Hu;
vvEPA
United States Solid Waste and . EPA530-R-94-002
Environmental Protection Emergency Response PB95-167235
Agency (5305) -. January 1995
One-time Waste
Estimates for Capacity
Assurance Planning
Capacity Planning Pursuant
to CERCLA Section 104(c)(9)
EPA Headquarters Library
Recycled/Recyclable
Printed on recycled paper that contains at
least 50% post-consumer recycled fiber
-------
r
-------
TABLE OF CONTENTS
BACKGROUND AND SUMMARY
CHAPTER 1. SUPERFUND REMEDIAL ACTIONS
CHAPTER 2. SUPERFUND REMOVAL ACTIONS
CHAPTERS. RCRA CORRECTIVE ACTIONS
CHAPTER 4. UNDERGROUND STORAGE TANKS CONTAINING HAZARDOUS
SUBSTANCES
CHAPTER 5. STATE AND PRIVATE CLEANUPS
2
5
14
22
36
46
-------
-------
BACKGROUND AND SUMMARY
Section 104(c)(9) of CERCLA requires states to assure adequate capacity for the treatment
and disposal of hazardous wastes that are reasonably expected to be generated within a state for 20
years before any remedial action is provided, by EPA under section 104. This assurance, the basis of
which is in the form of Capacity Assurance Plans (CAPs), must be provided in a contract or
cooperative agreement entered into between the state and the Administrator. If such an assurance is
not provided, no Superfund financed remedial actions can be provided.
States must provide an assurance that addresses any hazardous waste (i.e., recurrent and
remedial) reasonably expected to be generated within the state. The 1993 Guidance for Capacity
Assurance Planning addresses the issue of how states should make the capacity assurance for
recurrent wastes. This particular report is part of,the Agency effort to assist states in assuring
capacity for one-time wastes. The Agency began working on this effort over two years ago in
response to states' concerns over the difficulties they faced when developing one-time waste
projections for their 1989 CAPs. Specifically, the National Governors' Association's (NGA) CAP
Policy Development Workgroup made a recommendation to form a workgroup of state and EPA
representatives to develop approaches to calculate future one-time waste generation. The proposals
developed by'this.workgroup.provided the basis for an effort EPA subsequently undertook with a
research group at Oak Ridge National Laboratories/University of Tennessee. The methodologies
developed from this collaborative effort were revised after consultation with the appropriate EPA
program offices, a presentation to the NGA CAP Policy Development Workgroup, and comments
received from the states; . . '
This report contains detailed descriptions of.the methodologies the Agency used to develop
tonnage estimates representing twenty years of off-site shipments to commercial Subtitle C hazardous
waste management facilities. The Agency used the one-time waste estimates which appear in
Appendix A when it conducted the national assessment of all states' CAP data. .This report discusses
the methods for calculating wastes associated with the five major sources of remediation activities:
Superfund remedial actions; Superfund removal actions^RCRA Corrective Actions; Underground
Storage Tanks cleanups; and State and Private cleanups. The Agency will niake publicly available in
the fall of 1,994 another report which describes how states can reduce the generation of these wastes
through the promotion of on-site treatment using conventional and innovative technologies.
The five methodologies identify for each source of remediation: (1) the potential sources of
contamination (e.g., the number of tanks containing hazardous waste, the NPL sites that have the ,
potential to send .waste off-site); (2) the type of .contamination (e.g., organics, metals) to determine
the appropriate treatment; (3) the probability that the waste generated at these sites will be sent off-
site for treatment and disposal; (4) the waste tonnages that will likely be sent off-site; (5) the tonnage
of treatment residuals generated from treatment of these wastes; (6) the probability of disposal of the
waste and residuals in Subtitle C versus Subtitle D landfills; and (7) the distribution of waste tonnages
twenty .year period. ' - - ' . -
All of the methodologies presented in this report contain cross-cutting assumptions that apply
nationwide and are derived from both the analysis of historical data on cleanups and the interpretation
of the impacts of current Agency policies on one-time waste cleanups. The supporting documentation
for the assumptions can be found in RCRA Docket F-92-CAGA-FFFFF The primary assumptions ,
include-the following: . ' . - /''..'
-------
Since wasfewater contamination at clean-up sites is typically treated oh-site
using pump and treat technologies, all the methodologies in this report assume
.that wastewaters at remedial sites will be treated on-site and that residuals
from such treatment are negligible. .
Because data indicate that the majority of remediation wastes shipped off-site
to Subtitle C managements are sent to incinerators, landfills, and/or, facilities
that stabilize wastes, these were the only managements considered in the
methodologies. Moreover, since hazardous remediation wastes contaminated
with organic constituents and shipped off-site to commercial Subtitle C
facilities typically are treated by incineration and those contaminated with
metals are typically stabilized, the methods.assume organic wastes are
incinerated and metal wastes are stabilized.
Since the 1993 CAP Guidance addresses incinerator ash and stabilized
residuals shipped to Subtitle C landfills, the one-time waste methodologies
also considers these residuals. However; because residuals generated by
treatment of contaminated media generally have a higher inorganic content
than residuals from recurrent waste, different residual factor were established
for one-time residuals. Residual ash amounts for wastes treated with
incineration were calculated by using a factor of 1.6 (i.e., soil into an
incinerator equals soil out of an incinerator) and 1.5 for waste destined for
stabilization (i.e., fifty percent increase in amounts disposed of in landfills
after stabilization). ' ~
For the purposes of the Agency's assessment of capacity, States were asked to
account for waste demands from 1991 to 2013. States are responsible for
submitting Biennial Reporting System (BRS) data or its equivalent to
determine remediation quantities for 1991 (i.e., data submitted in CAP Table
3 for one-time wastes). EPA has developed waste tonnages for each year
from 1992 to 1999. From 1999 to 2013, EPA will assign the average tonnage
for the seven year period from 1992 to 1999.
EPA excluded from the methodologies remediation wastes generated by
federal facilities. EPA investigated primarily Department of Defense (DOD)
and Department'of Energy (DOE) facilities since they havi the majority of the
federal facilities sites that need remediation. An EPA analysis of clean-ups at
DOD facilities sites showed that most management of remedial waste occurred
on-site. EPA expects this practice to continue because of DOD policies which
promote on-site treatment, and the reality that many .cleanup wastes at DOD
facilities are dangerous to transported and require specialized management
(e.g., wastes with explosive contaminants).- Because many DOE remediation
sites are contaminated with mixed hazardous/radioactive wastes and these
wastes have been excluded from the CAP pursuant to the 1993 Guidance due
to transportation and human handling/exposure concerns, DOE facilities were
not considered in the methodologies.
-------
These general assumptions, as well as the specific assumptions developed for each
methodology, are based on nationally-available data. EPA'recognizes that states may have more
accurate, state^specific data for each methodology described in this report. Many states did send in
data or comments on these methodologies, which the Agency incorporated into this final document.
-------
1. SUPERFUND REMEDIAL ACTIONS
1.1 INTRODUCTION
This chapter presents the methodology used to estimate the total amount of one-time
hazardous wastes generated by Superfimd remedial actions for the years 1992 to 2013 on a state-by-
state basis. Superfund remedial actions are the actual construction and implementation of a Superfund
remedial design that results in long-term site cleanup. The Superfund program identifies sites where
hazardous substances have been, or might be, released into the environment; ensures that these
substances are cleaned up by responsible parties or the government; and evaluates damages to natural
resources.
The methodology that the Agency has developed uses data on existing National Priority List
(NPL) sites to estimate waste generation each year through 1999. EPA assumes an average annual
rate of waste generation from 2000 to 2013 based on the average annual waste volume in each State
from 1992 to 1999. Only those hazardous wastes requiring RCRA Subtitle C off-site commercial
treatment or disposal capacity are included in this study. This methodology does not take into
account potential changes to the Superfund program resulting from Superfund Reauthorization.
1.2 DATA SOURCES
EPA utilized numer;. .
-------
Exhibit 1-1
Data Sources For. the Superfund Remedial Action Methodology
Database/
Report
HAZDATA .
BASECOST
SUM ROD
NPL Technical
Data Files
CRES
(schedule of
NPL site ' '
events)
Guide to
Superfund
Sites
EPA ROD
Annual
Reports
*
Data Element
' ..
site name
site location
site type
ROD date
site characteristics
contaminant sources
contaminant types
volume
remediation technology
site name
volume
remediation technology duration
site name
site location
ROD date - ' ,
media type
contaminant types
site name
site location
site activity
media type
contaminant types
remediation technology
date in NPL
site name <
site location
event type ' '
estimated start year
estimated end year .
actual stan year
actual end year
site name .
site size
volume
contaminant type . -
site type
site name
site size
volume ;
contaminant type
site type
remediation technology , '
Data Years
Used
1987 to
early. 1990
,
1987 to
early 1990 '
1983 to '
early 1990
as of
February
1990
1982 to
early 1992
1982 to
early 1992
1982to ,
1991
'.
, Who Collected
ORNL/UTK
.
ORNL/UTK
Ontario Ministry
of the
Environment
EPA
Pasha
Publications
R. C.
DiGregorio/
Pasha
Publications
EPA
,
Data Sources
EPA RODs, NPL
Technical Data Files, and
EPA contacts
1
same as HAZDATA
(companion database)
EPA RODs
i
-
_
EPA RODs and EPA
SCAP1 1 report
EPA RODs + EPA
reports and contacts
1 ,
EPA RODs
-------
Exhibit 1-1 (continued)
Data Sources for the CERCLA Remedial Action Methodology
Database/
Report
HW!R RODs
Database
TIO report
(EPA, 1993b)
Data Element
. !
site name
site location
ROD date
volume
contaminant types
remedy location
name, location, contaminant type, and
media type for sites without RODs;
summary statistics on past RODs
-
Data Years
Used
1992 to
1993
1982 to
1991 and
projections
to 1996
Who Collected
ICF
Incorporated
EPA,
Technology
Innovation
Office
Data Sources
EPA RODs
EPA RODs
BASECOST contains waste volume information f9r individual remediation technologies
associated with the sites or operable units reported in the HAZDATA database. A total of 548 .
records comprise the BASECOST database. It also .includes estimates for the duration of the cleanup
under the recommended remediation technology.
SUMROD
This database contains data extracted from RODs signed between 1983 and early 1990. The
database was compiled by the Ontario Ministry of the Environment to assist them in developing soil
cleanup criteria. The database is organized on a compound-by-compound basis, and includes site
name and location, date the ROD was signed, media type, cleanup goal, and the site contaminants.
U.S. EPA ROD Annual Reports for FY 1990 and 1991 i '
RODs and ROD. Amendments for all Superfurid sites signed within ail EPA fiscal year are '
.documented in annual reports published by EPA. -The'annual reports for fiscal years 1990 and 1991
were used extensively for this project. Each abstract for RODs signed in 1990 and 1991 was " .
reviewed, and those containing the keywords "Off-site Treatment" or "Off-site Disposal" were
selected as sites with potential for generation of off-site wastes. Summary tables provided in these
annual reports also include overviews of site problems, selected remedies, clean-up criteria, and
estimated costs for all RODs signed between 1982 and 1989. For RODs signed during 1982-1989,
the keyword search was applied to the summary tables to identify those'that recommended off-site
treatment or disposal as their remedial technologies. RODs signed prior to 1986 were checked
against the current NPL, and sites that had completed their cleanup efforts prior to 1992 or had been
deleted from the NPL were removed from the final data set.
Hazardous Waste Identification Rule RODs Database
v This database was compiled by EPA's Office of Solid Waste in support of economic analysis
for future rulemakings regarding the Hazardous Waste Identification Rule (HWIR). The database
includes data on contaminated soil, sediment, debris, and waste (e.g., sludge) mixed with soil and/or
sediment from RODs signed in 1989 through 1992 and from some 1993 RODs. The database is
organized by site and includes data on waste volumes, contaminants, contaminant concentrations, and
-------
in-situ versus ex-situ management. Data for RODs signed in 1992 and 1993 were used by EPA for
the CAP one-time waste projections.
e
EPA pooled data from HAZDATA, BASECOST, SUMROD, the EPA ROD annual reports,
and the HWIR RODs database to form a single database for all RODs at NPL sites expected to
generate oneTtime wastes during the projection period of 1992-1999. However, these data are not
sufficient themselves for projecting future generation because RODs do not always contain waste
volumes, waste management methods, or whether the waste will be managed on site or off site.
Additionally, EPA has not completed RODs for all sites currently on the NPL. Therefore, EPA
supplemented RODs data with information from other sources described below.
1.2.2 NPL Technical Data Files
This database contains information for approximately 1,200 sites on the NPL'as of February
1990. The four major categories of data are Hazard Ranking System (MRS) scoring data, site
documentation data, administrative data, and auxiliary data. Data elements include site name and
location, site'activities, contaminated media type (e.g., soil, sediments, ground water), types of
contaminants, contamination impact, remediation technology, site ownership, and the date that the site
was added to the NPL. The database was used by English (1991) to help create the HAZDATA and
BASECOST databases.
1.2.3 1992-1993 Guide to Superfund Sites
This report, compiled and edited.by R. C. DiGregorio of Pasha Publications Incorporated
(DiGregorio, 1992), contains,status reports for over 1,200 sites listed in the NPL as of the end of'
1991. It provides site history and technical information such as the recommended remedial
technologies. EPA used this publication in conjunction with the EPA'ROD Annual Reports to obtain
supplemental information on site size, waste types, and waste volumes. In the event that
discrepancies among the sources were found, EPA relied on information reported in the EPA ROD
Annual Reports. . .
.' ' , - .
1.2.4 CERCLIS Remedial Event Schedule (CRES) Database
This database was prepared by Pasha Publications, Incorporated and includes 3,152 records,
each representing an event scheduled for the Superfund sites, as of early 1992. The,CRES' database
provided data on the actual or planned year of cleanups and was used to calculate the average
duration.of the steps in the remedial action process for sites whose schedules were not provided in the
RODs. .
1.2.5 EPA Technology Innovation Office Report, Cleaning Up the Nations Waste Sites: Markets
and Technology Trends (TIO Report) (EPA 1993c)
The TIO report provides data for individual NPL sites without RODs as of September 30,
1991. EPA used these data, including site name, location, media contaminated, contaminant types,
and planned ROD date, to estimate waste volumes for sites without RODs. EPA also used data from
the TIO report to estimate waste volumes for sites with RODs that did not contain volume data, and
to estimate the proportions of remedial action waste managed in different CAP Management
Categories. ' .
-------
1.3 METHODOLOGY
EPA used site-specific data to estimate State-by-State and year-by-year waste volumes from
1992 to 1999. It then assumed a constant annual waste generation in each State from 2000 to 2013
based on ah average of the volumes from the proceeding years. This approach is consistent with the
projection methodology for recurrent wastes.
1.3.1 Identify Sites with Potential Off-site Waste Generation
EPA compiled data for all.NPL sites that will potentially generate one-time wastes that will be
managed off site between the beginning of 1992 and the end of 1999. Two types of sites were
included: sites with RODs and sites without RODs. EPA-excluded sites with RODs where the
selected remedy will include on-site waste management only, and sites with ground water
contamination only. In addition, as stated in the Introduction, EPA excluded federal facilities from
this methodology. - ,
Sites without RODs were identified in the TIO report (EPA 1993c). Appendix A of the TIO
report lists all sites on the NPL! without RODs, as of September 30 1991. EPA used media
contamination data in the TIO report to identify sites expected to generate off-site wastes. In
particular, EPA assumed that sites identified in the TIO report which have only ground water
contamination will not generate off-site wastes, and sites with contaminated soil or sediment or other
hazardous wastes will have the potential to generate off-site wastes.
Because site data were compiled from several existing sources, EPA compared all data
sources to ensure that ROD data for a single site were not included twice. Sites may appear in the
,-data set more than once, however, if separate RODs were issued for different parts of the site.
