risk, changes in benzene
exposure were calculated at a nearby residence for each control
option.  Table 3.6 shows these results.  The baseline benzene
concentration was 0.62 ug/M3.   The different pollution prevention
options reduce this concentration to between 0.61 ug/M3 and 0.28
ug/M3.

By looking at this drop in potential benzene exposure as a
surrogate for risk reduction, a more quantitative measure of the
effectiveness of each option can be developed.  The third column
in Table 3.6 shows this exposure reduction.  Here the existing
facility contributes 100 percent of the controllable benzene
exposure.  As benzene concentration at the nearby residence
decreases, exposure also drops.  From the values in this Table,
reducing barge loading emissions has the largest potential to
reduce benzene exposure.  Many of the other options had small or
minimal impact on benzene concentrations at a residence, and
therefore, small or minimal potential to reduce relative risk.

Figure 3.6 shows the same kind of information for the impacts of
adding secondary seals to gasoline storage tanks.  Again, the
most outlying isopleth has moved inwards.  Figure 3.7 shows the
new isopleths that result from implementing barge loading
controls, adding secondary seals to gasoline storage tanks and
upgrading the blowdown stacks to flares.  The new isopleths have
moved to nearly within the Refinery property.  Qualitatively they
show a significant change over a fairly wide area, indicating a
potentially effective control strategy.

Figure 3.8 shows a similar plot, but this time the changes
reflect the impacts of control options for the API Separator and
underground drainage system.  In contrast to Figure 3.7, the
curves show almost no impact on air quality has occurred in
residential areas.  This combination has a small effect over a
narrow area, and therefore appears to be a relatively ineffective
control strategy.

These results will be fine tuned by a more rigorous risk analysis
that considers total population exposures, rather than focusing
on a single residence.  In the interim, the benzene exposures


                               3-7

-------
were used as another piece of information in prioritizing the
different pollution prevention options.  Section 4.0 discusses
how this was done.

3.4  Risk Assessment

Risk assessment provides a systematic method for estimating risks
to exposed individuals or populations resulting from a particular
hazard—in this case the potential exposure to chemical
contaminants in the environment.  This component of the Project
will help evaluate the relative risk reductions associated with
alternative pollution prevention options.  Significant
uncertainty is often associated with this kind of analysis and
therefore, the results should not be interpreted as definitive.

Risk assessment typically begins with a characterization of the
risks associated with baseline or current releases.  The baseline
assessment gives an indication of the potential for human health
or ecological risk problems.  The predicted changes in emissions
and sources are then estimated and the expected risk from the
option scenarios is evaluated.  The risk evaluation is then based
on both risk reduction to the most highly exposed individuals and
to the exposed population as a whole.  One of the benefits from
pollution prevention is derived by estimating the reduction in
risk achieved by each technical option.

This risk assessment follows current EPA methods and established
Agency policy as outlined by thfe National Academy of Sciences
(NAS, 1983) and established as final Risk Assessment Guidelines
(EPA, 1986); EPA, 1992).  It involves four steps:  hazard
identification, determination of dose-response relations,
evaluation of human exposure and finally characterization of
risks.

This risk assessment did not attempt to analyze other
environmental effects associated with Refinery releases such as
their contribution to formation of ozone, acid rain, risks
associated with occupational exposure, transportation of products
or wastes, or the potential for accidental releases.  Rather,
this risk assessment considered primarily the changes in risk
associated with human exposure to airborne benzene vapors.

        3.4.1  Screening Analysis

Screening analysis was conducted for different exposure pathways
and chemicals of concern.  Using a screening level cut-off of a
one-in-a-million risk for a 70-year lifetime exposure for the
maximally exposed individual  (MEI), a set of carcinogenic
chemicals were identified for further analysis.  For
noncarcinogens, MEI exposure levels were compared to health
thresholds.  If the MEI exposure exceeded the established health
threshold, further analysis was done.  Because the risk

                               3-8

-------
assessment is not yet complete, the following discussion focuses
on the results of a screening analysis.   A final risk assessment
will be provided as a part of the support documentation for this
project (see Appendix C).   Preliminary results and rankings will
be reexamined after the risk assessment is complete and may
change.

Air:  Of the chemicals screened, nickel,  vanadium, methanol,
carbon tetrachloride, xylenes, toluene,  ethylene,  propylene,
naphthalene, ethylbenzene, polynuclear aromatic hydrocarbons
(PAHs),  1,2,4-trimethylbenzene and cyclohexane showed
insignificant risks and were not evaluated further.

For MTBE,  EPA has not formally established health effects and
reference dose.  A preliminary estimate of the threshold
reference concentration was developed for this risk assessment.
Estimated concentrations for the MEI location were 10-15 percent
of the reference concentration.  No further evaluation was done
for MTBE.

Two other chemicals and one mixture failed this initial screen:
VOCs, 1,3-butadiene, and benzene.  VOCs were present at about 0.2
ppm outside the Refinery boundary.  VOCs are a complex mixture of
hydrocarbons with an unspecified  (and variable) composition.  The
American Conference of Government and Industrial Hygienists
(ACGIH)  has set an exposure standard (TLV) of 300 ppm for
workplace exposure to gasoline vapors,  another hydrocarbon
mixture of unspecified composition (ACGIH, 1990).   ACGIH
threshold limit values should not be used to determine exposure
limits for the general population, but simply provide a benchmark
for this discussion.

