EPA/630/R-94/001
December 1993
REPORT ON THE TECHNICAL WORKSHOP ON
WTI INCINERATOR RISK ISSUES
Prepared by:
Eastern Research Group, Inc.
110 Hartwell Avenue
Lexington, MA 02173
EPA Contract No. 68-C3-0315
Risk Assessment Forum
U.S. Environmental Protection Agency
Washington, DC
Printed on Recycled Paper
-------
NOTICE
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use. Statements are the individual views of each workshop participant; none
of the statements in this report represent analyses or positions of the Risk Assessment Forum or the
Environmental Protection Agency (EPA).
This report was prepared by Eastern Research Group, Inc. (ERG), an EPA contractor, as
a general record of discussions during the Technical Workshop on WIT Incinerator Risk Issues. As
requested by EPA, this report captures the main points and highlights of discussions held during
plenary sessions and includes brief summaries of the work group sessions. The report is not a
complete record of all details discussed, nor does it embellish, interpret, or enlarge upon matters that
were incomplete or unclear. In particular, each of the four work group summaries was prepared at
the workshop by individual work group leaders based on their groups' discussions during the
workshop. Thus, there may be slight differences between the four groups' recommendations. ERG
did not attempt to harmonize all the recommendations.
u
-------
CONTENTS
Foreword v
SECTION ONE—OVERVIEW 1-1
General Summary 1-1
History of the Incinerator ..-. 1-3
SECTION TWO—CHAIRPERSON'S SUMMARY OF THE WORKSHOP 2-1
Dr. Eula Bingham
SECTION THREE—WORK GROUP SUMMARIES 3-1
Combustion Engineering Work Group .3-1
Dr. Barry Dellinger
Meteorology/Air Dispersion Work Group 3-13
Dr. Walter Dabberdt
Exposure Assessment Work Group 3-17
Dr. Thomas McKone
Toxicology Work Group 3-26
Dr. Mary Davis
SECTION FOUR—HIGHLIGHTS FROM COMMENTS .4-1
Peer Reviewers' Preliminary Comments 4-1
Observers' Comments , 4-4
APPENDICES A-F; EPA PREMEETING MATERIALS
APPENDIX A. CHRONOLOGY OF EVENTS REGARDING WASTE
TECHNOLOGIES INDUSTRIES
(AS OF NOVEMBER 19,1993) A-l
APPENDIX B.
APPENDIX C.
APPENDIX D.
APPENDIX E.
APPENDIX F.
REVIEWERS ; B-l
AGENDA C-l
WORK GROUP BREAK-OUT ASSIGNMENTS D-l
BACKGROUND INFORMATION FOR PEER REVIEWERS
AND CHARGE TO PEER REVIEWERS E-l
PREMEETING COMMENTS F-l
iii
-------
CONTENTS (cont.)
APPENDICES G-L; OBSERVERS AND OBSERVER MATERIALS
'age
APPENDIX G.
APPENDIX H.
APPENDIX!.
APPENDIX!.
APPENDIX K.
APPENDIX L.
FINAL LIST OF OBSERVERS G-l
OBSERVER TECHNICAL QUESTIONS H-l
OBSERVER COMMENTS 1-1
CITY OF EAST LIVERPOOL, OHIO, DEPARTMENT
OF PUBLIC HEALTH LETTER TO EPA
REGIONAL ADMINISTRATOR
J-l
ADDENDUM TO PETITION REQUESTING REVOCATION
OF TRIAL BURN PERMIT, OPPOSING ISSUANCE OF
PERMIT, REQUESTING PUBLIC HEARING, AND
REQUESTING THE PROMULGATION OF REGULATIONS
TO PROTECT PUBLIC HEALTH IN THE MATTER
OF WTI K-l
COMMENTS ON "WIT PHASE II RISK ASSESSMENT -
PROJECT PLAN," EPA ID# OHD980613541 L-l
IV
-------
FOREWORD
This report includes information and materials from a peer review workshop organized by
EPA's Risk Assessment Forum (RAF) for the Office of Solid Waste and Emergency Response and
Region 5. The meeting was held in Washington, DC, at the Holiday Inn Capitol on December 8-9,
1993. The subject of the peer review was a draft project plan prepared by EPA Region 5 for
assessing risk at an incinerator operated by Waste Technologies Industries (WTI) in East Liverpool,
Ohio. The peer review panel was convened to evaluate the project plan as the scientific foundation
for a risk assessment, which will be used in setting final permit conditions for the WTI facility.
Notice of the workshop was published in the Federal Register on November 17,1993. The
notice invited members of the public to attend the workshop as observers and provided logistical
information to enable observers to preregister. Over 100 observers attended the workshop, including
representatives from local, state, and federal government, industry, environmental groups, the press,
trade organizations, consulting firms, law firms, public interest groups, and citizens from East
Liverpool, Ohio.
In setting parameters for the meeting, EPA referred to the 1983 National Academy of
Sciences paradigm that distinguishes risk assessment from risk management, emphasizing that the
former draws on the scientific disciplines while the latter involves both the science-based risk
assessment and non-scientific considerations such as economics. In outlining the scope of the peer
review, EPA stressed that the draft project plan addresses risk assessment but not risk management
issues, and that the plan is not itself a risk assessment. EPA asked the peer reviewers to concentrate
on technical issues concerning the science, methods, expected uncertainty, and inferences and to
suggest immediate and long-term recommendations.
Prospective peer reviewers were nominated by government, environmental groups, and
industry. EPA's Council of Science Advisors, which developed EPA's peer review policy, selected
a balanced group of scientists representing the fields of toxicology, environmental fate and transport,
combustion engineering, atmospheric modeling, and exposure assessment from the pool of nominees.
Appendix B lists the 13 peer reviewers.
EPA sought comments from these scientific experts on the draft project plan for conducting
the multipathway risk assessment. The draft plan presents the proposed approach for assessing
human health risks associated with direct and indirect exposure to emissions from the WTI
incinerator. EPA will consider the expert recommendations to revise the project plan.
The workshop report is organized as follows. The report opens with a brief overview of the
workshop and history of the incinerator (Section 1), and is followed by the chairperson's summary
(Section 2) and the four work group chairs' summaries (Section 3), respectively. Finally, highlights
of peer reviewers' preliminary comments and observers' comments are provided (Section 4).
Appendices to the workshop report include EPA premeeting materials including the agenda, list of
peer reviewers, and premeeting comments (Appendices A-F) and observer materials including the
list of observers, and observer technical questions and comments (Appendices G-L).
-------
EPA will seek peer review of the draft risk assessment report at a second meeting of external
experts, currently scheduled for later in 1994.
Dorothy E. Patton, PhJD.
Executive Director and Chair
Risk Assessment Forum
VI
-------
SECTION ONE
OVERVIEW
General Summary
The workshop provided a forum for the expert peer review panel to discuss the scientific and
technical aspects of the draft project plan for the WIT Phase H risk assessment. Workshop
participants contributed useful and substantive suggestions and recommendations for improving the
risk assessment. Recommendations that can be implemented immediately and incorporated into the
WTI risk assessment as well as recommendations for future risk assessments were presented. Section
3 of this report provides summaries and recommendations prepared by the four work group
chairpersons.
Considerable discussion took place on the characteristics of the waste stream used in the trial
burn versus the waste stream incinerated during normal operations. Since the waste stream used
in the trial bum was not representative of waste materials to be routinely incinerated at the WTI
facility, the contaminant emissions modeled in the risk assessment would not be valid. Peer
reviewers suggested accepted methods for characterizing and estimating emissions. In the long term,
trial burn procedures must be designed to generate data acceptable for use in risk assessments. Peer
reviewers also emphasized the importance of characterizing and considering chemical form and
speciation in fate and transport and dose-response evaluations.
Peer reviewers noted that the discrepancy in dioxin emissions concentrations between the
trial burn and performance test were not explained adequately in the draft plan. It was suggested
that the data be reviewed to decide which data set is representative of the normal operation of the
facility and is most appropriate for use in the risk assessment. The dioxins/furans phase distribution
also needs to be reevaluated and a protocol developed to convert sampling data to an in-stack and
plume phase distribution.
1-1
-------
Moreover, procedures for estimating emissions under upset and accidental conditions are not
well developed in the draft plan; as presently structured, COMPDEP cannot simulate such events.
Although estimating the frequency of these occurrences can be accomplished using historical data
for both the WIT facility and other incinerators, estimating the extent and character of a release
during such an occurrence is more problematic. Existing methods rely on default values that will
not reflect the specific operating parameters of the WIT facility. Reasonable worst case release
scenarios should be formulated based on consideration of hypothetical events. Furthermore, fugitive
emissions also are not addressed adequately in the draft plan.
It was recommended that the COMPDEP model be modified to consider the adverse effects
associated with fumigation and calm conditions. The COMPDEP model does not adequately
simulate the complex atmospheric dynamics of windflow in the vicinity of the Ohio River valley and
the hilly terrain adjacent to East Liverpool, Ohio. Nonetheless, using representative wind and
temperature measurements will allow estimates to be developed of effective transport wind vector
at plume height. Additional local meteorological data from the WIT site and within the river valley
should be obtained and analyzed to provide a more representative estimate of wet deposition,
transport, stability, mixing height, and diffusion rates. Reviewers also pointed out that the
COMPDEP model cannot factor in ambient chemistry or volatilization processes that might result
in vapor-to-particle transformations and vice versa. Moreover, no EPA dispersion models are able
to simulate the effects of terrain-induced downwash.
In general, peer reviewers found that the draft plan provides a comprehensive approach for
evaluating human exposures to stack emissions from the WIT facility using available site-specific
data. Reviewers suggested, however, that the interactive modeling and measurements approach to
exposure assessment should be expanded to the extent possible. Reviewers were concerned that
neither the cumulative risks from existing and future exposures to other sources in the area are
addressed, nor are the regional impacts (i.e., beyond 50 miles) of the WIT facility in combination
with other combustion sources.
The Toxicology Work Group recommended that an ecological risk assessment, including
biological and environmental monitoring, be conducted in conjunction with the human health risk
1-2
-------
assessment. Since data are lacking on the ecosystem in the area of the WIT facility, however, it will
be difficult to gauge the level of assessment needed to evaluate ecological risk.
All four work groups expressed the need for more field data for validating models used to
estimate exposure and for describing uncertainties in the risk assessment quantitatively. It was also
suggested that the treatment of uncertainty and the issues of uncertainty and variability should be
integrated as much as possible into all aspects of the risk assessment. A three-tiered approach to
uncertainty analysis was proposed by the Exposure Assessment Work Group. First, the magnitude
and variation of inputs (i.e., assumptions and justifications) could be identified. Second, local or
global sensitivity analyses could be conducted to assess how model predictions are impacted by
model reliability and data precision. And, third, an uncertainty analysis using variance propagation
methods (e.g., Monte Carlo) could be used to map how the overall precision of risk estimates is tied
to variability and uncertainty associated with the models, inputs, and scenarios.
History of the Incinerator
The Waste Technologies Industries (WIT) incinerator, the subject of the peer review workshop,
is located in East Liverpool, Ohio, across the Ohio River from West Virginia and about a mile and
a half west of the Pennsylvania border. A permit to store and treat hazardous waste regulated under
Subtitle C of the Resource Conservation and Recovery Act (RCRA) was issued to the WTI facility
on June 24,1983. Because the original permit was appealed, however, it did not become effective
until January 25,1985. On November 30,1990, WTI began to construct the incinerator.
Based on intense interest in the WTI facility, EPA Region 5 initiated a risk assessment in 1991
before authorizing interim operations. Because site-specific information was unavailable, the risk
assessment was conducted using regional meteorological data and stack emissions data from other
incineration facilities. In accordance with EPA's Office of Solid Waste guidelines, only the direct
inhalation exposure pathway was assessed. This initial risk assessment, conducted by a contractor
and referred to as Phase I risk assessment activities, was completed and made available to the public
in July 1992. The risk assessment indicated that predicted exposure levels were below the level of
concern. Subsequently EPA's Office of Health and Environmental Assessment conducted an
1-3
-------
additional screening level analysis, generating preliminary risk estimates for four exposure scenarios,
each of which included indirect exposures through the food chain.
Since these initial Phase I risk assessments were completed, a full year of onsite meteorological
data as well as WIT incinerator-specific data from a trial burn and performance test have been
collected. In 1993, using these new data a multipathway risk assessment was initiated, referred to
as Phase II risk assessment activities. (Appendix A provides a complete chronology of events
regarding the WTI incinerator.)
Since EPA Region 5 first committed to performing a risk assessment for the WTI incinerator,
considerable advances and changes to EPA guidance for performing indirect exposure risk
assessments have been adopted. On December 3,1993, EPA's Science Advisory Board (SAB) met
to discuss these developments concerning EPA guidance. Preliminary results from the SAB review
are expected in early 1994. SAB recommendations for improving the general methodology will be
incorporated into the WTI risk assessment to the degree feasible given time and resource constraints.
1-4
-------
SECTION TWO
CHAIRPERSON'S SUMMARY OF THE WORKSHOP
Eula Bingham
Department of Environmental Health
University of Cincinnati
Cincinnati, Ohio
On December 8 and 9,1993, EPA convened a workshop to peer review a draft project plan for
conducting a risk assessment at the WI1 incinerator in East Liverpool, Ohio. Over 100 participants
attended the workshop.
The purpose of the workshop was to discuss the science, health and ecological, and technical
issues contained in the project plan for a risk assessment. Although many policy issues are relevant
to risk management, these were not within the scope of the workshop discussions.
Prior to the workshop, an EPA Region 5 contractor developed a draft project plan outlining
the approach for conducting a multipathway risk assessment. Peer reviewers studied the draft
project plan and provided EPA with written comments, which were compiled (see Appendix F) and
made available to all workshop participants. These preliminary comments, workshop discussions,
and observer comments were used to develop a final commentary on each of four subject areas
addressed during the workshop: process engineering, meteorology/air dispersion modeling, exposure
assessment, and toxicology (see Section 3). After modifying the project plan based on the results
of the workshop, EPA Region 5 contractors will conduct the risk assessment, which will be peer
reviewed during another meeting.
The benefit of a risk assessment is that, more than merely providing a numerical value for the
level of risk, it establishes a methodology that describes the potential health consequences for a
human population that might be impacted by, for instance, incineration technology. A thorough risk
2-1
-------
assessment must follow a methodology that is more than routine; rather, the analysis of the
particular situation (in this instance an incinerator) must be coupled with scientific rigor. For many
reasons, risk assessors cannot yet perform perfect risk assessments and we need to acknowledge that
fact when we perform and report the results. Thus the information and comments resulting from
the workshop should provide useful information to EPA and the contractor for improving the risk
assessment process for the WIT incinerator.
An important component of the peer review workshop was examination of the adequacy of the
plan and the quality of data. The review attempts to determine whether there are holes in the plan.
The quality of data regarding exposure, meteorology, and combustion will determine the quality of
the risk assessment. The adequacy of data on adverse health effects related to the various chemicals
in WIT incinerator emissions also is a critical consideration.
Several general points emerged from the peer reviewers' preliminary comments:
* Certain areas of the project plan are vague.
* Toxicologic endpoints may need farther consideration. Are they appropriate? Were they
selected carefully? Are they based on solid science? Will the reference doses (RfDs)
and other toxicity values used in the risk assessment merely be incorporated into
equations without consideration of their derivation?
* Ecological risks need to be assessed.
» The adequacy of the identification and analysis of chemicals of concern in the emissions
should be confirmed.
* An explanation should be included of why exposures from other incinerators have not
been fully described or attempts to access their data have not been adequately described.
* Accidental releases at the site, transportation to and from the site, and the health of
workers and the community may need further consideration.
Human health and ecological risk assessments should not be performed separately. The results
of an ecological risk assessment need to be correlated with the results of the human health risk
assessment. In this way, the assessment can take into account what ecologists call the "web of life."
2-2
-------
Indeed, the community where the WIT incinerator is located exemplifies the interactions implied
by this term.
The peer reviewers expressed heightened concern about the-level of uncertainty in all phases
of the risk assessment and discussed how this uncertainty should be addressed. Two other general
recommendations resulted from discussions in the four work groups:
The degree of uncertainty with regard to the risk assessment for such a complex
operation as the incineration of hazardous waste in this community must be expressed
along with the values that are developed.
Changes must be delineated concerning when and how technical data used in the various
components of the risk assessment are collected.
The review, which began as an exercise in critiquing a document—a task most peer reviewers
perform routinely—took on added significance as discussions turned to the human dimensions,
involving a community and children in a school very close to a hazardous waste incinerator.
2-3
-------
-------
SECTION THREE
WORK GROUP SUMMARIES
Combustion Engineering Work Group
Bany Dellinger, Chair
Department of Environmental Science & Engineering
University of Dayton
Dayton, OH
Elmar Altwicker
Department of Chemical
Engineering
Rensselaer Polytechnic
Institute
Troy, NY
Randy Seeker
Energy and Environmental
Research Corporation
Irvine, CA
The charge of this work group was to review the WIT risk assessment plan and assess the
proposed approach to the development of an emissions profile for input into the risk assessment.
The group addressed three general classes of emissions: general organic combustion by-products,
dioxins and furans, and metals. The group also addressed emissions during incinerator upsets and
from accidental releases.
In the group's review, every effort was made to incorporate the specific expertise of the work
group members and resource personnel, as well as comments from the observers. Many support
documents and supplementary information were made available to the group at the time of the
workshop. Because of time constraints, however, a full review and commentary on this information
could not be included in this report. In addition, some issues were raised that were discussed by the
panel and judged to be relatively unimportant. These issues were not included in the work group's
discussion because of the requested brevity of this report.
3-1
-------
Organic Emissions
1. Issues
It is well documented in the technical incineration literature that only 20 to 70 percent of the
organic emissions from hazardous waste incinerators have been chemically characterized and
quantified, even at the most extensively tested facilities. The lack of 100 percent quantification and
the potential risk incurred on the exposed population by the uncharacterized fraction is at the center
of the controversy surrounding siting of hazardous waste incinerators. To their credit, the authors
of the WIT risk assessment have made a serious attempt to account for this fraction in the WIT
emissions profile. The experimentally determined emissions data for the WIT facility appear to be
flawed in several respects, however; consequently they do not provide an adequate base from which
to estimate the emissions that have not been chemically characterized. Although there is some
technical merit to the approach developed to estimate these uncharacterized emissions, it is currently
incomplete and also flawed in at least one respect.
The trial burn conducted in March of 1993 was designed to define permit conditions for
destruction and removal efficiency (DRE) and metals limits, and was not specifically designed for
quantification of organic by-products. EPA is currently developing new guidance on how to conduct
trial bums for assessing emissions for use in risk assessments. The trial burn was not conducted on
a waste feed that is representative of the type of waste that is burned at the WIT incinerator. Only
five chlorinated materials were incorporated into the waste feed.. The chemical analysis was only
*
based on GC-MS scans and did not include specific compounds of concern for target analysis. A
number of nitrogen-containing species, including cyanides, were identified that were probably the
result of the incomplete combustion of the corn-cob product used to "simulate" the cellulose content
of hazardous waste. In addition, some chlorinated products were identified that are easily predictable
from the chlorinated compounds that were burned. The combination of a lack of a representative
waste feed and only cursory chemical analysis gives little information about the combustion
characteristics of the WIT incinerator.
To supplement the measured emissions from the WIT incinerator, the plan discusses how a
historical emissions data base for incinerators and data from the Biebesheim incinerator in Germany
3-2
-------
will be used to identify other emissions to include in the emissions profile. However, since the waste
feed composition of the WIT incinerator has not been characterized and that of the Biebesheim
facility is also not known, it is not clear how this approach can be successfully implemented.
To account for the chemically uncharacterized emissions, a "surrogate" emissions prorating
procedure is proposed. In this procedure, the estimated emissions for the chemically characterized
fraction of the emissions will be increased by a factor equal to the ratio of the "total hazardous
organic (THO) emissions" to the total characterized emissions. It is assumed in this procedure that
the mass weighted toxicity of the characterized fraction is equal to the mass weighted toxicity of the
uncharacterized fraction. Although reasonable, this assumption is by no means proven and the
actual toxicity is highly uncertain. Furthermore, the meaning of THO is not well defined. It was
apparently planned to use the THC emission minus the methane emissions as the input value for
the THO emissions. This would appear to be a significant error, since the non-methane THC value
does not include organic emissions containing more than five carbons, polynuclear aromatics (PNAs),
and other high-molecular weight compounds and THC monitors do not respond well to halogenated
hydrocarbons. Furthermore, it is difficult to see how other data (i.e., Biebesheim and historical data)
are to be incorporated into this approach.
2. Near-Term Recommendations
To address the issue that the waste feed during the trial burn was not representative of
the true feed to the incinerator, it is recommended that a waste feed chemical
composition profile be developed using the waste feed manifests for the facility during
its year of operation. The waste generators, suppliers, and chemical process experts
should be contacted in order to develop as detailed a waste feed composition profile as
possible.
Using this waste feed, combustion chemistry experts should be called upon to develop
a list of combustion by-products that may be .generated from its incineration and to
provide an estimate of the relative emission rate of these by-products. This procedure
should also include a self-consistent estimation of the DREs for the principal organic
hazardous constituents (POHCs) used in the trial burn. The actual trial burn DREs can
then be compared to the estimated DREs to "calibrate" the entire emissions estimate
and convert the relative emissions to predicted emissions rates. The total of these
predicted emissions would represent the "characterized" fraction of the stack emissions.
It is further recommended that the emissions of this characterized fraction be prorated
in a manner very similar to that proposed in the WTI risk assessment plan to address the
3-3
-------
uncharacterized fraction. The value of the THO parameter must be defined. A
reasonable, worst case assumption is that the THO is equivalent to the total of the GC
response for the VOST analysis for volatile organic compounds, the MM5 analysis for
semivolatile organic compounds (SOCs), and the total organic carbon (TOC) in the
particulate catch as determined by standard laboratory methods. The detailed analytical
reports from the trial burn conducted in March 1993 can be used for this purpose. This
method does not include methane and other low- molecular-weight, non-toxic
hydrocarbons; however, it will tend to overestimate the number of toxic organic
compounds in the organic fraction it does include. Consequently, it will yield a
conservatively high estimate of the toxic organic emissions. Considering .that the facility
is currently burning hazardous wastes, however, it is reasonable for a near-term
estimation procedure. It is recognized that other methods of estimating the THO
emissions can be envisioned. It is further recommended that to the extent possible, the
emissions should be estimated by other methods and presented in order to more clearly
identify the range of uncertainty in these calculations.
3. Longer-Term Recommendations
The trial bum procedure was not designed to provide input to the risk assessment. A more
realistic trial bum in which a more representative waste feed is used and more detailed products of
incomplete combustion (PIC) analysis is performed is highly desirable and therefore recommended
to reduce much of the uncertainty in the emissions assessment procedure. One might even consider
developing a tiered emissions estimation approach in which a largely calculational procedure is used
first to determine the potential for unacceptable risk, but also to use the more detailed trial burn
procedure as a second-tier backup approach.
It is painfully clear that little effort has been extended by the technical community to develop
procedures to provide reasonable estimates of the combustion emissions. In a short time, our work
group has identified several areas that can be readily improved. Additional thought and
consideration should be applied to these areas, and a comparison of the various emissions estimation
procedures should be performed to determine the sensitivity of the risk assessment to the procedure
employed. Such an approach will do much to identify the resulting uncertainty in the calculated risk.
3-4
-------
Dioxins/Furans (PCDD/F)
1. Issues
Quantities. The apparent state-of-the-art design of this facility should have resulted in low
emissions of these dioxins and furans. According to the March 1993 Trial Burn Results, however,
emissions of PCDD/F were relatively high (toxicity equivalents [TEQ] 20-64 ng Seconds-1, i.e.,
greater than 130 ng meters-3 PCDD/F) in spite of a two-point carbon injection system. Because
these emissions were a factor of 100 higher than that reported for typical hazardous waste
incinerators, a so-called enhanced carbon injection system was installed. During a three day
performance test of this new system in August 1993, PCDD/F stack emissions ranged from 6 to 39
ng m-3 (arithmetic average: 13). The revised, lower PCDD/F concentrations were proposed for use
in the risk assessment.
The description of the carbon injection system (location, rates) in the project plan has been
quite vague. In addition, it appears that other parameters (principally the nature of the waste
burned) changed between the trial burn and the performance test. Furthermore, it appears that the
sampling location may have been changed. Consequently, it is unclear to what extent the reduction
in PCDD/F emissions can be attributed to enhanced carbon injection. This concern is highlighted
by the observation of increased PCDD/F concentrations immediately downstream from the single-
point carbon injection system at the Biebesheim incinerator. PCDD/F emissions were eventually
reduced by a factor of 2 to 3 at the Biebesheim facility, presumably due to capture by downstream
carbon. Since it appears that PCDD/F formation is being promoted (and captured) on the carbon,
a more global environmental issue of how the contaminated carbon is reclaimed is also raised.
Phase Distribution. The phase distribution of PCDD/F is important with respect to the two
subsequent processes: atmospheric dispersion and deposition/exposure. In the work plan it has been
proposed to use the methodology of Bidleman to address phase distribution of SOCs in the
atmosphere. The approach seems to assume that no partitioning has or is occurring in the
process/stack itself. However, because the particles formed in the process may have special
characteristics (i.e., % carbon, % metals, surface area, diameter) different from ambient air aerosols,
this approach is highly questionable.
3-5
-------
The particle emissions from the WIT process are reported to be very low, approximately 0.0019
gr dscf-1; however, there has been only one measurement of the particle emissions and size
distribution. Thirteen percent of the measured particulate had a diameter greater than 1 pm (none
greater than 3 /xm) and 32 percent had a diameter less than 0.4 jura. Because so much depends on
the size distribution, one must ask: How representative is this sample? Regardless of subsequent
use of such measurements, one measurement is simply not enough to develop a theory of
partitioning and transport from the stack. In addition, there appears to be a discrepancy between
the mean particle concentrations measured by MM5 and the cascade impactor (0.0005-0.0014 and
0.0002 grscf-1, respectively), which strongly affects the total surface area for catalyzed PCDD/F
formation.
One important ramification based on this single data point is that these particles would also
be long lived, since both diffusion andimpaction removal mechanisms would be relatively ineffective
for particles of this size. In addition, the particle composition might be such that irreversible
desorption may occur rather than only physical adsorption/desorption. The linear Langmuir
isotherm method proposed (Bidleman, 1988) may not be applicable to such particles. In fact,
Pankow (1991) suggests as much by postulating a non-exchangeable fraction (of a SOC) in terms of
adsorption/desorption, which may still be extractable. At the stack temperature of 190°F,
considerable phase redistribution within the stack of SOCs to the particle surfaces would be
expected. Furthermore, since these particles would cool rapidly once emitted, and their
concentrations are expected to be at least 10 times higher than in the prevailing air, additional phase
redistribution would be expected in the plume. It may be necessary to consider the possible
displacement (more readily adsorbed species, such as chlorophenols, displacing PCDD/F), returning
compounds such as PCDD/F to the gas phase.