1.3.2 Estimate Volume of Waste to be Generated at Each Site
The total .quantities of hazardous waste expected to be generated from Superfund remedial
actions are generally,estimated during the Remedial Investigation/Feasibility Studies (RI/FSs), and
documented in the RODs. In the event that the total volume of hazardous wastes generated from a.
site was not specified in the ROD, EPA reviewed supplemental data sources:(e.g., 1992-1993 Guide
to Superfund Sites) for volume data. If no volume data were found in any of the available sources,
EPA estimated volumes based on the type of contamination at the site. The average volumes per site
for each contaminant type were calculated from volume-data in RODs signed from 1982 to 1991.
Data for this approach were available in Appendix A of the TIO report, and Exhibit 1-2 summarizes
these data. The average waste volumes presented in this exhibit were calculated using all RODs with
waste volume data from 1982 to 1991, except statistical outliers. These data include volumes, that
were managed on site or in situ.
' For sites with no contaminant data, the average volume assigned was the average volume for
all contaminant types, reflecting their frequency of occurrence. This approach was also used for all
-NPL sites without RODs.
-------
EXHIBIT 1-2
> Soil, Sediment, and Sludge Based on Contaminant Types
Contaminant Type
Metals
Volatile Organic Compounds (VOCs)
Semi-volatile Organic Compounds (SVOCs)
VOCs and Metals
SVOCs and Metals
VOCs and SVOCs
VOCs, SVOCs, and Metals '
Others
Average Volume Per Site
(tons)
75,400
13,700
27,600
67,000
49,200.
23,500
: 102,400 '
55,300
Source: Exhibit A-5 in, EPA 1993c, p. 121.
1.33 Calculate Year-by-year Waste Generation for Each State ' .
The timing of waste generation was based on actual remedial action schedules if available
(e.g., from the CRES database). If the actual or previously estimated dates.of remediation were not
available, EPA estimated the years of waste generation using average event durations calculated from
actual remedial action schedules in the CRES database. The estimated average durations of the
remediation activities are as follows: " "
Five years after a site is listed on the NPL its ROD is signed; , '
Three years after a ROD is .signed the remedial action begins; and
Remedial action lasts for two years. ,
Based on these results, -EPA identified the years in which each site with an incomplete
schedule is expected to generate waste. For example, sites added to the NPL in 1987 that lack a
cleanup schedule are expected to generate waste in 1996 and 1997 because the average duration
period from the NPL date to the ROD date is 5 years (i.e., 1992), the average duration of the
remedial design period between the ROD signed date and the beginning of the remedial action is 3
years (i.e., completed in 1995) and the average RA lasts 2 years (i.e., 1996 and 1997).
For calculating annual waste volumes, EPA assumed that waste is'generated at a constant rate
over the two-year remedial action, based on CRES data, as described above. Therefore, 50 percent
of the total waste volumes would be generated in each year of the remedial action.
10
-------
Staterby-State waste volume estimates for each year were made by adding waste volumes for
all sites on a State-specific basis. The locations of all sites are known from the RODs and other -
sources identified in Section 1.3.1.
/
1.3.4 Determine the Proportion of Waste Managed Off Site '
For each'site identified in Step 1 (Section 1.3.1), EPA estimated the proportion and volume of
the waste that is managed in off-site RCRA hazardous waste treatment and disposal facilities. For
sites with RODs, EPA used remedial descriptions in the RODs to determine management location.
For sites without RODs and sites whose ROD provided no information on waste management
location, EPA estimated the proportion managed off site based on an analysis of 1992 and 1993
RODs. This analysis, which was conducted in support of economic analysis for the forthcoming
Hazardous-Waste Identification Rule (HWIR), determined that approximately four percent of the.soil
and sediment (by volume) excavated at Superfund NPL sites with RODs signed in 1992 and 1993 will
be managed off site (ICF. Incorporated 1993b): EPA based this proportion on RODs signed in 1992
and 1993, rather than on a larger set of RODs (e.g., 1982 to-1993), because RODs signed in recent
years provide better information on'current remedial action technologies. '
EPA's methodology also includes an adjustment to account for the use of Corrective Action
Management Units (CAMUs) at Superfund remedial action sites. CAMUs create strong incentives for
on-site waste management and may significantly reduce the demand for off-site Subtitle C
management from Superfund remedial actions. A more detailed description'of CAMUs is presented
in Section 3.2.
Although the CAMU concept was developed under the corrective action program, it will
affect volumes of waste from Superfund remediations as well. The initial CAMU concept in the
proposed Subpart S rule was based in part on the existing Superfund area of contamination (AOCs)
concept (the proposed rule was issued.June 1990, 55 Federal Register 30738). The CAMU, as
finalized February 16, 1993 (58 Federal Register 8658), is broader than the AOC concept because it
allows consolidation of AOCs themselves into a single area for the purpose of remediation at
Superfund sites without triggering RCRA land disposal restrictions (LDRs). CAMUs may be used at
Superfund sites, because the CAMU rule is an applicable or relevant and appropriate requirement
(ARAR) for Superfund decisions. To adjust the estimated Superfund one-time waste volumes for
CAMUs, EPA multiplied off-site waste volume estimates by a factor of 0.43, which was derived from
background data for the CAMU rule RIA (EPA 1993d).
i . '
1.3.5 Allocate Off-site Waste to CAP Management Categories
EPA allocated waste to CAP Management Categories based on contaminant data contained in
the RODs. Contaminant types at sites were classified as containing metals only, organics only, or
both. For sites without available contaminant data from the RODs, EPA estimated the proportion of
wastes in CAP Management Categories based on the number of Superfund.sites contaminated with
metals, organics, or both from the TIO report:
27 percent contaminated, with organic constituents, only;
11 percent contaminated with metals only; and
62 percent contaminated with both. s
11
-------
EPA multiplied the waste volume at each site without contaminant data by these percentages to
calculate waste quantities in each contaminant ciass.
To use these contaminant classifications to allocate wastes to CAP Management Categories,
EPA assumed that: , . '
Wastes contaminated with organic constituents are treated by Incineration -
Sludge/Solids; . . . .
Wastes contaminated with metals are treated by Stabilization/Chemical
Fixation; and
Wastes contaminated with 'both contaminant types are treated by in both'
j categories.
To calculate the volume of waste residuals from incineration and stabilization disposed in
landfills, EPA assumed that all residuals from the treatment of listed hazardous wastes are managed in
RCRA Subtitle'C landfills unless the Agency received information otherwise from the states. This
assumption is based on the derived-from rule (40 CFR 261.3(c)(2)(i), which requires RCRA Subtitle
C management of any solid wastes generated from the treatment, storage or disposal of a listed
hazardous wastes unless and until the waste is delisted. EPA also assumed that residuals of treated
characteristic wastes do not exhibit a characteristic of hazardous waste, and are'managed in RCRA
Subtitle D landfills. EPA used 1991 BRS data for Superfund remedial action wastes (BRS Form GM,
Source Code A61)sto calculate proportions of remedial action wastes are that are listed hazardous
wastes'or mixtures of listed and characteristic hazardous wastes (ICF Incorporated 1993a):
67 percent of one-time wastes contaminated with only organics were listed;
. 10 percent of one-time wastes contaminated with only metals were listed; and
. 96 percent of one-time wastes contaminated with both were listed.
The treatment residuals for these w.astes are assumed to be managed in RCRA Subtitle C.
landfills. A residuals factor of 1.5 is multiplied by the waste volume stabilized to account for the
overall increase in volume resulting from the remedy. Incineration is assumed not to change waste
volumes (i.e., residuals factor of 1) because Superfund wastes are primarily soils which are not
significantly reduced in volume by incineration. These residuals factors are based on volume changes
for treated soils reported in the literature (Peretz, 1992).
12
-------
1.4 REFERENCES
DiGregorio, R.C. (Ed.) 1992. 92193 Guide to Superfund Sites, Pasha Publication, Inc.
Doty, C. B., A. G. Crotwell, and C. C. Travis, 1991. Cost Growth for Treatment Technologies at
NPL Sites. Prepared by the Oak Ridge National Laboratory, Oak .Ridge, TN. April 1991.
ORNL/TM-11849.
English, M. R., 1991. The Superfund Process: Site-Level Experience, Waste Management Research
and Education Institute, University of Tennessee, Knoxville, December.EPA 1993a. "List of
Superfund Sites Expected to Generate One-time Wastes Between 1992 and 1999," November.
EPA 1993b. Guidance for Capacity Assurance'Planning, OSWER Directive Number 9010.02, May.
EPA 1993c. Cleaning Up the Nation's Waste Sites: Markets and Technology Trends, Office of Solid
Waste and Emergency Response, Technology Innovation Office, EPA Publication 542-R-92-012,
-April 1993.
j
EPA 1993d. Regulatory Impact Analysis for the Final Rulemaking on Corrective Action Management
Units and Temporary Units. .Office of Solid Wastes. January 11, 1993.
EPA 1992a. ROD Annual Report, FY 1991: Volume 1, EPA Publication 9355.6-05-1, April.
EPA 1992b. ROD Annual Report, FY 1991: Volume 2, EPA .Publication 9355.6-05-1, April.
EPA 1992c. "Background Information: National Priorities List, Final and Proposed Rules," U.S.
EPA Intermittent Bulletin Vol. 2, No. 2, EPA Publication 9320.7-041, October.
EPA 1991, ROD Annual Report FY 1990, EPA/540/8-91/067, July. ' ' .
ICF Incorporated 1993a. "Analysis of 1991 BRS Data on the Management of Superfund Remedial
Action Waste." Memorandum to Bill Sproat, Radian, from John Trever and Mike Berg, ICF
Incorporated. November 23.
ICF Incorporated 1993b. "Approach for Estimating On-Site and Off-Site Percentages of Media '
Eligible for Exemption Under HWIP," draft memorandum to Lyn Luben, EPA/OSW/CABD, from
Dena Gittelman, John Trever, Josh Cleland, and Randy Freed, ICF Incorporated, September 24,
1993. " .
Peretz, J., 1992. "Basis and References for the Treatment Factors Used in the HAZRAM Model for
Projecting Secondary Treatment Demand," December 1992, in Hazardous Waste Residuals
Assessment Model, Undated.
Probst, K. N., Portney, P. R., 1992. Assigning Liability for Superfund Cleanups, An Analysis of
Policy Options, Resources for the Future Report, June.
Tonn, B./H. L. Hwang, S. Elliott, J Peretz, R. Bohm, B, Jendrucko 1993. Methodologies for
Estimating One-time Hazardous Waste Generation for Capacity Assurance Planning, Oak Ridge
National Laboratory and the University of Tennessee-Knoxville, October.
\
13 '
-------
2. SUPERFUND REMOVAL ACTIONS
2.1 INTRODUCTION
This'chapter presents the methodology used to estimate the total amount of one-time.
hazardous wastes generated from Superfund removal actions for the years 1993, 1999, and 2013 on a
State-by-State basis. Generally, these are short-term actions taken to respond promptly to an urgent
clean-up heed.' Removal actions can include cleanup or removal of released substances from the
.environment; actions in response to the threat of a release; actions that may be necessary to monitor,
assess,, and evaluate the release or threat; disposal of removed material; or other'actions needed to
prevent, minimize, or mitigate damage to public health, or welfare, or to the environment. Only
diose hazardous wastes requiring off-site commercial treatment or disposal are included in this study.
2.2 DATA SOURCES
EPA used three data sources for estimating one-time waste volumes from CERCLA removal
actions: ' .
(1) Superfund Emergency Response Actions, A Summary of Federally Funded
Removals, Sixth Annual Report-Fiscal Year 1991. United States
Environmental.Protection Agency, Office of Solid Waste and Emergency
Response, Washington, DC..EPA/540-R-92-020, PB92-963421, October '
, 1992; .
(2) 1991 Biennial Report data; and . .
(3). Cleaning Up the Nations Waste Sites: Markets and Technology Trends.
United States Environmental Protection Agency, Office of Solid Waste and .
Emergency Response, Technology Innovation Office, Washington, DC,
EPA542-R-92-012, April 1993. (TIO report)
The Superfund Emergency Response Actions Annual Report for 1991 provided brief
descriptions of all removal (i.e., emergency.response) actions completed in 1991 and summary data
for all removal actions from 1980 to 1991, including the number of removals in each State. The
removal or emergency response actions include a wide variety of activities such as supplying
alternative drinking water supplies, removing wastes from the site, and stabilizing wastes on site to
prevent releases, prior to planned remedial actions. EPA used these data to project the number of
future removals in each State and to identify a typical waste volume per site. The 1991 Biennial
Reports and the TIO report-provided data on the allocation of wastes to CAP Management Categories.
2.3 METHODOLOGY .
The methodology for estimating one-time waste volumes from CERCLA removal actions uses
historical data to project State-by-state volumes for each year from 1992 to. 1999. EPA assumed
constant annual waste generation from 1999 to 2013.
14
-------
-. I .
213.1 Removal Actions Nationally Each Year Through 1999
-Using regression analysis, EPA projected the number of removal actions nationwide each year
from 1992 to 1999 based on the number of removals each year from 1987 to 1991. EPA chose the
years 1987 to 1991'because the years prior to 1987 include the start-up years for the Superfund
program when the annual rate of increase in the number of removal actions was much higher than it
has been in recent years., A regression analysis based on the number of removals from 1980 tO-1991
would produce unrealistically high projections of future removals. For. example, such a regression
would project 563 removals in 1999, whereas the regression based on recent trends (i.e., 1987 to
1991) projects 289 removals in 1999. Exhibit 2-1 presents the number of removal actions each year
from 1993, to 2013 based on the approaches in this step of the methodology.
2.3.2 Number of Removal Actions in Each State
To project the number of removal actions completed in each State in future years, EPA
multiplied the estimated number of removal actions nationwide (described above) by the percentage of
all past removal actions in each State. This approach assumes that each State's share of future
removal actions, will be equal to its share of completed removal actions. State-byTState percentages of
completed removals were calculated by dividing the number of removal actions completed in each
State from 1980 to 1991' by the total number of removal actions completed nationwide during the.
same period. The percentages for all States add to 100 percent. Exhibit 2-2 lists the percentages
calculated for each State. These percentages are assumed to remain constant in the future. Thus, a
State's share of the number of removals completed nationwide is expected to be the same in 1993,
1999, and 2013.
2.3.3 Annual Volume Managed Off Site
EPA estimated the volume of hazardous wastes .from CERCLA removal actions by
multiplying the projected number of removal actions in each State (calculated in the previous step) by
(1) the percentage of removals that generate wastes for o'ff-site management and (2) the average
volume of waste managed off-site at a sample of removal action sites. These two factors are
described below. . '
Percentage of Removals that Generate Hazardous Wastes Managed Off Site
-, Many removal actions generate no one-time wastes (e.g., construction of fences or berms
around contaminated areas) or wastes managed on-site only. To eliminate these removal actions from
the one-time waste projections, EPA multiplied the projected number of sites in each State by 44
percent, the portion of removal actions expected to generate waste for off-site management. This
percentage was calculated by dividing the number of 1991 removals judged to involve off-site Subtitle
C management (92) by the total number of 1991 removals described in the annual report (208).
Because many descriptions of removal actions do not clearly identify the nature of off-site l
management, the percentage reflects some assumptions, specifically:
Off-site management was RCRA Subtitle C management unless otherwise
indicated by the report or unless the waste was clearly not a RCRA hazardous
waste.
15
-------
Off-site staging of waste was counted as off-site RCRA Subtitle'C
management because the wastes will eventually be treated and/or disposed.
Exhibit 2-1
Projected Number of Removal Actions Nationwide from 1992 to 2013
Year
1992
1993
1994
1995
1996
1997
1998
1999
2000 to 2013-
Number of Removals
Nationwide
268
271
"274
277
280
283
286
289
289
Average Volume of Waste Per Removal With Off-site Management
The average waste volume per removal is calculated from 1991 BRS data. EPA retrieved
data from the 1991 BRS for wastes from CERCLA Emergency Responses (Biennial Report Form
GM, source code A62) that were managed off "site. This produced waste volume data for 17. sites
with a total volume of 5,423- tons, and an average volumeter site of 319 tons. To calculate waste
volume estimates, EPA multiplied the average volume per site (319 tons) by the State-by-State and
year-by-year estimates' of the number of removal actions generating one-time waste for off-site
management.