No animal health studies have been completed with VOCs expected
from Refinery emissions.  One animal study using wholly vaporized
gasoline found statistically significant neoplasia in male rat
kidneys and female mouse livers from lifetime exposures above 292
ppm  (McFarland, 1984).  The rat kidney effects were subsequently
associated with a protein found only in this species, and are now
considered irrelevant to human health risks (Loehr, 1991).
Further studies are continuing at the Chemical Industry Institute
of Toxicology  (CUT) to determine the relationship of mouse liver
neoplasia to human health risks.

There is considerable debate, much uncertainty, and little data
on human health effects at the low VOC concentrations expected
around the Refinery.   Recent epidemiology studies of refinery
workers, who would be expected to receive a higher exposure to
VOCs than the general population, show a lower incidence of total
cancer cases than the general population  (Wong and Raabe, 1989).
However, these studies are not able to differentiate between
other potential'confounding factors such as the "healthy worker


                               3-9

-------
effect," cigarette smoking, diet, etc.  VOC impacts from the
Refinery were not evaluated further.

Benzene is a known carcinogen.  It poses a one-in-a-million risk
at a concentration of 0.12 ug/M3  for a lifetime  exposure.   Air
modeling results show the concentration of benzene at the
refinery boundary is 2.0 ug/M3.   At  the nearest  residence,  this
concentration is 1.5 ug/M3.   1,3-Butadiene is  also a  known
carcinogen.  It poses a one-in-a-million risk at 0.0036 ug/M3 for
a 70-year lifetime exposure.  Air modeling for this chemical
shows a concentration of 0.0057 ug/M3 outside  the  boundary  and
0.0050 ug/M3 at a residence.   Risks  from benzene and  butadiene
will be discussed in the final risk assessment document in
preparation at this time.

Surface Water;  The Refinery has a State water discharge permit
covering two outfalls:  a stormwater settling basin outfall (002)
and a combined treated process water mixed with once-through,
non-contact cooling water outfall (001).  The surface water
analysis used the results from samples of these streams tested
for 22 contaminants, total organic carbon, and such physical
properties as pH and temperature.

Surface water discharge concentrations were generally below the
analytic detection limits (Amoco/EPA, 1991c).   A screening level
analysis identified the potential for risks to either human
health or aquatic life based on established federal water quality
criteria.  This comparison was based on the EPA recommended
approach for determining reasonable excursions above water
quality criteria (EPA, 1991) .

The screening analysis using two data points for these stream
shows that copper exceeds EPA criteria for aquatic toxicity.
Specifically, the highest copper concentration measured in
outfall 002 was 250 ug/1—about 90 times the marine acute copper
criteria of 2.9 ug/1.  The highest copper concentration in
outfall 001 exceeded the marine acute copper criteria by 76
times.

Table 3.7 provides additional data regarding copper concentration
in the Refinery outfalls, collected as part of the Refinery's
Toxics Monitoring Program.  Over the past two years,  the copper
concentration in both outfalls have averaged between
66 and 76 ug/1.  Criteria are not the same as standards.
However, if Federal criteria are adopted as water quality
standards in Virginia, these concentrations would constitute an
exceedence.  It is interesting to note that water from the York
River has also exceeded these criteria on occasion, as shown by
VEPCO data in Table 3.7.  The source, bioavailability, impacts
and possible remedies for this copper 'level warrant further
study.


                               3-10

-------
Fish liver samples collected inside the Refinery's storm water
settling basin showed higher than background incidences of
lesions.  Concentrations of hydrocarbons and some metals in
sediments from the settling basin were also higher than expected,
probably resulting from the carryover of coke particles from the
coke yard through the surface drainage ditch system.   The
Refinery's i^crm water discharge continues to meet the NPDES
permit requirements.  The Refinery is evaluating this information
with the assistance of the Virginia Institute of Marine Science:.

Drinking Water;  Because the nearest drinking water source is
seven miles from the Refinery and the York River is too saline
for consumption, drinking water was not considered a potential
source of exposure.

Groundwater;  The groundwater contamination was fou,,d to be
minimal with no off-site migration (Amoco/EPA, 1991e).
Therefore, groundwater appeared to pose little or no risk and was
not analyzed further.

The sampling program did not find evidence of groundwater
contamination from on-site waste disposal.  Impacts from off-site
solid waste disposal on groundwater contamination were not
assessed.  However, Amoco has rigorous screening procedures for
selecting off-site disposal sites, as it is responsible for
future liability associated with such disposal.

    3.4.2  Uncertainty

Uncertainty in the risk assessment process arises from multiple
sources:  the use of animal study results, difficulties with
human studies, variations in individual responses to chemical
exposures, the impact of differing does rates, multiple
simultaneous exposure to chemicals, and the use of extrapolation
methods to estimate risks from high-exposure populations to low-
exposure populations.  These will be discussed in more detail in
the Risk Assessment Report.