Formation Mechanisms. In a high-temperature slagging kiln, PCDD/F measurements at the
combustor exit would be expected to be very low. Thus, one anticipates that the stack PCDD/F was
the result of downstream formation processes; however, since only stack measurements have been
made, it is impossible to know where PCDD/F formation occurred. The Biebesheim facility probably
is not a good direct comparison, since the three downstream sampling points at that facility showed
a progressive decrease in PCDD/F concentrations. Copper and iron were not measured in the trial
bum. Because they are known catalysts for the post-combustion formation of PCDD/F, variations
3-6
-------
in their concentration may have caused the stack concentrations of PCDD/F to have been different
in the trial burn and subsequent performance test.
2. Short-Term Recommendations
Because of the apparent differences in the trial burn and performance test procedure,
it is not clear which set of PCDD/F data is most representative of the normal operation
of the facility. It is strongly recommended that the combustion conditions, waste feed
composition, and sampling/analysis procedures used in the two tests be carefully reviewed
in order to determine which data set is most appropriate for use as input to the risk
assessment. It may be concluded that neither data set is appropriate or that some type
of combination of data sets should be used.
It is not appropriate to directly use the method 23 sampling result in terms of
fronthalf/backhalf catches to determine the PCDD/F phase distribution. The stack
conditions should be reexamined, and a protocol should be developed to convert the
sampling data to an in-stack and plume phase distribution.
The possible causes for the relatively high PCDD/F emissions should be investigated. If
at all possible, post-combustion concentrations of copper and iron should be measured
to determine if abnormally high concentrations are present that may be catalyzing
PCDD/F formation.
3. Longer-Term Recommendations
The work group contends that the PCDD/F emissions have not been adequately characterized
under normal operating conditions. Additional measurements should be made as soon as realistically
possible. Because the facility can accept so many types of wastes and already has a history of
elevated PCDD/F emissions, testing both a reasonable, worst-case waste composition and a normal
composition should be considered. Technical literature and predictive models should be used to
ensure that appropriate conditions are utilized in these tests.
3-7
-------
Metals
1. Issues
The characterization of metals emissions from the WIT facility will largely rely on the data from
the March 1993 trial bum. This trial burn was designed and conducted for metals using what is
considered to be a reasonable protocol for defining permit conditions that will maintain emissions
within required levels. Consequently, in contrast to the organic emissions data, this is considered
a good source of metals emissions data for the metals that were included in the testing. However,
quantifying metals emissions and assessing characteristics that were not determined in the trial burn
remain key issues. These issues involve the following:
Defining the emission rates for metals for which there are no WIT emissions data but
which are considered important to multipathway risk assessments (i.e., copper, aluminum,
barium, nickel, selenium, silver, thallium, and zinc).
Defining not only the emission rate of the metals as defined in the trial burn but also the
chemical form (speciation), particularly for cadmium, lead, chromium, mercury, and
arsenic.
Defining the size distribution of metal-containing particles and the uniformity of the
distribution of individual metals within these particles.
No clear approach to filling these data gaps was proposed in the project plan for the WIT phase II
risk assessment.
The techniques to be used for assessment of the emissions rates for those metals not tested in
the March 1993 trial bum were not well defined in the project plan and there was some apparent
confusion about acceptable techniques for conducting such an assessment. How will other data from
other facilities be used and related to the different design, operating, and metals feed characteristics
of WTI? Mass balances are generally not suitable to quantify the stack emissions from incinerators
because the trial bum protocol is not designed to close mass balances and because the emissions
represent only a trace level and a small fraction of the total metals fed to the incinerator.
Therefore, large errors arise in attempting to quantify the small emission values by substraction of
relatively large numbers (i.e., the metals feed rate and the ash metals retention rates). For this
3-8
-------
reason, an approach for estimating these emissions in the short term was developed in this report
for consideration by the project team.
2. Short-Term Recommendations
The most appropriate short-term approach is to rely on the WTI-data quantified metals
as an indicator of how unquantified metals will behave. Since metals are not formed in
the incineration process, the key feature on any estimating procedure is to define the fate
of the metal; that is, to define where the metal in the waste stream partitions throughout
the system. Individual behavior of metals can be grouped by volatility and
physical/chemical transformations that are expected to occur in the WTI facility. Data
from other metals with similar expected behavior can be used to estimate the fate and
partitioning of these metals. This metals partitioning information, when combined with
information on the metals concentrations in the waste, can provide a reasonable estimate
of the emission rates of these metals.
Additional features of the WTI system should be examined in making these estimates.
It is important to factor in the impacts of the higher temperature kiln, which promotes
vaporization of the metals; the particulate collection device's performance versus particle
size; and the different behaviors of the various physical/chemical forms of metals in
different waste streams. In addition, with high-performance particulate collection devices
of the type that are used in the WIT facility, some of the metal emissions may arise due
to trace impurities in the scrubber materials and water. Thus, its recommended that the
feedstocks should also be examined as potential sources for metals emissions.
There is also an apparent need to quantify the chemical form emitted for several key
metals if the risk assessment is to be completed. No data currently exists on the chemical
form (speciation) of the metals except for hexavalent chromium. These data should be
examined carefully to ensure their reliability, given the recent implementation of the
chromium VT sampling and analysis protocol. For other metals of importance, some
estimate of the chemical form could be made using detailed thermodynamic predictions.
EPA's Risk Reduction Engineering Laboratory (RREL) in Cincinnati and the Air and
Energy Engineering Research Laboratory (AEERL) in Research Triangle Park (the
Combustion Research Branch) has developed procedures for assessing the
thermodynamics of metals under incineration conditions.
The assessment of the particle size distribution of metals and the distribution of the
metal within particles is a more complicated process. There is currently no data available
from WTI on these emissions properties except for some general size distribution data
from the trial bum. These properties are both dominated by aerosol dynamics resulting
from the nucleation, condensation, coagulation, and agglomeration of vapors and
particles. The assumption used in the plan suggests the use of uniform distribution of
the metals throughout the particle sizes and within each particle. Some preliminary
assessment of such physics should be made by examination of results of mathematical
3-9
-------
predictions from aerosol dynamics codes such as MAEROS. Again AEERL and RREL
should be contacted as good sources of information on these codes.
3. Longer-Term Recommendations
In the longer term more reliable emissions rate data should be obtained on all of the metals
of interest and the physical and chemical form of the metal in the emissions. The emissions rates
of other metals should be obtained with available sampling and analysis protocols in a
straightforward manner.
The physical and chemical forms of the metals in the emissions require more complicated
analyses because of the trace amounts of paniculate matter in the stack and the even smaller
amounts of metals within particle size classes and individual particles. In the midterm the work
group recommends a detailed assessment of the needs of the other aspects of the comprehensive risk
assessment for these parameters by conducting sensitivity studies. Filling these gaps in the midterm
could rely on more detailed theoretical predictions using vapor/aerosol dynamics codes. If these
parameters are of significant importance, then sampling and analysis protocols will have to be
developed to allow then* direct experimental assessment in the field.
Upsets and Accidental Releases
1. Issues
Due to the proximity of this plant to schools and residences, one of the critical factors affecting
health impacts associated with emissions from the WIT facility is the assessment of emissions
occurring during abnormal operation and accidents. These conditions potentially include the
following scenarios:
• Transients due to non-steady state operation such as startup and shutdown.
• System upsets such as malfunctions or perturbations in equipment operation that could
result in a waste feed cutoff if severe enough to approach or exceed permit limits.
3-10
-------
Fugitive emissions due to leaks and spills.
Very low frequency events such as fires and natural disasters that could lead to accidental
releases.
The project plan is extremely vague on techniques to be used to assess the likelihood of these
events and the level of emissions that will occur during an event. The Addendum to the
methodology for assessing health risks associated with indirect exposure to combustor emissions does
supply some guidance on estimating emissions during certain operating conditions but does not
address fugitive emissions or accidental releases. The data base for emissions from startup and
shutdown, fugitive emissions, and accidents appears to be weak and the attempt to use these data
may weaken the credibility of the study. It may be better to review the data quality early in the
program and determine whether it is reliable enough to use for this purpose. An alternative
approach needs to be developed to address these important scenarios if the data are not adequate.
2. Short-Term Recommendations
The procedures for estimating emissions under upset conditions is not well developed.
Estimating the magnitude of emissions during an event requires an approach that
defines both the frequency of the occurrence and estimates the size of the release during
the occurrence. The frequency of the normal occurrences can be estimated from
historical data for both the WIT facility and other incinerators However, it is important
to ensure that past operation is reflective of the future (i.e., ensure that WTI is
intermittently feeding containerized heterogeneous solids that are prone to yield the most
unsteady combustion conditions).
Estimating the character and magnitude of the emissions is more problematical. One
could rely on the estimation procedure proposed for organic emissions in the Addendum
to estimate the chemical characteristics of the emissions. The techniques proposed in the
Addendum for estimating emissions during malfunctions, startup, and shutdown largely
rely on the California Air Resources Board default values. This procedure should be
referred to in order to determine the magnitude of the emissions; however, the procedure
should be closely examined in light of the specifics of operation of the WTI unit.
No guidance is provided in the Addendum on fugitive emissions or low frequency,
potentially catastrophic accidents. The project plan calls for use of unspecified EPA
models for assessing fugitive emissions. In the short term the work group recommends
consultation with the EPA Office of Air, which is assessing such emissions as part of the
requirements for implementation of the Emergency Response and Right-to-Know aspects
of the Clean Air Act Amendments of 1990.
3-11
-------
3. Long-Term Approach
Estimates of emission of toxic chemicals during incinerator upsets or transients suggest that
long-term-averaged emissions may increase by no more than a factor of two from normal operation.
Because the incinerator has an automatic waste feed cutoff system tied into the continuous
monitoring system, it is highly unlikely that a catastrophic emission will occur during an incinerator
upset. However, the work group strongly feels that the sudden release of toxic chemicals in an
accidental fire, explosion, or related event could have a significant impact on safety at the nearby
elementary school.
In the long term, a much more complete assessment of the emissions associated with accidental
releases must be made. Several sources should be consulted in the assessment of the emissions from
these releases, including literature on emissions from open fires, accidental release modeling
approaches used by industry and EPA in emergency response planning, as well as flammables and
explosives event models. The work group cannot offer any detailed approach other than to suggest
that an assessment of the types of accidents be based on historical incinerator data and that a
comprehensive analysis of the emissions from the more likely scenarios be estimated based on state-
of-the-art modeling procedures. Reasonable worst case release scenarios should be formulated based
on consideration of hypothetical events. This is clearly an area where much more research and data
is needed in order to make any sort of truly realistic estimate of risk.
3-12
-------
Meteorology/Air Dispersion Work Group
Walter F. Dabberdt, Chair
Surface & Sounding Systems Facility
National Center for Atmospheric Research
Boulder, CO
Halstead Harrison
Atmospheric Sciences
University of Washington
Seattle, WA
The design and siting of the existing hazardous waste incineration facility seriously limits the
ability to mitigate potentially adverse impacts associated with the atmospheric exposure pathway.
The facility is poorly located in a steep, narrow river valley where meteorological conditions inhibit
effective dispersion. The problem is further complicated by the lack of "good engineering practice"
in the design of the stack. Moreover, the risk assessment is additionally impeded by the lack of a
comprehensive suite of site-specific meteorological measurements. Given these limitations, the
following "short-term recommendations" are provided in the spirit of improving the atmospheric
dispersion aspects of the risk assessment project plan within the timeframe imposed by completion
of the Phase II assessment in April 1994. "Long-term recommendations" also are provided for
development of a more robust and representative assessment of risk without the arbitrary schedule
limitation. Accordingly, these comments and their subsequent incorporation into a revised risk
assessment project plan should not be viewed as setting a precedent or a model for future risk
assessments. Future risk assessments should be undertaken with atmospheric dispersion
considerations that are based on the types of improvements embodied in the suite of long-term
comments. Also, future risk assessments should be conducted prior to siting and construction of
other facilities in order to avoid the serious siting and design flaws that characterize the WIT East
Liverpool, Ohio, incinerator.
3-13
-------
Recommendations
1. Short-Terra
Neither the COMPDEP model nor the new ISC-COMPDEP model adequately simulates
the complex atmospheric dynamics of windflow in and above the Ohio River valley and
adjacent hilly terrain in the vicinity of East Liverpool, Ohio. Their treatment of ambient
diffusion and deposition/scavenging processes also is limited, though likely adequate for
assessment of exposure, dose, and risk over long periods with relatively uniform emission
features. A marginally acceptable evaluation of transport and diffusion might be
accomplished provided representative wind (and temperature) measurements are
available to estimate an effective transport wind vector at plume height (and to estimate
plume rise). The assessment will be compromised if site-specific 10- or 30-m wind
observations (or Pittsburgh soundings) are used to transport the plume. Either approach
can lead to significant errors in defining impacted areas. Instead, it may be possible to
provide a reasonable wind estimate through consideration both of the site-specific WIT
meteorological observations and the 10-, 46-, and 150-m meteorological observations
available in the river valley about 13 km from the Beaver Valley Nuclear Power Station.
Additional local meteorological data (particularly precipitation measurements) should be
obtained and analyzed along with the full set of WTI and Beaver Valley Nuclear Power
Station meteorological data to provide a more representative estimate of wet deposition,
transport, stability, mixing height, and diffusion rates. Analysis of these data should be
refined; for example, short-period (5-minute) averages and higher moments should be
routinely calculated and archived.
The COMPDEP model needs to be modified to more realistically consider the adverse
effects associated with fumigation and calm conditions. The proposed treatment that
ignores calms (except in the adjustment of the effective number of hours constituting the
annual average) is inappropriate and likely will lead to significant underestimation of the
annual average concentration at many near-field receptors. By analogy, ignoring
fumigation cases can result in underestimated concentrations at mid- and far-field
receptors.
Terrain-induced downwash is almost certainly expected to be a serious problem at the
WTI site (at least for moderate- and high-wind cases).. None of the EPA Guideline
Dispersion models is able to simulate these effects. EPA's Good-Engineering-Practice
(GEP) Guidelines for Stack Height Determination (Huber andSnyder, 1983) recommends
that either numerical modeling or fluid modeling should be undertaken in such cases to
determine GEP stack height. We are not aware of any numerical models that have
successfully been applied in this manner, although models exist with this potential. On
the other hand, wind tunnel studies have been applied in this way and should be
undertaken for this purpose in the case of the WTI risk assessment. Additional benefits
of such modeling are the quantification of the near- and mid-field three-dimensional
windflow within and downwind of the river valley and the quantification of the combined
effects of terrain and buildings on the near-field dispersion.
3-14
-------
Short-term increases in incinerator emissions can result from process upset conditions
and accidents. COMPDEP as presently structured cannot simulate these potentially
important events; the model should be modified to enable reasonable estimates of short-
term concentrations. COMPDEP can estimate concentrations (within the limitations
discussed earlier) attributable to fugitive emissions, however, and the model should be
applied for this purpose.
COMPDEP sensitivity tests should be conducted to demonstrate the magnitude of
variation in concentrations and deposition estimates to variation in model inputs.
Similarly, uncertainty analyses should be conducted to provide estimates of the error bars
of the model's concentration and deposition outputs.
The partitioning of emissions among the vapor and solid phases is a very important
aspect of the concentration/exposure/dose assessment process. The COMPDEP model
cannot consider ambient chemistry or volatilization processes that might result in vapor-
to-particle transformations (or vice versa). The potential significance of such
transformations is beyond the expertise of the work group, but ought to be addressed
independently by supplemental peer reviewers with expertise in this area.
In view of the short-term technical recommendations above, it is further recommended
that the April 1994 deadline be modestly extended to ensure incorporation of the
suggested modifications.
2. Long-Term
The Agency should develop new guidance that ensures that the appropriate suite of
atmospheric measurements (and subsequent processing and analysis) is made to support
comprehensive risk assessments (and other dispersion-oriented analyses) in complex
terrain. Such guidance should reflect advances in remote and in situ measurement
technology and the requirements of advanced non-steady-state models.
Additional site-specific surface and upper-level (i.e., boundary-layer) meteorological
measurements should be made at the WIT facility in spite of the apparent near-term
requirement for a completed risk assessment. Our experience is that these deadlines
frequently shift, and that such data typically become necessary at a later date.
Undertaking enhanced measurements at this stage is apt to be a cost-effective measure
when viewed in relation to a future requirement.
The Agency should abandon steady-state modeling applications and initiate development
of a robust, state-of-the-art meso-gamma scale (less than 50 to 100 km)
transport/diffusion/reaction/deposition model embodying such features as four-
dimensional data assimilation (meteorological and chemical) and nested grids. Such a
model should explicitly consider terrain effects on the windflow and time- (and space-)
varying emissions, precipitation, and chemistry. Extensions of limited and simple models
to extended and complex problems should be phased out.
3-15
-------
Uncertainties in the estimation of wet deposition are very large—particularly on the
meso-gamma scale and also for acute episodic conditions. The Agency needs to develop
and implement a plan that identifies the gaps in our knowledge and supports theoretical
and experimental studies to provide reliable estimates.
3-16
-------
Exposure Assessment Work Group
Thomas McKone, Chair
Lawrence Livermore National Laboratory
University of California
Livermore, CA
James Butler
Argonne National Laboratory
University of Chicago
Agonne, IL
M. Wilson Tabor
Institute of Environmental Health
University of Cincinnati Medical Center
Cincinnati, OH
George Fries
Agricultural Research Service
U.S. Department of Agriculture
Beltsville, MD
The proposed exposure assessment relates to the characterization of exposures attributable
to contaminants in the gas and particle phases of the atmosphere. Our comments are divided into
two categories:
Priority issues that are features of the project plan that can be addressed within the
available timeframe.
Specific comments on emissions characterization, exposed populations, exposure
concentrations, dose estimation, uncertainty and variability, evaluation of routine
fugitive and short-term episodic emissions, and other issues that can, in most cases,
be addressed within the available timeframe and would be useful to consider in
future risk assessments of this type.
Overview and Priority Issues
In spite of inconsistencies, omissions, and uncertainties noted below, the project plan
provides a comprehensive approach for evaluating human exposures to stack emissions. The plan
provides a process for addressing multiple routes of exposure, multiple chemical agents, and specific
characteristics of the population at this site.
3-17
-------
It should be recognized that the plan addresses human exposures in the region of concern
and only addresses exposures attributable to operation of a single kiln at the WIT site for 30 years.
Existing and future exposures to other sources in the area are not addressed. This issue becomes
relevant if there are future efforts to monitor exposure in this region. The proposed risk assessment
cannot be used to estimate total exposure for this population but only the exposures attributable to
WIT. Also, this plan does not address the regional impacts (i.e., beyond 50 miles) of this facility in
combination with other combustion sources.
Uncertainties need to be confronted quantitatively. The treatment of uncertainty and the
issues of uncertainty and variability should be integrated as much as possible into all aspects of the
report and not just included as an addendum to the section on risk characterization. Numbers that
have large statistical variances should not be reported as single values. Regression equations for
biotransfer factors should be presented with the standard error of the estimator. The three-tiered
approach to uncertainty analysis described below should be used.
Rather than focus on the magnitudes of the mid-range and high-end exposures, it would be
more appropriate to focus on the question of what is our confidence in specifying exposure
distributions within a population and among more highly exposed individuals within this population.
We should ask how precisely can we determine the probability that the range of dose in some
exposure group exceeds x, where x is an unacceptable value, rather than ask what is our conservative
estimate of risk.
Use of site-specific data is to be commended and encouraged. The concept of an interactive
modeling and measurements approach to exposure assessment, which is suggested by this plan,
should, to the extent possible, be expanded. That is, the exposure models described here could serve
as a portfolio of scenarios for contaminant-specific, multimedia dispersion in the environment.
Integrated measurement/modeling efforts of this type could serve as road maps to identify pathways
and populations for which detailed exposure survey studies are needed.
Compounds selected as surrogates for the risk analysis on the basis of quantity, toxicity, and
K^, should not necessarily be used to carry out validation studies unless it can be demonstrated that
these also are persistent compounds—as is implied by the exclusion of a persistence factor from the
3-18
-------
selection criteria. Also, evaluating exposures to short-term releases—accidents, fires, equipment
malfunctions, fugitive emissions—is necessary.
The integration of surface water, soil, and atmosphere is particularly well done. The
inclusion of the net effect of gaseous diffusion from air to water and volatilization from water to air
makes the model more credible and should improve its reliability. Where possible, information on
feedstocks and ash residuals should be used together with a mass balance to verify emissions
estimations.
Comments and Recommendations on Specific Components of the Plan
1. Emissions Characterization
Mercury, Arsenic, Cadmium, Lead, and Chromium. The assumption that all metals are in
the particulate phase is problematic specifically for mercury and also potentially for arsenic,
cadmium, and chromium. The models proposed in the plan are not suited to speciation of inorganic
metals. Mercury has a relatively high vapor pressure so it does not necessarily come out of the stack
bound to particles. There can be transformations between the particles and gas as the plume
migrates away from the plant. Bioavailability for plants, food-producing animals, and humans
depends on the chemical form of the metals. It is not clear how the transport and exposure models
will address the chemical thermodynamics and reaction kinetics that define chemical form.
Fugitive Emissions. The process of handling collected ash from the kiln should be included
in the estimation of fugitive emissions.
2. Exposed Populations
An important question is whether it is possible to derive sufficient site-specific information
within the timeframe of the project to allow reliable estimation of the numbers and types of activities
associated with various population subgroups. Estimates for high-end exposure should be more
reliable than estimates of central tendency.
3-19
-------
We identified a number of groups that might be considered high-end or sensitive exposure
groups. These are:
« Children who live near the facility and attend school near the facility.
» Elderly people (who apparently make up a larger than expected fraction of the East
Liverpool community).
* Those who are already highly exposed to metals (i.e., lead) and dioxin-like
compounds from other sources.
• Hunters of deer and waterfowl.
« Very active people with higher breathing and food consumption rates.
• Individuals who both work at the WIT facility and live near it.
There needs to be a survey of gardening patterns within the primary plume area—that is,
within East Liverpool itself and up and down the river valley where contaminants might be blown
much of the time. Much of the area surveyed for home garden practices is to the west of the facility
where the plume does not blow as often.
There is an issue concerning the consumption of local versus homegrown meat, milk, and
eggs. Homegrown roods are produced on land associated with a household and consumed within
that household; whereas locally grown foods are produced in home gardens and commercial farms
with air, soil, and/or water that have been contaminated by the incinerator and are consumed by the
other households in the local population. Also, slaughterhouse data should be used to assist in the
estimation of locally consumed meat.
3. Exposure Concentrations
The Biddleman model is used to determine the relative amount of chemical in gas phase
versus particles in air. There are empirical parameters used in this model that have been fit using
a small set of chemical data. It has been noted that these empirical values could be substantially in
error for several compounds. Since the vegetation-uptake model is very sensitive to the relative
amount of contaminant in vapor relative to particles, these values must be reevaluated.
3-20
-------
No discussion of exposure via household dust is included in the plan. Since this is a
significant route of exposure for sensitive subgroups, such as infants and children, it should not be
discounted. Dermal and ingestion pathways for outdoor soil do not necessarily represent how these
exposures occur inside houses. House dust originates from three sources:
Airborne particles that penetrate buildings in the exchange between outside air and
indoor air.
Surface soil and dust tracked into buildings on shoes or clothes, or by pets or other
vectors.
A variety of sources related to occupant activities, material degradation, and
household products.
According to the equation on the bottom of page 71 in the draft project plan, the transfer
of Chemicals from air to soil is by deposition, which does not explicitly include diffusion. In
Appendix B, Tables B-l and B-6, volatilization (or mass diffusion) is described as one of the
e
mechanisms by which chemicals are transferred from soil to air. In Table B-6, the mass transfer
coefficient is applied to soil as if the chemical concentration in air is zero. As proposed here, these
equations, when combined, will result in a failure to achieve conservation of mass and comply with
chemical thermodynamics for semivolatile organic chemicals. These equations should be modified
to avoid this problem. The approach used for the surface-water compartment could serve as an
appropriate guide for this modification.
Although it might seem obvious to some that there is no need to include ground water as
a potential exposure medium, the basis for excluding this exposure medium should be discussed and
justified in the risk assessment.
The COMDEP model does not count snow as a form of precipitation for purposes of
estimating deposition of contaminants from air to soil. This could result in underestimating the soil
concentration of particle-bound contaminants. This could be remedied by assigning a value other
than zero to the F factor for snow.
3-21
-------
The exposure concentrations included in this risk assessment should be reconciled with
concentrations compiled in several states. For example, the California Air Resources Board has
carried out hundreds of risk assessments for point source emissions and these risk assessments all
have a protocol for addressing indirect exposure pathways. New Jersey has confronted the issue of
long-range mercury transport.
4. Dose Estimation
A high-end exposure duration for breast feeding of 270 days is too low as a high-end value.
It is not unusual these days to have infants that are breast fed for a year or more, although this
would not be a likely primary source of food for an infant beyond 6 months.
Food consumption rates listed in Table A-9 are based on survey data that is at least 15 years
old. There has been a shift in dietary habits since the survey. Currently, pork and poultry
consumption are more significant relative to beef than indicated in these consumption data.
Nonetheless, the high-end values should still be realistic for the highly exposed population subgroups.
(Note that high-end numbers in Table A-9 appear to be duplicative with typical exposure numbers.
Data should be checked against the original references.)
The fish consumption data provided in Table 3 on page 19 are out of date. West Virginia,
Ohio, and Pennsylvania have stepped up efforts to carry out fish surveys. The more current
information in these surveys should be used.
5. Uncertainty and Variability
One of the issues in an exposure assessment that must be confronted is how to distinguish
between the relative contribution to outcome variance of true uncertainty versus inter-individual
variability (i.e., heterogeneity). Uncertainty or model-specification error (e.g., statistical estimation
error) can be modeled using a random variable with an identified probability distribution. In
contrast, inter-individual variability refers to quantities that are distributed empirically within a
defined population—such factors as food ingestion rates, exposure duration, and expected lifetime.
3-22
-------
The risk assessment project plan addresses variability by using central and high-end exposure
groups. Ideally, a statistical variance propagation method should be used to define the distribution
of potential dose. Nonetheless, this "high-end" approach seems adequate for characterizing
variability. The process for addressing uncertainty in the plan, however, is vague. Potentially large
uncertainties exist in model predictions and these uncertainties must be communicated to those using
the results of the risk assessment. For example, the atmospheric dispersion model COMDEP can
simulate atmospheric dispersion to only within a factor of two or more at this complex site. In the
next section, we offer an approach for addressing uncertainties in the risk assessment. In the
remainder of this section, we consider some of the more significant sources of uncertainty in the
proposed risk assessment.
How much uncertainty is added to the assessment by using data from similar incinerators
and/or from the trial bum as the primary basis for estimating the source of emissions? Is there some
way to consider more carefully the types of feedstocks expected so the precision and credibility of
this estimate can be improved? Because the source term magnitude is the primary parameter of the
exposure and risk assessment, any uncertainty in this value will be mapped onto all the results.