16
-------
Exhibit 2-2
Projected Removal Actions in 1993,. 1999, and 2013
State or Territory
Alabama
Alaska
American Samoa
Arizona3
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Guam
Hawaii
Idaho '
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Marianas1"
Maryland
Massachusetts
Number of
Removals
1980 to
1991
23
5 ,
- 7
15
14
84
'45
11 '
15
0
52
- 66
9
4
14
43
59
10
15
39
19
10
25
24
58
Percent
of All
Removals
1.33
0.29
0.41
0.87
0.81
4.87
2.61
0.64
0.87
0
3.02
3.83
' 0.52
0.23
0.81
2.49
' 3.42
0.58
0.87
2.26
1.10
Q.58
1.45
1.39
. 3.36
Projected Number of Removals
1993
3.6
0.8
Yl
2.4
2.2
13.2
7 A
. 1.7
- 2.4
0
8.2
10.4
' 1.4
0.6
2.2
6.8
9.3
" 1.6
2.4
6.1
3.0
1.6
3.9
3.8
' 9.1
1999 .
.3.9
0.8
' 1.2
2.5
2.3
14.1
7.5
1.8
2.5
0
8.7
11.1
' 1.5
0.7
2.3
7.2
9.9
1.7
2.5
6.5
3.2
1-7.
4.2
4.0
9.7
2013
3.9
0.8
1.2
2.5
2.3
. 14.1
7.5
1.8
2.5
0
8.7
11.1
1.5
0.7
2.3
7.2
9.9
1.7
2.5
6.5
3.2
1.7
4.2
4.0
9.7
17
-------
Exhibit 2-5 (continued)
Projected Removal Actions in 1993, 1999, and 2013
State or Territory
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Number of
Removals
1980 to
1991
73
13-
29 .
69
11
14
7
45
96 '
8
114
74
4
60
15
. 11
118
3
- 11
30
8
... 14
97
11
7
Percent
of All
Removals
4.23
0.75
1.68
;- 4.00
0.64
0.81
0.41
' 2.61
5.57
0.46
6.61
4.29
0.23
3.48
0.87
0.64
- 6.84
0.17
0.64
1.74
0.46
0.81
5.63 .
0.64
0.41
Projected Number of Removals
1993
11.5
2.0
.4.6
10.8. ,
1.7
2.2
1.1
7.1
15.1
1.3
17.9
11.6
0.6.
9.4
- 2.4
1.7
18.50
0.5
1.7
4.7
1.3
, 2.2
15.2
1.7
1.1'
1999
.12.2
. 2.2
4.9 ,
11.6
1.8
2.3
1.2
-7.5
. 16.1
1.3
19.1
12.4
0.7
10.1
2.5
. 1.8
19.8
0.5
1-8
5.0
' 1,3
2,3 '
16.3
1,8
1.2
2013
i
12.2
2.2
4.9.
11.6
1.8
2.3
1-2
7.5
16.1
1.3
19.1
12.4
0.7
10.1
2.5.
1.8
19.8
0.5
1.8
5.0
1.3
2.3
16.3
,1.8
1.2
18
-------
State or Territory
Virginia
Virgin Islands
Washington
West Virginia -
Wisconsin
Wyoming
Total
Number of
Removals
1980 to
1991
10
3
i9'
52
24 .
8
1,724
Percent
of All
Removals
0.58
0.17
1.10
3.02
1.39
0.46
100
Projected Number of Removals
1993
- 1.6
0.5
3.0
8.2
3.8
1.3
271
1999
1.7
0.5
3.2
8.7
4.0
1.3
289
2013
1.7
0.5
3.2
8.7 -
4.0
1.3
289
,' Includes two removals within the Navajo Nation.
b Formerly the Pacific Trust Territories (excludes Guam).
2.3.4 Allocation of Wastes to CAP Management Categories
EPA allocated waste to CAP Management Categories based on waste codes for removal action
wastes reported in the 1991 BRS. This step uses the same data that were used to determine the
average volume of waste per removal. EPA used these data to identify percentage of the waste
bearing waste codes for metals, organics, or both: ,
16 percent contaminated with organic constituents only; . -
64 percent contaminated with metals only; and
' 20 percent contaminated with both.1
To use.these data to allocate wastes to CAP management categories, EPA assumed that:
Wastes contaminated with organic constituents are managed in Incineration-
Sludge/Solids;
Wastes contaminated with metals are managed in Stabilization/Chemical
Fixation; and )
*/ Wastes contaminated with both contaminant types are managed in both
' categories. -
To calculate the volume of residuals managed in RCRA Subtitle C landfills, EPA assumed
that the following wastes are managed in Subtitle C landfills:
28 percent of all residuals from incinerating organics;
: 30 percent of all residuals from stabilizing metals; and
19 ' -
-------
. 95 percent of residuals from incineration followed by stabilization of mixed
organic-and metal wastes. . t . .
\ ,
The remaining residuals-are assumed to be managed in RCRA Subtitle D landfills. These factors are
based on analysis of waste codes and management types for removal action wastes in the 1991 BRS.
(ICF Incorporated 1993) EPA developed these portions by assuming that all treatment residuals of
characteristic wastes are managed in.Subtitle D landfills and all treatment residuals of wastes
containing listed wastes or listed and characteristic wastes are managed in Subtitle C landfills. (EPA
used a similar approach for Superfund remedial action wastes.) v ,
A residuals factor of 1.5 is multiplied by the waste volume stabilized to account for the
increase in volume resulting from the remedy. Incineration is assumed not to change waste volumes
(i.e., residuals factor of 1) because one-time wastes are dominated by soils which are not significantly
reduced in volume by incineration. These residuals factors are based on the results of a literature
review (Peretz, 1992). -
EPA multiplied the percentages of waste in CAP Management Categories and the residuals
factors by each State-by-State and year-by-year one-time waste volume estimate to determine the
capacity demands for each State in each year through 2013. .
20
-------
2.4 REFERENCES
EPA 1993. Cleaning Up the Nations Waste Sites: Markets and Technology Trends, United States
Environmental Protection Agency, Office of Solid Waste and Emergency Response, Technology
Innovation Office, Washington, DC, EPA542-R-92-012, April 1993. (TIO report)^
EPA 1992, Superfund Emergency Response Actions, A Summary of Federally Funded Removals,
Sixth Annual Report-Fiscal Year 1991. United States Environmental Protection Agency, Office of
Solid Waste and Emergency Response, Washington, DC, EPA/540-R-92-020, PB92-963421, October
1992. '
1
ICF Incorporated 1993.. "Analysis of 1991 BRS Data on the Management of Superfund Removal
Action Waste." Memorandum to Bill Sproat, Radian, from John Trever and Mike Berg, ICF
Incorporated. November 23.
Peretz, ].., 1992: "Basis and References for the Treatment Factors Used in the HAZRAM Model for
Projecting Secondary Treatment Demand," December 1992, in report entitled Hazardous Waste
Residuals Assessment Model.
21
-------
3. (RCRA CORRECTIVE ACTIONS
3.1 INTRODUCTION
To estimate the amount of hazardous waste that will require treatment and disposal capacity at
commercial hazardous waste management facilities as a result of RCRA corrective actions, EPA
v identified the universe of RCRA facilities subject to corrective action requirements, developed a '
method to estimate the extent of contamination at each facility, forecast management practices for
cleanup wastes,.and predicted the nmmg and duration of remediation. This chapter explains the step's
EPA took to obtain its state-by-state results. '
3.2 BACKGROUND AND DATA SOURCES
3.2.1 Regulatory Background of RCRA Corrective Action
Under RCRA, Congress authorized EPA to promulgate regulations addressing the problems
associated with thelmproper management of hazardous wastes. In 1984, Congress enacted the
Hazardous and Solid Waste Amendments (HSWA), which significantly expanded the requirements.
In particular, sections 3004(u) and (v) of the amended statute require corrective action for both on-site
and off-site releases to all environmental media from solid waste management units (SWMUs) at
RCRA hazardous waste treatment, storage, and disposal facilities (TSDFs). EPA codified the
corrective action mandates in its regulations at 40 CFR 264.101. EPA Regions and authorized States
(currently 18 states) are implementing the corrective action program and are expected to continue
characterizing, ranking, and remediating existing contamination at TSDFs well into the next century.
The corrective action program will also address future contamination that occurs.
\ - - , .
On February 16, 1993, EPA promulgated the CAMU/TU final rule (58 Federal Register
8658). This rule established .two new types of units that will be used to facilitate remediations under
RCRA corrective action authorities. Both tend to reduce, though not necessarily eliminate, the
volume of waste sent off site to commercial facilities. A TU is a unit that allows the owner or
operator at a facility to treat or store remediation waste, for a limited period.of time, without
complying with RCRA land disposal restrictions (LDRs) and minimum technology requirements
(MTRs). A CAMU is an area within a facility that is designated by the Regional Administrator for
the purpose of implementing corrective action remediation. A CAMU may include non-contiguous
areas of contamination. Potentially, all cleanup waste and soil generated at a facility undergoing
corrective action could be managed in a single CAMU. Alternatively, more.than one CAMU can be
used at a'facility, with remediation wastes and contaminated media,moved from .one CAMU to
another without triggering the LDRs.' Fn the absence of CAMUs, the hazardous waste that is
excavated at a' facility would have to meet land disposal restrictions treatment standards before being
, land disposed .'.'.. ;
EPA developed an approach to estimate the impact of .Corrective Action Management Units
(CAMUs) on remediation wastes shipped off-site for Subtitle C management by using data presented
in the CAMU rule and RCRA-corrective action RIA. The 43 percent factor equals the estimated
annual volume of soil triggering the LDRs-at corrective action facilities implementing the CAMU
planning builds directly on EPA's RlAs for the corrective action and CAMU/TU rules (an EPA
concept that appears in the final CAMU rule (0.47 million tons per year) divided by the estimated
annual volume of soil triggering the LDRs at corrective action facilities that would?be cleaned up
' '1
: ' . . ' '"22
-------
following the CAMU concept that appears in the proposed CAMU rule (1.1 million tons per year), j.
These soil estimates were generated by the RCRA corrective'action RIA model, which is based on
detailed site-specific data.for a stratified random sample of RCRA corrective action facilities. For
For more information, see CAMU final rule published on February 16, 1993 (58 Federal Register
8658).
3.2.2 Corrective Action and CAMU/TU RIAs .
EPA's methodology for estimating one-time hazardous waste generation for capacity assurance '
planning builds directly on EPA's RIAs for corrective action and CAMU/TU,rules (EPA 1993a and
1993b). These RIAs are available for public review.
RIA Sample Selection ' ,
EPA derived the sampling frame of 5,397 non-federal facilities from the Hazardous Waste
Data Management System (HWDMS) and the Corrective Action Reporting System (CARS) (now
superseded by the RCRA Information System (RCRIS)).1'2 Using a cluster sampling design,1 EPA
sampled the universe of non-federal facilities across three strata based on facility size and RCRA
Facility Assessment (RFA) status:3
Large facilities;
Not large facilities with RFAs completed; and
Not large facilities without RFAs.
1 '
> . Facilities in the "large" stratum were identified by EPA Regional officials as being the most
important facilities in their Region in terms of their need for remediation, based on the facility size
and extent of contamination. Facilities classified as "not large" were stratified by RFA status. RFA
status is indicative of the likelihood that corrective action will be required, because RFAs tend to be
completed sooner at facilities with serious contamination. Facilities in both"the "large""stratum and.
the "not large with RFA" stratum were sampled at a higher rate than their actual occurrence in the
universe, so that more detailed information on corrective action costs could be obtained for the RIA.
Exhibit 3-1 provides information about the 70 non-federal facilities in the sample', as well as waste
generation and management data.
1 For more information on the RIA frame and sampling strategy, see EPA 1993a.
2 The corrective action RIA also considered federal facilities, but these have not been included in
EPA's analysis of one-time capacity demand for several reasons. First, the RIA sample considered
only a small number of federal facilities (9 out of 359 identified), and consequently the RIA results
provide a limited basis for projecting year-by-year capacity demand at the State level. Second, many
types of the wastes (e.g., explosives and mixed hazardous/radioactive waste) generated at federal
facilities require types of specialized management that are outside the scope of the CAP process.
3 RFAs are the first step in the corrective action process. Subsequent steps include RCRA facility
investigations (RFIs), corrective measures studies (CMSs), and, finally, remediation.
'23
-------
Exhibit 3-1
Characterization of Sample Facilities in Corrective Action RIA
FACILTY
IDENTIFIER
FACILITY 1
FACILITY 2
FACILITY 3
FACILITY 4
FACILITY 5
FACILITY 6
FACILITY 7
FACILITY 8
FACILITY 9
FACILITY 10
FACILITY 11
FACILITY 12
FACILITY 13
FACILITY 14
FACILITY 15
FACILITY 16
FACILITY 17
FACILITY 18
FACILITY 19 .
FACILITY 20
FACILITY 21
FACILITY 22
FACILITY 23
FACILITY 24
FACILITY 25
FACILITY 26
FACILITY 27
FACILITY 28
FACILITY 29
FACILITY 30
FACILITY 31
FACILITY 32
FACILITY 33
FACILITY 34
FACILITY 35
FACILITY 36
FACILITY 37
FACILITY 38
FACILITY 39
FACILITY 40
FACILITY 41
FACILITY 42- '
FACILITY 43
FACILITY 44
FACILITY 45
FACILITY 46
FACILITY 47
FACILITY 48
FACILITY 49
FACILITY 50
FACILITY 51
FACILITY 52
FACILITY 53
FACILITY 54
FACILITY 55.
FACILITY 56
FACILITY 57
FACILITY 58
FACILITY 59
FACILITY 60
FACILITY 61
FACILITY 62
FACILITY 63
FACILITY 64
FACILITY 65
FACILITY 66
FACILITY 67
FACILITY 68
,
FACILTY
SIC
2491
2491
- 2812
- 2812
2860
2869
2879
2879
2899
2911.
2911
2911
2911
2911
3000
3339
3480
3662
3672 '
3674
3728
3760
3820
4953
4953
4953
.2491
2491
2800
2821-
2821
2834
2834
2844
2869
2911 .
2911
3069
3316
3316
3470
3482
3669 ,
3691
, 4214
4230
4953
4953
4953
4953
5169
8221 '
8221
2047
2491
2816
2860
2369
2911
3354
3489
3568
3674
3699
3827
3840
3840
' 4953
FACILITY
PERMIT
STATUS
1
1
' 1
4
1
4
0
1
1
1
- 1 .
1
2
2
2'
' 1
. '4
1
1
2
1
2
1
'1
' 1
4
1
. 1
2
2
1 '
2
1
2
0
6
2
1
1
' 1.
1
0
1
9
1
2 '
2
2
1
4
,6
2
2
4
4
0
4
2
4 '
> 0
4
6
. 1 .
INCINERATION
QUANTITY
(TONS) ,
58
0
0
0
0
206, IK
^ 0
0
2,216
300,475
' 0
699
0
0
0
0
, o
0
0
31,084
0
o
0
0
1,903
27.829
0
- 0
0
0 '
0
0
0
. 121
. 0
0
0
116
, 96
0
0
0
0
' 0
25
0
0
0
0
0
0
0
0
0
! 0
864
0
0
0
. 0
2,544.
-0
0
0
0
0
0
0
24 .
STABILIZATION
. QUANTITY^
(TONS)
0
0
0
0
0
0
67,164
0
0
159,976
0
' 0
143,807
178
0
1,510
192,888
0
37
0
. 0 .
45,759
0
0'
' 1,903
55,658
0
0
. 0
0
0
0
o -
0
10,049
149,927
15,413
. 0
101
.0
5,994
226,404
0
0
43
62,333
180,722
0
0
0
0 ' '
25
0 .
0
3,613
0
0
0
, 56,837
0 .