3.5 Public Perceptions

The Workgroup wanted to include information about York County's
perception of the impact the Refinery has on the surrounding
environment and its role in local environmental concerns.  They
believed this information might be useful in characterizing the
viability of potential pollution prevention options.  This
information would also be useful in developing a strategy to
communicate study results within the community.  This information
gathering was not considered a means to have a dialogue with
citizens or public officials about environmental management;
simply as a way to assess current opinions.  Three distinct
activities were undertaken to collect this information—thought
leader interviews, two focus group meetings, and a telephone

                               3-11

-------
survey of 200 households.  Based on these three sources of
information, several conclusions can be drawn:

The thought leader and focus group interviews concluded that land
development is the major quality of life concern in the Yorktown
area.  People cited traffic, problems with sewer,  water and
general development as the major detractors from the current
quality of life.  These same factors were considered the most
likely cause of any future decline in the quality of life over
the next few years.  The telephone survey quantitatively
demonstrated that over half of tho people interviewed listed
traffic as the major problem and another 25 percent listed
development as the major problem, while the remaining 25 percent
split their concerns between pollution and sewer and water
problems.  When specifically asked about air, water and disposal
of solid waste, residents indicated that they did have a concern
with respect to the oil refinery but those concerns were not at
the top of people's list.

In both the telephone survey and the focus groups, people voiced
their concerns over air, water and waste pollution.  There were
no immediate concerns that the members of the community raised
but nonetheless, when asked, some general concerns were
expressed.

Probably the strongest message underlying this assessment is that
there is not a strong base of support for the Refinery.  People
in general believe that the Refinery probably complies with the
law and that compliance with the environmental laws probably
provides some modicum of protection for the community.  The
Refinery could be caught up in the controversy over the adequacy
of environmental laws, especially if an accident or a spill takes
place.  Probably the most telling indication is the generational
split between young and old; young (under 40) being more
suspicious, probably more concerned with hazardous waste issues
than other kinds of pollution, less confident about the future,
etc.

Another interesting observation is the unorganized, and certainly
unstructured way in which people gain information about
environmental problems—whether it is compliance or pollution.
There seems to be no authoritative source of information about
environmental problems that a majority of people or even a large
number of people accept as reliable.  On the subject of solid
waste, people are suspicious that they will never know what is
contained in wastes until enough generations go by and  problems
surface.  The lack of strong community support for the Refinery
as a positive force in the community and the lack of any real
authoritative source on environmental problems creates the
potential for disastrous community relations.  There is no
effective channel to communicate with the public if problems
arise.

                               3-12

-------
In general people feel that there are more pressing problems in
the community than the Refinery.  Of possible releases to the
environment, the community is more concerned about air pollution
than water or hazardous waste from this facility.  The strong
message that comes through is the desire to establish a credible
"communication channel" that the community can use, if needed.
Another area of concern that surfaced in the focus groups is the
general problem of reduced yields in the Chesapeake Bay.
Although there is no clear linkage between the existence of the
Refinery and the yields, there is no authoritative or credible
source that the community can tap for information.  This may be a
problem of perception that the Refinery may have to deal with in
the long run.
                               3-13

-------
                            Table 3.1

            Pollution Prevention Options  Identified at
                    the Williamsburg Workshop
Flare
Give away excess gas now flared
Redirect sour water stripper vent gas from flare to sulfur plant
Use gas from crude vacuum unit as fuel rather than as flare
Adjust process conditions to reduce flare gas generation

FCU

Use low attrition catalyst
Use Sox Reduction additive  for FCU catalyst
Use more selective catalyst
Capture catalyst fines
Hydrotreat FCU feed
Use Oxygen enrichment in regeneration unit
Integrate regeneration energy with air blower
Eliminate fluid-bed reactors

Fugitives

Implement a leak detection and repair (LDAR) program
Reduce barge loading and other transfer/handling losses
Cover sources of volatiles, including double seals on tanks
Track methane as a greenhouse gas

Refinery Water System

Reduce water content of sludge sent to coker
Enclose inside battery limits  (ISBL) drains
Contain cleaning fluids to reduce evaporation
Redesign desalter system
Reroute desalter effluent
Keep soils out of sewer
Segregate process water effluent and pretreat before discharging
Minimize oil/water contact
Use natural gas for stripping in place of steam
Find substitutes for filter aids
Flash difficult emulsions
Optimize sour water system (SWS)
Use stripped sour water as FCU water wash
                               3-14

-------
                       Table 3.1 Continued

Energy Integration

Achieve better integration with Virginia Power (VEPCO),  including
   giving away flare gas
Optimize plant-wide energy use
Use oxygen enriched air in furnaces

Other

Perform on-line sampling to improve process control
Remove oxygen from feed streams to eliminate heat exchanger
   fouling
Recover sulfur in solid instead of a liquid phase
Filter crude to reduce tank sediments

Corporate Ideas

Produce a single grade of gasoline
Eliminate non-biodegradable products
Desalt at the wellhead and reinject brine and sediment
Use more paraffinic feedstocks
Use renewable feedstocks
Institute overall product stewardship
Co-locate facilities to maximize recycling
Reprocess used lube oil

Ideas Currently Being Evaluated

Eliminate coker blowdown pond
Move sewer and sewer gas adsorption system above-grade
Segregate process water from rain water
Enclose or redesign API separator
Redesign product for lower emissions (lower vapor pressure)
Contain waste streams from cleaning operations

General Comments

Track progress in waste minimization to evaluate improvements
Beware of multi-media transfers
                               3-15

-------
                            Table 3.2

        Selected  Pollution  Prevention Engineering Projects

The following projects were identified for further study as a
result of the Pollution Prevention Workshop in Williamsburg and
subsequent Workgroup meetings.