There are a number of partition and biotransfer factors that are included in the model and
many of these are likely to be quite uncertain. The impacts of the uncertainty in these inputs should
be addressed by sensitivity and uncertainty analyses. Among the inputs that are likely to have high
uncertainty factors are:
Octanol-water partition coefficients
Organic-carbon partition coefficients
Soil-water and sediment-water partition coefficients (K,j)
Henry's law constant
Mass-transfer coefficients in water and air
3-23
-------
Steady-state bioconcentration factors for
plant root concentrations relative to soil
concentrations
plant leaf concentrations relative to air concentrations
fish concentrations relative to water/or sediment
concentrations
Steady-state biotransfer factors for
milk or dairy-product concentration relative to
contaminant intake by cattle
meat concentration relative to contaminant intake by animals
egg concentration relative to contaminant intake by chickens
breast milk concentrations relative to contaminant
intake by mothers
Contaminant biodegradation factors in soil
6. A Tiered Approach to Uncertainty Analysis
a. The results of this risk assessment will be more credible if the variance in the input values
is clearly stated and the impact of these variances on the final estimates of risk is assessed. At a
minimum, this can be done by listing the estimation error or the experimental variance associated
with the parameters when these values or their estimation equations are listed in tables. It would
help to define and reduce uncertainties if a clear summary and justification of the assumptions used
for each aspect of the model were provided. In addition, it should be stated whether these
assumptions are likely to result in representative values or conservative (upper bound) estimates of
exposure.
b. A sensitivity analysis should be used to assess how model predictions are impacted by
model reliability and data precision. The goal of a sensitivity analysis is to rank the input parameters
on the basis of their contribution to variance in the output. Sensitivity analyses can be either global
or local. A global sensitivity analysis quantifies the effects of variation in parameters over their
entire range of values. Such an analysis requires an uncertainty analysis as a starting point. The
variance in the outcome is compared to the variance of the inputs. A local sensitivity analysis is used
to examine the effects of small changes in parameter values at some defined point in the range of
outcome values.
3-24
-------
c. Variance propagation methods (including but not necessarily Monte-Carlo methods)
should be used to carefully map how the overall precision of risk estimates is tied to the variability
and uncertainty associated with the models, inputs, and scenarios. This type of approach gives the
decision maker flexibility to address margins of error; to consider reducible versus irreducible
uncertainty; to separate individual variability from true scientific uncertainty; and to consider
benefits, costs, and comparable risks in the decision-making process.
7. Short-Term Releases
There are three types of events that should be included in this category:
• Upset conditions
• Fugitive emissions
• Accidents
These events could result in larger annual releases than the routine emissions to which the majority
of the report is devoted.
In the case of accidental releases, it will be necessary to make use of a combination of fault-
tree studies, local transportation accidents data, reviews of operating experience, and reviews of past
experience to determine both the frequency of accidents and the chemical source terms associated
with these accidents.
3-25
-------
Toxicology Work Group
Mary Davis, Chair
Robert C. Byrd Health Sciences Center
West Virginia University
Morgantown, WV
Eula Bingliam
Department of Environmental
Health
University of Cincinnati
Cincinnati, OH
Pirn Kosalwat
KBN Engineering and Applied Science
Gainesville, EL
Thomas Gasiewicz
Department of Environmental
Medicine
University of Rochester
Rochester, NY
The group considered the potential for adverse health impacts from the WIT plant. The
project plan attempts a comprehensive evaluation of the facility and considers multiple chemicals
and multiple pathways of exposure. The group found difficulties with the plan in several areas.
These are generally grouped into categories of insufficiencies, vagueness, and scientific underpinnings
of the assessment.
Insufficiencies
1. Environmental Risk Assessment
Environmental effects of the facility have not been addressed. The work group recommends
that an ecological risk assessment be conducted, including at least the following elements:
• Identify the contaminants of concern.
• Characterize release, migration, and fate.
" Identify the routes of exposure.
3-26
-------
• Identify the exposed habitats, such as forest preserves, streams, rivers, wetlands, ponds
and lakes, grasslands, farms, and residential areas.
• Identify the exposed populations, including endangered species, sensitive species (e.g.,
reproductive/migratory sensitivities), foodweb-based species, and commercial and sport
species.
• Determine the ecological effects and perform an environmental evaluation, considering
community structure, species diversity, population dynamics, and effects that could
impair reproductive and behavioral processes.
• Characterize the ecological risks.
Without having information about the ecosystem in the area, it is difficult to estimate the
level of assessment needed to evaluate ecological risk. The chemicals to be evaluated for ecological
risk should be selected based on their toxicity to potential receptors, their persistence in the media
where potential receptors will be exposed, and their bioconcentration/biomagnification in the food
chain. These indicator chemicals may not necessarily be the same as those identified for human
health risk assessment. The indicator chemicals and likely exposure pathways are used to select
appropriate endpoints for the assessment, such as reduction of key populations, species diversity, and
disruptions of community structure.
The ecological risk assessment should be an ongoing process that parallels the operation of
the facility, and it should include biological and environmental monitoring.
Primary and secondary sources of ecological information include:
• Primary
- State natural resource departments
- U.S. Fish and Wildlife Service
- U.S. EPA's Aquatic Toxicology Laboratory (Wheeling, WV)
- Ohio River Sanitation Commission (ORSANCO)
• Secondary
- Army Corps of Engineers
- U.S. Geologic Survey Regional offices
- Ohio River Basin Consortium for Research and Education (ORBCRE)
- Other (U.S. Department of Agriculture, Farm Bureau, industry)
3-27
-------
2. Incremental Risk
Exposure to emissions from the WIT plant occurs against a background of previous and
ongoing exposures to a variety of other chemicals, including chemicals of concern for the WIT
facility. The additional burden from the WIT plant may move total exposures to the steep part of
the dose-response curve. Thus modeling just the exposure from the WIT plant would underestimate
the impact of the plant. For those chemicals that are air pollutants, the total exposure in air should
be modeled. For the other, existing body burden should be considered, especially for chemicals of
concern that are retained within the body (lead, dioxin).
Another source of incremental risk will occur when the second kiln (for which the plant is
permitted) is brought into operation. The current risk assessment does not fully characterize the
risks planned for the entire facility.
3. Utilize Data from the WIT Facility and the Area
Because the WIT incinerator has been in operation, samples should be collected, from the
facility and areas predicted to be impacted, and analyzed to validate the model. If necessary, the
model should be adjusted to make it more reliable. Similarly, information on emissions from the
plant itself should be used in the risk assessment, in preference to results from similar or not-so-
similar facUities at other sites.
The WIT facility has had, and will continue to have, noncompliance events. The nature of
these events should be investigated and incorporated into a model of noncompliance episodes (using
data from Ohio EPA and its Twinsburg office) to give a realistic accounting of the hazards.
4. Health Effects Data from European and Scandinavian Facilities
The project plan calls for using exposure and emissions data from other hazardous waste
incinerators, particularly those that are similar in design to the WIT facility. The risk assessment
also should include discussion of any health effects data from those other facilities.
3-28
-------
5. Discussion of Uncertainties in the Risk Assessment Process
The group was particularly concerned that the numerical estimates be well documented in
the text and summary. In particular, it is important to get away from focusing on a number and to
recognize the variety of uncertainties inherent in the process. This discussion should not be
restricted to the section on uncertainty, but rather should be addressed as the parameters with
uncertainty arise. This would include, for example, a discussion of the quality and reliability of the
number used for a particular toxic endpoint, the limitations of not knowing all the chemicals to be
emitted, or weakness from using data on meteorological conditions collected over only one year.
A brief discussion with caveats on uncertainties and weaknesses of risk assessment should be
included in the introductory material and where numerical risk assessments are presented (perhaps
as a footnote to a table).
6. Compounds to Be Evaluated
The speciation of metals needs to be determined before the risk assessment for the metals
is done. The spectrum of toxic effects and doses that produce those effects are different for
different forms of metals. The risk assessment should be based on actual exposures. Additional
products of combustion should be included: benzo(a)pyrene, benzo(b)fluoranthene, chrysene,
dibenzo(a,h)anthracene, fluoranthene, anthracene, nitrogen heterocyclics, and sulfur heterocyclics.
Also, nickel, copper, and aluminum should be evaluated.
Scientific Underpinnings
1. Reference Sources
EPA documents are a major source of data used in the project plan. These references often
are not readily available. It was not clear, therefore, if these documents include discussions of data
from the appropriate primary sources. Also, it appears that other data bases were not used as
sources for toxicologic information. Other potential sources of toxicology information include:
3-29
-------
" EPA's Integrated Risk Information System (IRIS)
» EPA's Health Effects Assessment Summary Tables (HEAST, Environment Criteria and
Assessment Office, Cincinnati, OH)
• National Library of Medicine (NLM) and Toxicity Information Online (TOXLINE)
• Registry of Toxic Effects of Chemical Substances (RTECS) (on-line through NLM)
• Agency for Toxic Substances and Disease Registry (ATSDR, Atlanta, GA) (toxicologic
profiles for specific chemicals)
» EPA Health Effects Assessment Documents
• Air Toxics and Risk Assessment, E.J. Calabrese and E.M. Kenyon (Lewis Publishers,
Chelsea, MI)
» American Council of Governmental Industrial Hygienists (ACGIH) Documentation of
the Threshold Limit Values and Biological Exposure Indices (Cincinnati, OH)
2. Isopleth Boundaries
The use of dioxin cancer rates for setting the isopleths should be justified scientifically. The
endpoint chosen for setting the isopleth boundaries will determine the extent of the population
considered to be at risk. For example, if dioxin cancer risk from consuming homegrown beef is not
sufficiently sensitive, then the population considered will be lower than the actual population. Other
effects should be modeled to validate the use of dioxin cancer risk, as dioxin may not accurately
reflect the distribution of other compounds.
3. Toxicokinetic Parameters
There are toxicokinetic parameters, such as absorption and bioavailability, available for some
of the chemicals and this information should be incorporated where possible. Where such data are
not available, the assumptions and justifications for them should be given.
3-30
-------
Items That Are Vague
1. Endpoints
It was not clear what endpoints would be used for which chemicals and subgroups. Different
endpoints will be needed for different chemicals, and individuals may be at risk for different effects
depending on age. The risk assessment should use the most sensitive endpoint, which may not be
cancer, and give consideration to other endpoints appropriate for each chemical being evaluated.
It would be useful to discuss, perhaps in the uncertainty section, how the dioxin cancer risk compares
to other risks (e.g., immunosuppressive or reproductive and developmental toxicity) for dioxin, since
these endpoints occur at lower doses than cancer.
The risk assessment should evaluate both acute and chronic respiratory effects of particulates
and acid gases. Coal-fired power plants or other significant sources of acid gases are potential
sources of data or experience with this issue. On December 9, the New York Times reviewed an
article in the current (12/9/93) New England Journal of Medicine, which described substantial
increases of chronic respiratory disease and decreases of life expectancy in areas of the country with
high concentrations of fine particles, compared to areas with low concentrations. Based upon the
Times' summary, peer reviewers noted that this reinforces the importance of a full evaluation of the
health effects of particles.
2. Multiple Risks
Many chemicals will be evaluated for risk, but it is not clear how the results of those
evaluations will be combined to characterize the total risk an individual will experience. Potential
additive or synergjstic effects should be considered when appropriate data are available.
3. Unavailability from Inhalation Exposures
Toxic effects from inhalation exposures appear to be based on assumptions that are neither
given nor justified, particularly bioavailability from agents adsorbed onto particulates. This is a key
factor for dioxin and metals since they adsorb to particulates.
3-31
-------
4. Characterization of Releases
Emissions from European and Scandinavian facilities of similar design are to be utilized in
the risk assessment; however, the data were not given. Knowing the potential exposures is crucial
to evaluating the lexicological aspects of the risk assessment plan and not having this information
greatly complicated our task.
Confidentiality of wastes creates a problem for predicting emissions. Composition of waste
could be "bulked" together to protect confidentiality while indicating what is being burned. Since
the facility has been incinerating waste for a period of time (close to a year), these data should be
included and used wherever feasible; for instance, to determine the compounds that are being
emitted, rather than relying on data from a trial burn or from European and Scandinavian facilities.
5. Non-routine and Fugitive Emissions
The group is very concerned that potential emissions from non-routine conditions be
considered as a major part of the risk assessment and not an afterthought. These may be the most
important exposures and they should be fully evaluated. Methodologies are available to provide
estimates of risks for accidents and spills (failure analysis) and should be incorporated. In assessing
fugitive emissions, the quantity of materials on site and the impact of environmental conditions on
volatility should be considered. For process upsets the model should include how "upset" the
processes are and how this affects the different products emitted. Modeling of accidents is crucial.
While accidents, especially severe accidents, may be unlikely events, the group is concerned that high
exposures could occur away from the containment devices. The plan should describe the scenarios
that will be modeled for accidents and if they are the worst case scenarios. The assumptions or
parameters to be included should be stated. For example, the amount of the waste stored on site
that is assumed to be involved in the accident.
3-32
-------
6. Lead
The group was concerned that the Phase I risk assessment stated that the air quality standard
for lead would be violated. What does that mean and how will it be handled in the new risk
assessment? Also, the updated biokinetic uptake model for lead should be used.
Suggestions for Future Risk Assessments
1. Ongoing Assessment
A risk assessment needs to be an ongoing process, with collection of data for validating the
models used to estimate exposure.
2. Pre-operational Monitoring
Monitoring for background/current exposures during the permit process should be required
so that data are obtained before the plant operates. This will allow assessment of incremental risks.
The State of Florida's modeling requirements could be used as a model.
3. Site Visit
A visit to the WIT site would allow full appreciation and consideration of all site-specific
factors affecting the assessment of risk. Site visits should be required for future risk assessments,
and early in the risk assessment process.
Priority Items
The group suggested that highest priority be given to: . '
• Characterization of releases
• Consideration of non-routine and fugitive emissions
3-33
-------
» Conducting environmental risk assessment
• Consideration of incremental risks
• Use of sampling data from the WIT facility and surrounding area
• Including discussion of any health effect data from European and Scandinavian facilities
3-34
-------
SECTION FOUR
HIGHLIGHTS FROM COMMENTS
Peer Reviewers* Preliminary Comments
Prior to the workshop, each peer reviewer was asked to prepare written comments on the
utility of the draft project plan as the foundation for a scientifically responsible risk assessment for
the WTI incinerator. Reiving on their technical knowledge and best professional judgment, peer
reviewers responded with comments on:
• Features of the plan expected to lead to a strong risk assessment.
• Features of the plan that can be improved within the Phase n risk assessment timeframe
for meeting regulatory schedules.
• Issues that ideally (i.e., without time and resource constraints) would be useful to
consider in future risk assessments of this type.
• Specific issues (i.e., emissions characterization, hazard identification and dose-response
evaluation, exposure assessment, and risk characterization) related to the panel's areas
of expertise.
Appendix E provides the charge to peer reviewers and background information. Peer reviewers'
premeeting comments are presented in Appendix F.
In general, the peer reviewers found the draft project plan to provide a comprehensive
framework for conducting a risk assessment at the WI1 incinerator. Nonetheless, the comments did
raise a number of issues for consideration at the workshop that might necessitate adjustments to the
risk assessment.
For example, metals, organics, and dioxins/furans emissions characterization as well as
process engineering and emission scenarios were presented as major issues to be discussed in the
4-1
-------
combustion engineering work group. The lack of information on emissions characterization and
operating conditions for the WIT incinerator left many questions unanswered:
* Are metals volatilized?
• What is the form and fate of mercury?
• What method will be used to prorate surrogate emissions?
« Why were dioxins/furans emissions high before addition of the carbon injection system?
» How will compounds included in the risk assessment be selected?
• Where are the particulate surface area and gas-solid partitioning data?
Another issue concerned characterization of waste feed. The draft project plan provided no
discussion of waste feed characteristics for the WTI trial burn, the Biebesheim facility in Germany,
or waste proposed to be burned routinely at the WTI facility. The nature of the waste feed has
significance for the frequency of upsets and concomitant emissions, as well emissions from accidental
fires, all of which the peer reviewers felt need to be addressed.
Although primary meteorological measurements are available at the WTI incinerator and
near ground level, several meteorological and atmospheric dispersion issues were identified in the
preliminary comments:
• Downwash effects.
» Terrain effects.
• Deposition of fine particles.
• Plume rise as it relates to downwash from buildings and the terrain.
• Ambient meteorological measurements to understand the effects of vertical structure,
winds, temperature, etc.
" Model selection/appropriateness.
• Acute/short-term processes (e.g., upset conditions).
4-2
-------
The terrain surrounding the WIT facility presents challenges for conventional dispersion modeling.
Since topography can effect the transport and diffusion of emissions, downwash from terrain can
have a greater impact than from buildings. Most models, however, do not account for downwash
from terrain.
*
Peer reviewers found that the draft project plan proposes a comprehensive approach for
evaluating human exposures from stack emissions at the WT1 incinerator. They generally approved
of the multimedia, multipathway exposure model presented in the plan. Nonetheless, they did
identify the following issues related to exposure assessment for further discussion and clarification:
Exposed populations
- Use of high-end and central tendency
- Proper representation of children and occupational groups
- Local versus homegrown produce
- Regional exposures greater than 50 miles from the source
Exposure concentrations
- Reliability of intermedia transfer factors
- Partitioning between particles and gas in air
- Transfer from soil and air to plants
Dose estimation
- Process for developing potential dose estimates
- Current food consumption data
- Local data on contact rates
Uncertainty/variability
- Site-specific data to better characterize variability
- More field data to validate models
- Variance propagation, error analysis, Monte Carlo
Other issues
- Selecting surrogate emissions
- Monitoring for mercury
- Justification for excluding ground water
- Fugitive emissions
Because discussion or scientific documentation on toxicology in the draft project plan was
insufficient, peer reviewers found it difficult to comment on this topic. Yet, a major issue for several
reviewers was identification of effects or lexicological endpoints to be modeled and evaluated. It
4-3
-------
was suggested that although cancer appears to be the focus of health effects discussions, toxic
endpoints other than cancer might be more sensitive for a particular chemical. Additional
compounds that should be considered in the risk assessment include nickel, nickel oxides, aluminum,
copper, and polyaromatic hydrocarbons (PAHs). Other toxicologic issues identified include:
» Incremental risk
- Consider the risks added to existing exposures
- Background exposure for well-retained contaminants
» Lead methodology
- Use updated biokinetics model
• Mercury
- Use adequate assessment methods for methylmercury
- Methylmercury RfD provides inadequate protection for in utero
Observers' Comments
Observers were given two formal opportunities to provide information during the workshop.
First, a form was provided at the registration desk on which observers could write topic-specific
technical questions or comments. Completed forms were handed to the appropriate work group
chairs and questions and comments (see Appendix H) were considered in work group discussions.
Second, the workshop agenda included a time for observers to make public statements during the
plenary session held on Wednesday, December 8. Observers were asked to sign up if they intended
to make a statement (see Appendix I). At the discretion of each work group chair, observers also
were provided time during work group sessions to participate in discussions.
Observer comments raised issues concerning the technical and scientific aspects of the draft
project plan as well as the political and personal impact aspects of the WTI incinerator. Highlights
of observer comments include:
It may not be worthwhile conducting a risk assessment for the WIT incinerator, since
the desire to conduct a valid risk assessment does not imply the ability to conduct a
valid risk assessment.
Actual risks and high-end risks should be determined using a probabilistic assessment.
4-4
-------
• The proposed air model, COMPDEP, cannot predict air concentrations of contaminants
outside of the valley.
• The risk assessment cannot be conducted without a lot of input from WTL
• EPA should wait until the Addendum to the indirect exposure guidance is issued before
conducting the risk assessment.
• Conservative values and the level of uncertainty associated with the risk assessment
results should be clearly stated.
• Data quality objectives should be determined for all data collection efforts.
• Site-specific data should be used in the risk assessment.
• Synergjstic and cumulative effects of contaminants should be assessed.
• Scenarios for upset operating conditions should include failure of pollution control
equipment and all equipment over time.
• Waste feed rates and other operating parameters for trial burns should be evaluated and
emissions should be characterized.
• The persistence and potency of dioxins/furans should be evaluated.
• The individuals evaluating the WTI incinerator should be urged to visit the facility.
• Concern was expressed over the process used to select the peer review panel and the
purpose of the workshop now that the incinerator is operating.
» The WTI incinerator should be not be operating while the risk assessment is being
conducted since many uncertainties have been identified regarding the site and emissions
data.
» The State of Ohio passed a law in 1984 that requires incinerators to be at least 2,000
feet from schools and residences if sufficient uncertainty for safety and harm exists. If
it can be shown that no harm exists, the 2,000 feet rule is eliminated.
During the workshop, observers provided the peer review panel with three documents related
to the WTI facility (see Appendices J, K, and L).
4-5
-------
-------
APPENDICES A-F
EPA PREMEET1NG MATERIALS
-------
-------
APPENDIX A
CHRONOLOGY OF EVENTS REGARDING WASTE TECHNOLOGIES INDUSTRIES
(AS OF NOVEMBER 19,1993)
A-l
-------
12 C7>'1993 09:35 FROM EPA-OFFICE PUBLIC AFFfllRS TO
82022600135 P.02
CHRONOLOGY OF EVENTS REGARDING WASTE TECHNOLOilES INDUSTRIES
As of NovMber 19, 1993
09/04/81 WTI applies for permit.
11/13/82 U.S. EPA opens public comment period regarding draft permit, which
remains open until 01/03/83.
12/15/82 U.S. EPA holds public hearing regarding draft permit.
06/24/83 U.S. EPA Issues permit and Response To Comments.
Ot/09/83 State of West Virginia petitions for review of permit decision.
03/29/84 Administrator remands permit pending final decision on petitions.
04/19/84 U.S. EPA issues public notice of second public comment period for
West Virginia.
12/17/84 Administrator denies petitions.
01/25/85 U.S. EPA makes permit effective on this date.
04/20/90 U.S. EPA Issues Notice of Violation for failure to respond to
Information request. S9500 settlement.
10/02/90 WTI requests prior approval to proceed with a Class 1 permit
modification to change Attachment X of the permit, regarding tank
thickness and materials of construction.
10/19/90 U.S. EPA grants prior approval of Class 1 permit modification
described above.
10/29/90 WTI requests Class 2 permit modification to add spray dryer to
Incineration system.
11/02/90 WTI publishes public notice regarding Class 2 for spray dryer.
11/19/90 HTI holds public Information meeting regarding spray dryer.
11/30/90 WTI begins construction of the Incinerator (stack foundation).
Kote that test piles, grading, and relocation of underground
utilities had begun 1n 09/90.
01/18/91 U.S. EPA Issues Notice of Deficiency regarding spray dryer
modification and requests extension of review time for Class 2
permit modification to 30 days after receipt of complete
application. U.S. EPA upgrades modification to Class 3 1n
February.
-------
12/07/1993 39:3? FROM ERA-OFFICE PUBLIC fiFFflIRS TO
S20226001SS P.03
-2-
08/23/91 U.S. EPA opens public comment period regarding,addition of spray
dryer. '
08/27/91 U.S. EPA issues draft permit modification and fact sheet regarding
addition of spray dryer to the RCRA permit.
09/09/91 Citizen's roundtable.
09/24/91 U.S. EPA and OEPA jointly hold public Information meeting in East
Liverpool. *
09/25/91 U.S. EPA and OEPA jointly hold public hearing 1n East Liverpool
regarding permit modification to add spray dryer. Hearing nas to
be canceled due to disruption by protesters.
01/02/92 U.S. EPA inspects WTI facility and discovers illegal construction
of the spray dryer.
01/09/92 U.S. EPA files Administrative Complaint regarding above v'o'uion-
$156,250 proposed penalty.
01/16/92 WTI requests temporary authorization to operate spray dryer.
01/30/92 U.S. EPA Issues Consent Agreement and Final Order regarding
construction of spray dryer without a RCRA permit: $129,000 fine.
02/03/92 U.S. EPA Issues permit modification to add spray dryer and :c add
the Columblana County Port Authority as an additional owner.
03/05/92 The State of West Virginia, the City of Pittsburgh, the CoTu5*lana
County Port Authority, and several others appeal Region 5's
Issuance of the 2/92 permit modification.
03/13/92 WTI again requests temporary authorization to operate spray cryer
because the first request was deemed procedurally deficient.
04/21/92 Attorney General of West Virginia files suit against WTI,
U.S, EPA, and Ohio EPA. The City of Pittsburgh and several
citizen groups intervene.
05/07/92 Hearing 1n front of House Subcommittee on Administrative •-«* *nd
Governmental Relations. Region 5 learns about a
September 21. 1990, contract apparently making Von Roll
(Ohio), Inc., an operator of the facility.
6/17/92 U.S. EPA approves Training Plan.
07/07/92 U.S. EPA requests Information under RCRA Section 3007 to a*4 •«
Its Investigation Into the ownership of the facility. Ad
-------
12/07X1993 09:38 FROM EPfl-OFFICE PUBLIC fiFFflIRS TO
32022600135 P.34
-3-
07/09/92 U.S. EPA issues Phase 1 of the two-phased Risk Assessment.
07/09/92 U.S. EPA grants temporary authorization to operate spray dryer for
180 days.
07/15/93 U.S. £PA's Inspector General issues a. Special Report regarding th«
WTI permit; based upon the Region's response, the audit is closed
on 11/25/93.
07/24/92 U.S. EPA Environmental Appeals Board (EAB) rules regarding spray
dryer and addition of the Port Authority. The spray dryer
modification is upheld. Port Authority issue is remanded-to
Region 5 for resolution.
07/30/92 WTI completes construction.
08/24/92 U.S. EPA inspects facility to determine whether construction had
been completed as required under the RCRA permit.
09/01/92 In the West Virginia case, WTI stipulates to give 7-day notice to
all parties before receiving any hazardous waste on site.
Temporary Restraining Order is lifted.
09/08/92 U.S. EPA notifies WTI of Its preliminary determination that von
Roll (Ohio), Inc., must be added to the permit as an additional
operator, and that Von Roll America, Inc., must be added as an
owner.
09/11/92 Colutnbiana County Port Authority sells the incinerator property
to WTI. .
09/15/92 U.S. EPA meets with representatives of WTI at U.S. EPA
Headquarters.
09/22/92 U.S. EPA meets with concerned citizens at Region 5 offices.
09/25/92 U.S. EPA issues letter regarding results of th« August 24-25
permit compliance Inspection. U.S. EPA approves Closure Plan
dated 7/31/92, Inspection Plan dated 8/19/92, and all piping and
Instrumentation drawings.
09/30/92 U.S. EPA approves Waste Analysis Plan dated 8/7/92, and
Contingency Plan dated 8/10/92, and authorizes WTI to begin
accepting hazardous waste and start the shakedown period.
10/01/92 Grant/Zusman Issue memo regarding issues of ownership and
operational control.
10/02/92 U.S. EPA opens 30-day public comment period on issues related to
ownership and operational control.
-------
12/07/1993 39:39 FROM EPfi-CFFICE PUBLIC SFFfllRS TO
S2022600186 P.35
10/09/92 U.S. EPA imposes interim stack emission levels and feed rates for
toxic and carcinogenic metals for the shakedown period.
10/08/92 Judge in. the West Virginia case dismisses U.S. EPA and Ohio EPA.
10/09/92 WTI gives its 7-day notice that it will begin receiving hazardous
waste.
11/03/92 HTI submits signed Part A application which includes
Von Roll (Ohio), Inc. Public comment period closes.
11/12/92 Judge in the West Virginia case allows WTI to begin the shakedown
period under its permit.
11/13/92 WTI begins receiving hazardous waste and begins the shakedown
period.
12/07/92 Senator Al Gore and other legislators request the GAG to conduct a
detailed review of several major WTI issues.
01/06/93 Ohio EPA approves Trial Bum Plan Revision 4.