0
0
- o
0
0
2,098
0
0
i
LANDFILL
QUANTITY
(TONS)
0
0
/ o
0
7
0
0
0
176
11,722
3
0
0
193
0
5
6,192
0
87
0
0
0
268
120
11,417
0
0
0
0
0
137
0
-0
0
0
0
15,413
10
0
0
6,957
0
0
385
43
' 0
I 0
0
0
0
0
- 0
0
0
. 0
1,079
0
0
0
0
0
0
0
0
0
0
0
0
-------
Remedy Selection Process for RIA Sample , ,
In order to account for the complexity of the decisionmaking process when simulating the
^selection of remedies, EPA developed an approach that relied on panels of experts to select remedies
at the sample facilities. To simulate the type of interactions between EPA and those responsible for
vthe facility cleanups that occur in real-world situations, two kinds of expert panels were convened:
* Policy Panel: This panel represented the'role of the regulatory agency in
setting remedial objectives, requesting additional technical information from
the technical panels on the performance of proposed remedies, and making
final remedy selection decisions.
« Technical Panel: This panel was charged with, develop ing one or more
- ,, - technical remedies for each facility, based on guidance from the policy panel,
and estimating the costs of the remedies. Technical panels were encouraged
to develop a range of remedies, including these that would represent the
facility owner or operator's preference to propose the most cost-effective
remedies that would meet the proposed corrective action regulatory objectives.
The policy panels consisted of Regional EPA and State regulatory staff with extensive
experience in implementing the corrective action program. Each pplicy panel consisted of six
individuals, usually representing a variety of EPA Regions and States to reduce regional biases.
The technical panels consisted of national remediation experts selected for their facility-
specific remedial design experience. Each technical panel comprised individuals representing several
disciplines:
Hydrogeology; x - -
.Geology; .
Geochemistry; , -"_.
.Soil science;
Civil, chemical, or environmental engineering; and
Chemistry. )
The technical experts were identified through a competitive search across many well-
recognized remediation firms in the United States. Many of the experts had significant RCRA field
experience, Awhile most had extensive experience providing investigation and remediation support
under the Superfund program. Each technical panel consisted of six members selected to represent a
balance of key disciplines listed above. It was always critical that each pane! had one or more
hydrogeologists and one or more engineers and waste treatment experts. For the most part, the
panels divided the work on each facility along lines of technical expertise.
The remedy selection expert panel sessions were conducted over the course of eight weeks in
1991 and 1992. The process' involved the use of one policy panel and two technical panels during
each of two four-week sessions. The panels evaluated information on the extent of contamination at
59 of the 79 sample facilities (including nine federal facilities) where corrective action was projected
to be necessary. The panels did not review the remaining 20 facilities in the sample, as the Agency
determined that no further action would be necessary at these facilities because of the absence of
contamination. >' .
i '
. 25 -
-------
In the first step of the remedy selection process for a sample facility, the panel members were
presented with information characterizing the extent of contamination at each facility in the absence of
corrective action (i.e., the baseline extent of contamination). This information included overviews of
historical facility operations, waste generation activities, permitting and enforcement status, financial
condition, and SWMUs. EPA described the wastes managed in the units and the constituents of most
. concern in the various media (e.g., soil, air, surface water, and ground water). EPA determined
which constituents were of most concern based on the degree to which they exceeded action levels for
various media, and on the distance the contamination had traveled from the point of release. .When
available, the Agency preferred to use monitoring data in characterizing the extent of contamination.
For example, soil samples and ground water sampling data were available for a number of facilities
that had reached the RFI stage. A multimedia model was used to estimate the extent of contamination
when^monitoring data were not available to estimate current contamination at a facility, and to predict
future contamination. The panels were provided maps presenting the locations of SWMUs at the
facility and delineating contaminant plumes. This information was often accompanied by a short
summation of facility issues by a facilitator to expedite the panel process.
Next, the policy panel reviewed the facility data and developed remedial objectives for each
SWMU and for facility-wide environmental contamination (soils, ground water, surface water, and.
air). -In developing facility-wide objectives, the panels followed the framework of.proposed corrective
action regulations and indicated target cleanup levels that remedies would have to meet, broad source
control objectives (e.g., on-site treatment, off-site treatment, capping wastes in place), and timing
objectives. In developing these broad objectives, the policy panel identified the extent of current
exposures at the facility and made assumptions concerning the potential future use of the facility.
Following the intent of the proposed corrective action regulations, the panel assumed that those
facilities with a greater current or future exposure potential would be required to develop more
stringent remedial alternatives commensurate with the threat. The policy panel typically expressed
remedial objectives as goals rather than specific technologies.
The completed facility remedy pbjectives were then presented to the technical panel, which
developed detailed technical options for remediating the facility based on these pbjectives. In
developing remedies, the technical panels had access to a full library of reference, materials on
treatment technologies (including innovative technologies), engineering design information, .
engineering costs, and, for ground water extraction remedies, plume capture computer models. Using
these materials, they proposed technical remedies for each facility for remediating ground water,
excavating and treating soils, and remediating any other site problems requiring corrective action. . .
Where more than one remedial alternative was feasible, the technical panels presented alternatives for
consideration. Finally, the technical panel'qualitatively evaluated the performance of each remedial .
alternative and developed rough cost estimates to allow the policy panel to consider cost as a remedy
selection factor. " '
After receiving the remedial alternatives from the technical panel, the policy panel sometimes
requested that additional alternatives be evaluated, or requested minor modifications to a proposed
remedy. The technical panel developed this additional information and submitted it to the policy
panel. Based on the final information provided by. the technical panel, the policy panel selected a
final remedy for the facility. After the policy panel selected a final remedy, the technical panel
generated its final cost estimate. In the course of estimating costs, the .technical panel developed
sufficient information for EPA to estimate the volume of hazardous waste that would-be generated at
a facility. _ *
26
-------
3.3 METHODOLOGY
This section explains how EPA used the results generated by the expert panels to assign waste
generation and management characteristics to all 5,397 facilities in the universe of RCRA facilities
(i.e., the non-federal facilities identified in the RCRIS database). Furthermore, it describes EPA's
approach for determining when facilities would commence corrective action remediations and for
aggregating results to obtain State-by-State estimates for commercial demand for Subtitle C .
incineration, stabilization, and landfilling in 1993, 1999, and 2013,
3.3.1 Match Waste Generation and Management Practices At Sample and Non-Sample Facilities
EPA's first step in developing a matching process was to identify factors that would predict of '
the likelihood that corrective action will be needed at a facility and, if corrective action should be
performed, the volume of wastes likely to be generated and managed off site. The following seven
factors were considered: i. .
'(1) Number and type of solid waste management units (SWMUs). The more
SWMUs that exist at a facility, the greater are the opportunities for releases to
the environment that require corrective action. Thus, the number of SWMUs
is likely to be positively related to both the likelihood that corrective action
will be needed and the amount of off-site capacity demand.
(2) Stage in corrective action process (e.g., RFA completed). The further a
facility has progressed in the process, the more likely it is that corrective
action remediation will occur.
(3) Facility size. Large facilities are probably more likely than small facilities to
need corrective action because, on average, they contain more SWMUs and
corresponding opportunities for releases. Corrective actions at large facilities
may also tend to contaminate larger volumes of soil than small facilities
because releases may spread further (e.g., to the facility boundary) before the
cleanup begins.
- (4) Types of wastes handled at the facility. The volume of contamination is
influenced by the fate and transport characteristics of a waste. The corrective
action waste management methods (e.g., in-situ, ex-situ on site, and ex-sitii
off site) also depend on'the waste types. Thus, facilities that handle similar
wastes may tend to generate similar volumes of corrective action wastes
managed off site.
(5) Waste management practices at the facility. This factor influences the
likelihood of releases and thereby affects the likelihood that corrective action
is required (e.g., corrective action may be more likely when wastes are
managed in a surface impoundments than in storage tanks).
(6) " Facility age. Old facilities, on average, may generate greater volumes of
corrective action waste because they have had more opportunities (i.e., more
time) than new facilities for releases to the environment to occur and because
waste management practices have improved over time.
'
..--. 27
-------
(7) Soil, hydrological, and climatic conditions. These factors affect the fate and
transport.of wastes released into the environment and therefore influence the
volume of wastes that must be managed off site.
' " \ .
EPA faced 'two major limitations in establishing a set of variables that could be evaluated for
possible use in a matching scheme. First, quantitative data are available on only a limited number of
parameters for non-sample facilities. Second, the variables available for both sample and non-sample
facilities are related only indirectly to the amount of corrective action waste likely to be generated at a
RCRA facility and managed off site. EPA was able to identify four variables that were both
uniformly available for non-sample facilities and at least indirectly related to the likely capacity
demand. - ,
(1) RIA Sampling Strata. This variable is a relatively strong indicator of the
number of SWMUs and facility size. The corrective action RIA explicitly
considered strata in developing its sample set of facilities, and within each
strata the RIA shows considerable differences in the number of SWMUs (EPA
1993a). Large facilities have roughly 1,300 SWMUs on average. Not large
facilities that have completed RCRA Feasibility Assessments (RFAs) have
roughly 790 SWMUs, while not large facilities that have not completed RFAs
have roughly 180'SWMUs. Because the strata variable distinguishes not large
facilities that have or have not completed RFAs, it indicates a not large
facility's stage in the corrective action process. The relationship between
strata and stage in corrective action process has been rated as moderately
strong, however, because the sampling strata do not supply information about
the corrective action stage of large facilities.
' . i ' . p
RIA sample strata appears to be the best of the four available variables for
' matching sample and non-sample facilities, largely because the factors for
which it was rated strong or moderate number and type of SWMUs,
facility size, and stage in corrective action process are particularly good
indicators of capacity demand from remediation, relative to the other
indicators.
(2) Industry.. Industry is strongly related to die types of waste generated at a
facility because of the common chemical inputs, outputs, and processes.
While industry is an indicator of waste management practices at a facility, ,
EPA judged this relationship to be moderately strong because a wide range of
systems can be used to manage similar wastes. In addition, the type of
industry occurring at a facility tends to be somewhat correlated with its age
because facilities producing similar products tend to face similar economic and
financial environments.
i i
(3) Permit Status. A facility's likelihood of requiring corrective action can
sometimes be inferred by its permit status. For example, closing facilities
required to obtain post-closure permits are more likely to require corrective
action than closing facilities not required to obtain post-closure permits,
because such permits indicate that hazardous waste has been managed in-land-
. based units and will remain on site after closure. Permit status is also
28
-------
moderately correlated with the number and type of SWMUs at a facility; for
example, facilities "with a permit by rule are likely to have few SWMUs.. ^
. (4) Location. This factor is related to the soil, hydrological, and climatic
conditions at a facility. This relationship is rated as moderately strong
because a variety of soil, hydrological, and climactic conditions may occur
within a particular State or EPA Region.
EPA developed a matrix to organize the evaluation of these four variables. See Exhibit 3-2.
For each combination of the four variables and the seven factors predicted'to contribute to off-site
capacity demand, EPA assigned a strong; moderate, or weak rating to express the strength of the
relationship, as described above. ' * ,
Exhibit 3-2
Relative Strength of Relationship Between Potential Predictors of
Corrective Action Volumes and Variables Used in Similarity Comparisons
Predictive Factor of
Capacity Demand
Number and Type of
SWMUs
Stage in Corrective
Action Process
Facility Size
Waste Types
Waste Management
Practices
Age
Soil, Hydrologicai, and
Climactic Conditions
Overall Evaluation
Variables for Matching Sample and Non-sample Facilities '
RIA Sample
Strata
Strong
\
Moderate
Strong
Weak
Weak
Weak
;
Weak -
Strong
*
Industry
Weak
Weak
Weak
Strong
Moderate
. Moderate
Weak
Moderate
»
Permit Status
Moderate
Moderate
Weak -
Weak
Weak
Weak
Weak
Moderate to Weak
Location
(EPA Region or
State)
Weak
Weak
Weak
Weak
Weak
\
Weak
Moderate
Weak
Because strata appears to be the most relevant factor in predicting capacity demand, EPA
determined that only sample facilities belonging to the same strata as the non-sample facility should be
considered further in identifying the most appropriate sample facility for transferring waste generation
and waste management data to a non-sample facility.
Following strata in order of importance are SIC code, permit status, and location,
respectively, as shown in Exhibit 3-2. Based on these results, EPA decided that industry should be
29
-------
considered twice as important as permit status and that permit status should be considered twice as
'important as location. To implement this system, EPA started by choosing 1000 points as a
maximum value to award a sample facility when it matched the three-digit SIC code (i.e., industry) of
a non-sample facility. In turn, EPA set 500 points as the maximum.value for permit status, and 250 ,
points as the maximum value for location. Thus, the maximum total score is 1750 points. .
f ' ~ .
For each of the three factors industry, permit status, and locale used in scoring the
similarity between sample facilities and a given non-sample facility, EPA used three different fractions
of the maximum points possible for evaluating combinations of characteristics for a sample facility
and a non-sample facility: . >- . '
(1) All Points. When a sample facility and a non-sample facility had the same
value for the factor being considered, the maximum value was assigned.
If a sample facility had the same three-digit SIC code as a non-sample
facility, it was awarded 1000 points.3' .
If a sample facility had the same permit status as a non-sample.
, facility,4 500 points were awarded to the sample facility.
' If a sample facility was.in the same State asrthe non-sample facility,
250 points were awarded. . '
' (2) No Points. When a sample facility and a non-sample facility were dissimilar
with regard to the variable being considered, no points were assigned.
s '
(3) Half Plus One Points. When a sample facility and a non-sample facility had
a similar but not identical value for the variable being considered, one more
than half of the maximum number of points were assigned', so that even a
partial match for a given variable was more significant in determining a match
than any next less important variable.
For a sample facility with the same two-digit but not three-digit SIC'
code as a non-sample facility, 501 points were awarded.5
\
For a sample facility with a similar permit status as a non-sample
. facility, 251 points were awarded.
3 Because a significant number of facilities perform industrial activities that could be classified
under multiple four-digit SIC codes and each facility is assigned only one four-digit code, three-digit
codes are used to compare the industrial activities at sample and non-sample facilities.
4 For an explanation of how "same" and "similar" permit statuses were determined, see ICF
Incorporated, 1993. ' '
5 A match at the one-digit SIC level received no points because this match is insufficiently
indicative of similarity in waste management and waste management characteristics.
30 "; '
-------
For a sample facility in the same EPA Region but not the same State
as a non-sample facility, 126 points were awarded.
These results are summarized in Exhibit 3-3.
Exhibit 3-3
Decision Rules for Assigning Points to Sample Facilities
Variables Compared
Points Allocated
Sampling Strata ' ,
Only sample facilities with same stratum as a non-sample facility
are considered for further evaluation ,
Industry "
i
Same 3-digit SIC
Same 2-digit SIC, but different 3-digit SIC
1000
501
Permit Status
Same permit status
Similar permit status
500
251
Location
Same State
Same EPA Region, but different State
250
126
The matching process also included the following rules:
In cases where more than one sample facility received the same highest score
for a non:sample facility, the sample facility to be matched with the non- -
sample facility was selected randomly from among the tied sample facilities.
Over 85 percent of large facilities and over 65 percent of not large facilities
had a unique sample facility with the highest similarity score. The average
highest similarity score for the sample facilities matched to non-sample
facilities using this process is 745 out of a maximum possible score of 1750.
As stated above, all sample facilities had the same stratum as their matched
non-sample facilities. , .
* Non-sample facilities that were among the sample facilities were matched to
themselves.
31
-------
EPA,considered and rejected two other approaches for projecting one-time corrective action
waste volumes: '
(1) EPA determined that using Monte Carlo modeling to match non-sample
facilities to RIA sample facilities within each of the three strata would require
too large a level of effort and would not provide enough-flexibility to allow
for modelling assumptions to be altered.
. (2) EPA evaluated an approach that would have projected volumes by matching
waste generation data from the 1986 National Screening Survey (also known
. as the GENSUR) with waste management.data from the 1987 National Survey
of Hazardous Waste Treatment, Storage, Disposal, and Recycling Facilities
(also known as the TSDR Survey). EPA decided not to pursue this approach
because these data sources are dated in-comparison to the, sample facility
database developed for the corrective action RIA. .;
3.3.2 Simulate the Timing of Corrective Action Remediation
The timing and number of corrective action remediations within each State depends on many
variables; several of them are difficult to project into the future.' Important determinants of the pace
of corrective action implementation within a State include the EPA Regional and State strategy for
implementing the corrective action program, the number and type of facilities within the State, and
the available budget. '
EPA simulated the timing of corrective action remediations within each State by using data
developed by the Office of Solid Waste to estimate the proportions of facilities that would progress far
enough through the corrective action process to commence remediation in each of three time periods:
(1) 1992. The percentages for the first period, as shown in Exhibit 3-4 reflect
actual progress at the Regional level in implementing the corrective action
' . program (EPA 1991 and EPA 1993e). ^Specifically, these percentages are /
based on the average rate of progress from 1989 through 1992.