1.  Reroute Desalter Effluent;  Hot desalter effluent water
    currently flows into the process water drainage system at
    Combination unit.   This project would install a new line and
    route this stream directly to the API Separator.  This
    reduces volatile losses from the sewer system by reducing
    process sewer temperature and oil content.  Volatile losses
    at the API Separator increase slightly.

2.  Improve Desalter System;  Evaluate installation of adjunct
    technology (e.g.,  centrifuge, air flotation, or other
    technology)  on desalter water stream prior to discharge into
    the underground process drainage system.  This reduces oil
    and solids waste loads in the sewer system, affecting the
    waste water treatment plant and volatile losses from the
    drainage system.

3.  Reduce FCU Catalyst Fines;  Evaluate possible performance of
    more attrition resistant FCU catalyst to reduce fines
    production.   (Subsequent review with catalyst vendors
    indicated the Refinery was already using the most attrition
    resistant catalyst available.)  Two other fines reduction
    options were considered.

3a. Replace FCU Cyclones; Assess potential for reducing emissions
    of catalyst fines (PM10) by adding new cyclones in the
    regenerator.

3b. Install Electrostatic Precipitator at FCU;  Assess potential
    of electrostatic precipitator in reducing catalyst fines
    (PM10) emissions.

4.  Eliminate Coker Slowdown Pond;  Change operating procedures
    for coke drum quench and cooldown so that an open pond is no
    longer needed.  This reduces volatile losses from the hot
    blowdown water.

5.  Install Seals on Storage Tanks;  Double seals or secondary
    seals will reduce fugitive vapor losses.

5a. Secondary Seals on Gasoline Tanks;  Install secondary rim
    mounted seals on tanks containing gasoline.
                               3-16

-------
5b. Secondary Seals on Gasoline and Distillate Tanks;   Install
    secondary rim mounted seals on tanks containing gasoline and
    distillate material.

5c. Secondary Seals on ALL Floating Roof Tanks:   Install
    secondary rim mounted seals on all floating roof tanks.

5d. Option 5c + Internal  Floaters on Fixed Roof Tanks;   Install
    secondary rim mounted seals on floating roof tanks  and
    install a floating roof with a primary seal on all  fixed roof
    tanks.

5e. Option 5d + Secondary Seals on Fixed Roof Tanks;  Install
    secondary rim mounted seals on all floating roof tanks and
    the install a floating roof with a primary and secondary seal
    on all fixed roof tanks.

6.   Keep Soils out of Sewers;   Use road sweeper to remove dirt
    from roadways and concrete areas which would otherwise blow
    or be washed into the drainage system.  Develop and install
    new sewer boxes designed to reduce soil movement into sewer
    system, particularly  from Tankfarm area.  Estimate  cost for
    installation on a Refinery wide basis.  Both items  reduce
    soil infiltration, in turn reducing hazardous solid waste
    generation.

7.   The Benzene Waste Operations NESHAP requires control of
    benzene emissions from refinery wastewater sources.  Three
    separate projects (7A, 7B, and 7C) were identified  to meet
    these requirements.   Specific design and construction
    features of these projects will provide for compliance with
    some future regulations, such as storm water permitting, RCRA
    remediation, the Primary Sludge rule and land disposal
    restrictions.

7A. Convert Slowdown Stacks;  Replace existing atmospheric
    blowdown stacks with  flares.  This reduces untreated
    hydrocarbon losses to the atmosphere, but creates criteria
    pollutants.

7B. Drainage System Upgrade;  Install above-grade, pressurized
    sewers, segregating storm water and process water systems.

7C. Upgrade Process Water Treatment Plant:  Replace the API
    Separator with a covered gravity separator and air  floatation
    system.  Capture hydrocarbon vapors from both units.

8.   Change Sampling Systems;  Install flow-through sampling
    stations (speed loops) where required on a refinery-wide
    basis.  These replace existing sampling stations and would
    reduce oil load in the sewer or drained to the deck.
                               3-17

-------
9.  Reduce Barge Loading Emissions;  Estimate cost to .install a
    marine vapor loss control system.  Consider both vapor
    recovery and destruction in a flare.

10.  Sour Water System Improvements:  Sour water is the most
    likely source of Refinery odor problems.   Followup on
    projects previously identified by Linnhoff-March engineering
    to reduce sour water production, and improve sour water
    stripping.