01/08/93 U.S. EPA approves Trial Burn Plan Revision 4 and imposes two
conditions prior to limited commercial operation.
01/12/93 Greenpeace and others file suit in Federal District Court
(H.O. Ohio) against U.S. EPA, Ohio EPA, and WTI to prevent the
trial burn from proceeding; a Temporary Restraining Order
-------
12/07/1993 39=40 FROM EPP-QFFICE PUBLIC AFFAIRS TO
32022S0318S P.ES
-5-
03/22/93 United States Supreme Court Justice John Paul Stevens denies an
emergency request to overturn the Sixth Circuit'Court stay.
03/26/93 U.S. EPA informs WTI that trial burn run 8 must be re-done because
of sampling problems which rendered it inconclusive.
03/30/93 HTI repeats trial bum run 8.
04/01/93 WTI notifies U.S. EPA of simple Class 1 modification regard-ing
trial burn POHCs, changed to reflect approved Trial Burn Plan.
04/02/93 HTI certifies that It met carbon monoxide and paniculate emission
Units during the trial burn.
04/02/93 WTI faxes a report to the U.S. EPA showing that the incinerator
failed to achieve the required Destruction and Removal Efficiency
(ORE) of 99.99% during two of the nine trial burn test runs.
04/06/93 U.S. EPA. acting on the carbon monoxide and particulate
certification, authorizes WTI to begin the post trial burn onase
of operations.
04/12/93 U.S. EPA Imposes restrictions on WTI, precluding it from operating
under the conditions maintained during the two failed trial burn
test runs.
04/12/93 C€PA Issues letter allowing WTI to go back Into operation.
04/14/93 U.S. EPA gives prior approval to proceed with Class I permit -rod
requested on 03/16/93.
04/26/93 WTI reports failure during the trial burn to meet the stack
emission limits for mercury during 2 days of the trial burn.
04/26/93 GAD audits Region 5 from 04/26 through 04/30.
05/06/93 U.S. EPA meet* with concerned citizens and representatives of
Greenpeace in Region 5 offices.
05/06/93 Greenpeace requests the Environmental Appeals Board to review the
WTI aatter and halt limited commercial operation.
05/07/93 U.S. EPA issues revised interim stack emission limits and waste
feed rates for toxic and carcinogenic metals, in response to the
report from WTI that mercury emissions exceeded the allowable
Units.
05/08/93 WTI submits trial burn results.
05/27/93 WTI brings facility down for rebrfcklng of kiln.
-------
12/07/1993 11:43 FROM ERA-OFFICE PUBLIC AFFAIRS TO
82022600186 P. 32
06/16/93
06/17/93
06/17/93
06/18/93
06/21/93
06/18/93
06/24/93
06/25/93
06/28/93
07/01/93
07/02/93
07/06/93
07/08/93
07/19/93
U.S. EPA Issues letter expressing concern over dioxin levels in
the trial burn report and requesting details of how WTI wm lower
dioxin emissions, U.S. EPA asks WTf to only bum low chlorine
wastes until matter is resolved.
WTI responds with proposal regarding reduced chlorine feed.
4* pr10r aPP^al to implement a Class I permit
modification regarding minor changes to the Contingency Plan.
U.S. EPA sends letter telling WTI that Its reduced chlorine,
is unacceptable, and suggests facility not go back into operation
until after a meeting in Chicago.
EAB denies review of petitions by (1) Greenpeace, Swearingen. and
Spencer, and (2) City of Pittsburgh and the State of West vi-qinia
for lack of Jurisdiction. Issues were <1) April 6, 1993, U.S. EPA
letter and (2) April 12, 1993. U.S. EPA letter.
Ohio Attorney General's Office issues the results of its
litigation 1nto tn« background of WTI, including its conclusion
that the partnership had dissolved.
HTI meets with U.S. EPA in Chicago, U.S. EPA expresses conce-n
over d1ox1n/furan levels. WTI Interested in meeting new dioxin
levels and 1n pursuing Class 2 pernrtt modification -to allow
Installation of enhanced carbon injection system (ECIS).
WTI submits Class 2 modification request for ECIS.
WTI submits request for temporary authorization to install, test,
and operate ECIS.
WTI notifies of requested Class 2 permit modification for adding
labpacks to the WAP. adding waste codes to the list of acceo-.aole
wastes, and for modifying the Trial Bum Plan to perform a new
test similar to condition 2 of the original trial burn.
Public notices published by WTI regarding all three Class Z parslt
modification being requested.
GAO audits Region 5 from 07/06 through 07/08.
U.S. EPA Issues temporary authorization regarding ECIS.
Sreenpeace/Swearingen/Spencer file in the D.C. Circuit Court of
Appeals regarding the U.S. EPA's decision to allow post tria^ Burn
operation.
-------
12/07X1993 11:44 FROM EPfl-OFFICE PUBLIC flFFflIRS TO
82022600136 P.33
07/23/93 WTI (via Waterman) notifies U.S. EPA of simple Class 1 permit
modification to resolve 3000 pound per square foot permit language
issue. U.S. EPA ORC informs Waterman that this is not
appropriate.
07/23/93 City of Pittsburgh files in Third Circuit Court of Appeals
(Philadelphia) regarding the U.S. EPA's decision to allow post
trial burn operation (one petition challenging 04/12/93 decision
and a second petition challenging EAB's 06/12/93 decision).
07/27/93 WTI holds public information meeting regarding Class 2 permit
modifications.
08/05/93 WTI begins 3-day ECIS performance test.
08/05/93 WTI submits Class 1 permit modification request to add Von Roll
America, Inc. ("VRA"), as an owner of the facility.
08/11/93 U.S. EPA authorizes WTI to go back into operation based on
preliminary ECIS test results.
08/24/93 U.S. EPA approves Class 1 permit modification adding
Von Roll (Ohio), Inc., as an additional operator; announces a
tentative decision and public comment period regarding adding VRA
a* owner; and files an enforcement action for failure to noti*y of
operator change and certain minor storage violations.
10/06/93 Court grants motion to transfer Third Circuit appeals to the 3.C.
Circuit.
10/20/93 U.S. EPA accepts revised interim feed rate limits for toxic
carcinogenic metals.
10/28/93 U.S. EPA approves Class 2 permit modification requests for
(1) modified trial burn test condition 2; (2) adding seven
listed K-serics wastes; and (3) permanent operation of ECIS.
Proposed additions of labpacfcs to Waste Analysis Plan is not icted
on.
10/28/93 Court grants motion to consolidate Third Circuit appeals wiw» 3.C.
Circuit appeals.
11/19/93 Sixth Circuit Court of Appeals overturns the 03/05/93 Clevelr*
District Court preliminary injunction.
-------
APPENDIX B
REVIEWERS
B-l
-------
-------
£
LU
U.S. Environmental Protection Agency
Technical Workshop on
WII Incinerator Risk Issues
Holiday Inn Capitol
Washington, DC
December 8-9,1993
Reviewers
Hlmar Altwicker
Professor
Department of Chemical Engineering
Rensselaer Polytechnic Institute
110 Eighth Street
Troy, NY 12180
518-276-6927
Fax:518-276-4030
Eula Bingham
Professor
Department of Environmental Health
University of Cincinnati
3223 Eden Avenue (056)
Cincinnati, OH 45267-0056
513-558-5728
Fax: 513-558-4397
James Butler
Environmental Health Scientist
Argonne National Laboratory
University of Chicago
9700 South Cass Avenue (EAD/900)
Argonne, IL 60439
708-252-9158
Fax: 708-252-4336
Walter Dabberdt
Manager, Surface & Sounding Systems Facility
National Center for Atmospheric Research (NCAR)
3450 Mitchell Lane
Boulder, CO 80307
303-497-8819
Fax: 303^97-8770 • .
Mary Davis
Professor of Pharmacology & Toxicology
Robert C. Byrd Health Sciences Center
West Virginia University
One Medical Center Drive
Morgantown,WV 26506-9223
304-293-3414
Fax: 304-293-6854
Barry Dellinger •.- ;
Head, Department of Environmental
Science & Engineering
University of Dayton
300 College Park
Dayton, OH 49469 .
513-229-2846
Fax:513-229-3433
George Fries
Research Animal Scientist
Agricultural Research Service
U.S. Department of Agriculture
Building 201 - Room 4
Beltsville Agricultural Research Center
Beltsvffle, MD 20705
301-504-9198
Fax:301-504-8438
Thomas Gasiewicz
Associate Professor
Department of Environmental Medicine
University of Rochester
575 Ehnwood Avenue
Rochester, NY 14642
716-275-7723
Fax:716-256-2591
Halstead Harrison
Associate Professor
Atmospheric Sciences (AK-40)
University of Washington
Seattle, WA 98195
206-543-4596
Fax: 206-543-8308
PimKosahvat
Senior Scientist
KBN Engineering and Applied Science
1034 Northwest 57th Street
Gainesville, FL 32605
904-331-9000
Fax: 904-331-3368
-------
TbomasMcKone
Senior Scientist
Lawrence livenaoie National Laboratory
University of California
7000 East Avenue (1^453)
LJvcmoie,CA 94551
510-422-7535
Fax:510-423-6785
Randy Seeker
Senior Vice President
Energy and Environmental Research Coipoiation
18 Mtsoo Street
Irvine, CA 92718
714-859-8851
Far 714-859-3194
M.Wfl«m Tabor
Associate Professor and Director of EAC
Institute of Environmental Health
University of Cincinnati Medical Center
3223 Eden Avenue
Cincinnati, OH 45267-0056
513-558-0515
Fax: 513-5584397
-------
APPENDIX C
AGENDA
C-l
-------
-------
£
111
U.S. Environmental Protection Agency
Technical Workshop on
WTI Incinerator Risk Issues
Holiday Inn Capitol
Washington, DC
December 8-9,1993
Agenda
Wednesday, December 8
7:30AM
8:30AM
8:40AM
8:55AM
9:10AM
9:30AM
9:50AM
10:10AM
10:30AM
10:45AM
Registration and Onsite Check-In
PLENARY SESSION
Workshop Chair:
Dr. Eula Bingham
University of Cincinnati
Welcome
Dr. Dorothy Patton, U.S. EPA, Risk Assessment Forum
Issues and Objectives/Preliminary Peer Review Comments
Dr. Bingham
EPA Risk Assessment Activities for the WTI Incinerator
Dr. Harriet Croke, U.S. EPA, Region V
Combustion Engineering
Dr. Barry Dellinger, University of Dayton Research Institute
Meteorology/Air-Dispersion Modeling
Dr. Walter Dabberdt, National Center for Atmospheric Research
Exposure Assessment
Dr. Thomas McKone, Lawrence Livermore National Laboratory/
University of California
Toxicology
Dr. Mary Davis, West Virginia University Medical Center
BREAK
Work Group Break-Out Sessions
(over)
-------
12KWNOON
IrOOPM
3JOPM
3:45PM
4:15PM
5.-OOPM
LUNCH
Work Group Break-Out Sessions (con't)
BREAK
Plenary Session
Dr. Bingham
Information Exchange
Work Group Chairs
Observer Comments
ADJOURN
Thursday, December 9,1993
8:OOAM Work Group Break-Out Sessions (con't)
Plenary Session:
Work Group Chairs' Reports and Discussion
Dr. Bingham
9:30AM Combustion and Engineering
Dr. Dellinger
IChOOAM Meteorology/Air-Dispersion Modeling
Dr. Dabberdt
10:30AM BREAK
Plenary Session:
Work Group Chairs' Reports and Discussion (con't)
10:45AM Exposure Assessment
Dr. McKone
11:15AM Toxicology
Dr. Davis
11:45AM Closing Plenary Session: Summary
Dr. Bingham
12:30PM ADJOURN
-------
APPENDED
WORK GROUP BREAK-OUT ASSIGNMENTS
D-l
-------
-------
Technical Workshop on
WTI Incinerator Risk Issues
WORK GROUP BREAK-OUT ASSIGNMENTS
WorkGroup: COMBUSTION ENGINEERING
Barry Bellinger, Chair
Elmar Altwicker
Randy Seeker
EPA Resource Person: Clyde Dempsey
Work Group: METEOROLOGY/AIR-DISPERSION MODELING
Walter Dabberdt, Chair
Halstead Harrison
EPA Resource Person: Pamela Blakely
WorkGroup: EXPOSURE ASSESSMENT
Thomas McKone, Chair
James Butler
George Fries
Marvin Tabor
EPA Resource Person: John Schaum
WorkGroup: TOXICOLOGY
Mary Davis, Chair
Eula Bingham
Thomas Gasiewicz
Pirn Kosalwat
EPA Resource Person: 'Linda Birnbaum
-------
-------
APPENDIX E
BACKGROUND INFORMATION FOR PEER REVIEWERS
AND
CHARGE TO PEER REVIEWERS
E-l
-------
-------
U.S. Environmental Protection Agency
Technical Workshop on
WTI Incinerator Risk Issues
December 8-9, 1993
Background Information for Peer Reviewers
and
Charge to Peer Reviewers
-------
BACKGROUND INFORM&TION FOR PEER REVIEWERS
WTI Risk Assessment
EPA risk assessment activities for the WTI incinerator
consist of two phases. In 1992, EPA Region V conducted a risk
assessment before authorizing interim operations of the WTI
incinerator (Phase I). This risk assessment, which was developed
by a contractor, focused on direct exposure through the
inhalation pathway in accordance with EPA's Office of Solid Waste
guidelines. Early this year, EPA's Office of Health and
Environmental Assessment conducted an additional screening level
analysis, with preliminary risk estimates for four exposure
scenarios, each of which included indirect exposures through the
food chain.
For Phase II, which Region V expects to complete next
spring/ EPA has undertaken additional studies to evaluate
exposure pathways. Again working with contractors,
meteorological data collected over a one-year period from the WTI
site will be utilized, as well as emissions monitoring data from
the recent trial burn and subsequent performance test. These
data will be used to perform air dispersion and deposition
modeling for emission constituents. Other site-specific data
related to exposure are being collected for. the risk assessment
so that it will reflect the characteristics of the site to the
extent possible.
Peer Review
The Agency has established a special two-part process for
external peer review of Phase II risk assessment activities.
On December 8 and 9, EPA will hold the first of two external
peer reviews of Phase II risk assessment work. For this December
meeting, EPA is asking several scientific experts to comment on
the draft plan for conducting the risk assessment. EPA will use
the expert recommendations to complete the plan and conduct the
risk assessment. At the second meeting, currently projected to
be held next spring, EPA will seek expert peer review of the
draft risk assessment report.
For each meeting, EPA will convene a peer review panel of
independent scientists from the fields of toxicology,
environmental fate and transport, combustion engineering,
atmospheric modeling, and exposure assessment. These experts
will focus on scientific data, methods, and analyses, along with
the related assumptions and uncertainties that describe the risk
at the site.
(over)
-------
_ 2 —
Scope of the Peer Review
Although the WTI incinerator presents many diverse issues,
this peer review is sharply focused on science-based information,
analysis, and recommendations that the Agency can use to complete
the draft plan so that the resulting risk assessment reflects
sound scientific principles. With this in mind, several
limitations are important.
First, the project plan and risk assessment relate to
scientific and technical issues only. Larger issues involving
past and future permitting decisions, which are based in part on
the risk assessment and in part on other considerations, will be
examined after the April peer review.
Second, EPA does not expect the peer reviewers to
reach consensus on all issues raised in the workshop. Rather,
the purpose of this workshop is to collect expert opinions on
reasonable approaches to some difficult and potentially
controversial scientific issues.
Finally, EPA does not expect to resolve all uncertainties in
the data and methods that will be used for this assessment.
However, with the help of the peer reviewers, EPA expects to
identify the most important areas of uncertainty, useful options
for addressing the uncertainties, and the expected impacts of
these uncertainties on the planning and conduct of the risk
assessment. In this regard, while additional long-term research
may well be needed to resolve some questions, EPA is seeking
practical, readily implemented suggestions for completing the
project plan.
Premeetinq
As a reviewer for the draft project plan, EPA is asking you
to use your technical knowledge and your best professional
judgment to comment on the project plan as a foundation for the
WTI Risk Assessment. While we invite you to comment on any
technical aspects of the project plan, we encourage you to focus
on the questions outlined in the enclosed "charge" (attachment
4).
-------
CHARGE TO REVIEWERS
FOR THE WTI RISK ASSESSMENT PROJECT PLAN
The WTI facility has attracted considerable attention at a
variety of levels. Our concern in this exercise is restricted to
the' technical level of those concerns. The Agency's goal is to
generate an assessment of possible risks associated with the
operation of the WTI facility that is as complete as is
reasonably possible within time and resource constraints and
which will provide a sound technical basis for making the risk
management decisions called for in this case.
As a reviewer of the WTI risk assessment project plan, you
should use your best technical knowledge and professional
judgment to comment on your- expectations for completeness and •
reasonableness in the assessment that results from this plan.
Your comments will be considered in revising and carrying out the
plan.
This charge has two parts. In the first part, EPA is asking
for general impressions of the plan, giving special attention to
the Agency's schedu'le for completing this risk assessment. In
the second part, EPA is asking each reviewer to focus on several
specific issues in his or her area of expertise with comments on
other areas invited but optional.
Part I. EPA plans to complete the risk assessment by April to
meet regulatory timetables.
Please comment on features of the plan expected to lead to a
strong risk assessment.
Please comment on those features that can be fixed within
the available timeframe.
Please comment on any issues that ideally - without time and
resource constraints - would be useful to consider in future
'assessments of this type.
Part II. While we invite you to comment on any and all technical
aspects of the plan and how it might be improved, we are
particularly interested in your comments (and associated
rationale) on the following issues:
Emissions Characterization
Emissions Characterization includes identification of
substances of concern and the development of emission rates for
these contaminants.
-------
Particulates and a suite of metals, acid gases, and
dioxins/furans have been selected for quantification.
Would you recommend any additions or deletions to this
list (pp. 29-30)?
Three criteria - quantity, toxicity and
bioaccummulation potential - have been proposed for use
in selecting surrogate organic substances for
quantification in the risk assessment. Please comment
on the adequacy of these criteria and the manner in
which they will be applied to the master list of
substances of potential concern to yield a final list
for quantification in the risk assessment .(pp.. 31-34) .
Estimation of chemical emission rates will be based on
the data from the trial burn and performance test, data
from similar rotary kiln incinerators and/ where
appropriate, data from other hazardous waste
incinerators and other types of combustion units.
Please discuss your view of extrapolating and combining
these data to estimate emission rates(pp. 34-44).
Several emission scenarios (e.g., routine, day-to-day,
operations; startup/shutdown process upsets; and
accidents) have been identified.. Please comment on
whether these scenarios sufficiently bracket the range
of possible emission situations. In your view, does
the emissions characterization reflect the wide range
of operating conditions over the useful life of the
facility (pp. 38-39, 93-96)?
A critical aspect to estimating indirect exposures is
the vapor phase/particle partitioning. Please comment
on the approach to partitioning emissions between the
vapor and particulate phases (pp. 44-45).
Hazard Identification and Dose-Response Evaluation
In the risk assessment, both carcinogenic and non-
carcinogenic health effects will be evaluated. Please comment on
the issues listed below: ' •
1. the planned use of dose-response data in the assessment
and any issues of chemical form and speciation (pp. 47-
54) ,-
2. the chronic and acute toxicological endpoints that
should be selected in light of what is known about
emissions from such facilities (pp. 48-51);
-------
the planned use of the Uptake Biokinetic model for
characterizing lead toxicity (p. 54);' and
the assessment of human health effects, but not
ecological effects.
Exposure Assessment
Exposure assessment involves identifying the potentially
exposed populations and quantifying the magnitude, frequency, and
duration of exposure.
Exposed Populations
1. An approximate distribution of individual risk
estimates will be developed based on site-specific
information concerning the location of specific sub-
groups with respect to the WTI facility, typical
activity patterns for each sub-group, and the number of
individuals that comprise each sub-group. Please
comment on the use of this approach for developing
central tendency and high-end estimates of exposure
vis-vis EPA's Exposure Assessment Guidelines (pp. 55-
56).
2. To further characterize the variability in exposure,
the plan proposes to further examine the high-end sub-
groups in each concentration isopleth by combining
typical and high-end exposure factors. Please comment
on this approach (pp. 56-58).
3. The study area will initially be defined based on a
cancer risk level of less than 10'7 estimated from the
consumption of beef containing dioxins/furans. Please
comment on the approach to delineating the study area
from a technical point of view (pp. 58-59).
4. The plan identifies several population sub-groups that
will be the focus of the risk assessment. Please
comment on the use of these sub-groups to quantify the
potentially exposed population around WTI (pp. 59-60) .
5. Are there additional exposure routes for any of the
population sub-groups that need to be been considered
(pp. 60-63)?
6. The plan will characterize the exposure distribution
through point estimates of central tendency and high-
end exposure situations. We recognize that other
approaches are possible for characterizing variability
and uncertainty, e.g., Monte Carlo simulation. Please
-------
comment on whether Monte Carlo simulation should be
attempted, given the nature of the data to be collected
and the resources that would be required. If so, what
aspects of the risk assessment should be the focus of
the analysis?
Exposure Concentrations
_ The estimation of exposure concentrations will be performed
using fate and transport models that simulate the transport of
stack gas constituents in the environment and the resulting
concentrations in various media.
1. The COMPDEP atmospheric dispersion model will be
applied to the WTI facility. Please comment on the
selection of this model (pp. 64-67).
2. Please comment on the approach used for characterizing
the pollutant distribution (surface versus volume
distribution) identified within the project plan
(p. 65).
3. For very small particles (less than .05 micron
diameter)., the deposition velocity increases with
decreasing diameter. Since the final outpoint for the
particle size data collected at WTI was 0.4 micron, a
sensitivity analysis will be performed to characterize
the total mass of particles reported to be less than
0.4 micron (33%). The sensitivity test will assume
that the total mass of particles below the 0.4 micron
diameter will be modeled to represent .03 micron
diameter particles (to account for the increase in
deposition velocity). Please comment on this approach
(p. 66) .
4. One year of on-site meteorological data will be
supplemented with data from the nearest National
Weather Station. Please give your view on the approach
for predicting long-term meteorological conditions
around the WTI facility (pp. 67-68).
5. Please_comment on the various modeling equations used
to estimate the concentrations of contaminants in the
different media and their applicability to the
different chemical classes to be addressed in the risk
assessment (pp. 71-83; Appendix B).
Dose Estimation
Please comment on the specific dose procedures for
estimating the "average daily dose" and the proposed
values for the exposure factors. Each of these factors
-------
has a range of possible values based on the population
sub-groups. Please comment on the values selected as
representative of typical and high-end exposure for
each of the sub-groups (pp. 85-91; Appendix A).
Bymltiation of Routine Fugitive and Short-Term Episodic Emissions
Please comment on the approaches that will be followed
to evaluate other than routine emissions from the stack
and how these estimates should be integrated with the
risks from routine operations (pp. 93-96).
Risk Characterization.
Risk Characterization -involves .integration of toxicity and
dose-response information with exposure estimates and provides an
evaluation of the overall quality of the assessment, including a
prominent display of critical uncertainties.
1. Two different approaches are proposed for estimating
population risk- one procedure for non-food chain
pathways-and another for food chain pathways. Please
comment on these approaches (pp. 97-101).
2. Please comment on the approach to characterizing
uncertainty (qualitatively as well as quantitatively) .
What additional steps are available to make the
uncertainty analysis more quantitative (pp. 101-105)?
-------
APPENDIX F
PREMEET1NG COMMENTS
F-l
-------
-------
United States
Environmental Protection Agency
Technical Workshop on
WIT Incinerator Risk Issues
Premeeting Comments
Washington, DC
December 8-9,1993
-------
ORDER OF PREMEETING
COMMENTS
Elmar Altwicker
Eula Bingham
James Butler
Walter Dabberdt
Mary Davis
Barry Bellinger
George Fries
Thomas Gasiewicz
Halstead Harrison
Pirn Kosalwat
Thomas McKone
Randy Seeker
Marvin Tabor
-------
Elniar Altwicker
-------
-------
Altwicker
Comments on III. Emissions Characterization, p. 27ff
The distinction (if any) between organic residues and PICs (p. 29, Table 7) is not
clear. Since the list includes chemicals detected in the March 1993 trial burn (p.
28) it would have been useful to have the actual trial burn results (what were the
sampling/detection limitations). It seems a bit arbitrary to include certain
chlorobenzenes and not others, whereas Polychlorinated Biphenyls (PCBs) were
included apparently generically, though surely not all PCB isomers were detected.
A total of only three PAHs (polynuclear aromatic hydrocarbons) is surprising.
I suggest the inclusion of Cu and Fe in the metals portion of the table (p. 30) ;
the two metals and their salts may play a role in dioxin/furan formation (PCDD/F).
Some of the PICs may be POHCs (primary organic hazardous compounds) so the trial
burn basis should have been clearly stated. On page 35, the composition (but not
relative proportions) is given as carbon tetrachloride, monochlorobenzene ,
trichloroethene, and 1,2,4-trichlorobenzene, but not the actual DREs (destruction
and removal efficiencies); but monochlorobenzene could give 1,2,4-trichlorobenzene
(and vice versa).
Particulate matter (paragraph 2, p. 31; Table 7) is to be quantitatively evaluated in
the risk assessment (what criteria will be used?: mass, size distribution, species
distribution, % carbon, etc.).
Quantity, toxicity, and bioaccumulation potential (p. 31) are to be the basis for the
risk assessment. What about photochemical smog potential and visibility
(especially the latter in light of the apparent level of particle emissions, their size,
and the location of the facility, p. 3. I will return to these points later).
The operating conditions of the trial burn should be provided (p. 35/36). At least
four variables were varied for three operating conditions; was a statistical design
used? Details of the variability of the measurements (p. 37) were not provided.
Some of the compounds in Table 8 (p. 36) clearly are PICs; the sentence "other
organic emissions, such as products of incomplete combustion, were not measured"
makes no sense, therefore, rlt seems to be implied that there were other organic
emissions, in addition to those listed in Table 8, and that they might have been
significant.
-------
Altwicker
A summary of the most important findings from the HIM-facility, Biebesheim,
Germany, before and after "... significant modifications ..." should have been made
available. Among other things the sampling locations are clearly identified in that
report. How are "past published historical emission data" to be assessed (p. 38)?
Were the operating conditions of the performance tests after installation of the
enhanced carbon injection system (p. 40) comparable to those used in previous trial
burns? Why is this system is not shown in Figure 4 and 5? The tabular results in
Figure 5 presumably pertain to the carbon injection system in operation (cf.
particulates). It is between the spray dryer and the electrostatic precipitator (ESP).
PCDD/F-sampling was conducted at the same sampling point?
The discussion on PCDD/F emissions could have been more concise. On p. 40 total
TEQ rates are given in ng s'1; later total PCDD/F are given in concentration units. No
indication is given of the congener profiles; are they "typical"
combustion/incinerator profiles? What was the PCDD/PCDF ratio? The phase
distribution?
It needs to be noted that the HIM-facility, Biebesheim reports higher PCDD/F
concentrations after the boiler upon carbon injection ahead of the boiler (factor 2-
3), though stack concentrations are lower by about the same factors, suggesting both
participation of carbon in formation as well as increased particulate adsorptive
capacity.