(2) 1993 to 2002. Based- on the pace, of remediations in the last several years,
EPA projects that by the end of 2002, roughly 20 percent of the facilities
requiring corrective action will have begun remediation (EPA I993e).
, Because some facilities started remediation prior to 1993, EPA has assumed
an annual rate of new remediations of two percent over the period. In other
words, 20 percent of all facilities will commence corrective action remediation
in this period. The Agency, did not differentiate among Regions in applying
this percentage.
(3) 2003 to 2013. Lacking other data, EPA assumed that this same rate of
remediation starts would continue through the third period, 2003 to 2013.
Thus, 22 percent of the facilities requiring corrective action will begin
remediation during this 11 year period. x
'32
-------
Exhibit 3-4
Estimated Corrective Action Remediation Starts over Time
EPA Region
I
H
III
IV
V
VI
VII
VIII
IX
X
Percentage or Percentage of
remediations remediatioos started
started in 1992 1993 - 2002
. 0.0 . ,
0.5
0.3
- . 0.3
0.3
0.7
0.4
0.9
- ; 0.1
2.4
20
-20
20
20
20-
20
20
20
20
20
Percentage of
remediations started
2003 - 2013
22
22
'22.
22
22
. - 22
22
22
-' . 22
22
In Exhibit 3-4, the figures for each Region do not sum to 100 percent for two reasons: (1)
some facilities commenced remediation prior to 1992; and (2) not all facilities needing corrective
action will commence remediation by 2013. ' '
To apply these Regional percentages to specific states, EPA used the following four-step
procedure:
(1) For each State; EPA identified the number of facilities (using the matches to
sample facilities) that will generate a demand for capacity in each CAP
Management Category (before, during, or after the 1992-2013 timeframe).
For example, a hypothetical State with 200 RCRA facilities might have 10
facilities that will create a demand for incineration, 20 for stabilization, and 5
for landfill.
(2) For each of these CAP Management Categories, EPA determined the average
amount and duration of the demand by the facilities in each State with such a
demand. Using the example above, the average demand for incineration
would be the total demand for incineration by all 10 facilities with such a
demand divided by 10. The average duration was determined in the same
manner and rounded off to a whole number of years. Most durations were
one year; a few were two years.
(3) EPA calculated the number of facilities that would create a demand for each
CAP Management Category in each projection period in each State, using the
' - ' .33 ' -
-------
Regional rates of new corrective action remediations described above. For
example, if the hypothetical State identified above is located in Region I, then:
0 percent of the 10 facilities with a demand for incineration (0
facilities) would be allocated to 1992; .
.. --. 20 percent (2 facilities) would be allocated to the period from 1993 to
2002; and ,
22 percent (2.2 rounded to 2 facilities) would be allocated to the
. period from 2003 to 2013.
(4) Finally, EPA randomly assigned starting years to each of the facilities with a ,
demand for each CAP Management Category in each projection period using a
computer-driven random number generator. For example, each of the two
facilities with a demand for incineration during 1993 to 2002 would be
randomly assigned to a starting year in that period. If the average duration of
" the demand' is two years, the facility would show a demand for incineration in
the randomly assigned year and the following year.
This procedure has several advantages.. It reduces significant year-to-year fluctuations in.
, demand by using the average demand for capacity by CAP Management Category. It also avoids the
need to predict when corrective action will start at each facility.
3.3.3 Aggregate Volumes by CAP Management Category
EPA used the methodology described above to'project the demand for incineration,
'stabilization, and landfill capacity demand in all years from 1992 through 2013. The demand for
each type of capacity in each year was summed across, facilities to determine the total demand in each
year.
These results show that the projected volume of one-time waste requiring disposal at Subtitle
C landfills is small relative to the volumes for incineration and stabilization. The expert panels that
selected remedies for sample facilities were able to specify disposal at hazardous or nonhazardous
landfills, and they often chose disposal in nonhazardous landfills, evidently because many
was'iestreams were characteristic hazardous wastes (i.e., exhibited one of the characteristics indicated
in 40 CFR 261.23 ignitability, corrosivity, reactivity, or toxicity) and required no further Subtitle
C management after decharacterization through incineration and/or stabilization.
34
-------
3.5 REFERENCES
EPA 1993a. "Corrective Action Program Accomplishments" briefing given to the Office Director by
PSPD staff. June 18. - . '-
)
- EPA 1993a. Regulatory Impact Analysis for the Final Rulemaking on Corrective Action for Solid
Waste Management Units: Proposed Methodology for Analysis, draft. Office of Solid Waste.
March.
EPA 1993b. Regulatory Impact Analysis for the Final Rulemaidng on Corrective Action Management
Units and Temporary Units. Office of Solid Waste.- January 11..
EPA 1993c. List of Non-Federal RCRA TSDFs Included in EPA's Analysis of One-Time Waste from
Facilities Undergoing Corrective Action Remediations. November.
EPA 1993d. Cleaning Up the Nation's Waste Sites: Markets and Technology Trends. Office of
Solid Waste and Emergency Response, Technology Innovation Office, Washington, DC, EPA542-R-
92-OJ2. April. ,
EPA 1993e. '"Corrective Action Program Accomplishments" briefing, given to the Office Director by
PSPD staff. June 18. : .
; EPA 1991, "RCRA Implementation" briefing, given to the Office Director by PSPD staff.
December.
ICF Incorporated 1993. "Explanation of Permit Status Categories," memorandum to Robert
Burchard, U.S. EPA, OSW/WMD, from John Trever, Reid Harvey, and Mike Berg, December 22.
Tonn, B., H. L. Hwang, S. Elliott, J. Peretz, R. Bohm, B. Jendrucko 1993. Methodologies for
Estimating One-time Hazardous Waste Generation for Capacity Assurance Planning, Oak Ridge
National Laboratory and the University of Tennessee-Knoxville, October.
35
-------
4. UNDERGROUND STORAGE TANKS CONTAINING HAZARDOUS SUBSTANCES
4.1 INTRODUCTION .'.
This chapter presents the methodology used to estimate the total amount of one-time
hazardous waste generated from cleanups of releases from underground storage tanks (USTs)
containing hazardous substances for the years 1993, 1999, and 2013. Section 4.2 describes the data
sources used to develop these one-time waste estimates.
4.2 DATA SOURCES
,EPA used three data sources for estimating one-time waste volumes from cleanups of releases
from USTs containing hazardous substances: ' '
\
(1) Underground Storage Tanks:. Resource Requirements for Corrective
Action. Donna Synstetien Bueckman, Sunita Kumar, and Milton
Russell, University of Tennessee, Knoxville, December 1991;
(2) Survey of Underground Storage Tanks for 1990 and 1991. Conducted
by CRM Associates for EPA's Office of Underground Storage Tanks;
and , . ,
(3) Chemicals Stored in USTs: Characterises and Leak Detection.
United States Environmental Protection Agency, Office of Research
and Development, Washington, DC, EPA/600/2-91/037, August 1991;
and
The report by Bueckman et al. provided a detailed description of a cost estimation method
and a simulation model developed by its authors, including the model's parameters, justification for
those parameters, and the source of data used to establish the baseline .numbers and project future
numbers of regulated petroleum and hazardous substance tanks. EPA adopted many of the model's
assumptions and parameters, such as values for release rates and changes over time in the UST
population.
The 1990 and 1991 surveys of USTs conducted by CRM Associates for the Office of
Underground Storage Tanks (the OUST report) provided data on the number, age, construction,
contents, and level of protection of USTs by State. EPA used these data in establishing the size of
the hazardous substance UST population in the base year and in projecting future numbers of tanks
over time. . . , " i ;
The Office of Research and Development's (ORD) report, Chemicals Stored in USTs:
Characteristics and Leak Detection, provided another source of data on (1) the number of USTs
containing chemicals (i.e.,.hazardous and non-hazardous substances other than petroleum) in several'
States, the percentage of those USTs that contained hazardous substances, and (2) the types of
hazardous substances'stored in these USTs. EPA used the first set of data in this report to develop a
factor for predicting the percentage of the UST population identified from the OUST report data
whose releases would generate RCRA hazardous waste. EPA also used these data to estimate the
number of hazardous substance USTs in the'States for which data were provided in that report. EPA
' ": 36 ' '
-------
used the second set of data in this report to predict the management practices for RCRA hazardous
waste generated from cleanups of such releases.
4.3 METHODOLOGY
Essentially,7 this model is based on four independent variables: (1) the number of hazardous
substance USTs, (2) the percentage of hazardous substance USTs with releases, (3) the average
volume of hazardous waste resulting from a release that is managed off-site, and (4) the allocation of
off-site waste volumes to appropriate CAP Management Categories.
4.3.1 The Number, Age, and Protection Status of Hazardous Substance USTs
Estimating one-time hazardous waste generation from UST cleanups requires data on the
number of USTs containing hazardous substances in each State broken down by tank age and
protection status. The ORD report contains data on the number of hazardous substance USTs in 14
States (EPA. 1991, p. 14, Table t) and estimates that all hazardous substance USTs comprise
approximately one percent of the total tank population. (EPA 1991, p. 4 and EPA 1991,
Appendix A). Using this assumption, the number of hazardous substance USTs was estimated for the
remaining States using data from the total number of USTs by State in the annual OUST surveys.
The tank age and protection status for the hazardous substance USTs in all States was then estimated
based on the OUST survey data.
'*
In estimating the age and protection status of USTs, the method, relying on an approach
developed by Bueckman et at., first assigns tanks to age categories and then subdivides the
categorized tanks into two groups: protected and unprotected. These steps were performed using
data on tank characteristics that were collected as part of the OUST survey. The age categories to
which tanks are.assigned are: 0-5 years, 6-10 years, 11-15 years, 16-20 years, and greater than 20
years. A tank was considered protected if it was classified as having cathodic protection, having an
interior lining, being constructed of fiberglass-reinforced plastic, or having other protection. It is
important to note that .this protection status simply indicates compliance with petroleum UST -
regulations^ and does not indicate compliance with the more stringent secondary containment
protection status required for hazardous substance USTs by 1998.* If the tanks protection status was
"none" or "unknown," then it was classified as unprotected. The effective life of a tank is assumed to
be 20 to 25 years, irrespective of protection status.
The OUST survey of USTs undertaken in 1991 is taken as an indicator of the number, age
distribution, and protection status of USTs by State at the beginning of 1992. For States not covered
by the ORD report, these " 1992" data on the number of tanks'have been compared with the "1990"
data used by Bueckman et al. as a check on consistency. In those, instances where the two data sets
were not consistent, the 1992 data were used in place of the 1990 data.7 The 1990 data were
5 According to 40 CFR 280.21, all existing tanks are required to be upgraded, or protected, to
prevent releases due to structural failure or corrosion by December 22, 1998.
1 For several states, the 1992 number of tanks reported, either in total or for a specific age
category, was significantly larger than the 1990 number of tanks. Because the 1992 data are more
current, they are assumed to be relatively "more correct" than the 1990 data.. Hence, the 1992 data
were used to adjust the 1990 input data set. ' '
37 . -
-------
replaced in whole or in part with the 1992 data for the following States: Alaska, California, Georgia,
Hawaii, Kansas, Kentucky, Louisiana, Maryland, Michigan, Nevada, New Jersey, Puerto Rico, and
Wisconsin. The assumption that one percent of USTs contains hazardous substances was then applied
evenly across the tank age and protection status categories for the States not covered by the ORD N
report to produce estimates of the number of tanks containing hazardous substances in each age-
protection status cohort for each State in the base year. The distribution of States covered'by the
ORD report was assumed to be the same as the distribution for all other States combined. The
estimated number of hazardous substance tanks in 1990 by State, age, and protection status is
' presented in Exhibit 4-1. (This exhibit does not address tanks after the projection period ending in
1999 because, as discussed in Section 4.3.2, all hazardous substance USTs must have secondary
containment protection by the end of 1998 and all such tanks are assumed to have no releases
requiring off-site waste management.)
, The tanks in each age-protection status cohort are then "aged" over the projection horizon by
five year intervals using assumptions developed by Bueckman et al for petroleum USTs. Although
Bueckman et al also developed an algorithim to age the protected tanks, it is not used in this
methodology. Protected tanks, as noted above and discussed further below, are assumed to have no
releases and therefore do not create a demand for off-site waste management. '.
In their aging, unprotected tanks may be subjected to one of four actions during the projection
period:
Remain open without upgrading (until the end of 1998, the regulatory deadline for
secondary containment protection); .
Close according to formal closure procedures without being replaced;
Be replaced with a new protected tank; or .
Add protection and thereby become a protected tank.
38
-------
Exhibit 4-1
Estimated Number of Hazardous Substance USTs
by State, Tank Age, and Tank Protection Status
State
Alabama
Alaska
American Samoa
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Guam
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky '
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Northern Marianas
Protected Tanks
0-5
years
3
, 0
0
4
3
42
3
14
0
. I
5'
10
0
0
' . 4
86
16
3
1
3
0
8
7
3
17
5
10
18
5
3
1
. 3
18
1
31
32
2
0
6-10
years
' 8
0
0
I
1
56
'1
18
0
0
5
10
0
1
0
94
23
5
0
0
0
.30
2
1.
7
15
4
20
1
1
0
1
, 13
0
19
58
1
0
11-15
years
4
0
0
0
2
54
2
17-
0
0
" . 5
' 35
0
0
2
86
19
3
1
0
0
. 56
1
1
7
10
3
9
1
0
0
1
11
0
11
80
0
0
16 - 20
years
6
0
0
i
i
- 40
1
15
0
: 0
3
1
0
0
0
7!
22
3
0
0
0
12
2
1
8
5
2
J-
l
1
0
0
9
0
13
33
0
0
> 20
years
, 2
0
0
o
0
v 333
1
50
1
0
10
2
0
0
' 4
76
, 121
3
1
3
2
82
4
1
10
22
1
16
4
1
0
0
16
0
, '24
50
0
0
Unprotected Tanks
0-5
years
l
l
0
7
3
66
4
2
:
. 2
5
3
0
0
3
63
18
1
1
4
0
0
16
4
21
1
4
18
11
1
2
3
16
- 1
62
8
3
0
6-10
years
2,
5
0
5
2
92
7
6
I
1
11
3
0
2
-, o
231
25
9
0
1
0
0
16
3
43
6
9
46
14
2
0
8
30
1
143
19
3
0
11-15
, years
2
1
0
2
5
88
12
8
1
1
13
16
0
1
5
339
26
8
3
1
0
0
10
5
. 58
7
13
28
7
1
2
7
31
1
145
- 33
1
0
16 - 20
years
3
0
0
5
1
65
6
6
0
0
12
1
0
0
1
382
31
8
0
1
0
0
19
7
67
3
6
22
9
3
0
5
28
1
190
17
0
o--
> 20
years
i
2
0
' 2
0
1858
3
' 39
8
2
43
. 1
'-'- o
1
14
631
205
8
14
19
5
4
42
8
111
27
4
' 58
24
6
2
3
- 57
1
299
43
1
0
39
-------
Exhibit 4-1 (continued)
Estimated Base Year (1990) Number of Hazardous Substance USTs
by State, Age of Tank, and Tank Protection Status
State
- . \
Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Rhode Island
South Carolina
South Dakota
Tennessee
Texas '
Utah
Vermont l
Virgin Islands
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Totals
Protected Tanks
0-5
years
21
2
1
16
5
2
15
0
6
11
3
1
0
.20
4
. 1
13
3
486
6-10
years
23
1
4
. ' 6
2
2-
26
0
1
31
[
., 2
0
5
3
1
14
1
519
11-15
years
22
, ' 1
2
6
4
2
6
0
1 .