11.  Institute LDAR Program;  Institute a leak detection and
    repair program for fugitive emissions from process equipment
    (valves, flanges, pump seals, etc.) and consider costs and
    benefits.

    a.  Annual LDAR Program with a 10,000 PPM hydrocarbon leak
        level

    b.  Quarterly LDAR Program with a 10,000 PPM hydrocarbon leak
        level

    c.  Quarterly LDAR Program with a 500 PPM hydrocarbon leak
        level
                               3-18

-------
                                                          Table 3.3




                                           AMOCO/EPA Pollution Prevention  Project
OPTION
1
2
3a
3b
4
5a
Sb
5c
5d
5e
6
7A
7B
7C
8
9
10
lla
lib
lie
Reroute Desalter Water
Improve Desalter System
Replace FCU Cyclones
Install ESP at FCU
Eliminate Coker B/D Pond
Install Sec. Seals on Gasoline Tanks
on Gasoline and Distillate Tanks
on all Floating Roof Tanks
Option 5c & Floaters on Fxd Tanks
Option 5d & Sec. Seals on Fxd Tanks
Decrease Soils in Drainage Systems
Slowdown System Upgrade
Drainage System Upgrade
Water Treatment Plant Upgrade
Modify Sampling Systems
Reduce Barge Loading Emissions
Sour Water System Improvements
Annual LDAR Program @ 10000 PPM
Quarterly LDAR Program @ 10000 PPM
Quarterly LDAR Program @ 500 PPM
NET
RELEASE
REDUC-
TION
TONS/YR
52.4
U/D
245.0
442.0
130.0
474.7
482.1
541.0
591.7
592.2
530.0
5096.0
112.5
58.0
63.0
768.0
18.0
319.5
510.5
705.5
EFFICIENCY
90.0%
U/D
48.0%
87.0%
5C.O%
75.0%
76.0%
85.0%
93.0%
94.0%
50.0%
98.0%
95.0%
95.0%
100.0%
98.0%
100.0%
40.0%
64.0%
89.0%
MATERIAL
VOC'B
U/D
Catalyst Fines
Catalyst Fines
VOC's
VOC's
VOC's
VOC'B
VOC'B
VOC'B
Listed Hazard Waste
VOC'B
VOC'B
VOC's
VOC's & Heavy HC
VOC'B
Sulfide, NH3, &
Phenols
VOC's
VOC's
VOC's
TIMING
YEARS
1-3
U/D
4-7
4-7
1-3
>7
>7
>7
>7
>7
4-7
4-7
1-3
1-3
4-7
1-3
1-3
<1
<1
<1
POLLUTION
PREVENTION MODE
Recycle
U/D
Recycle/Disposal
Disposal
Source Reduction
Source Reduction
Source Reduction
Source Reduction
Source Reduction
Source Reduction
Source Reduction
Treatment
Treatment
Treatment
Source Reduction
Recycle
Recycle/Treatment
Source Reduction
Source Reduction
Source Reduction
U/D = Undefined

-------
                                                          Table 3.4a
                                            AMOCO/EPA Pollution Prevention Project
                                                      Financial Summary
OPTION
1
2
3a
3b
4
5a
5b
5c
Sd
5e
6
7A
7B
7C
8
9
10
lla
lib
lie
Reroute Desalter Water
Improve Desalter System
Replace FCU Cyclones
Install ESP at FCU
Eliminate Coker B/D Pond
Install Sec. Seals on Gasoline Tanks
on Gasoline and Distillate Tanks
on all Floating Roof Tanks
Option 5c & Floaters on Fxd Roof Tanks
Option 5d & Sec. Seals on Fxd Roof Tanks
Decrease Soils in Drainage Systems
Slowdown System Upgrade
Drainage System Upgrade
Water Treatment Plant Upgrade
Modify Sampling Systems
Reduce Barge Loading Emissions
Sour Water System Improvements
Annual LDAR Program 010,000 PPM
Quarterly LDAR Program 610,000 PPM
Quarterly LDAR Program 8500 PPM
PV OF
CAPITAL
M$
1,000
Undefined
8,300
9,100
2,000
212
262
364
1,532
1,682
213
4,521
18,800
22,500
67
4,700
605
5
5
5
PV OF
O&M
M$
854
Undefined
8,885
10,413
1,590
203
253
349
1,468
1,610
594
3,712
14,951
19,203
67
4,287
517
421
621
872
ANNUALI ZED
COST
M$/YR
317
Undefined
2,939
3,337
614
71
88
122
513
563
138
1,408
5,772
7,132
23
1,537
192
73
107
150
COST-
EFFECTIVENESS
$/TOM
6,050
Undefined
11,996
7,550
4,723
150
183
226
867
951
260
276
51,307
122,966
365
2,001
30,667
229
210
213
ote: All cash flows are discounted at 15%,lb Year project  life.
     Capital spending for all projects is assumed to begin  in  1991.
     O&M •= operating and maintenance coets, depreciation, indirect coate, taxes,
end insurance.