On page 42 (under iii.) an aerosol mechanism is alluded to; this may play a role in
PCDD/F formation. If I read the data in Figure 5 correctly, 0.01 s scf'1 of particulate
matter (v., p. 43/44) is emitted. This represents 99.16% efficiency (15 vs. 21, Figure
5), or 99.8% (16 vs. 21, Figure 5). However, using -30 scf nr3, as a conversion factor
0.01 g scf'1 is equivalent to 3 gnr3, rather high emissions. The total available
surface area of that particulate matter may make the subsequent discussion on
partitioning (p. 44) moot.
One gets the impression from reading the opening paragraph to Section 4 (p. 44)
that emitted constituents will partition subsequently. This clearly cannot be the
case. On the next page, this impression is dispelled somewhat. What is the basis for
the assumption (second paragraph on p. 45) that metals are homogeneously
dispersed throughout the entire particle? Documentation is lacking.
On page 66 it is stated that a significant fraction of the total particle mass (about
1/3) is <0.4 u.m in diameter. Assuming that this is equal to 1 g nr3, a spherical
-------
Altwicker
particle (0.4 u.m), and a density of 1.4 g cm'3 (Biddleman, 1988) a total surface area
of 0.11 cm^ cm"3 can be calculated. This is four orders of magnitude greater than
the highest value cited for urban air particles, 1.1 x 10'^ cm^ cm"3 (p. 44).
Presumably, these particles, after emission, would rapidly cool to ambient
surroundings prior- to significant plume dilution (it should be possible to calculate
the rate of cooling and compare it to the rate of dilution). It seems, therefore, that
a thorough understanding of the nature of the stack particles and their immediate
behavior downstream is more important than the analysis proposed on p. 44/45.
Are these particles not mostly carbon? Their surface area is likely to be much higher
than calculated above. Fly ash surfaces can range from 1-15 m^ g-1, while carbon
pockets may have surface areas 0(100) m^ g-1, and activated carbons can be much
higher. Swedish investigators recently furnished us with a fly ash cut (< 1 u.m) that
has a surface area of 5000 m^ g-1.
Though the temperature at which the ESP operates is lower than that used in
laboratory studies of PCDD/F-formation, carbon is nevertheless a PGDD/F-precursor
and -potential slow buildup should be monitored (cf. above).
Biddleman (1988) gives
PL
In -± = ASf (Tm - T)/RT
where
?? = vapor pressure of the sub-cooled liquid = P! (p. 45)
30 _
= crystalline solid vapor pressure = Pc (p. 45)
Therefore, the equation on top of page 45 should read
,
ln
Biddleman (1988) cites Asf/R = 6.79 as an average and gives three references; he does
not say that it can be satisfactorily estimated. In fact, Mackay, et al. (1982) state
that it is an average empirical value and "may be substantially in error for certain
compounds." If precise partitioning between phases becomes important for certain
compounds from the risk assessment, it may be useful to iterate on this value. If PI
-------
Altwicker
(or P£) replaces P* for dioxin (TCDD, P§ = 5 x 10'13 atm), the value would be 745 x
P* (Biddleman, 1988) which is still small compared to c x ST (1.7 x 10'4 atm cm x
0.11 cni^ cm'3), where ST is the particle surface area per unit volume (p. 44). Thus,
almost all TCDD (and PCDD/F) would be expected to be on particles when leaving
the stack or shortly thereafter; interaction with the "general atmosphere" would not
be required. This analysis is not contradicted by apparent gas phase PCDD/F when
sampling. Desorption and readsorption can occur. The Biddleman analysis
apparently has the assumption of uniform surface site distribution built into it. It
is more likely that adsorption would be described by a Freundlich isotherm unless,
as noted above, the emitted particles are mostly carbon.
The WIT-stack is less than good engineering practice (GEP), p, 66. Shouldn't that be
corrected?
The section C, Releases from On-Site Accidents (p. 95), could be expected to include
more specific considerations of failure or malfunction modes of the air pollution
control units, i.e., if carbon injection system is out-of-service, can the ESP be
operated differently. What are the rapping cycles of the three fields, how readily are
they modified? In other words, can a probable scenario of malfunctions be
prepared? Actual accidents would be only one consideration.
How important is the nature of the carbon injected in the efficiency of the
adsorption process? Its prolonged storage? Again cf. above (HIM-facility results).
In light of the discussion on the low end of the particle size distribution (p. 66) the
strategy outlined on p. 4, item 3., Exposure Concentrations (attachment 4, Charge to
Reviewers) seems a bit arbitrary. Why not use several diameters below 0.1 nm or an
assumed distribution?
-------
CORRECTION TO PREMEETING COMMENTS BOOK
Elmar Altwicker has provided us with the following correction to a calculation on Page 3 of his
premeeting comments:
0.01 g scf-i x 30 scf nr3 = 0.3 g nr3 and
not 3 g nr3
Consequently, 1/3 equals 0.1 g nr3
Assuming a surface area of 3 m2g-t (an average value for MSWI-fly
ash; it is probably that hazardous waste incinerator ash would be
higher) one can calculate »
0.1 g
m
3 m2 • • _ m2 . cm2
III »-
-T = 3 x 10-3
_ — «••> ™1 = J A 1U -* 7
g mj cm3
Compared to Biddleman's urban value of
cm2
1.1 x 10-5
cm3
i.e., the surface area available on particles leaving the stack is at least
100 x greater than that from "average urban particles".
This correction also affects the calculation shown at the bottom of
page 04, but not the conclusion.
-------
En I a Bingham
-------
Eula Bingham
Comments on WTI Risk Assessment Project Plan
My review of the WTI Risk Assessment Project Plan raises the
following issue that I do not see addressed:
1. EPA guidance documents were not provided to the reviewers, how were
these prepared and what was the peer review?
2. The Dioxin document needs to be finished and the interim document
should be shared with WTI and reviewers of WTI Risk Assessment Plan.
3. Is WTI contacting German and Swedish regulatory authorities re any
data or health complaints?
4. Where are the WTI incinerators located in Germany and Sweden? It is
very vague in plan.
5. When the models of wastes are used what will be the theoretical
sources? I am particularly concerned with decreasing and plating
operations.
My reaction to the plan is that it only seeks to respond to a legal
mandate using EPA guidelines. My greatest concerns are that EPA in its
guidelines for risk assessment has many areas where there are no
toxicological data and inadequate toxicological data on possible target
organs. To be very specific the air pollution aspects of this project
appear deficient. Most risk assessments are geared toward water and
ingestion. The contaminants and combinations of contaminant toxicology
via inhalation is largely unknown for the emissions.
Are there any data from human populations exposed to such emissions
that provide data as to eye or nasal irritation, increases in upper
respiratory and respiratory problems, asthma? There are investigators,
many of them physicians who should help review the plan and offer advice.
EPA guidelines for risk assessment do not address such effects on human
populations and these adverse health effects are not ordinarily Tisted in
the data bases. Drs. Carl Shy and Frank Speizer are two possible
contacts.
Page 101 provides a pertinent example of my great concern over this
plan. The description notes "a paucity of toxicological data" and a few
lines below "by discussing the limitations of the toxicity data
presenting the greatest risk".
How will you know? It is probably the dioxins but thats a guess!
-------
Eula Bingham
The plans for carcinogenesis risk assessment follow EPA guides as
far as I can determine. These risk assessments should include nickel. It
is a carcinogen (IARC, 19 ). The burning will most likely result in
nickel oxides or other nickel compounds depending on what is in the mix.
-------
James Butler
-------
James P. Butler
COMMENTS ON THE WTI RISK ASSESSMENT PROTECT PLAN
Based on my review of the WTI Phase H Risk Assessment Project Plan (11/93
draft), the Charge to Reviewers, and selected background papers, I can offer the
following comments on technical aspects of the proposed methodology. Comments are
organized into three main areas: Parts I and n correspond to the Charge to Reviewers,
including headings and comment numbers, and Part ffl contains general comments and
recommendations about scientific issues that were not addressed in the charge.
PARTI
My overall impression of the project plan is that it is a well organized and clearly
written document. Although the plan has some inconsistencies and omissions
(discussed below), in general it presents a comprehensive approach for evaluating
human health risks. The level of detail in most sections is appropriate for illustrating
the risk assessment methodology.
There are a number of features of the project plan that should strengthen the risk
assessment approach. Clearly, the planned use of extensive site-specific data for facility
emissions, population exposures, and meteorological conditions will contribute greatly
to increased confidence in the risk assessment results. Evaluating acute health risks
from short-term emissions (accidents, spills, fire, equipment malfunctions, fugitive
emissions, etc.) is a valuable extension of the risk assessment methodology.
Features that can be fixed within the available time frame are highlighted in the
individual comments in Parts II and in.
Without tune and resource constraints there are several key issues that I would
consider for a more detailed analysis in the future: long-range transport, ecological
impacts, Monte Carlo analyses, and model validation. The significance of these issues
is discussed below in the individual comments sections.
PART II
Emissions Characterization
1) The preliminary list of contaminants of concern appears to be comprehensive. The
only addition to the list that I would recommend is benzo(a)pyrene (BaP). As a
byproduct of incomplete combustion, BaP has been detected in incinerator emissions,
-------
James P. Butler
albeit at relatively low concentrations. Because it is classified as a probable human
carcinogen, extensive data exist on BaP toxicity and environmental levels. For these
reasons, it is appropriate to include BaP in the preliminary list of organic PICs, using
an estimated emission rate.
2) The selection of surrogates for organic compounds (excluding dioxins and furans)
is proposed based on emission rates, toxicity, and bioaccumulation potential. This is
an adaptation of the commonly used concentration/toxicity screening procedure.
Because of the importance of food chain pathways, it is appropriate to extend the
screening procedure in this manner. Although the approach appears to generally be
sound, several modifications are recommended. The arbitrary cutoff of 10 chemicals
with the highest scores should be somewhat flexible. It is reasonable to consider
extending the list to include chemicals with "borderline" scores that may have other
important characteristics, e.g., high mobility or persistence in the environment. While
the list of criteria could easily be extended, the screening procedure risks becoming
unwieldy and somewhat contrived. Flexibility in the organic surrogate selection process
is a compromise that will also ensure that a potentially significant chemical will not be
eliminated unless its score is considerably lower than the top ten, not just under the
cutoff. Another recommendation is to pick the more conservative QCB scores for each
organic compound on the basis of K^ or BCF (if available). The bioaccumulation
potential of some chemicals in fish could be as significant as in meat and milk.
5) My only comment on the approach to partitioning between the vapor and
particulate phases stems from the statement (p. 45) that all metals will be assumed to
be entirely in the particulate phase, which "...is likely to significantly overestimate the
deposition rates for mercury..." While there is some controversy about how much
HgCl2 is adsorbed on particles as the flue gas is cooled, it is true that almost all mercury
leaves the stack in the vapor phase. It is also true that the assumption that mercury is
associated 100% with particulate matter (e.g., soluble forms of mercury such as HgCl2)
will overestimate deposition rates and result in conservative estimates of health effects
in the immediate study area. My concern is that long-range transport of elemental
mercury is totally ignored, although it has been reported that less than 10% of mercury
emitted from a point source is deposited locally (Lindqvist and Rodhe, Tellus 37B: 136-
-------
James P. Butler
159,1985). In addition to the potential for significant human health effects (via fish
ingestion), dispersion of mercury on a regional or global basis can contribute to adverse
ecological impacts. Although this is clearly an area for further research, the issue of
long-range transport of mercury should still be addressed in the risk assessment, at least
qualitatively. [The issue of ecological risks is also discussed in the next section,
question #4.]
Hazard Identification and Dose-Response Evaluation
1) The brief discussion about mercury speciation (p. 54) includes the point that some
of the mercury emitted from the stack of the WTI facility could be converted to the toxic
organic form, methylmercury. A statement is made that "The extent of such conversion
will be carefully evaluated in the risk assessment..." This is too vague to offer any
assurance that the complex process of methylation of inorganic mercury in the
environment will be adequately evaluated. Simply estimating mercury loadings in a
particular water body is not enough to accurately predict levels in fish; there are many
physical, biological, and chemical factors that play a role in mercury uptake and
bioavailability. An extensive review of the literature is necessary, preferably coupled
with some modeling (e.g., EPRI is currently funding model development for aquatic
impacts of mercury emissions from utilities).
2) The toxic endpoints for methylmercury (Me-Hg) were recently reviewed by Stern
(Risk Analysis 13: 355-364,1993). The current RfD for Me-Hg is 0.3 ug/kg/day. This
value, apparently under review by EPA, is based on the lowest observed effect level
(LOEL) dose for paraesthesia. Stern offers a convincing analysis suggesting that the
current RfD does not provide an adequate level of protection against in utero
developmental effects, and that a significant fraction of women of childbearing age may
be exposed to Me-Hg at levels exceeding a reasonable margin of safety for adverse
developmental effects. Consequently, I strongly encourage the use of a RfD for Me-Hg
of 0.07 ug/kg/day based on a developmental endpoint.
4) To focus exclusively on human health effects and ignore ecological effects would be
a weakness in any risk assessment. This is the case for the WTI risk assessment as well.
The. WTI facility is in an area with many water bodies and large tracts of land reserved
-------
James P. Butler
for state parks and game lands. Unfortunately, there is limited guidance on conducting
ecological risk assessments. In addition, much of the data needed to evaluate ecorisks
would not be available until field surveys and sampling were performed. Because of
the schedule, conducting a comprehensive ecological risk assessment is not practical.
In lieu of this, it would be useful to at least identify any sensitive ecosystems and
evaluate effects in a qualitative manner by reviewing information that is readily
available (similar to a NEPA Environmental Assessment or Environmental Impact
Statement).
Exposure Assessment
Exposed Populations
1) The approach to developing a distribution of individual risk estimates seems to be
consistent with EPA's Exposure Assessment Guidelines. Conceptually, I agree with the
approach to developing central tendency and high-end (upper 10%) exposure
distributions to avoid focusing on just "worst case" estimates, provided that all highly
sensitive population subgroups are included in the analysis.
2) A sensitivity analysis to estimate exposures for the more highly exposed individuals
in high-end population subgroups is a reasonable approach to characterizing the
variability in exposures.
3) The rationale for defining the perimeter of the study area on the basis of dioxins and
furans in locally-raised beef is that this was the most significant exposure pathway
identified in a "screening level assessment." However, it is my understanding that this
screening level assessment focused on only dioxins and furans. While I agree that this
may indeed be the most important pathway, this preliminary conclusion should be
confirmed by looking at several other contaminants/pathways. Additional "driving"
exposure pathways could include mercury intake through subsistence fishing or lead
exposure from soil ingestion by children. Therefore, I recommend that several more
bounding estimates be developed for noncarcinogens to complement the dioxin
screening analysis. While the perimeter of the study area may well remain the same,
this should be confirmed.
-------
James P. Butler
4) The list of population subgroups included in the exposure assessment covers most
of the potentially exposed individuals around the WIT incinerator. If there are any
hospitals or nursing homes in the study area, a "special population segment" should be
added for sensitive individuals. For example, exposure to acidic aerosols and gases
could present a greater risk for individuals with chronic respiratory diseases. I would
also consider additional pathways for subsistence fishermen. This could be a significant
group that should not have any potential exposures underestimated. It is unlikely that
fish consumption is the only food chain exposure pathway of importance for
subsistence fishermen. All the other adult subgroups include at least two of the three
food chain pathways (meat, milk, vegetables). Likewise, it seems reasonable that
subsistence fishermen could also obtain milk from locally raised cows and vegetables
from home gardens.
5) The groundwater exposure pathway is not included in the project plan. As pointed
out in the Addendum to Methodology for Assessing Health Risk Associated with Indirect
Exposure to Combustor Emissions (EPA, 1993), the conclusion that incinerator emissions
do not impact groundwater was based on an analysis of two combustors, which may
not be valid for other combustors having different conditions. While it seems unlikely
that this is a significant pathway, relevant site-specific conditions (e.g., recharge rates,
contaminant solubilities, etc.) should be evaluated to confirm that they are similar to
conditions in the original groundwater analysis (EPA/600/6-90/003,1990).
6) A Monte Carlo probabilistic analysis of exposure and fate/transport models would
provide more information about overall exposure distributions. However, the validity
of an uncertainty analysis using Monte Carlo simulations depends on the probability
distributions for trie input parameters and sufficient data are often not available to
adequately develop these distributions. Given the number of models, it would also be
a time consuming task to compile comprehensive parameter distributions for the WTI
risk analysis. Because a sensitivity analysis will be performed on highly exposed
subgroups, the use of point estimates of central tendency and high-end exposures may
be adequate for this risk assessment. However, in the future a risk assessment of this
nature would be strengthened by including probability distributions for key intakes and
concentration estimates.
-------
James P. Butler
Exposure Concentrations
4) Wet deposition is an important removal mechanism for certain forms of mercury
(e.g., HgCl2) in the atmosphere. The use of hourly data for precipitation type and
intensity should improve the estimates for washout of mercury in the study area.
However, it is important to use precipitation data collected locally or from a site with
similar rainfall patterns. Because the project plan calls for Pittsburgh Airport
precipitation data to be used, it is important to compaVe precipitation patterns. In
addition, when modeling a source that would be operating for many years, it is
preferable to use precipitation data that were also collected over a number of years.
5) My only comment about the fate and transport models relates to the estimation of
breast milk concentrations. The point that metals are not expected to accumulate in
mother's milk should not be used to rule out mercury which can be converted to the
lipophilic organic form, methylmercury.
Dose Estimation
I recommend revisiting the exposure assumptions for fish ingestion by subsistence
fishermen. The typical and high-end values for fish ingestion rates may be too low in
light of a recent EPA study (under review) of a fish-eating, Native American population
along the Columbia River. For example, it was reported that the average fish
consumption rate was over 57 g/day, while the typical exposure to locally-caught fish
in the project plan is estimated to be 30 g/day. There is also a minor inconsistency to
correct: the high-end exposure duration for breast feeding was given as 270 days in the
text and 365 days in Table A-21. Without data to support either value, I would select
365 as the high-end assumption.
Evaluation of Routine Fugitive and Short-Term Episodic Emissions
In the project plan it is assumed that fugitive emissions will consist mainly of
volatile organics emitted during routine operation of the incinerator and bulk waste
handling and storage. Potential inhalation exposures to collected fly ash and bottom
ash should also be evaluated. Handling, storage, and transportation of this residual
waste to its disposal site could generate fugitive particulate emissions that present a
risk, especially to workers.
-------
James P. Butler
PART III
The following are general comments that do not directly correspond to the
questions or issues listed in the Charge to Reviewers:
1) The general approach to dose estimation involves assessing incremental intakes of
chemicals emitted from the facility. For some contaminants (e.g., mercury, lead) it is
essential to factor in estimates of existing body burdens and intakes from other sources.
The project plan needs to address the issue of background exposures in the population
subgroups, especially for compounds that are retained in the body, have relatively low
thresholds, or have other significant sources in the environment.
2) Because of the importance of food chain pathways, dioxins and furans are often
mentioned in the project plan when discussing contaminants of concern, models,
chemical properties, study boundaries, etc. In addition, there is existing guidance on
dioxin-like exposure assessments which is useful to adapt and build upon. This
comment is simply a reminder to focus the same level of attention and resources on
other important chemicals emitted from the facility. It is obvious from a number of my
comments that I believe several inorganics (mercury, lead, and possibly hexavalent
chromium) are potentially of equal or greater significance than dioxins when evaluating
incinerator emissions.
3) I agree with the discussion on P. 81 that available field data are very limited. There
are, however, some data that may be useful from a study on mercury in precipitation
samples collected in the vicinity of municipal solid waste incinerator in New Jersey
(Greenberg et al., Proceedings of the EPA/AWMA Symposium on Measurement of Toxic and
Related Air Pollutants, 1992).
4) Related to the above comment (#3), the paucity of field data makes it virtually
impossible to validate the fate and transport models that are to be used in this risk
assessment (as well as others). To verify that the models and assumptions were
reasonably accurate, it would be worthwhile to periodically monitor for key indicators
at selected locations within the study area. Examples could include: measurements of
dioxins in cow's milk, methylmercury in fish, lead in surface soils (over time). These
-------
James P. Butler
type of monitoring data would eventually help reduce uncertainties about the models
and input values.
5) A minor point: it would be helpful to have better plates. Some of the symbols on
the blueprints are not unique which makes them difficult to use. GIS-type maps with
either color or better shading/symbols would be much clearer.
-------
Walter Dabberdt
-------
November 29, 1993
Vtelter F. Dabberdt
Preliminary Comments on Atmospheric Dispersion and Deposition
Issues of the WT1 Phase II Risk Assessment Project Plan:
The proposed risk assessment (the "Plan") appears to estimate concentrations and
potential doses on the basis of annual averages, but for different gross emission
strategies. Are acute effects important; and if so, on what time scales?
Apparently, California assumes worst-case emissions last for one hour and are
uncontrolled (represented as a 100-fold increase over 'normal1 emission rates). If
an accidental release occurred under stable conditions, then the single-hour
. concentration would be roughly equivalent to 103 hours (-40 days) of normal
annualized exposure. Are dose responses independent of integration time? If not,
when are these non-linearities significant?
Several questions arise in the proposed use of the so-called COMPDEP model?
Is or will it actually be available and adequately tested for use in this risk
assessment? If it is not available, then what dispersion model(s) will be used?
COMPIex-1 is a screening model while ISC is not; are the two models compatible
for this type of assessment? How is plume rise determined? Is mixing depth
explicitly treated? One might infer it is not from the COMPDEP description in the
Addendum to Methodology for Assessing Health Risks Associated with Indirect
Exposure to Combustor Emissions* (the "Addendum"). The Plan describes the
treatment of building wake and downwash effects in COMPDEP, but the
Addendum indicates these effects cannot be considered when COMPDEP is
running in the COMPLEX-1 model mode; this apparent discrepancy needs to be
resolved in favor of building effects for all terrain types. In fact, neither ISC nor
COMPLEX-1 is capable of estimating downwash induced by topographic effects
on the ambient wind flow. The site description in the Plan indicates that local
topographic relief may be important. ISC can produce short-period peak average
+EPA/600/AP-93/003, Nov. 1993, External Review Draft
-------
November 29, 1993
Walter F. Dabberdt
concentrations for 1-, 8-, 24-, and longer periods, but COMPLEX-1 apparently
cannot; if correct and if acute exposures must be quantitatively assessed, then this
apparent intramodel conflict must be resolved. The COMPLEX-1 model cannot
simulate terrain-induced wind channeling and meso-y scale fiow curvature, yet the
discussion of the site in the Plan correctly states that the terrain features may have
a significant effect on the wind flow. This could result in significant geographic
misplacement of the concentration estimates. The Addendum indicates that
COMPDEP does not explicitly treat calm wind conditions (n hours), except to
calculate the annual average as (365x24) - n. This approximation can be especially
inaccurate if acute exposures are to be considered. The Plan indicates that
fugitive emissions are to be modeled, yet this is not reflected in the air dispersion
portion of the Plan. And lastly, the Plan indicates (p. 24) that acute on-site effects
will be considered qualitatively, while off-site acute effects will be considered
quantitatively. It would appear that the opposite treatment should be given to on-
!
and off-site effects. Further, the air dispersion portion of the Plan does not discuss
either type of acute effect.
Section VI (Evaluation of Routine Fugitive and Short-term Episodic Emissions) of
the Plan provides a good overview of the types of short-term emissions scenarios
that will be considered. However, the implication is that this will be done
qualitatively (reinforced by the lack of consideration of short-term effects in the air
dispersion section).
There is an apparent discrepancy between the Charge to Reviewers and the Plan
in the manner in which very small particles will be treated.
The uncertainty analysis portion of the Plan is vague, and mainly repeats the
criteria given in the Federal Register. In the case of atmospheric dispersion, it is
unlikely that a screening model like COMPDEP can simulate atmospheric
concentrations from a short stack adjoined by relatively tall buildings to within a
factor of two in moderately complex terrain.
-------
Mary Davis
-------
Mary E. Davis
Parti.
Features of the plan expected to lead to a strong risk assessment.
The model is comprehensive and considers multiple routes of exposure and multiple agents and
specific populations.
Features that can be fixed within the available time frame
Incorporate ambient concentrations of acid gases, to model total exposure.
For process upsets and accidents, model a likely case, not constrained by permit limits.
Estimate exposure through ingestion of locally grown deer.
Estimates of gardening practices in the plume area. Much of the area surveyed is to the
west of the facility and therefore not at as great a risk.
Count snow and ice as forms of precipitation in the COMPDEP model.
Assess the accident rate on the local roads that will be used in delivery of waste to the
plant.
Issues that ideally - without time and resource constraints - would be useful to
consider in future risk assessments of this type.
Devise an air quality model that reflects the conditions and circumstances of a steep river
valley such as those that occur throughout the Appalachian Mountain region.
Part II.
Emissions Characterization
1. Additions to list of substances to be quantitatively evaluated in the risk assessment
This is not an addition of an agent, but a concern on how acid gases will be quantitatively
evaluated in the risk assessment (pg 31). Does that evaluation consider just that which is emitted
by the WTI plant or will it consider the WTI emissions as added on to current acid gas exposures?
This part of the country receives very high amounts of acid precipitation (similar to the
Adirondack^) and so I am concerned that the increased emissions will have substantial adverse
effects on lung function if the total exposure reaches the steep part of the dose-response curve.
4. Emission scenarios
It is not clear to me how the process upset and accidents are incorporated into the risk
assessment. It seems that there are two scenarios: normal operation and abnormal events (or
-------
Mary E. Davis
maybe worst case) to provide a conservative estimate. Are the process upsets and accidents
included in this conservative estimate scenario or are there other scenarios to be evaluated? I
question this because the conservative estimate assumes that emissions will be within the permitted
limits and accidents cannot be assumed to stop themselves when the permitted limits have been
reached.
Hazard Identification and Dose-Response Evaluation
2. Use of dose-response data in the assessment
It is not clear to me what effects will be evaluated for which chemicals. Many of the
chemicals have both cancer and non-cancer effects, and cancer does not necessarily have the most
sensitive dose-response curve. Will multiple effects be evaluated for the chemicals? If not, how
will it be decided which effects will be evaluated? For instance, for 2,3,7,8-TCDD, enzyme
induction, immunotoxicity and reproductive toxicity occur at lower doses than cancer.
For lead, it is stated if predicted child blood levels are below levels of concern then adult
lead exposure will not be evaluated. What is the value for the child blood level of concern and how
was it derived? Specifically, does it reflect current research indicating that adverse effects on
intellectual performance occur at blood lead levels greater than 10 ug/dl?
3. Use of Uptake Biokinetic model for lead toxicity
I haven't yet seen the Uptake Biokinetic model so I can't comment on it. I have requested a
copy of it, and hope to receive it in time to read it before the workshop.
4. Assessment of human health effects, but not ecological effects
I do have concerns that only human health effects are being modeled. Humans are not
necessarily the most sensitive species (indeed for some dioxin effects humans are relatively
resistant) so protecting humans will not necessarily protect other life forms that inhabit the area.
Exposure Assessment
Exposed Populations
3. Delineating the study area by cancer risk associated with consumption of dioxin/furans in
beef
Adverse effects on the immune system and reproductive function occur at doses of dioxin
below those that cause cancer. Therefore, defining the study area strictly on cancer effects may
artificially limit the size of the study area.