'29
0
0
0
5
7
1
13
0
520
16 - 20
years
18
i
. i
. 3
1
1
6
0
1
11
0
2
0
2
3
1
18
0
327
> 20
years
17
1
2
6
0
1
,12
0
.3
25
1
0
0
5
12
2
16
0
943
Unprotected Tanks
0-5
years
20
3
1
35
5
. 3
6.
0
8
11
2
. ' 0
0
38
4
3
13
4
512
6- 10
years
18
9
15
. 42
4
4
26
2
8
72
5
. 1
0
58
8
4
31
3
1056
11-15
. years
34
7
11
70
12
5'
6
4
' 11
84
1
0
0
58
20
3
38
0
1245
16-20
years
34
5
9
48
3
4
8
1
, 12
32
1
1
0
32
10
5
55
2
1161
> 20
years
75
7
17
94
1
10
18
1
28
92
9
0
0
83
43
:i
58
1
4094
40
-------
After tank protection is determined, the tanks are advanced, to the next age category and regrouped '
into protected and unprotected tanks.
The assumptions used in this aging process are presented in Appendix A of the report by
Bueckman et al, and are reproduced here as Exhibit 4-2. The exhibit shows the portion of existing
protected and unprotected tanks of various ages that will be subject to one of the available actions .
during certain time periods. - For example, the exhibit shows that of all unprotected tanks age 0 to 5
years during the period from 1990 to 1994, 60 percent will remain open, 7.5 percent will close, 2.5
percent will be replaced, and 30 percent will be protected during that period. (See Bueckman et al.
report for an explanation of how these percentages were derived.)
Accepting this algorithm results in the following assumptions and constraints: tanks
containing hazardous substances will behave (perform, age and degrade) like tanks containing
petroleum, new tanks age 0-5 years are limited to replacements, "no significant growth in the demand
for USTs is anticipated over the time period covered in this estimation" (Bueckman et a/., p. 29), the
. majority (96 percent) of all closures will occur between 1990 and 1999 based on the assumption of
compliance with existing regulations (Ibid., p. 27), and 30 percent of all unprotected USTs will have
been upgraded or replaced after five years.8 (Ibid., p. 25) As noted earlier in this section, the UST
regulations require that all USTs be protected and have secondary containment by the end of 1998
and, therefore, the model assumes that no unprotected tanks wili exist after 1998.
4.3.2 The Percentage of USTs with Releases
To project the percentage of hazardous substance USTs with releases in each time period for
each State, EPA adopted the release rates and approach used by Bueckman et al for unprotected
tanks. They assumed that age will affect the probability of tank failure and that "a release can occur
from a spill, an overfill or a leak and may be above or below the ground." (p. 37)
The release rates for unprotected tanks by age cohort and protection status are shown in
Exhibit 4-3. These factors represent the Bueckman report authors' synthesis of information from a
variety of sources, with particular emphasis given'to tank testing information and cause of release
Exhibit 4-2
Assumptions Used in Aging the UST Population
(Portion of USTs)
Unprotected Tank
Age (years)
0-5
6-10
11-15
Remain Open
0.60
0.60
0.60
Close
0.075
6.075
OAO
Replace
0.025
0.025
0.15
Add Protection
0.30
0.30
0.15
s The speed with which any action is taken over time is controlled by the figures presented in
Exhibit 4-4. If new data suggest that the rate of upgrading or replacing unprotected tanks will be
greater than 30 percent during the period from 1990 through 1995, this can be reflected in the model
by changing the appropriate aging algorithm parameter.
41
-------
Unprotected Tank
Age (years)
16-20
> 20 or unknown .
Remain Open
. 0.60
0.00
Close
0.10 .
0.40
Replace
0.25
0.60
Add Protection
0.05
0.00
Unprotected USTs; 1995-1999
0-5
6-;o
11-15
16-20
>20 tit unknown
0.00
. 0.00
0.00
0.00
0.00
0.00
0.10
0.20
0.25
0.30
0.00
0.20
. 0~50
0.70 -
0.70
0.00
0.70 . .
0.30
0.05
0.00
Remain Open Active or inactive tank, registered status is "open."
Close Formal closure procedure, services of the UST are not replaced.
Replace - Services of an UST that was closed are reopened with a new tank system on the same site or elsewhere. i
Add Protection -- An unprotected UST may be upgraded or retrofitted to obtain protected status.
information.9 Weighting these release rates by the percentage of tanks found in the applicable age
status cohorts for a particular time period produces a set of weighted release factors that are then
applied to the total number of tanks in that time period in each State. The result is the number of
LJSTs within each age group in each State that have releases in the particular time period. These
numbers are then summed to yield the total number of USTs with releases in the time period.
For protected tanks, EPA used an approach to account for the added protection provided by -
secondary containment,10 which according to Bueckman etal. "may reduce.release rates to virtually
zero." (p. 104) The hazardous substance UST protection standards require secondary containment
systems that will. (1) contain released regulated substances until they are detected and removed, and
(2) prevent the release of regulated substances to the environment at any time during the operational
lives of those systems.11 Under'these conditions, the probability should be close to zero that a
release from a protected hazardous substance UST system would contaminate soil or groundwater.
All hazardous substance USTs must comply with these standards by December 22, 1998. Therefore,
for'purposes of estimating generation of hazardous waste from hazardous substance UST cleanups, the.
,9 See, for example, EPA 1987a, EPA 1987b; and the discussion in Bueckman et al., pp. 37-40;
10 In the preamble to EPA's proposed technical standards for USTs, secondary containment was
defined as "a system installed around an UST that is designed to prevent a release from migrating
beyond the secondary containment system outer wall (in the case of a double-walled tank system) or
excavation area (in the case of a liner or vault system) before the release can be detected. Such a
system may include, but is not limited to, impervious liners (both natural and synthetic), double-walls
or vaults." (52 Federal Register 12772, April 17, ,1987) ' '
11 40 CFR 280.42(b). According to 40 CFR 280.12, "UST system" is defined to include the
underground storage tank, connected underground piping, underground ancillary equipment, and any
containment system. ' ... , .
42
-------
current methodology assumes that no hazardous substance UST systems will have releases beginning
in 1999. . .
Exhibit 4-3
Release Rates for Unprotected Hazardous Substance USTs
Over Five-Year Projection Period
Age Cohort (tank age in years)
0-5.
6-10"
11-15
16-20
>20
Release Rate (% of all tanks)
0.5
0.5
. 5.0
10.0
25.0
Source: Bueckman'ef al., p. 39. -
4.3.3 Volume of Hazardous Waste Generated and Managed Off Site
Capacity assurance planning requires projecting the volume of hazardous waste (e.g.,
contaminated soil) requiring off-site (i.e., commercial) treatment and disposal. The methodology
assumes an average of 150 cubic yards of contaminated soil excavated and managed off-site per
leaking tank. This estimate relies on data from estimates for petroleum USTs. First, using, data from
the Bueckman report, a weighted average of 280 cubic yards was estimated to'be remediated and
managed off site per release site. Applying a tanks releasing-per-site conversion factor weighted by
the number of tanks in each age and protection status category and their release rates to this volume,
approximately 150 cubic yards of soil was estimated to be excavated and managed off site per release. n
12 See Bueckman et al. Using data from the EPA Computerized On Line Information System and
best engineering judgement, the Bueckman report calculated tanks-per-site conversion factors for each
age and protection status of tanks. We weighted these factors according to the baseyear (1990)
population of tanks and projected release rates to develop a tanks releasing-per-site conversion factor.
Specifically, the Agency multiplied the number of unprotected tanks in each age category in the
baseyear by their respective release rates to determine the total number of releases projected in each
category (tanks x releases/tank = releases). Then the Agency multiplied the number of-releases in
each category by the releases per contaminated site factor from Bueckman et al, added the results for
all age categories, and divided the sum by the total number of releases for all categories ((releases x
releases/site)/releases = releases/site). The resulting factor, approximately 1.9, represents the
average number of tanks releasing at a contaminated site. This number is relatively high because a
large fraction of releases are at old sites where more than one tank has had a release.
43
-------
4.3.4 Allocation of Wastes to CAP Management Categories
EPA predicted the practices that would be used for managing hazardous waste generated from
UST cleanups based on information provided in the Office of Research and Development's report.
(EPA 1991) Based on that report, excavated waste will be contaminated predominately with organic
solvents. Organic compounds (including solvents and monomers) were stored in 81 percent of the
tanks that contained hazardous substances. Of that 81 percent, 60 percent is accounted for by five
common solvents: acetone, toluene, methanol, xylene, and methyl ethyl ketone. (EPA 1991, pp. 13-
14) Using this information and the knowledge that contaminated soil would likely be contaminated
with only one constituent,13 EPA predicts the following breakdown of waste management practices:
Incineration and landfill
Incineration followed by stabilization
and landfill >,
Stabilization and landfill
80%
10%
10%
; 100%
This approach assumes that die residuals of incineration and stabilization are managed in
RCRA Subtitle C landfills. The rationale for this approach is that the vast majority of one-time
hazardous wastes from hazardous substance UST cleanups are likely to be RCRAJisted wastes,
residuals of which are hazardous wastes under the derived from rule (40 CFR 261.3(c)(2)(i)).
(Treatment residuals of characteristic wastes, on the other hand, are not hazardous wastes if they no -
longer exhibit a characteristic of a hazardous waste.) EPA believes that wastes from hazardous
substance UST cleanups are listed wastes because USTs contain commercial chemical products, which
are most likely to bear P or U listed hazardous waste codes (e.g., off-specification, discarded, or
spilled products) under the RCRA waste identification system. (EPA 1991)
< ' ' ' ' . "
EPA recognizes that these assumptions, because they do not allow for increasing use of
recovery technologies or on-site remediation technologies, may overstate future demand for RCRA
hazardous waste treatment and disposal capacity. EPA has identified a growing trend towards greater
use of on-site treatment technologies such as soil vapor extraction (SVE) and bioremediation for soil
contaminated with hazardous substances; particularly volatile organic compounds. Also, EPA's
Office of Underground Storage Tanks has been undertaking a campaign to encourage and greatly
increase the use of on-site technologies at UST cleanups wherever feasible with the specific goals of
decreasing .the costs of cleanups and reducing the amount of contaminated material that must be
disposed of off site. In fact, new technologies, primarily used on site, might substantially reduce the
need for off-site treatment in the future. . ,
13 Most UST cleanups will address a release from a single tank, where multiple tanks have leaked;
they are likely to have contained the same chemical product.
44
-------
4.4 REFERENCES
i
Bueckman et al. 1991. Underground Storage Tanks: Resource Requirements for Corrective Action.
Donna Synstelien Bueckman, Sunita Kumar, and Milton Russell. The University of Tennessee,
Knoxville, December.
EPA 1993. Cleaning Up the Nation's Waste Sites: Markets and Technology Trends: United States
Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, DC,
EPA 542-R-92-012,"April.
EPA 1991. Chemicals.Stored in USTs: Characteristics and'Leak Detection. United States
Environmental Protection Agency, Office of Research and Development, Washington, DC,
EPA/600/2-91/037, August. . '
EPA 1989. 1989 Annual Survey of Under ground Storage Tanks. United States Environmental
Protection Agency, Office of Underground Storage Tanks, Washington, DC, December.
EPA 1987a. Causes of Release from UST Systems, United States Environmental Protection Agency,
Office of Underground Storage Tanks, September 30. .
EPA 1987b. Summary of County/City Reports on Releases from Underground Storage Tanks, United
.States Environmental Protection Agency, Office of Underground Storage Tanks, February 20.
ICF 1993. Memorandum on Waste Generation and Management Factors Derived From ORD Report
Chemical USTs (EPA Contract No. 68-W3-0001, Work Assignment 22, Task 1). Prepared by Ralph
Braccio, John Trever, and Mike Berg of ICF Incorporated, and addressed to Bill Sproat of Radian
Corporation. October 28. .
Tonn, B., H.L. Hwang, S. Elliott, J. Peretz, R. Bohm, B. Jendrucko, 1993. Methodologies for
Estimating One-Time Hazardous Waste Generation for Capacity Assurance Planning, Oak Ridge
National Laboratory and the University of Tennessee-Knoxvillei October.
45
-------
5. STATE AND PRIVATE CLEANUPS
5.1 INTRODUCTION
. This chapter presents the methodology EPA used to estimate the total amount of one-time
hazardous waste generated from State and private cleanups for the years 1992 through 2013 on a
State-by-State basis. 'State and private cleanups are site remediation activities that are conducted and
overseen by State and local agencies and private firms, excluding Superfurid remedial and removal
actions, RCRA Subtitle C corrective actions, and UST cleanups.
5.2 DATA SOURCES ...
EPA used the following three data sources for estimating one-tirrie waste volumes from State
and private cleanups: '
' \.-
(1) EPA Superfurid remedial action'waste estimates from Chapter 1;
(2) EPA RCRA corrective action waste estimates from Chapter 3; ' -
(3) Preliminary EPA projections of the national volumes of contaminated media
generated annually that were prepared in support of the development .of a the
not-as-yet proposed Hazardous Waste Identification Rule (HWIR), (See ICF
!992a, ICF I992b, ICF 1992c, and ICF I993b.)
The first two sources. Chapters 1 and 3 of this report, contain EPA's one-time waste
projection methodologies and results for Superfund remedial action and RCRA corrective action waste
estimates. The State and private cleanup methodology also uses these estimates'in conjunction with a
ratio of State and private cleanup volumes relative to Superfund remedial action and RCRA corrective
action volumes that were developed from EPA analyses for HWIR. These analyses included the use
of the decision science technique of expert judgment elicitation to estimate the volumes of
contaminated media generated annually, including national-level estimates for State and private
cleanups, Superfund remedial actions, and RCRA corrective actions.14
5.3 METHODOLOGY
t » '
The methodology for estimating one-time waste volumes from State and private cleanups
consists of three steps. . .
5.3.1 Identify Ratio of Cleanup Volumes
As part of an analysis to predict the quantities of contaminated media potentially affected by
HWIR, EPA projected.the annual generation of contaminated soil from Superfund remedial actions,
14 See Spetzler, C.S. and Stael Von Holstein, C.-A.S., "Probability Encoding in Decision
Analysis.'1 Management Science, Vol. 22, No. 3.; Stael Von Holstein; C.-A.S. and'Matheson, I.E.,
A Manual for Encoding Probability Distributions, SRI International, Palo Alto, C A., 1979; and .
Morgan, M.G. and Henrion, M., Uncertainty: A Guide to Dealing with Uncertainty in Quantitative
Risk and Policy Analysis, Cambridge University Press, 1990.
46 - .
-------
RCRA corrective actions, RCRA closures, State Superfund, and voluntary cleanups. These estimates
show that the volume of waste from State and private cleanups is equal to approximately 11 percent
of the volume of waste from Superfund remedial actions and RCRA-corrective actions. The
remainder of this section describes the process that EPA used to develop the preliminary national
volume estimates from which the ratio was derived. *
EPA used a two-part process to develop the national volume estimates. First, EPA reviewed
available data sources to develop initial estimates of contaminated media volumes. Secondly, EPA
conducted structured interviews, using expert judgment elicitation and, a decision science technique,
10 revise the initial estimates. \ -
Calculate Initial Waste Volumes
. To calculate initial waste volumes for several sources of contaminated media and for each
type of cleanup, EPA developed the following key parameters:
Number of sites nationwide;
Percentage of sites with contaminated media;
Pace of remediation;
Average volume of contaminated media per site; and
Portion of-the volume excavated.
For each type'of remediation, EPA derived initial estimates for each parameter from review of
various data sources. For example:
CERCLA Remedial Actions. Estimates of the total volume of soil from
CERCLA sites were based primarily on Records of Decision (RODs) from
1989, 1990, and 1991. The total number of CERCLA sites was estimated
using the Comprehensive Environmental Response, Compensation, and
Liability Inventory System (CERCLIS). The percentage of sites with
, , contaminated soil was based on a review of ROD abstracts in the 1990 ROD
Annual Report (EPA 1991).