-------
                                                           Table  3.4b
                                            AMOCO/EPA  Pollution  Prevention Project
                                                       Financial Summary
OPTION
1
2
3a
3b
4
5a
5b
5c
5d
5e
6
7A
7B
7C
8
9
10
lla
lib
lie
Reroute Desalter Water
Improve Desalter System
Replace FCU Cyclones
Install ESP at FCU
Eliminate Coker B/D Pond
Install Sec. Seals on Gasoline Tanks
on Gasoline and Distillate Tanks
on all Floating Roof Tanks
Option 5c & Floaters on Fxd Tanks
Option 5d & Sec. Seals on Fxd Tanks
Decrease Soils in Drainage Systems
Slowdown System Upgrade
Drainage System Upgrade
Water Treatment Plant Upgrade
Modify Sampling Systems
Reduce Barge Loading Emissions
Sour Water System Improvements
Annual LDAR Program 010000 PPM
Quarterly LDAR Program 010000 PPM
Quarterly LDAR Program @500 PPM
PV OF
CAPITAL
M$
1,000
Undefined
8,300
9,100
2,000
259
321
445
1,827
2,003
337
5,095
18,800
22,500
76
4,700
605
5
5
5
PV OF
O&M
M$
1,502
Undefined
14,738
18,153
2,807
426
531
734
3,018
3,306
1,207
7,303
26,388
33,808
129
7,531
909
695
1,045
1,478
ANNUALIZED
COST
M$/YR
329
Undefined
3,O29
3,583
632
90
112
155
637
698
203
1,630
5,941
7,403
27
1,608
199
92
138
195
COST-
EFFECTIVENESS
$/TON
6,279
Undefined
12,363
8,106
4,862
190
232
287
1,077
1,179
383
320
52,809
127,638
429
2,094
11,056
288
270
276
Note: All cash flows are discounted at10%, 15 year project life.
      Capital spending for all projects is assumed to begin in 1991.
      O&M «= operating and maintenance costs, depreciation, indirect coats, taxee, and  insurance.
                                                             3-21

-------
                                                         Table 3.4c
                                           AMOCO/EPA Pollution Prevention  Project
                                                      Financial  Summary
OPTION
1
2
3a
3b
4
5a
5b
5c
5d
5e
6
7A
7B
7C
8
9
10
lla
1*0
lie
Reroute Desalter Water
Improve Desalter System
Replace FCU Cyclones
Install ESP at FCU
Eliminate Coker B/D Pond
Install Sec. Seals on Gasoline Tanks
on Gasoline and Distillate .Tanks
on all Floating .Roof Tanks
Option 5c & Floaters on Fxd Tanks
Option 5d & Sec. Seals on Fxd Tanks
Decrease Soils in Drainage Systems
Slowdown System Upgrade
Drainage System Upgrade
Water Treatment Plant Upgrade
Modify Sampling Systems
Reduce Barge Loading Emissions
Sour Water System Improvements
Annual LDAR Program 010000 PPM
Quarterly LDAR Program @10000 PPM
Quarterly LDAR Program 6500 PPM
NET
PRESENT
VALUE
H$("
-995
Undefined
-8,810
-11,773
-1,865
3
-74
-229
-1,838
-2,137
-130
-5,585
-18,767
-23,731
-41
-4,321
-837
13
31
10
NET
ANNUALIZED
COSTP)
M$
-131

-1,158
-1,548
-246
0.4
-10
-30
-242
-281
-17
-734
-2,467
-3,120
-5
-568
-110
2
4
1
EMISSION
REDUCTIONS
TONS/YR
52.4

245.0
442.0
130.0
474.7
482.1
541.0
591.7
592.2
530.0
5,096.0
112.5
58.0
63.0
768.0
18.0
319.5
510.5
705.5
NET
COST-
EFFECTIVENESS
$/TON0>
-2,500

4,728
3,502
-1,886
1
-20
-56
-408
-4/4
-32
-144
-21,933
-53,793
-86
-740
6,114
6
8
2
RATE OF
RETURN*4'





10
7
3


3



1


14
19
13
otes:    1. Net present  value ofall  cash  flow,  including initial  capital,  operating  and  maintenance  expenses,  taxes,
           depreciation,  indirect  costs,  revenues,  etc.
        2. Stream  of equal payments,  discounted  at  10%,  and a  project  life of  15  years which equal the  net  present
           value.
        3. Net cost-effectiveness  reflects  value of recovered  product  (5 • 75/gallon);  calculated by dividing net annualized
           cost  by Emission Reductions  and  multiplying by 1000.
        4. The rate of  return  is the  internal rate  of  interest at which  the costs and benefits  are equal  over  the life  of
           the project.
                                                            3-22

-------
                   Table 3.5
Maximum Annual Average Predicted Concentrations
  For The Yorktown Refinery for  BTEX Chemicals
CHEMICAL
BENZENE
TOLUENE
ETHYLBENZENE
XYLENE
INSIDE
PLANT
FENCELINE
CONG.

-------
                                                Table 3.6
                        Maximum Annual Average Benzene Concentrations and Benzene
                      Exposure  Associated with Var^ jus Pollution Prevention Options
OPTION (SEE TABLE 3.2
FOR DESCRIPTIONS
0
1
4
5 a
5e
7A
7B
7C
7C +7B
9
lla
lib
Base Case
Desalter Control
Coker Pond Control
Secondary Seals on
Gasoline Tanks
Secondary Seals on all
Tanks
Slowdown Stacks to
Flare
Drainage Controls
Trtm't Plant Upgrade
Combination
Capture Barge Loading
losses
Annual LDAR
Quarterly LDAR
BENZENE CONC. AT
NEARBY RESIDENCE
(ug/M3)
0.62
0.61+
0.61
0.51
0.50
0.55
0.59
0.59
0.56
0.28
0.61
0.60
PERCENT OF BENZENE
EXPOSURE COMPARED TO
BASE CASE
100
99
98
82
80
89
95
95
90
45
98
97
PERCENT REDUCTION
IN BENZENE
EXPOSURE
-0-
1
2
18
20
11
5
5
10
55
2
3
Jotes:  1. Options 5a and 5e cover the  range of control which  includes Option 5c,