-------
Mary E. Davis
5. Additional exposure routes that need to be considered
Ingestion of dioxin/furan in locally raised beef is the driving factor, and appropriately so
since it was identified in the screening risk assessment as the most sensitive. Did the screening risk
assessment consider ingestion of dioxin/furan in locally obtained venison? The Introduction
indicates that there are large tracts of land in state parks and game lands. I suspect, based upon
my experience of living in West Virginia, that consumption of local deer is substantial and may be
more significant than subsistence fisherman. A buck yields approximately 40 pounds of boneless
meat, which will provide a family with quite a few meals.
Gardening estimates were from areas in Ohio primarily west of the plant. What about
those areas under the plume?
Exposure Concentrations
1. Application of COMPDEP atmospheric dispersion model to WTI facility
According to the Overview of the COMPDEP Model document, snow does not count as a
form of precipitation. According to page 7, the F value is determined from the type of precipitation
and a value of 0 applies to the precipitation types of snow, ice or none. This is used if wet
deposition is to be calculated. It would seem to ignore a substantial portion of the precipitation
that occurs during the cold months of the year. What is the logic for this? Snow picks up acid
gases, and the acid shock from the snow melt in the spring causes adverse effects on trout streams.
I suspect snow picks up other pollutants as well.
Is the WTI stack above the ridge line? How does the model handle inversions?
Estimation of Exposure Dose
In general, how is exposure to chemicals on particulates estimated? Is that included in
inhalation of air, by assuming that all of the chemical that is on the particle is available?
Evaluation of Routine Fugitive and Short-Term Episodic Emissions
The risk assessment makes the assumption that accident experience nationwide will be
predictive of the accident experience in the vicinity of the WTI plant. I do not consider that to be a
valid assumption. I have not been to the WTI site nor driven the roads in rural Ohio. I have driven
rural roads in Pennsylvania within 30 km of the WTI site, and I have driven rural roads in West
Virginia a lot. There is a reason why one sees so many 4-wheel drive, high profile vehicles in this
area and that is that the roads often leave a lot to be desired. The Introduction indicates that "The
topography of area is gently rolling, except in the immediate vicinity of the site -where the Ohio
River,..forms a steep river valley." The more relevant factor is accident experience in this area,
and specifically the routes to the plant and specifically vehicles of the type and weight that will be
-------
MaiyE. Davis
bringing the hazardous wastes to the plant. Furthermore, the ability of the existing roads to stand
the additional load of delivery of wastes to the site should be considered. That is, will use by
trucks delivering waste to the plant cause the quality of the roads deteriorate such that accidents
are more likely?
-------
Barry Bellinger
-------
Barry Dellinger
The design of the WTI incinerator appears to be generally appropriate for
destruction of hazardous wastes. Based on the limited process description provided for the
pre-meeting review, the design incorporates most common features of modern, well-
designed hazardous waste incinerators and, in some areas, state-of-the-art technology.
However, its does appear to be sited unusually close to a number of schools; and the
location near the river .which may limit plume dispersion, does suggest that special
scrutiny of the incurred risk is necessary.
From an emissions perspective, there appears to be several unusual features about
the WTI incinerator. In spite of its modern, original design, dioxin emissions were
apparently unacceptably high. Even with the addition of an activated carbon control
system, the emission rate appears to be higher than that reported for many hazardous waste
incinerators. The reason for this is not clear. The description of the operating conditions is
insufficient to judge if the emissions are primarily a result of a combustion inefficiency or
are more directly related to the composition of the waste feed.
The emissions characterization from the April 1993 performance test is also highly
unusual. I have personally never seen so many nitrogen containing compounds reported
from such a test and the range of structures of chlorinated compounds is greater than that
typically reported. Since the stack test protocol was not provided, one does not know the
absolute nor relative emission rates of these compounds. It is entirely possible that the
reported emissions profile is a result of the particular sampling and analysis protocol used
in this study, i.e. specific chemical analyses may have been performed that are not usually
performed for incinerator tests. The most important factor in these emissions may be the
waste feed composition, which was not provided for pre-meeting review. The emissions
suggest a high concentration of chlorinated and nitrogenated compounds in the waste feed.
Were, perhaps, pesticides being burned? If so, then chlorinated, nitrogen containing
heterocyclic compounds may also have been formed that were difficult to analyze for.
Stack test protocols, waste feed compositions, and incinerator operating and design
parameters are essential for a proper evaluation of the stack test data and extrapolation to
burning of future wastes.
The essence of the controversy over hazardous waste incineration as I see it is
whether we are accounting for all of the toxic emission in the risk assessment This point
-------
Barry Dellinger
is not propoerly discussed in the risk assessment protocol. The plan discusses using a PIC
data base from similar incinerators, one compiled for incineration in general, and the PIC
results from the 1993 stack test. How the PIC analysis for the stack test was designed and
conducted is not discussed. Instead, the protocol focuses on the use of a limited number of
surrogate PICs selected from all of these data bases and how the measured emission rates
will be prorated to account for something called total hazardous organic emissions (THO).
Although I believe that this may be an appropriate strategy, there is no discussion of how
THO will be determined. Is it from gravimetric analysis of a total organic sampling train?
Will identifiable, non-toxic organics (e.g. methane) be subtracted from this total? Will it be
based on mass balance discrepancies? Will it be based on only identifiable GC or HPLC
peaks? This point must be fully addressed in order to have any confidence in ihe risk
assessment
More discussion on accidental releases is needed. How does the waste feed cutoff
work? How will the calculation of emissions from the combustion of hazardous waste
already in the system during the upset be handled? What upset frequency is assumed? In
the same general category, one could include a fire at a handling facility. How will the
emissions from the fire be estimated? What frequency of occurrence is assumed?
In summary, the estimate of organic emission appears tobe a fairly complete,
standard approach. Considering the apparently unusual problems associated with this
facility, an even more complete approach is indicated. Much more attention needs to be
paid to the waste feed composition during the tests on which the emissions measurements •
are made. Then a fuller description of the source and composition of all of the wastes that
may be accepted by the facility must be derived. Only then, can a realistic estimate of the
true operational emissions be made.
-------
George Fries
-------
George F. Fries
PartL
The strongest feature of the plan is the is the comprehensiveness of coverage of potential
emissions and the major exposure scenarios that might occur. No important compound or
exposure scenario appears to have been omitted. Viewed from the perspective of movement
of persistent chemicals through the food chain, the strong emphasis on dioxins and related
compounds is appropriate. A major strength of the plan is the heavy reliance on site-specific
information. This may be the most difficult feature of the plan to carry out in practice
because of the fairly short period of time available to complete Phase II. The default values
presented in Appendix A are generally conservative. Thus, a failure to obtain all of the site
specific information will probably not be too serious when evaluating the high-end subgroups
because the conservatism of the default values will encompass all realistic scenarios. Reliance
on the default values, however, will tend to over estimate exposure levels of the median
population exposures or of the total population at risk.
Part II.
Emissions Characterization
The types of compounds that may be present in incineration emissions is outside the area of
my expertise. Thus, I cannot recommend additions or deletions to the list.
The three criteria suggested - quantity, toxicity and bioaccumulation of potential - are the most
logical to use in selecting the surrogate chemicals for quantification and risk assessment.
Many of the compounds in the Organic Residues and PIC list are not likely to bipaecumulate
or be transported into the food chain because of a high degree of volatility and/or
susceptibility to metabolic degradation. Some judgement could be used in addition to the log
in selecting the compounds for bioaccumulation potential in milk and meat. The value log
is good for determining the distribution between fat and other tissues, but it is not
particularly reliable for predicting metabolism of a compound. Thus, for some non-
i
chlorinated aliphatic compounds like the polycyclic aromatic hydrocarbons (PAH),
bioaccumulation might be predicted from the log K,,,,. In practice there is little
-------
George F. Fries
bioaccumulation of PAHs because .these compounds are readily metabolized by animals into
water soluble products that do not enter the human food supply. Therefore, the emphasis on
dioxin-like compounds in this assessment is appropriate. Most of the other organics will only
be serious if high air concentrations are predicted and inhalation is projected as a serious route
of exposure.
It is appropriate to use both the trial burn and performance data from similar rotary kiln
incinerators and other types of incinerators where appropriate. The combined data should
provide stronger conclusions than using either type of data to the exclusion of the other. The
data from the test performance and trial burns are limited. Any conclusions drawn from this
limited data will be strengthened if the conclusions are consistent with those drawn from the
long-term experience of comparable incinerators.
I am not familiar enough with the characteristics of incinerator emissions to comment on the
other points.
Hazard Identification and Dose Response Evaluation
The use of dose response data in the assessment appears to be appropriate. Based on present
knowledge, cancer risk is probably the most important criteria for most chronic toxicological
endpoints. The situation with dioxin-like compounds may change when the EPA reassessment
is completed. If this occurs before the completion of this assessment, appropriate adjustments
will be required.
The use of the biokinetic model for lead uptake appears appropriate. Children as the most
sensitive group and are likely to be the most exposed group because of their greater contact
with soil. Thus, adult exposure should not be a concern if the assessment does not reveal a
risk to children.
The concept of assessing human health effects but not ecological effects is appropriate.
Generally, one is concerned about individual risks when assessing human health, but
population risks are the most important when assessing ecological effects. Even if an
-------
George R Fries
i
individual animal was impacted adversely, it is difficult to visualize the scenario in which a -
population of animals in the region could be impacted significantly.
Exposure Assessment
Exposed Populations
I agree with the general approach for deriving the approximate distribution of individual risk
estimates. An important question is whether it be possible to derive sufficient site-specific
information within the time frame of the assessment to allow reliable estimation of the
numbers and typical activities of individuals in each subgroup. As noted earlier, estimates for
the high-end should be much reliable than estimates of central tendency. This is not a serious
concern if the high-end risk for individuals is de minimus.
Intuitively, these population sub-groups appear to be appropriate. The subsistence farmer
utilizing home-produced meat, milk and garden produce would have the highest potential
exposure among the various subgroups. There is one other potential subgroup that could have
a exposure level comparable to the subsistence farmers. Consumers of locally grown products
from a single source, such as an individual who obtains custom slaughtered beef from a
neighbor might have exposures comparable to subsistence farmer even they would not be
classed as farmers. I do no know if this would be a common practice or involve a large
number of people.
The use of beef consumption to delineate the area of concern for the high-end exposure group
(subsistence farmers) is appropriate because all previous modeling of environmental exposures
to dioxins and furans have indicated that beef consumption would typically present the greatest
exposure. This route may well account for more than 50% of the projected dioxin and furan
exposures from a localized source. Beef concentrations may not the best delineator of areas
for other subgroups because of marketing patterns.
I am not aware of any additional significant exposure routes for any of the population
subgroups. One might hypothesize a few rare food chain situations that would add to the
-------
George F. Fries
exposures of a few individuals. For example, range chickens have higher exposures to
contaminants in soil than the typical commercial poultry flock. There is a scarcity of data to
evaluate the potential of this source of exposure, which is presumed to be small and
unimportant.
The suggestion of Monte Carlo approaches is interesting but a little more information is
needed "to determine whether the system is well enough understood to apply these techniques.
Basically, I do not believe that enough is known about the probability distributions of some of
the parameters. Time resources that would be required to obtain this information is probably
too great for the modest benefits.
Exposure Concentrations ,
I am not experienced in emission characteristics, modeling deposition, and the COMPDEP
model. Therefore, I have no comments on these issues.
General equations for estimating chemical concentration hi various media are appropriate. The
most important question is the selection of the appropriate parameters for the constants in the
equations. Also, there may be a lack of reliable parameter values for some compounds.
Generally, 2,3,7,8-TCCDand other persistent compounds provide a worst case for the
.persistence and transmission of chemicals through the food chain. Specific comments on some
of the parameter values in the tables of Appendices A and B are later section of this writeup.
Dose Estimation
These topics are covered, at least by inference, in other sections.
Routine fugitive and short-term episodic emissions
The issue of fugitive emissions is handled adequately. Fugitive emission of persistent
compounds that bioaccumulate will provide a proportional addition to the environmental and
exposure levels. There is no particular significance of a brief episode of high level emission
-------
George F. Fries
of these compounds beyond the proportion it adds to the total environmental load because
these compounds are usually evaluated in terms of the average or total exposure over a long
time period, generally the effect of this would be averaged out during its environmental
processes. Short term emissions of acutely toxic compounds should be evaluated in terms of
air concentration and inhalation exposure.
Risk Characterization
The two different approaches for estimating population risks are justified. If the area is
concern is a significant net exporter of food to areas outside of the study, an important part of
the total population risk will involve individuals outside of the study area. This should be
evaluated. Risks through non-food chain pathways appear to be negligible and I am not aware
of an appropriate methodology. Uncertainty issues are addressed elsewhere.
Specific Comments on Appendices
Table A-9
The food consumption rates are based on survey data that is at least fifteen years old. There
has been a shift in dietary habits since the survey. Pork and poultry are more significant
relative to beef than indicated in these consumption data. The high-end consumption rates,
however, should still be realistic for the more highly exposed subgroups in the assessment.
Table A-10
Slaughter data for the study area may be available from the Food Safety and Inspection
Service (FSIS) of the United States Department of Agriculture. All meat in interstate.
commerce must be from animals slaughtered in federally inspected plants. Approximately ;
97% of animals were slaughtered under federal inspection in 1991 (Agricultural Statistics,
1992 edition). Thus, these statistics should be available the area unless the number of plants
are insufficient to avoid disclosing confidential business information. If available the FSIS
-------
George F. Fries
data also might provide inferences on the potential movement of cattle into or out of this area
for slaughter.
Tables B-l and B-2
The greatest uncertainty in these tables involve the soil loss constant. These data are not
available for many compounds. Even when estimates are available for compounds like TCCD,
there is a great deal of uncertainty associated with the values .
Table B-7
Most of the input values in this table are quite conservative. For example, soil intake by cattle
is largely a factor of whether or not the animals are pastured. Many commercial dairy herds
do not use pasture. Soil intake is also reduced if dairy cattle are confined to concrete as
occurs on many commercial dairy farms. Similarly, many hogs are raised in total confinement
and would not have exposure to soil. The soil intake figures are also conservative in the sense
that the cattle figures involved studies of year-round grazing, a condition that would not
necessarily be appropriate for a northern area like Ohio. The default value for grain
consumption by dairy cattle is also on the low side. The 1992 edition of Agricultural Statistics
(US Government Printing Office, 1992) indicates that the average amount of grain fed to dairy
cows in 1991 was 5440 pounds, 4810 pounds and 6130 pounds in Pennsylvania, West Virginia
and Ohio respectively. Converting these figures to the metric system yields average grain
intake (dry weight) was about 6.2 kg/day rather than the 2.6 kg/day used as a default value.
Grain is presumed to have lower concentrations of persistent environmental contaminants than
hay and silage. I am not raising these issues in order to suggest changes the parameter values
but rather to point out that the default values in this table and in many other tables are highly
conservative and would tend to yield high-end exposure estimates.
Table B-9
Specific biotransfer factors should be used when they are available as in the case of TCCD.
Many compounds with a high predicted biotransfer value are subject to metabolic degradation
-------
George F. Fries
by animal will not be bioconcentrated. This is particularly true of compounds like the simple
nonhalogenated hydrocarbons and the PAHs. It should also be noted that the log Kow
overestimates the transfer of some of the more highly chlorinated compounds as is illustrated
in the dioxin data cited in the body of the plan. To the extent possible, the biotransfer factors
should be based on actual literature values rather than the calculated value when the literature
value is available.
Table B-10
This equation does not appear to account for the possible transfer of compounds in the vapor
phase from soil via air to plants. This may be a significant factor in the air concentrations in
the microenvironments of plants after a long-term buildup of persistent compounds in soil.
Table B-12
As with annuals, specific literature values for the plant uptake it should be used in place of the
equation if the data are available.
Table B-16
The term "silage" undoubtedly applies only to corn and does not apply to grasses or legumes
that may be ensiled. In this case, the equation for pasture grasses and hay should be used,
even though the material is referred to as silage in some circles.
Table B-17
There is a great deal of uncertainty in these values. I believe the environmental half-life is
based on dry deposition of particulates. Half-lives after wet deposition or vapor adsorption
could be considerably different. For example, lipophilic chemicals adsorbed to waxy plants
from the vapor phase might have much longer half-lives.
-------
Thomas Gasiewicz
-------
Gasiewicz, T.A.
COMMENTS ON "WTI RISK ASSESSMENT PROJECT PLAN (OHD980613541)
Thomas A. Gasiewicz, Ph.D., Associate Professor, Department of Environmental Medicine,
University of Rochester Medical Center, Rochester, New York 14642
I. INTRODUCTION
1. p. 6, 2nd paragraph, 1st sentence: Here a statement should be made to indicate whether
any data is supplied with the particular waste material to indicate its makeup and or
chemical composition. What are the requirements in this regard?
2. p. 6, 2nd paragraph, line 5, "..sampled..": What is the material being sampled for?
Specific information would be useful.
3. p. 14, line 6 from bottom, "...except for consumption of beef..": One would assume that
this refers to the data from "a subsistence farm", but it should be explicitely stated.
4. pp. 14-20, "F. Site-Specific Data...": It might be noted here whether or not the analysis
programs described in this section will be continued in the future. As indicated, these
programs provide very useful base data. They wiU be of greater value for the purpose of
risk assessment of the WTI facility if they are continued and possibly expanded. If the
continuance of these programs are mentioned later in the text, some indication should be
made here as to where.
H. EMISSIONS CHARACTERIZATION.
1. p. 28, 1st paragraph, "The list was created as follows...": The indicates that the list was
generated partially based on "Chemicals detected in the March 1993 trial burn.." and
those "Organics detected most frequently and in the highest comconcentrations during a
comprehensive EPA...study..". It is difficult for the reader to know which chemicals were
detected vs which were chemicals were tested for. It might be useful to provide a listing of
those chemicals for which an analysis was performed but which were not detected in any
significant concentration to make the listing in Table 7.
2. The following should be considered for addition to the list:
aluminum
copper
certain other polycyclic aromatic hydrocarbons (PAHs) including: benzo(a)pyrene,
benzo(b)fluroanthene, chrysene, dibenzo(a,h)anthracene, fluroanthene, and
anthracene
Although some of the above might rank lower on an inherent toxicity list, the relative
output may be considerable. This might be especially true for the PAHs since petroleum
wastes, tars, and creosols will be incinerated (see Appendix C).
-------
Gasiewicz, T.A.
3. pp. 31-34: The criteria - quantity, toxicity and bioaccumulation potential - are
appropriate for the quantification of risk assessment. Furthermore, the manner in which
the organic chemicals will be ranked and selected as surrogates appears appropriate.
4. Development of Chemical-Specific Emission Rates: This reviewer generally agrees with
the approach as outlined in pp. 34-44. It is important that the trial burn data be used as
the basis for the determination of the emission rates. However, realizing that with time and
differences in waste feed there may be considerable variability in these emission rates,
other data should be used. The data from the Von Roll facility in Germany might be the
best for this since it has similar characteristics to the WTI facility. A comment should be
made in the text as to why the Germany facility has been selected vs other Von Roll
facilities. Is it possilble to obtain data from other Von Roll facilities with similar
characteristics? Are the other Von Roll facilities very different?
It would be useful to specifically point out any known significant differences in the
Von Roll Germany facility that may exist and potentially affect the emissions or the
variability observed. If none are known then it should be specifically mentioned.
This reviewer also agrees with the proposed use of the data from other types of
hazardous waste incinerators and combution units - but only, as the document indicates, if
somehow they can be reliably extrapolated to the WTI facility since this data and burn
conditions may not be approriate for the WTI site. This may give an overly conservative
risk assessment. On the other hand, even if specific data cannot be extrapolated, the data
might be useful in providing qualitative information on expected potential sources of
variability.
The considerations for the dioxins/furans appear appropriate. However, the
"emission limit for total dioxins/furans of 30 ng/dscm...recently imposed on WTI.."
needs to be explained. By what agency was this imposed by? Some specific reference
should be given.
For "Other PICs and Organic Residues" (p. 41), it is not clear why data on organic
chemical residues and PICs from "similar facilities" (such as the one in Germany) are
"not sufficient". This is stated specifically for the WTI facility, but not for similar Von
Roll facilities. If the data from other hazardous waste facilities (EPA, 1989c) and/or the
Louisiana facility (USEPA, 1987a) are used, what will be the specific rationale to or to not
extrapolate to the WTI facility? Will this give a very conservative estimate of emission
rates for the PICs and other organics?
The considerations for the metals appear to be appropriate.
For the acid gases (p 43), how were the "expected stack emissions of acid gases"
projected? What were the assumptions and what were they based on? Some specifics
should be given.
5. Emission scenarios:
The first scenario (pp. 38-39) based on the average emissions for continuous
-------
Gasiewicz, T.A.
operations is appropriate.
The second scenario (p. 39) seems to be somewhat self-limiting if only the applicable
regulatory limits were considered. Rather, it would seem appropriate to consider the high
end of the variability estimations based on real data from other incinerator facilities as
extrapolated appropriately to the WTI facility for this second scenario also. This may or
may not be more conservative than use of the regulatory limits. Nevertheless, the real
and/or extrapolated variability data should be used as an additional scenario.
The potential short-term episodic emissions listed for consideration on p. 93 are
certainly appropriate. It is not clear why any data from other Von Roll facilities are not to
be considered for the "Startup/Shutdown, Upsets, and Malfunctions" emissions. It would
seem appropriate to consider data from a facility with similar engineering characteristics.
This same consideration would not necessarily apply to the other short-term episodic
emission scenarios.
Page 94 indicates that acute and/or subchronic toxicity data and short-term regulatory
standards will be evaluated. There is some concern here that the method for choosing
surrogate compounds and chemicals on pp 31-34 does not choose on the basis of acute
and subchronic toxicity data but rather chronic exposure toxicity data. Should surrogates
be chosen on the basis of acute and sub-chronic toxicity data also?
The scenarios for "Routine Fugative Emissions", "Releases from On-Site
Accidents", and "Release from Off-Site Transportation Accidents" do well to consider
past data and appropriate model systems. However, will the WTI facility handle any
particular hazardous waste that are unique to this facility and not considered by other
types of data, analyses, or models? If so, do other conditions/scenarios unique to the
particular waste need to be considered?
The scenarios do not appear to comment on or consider any changes in emissions
that will occur with aging of the facility. Assuming routing maintenence, is this a potential
consideration or not? In either case, it should be addressed.
6. Vapor phase/partical partitioning: The physical chemistry is beyond the expertise of this
reviewer; no comments have been made.
m. HAZARD IDENTIFICATION AND DOSE-RESPONSE EVALUATION
1. Planned use of dose-response data: The use of SF, RfD, and RfC data appear to be
appropriate. However, it should be emphasized that for non-carcinogenic effects, data for
both chronic and acute and sub-acute exposures and toxic end-points will be considered.
This is due to the different types of emission, and thus exposure, scenarious that will likely
occur. Thus, if an RfD is given only for a chronic exposure, additional data for acute
exposure may have to be evaluated.
2. Chemical form and speciation: It is important and appropriate that the chemical forms
and species be considered in the dose-response evalutation. While few specifics are given
-------
Gasiewicz, T.A.
in the text, the general plan as applied to the examples discussed appear to be appropriate.
For example, the exposure assessment process (but not the final risk analysis) should
consider the different pharmacokinetic properties of the dioxin/furan congeners.
3. Table 9: As indicated above, it should clearly be indicated that both acute and chronic
lexicological endpoints be considered for the "non-carcinogenic" effects. Obviously the
exact endpoint chosen will vary with the chemical. Given the chemicals, the selection of
"Carcinogenic" vs. "Non-Carcinogenic" appears appropriate.
4. Characterization of lead toxicity: This reviewer is not an expert on lead. However, in a
brief discussion with someone from our department who is more of an expert (Dr. D.
Cory-Slechta), she indicate that the discussion here may be misleading. There are
certainly EPA recommendations for blood lead levels. Furthermore, she pointed out that
the models of Harley and Kneip are rarely used due to updates by other investigators.
Thus, it seems some revisions of this particular section and the approaches described are
necessary.
5. Ecological effects: The potential eco-toxicology of the particular chemicals listed (with
the possible exception of the dioxins) is beyond this reviewer's area of expertise. For the
dioxins, the data suggests that a variety of animal species including fish and certain other
wildlife may be more sensitive to the toxicological effects of these compounds than
humans. In my opinion, it would be a long-term mistake if we do not consider the
ecological effects of these emissions as well. For example, it was noted (p. 14) that there
was found to be a risk of 105 for humans considering a scenario in which all beef
consumed was from the home stock. This "subsistence" model as applied to humans
may likely be the real model for local wildlife. It would be useful to have a model for
deposition and contribution to the local lakes and streams, as well as an estimate (or at least
a statement) of effects on local wildlife. If this is not performed, the authors should
indicate why.
IV. EXPOSURE ASSESSMENT.
1. Approach for developing central tendency and high-end estimates of exposure (pp. 55-
56): Conceptually the general approach as outlined in the 1st paragraph of p. 55 is good.
Furthermore, the descriptors of "high-end" exposure and "central tendency" exposure
groups are clear. The area will be divided by isopleths based on air dispersion modeling.
Within these areas sub-groups will be identified. However, it is not clear to this reviewer
what the basis is for the initial separation and classification of the specific population sub-
groups (discussed in 1st full paragraph, p. 56). What is the rationale for choosing these
particular sub-groups? Is the classification based on residence, occupation, and / or
behavior patterns which gives a sense of perceived differential exposure? If this is the
case, the examples of resident, non-farmer, farmer, children, subsistence farmers, and
-------
Gasiewicz, T.A.
resdents with home gardens may be very oversimplified. For example, there may be
children that live on subsistence farms vs. a residence without a garden, vs a residence with
a garden. Furthermore, there may be residents living in one sub-area but working or
taking food from another sub-area. Thus, the final division of the sub-groups may be
much more complex. Thus, although the general concept of defining sub-groups is good,
some rationale needs to be given on what the basis of this is.
Once the sub-groups are chosen, exposure estimates for each sub-group would be
performed by "combining exposure factors" which reflect activity and behavior patterns.
This might be viewed as confusing since it multiples the "25 exposure estimates" by the
different types of behavior patterns.
Thus, overall the general approach for developing "central tendency" and "high-
end" exposure estimates seems to be appropriate. However, the description of this general
approach as given in V.a is very confusing since more explanation than is needed is given
in this section. My suggestion is that the general approach section should be kept to a
minimum, avoiding details that might lead to some confusion. These details can be added
in later sections.
2. The general concept of the sensitivity analysis described on p. 56 (last paragraph) into p.
57, might be useful in that a range for the high-end exposures will be obtained. This is
also important since in the initial analysis only the sub-group exposure, and not of
individuals comprising the sub-group, will be estimated. Here behavorial patterns and thus
exposure estimates of specific individuals will be determined. However, insufficient detail
is given for a complete understanding as to how this will be performed, i.e. a more detailed
example would be useful. Furthermore, one would wonder whether this appropriately and
accurately depicts the uncertainty or error of the estimates involved. Again, there is
insufficient discussion and detail to make a judgement on this.