RCRA Corrective Actions. Estimates of the percentage of facilities with
contaminated.soil, percentage of facilities excavated, and average quantity
excavated were based on work for the,regulatory impact analysis (RJA) for the
final Subtitle C corrective action rules (ICF I992c). The value for number of
RCRA facilities was estimated from the Resource Conservation and Recovery
Inventory System (RCRIS).
(
. State Superfund and Private Cleanups. Finally, the total number of State
Superfund and voluntary cleanup sites was estimated using State andt Private
Sector Cleanups (Day, S.M. et af) and professional judgment was used to
, estimate values for the other parameters. -
47
-------
Expert Elicitation v . . , y
Expert excitation was used to revise the initial estimates on the quantities of contaminated
soil.15 For each remediation category, EPA identified individuals with expertise in the various
remediation categories. These experts were asked to review the initial estimates of contaminated
volumes and provide their own estimates and associated confidence .intervals for each of the
parameters used to construct the estimate. In those cases where the expert disagreed with the initial
estimates, EPA substituted the expert's judgments for the original figures. When the expert agreed
with or had no basis to modify the initial estimates, EPA retained the original figures.
The experts' responses were usually given in terms of subjective probability distributions.
Experts were asked to provide high, low, and mean estimates. EPA interviewers men asked the
experts to judge the percent chance that the actual number would fall above or below the estimate.
For example, an expert might estimate a low value with a 10 percent chance that-the actual number
would fall below the estimate and a high value with a 10 percent chance that the actual number would
fali ^above.the estimate. These estimates and confidence intervals were used to derive a normal
statistical distribution with a mean (i.e., a standard bell-shaped curve). When the expert provided
only a low and a high number, EPA assumed a uniform distribution between the two extreme values.
That is, the actual number was equally likely to occur at any. point between the estimates, rather than
at a mean. These estimates were then entered into Demos, a probabilistic modelling software, to
generate and mathematically manipulate the probability distributions.
Demos generated and plotted distributions of variables that depend on other probabilistic
values by taking random samples of values from each input distribution. For instance, the annual
quantity of contaminated soil from a given source (e.g., Superfund remediations) is equal to the total
contaminated soil generated by that source multiplied by the assumed pace of remediation. The total
contaminated soil generated is a probabilistic value and the assumed pace of remediation is a given
value. Demos calculated the annual quantity of contaminated soil by generating a random value from
the distribution of the total contaminated soil generated and multiplying it by the assumed pace of
remediation. By repeating this process 20 times, Demos generated a probability distribution for the
annual quantity of contaminated soil. In this way, Demos generated cumulative distributions for each
source category. . ,.
In addition, EPA's methodology includes an adjustment to account for the potential use of
Corrective Action Management Units (CAMUs) at Superfund remedial action and RCRA corrective .
action sites. For Superfund remedial-action projections, the methodology, reduces the demand for off-
site management by 43% for all States. This adjustment is based on past work for the RGRA
corrective.action RIA.
> ' ^_
. Exhibit 5-1 below presents the annual median estimates (and the corresponding percent of
total soil) for each of the source categories. The estimate for corrective action is not based on the
results of HWIR expert elicitation, because those results assumed the use of CAMUs. Instead, the"
corrective action figure of 1,700 thousand tons per year is based on EPA analysis conducted for the
,15 See ICF Incorporated 1992a and 1992b for further detail on this approach. In addition, ICF
1993b identifies the experts who were interviewed and EPA's current plans for refining the estimates,
primarily to develop five-year, instead of 20-year projections.
. ' 48 ~. . .
-------
CAMU final rule. -The results in Exhibit 5-1 are not directly comparable to the CAP projections
presented in this and other chapters for two reasons:
(1) The volumes in Exhibit 5-1 represent all excavated wastes, 'whether they, are managed
onsite or offsite. In contrast, the CAP projections address wastes managed offsite
only. * v
(2) The volumes in Exhibit 5-1 represent contaminated soil only. The CAP projections
address contaminated soil and other types of one time waste (e.g., debris).
Exhibit 5-1 . ' ;
Annual National Volumes of Contaminated Soil Generated
Projected by Preliminary HWIR Analysis
Source of Contaminated Soil
Superfund Remedial Action
RCRA Corrective Action
State Superfund
Voluntary
Total
Annual Volume
Generated
(thousands tons)
900
1700
90
190
2880
Percent of Total
31
59
3
7
100
As shown in the exhibit above, State and private cleanups comprise approximately 10 percent
of the total volume of contaminated soil.16 Using the data presented above, State and private
cleanup waste represents 11 percent (10/90 x 100) of the combined volume of Superfund remedial
action and RCRA corrective action wastes. As described below, this ratio was applied to CAP
projections for Superfund remedial actions and RCRA corrective actions to project state-by-state
volumes for State and private cleanups.
\ - - '
5.3.2 Apply Ratio to Superfund Remedial Action and RCRA Corrective Action Projections
As shown in exhibit 5-1, State and private cleanups represent 11 percent of the combined
volume of Superfund remedial action and RCRA corrective action wastes. Thus, the one-time waste
projection methodology for State and private cleanups multiplies the projected annual average volume
for Superfund remedial actions and RCRA corrective action wastes (presented in Chapters 1 and 3 of
this report) by 11 percent.
16 This methodology assumes that the relative amounts of contaminated soil at Superfund remedial
action, and State and private cleanups are the same as the relative amounts of all types of one-time
wastes (e.g., contaminated soils and debris) generated at these cleanups. This simplifying assumption
was used because data on one-time wastes other than contaminated soil are not available for state and
private cleanups. _. . ~
49
-------
This method assumes that if a State has a high (low) volume of waste managed off site from
Superfund remedial actions and RCRA corrective actions, it will have a high (low) amount of State
and private cleanup waste managed .off site. EPA believes that this assumption is reasonable for
several reasons. . \
A State with a relatively large volume of Superfund remedial action waste is
likely to have a relatively large volume of State and private cleanup waste.
A high volume of waste generated by Superfund and RCRA cleanups will tend
to be positively correlated with the presence of certain industries that generate
large volumes of hazardous waste (e.g., chemicals, manufacturing). EPA
believes that the volume of waste generated by State and private cleanups is -
also likely to be positively correlated with these same industries.
Under the Superfund and RCRA programs, States, have-some input into
decisions affecting the volume of waste managed off-site, such as the choice
of on-site or off-site remediation technologies. EPA believes that States are
likely to be consistent in such policies between their Superfund and RCRA
programs and State and private cleanup programs.
Because the methodology depends on the output of the projection approaches for two other sources of
one-time waste, the estimates for State and private cleanup waste necessarily embrace all the relevant
assumptions used in these other methodologies. .
5.3.3 Allocate Waste Volumes to CAP Management Categories '
EPA allocated State and private cleanup waste to CAP Management Categories by assuming
that this waste is managed similarly to the waste from Superfund remedial actions. This approach is
reasonable because State Superfund cleanups, for example, are often conducted at inactive or
abandoned facilities contaminated by a variety of hazardous wastes, like federal Superfund remedial
actions. ' -".-.
* ' .
Combining this assumption and the prior step in the methodology, EPA allocated State and
private cleanup waste to CAP Management Categories based on the combined Superfund remedial
action and RCRA corrective action waste volumes managed in each Category in each state over the
periods 1992 to 1999 and 2000 through 2013. For example, if a state incinerated an average of
1000 tons/year of Superfund remedial action and RCRA corrective action waste from 1992 to 1999,
EPA assumed that the state incinerated an average of 110 (1000 x .11 tons of State and private waste
over the "same period. , . .
In this calculation, EPA used state-by-state averages of the volume of Superfund remedial
action waste and RCRA corrective action waste in each CAP Management Category for the projection
periods from 1992 to 1999 and -2000 to 2013'to reduce the impact of the significant year-to-year
fluctuations in the projected volumes of Superfund and RCRA corrective actions waste in many states.
That is, annual,state and private cleanup volumes in each state were summed for each CAP
Management Category in each projection period and then divided by eight and 14 years respectively
to derive an annual combined average volume, which was then multiplied by 11 percent.
50
-------
5.4 REFERENCES
Day, S.M., E. Zeinelabdin, and A. .Whitford 1991. State and Private Sector Cleanups. University
of Tennessee, Knoxville, TN, December.
EPA 1993. Regulatory Impact Analysis for the Final RuUmaking on Corrective Action Management
Units and Temporary Units. Office of Solid Wastes. January 11. .
EPA 1991. ROD Annual Repot. tY 1990. United States Environmental Protection Agency, Office of
Solid Waste and Emergency Response, Washington, DC. EPA/540/8-91/067. July.
ICF Incorporated 1993a. "Analysis of 1991 BRS Data on the Management of Superfund Remedial
Action Waste." Memorandum to Robert Burchard, EPA/OSW/WMD, from John Trever, Nikki
Feuerstein, and Mike Berg, ICF Incorporated. November 30. '
ICF Incorporated 1993b. "Improved Contaminated Media Data - Draft." Memorandum to Lyn
Luben, EPA/OSW/CABD, from Mike Berg and Josh Cleland, ICF Incorporated. July 28.
ICF Incorporated 1992a. "Expert Elicitation Approach for Contaminated Media." Memorandum to
Lyn Luben, EPA/OSW/CABD, from Theresa Muilin, Josh Cleland, and Mike Berg, ICF
Incorporated. July 31.
ICF Incorporated 1992b. "Suggested Experts and Data to be Elicited for Contaminated Soil and
Sediments Analysis." Memorandum to Lyn Luben, EPA/OSW/CABD, from Theresa Mullin, Josh
Cleland, and Mike Berg, ICF Incorporated. August 10. .
ICF Incorporated 1992c. "Hazardous Waste Identification Rule: RIA Methodologies and Findings."
Briefing prepared by ICF Incorporated for EPA, Office of Solid Waste, Communications, Analysis,
and Budget Division, page 23. December 17.
51
-------
52
-------
I
X-
00
>+
o
S
CO
>»+-+
n
a*
».
-+^4.^4..
-
u
I
"
S
+ ! +
3
il^
li§
_*
__+__.
03
?
_>_
+ _.(._.»._.(.
C>
«
^.-4. +.
+
_+_.
f .
./
-------
a
x
S
'» -I +
4
__.
g
^
H
S
+++*»
o o
o o
=-+
~*"^i
11
+. 4. + + 4 + + +.
g
,+ i.
+ ++++ ++++
*
o
M
H
»-*
j
is
!§
g
I
,3-
M
ri
g
n
+++^+.+h .
g
M
S
t-t
ri
5-+TT
- + -+ --.+ - 4- 4-
s
§
J!
+
=*=-
-+-
^
M
bi
U
g
-------
^ -4- 4- 4- 4- + H 4- 4-
+ + 4- 4-4- + 4- + +
* +
g
+ * *
d
b.
|So
a
+_+
P iJ
3
*-5-
d"+~
4. .4. ^ 4. i 4. + + 4.
*.-f
++-^+
< M 01
3"1
=*
H > I
4- + 4-
+ H 1-
-------
s
H
H » +
=-
8
IN
t?35
++^+
+ ^+++4-
+ * +
S
.P
+++++++»+
''"z"1'
o
o
m"
0
2
H
i
u
T
I
h ^-^
"I
B:
-*»"
3
H
4
HI
-------
X
B>
ACITY
H + + * + '+ +
KC
AB
,+ __+__ + _ + _+
r»
P-
DC Q O
w 55 A
Ou E |H
Sl^
+__++__ ++
S
s
(H
J
tH
1
H « + + *-*^ + »
* * + + +
ACTION
s
ON
KC
» -I
«'
i!
§
ZE
*e-
^>:
sei
» + H -f
+ -f. 1- * « +
( » h -f^-»- + +
g
-------
«-₯ *
+ + +>+ ) + +
i^
S M SB
'* X H
= 9 y
M S < .
+-*
"^f-+*+^+
ilg
ill
+ + +-*+< *« +--T-+-£+r-*--
o
ION
+r
st
s"
g
s
+++++-
g
.M]'
M
a.
9E
0
CO
CM
§
i
U
S3
-------
Oi
g
o
g
U
H
2
^ +-^+
^-*+
g
in
13.
00
g
JT*-^
++.fr-~ +
<» "aT
o"*"5-
ON
++*
g
5
+ + _,*._ +
H + +
a
§
?='
K£f
§
*
K
g
-------
i
+.
H
*~ "rr*^*
m
lis
is -
g
H
J
IH
s_
is-
is-
+ +^+
+ +*
M
++
s
g
o
m"
K + +« 4-
0>
f
C4
-* +-
-------
»"
+ + +
ii
s
1
i^
iSo
+ » + 4- +
.+
N
-*?
_+
Q
.+ .».+ + + 4- + » -f + + +
« +
<->
+ » +
+
+ + -t +
»+-*
O
U < I
*"> S
m
4-++-
m n
2
o
-------
+ »»-I +
3
|s§
(-t ++
VI [
_+_
O I O
. I
+
0 1 0
,;-+_+
J
!^S
e * o * o
*+
O
0000
t- +» H
- s
s
-+--,--4
-+*-++-
H:
Ig-S
§
-------
1
n
N
M
« +-^-r-T*-!"* -T*-
M
O
CO
n
N
d
-------
B
+
i+«
+
-+-
d
||B-
a, S H
=> D u
Kl 5 <
+-»--^-»-+
+
H +-
o*
«>
Si
ACTION
s
+ + +
+.«.+
tx
«
-*» +
O
M
5
M
»J
^«
1
₯*-
M
s
»-t
J
s
»»»»
+
««!
SSI
s
tn p
2
m
"*"ZT
-j
H
I
-------
(D
i-l
O
ea
N
O O
-+ *f^+1 »
17* »
*-f»-I +-f+
o
LI2ATION|
^H +!( 1- -t
h +
»-^ ++-
g
n
s
1
§
g
IH
H
it
i
§
»4>
O
e o
»
m r-
-4-^*+-
§
I
IH
I
d
IH
(K
ill
S
s
£i!
«B
^i
-------
+ »-I ( ++ +
ils
O. S H
g§^
+
e o
+*++
§
HC
<. + .4. 4._,._^.__^ T
NC
s
as
B p
B S
M
J
|n|
I
a««
in
+ H f-^+-
+_
§
HC
-5"1"--
*»*-»+^-»--
"B:
KS!
STAB
d
'
ES
-------
i
o
«
h+<
>=-+=-
+ _+_.(.__.,..
2
-4. + + + + + 4. +-
h-^+«
-++_+__.,.__.,.-
-+-i-f
_^.
e o
+ + + + + + + +» + + * +T-+-T-+
|s
5-
X
e
S.
"TTt-
S
-***
g
-+-++-++-«-*=-+*-
§
S
m
i
g
t-*
H
g
3
d
M
Hoiiov
ii§
S|E
s§y
HOIIOV
i^i
-------
s
8
4- ++
m
n
s
+*»-^+ + + + ₯ + + »+
i
3
1^
9
5 g
» + ++
'++* -r+^
s
8
*
3
5
+
d-
ass
s
++»
+++ + *..^*
+
5
<
*+
«
US.
0
-*+^+-
4. J4. __+. __+. _+ .
.d
I
-------
U Q
v- as
ol p
-------
»»-
r«
S^-S^-ST
« + H +
*
m
V O O O
-+ »+-
-»+ -
=-*
+*
»-»+-
-+_+.
a
s
»»+ (
*n o o n *n
§
IM
y
g
*-*
H
*-*
3
T
i^
!§S v
i|g'
-»+ *=*
O
I
M
1
+^
s
j
M
Db
s
8
d
b
O
1
g
ACTION
iB'K'
-------
1
+ + *-+.^ + +
g
S
lie
EM M ^H
OH £ H
a u u
w 2 <
*4
CO
*+»-t-
+++++
o
5 d
M
in
tn
i^
U Q M
B. S H
MS ^
f ^-f ^+^4 +^+--+-i-4-
a
co'
§
332
^« & w
3 W D
+ + * h
S
i£i
r «
CO
g
H
+ ^+ *
CO
(M
CM*
+._-+._+_+.