       2. Options 3a,  3b, 6, 8, and  10 do not effect  benzene  emissions.
                                                  3-24

-------
                            Table 3.7
      Copper  Concentrations  in Yorktown Refinery Effluents
                     (concentrations  in ug/1)
Date
8/13/91
9/9/91
4/2/91
1/29/91
10/30/90
8/26/90
6/25/90
Average
Outfall 001
NA
3
192
64
<50
40
96
66
Out.all 002
<1
NA
128
93
<50
30
127
76
Notes:

Virginia Electric Power has reported the following values for
their fresh water intake from the York River:
4/16/91
1/15/91
10/23/90
2
1
8
                               3-25

-------
             Figure 3.1
             Yorktown  refinery
             Simplified Flow Diagram
             with  Release Sources
             and Control Options
                     Crude oil
Wa«"  +OO
fo sewer
                           ;  TJI;
                           Q_    ,,  .  .  |\ Catalytic cracking
                           DesuHurization [)             "
                      Spent
                      catalyst
                      T
\jlinll J
                    7b rtcov-e/y,
                    recycle, and
                    landfill
                               QQ
          /-/"	VV^~N>
          (  Fugitive HC   \
       •SElli
     Uttra form ing

        Water to sewer
T
                                         f ugrtnv A/C ) *
                                                                 A
                                                                                                     H
                                                                                             Slowdown
                                                                                             nocks
                                                   , To recovery.
                                                    recycle and landfill
                                        Sulfur recovery
             ^ Refinery
          |l_   fuel gas
' Vipor recovery
                                          QQ
                                                           Polymerization—MTBE
                               CZI)  Treating and blending
                               v—'        tanks
                           Furnace oil
                                ^-V^X
                                f  ftyp/bve  1
                                XHC It J
                                 X^VK
                                      i  \
                                                    Gasoline
                                                                                                                Treated water
                                                                                                                to Vbrtc /Iv
                                                                                              To recycle and landfill
                                                                          Sulfur
                                                   Marine
                                                   loading
                                                                            Pollution Prevention
                                                                            Project Numbers
                                                                            refer to Table 3.2

                                                                            Solid
                                                                            Airborne emission


                                                                            Surface or
                                                                            ground water
             Petroleum coke
                                  Liquefied
                                  petroleum gas
                                  (LPG)

-------
                                                                                                                      .v '!"•:  ..'"Ml K.A I?- [C':iNj>:   •   ,  ., /:;,»
                                                                                             1 \'l
            :y.7irvv<
,.

LJ
I—
     41
i	
                                           •                  2T-                             T           '
                                                                                                          !       !       !       .[.    .  \

-------
Figure 3.3  SOURCE CULPABILITY FOR BENZENE
                                  IN YORK RIVER
          CO
          z
          o
          I
          LU
          O
          z
          o
          o
7.00-
8.00-
5.00-
.00-
3.0O-
2.00-
.00-






—_^^_
£
^
|
^
^
^
s
^
^
•^
•4^
<«.
^>
-*•
•*.
-*•
^
^
^
^
I
i
-i
^
-)•

g
rs
~




''/.
1
^
^
^
•^
-*-
*-
^«
^
4-
^
^
^
•^
t-
^
^
^^
*•
^-







-1U
r90
•80
^70
•50
•40
•30
•20
•10
FUOmVES  API   BARGE  SEWER COOLING  TOTAL

     TANKS  H2OTANK COKEfl   LAND SLOWDOWN

                 SOURCE
                                                 z
                                                 o
                                                 H

                                                 CQ
                                                 cr
                                                 t-

                                                 o
                                                 o
UJ
o
a:
UJ
a_
                           CONCENTRATION
                          PERCENT OF TOTAL
                               AT FENCEUNE ON UVND
          o

          I
          UJ
          u
          2
          O
          o
1.4O-
1.20-
1.00-
0.8O-
0.80-

0.20-

g




ffil

^
•4MB
Ta
3





«•

^
^
-4
*4
kH


«n, VF\

fy
't
^
^
pv
^
^
^
^
-*-
^
-^
^
-»-
-*-
4-
•r-
-f
T-
t







- im
•90
-ao
•70
•40
DO
•20
•10
                             COOUNO

     TANKS   K2OTANK  COKER    LAND  BIOWOOWM


                  SOURCE
                                                 O
                                                 o
                                                 cr
                                                 UJ
                                                 Q.
       E23 CONCENTRATION
                                                    TOTAt
                               NEARBY RESIDENCE
          Z
          O
          o

          o
          o
a/o-
0.60-
0.50-

0.30-
0.20-
0.10-



^

/
y
/
i


/
/

-------
                    Figure 3.4
       Histogram of Benzene  Emissions
 With and Without  Marine Loading  Controls
Benzene, tons/year
 67.2
                      15.0
                  10.5
  Total   B/D Stacks
Tanks    Marine Waste Water  Coker
 Emissions Source
Fugitive
             Base
       Marine Vapor Control