3. Delineating the study area (pp. 58-59): The general approach of using the dioxin data for
the determination of the study area and specific concentration isopleths is good. However,
this model would likely only apply to those chemicals associated with paniculate matter
since it is highly likely that the dioxins will be mainly, if not exclusively, associated with
participates. Here the assumption has been made, but not stated, that the emissions
partitioning of the dioxins is the same as for the other chemicals under study. Obviously,
the final determination of this rests on the results of the "Determination of Emissions
Partitioning" as described on pp. 44-45. Nevertheless, the model of the dioxins will apply
to only those chemicals that have a similar emissions partitioning. This should be clarified
and explained. If this is indeed the case, do other mbdels need to be investigated for other
chemicals such as the organics and acid gases? What are the specifics of these models? If
the assumption is correct, the basis of it should be stated.
4. Identification of sub-groups: The classification of the sub-groups appears appropriate.
-------
Gasiewicz, T.A.
However, it is not clear as described in the text whether the groups of children who attend
school in the area, breast-feeding infants, and subsistence fisherman, are considered as
additional sub-groups to the 8 listed above, or are these considered as a special
classification of the 8 main subgroups. Some explanation is needed. (It is possible that a
subsistence fisherman may be in one of the adult farming or non-farming sub-groups.
Furthermore, the subsistence fisherman may fish in one sub-area while living in another.)
5. Exposure Routes: All of the significant exposure routes have been considered.
6. Monte Carlo simulation: This statistical anlaysis is on the border of this reviewer's
expertise. I will leave this to others with greater expertise.
V. EXPOSURE ASSESSMENT:
1. COMPDEP atmospheric dispersion model: This reviewer lack expertise in atmospheric
modeling and particle physics. No specific comment are made.
2. Characterization of pollutant distsribution: No comment.
3. Size distribution of particles: No comment
4. Meteorological data: Do all of the data obtained from the 3 meteorological tower
accurately mimic or predict the conditions at the WTI stack? The stack is approximately
46 meters in height. The meteorologicl towers are 10, 10, and 30 meters in height. Can
data from the Pittsburgh airport be extrapolated to the WTI sight? The assumptions here
should be stated, and the basis of these explained.
5. Modeling equations: This reviewer is not familiar with the derivation of the equations
used, so no comments can be made. It is extremely important, as indicated on p. 81, that
the estimated concentrations in soil, meat, milk, vegetation etc., be compared to real data
obtained at the site or other sites. This comparison will approximate the goodness of the
equations and assumptions used.
VL DOSE ESTIMATION:
1. Inhalation: The given equations do not take into account particle size distribution,
deposition of small particle sizes in to lung and absorption from lungs. Is the assumption
here that all contained in air as vapor or associated with particles absorbed? Although the
text on page 84 suggests that this is indeed the assumption, it is not explicately stated. If
this is the assumption it is certainly a worse case scenario, and it should be stated as such.
Although there may not be good absorption data (from lungs, skin, and GI tract) for all of
the compounds under-study, there certainly is data for some of the chemicals to give more
realistic estimates. For example, some of this data has been used for the dioxins in other
-------
Gasiewicz, T.A.
risk evaluations. The rationale for not using available data should be stated.
The same comments can be made for other sources, i.e. beef etc.
2. Irigestion of fish: Does the value for the chemical concentration in fish assume an equal
distribution throughout the body? This appears to be the assumption based on Table B-
21. Is the CF value on p. 89 a conservative estimate if only (likely) fish muscle is eaten?
3. Exposure factors: For all of the exposure factors an assumption is that children both
reside and attend school within the same exposure area. This should be stated as a
conservative assumption. As indicated above when absorption factors are known for
particular chemicals they should be used.
Both the dermal and ingestion exposure to soils may be incresed for residents with
gardens.
All of the food consumption data assumes (not a stated assumption) that individuals
living in a particular exposure area consume predominantly some percentage of the food
produced within that particular area. Is this a reasonable assumption? What is the basis for
this?
The values for other factors appear to be well documented or reasonable assumptions
indicated.
VH. ROUTINE FUGITIVE AND SHORT-TERM EPISODIC EMISSIONS:
1. See H.5. above.
VHI. RISK CHARACTERIZATION. .
1. Population Risk Estimation: Additional explanation is needed. The use of examples and
an explanation for the interpretation, significance, and use of the final risk number would
be useful. A statements of the assumptions, especially for the food chain pathway, would
also be appropriate. In this regard, it is not clear how the method for risks from the food
chain pathways estimates the total cancer risk to the total consuming population. Does this
assume exclusive exposure to the food produced in the study area?
2. Uncertainties: The general discussions on the Scenario, Parameter, and Model uncertainies
appear appropriate. As indicated much of the unknown quantitative uncertainty is
associated with the Parameter Uncertainty category, and it would be useful to expand the
approach to include more quantitative uncertaintly evaluation for each parameter. The
approach that is given is to identify the parameter that has the greatest uncertainty. It is
not clear how this approach can be incorporated into a quantitative uncertainty analysis
for the whole risk assessment process. The sensitivity analysis described may be useful,
but it may not accurately represent the real uncertaintly issue since the "low-end"
exposure limits are not addressed. Determination of a high end / typical ratio may really
only flush out parameters likely to have the greatest uncertainty.
-------
i Gasiewicz, T.A.
One approach to make the uncertainty analysis more quantitative would be to
accurately detemine the quantitative uncertainty for individual parameters. Using the
appropriate statistical approach, the quantitative uncertaintly for each parameter would
contribute to the uncertainty of the final number calculated. This is on the border of this
reviewer's expertise in statistics, and this may be better addressed by another. Nevertheless
it seems possible that the incorporation of quantitative error esitmates for each parameter
would be useful for making the uncertainty of the whole risk assessment process more
quantitative. Furthermore, this would provide data to really determine which parameter
has the greatest uncertainty. This should be discussed in more detail.
-------
Halstead Harrison
-------
Review of
"WTI PHASE II RISK ASSESSMENT PROJECT PLAN"
[EPA ID # OHD980613541],
With Attention to Dispersion and Deposition
Modeling
Halstead Harrison
University of Washington
Atmospheric Sciences [AK-40]
Seattle, WA 98195
(206)-543-4596 Office
(206)-543-0308 FAX
November 23, 1993
Abstract:
The WTI phase-II plan proposes to estimate maximum
cancer risks that may be attributable to the WTI
incinerator project, through a sequence of observations
and models to estimate a suite of toxic emissions,
their dispersion and deposition, their exposure paths,
and their separate toxicities. In this review I shall
comment briefly on the control criteria, and on
estimating dispersions and depositions, with an emphasis
on uncertainties.
Control Criteria:
I did not see a clear statement of a parameter that
is to be used to judge whether the WTI project is
acceptable, or a protocol to respond to various
estimates of that parameter.
I presume that this parameter might be something like
a lifetime cancer risk for a most exposed individual.
The phase-I assessment cited a "maximum estimated
inhalation cancer risk for any single organic stack gas
constituent" of approximately "1 x 10 6, for dioxins
and furans." The cumulative risk of all toxins, by all
paths, was not estimated, nor were protocols discussed
about what to do with various outcomes.
I assume that to do this is a principal task of phase-II.
-------
H. Harrison
No criterion is stated in the phase-II plan, however,
as to what risk may be acceptable, what uncertainty
may acceptable in estimating that risk, and what
ameliorating steps might be required with various
modeled outcomes.
I recommend that these issues be addressed before the
phase-II project proceeds much further. This will
have the double benefits of focusing the modeling
efforts, and of defending later regulatory decisions
from criticism that responses may have been influenced
in a posteriori ways by the models' predictions.
Here is a not-unlikely outcome: suppose the phase-I
estimate of 1 x 10 s is about right for the single most
toxic stack toxin through an inhalation path. Suppose
also that when all paths and all toxins are properly
summed in the phase-II effort this number grows to
1 x 10~B, with a standard error of a factor of ten.
This would suggest that there would be about a 16%
chance_that the "true" risk would be greater than
1 x 10~4.
Is this risk acceptable? How should this be decided?
If the risk is not acceptable, what steps should then
be taken? If this a priori guess of the risk is not
acceptable, should steps be taken now to improve the
phase-II effort? What steps?
The central idea is that phase-II should be designed
to assist wise management.
Dispersion and Deposition:
The phase-II plan has three tasks, in sequence:
estimating the emissions, modeling their dispersion
and deposition, and estimating doses and risk levels.
For the second of these tasks it is proposed to use a
COMPDEP model with one year's meteorological data, taking
the advantage of three "Met" towers on or near the site.
Recommendations for the use of these data "will be
developed". Off-site data from the Pittsburgh airport,
25 miles to the southeast, will be used for stability-
class and precipitation.
-------
H. Harrison
It is my experience that remote Met data are marginally
useful, at best, and are often misleading. This is
true especially with airport data, where wind speeds
bias high by 1-2 meters per second. Please take this
into account, and do not accept airport-local stability
classes. The characteristic horizontal scale length
for fluctuations in rain rates is much less than 25 miles.
Please use every effort to get local rain data.
COMPDEP uses Monin-Obukhov similarity theory to extract
vertical mixing parameters <[K*, L] from the vertical
velocity and temperature gradients. The local Met
towers will greatly assist this. Please be aware,
however, that the field experiments on which M-O
parameter!zations depend were biased towards relatively
high wind speeds, U > 2 m/s, and required averaging
times of tens of minutes. These speeds and times
in turn required stationary upstream conditions over
fetches of 2-4 km. With lower wind speeds the fetch
requirements grow proportionately. Long-fetch
conditions are untypical in the complex river-valley
terrain near the WTI incinerator.
M-O theory is therefore increasingly unsatisfactory at
low wind speeds and in complex terrain: exactly those
conditions when the potential for high air-pollution is
greatest.
Further, COMPDEP is a variant "Gaussian Plume". Models
of this family make many assumptions which are also not
appropriate at low wind speeds in constricted terrain.
In these circumstances, uncertainties of factors of two
— as cited on page 105 of the draft phase-II plan ...
are optimistic.
Other models can be suggested [I shall do so if
requested], and some of these are probably superior
to COMPDEP. Many details of parameter selection may
be discussed [I shall do so if requested], but I do
not wish to get lost here in parochial nitty-gritty,
for the following important reason:
The phase-II plan emphasizes the ordinary operations of
the WTI incinerator. But extraordinary operations are
more critical. Most of the risk-exposure from the WTI
-------
H, Harrison
will be concentrated among few episodes. Most of these
will be associated with "stagnation events" that are
characterized by near-surface winter-time radiation
inversions ... and by episodes of piant-mismanagement.
Chernobyl was safe, most of the time.
Thus the air-pollution modeler's art, with much
discussion of minutae concerning Huber-Snyder downwash,
and Sehrael-Slinn deposition velocities, and Monin-Obukhov
lengths, and so on, can get frustratingly irrelevant in
the presence of one juicy stagnation event, perhaps
coupled with one briefly careless plant foreman.
For this reason I emphasize again that uncertainties
of factors of two ... as cited on page 105 of the draft
plan [with a reference to an EPA study of dispersions,
only, in flat terrain] are optimistic with the present
effort, which requires the modeling of both dispersions
and depositions in complex terrain and stagnant air.
Recommendations
1. Please state clearly what risk parameter will govern
regulatory response, and consider ahead of time what
that response should be for various outcomes of the
phase-II study.
2. Please include within the phase-II project an explicit
effort to estimate uncertainties in the governing
risk parameter, with emphasis on the incidence of those
relatively few events with highest risk.
-------
Kosalwat
-------
PIM KOSALWAT
I generally agree with the risk assessment (RA) plan for the
WTI facility. In 1992, KBN used a similar plan to conduct a
multiple pathway human health RA for a proposed hazardous
waste incinerator in South Florida. My comments on the
project plan are as follows:
Exposure Assessment
Exposed Populations
1) The ages of children who attend school in the area
(p. 60) should be lower to include 5 and 6 years
old since a typical elementary school has
kindergarten classes (the youngest in the classes
would be 5 years old). Consequently, the body
weight used in the RA should be adjusted
accordingly.
2) According to Figure 2 (p. 4) of the RA plan, it
appears that there is an unusually high number of
schools near the WTI facility. Do these include
i
pre-schools or day-care centers? If such
facilities exist near the incinerator, children of
ages 2-5 years old should probably be used to
represent the special sub-group, instead of those
in elementary school, unless the elementary school
being considered is significantly closer to the
incinerator than the preschool.
3) One sub-group which is not being considered is WTI
workers who reside near the incinerator. It is
reasonable to assume that persons working at the
facility would also live in the vicinity and could
incur additional exposure to facility emissions at
home. Therefore, the risk analysis should
quantify chemical intake/uptake for both
occupational exposures and exposures potentially
incurred at home.
-------
PIM KOSALWAT
Exposure Concentrations
In the estimation of soil concentrations (CS), the
default value for total time of deposition (Tc) is 30
years. Since the length of lifetime (LT) exposure for
cancer risk estimates is 25,550 days (i.e., 70 years),
the Tc should probably have a default value of 70 years
unless the expected life-time operation of the
incinerator is 30 years.
General Comments
1) I believe that ecological risk assessment should be an
integral part of any decision made on the operation of
the WTI incinerator, particularly when there appears to
be numerous streams within the 50-km radius of the
facility (Plate-2 map). In addition, it is stated in
the project plan that there are large tracts of land in
the area that are reserved for state parks and game
lands. Locations of streams or other surface waters,
as well as other ecological habitats (e.g., forested
areas, grasslands, floodplains, and wetlands) near the
incinerator should be identified. There should be a
screening of vegetation, aquatic life, and resident or
migratory wildlife in the area. It should also be
determined if threatened or endangered species exist in
the area. Then, the habitats and populations that
potentially could be affected by the stack emissions
should be determined. I recognize that it would be
very costly and time consuming to include several
indicator chemicals. Perhaps, a single indicator
chemical with the highest bioaccumulation or
biomagnification potential through trophic levels in
food chains could be used in the assessment. For
example, mercury could potentially be used as an
indicator chemical in this ecological risk assessment.
Although emitted as inorganic form, mercury can be
methylated under both aerobic and anaerobic conditions
-------
2)
PIM KOSALWAT
by microorganisms indigenous to soils and water.
Methylmercury is highly stable and its ionic properties
allow it to penetrate membranes in living organisms
(Eisler, 1987). Methylmercury in surface waters is
rapidly accumulated by aquatic organisms;
concentrations in carnivorous fish at the tops of
aquatic food chains are biomagnified on the order of
10,000 to 100,000 times those concentrations found in
ambient waters (ATSDR, 1989). Environmental and
biological sampling performed for the ecological RA can
also be used as a baseline for ecological monitoring of
the incinerator (pre-operation vs post-operation).
According to the project plan, a variety of industrial
operations exist in the Ohio River Valley in the
vicinity of WTI. The final decision made on the WTI
incinerator should also consider risks from other
industries currently in operation. Each industry may
not pose unacceptable risks to humans and the
environment in the area, but by adding another source
of potential contamination, the risks collectively may
be unacceptable.
REFERENCES
Agency for Toxic Substances and Disease Registry (ATSDR).
1989. Toxicological Profile for Mercury. U.S. Public
Health Services, Washington, DC.
Eisler, R. 1987. Mercury Hazards to Fish, Wildlife, and
Invertebrates: A Synoptic Review. U.S. Fish and
Wildlife Service, Patuxent Wildlife Research Center,
Laurel, MD.
-------
Thomas McKone
-------
T.E. McKone
Comments on: WTI PHASE II RISK ASSESSMENT PROJECT PLAN
(November 1993 Review Draft)
Thomas E. McKone
My comments are divided into two parts. The first part addresses general
features of the plan that are expected to lead to a strong risk assessment, features of the
plan that can be fixed within the available time frame, and issues that would be useful
to consider in future risk assessments of this type. The second part provides specific
comments on emissions characterization, hazard identification and dose-response
evaluation, exposure assessment, evaluation of routine fugitive and short-term episodic
emissions, and risk characterization.
PARTI
GENERAL COMMENTS ON THE OVERALL WORK PLAN
Features of the plan expected to lead to a strong risk assessment
For lipophilic contaminants, such as dioxins, furans, polychlorinated biphenyls,
pesticides, and for metals such as lead and mercury, exposures through food have been
demonstrated to be dominant contributors to total dose within non-occupationally
exposed populations. Thus, it is both significant and scientifically valuable that the
WTI risk assessment confronts these multiple exposure pathways. The overall exposure
model proposed here is comprehensive and includes all plausible multiple exposure
pathways with some minor exceptions as noted below.
The extensive scientific peer review being given both to this document and the
EPA guidance documents upon which it is built leads to a strong risk assessment.
As described in the work plan, there is an effort to monitor fish populations to
determine, whether the WTI facility will increase contaminant levels in fish. This
component of the plan is to be commended and should, if possible, be extended to
terrestrial biota.
The collection of site-specific food-consumption data is an important
contribution to this and other incinerator risk assessments and this type of data
collection should be encouraged.
-------
T.E. McKone
The integration of surface water, soil, and atmosphere is particularly well done
and represents a major improvement over earlier incinerator risk assessments that I have
reviewed. The inclusion of the net effect of gaseous diffusion from air to water and
volatilization from water to air makes the model more credible and should improve its
reliability.
Features of the plan that can be fixed within the available time frame
For me, one of the most significant issues in the plan that can and should be
addressed is the inherent lack of reliability associated with both measured data and
models used to determine intermedia transfer factors (ITFs). There are a number of ITFs
proposed for use in the risk assessment including—octanol-water partition coefficients
(K0-w); organic-carbon partition coefficients (Koc); soil-water and sediment-water
partition coefficients (K^); Henry's law constant; mass-transfer coefficients in water and
air; steady-state bioconcentration factors for plant root concentrations relative to soil
concentration, plant leaf concentration relative to air concentration, and fish
concentration relative to water/or sediment concentrations; steady-state biotransfer
factors for milk or dairy-product concentration relative to contaminant intake by cattle;
meat concentration relative to contaminant intake by animals; egg concentration relative
to contaminant intake by chickens; and breast milk concentrations relative to
contaminant intake by mothers; and contaminant biodegradation factors in soil.
My own recent work on the reliability of methods for measuring and estimating
these types of ITFs reveals that the reliability for determining IFF values has an
associated error factor in the range of 1.5 to 10 depending in part on whether the ITFs
are measured or estimated. I have found that overall variance in quantitative estimates
of ITFs comes from several factors including (1) variability among experiments; (2) our
ignorance regarding the processes of metabolism and chemical partitioning; and (3) the
reliability with which we can measure both the outcome (biotransfer or partition factor)
and the explanatory variable (i.e., Kow). It is likely that the lack of reliability for
determining these ITFs can be a major contribution to the overall variance in the
exposure estimates. In the time frame of the WTI risk assessment, there is little that can
be done to increase the reliability of the ITF estimates. However, the results of this risk
assessment will be more credible, if the variance in these values is clearly stated and the
impact of these variances on the final estimates of risk is assessed. At a minimum, this
-------
T.E. McKone
can be done by listing the estimation error or the experimental variance associated with
the UF parameters when these values or their estimation equations are listed in tables.
Another issue that should be addressed before following this plan is the
potential violation of the law of conservation of mass and the laws chemical
thermodynamics. The specific instances where this might happen are described in my
more detailed comments below. Since there is no uncertainty about the validity of these
two laws, every effort should be made to make the models comply. In general, the
l
proposed transport models are mass conserving. However, I noted that contaminants
are allowed to volatilized from soil and plant surfaces, but it is not clear that the lost
chemical is then treated as an input to the atmospheric transport model, thus violating
conservation of mass. With regard to chemical thermodynamics, it should be recognized
that soils, surface waters, sediments, and (in particular) plant tissues are all secondary
to the atmospheric source of contaminants and, thus, are not likely to have an annual
average chemical potential (that is "fugacity") that exceeds the chemical mass potential
in the atmosphere. Because of the way contaminants are allowed to transfer from air to
both plants and soils by a combination of deposition and diffusion and be lost by a
removal process that is defined by a rate constant that is not clearly linked to the mass
potential in these compartments, there is a possibility that these compartments could
become chemical traps that receive chemicals but do not exchange them in a way that
maintains chemical equilibrium. This problem is easy to fix by placing some limits on the
chemical potential of any compartment relative to the air. I note that the model for the
surface water is carefully constructed to maintain both conservation of mass and
comply with rules of chemical thermodynamics. This care should be extended to the soil
and plants compartments.
The treatment of uncertainty and the issues of uncertainty and variability should
be integrated as much as possible into all aspects of the report and not just included as
an addendum to the section on risk characterization. Numbers that have large
statistical variances should not be reported as single values. Regression equations for
biotransfer factors should not be presented without including the standard error of the
estimator in these equations
-------
T.E. McKone
Issues that would be useful to consider in future assessments of this type
There is too much emphasis in this whole approach to answering the question of
what is the mid-range and high-end exposure. I think it would be more appropriate to
focus on the question of what is our confidence about specifying exposure distributions
within a population and among more highly exposed individuals within this population.
Not so much how many people are above a limit, but how precisely can we determine
the probability that the range of dose in some exposure duration exceeds x, where x is
an unacceptable value.
The concept of an interactive modeling and measurements approach to exposure
assessment, which is suggested by this plan should, to the extent possible, be expanded.
That is, the exposure models described here could serve as a portfolio of scenarios for
contaminant-specific, multimedia dispersion in the environment. Integrated
measurement/modeling efforts of this type could work like road maps to identify
pathways and populations for which detailed exposure survey studies are needed.
Risk assessments often operate under the premise that, with sufficient funding,
science and technology will provide an obvious and cost-effective solutions to
contaminant problems. However, in reality there are many sources of uncertainty and
variability in the process of human health-risk assessment. These uncertainties often
confuse the selection of effective solutions. Many of these uncertainties and variabilities
are not reducible. Thus, ultimately, I believe it is necessary to make use of variance
propagation methods (such as Monte-Carlo methods) in order to carefully map how the
overall precision of risk estimates is tied to the variability and uncertainty associated
with the models, inputs, and scenarios. This type of approach gives the decision maker
flexibility to address margins of error; to consider reducible versus irreducible
uncertainty; to separate individual variability from true scientific uncertainty; and to
consider benefits, costs, and comparable risks in the decision making process.
-------
T.E. McKone
PART II
GENERAL AND SPECIFIC COMMENTS ON THE INDIVIDUAL COMPONENTS OF
THE WORK PLAN
Emissions Characterization
Three criteria—quantity/ toxicity, and bioaccumulation potential have been
proposed for use in selecting surrogate organic substances for quantification in the risk
assessment (pages 31-34). In my view, the use of surrogate chemicals is appropriate
and necessary. However, it may be necessary to include one other factor—persistence—
in this scheme. As it is structured now, the scheme does not distinguish chemicals that
are highly susceptible to volatilization or to photolysis. Allowance for such compounds
should be made.
It is appropriate and necessary to consider the partitioning between solid and
gas-phase emissions using the model posed by Junge. In the adjustment for an
appropriate vapor pressure, the report does not make clear whether the temperature
used is the ambient temperature in the stack or the ambient temperature of the
atmosphere in the region surrounding the WTI facility. This should be made clear. In my
view, the temperature of the atmosphere should be used in any effort to determine
deposition to soil and plants and vapor uptake by plants. Also, there are potentially
large estimation error associated with the partition model proposed here and these
estimation errors should be quantified and noted.
Hazard Identification and Dose-Response Evaluation
(no specific comments)
Exposure Assessment
Exposed Populations
The use of an approximate distribution of individual risk estimates based on
site-specific information is an appropriate way of characterizing the inter-individual
variability in risk according to the guidance provided in the EPA Exposure Assessment
Guidelines. However, this approach does not give insight into the effect of uncertainties
in the intermedia transfer factors (ITFs). The ITFs—such the biotransfer factors in milk,
-------
T.E. McKone
meat, eggs, and breast milk—are derived from either controlled experiments or more
often from statistical estimation equations. The fact that median estimates are used
limits the ability of the exposure model to truly characterize our uncertainty about the
upper and lower range of the exposure expected for either the central tendency or highly
exposed individual.
At minimum, an uncertainty propagation method, such as Monte Carlo should be
used to characterize the impact of the potentially large uncertainties in the estimates of
biotransfer factors and other ITFs on the precision of mid-range and high-end estimates
of exposure within the population.
Exposure Concentrations
Only one year of meteorological data has been collected in order to characterize
the dispersion of contaminants within the atmosphere around this facility. In the
context of a risk assessment that applies to health effects over 30 to 40 years (the life of
the facility), this is really only a "snapshot" of the behavior of the atmosphere. It is not
clear that this one-year sample will reflect the variance in seasonal distribution of wind
dispersion and atmospheric conditions.
The COMDEP model accounts for deposition primarily as a physical process
that can be represented with the equivalent of a velocity. It has been clearly
demonstrated in the scientific literature that this approach is valid for particles and
chemicals attached to particles, such has metals, radionuclides, speciated organic and
inorganic compounds, and polar and nonpolar nonionic compounds with extremely low
vapor pressure. However, for nonionic, nonpolar, semivolatile organic compounds
simple deposition is clearly not the only process by which contaminants are transferred
from air to soil and plant surfaces. These latter compounds are often transferred from
air to lipid (i.e. plant leaves) and to organic-carbon (i.e. soil) surfaces by a mass-transfer
process that is governed by chemical potential and by mass-transfer resistance. It is
often not possible to use a simple deposition model for these chemicals. It is not clear
that the COMDEP model handles these types of mass-transfer processes appropriately.
A detailed discussion of these types of air-soil and air-plant exchanges is covered in a
number of papers, for example:
Mackay, D., and S. Paterson (1991) "Evaluating the Multimedia Fate of Organic
Chemicals: A Level HI Fugacity Model," Environ. Sci. Technol., 25,427-436.
-------
T.E. McKone
Riederer, M. (1990) "Estimating Partitioning and Transport of Organic Chemicals in
the Foliage/Atmosphere System: Discussion of a Fugacity-Based Model," Environ.
Sri. Technol. 24, 829-837.
According to the equation on the bottom of page 71, the transfer of chemicals
from air to soil is by deposition, which does not explicitly include diffusion. In
Appendix B, tables B-l and B-6, volatilization (or mass diffusion) is described as one
the mechanisms by which chemicals are transferred from soil to air. In table B-6, the
mass transfer coefficient is applied to soil as if the chemical concentration in air is zero.
As proposed here, these equations, when combined, will result in a failure to achieve
conservation of mass and comply with chemical thermodynamics for semi-volatile
i
organic chemicals. These equations should be modified to avoid this problem. The
approach used for the surface-water compartment could serve as an appropriate guide
for this modification.
The algorithm for estimating the contaminant concentration in animal tissue on
page 72 does not account for inhalation of contaminants by animals. Since inhalation is
considered a primary exposure pathway for human exposure, the role of this pathway
for contributing to animal tissue concentration should not be ignored.
The total contaminant concentration in plant tissues is expressed as the sum of
three contributions—root uptake, direct deposition, and air-to-plant transfers. This
model seems to be representing these uptake as primarily one-way transfers, with the
plants acting as a sink. Also, these three processes are represented as additive in
determining plant-tissue concentrations, even though root uptake of many contaminants
is confined to roots, air-to-plant uptakes are to leaves and plant surfaces, and
deposition is primarily to above ground surfaces. Plants are not necessarily well-mixed
homogenous systems as is implied here. The approach in the plan is illustrated in the
first diagram below.