+ «
U
§
-------
+-.T
o
\1~
i§2
+
m
+ -^++
o
-+^.+ +
S!
+ +7 » »
O
+ ++++
T4 -t- ^ -I -r
f « * +
"*""*
O
'Bs
!SJ
!£i
i EE
+ »++
*-t-+
f
"8
-+-
-------
B"
M 3
5 S
+++*+
f+++»-*+
0
+^+-»«» -^+
m
S
*=*=+=-+= *
m-
~ +
+₯
+
i-+ +
'**"zr*
M
S
+.++
y
g
_+_T_+_
+~
+++
e
n
-4+++-
"*
N
ifi
e
-»f-+*=»*
§
3
M
b.
B*
8-
-------
X
S"
X
II
H + +
'5
+ + + +
"as"1"*
o
K* Bf "
Sla
M S <
+ +*«»*
S
+
+ +- + -t +. .. J. 4. J. + J-_J. +. +.^H >
j§
_+._.
CO
f>
IM
CD
a
h+
g
_+..^ +-
j
*
g
I
g
5
f^
M .
d
b.
iS!
e
«
M
J
HH
b<
,-*^+-
m
-H » +-
£
3
M
09
<«u:
»ii
3
d
-------
-.(. _+_.!.,
r+
+_+_+.+*.
8
8
8
H
M
y
-+*
§
s
d
to
+. »
a
is.
O
5-+-S-
g*
o o i "o"
+ 4-
+++-
T+-S+
- -+ *-
3
I
.
51 r^
1 -H
31 «
1 .4
S-l
H I
Si
06
K
-------
++ *-
.
O
5-*-
n
M
0>
f + H ( -f +
ii§
in
-------
S'
w
IH
0
N
-+ > H 1- > +
S
-+ ^4.^+*
-^ -4- « + + *
g
*-l
H
S
i
IH
I
s
FILL
r-1 I M
f-l I
333
INC
-*~
+-5-
~"₯
g p
in m o>
wi o
S
ACTION
5
NC
T
*
++-»
.T*-*J
IOHI
«
m
-+ h»+-
( +*-
M
to
-»-+ +-
5
STAB
FILL
^B
S5f
*« ti <
-------
i
PI
i
o
J. ^
»
g
H
i
++++.
.W3
o
I-+
f^+*'
+_.+_.«. ^4,
§
> jo en
Sis
0
i
4. ^ +
n
-------
o
z
PI
Sea
-t
i 1-1 >
g?1
l-t I
VI I
i!
SgS
>H a a
m
Rao M
D G
*-4 Z II
MS
3
s
g
i
I
6
«
IH
r
6
3
i
5
t~4
t*
(M
M
bl
«
u
W
*-*
M
W
«
*
"tu
-------
H
t-l
5
2 g
O) PI
M
r-
w
s
i
f+<
3!
;H (
&
M
r
n
.»-!
p
i
n
M .
S
i-«
i
HI
"5
M
+ + *
>0
5
Ul
M
10
«
±U
g
S
i
*
g
-------
i
n
198
>'-! >
VI
i
r*
i
n
s
>fl
M
P
1
++ -f-
g
t-t
r
H-f
HI
H-t
g
a?
s!
i
. + 4-^ + 4 4- + + * +-
i
+-=+
ff
i»
.4. ^+-1^4.__ .^.+^.^..^.^._^. -^-+-
-+* +
M
Ul
ff-
N
to
s
. ,4.-4 ^4- + +-
ro
- + -^ 4^- 4-^- + -f +-^- +
H
P
Ut
i
-
§
-+ + ^+ 4-
i
M
"I
K
B)
-------
i
M
§
_
i
M
P
(B
*
fat
«^-
eo
3
to
__ ,
^
o
"**
i
M
P
*>
to
,±L.
|
M
3
H4
3
*^
M
§
fa)
s-
10
M
. -H
"
i
1
H4
P
"
10
10
0
1
3
P
. .
w
3-3
o »
O «
5 r
C9
1
&
M
r*
M
S
i
(O
H
. .
M
^J
H*
*^«
M
H*
!i
i (
M
3
>-*
H
l-i
S
>. ^~
.
.-^«
O*
i
P
i
M
^
Tlfl
.JL.
*
(O
'
1
*4
^^
P
9
4
O
ii
H
P
1
M
t^
M
H
i
*
, .
H*
o
?s
:|
M
S
tH
§
H
»-*
i
H*
o>
^4
.-^,
O
i
P
4
H4
r*
f*
o
o
1
M
P
,_,
O
O
*l
M e
SF
H
H
t
a
&
M
r*
M
S
i
.^,
/
»
^.
x*
A
a
M
M
|
t^
P
M
'^'
H*
i
Sc
S '
M
S
M
H
H4
i
*
**i
«
<^^~
*-
^>
M
0
U
I
-------
P"
Bl
g
P
-+ !-
-+J +-
is"
;n «»
Su> to
S >
H 3 fi
os v> »
r
r*
s
W
S
-++++.
i
+«
-+ :» ^-^* + + + +-
-i- + + .
O
M
000
i
+ + »r + + +
+ + ++ + *
o o o o o
§
S
» »-f
&
M
r1
M
I
o
*-t-
Ji^Ji.
3
£
+. 1-*
+ » -I
spj
§
H
i
* +
._+_
+ -t- +
.a
a-
2
-------
i
ra
H
M
i
§
*^ +
-
S
i-as
iga
o
+«
J2.'+J2.
o
++
,±.+i.
i
s
o
-»-f
+_^.^+^4.^+_ +
S
»+ +
+^.+^
+
i
HI
M
S
+ +
ts
i-t
R
~s
-------
o
i
3 M
!M$2
iss
'
vt
>
n
CO
i
M
(0
B
P
-» »-
-f + -
w
i
M
M>
1-4
e
s
-r--f
to
sc
3
s
tj
o
+ +
\
M
»
%
o
+
M
eo
SI
as
3E
i
-------
i
n
8
s!
3
8
o
+-f
1'"
i"
S
s
§
+ f+^ +
+++:+4-i+;;»-f-f^+^.+
i
o
*++
i
-------
£
1.3 g
S*
F
I
M
r-
n
C W B
n en >
§
t-
«-
s
i
I!
I
l-t
r-
M
E
i
M
Ul
.+
0"
»
*
ff.
<«J
M
*
«
M>
-a
w
o
w
* +-
K>
w
«
w
u
Ul
«
>i
F^
I
o
+-2.
M
w
i
C
I
o
^+-+i
++
i
ot
K
-------
1-1
i
, B
4--*++
s
+
-2-f-S-
Cf>
I
f*
HI
s
X
+ + > +'
+ «. +
f -f ^ + r -r * + 1- -f + K + + 4 + ' +
OJ
E
n
i
i
*+H
§
is:
U
|
r
HI
§
o
,-+i
'§
o
o
-f H 1
+**-
0>
10
u
.0
*
OP
M
U
-------
§
++ +
+ ++
10
'a> <"
w ^^^_
!-13
s
h +
s
1'
8!
g
M
s
M
g
+-=.++++.
+_+'.^_+.
+t h
s
+-f +
HO
+
w
§
+ +^ +
i+i
ACTION
I
l-l
r
M
§
M
._+JL+.
4+ +
i
pii
a
s
t-i-f-^ 4+^+
.»._+_+_.».
U* 4 'W
W 1 'W
f -f+ +
R
-------
4. r »
in
a
++-f-f
'(~
o
»-f
g
1-4
o
o
I
s
+ »: + +
ro
N
-+++-^-+
-------
133
15 >
I > m
c in I
25
M W
i
sli
i
X
i
P
8
r*
i-i
P
w
i
**
w
P
_+ +_+_+-_+-; -
OH
OH
a-
O
n
i
n
i
OHJ
i
- + * -+ -fT+ -f + +-
*-i
hH
S
M
S
P
g
P
H
S
1*4
a
-5
CO
Ul '
"++ 1-
++
++-,
_.+^.+++*.»+^+-Z.+J1.+
s
-t-^+-^-++H^+-
"£
-H + -» -r-4-» -f -f-
.4. + ^ K *. :+
N
CD
-------
j> Of
;« *>
C
s
h-l
i
i
>+.
* H
&
M
I-
a
o
O
++
W >J
-±.+^..
i
O
+.+Ji^.
+-f +
S
i
+ * 4-
JL..
w
g
1-1
Pc
ti
-4
*
i
+++
s
+
M
i+£^.
S
ng'
§lj
"Ei
jo;^^
3
§
-------
«-:*-
>-> i r
r* i «H
-iS
1 hH
:§
St O
t (-«
*£ I W
I
*A I O
O
+ +
+ -f *
* « »
s'
i!
+ 1- -f + -f
Ul
-(.++ +
i
g
++++
£
+«+
»++
M
M
xj
+ * + +-^-+ +₯
-+i
i
+*
-------
n
X4
^^
4-
oo
^*
M
1
M
r
g
I
en
I
t-l
S
' U
*
«s
M
"lO
N
e
-* 4-
M
<«
»
-44^-4 44 4-
ii
I
P
**
S
1.
en
M
*-
»
S
«
«
10
+ +
4 4 4 4
4^.4
8
H
X
I
-------
\%$
I£S
; ««>
*)
M
P
S
»-+
f
»-i
S.
i-»
_il.
-
o
i
o
*
-++-
--1-f^^ +
Ul
to
+*+>
+-(. t^+
i
ACTION
»^
M
P1
i
t»1-:+
M
*
t t t t-^-f- +---*-="i-=- -f -t-^-*
§gi
sij
ai
t 4-4"++-*-
O I O I O
i
s
-------
g
n
+-
S
++
+__+-__ +
*+-f
-+i
o
+-t-^-t-+-»..
o
-S-^i
P
»
+ + + (.+.
o
-1^4. +
O
+ +
IS1
i
r
i
+ * + +
Ill
g
H
P z
o
+++
f *+++- +
5
S
£
-------
I
s
133
I -* >
t£§
r1
£1++.=.«.-
§
g
*-=+
. ^-T^-I.-
+i-+++^
+ Jf + + +: +.
3!
+ + ,^+ 4 ++ +
o o o
i!
+ .(._.,._ +
*
p
+ + 4. +
i
iE
m
P'
f T -f
-------
i
i-
*q
M
N
5
**
*.
00
">*
w
H*
m
W
i
IH
P
fc*
S
*
\l
3
^,
*-
w
1
t-
HI
J*
O
M
*
S
!_
!*
s
N
m
***
8
B
9
n
M
i
M
i
to
10
s
^**
*i
"sj
5
£j
sj
»>
i
2
p*
fr*
£
XI
3
' N
t*
J
i
M
P
^4
s
M-
33
o *
§3
» *
ZP
ca
|
r
5
|3'
qff>
»-»
35
o
|H
i
«
9E
0
n
**
o
ro
r*.
f
n
§
M
P
M
b»
xj
Ut
"w
5
m
u
M
|
M
V*
M
M
*
_
o
9 g
i|
! §
a
NOIIOV
|
*-«
H
§
M
A
M
O
»0
sj
-
* S
M
O
H
M
s
- M
n
H-
ON
f
to
1
P
M
W
o>
KJ
OB
OJ
§
M
^*
ff*
*
s
^
0>
ACTION
1
p
HI
s
a>
CD
t0
IX
«
SUFERFUN
DffMPMTAt
O
|
1
§
1
^0
a>
OH
N
2
vD
U
ND
*
to
O
w
m
*4
1-4
i
n HI
z n
i i
1
s
1
a
- ||
HI W
§-
w n
w %
s
1
C3
' S
z
9
9
B
-K
§
5
-------
HOIiOV
'
3
sl!
M
LL!
1
ONJ
+-
ON
<
UI
e
-( » 4-
ro
>o
09
01' M
S
Ul.
^4-^-4-^-4 4
s
0
TI
1
ON
.4. +_+._+ +-
FILL
fe
N«
r»
M
S
M
o
*-+-
I
M
t-1
|
+4-4+i.
TIIJ
NO
M
E
*-i
t-
H*
S
0
-4.4+__.,.-
4.^+_+_+ +_.
.+_,
.4,.^. +^4- 4~
-++ +44-
00
si
«
.4. 4 4. 4-
(0
UI
o
o
u
w
w
w
%<
..^.4.^.^. f_i+_^.+_^+_^+-^_+^^,
K
to
~ X
"fi
t
^
2
§
5
-------
ALL
STAB
S3
tgs
E
I s P
.4.+i--
FILL
u M 10
M W (0
.±.4,^.4.^.4.^.4.^.4-
10
Ut
u
.4._4._ 4.-
8
i
4. 4.+ 44- + 4._4__4, +^+ +4- »-4- +
Ul O O> » f
.^4. ). 4,__4__. ,
NC
HO
s
4. 4-+ +
S)
U
*-^- + -^-+^+'
figg
i-i a *
sis
*?1
fi|!
sij
2 w '
+ + -f
4. 4-^ + h 4 -f
^._*
ACITY
BY
(
-------
8
K3
i$a
3».
gg'£
ON
&gs
>H S rO
si!
«?5
1
8
1
i
1
-
H
1
i
§
-+_++_ .
. __^. .4.-
O
+H
S
M
.
g
- + K
r*
r*
i
«
u
.+__^^4._
-++ +_+_
f >--*"
IM
- ^_ + i^.-
»+^"f + + ^+-
09
M
M
-------
-++
+-f«
e e e e
O
i
+'"
+__+__+__ +
o
i
ACTION
£
o
+i+
e e o o
+ +.++
0000
+__+_+_+__+_ +
SB v>
a <=
HIT)
'ii
o
*- +
kj
M O 01
< En
1 3
-------
. g
m
r;
(I)
a,
d
J. CO
< 5)
:i
S1 5
B ?
M
>
_
§
-+++-
P
s
i
1
H
IH
g
- + ^.__ + +-
.^_4«-
e
M
t"
.-4. + +_+-
.^.+^2..
.++_+.
+'
I
r"
§
i
-
R
§
£
«
Co
a
tn
-------
I
g'
s
W
O
+ + -f
i
-+ +
+ 4. ++ + +
133
> 1-1 >
I£S!
I.
+++ +
§£?
as s
U Z
f » +
3
4 ++. 1 + 4
i
* -f
+ 1--f+*
S
3 £
o
-(»»
+ +-f
i -
+»
'R
-------
1.33
i >-> >
!g3
I n (h
n
GB
S
i
8
s
s
i-»
r*
M
S
ZAT
S
i
tn
tf
8
a
i
si
2
S!
-444++4-
i-+-
t-444-=-444 44'
«
e .
10 '
a
-»-444 4-^+-
«
*>
.+^.+^+-
K>
»«*
t.-^-4^-4-=-44<
-+4-^44-
(4*
8
g
+«
o
o» i j
S . »-l
-------
n
FILL
5
Ht
r*
w
S
^*
§
.+__ 4.^+^+-^++
+_+_+..
^*
2
f+'
*_
n
M
S
l-i
o
+ +.
+++ +
to
»
+^i +-=i+^_+-i-+_1_+J±+
s
M
0>
+ K
S
..»«Ji
Si
r
M
S
H4
i
+»
§:
4 + 4_- +.^.__+++;-f_
\f>
te
*-
o
*1.
++
+^-»++
B
a
S
K
-------
g
M
i
-TIV
£
I
Mt
g
iP
§S
Ci I
Ml
-
n
HI 5 *Q
ill
p
g
-
i
n
i
P
S
g
.+4+ + » + + + -f + +-
O
++
. +_ -
-+ + +
+ +-4-
.^.+i
+ __ +__ +__+ +__ +
M
-^
O
b>
Vj
K
Ul
t
+_+_.+
nS =
H 3 "0
m
§
+ + « +
+^+ +
5 S
i -
O to
5
S
-------
5
s
"i
£*-
+«+»'
132
i<5
i > n
S
o
++'
H
r
S
K +A
s
Ht
H|
i
H
§
n
M
r-
H
i
fi
Iff!
?!
I
l-l
f"
n
1-1
r"
t*
. -
s
tn
>
m
ra
S
o
n
>
. -f + 1- -f + + + ++ * * -»~+
IO
*
-------
------- |