-------
                 FIG1.. IRC  .:' . l-: .
                                                        ANN'.iAL  AVF.RAGF  •."A'.1N'.'>_NTRATK';N::-; ' ., Q /m
                                                        f'A-vFi  [N'T.
367000
   4121500  -
                   36t>000    36 9000    370000     .V 7 1000    37k000    373000     374**»«    375000
J76800    277000
                                                                                                            4121500
   4129600 ^
                                                                                                            4120G00
Q
(Y
C
c
   4119600

   4117500
                     *v                     . •    •^.-,  ..-        .
                   "^  *>>  '• •  -    '     •  //  'X  '    - ';f- -V* -W-  • v-y^-.  ••:-;-
                    ''..  V  ;    '<•   -XV"-)'-   •  --*'-\-.vA--.^  -.  r--*
                   '.  tj  " . ..YV«,'...V-  Ar K-^,, ^; V,,,.-.^.,' :^...^. .v/' -•-,'     ,-
                     •	-^	llf..-.J	L_. .. L	1	^1	1-	1	_L_
                                                                      i     I  ...  L. •   !.._.:
                                                              373000 .   374i^w    375000

                                     .;Tr:  EAST  COORDINATES
         	1 4! ! 030

          377000

-------
    4121 ?v
4i2v*?vw
'oJ
K
<
z
H-l
o
iy

                                                   II	IIIIIHII    • •"-'*-'• -''  •
                                       Ill^^^gfe^:

                                       U11^  d  *  -1  -'-"   rS           ^
                                                              •3»r

                                                                                                                                             •* 1 21
                                                                                                                                              -1  4 1
                                                                                                                                                 411
                                                                                                                                                 41

                                                                                     X  ..

                                                                                                               !  •    L .. .  .1...L.L
                                                                                                                                                 •* 1  1 '

-------
                                                    ••jy.t "V-]F   A^N'JA.L   AV'FJVA'V   ' .v''N'V.LNTF«-.'A.'
                                                            'V A,_<:,
                                                                            .'A-
                                                -Jl  JUiUGur   STACK'-:
                     368000     349000     370000     371000
                                                 -__r
   4121500
• 1 ••••-!  "

-------
                                                                                                                 003SI Iv
                           ... -^ >*  .' s"    ,	1
                        x^-«t««N   /s-     '    I
;,.t\j; ~; T   \y>j i tsi T

-------
                         Figure 3.9
            Projected  Rates of Return
  Pollution  Prevention Projects  - Yorktown
      25
      20
      15
      10
         Percent
          Historical Rate-of-Return for Refinery Projects
                                           i  i  i m
       0
          1    3b   5a    5c    5e   7a   7c    9
           3a    4   5b   5d   6    7b   8    10  11b
                        Project Number
                    (See Table 3.2 for Descriptions)
Notes 1. Projects for which no values are shown have negative returns.
     2. Rate of return is the rate at which benefits and costs are equal for the life of the project.

-------
4.0  RANKING METHODS AND RESULTS

This Section describes the process used to rank pollution
prevention options discussed in Section 3.0.

4.1  Option Characteristics

As noted in Section 3.0, the Workgroup selected 12 options from a
long list of 'options identified at the Williamsburg Workshop.
Important characteristics of the 12 options, and their
alternatives, are summarized in Table 4.1.  For three options—3,
5 and 11—only one of the several alternatives considered would
be implemented.

Two options reduce solid wastes (catalyst fines and listed
hazardous wastes),  while the remaining ten focus on air emissions
(VOC, HC, H2S, and NH3).  Five of the twelve options employ
source reduction to reduce releases.  Capital costs range from a
low of $10 M to a high of $22.5 MM1.   Annual costs,  based on
discounting capital, operating and maintenance costs at a 10
percent discount rate, range from $30 M to $7.4 MM.  The 250-fold
ratio between largest and smallest annual costs is significantly
less than 2,250-fold ratio between largest and smallest capital
investment.

    4.1.1   Cost Per Ton

The cost per ton reduction in environmental releases (cost-
effectiveness of release reduction excluding revenues)  is given
for each option in Table 4.1.  The most cost-effective options,
on the basis of release reductions, are three LDAR alternatives
(lla, lib, lie) requiring an average of $278 per ton and three
secondary seal alternatives  (5a, 5b, 5c), which require an
average of $236 per ton.  Three other options also are low cost,
$429 per ton or less.  However, the cost for each of the other
seven options is $2 M per ton or more.  The Treatment plant
upgrade represents the least cost-effective option, requiring an
annual cost of $128 M for each ton of VOC recovered.  Another way
of viewing an option's cost-effectiveness is to determine the
price at which the recovered material would have to be sold in
order to offset the cost of recovery.  As shown in the Recovery
Cost colunn of Table 4.1, the recovered material would have to be
sold at $0.90 per gallon for the LDAR and secondary seal options
to break even, and at $415 per gallon for the most expensive
option, treatment plant upgrade.  For comparison, the Refinery
price is about $0.75 per gallon.
     1h = thousands;  MM = millions

                               4-1

-------