With this sort of model there is a danger that one can define a chemical
concentration in plant tissues that has a higher chemical activity (i.e. fugacity) than
either the air or soil compartments that are contaminating it—a situation that is highly
unlikely. This can occur unless the plant-surface loss mechanism is selected-in a way
that maintains the plant in chemical equilibrium with its surroundings (that is, solubility
-------
T.E. McKone
Deposition
1
X
Root uptake
•Plant-
i
Air-to-pant
partitioning
Plant-surface loss
coefficient
limits in plant tissues are not exceeded relative to solubility limits in air and soil). The
condition illustrated above could violate chemical thermodynamics and could be
avoided if the potential mass-transfer pathways are structured as shown in the diagram
below. This diagram is based on the simple, but more realistic chemical potential
models proposed by
Bacci, E., D. Calamari, C. Gaggi, and M. Vighi (1990) "Bioconcentration of Organic
Chemical Vapors in Plant Leaves: Experimental Measurements and Correlation/'
Environ. Sci. Technol. 24, 885-889.
Paterson, S., and D. Mackay (1989) "Modeling the Uptake and Distribution of
. Organic Chemicals in Plants/' in Intermedia Pollutant Transport: Modeling and Field
Measurements, Allen, D. T., Y. Cohen, and I. R. Kaplan, Eds. (Plenum Press, New
York, NY), pp. 283-292.
Riederer, M. (1990) "Estimating Partitioning and Transport of Organic Chemicals in
the Foliage/Atmosphere System: Discussion of a Fugacity-Based Model," Environ.
Sd. Technol. 24, 829-837.
The model for plant uptake from soil described on page 77 does not account for
rainsplash transfer from soil to plant surfaces. A study of radioactive-fallout in soils
has shown that significant amounts of radioactivity can be transferred from soil to plant
tissue surfaces [Dreicer, M., T.E. Hakonson, G.C. White, and F.W. Whicker, 1984,
"Rainsplash as a Mechanism for Soil Contamination of Plant Surfaces," Health Physics
46,177-187]. In this study, it was shown that the value of the ratio of contaminant
concentration on plant surfaces in mg/kg(plant dry mass) to contaminant concentration
in soil is 1.7 x 10"2 with a geometric standard deviation of approximately 2.5.
-------
T.E. McKone
Plant-surface loss
coefficient
Resuspension
and
rainsplash
The recommended regression equations for plant-soil bioconcentration (Bri), beef
biotransfer factor [Ba(beef)L and milk biotransfer factors IBa(milk)] reported in Tables
B-9, and B-12 appear to be based on the work of Travis and Arms (1988). Although,
there appear to be no other estimation methods available for predicting these types of
bioconcentration, it should be recognized that these regression equations have large
estimation errors and were derived primarily from pesticide data. Their applicability to
incinerator emissions might be questioned. In a recent paper (McKone, T.E., 1993, "The
Precision of QSAR Methods for Estimating Intermedia Transfer Factors in Exposure
Assessments," SAR and QSAR in Environ. Res. 1, 41-51), I have determined the standard
error of the estimate for the three Travis and Arms (1988) regression equations to be as
follows
log Bq = 1.588 - 0.578 log Kow ± 0.73 (for plants)
log BaCbeef) = log Kow - 7.6 ± 0.95 (for beef)
log Ba(milk) = log J^w - 8.1 ± 0.84 (for milk)
-------
T.E. McKone
These correspond to respective estimation-error factors of 5.4, 8.9, and 6.9. This means
that, for estimating Ba(beef), for example, there is a 16% probability that the true value
of the biotransfer factor is more than 8.9 times higher than the median estimate provided
by the equation above and that there is a 5% probability that the true value of Ba(beef)
is more than 33 times (=0.891-65) larger than the median estimate from the regression
equation. These types of estimation errors could have a significant impact on the
outcome of the propose risk assessment and should be addressed. Also, the estimation
error for meats other than beef should be higher than the estimation error in beef, due to
inter-species extrapolation.
Dose Estimation
The procedures used to characterize average daily potential dose are
appropriate and consistent with EPA guidance. It is particularly useful to distinguish
between the exposure medium concentration and the human contact rates with these
media in constructing these models.
Evaluation of Routine Fugitive and Short-Term Episodic Emissions
It is appropriate and necessary that the routine fugitive emissions and short-term
episodic emissions be used to determine total potential dose as part of the risk
assessment.
Risk Characterization
At best, mathematical models only approximate real systems, and therefore their
predictions are inherently uncertain. In evaluating the reliability of risk assessment
models two questions must be asked, (1) how large is the uncertainty in the model
predictions and (2) how much confidence can be placed in the results? To address these
questions, exposure and risk should be presented so that uncertainty in risk and
exposure can be characterized by expectation (mean) and spread (variance). Thus, I
believe there is a need to include both a sensitivity analysis and a variance propagation
analysis to determine the reliability of the risk estimates and to identify what are the
components of the analysis that limit overall reliability.
-------
T.E. McKone
As I noted above, the ETFs—such the biotransfer factors in milk, meat, eggs,
breast milk, etc.—are derived from either controlled experiments or more often from
statistical estimation equations and their exact values are often highly uncertain. The
risk characterization proposed here seems to be more focused on characterizing
individual variability rather than issues of true uncertainty (i.e. the value of ITFs). If
only median estimates are used for ITF values, there will be a significant limit on the
ability of the exposure model to truly characterize our uncertainty about the upper and
lower ranges of the exposure expected for either the central tendency or high end of
exposure individuals.
Many intermedia transfer parameters (such as KOC, vapor pressure, soil-erosion
rates, and biotransfer factors, etc.) are based on empirical-estimation techniques. These
types of empirical estimation equations often have very large estimation errors—a
geometric standard deviation of as much as an order of magnitude is common. In order
to determine the reliability of the risk assessment there is need to determine how these
empirical estimation techniques limit the precision of the transport and uptake models.
There are two approaches for determining the impact of these estimation errors on
model reliability—(1) validation studies that compare model predictions to the
i
measured relationship between emissions and concentrations and
(2) sensitivity/uncertainty analyses that identify critical links between estimation errors
and model precision. Both activities need to be a continuing part of the risk assessment
•
process for the WTI incinerator.
It would be useful to make clear the distinction between the mean and the median
particularly as it applies to the distribution of exposure among individuals. It should be
noted that the median exposure is the value for which half of the population receives an
exposure greater than this value and half receive an exposure less than this value. In
contrast, the mean is the exposure that we expect to be associated with an individuals
drawn at random from a population, when a large number of individuals is selected. In
health risk assessments, the median of risk is likely to be more valuable than the mean
for addressing questions that relate to individuals. Whereas, issues that relate to
populations and allocation of public-health resources are likely to focus more on the
mean, that is the actual expectation of detriment within a specified population.
-------
Randy Seeker
-------
W. R. Seeker
Emissions Characterization
The most critical components of this characterization are the
• selection of substances to quantitatively evaluate (defining all species of concern)
• determining average emissions rates for these substances (defining how much of each of
these species is actually emitted during normal conditions within the permit operations)
• determining emissions under normal and accident upset conditions
In summary, the approach proposed is to use procedures to define the most important substances
from different routes of exposure and to use emissions data primarily from the WTI facility. If
data are not available from the WTI facility, then related data from similar facilities will be used.
This general approach as described in the project plan, appears to be well founded and structured.
However there exists several points that need to be addressed more fully in order to completely
define the emissions characterization. These points are presented below in no particular order:
1. Relative to the selection of the substances of interest listed hi Table 7, another category is
recommended to define other potential substances. There should be an examination of theoretical
chemical pathways that might exist for the particular constituents in the waste to be treated at this
site. Products of incomplete combustion pathways should be examined to define other possible
substances that are chemically probable even if they have not yet been quantified. A combustion
chemist should be consulted to define these other potential substances.
2. The use of the word surrogates is questioned. Surrogates usually means the use of compounds
other than those actually present. The understanding of the approach defined in the project plan is
that the actual substances found hi the stack will be used wherever possible. The use of the word
surrogate implies a greater reliance of substitutes than is actually envisioned. Surrogates
terminology should be reserved for those compounds that are actually selected to represent others.
3. There is concern about using prorated emissions rates based upon the total hazardous organic
emissions developed from the limited test data from EPA tests in 1987. It is not clear from the
discussions of this approach on page 34 of the project plan, how these results will be used. While
the facilities are both rotary kilns, many critical features of the units are different including the air
pollution control systems. How will the total hazardous organic emission rate actually be defined?
Will the ratio of total hydrocarbon emissions from the two units be used to relate the different
-------
W. R. Seeker
operating performance of the two units? How will the emissions rates be scaled? These testing
program where still only able to quantify 60-90 % of the total organics measured. In addition, it is
not clear that this approach represents the best method for quantifying emissions of hazardous
constituents not quantified by direct measurements. The emissions, as actually measured from the
WTI unit, is the most reliable measure of an individual substance emissions rate. The use of the
prorated approach which increases the measured rate of that substance based upon other data from
another, albeit somewhat similar unit, is artificial and does not likely stand up to scientific scrutiny.
It would appear to be better to quantify the risks for which reliable data are available first, and then
to examine other substances separately so that the uncertainties do not become mixed and
intractable. If it is determined that these unqualified substances have to be emitted in very large
concentrations in order to be of importance, then they can be ignored on the basis of the total
hazardous emissions arguments. Clearly not all substances can be quantified by either method but
it would seem more scientifically correct to examine the uncertainties associated with the absence of
data than to artificially change reliable data.
4. There is a need to carefully examine data from other emissions sources and not just rely on
compilations such as the 1989 background documents. It is not clear that there has been a
systematic evaluation of these data. They can have variable QA/QC, test reporting inaccuracies,
reliability, and sampling and analysis protocols. Often the emissions factors reported in
compilations are highly variable in reliability and quality. The approach should include going back
to original test reports and making an evaluation of the reliability and representativeness of the data
before they are used. For metals, mass balances should be examined and it may be reasonable to
compare the theoretical partitioning of metals to see if the data are theoretically consistent. Only by
examining the actual test reports is it possible to judge the appropriateness of the data for this
application.
5. There seems to be a mix of logic in the selection of emissions characterization. There are several
scenarios/assumption categories that need to be made more consistent in reasoning and
nomenclature. For example, there is the first scenario (average emissions as measured), second
scenario ( referred to as the conservative estimates), prorated data using total hazardous organic
emissions data, scenarios for upsets and scenarios for accidents. It would be useful if these were
summarized in one location and put into a consistent reasoning that builds on each new added
complexity. This is particularly important since the public will view the risk assessment numbers
from this study with a certain level of authority and these different risk assessments based upon
-------
W. R. Seeker
different assumption could lead to significant confusion. The various scenarios and the logic
associated with selecting them needs to be self consistent and very clear. The second scenario is
i
referred to as conservative in several places. That is not a good descriptor of the scenario even
though it is likely more conservative than the first scenario. Even scenario one is conservative
therefore this nomenclature does not reflect the true meaning. If this scenario is designed to
represent the permit limits then it should be stated more clearly. Finally, the integration of these
scenarios with the uncertainty analysis to be carried out in the risk assessment should be better
defined.
6. The database for emissions from start up and shut down, fugitive emissions, and accidents
appears to be weak and the attempt to use these data may weaken the credibility of the study. It
may be better to review the data quality early on in the program and determine whether they are
reliable enough to use for this purpose. An alternative approach needs to be developed to address
these important scenarios if the data are not adequate. One alternative approach would be to define
how much emissions would have to increase hi order to have a significant risk and then to examine
if any scenarios could lead to this level of emissions. This approach could eliminate the need to
quantify emissions during these transients and the available data could be used to determine if the
order of magnitude of the emissions increases are troublesome.
7. The use of data from the Biebesheim Germany facility needs to be done with great care due to
the differences hi sampling and analytical protocol between Germany and the United States and
differences in waste feed characteristics. The facility could be a good candidate as a source of data
that are representative of this type of equipment but often differences in protocols and waste can
significantly bias results.
8. There are two sets of test data from the WTI facility for tests conducted in March 1993 and
August 1993. The plan currently states that both sets of data will be used. In particular, the
PCDD/PCDF data from the August 1993 results will be used while most of the other emissions
data will be from the March 1993 tests. It is extremely important to compare and contrasts the
results of these two tests in order to ensure that they are consistent. Are there differences in
operating conditions? Are the only differences in the tests the carbon injection locations and rates?
If the carbon injection system was run in the old manner (as it was in the trial burn) would
PCDD/PCDF increase back to these higher levels or are there other changes in operating
conditions that could be important? These are critical questions that must be addressed due to the
-------
W. R. Seeker
importance of the PCDD/PCDF emissions levels on risks, and therefore the plan must address how
these results will be examined.
9. It is not clear how other metals that have not been measured in the stack of the WTI facility will
be quantified. How will other data from other facilities be used and related to the different design,
operating and metals feed characteristics? It is important to factor in the impacts of the higher
temperature kiln which promotes vaporization of the metals, the ESP performance vs particle size,
and the different behaviors of different physical/chemical forms of metals in different waste
streams.
10. The method for determining emissions solid/vapor partitioning appears to be a simple
extrapolation using Langmuir adsorption isotherms. These isotherms have a number of simplifying
assumptions such as less than monolayer build up of material, homogeneous sites, etc. The data
from Bidleman should be examined carefully to see if these assumptions are holding up in practice.
The plan indicates that for dioxin there does need to some modifications to the simple models. A
more thorough examination of this subject is recommended.
8. Miscellaneous comments:
• The statement on page 38 second paragraph, " The data from such facilities would be
included ... only if they can reliably be extrapolated..." needs further explanation.
How will this be determined?
• Page 36 Table 8. TICs for tentatively identified compounds is an unnecessary acronym
• Particulates is not a noun. It should either be particles or particulate matter.
• The statement on page 35 third paragraph, " The trial burn,... under operating conditions
intended to represent conditions worse than normally anticipated." is not
necessarily correct. The trial burns are generally designed to define the acceptable
operating envelope but are not intended to be necessarily worse than normal.
• The statement on page 42, second paragraph, " Although most of the metals would be
captured in solid form in the slag and ash..." should be investigated carefully. The
more volatile metals such as Arsenic, Cadmium, Selenium and Lead will be highly
vaporized in this slagging high temperature kiln. A significant fraction of these may
escape capture in the electrostatic precipitator.
• Of particular importance is hexavalent chromium but there is no discussions of the
reliability of these data or how the reliability will be evaluated. Table 8, page 36
-------
W. R. Seeker
indicates that hexavalent chromium was measured in the March 1993 trial burn.
More details of how these data will be assessed for reliability and sampling/analysis
protocols would be useful. How do these data compare with other hexavalent
i
chromium data sets?
• The approach to be used for examining the fate of mercury described on page 54 is not
clear. How will the forms of mercury be determined and how will the
transformations from inorganic to organic mercuric forms be assessed?
• Relevant to scenario uncertainties discussed on page 102, wouldn't it be appropriate to
define what types of future activities would result in significant risks?
-------
Marvin Tabor
-------
M.W. Tabor
COMMENTS
WTI RISK ASSESSMENT PROJECT PLAN
PART I. General Features
Generally, the project plan is designed to meet
currently accepted practices, as defined by USEPA, for the
conduct of an adequate risk assessment of the impact of a
facility like the WTI facility. Strengths of the plan
include: l] general characterizations of "local"
populations in terms of likely routes of exposure, "local"
activities, including agricultural, and geography; and 2]
pilot data on the facility operation with a preliminary
indication of the characteristics of the emissions to. be
expected from the incineration of defined waste streams.
The . plan could include additional considerations• of the
types of chemicals expected in the emissions from the
incineration of the anticipated wastes, e.g. waste paint
and byproducts or process sludges from chemical
manufacturers, that the facility will be processing. More
consideration to emissions from incineration of the
l
broader classes of wastes described in the facility's RCRA
permit would enhance the risk characterization. This is
particularly important if this facility will be processing
these materials during the actual operational lifetime of
the incinerator. A general weakness of the plan is the
lack of a critical evaluation of the peer-reviewed
scientific literature in terms of the emissions, fate,
transport, exposure and expected toxicological end-points.
-------
M.W. Tabor
COMMENTS
Additionally, the literature cited in the document
generally was inadequate, incomplete [e.g. parts of
references missing like with Hulster & Marschner], and
unavailable, making a thorough evaluation of the plan
difficult.
PART II. Specific Features
Generally the emission characterization should include
those compounds and/or classes of compounds expected in
the emissions from incinerating the types of wastes
expected at the facility and those compounds found in the
trial burn emissions. - Substances such as hydrogen
cyanide, oxidized organics [e.g. aldehydes and acids], and
nitrogenous organics [e.g.nitro, nitroso, amino compounds]
should be included. A more careful evaluation of data
reported from previous studies would provide•a more
accurate reflection of expected emissions and subsequently
lead to a better risk assessment.
The specific approaches to dose-response evaluation
should possibly be expanded to at least include more
information on arsenic and cadmium, since both metals are
characteristically found in stack emissions from many
types of incineration facilities, and on a number of the
expected organics listed in Table 9.
In terms of the exposed populations selected for the
exposure assessment, are the demographics of the region
adequately reflected in these selected groups? For
example should individuals with higher breathing rates,
e.g. training athletes, joggers, hard-laborers, etc., be
-------
M.W. Tabor
COMMENTS
reflected as a sub-population? The selection of the
population sub-groups for the assessment was not entirely
clear from the information presented.
•In terms of the exposure routes, no discussion of
exposure via household dust was included. Since this is a
significant route of exposure for one sensitive sub-group,
i.e. infants and young children, the question of inclusion
of this route of exposure is raised for discussion.
In terms of environmental and exposure concentrations
of the substances of concern, the approaches described for
the environmental media are according to generally
accepted practices. However, some details are not
entirely obvious. For example, a clear summary and
justification of the assumptions used for each aspect of
i
the modeling should be highlighted. Additionally, some of
the information to be utilized, e.g. the one year
meteorological data, are questioned. Are these data
typical of the region, reflecting expected conditions and
i
trends? Another issue is the question of the
biotransformation of substances during food chain
accumulation. Should the modeling equations reflect
potential losses due to biotransformation rather than just
focusing on bioconcentration? In terms of dose "to the
individual, will discriminations be made among absorbed
dose, internal dose, bioeffective dose, etc.? These are
examples of the type of issues that need further
clarification in the exposure assessment. Furthermore, it
is stated [p.101] attempt will be made "to identify and
-------
M.W. Tabor
COMMENTS
evaluate the important sources of uncertainty., in the
exposure assessment". This is crucial to .the overall
credibility of the final risk assessment.
-------
-------
APPENDICES G-L
OBSERVERS AND OBSERVER MATERIALS
Because some of the materials in these appendices were distributed by observers at the
workshop, visual quality may vary.
-------
-------
APPENDIX G
FINAL LIST OF OBSERVERS
G-l
-------
U.S. Environmental Protection Agency
Technical Workshop on
WI1 Incinerator Risk Issues
Holiday Inn Capitol
Washington, DC
December 8-9,1993
List of Observers
Preregistered and Registered On Site
Vincent Adams
Program Manager
U.S. Department of Energy
Building 1435K (MS-7345)
K35 Plant
Oak Ridge, TN 37830
615-576-1803
Fax: 615-241-3703
Elmer Akin
Chief, Office of Health Assessment
Waste Management Division
US. Environmental Protection Agency
345 Courtland Street, NE
Atlanta, GA 30365
404-347-1586
Fax:404-347-5205
M. Joy Allison
Route 1
P.O. Box 855
Chester, WV 26034
304-387-2259
Ann Anderson
Regional Manager
A.T. Kearney, Inc.
222 West Adams Street
Chicago, IL 60606
312-223-6230
Fax:312-223-6372
Lauren Anderson
P.O. Box 180
Montclair,NY 07042
201-492-7701
Robert Beliles
Toxicologist
3002 PhyKmar Place
Oakton.VA 22124
202-260-3018
Fax:202-260-3808
Mitchell Bernstein
Partner
Skadden, Arps, Slate,
Meagher, & Flom
1440 New York Avenue, NW
Washington, DC 20005
202-371-7220
Fax: 202-371-7959
Linda Birnbaum
Director, Environmental
Toxicology Division
Health Effects Research Laboratory
U.S. Environmental Protection Agency
(MD-616)
Research Triangle Park, NC 27711
919-541-2655
Fax: 919-541-4324
Pamela Blakley
Air Emissions Expert
U.S. Environmental Protection Agency
77 West Jackson Boulevard (HRP-8J)
Chicago, IL 60604
312-886-4447
Fax: 312-353-4788
John Blandamer
Aptus, Inc.
P.O. Box 1328
Coffeyvffle, KS 67337
316-251-6380
Fax: 316-251-7498
Ruth Bleyler
Risk Assessor
Waste Management Division
U.S. Environmental Protection Agency
JFK Federal Building (HPR-CAN1)
Boston, MA 02203
617-573-5792
Fax:617-573-9662
Fred Blosser
Program Advisor
Office of Research & Development
U.S. Environmental Protection Agency
401 M Street, SW (MC-8101)
Washington, DC 20460
202-260-7449
Fax:202-260-9761
Jane Bobbin
Legislative Assistant
Office of Congressman Alan Mollohan
2242 Rayburn Boulevard
Washington, DC 20515
202-225-4172
Fax: 202-225-7564
Karl Bremer
Chief, RCRA Permitting Branch
U.S. Environmental Protection Agency
77 West Jackson Boulevard (HRP-8J)
Chicago, IL 60604
312-353-0398
Fax: 312-353-4788
Cyndy Bryck
Manager, Solid Waste Issues
Chemical Manufacturers Association
2501 M Street, NW
Washington, DC 20037
202-887-1290
Fax:202-887-1237
David Burmaster
President
Alceon Corporation
P.O. Box 2669
Cambridge, MA 02138
617-864-4300
Fax: 617-864-9954
-------
Dorothy Canter
Science Advisor
Office of Solid Waste &
Emergency Response
U.S. Environmental Protection Agency
401 M Street, SW (MC-5101)
Washington, DC 20460
202-260-3100
Fax: 202-260-8929
Alan Carpien
Attorney
Office of the General Counsel
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202-260-7217
Fax: 202-260-0584
David Cleverly
Environmental Scientist
Office of Health &
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW (MC-8603)
Washington, DC 20460
202-260-8915
Fax: 202-260-1722
Maxine Conner
Division Secretary
Martin Marietta Energy Systems, Inc.
P.O. Box 2003
(K-303-8, MC-7314)
Oak Ridge, TN 37831
615-576-2474
Fax: 615-576-8582
Pat Costner
Research Director
Greenpeace
P.Q. Box 548
Eureka Springs, AR 72632
501-253-8440
Fax: 501-253-5540
Elizabeth Cox
Environmental Scientist
U.S. Food & Drug Administration
200 C Street, SW (HFS-246)
Washington, DC 20204
202-254-9597
Fax: 202-254-3982
Harriet Croke
Chief, Ohio Permitting Section
U.S. Environmental Protection Agency
77 West Jackson Boulevard (HRP-8J)
Chicago, IL 60604
312-353-4789
Fax: 312-353-4788
Edmund Crouch
Senior Scientist
Cambridge Environmental, Inc.
58 Charles Street
Cambridge, MA 02141
617-225-0810
Fax: 617-225-0813
Clyde Dempsey
Acting Director
Waste Minimization, Destruction &
Disposal Research Division
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, OH 45268
513-569-7546
Fax: 513-569-7549
Arnold Den
Senior Science Advisor
U.S. Environmental Protection Agency
75 Hawthorne Street
San Francisco, CA 94105
415-744-1018
Fax: 415-744-2499
Robin Depot
Project Manager
Northeast Maryland Waste
Disposal Authority
25 South Charles Street
Baltimore, MD 29201
410-333-2730
Fax: 410-333-2721
Niaz Dorry
Campaign Director
Greenpeace Toxics Campaign
1436 U Street, NW
Washington, DC 20009
202-462-1177
Fax: 202-462-4507
Craig Evans
Technical Support Section
Pennsylvania Department of
Environmental Resources
400 Market Street
P.O. Box 8468 - 12th Floor
Harrisburg,PA 17105-8468
717-787-9256
Fax: 717-772-2303
William Farland
Director, Office of Health &
Environmental Assessment
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202-260-7315
Fax: 202-260-0393
Kaha Fatah
Greenpeace
1436 U Street, NW
Washington, DC 20009
202-319-2598
Sarah Foster
Weinberg Consulting Group, Inc.
1220 19th Street, NW - Suite 300
Washington, DC 20036-2400
202-833-8077
Fax: 202-833-7057
Stephen Fotis
Associate
Van Ness, Feldman & Curtis
1050 Thomas Jefferson Street, NW
Seventh Floor
Washington, DC 20007
202-298-1908
Fax: 202-338-2416
Shiva Garg
Environmental Engineer
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
703-308-8459
Fax: 703-308-8460
Richard Gillam
Environmental Engineer
Waste Management Division
U.S. Environmental Protection Agency
345 Courtland Street, NE
Atlanta, GA 20265
404-347-3433
Fax: 404-347-5205
Robert Giraud
Senior Engineer
DuPont Company
655 Papermill Road
Newark, DE 19711
302-366-6906
Fax: 302-366-4123
Robert Golden
Principal
ENVIRON Risk Sciences
1000 Thomas Jefferson Street, NW
Suite 506
Washington, DC 20007
202-965-5188
Fax: 202-965-5187
-------
Manuel Gomez
Designated Federal Official
Science Advisoiy Board
U.S. Environmental Protection Agency
401 M Street, SW (MC-1400F)
Washington, DC 20460
202-260-2563
Fax:202-260-7118
Laura Green
President and Senior Scientist
Cambridge Environmental, Inc.
58 Charles Street
Cambridge, MA 02141
617-225-0810
Fax: 617-225-0813
Gary Gross
Environmental Engineer
U.S. Environmental Protection Agency
841 Chestnut Building (3HW53)
Philadelphia, PA 19107
215-597-7840
Fax:215-580-2013
Brian Harris
Assistant Campaigner-at-Large
Greenpeace
1436 U Street, NW
Washington, DC 20009
202-319-2598
Fax:202-462-1177
Jack Heller
Master Consultant
U.S. Army Environmental
Hygiene Agency
Aberdeen Proving Ground
MD 21010-5422
410-671-2953
Fax: 410-671-5237
Rick Hind
Legislative Director
Greenpeace Toxics Campaign
1436 U Street, NW
Washington, DC 20009
202-462-1177
Fax:202-462-4507
Dwight Hlustick
Waste Combustion Specialist
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (5303W)
Washington, DC 20460
703-308-8647
Fax: 703-308-8617
Bob Holloway
Chief, Combustion Section
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (MC-5302W)
Washington, DC 20460
703-308-8461
Fax: 703-308-8604
Adam Johnston
ENVIRON
4350 North Fairfax Drive
Suite 300
Arlington, VA 22203
703-516-2300
Fax: 703-516-2345
J. William Jordan
Program Director
A.T. Kearney, Inc.
225 Reinekers Lane
Alexandria, VA 22314
703-739-1608
Fax: 703-836-0547
Anthony Kahafy
Environmental Engineer
U.S. Environmental Protection Agency
26 Federal Plaza (2AWM-HWF)
New York, NY 10278
212-264-9401
Fax: 212-264-8100
Metvin Keener
Director, Federal Affairs
Laidlaw, Inc.
655 15th Street, NW - Suite 300 '
Washington, DC 20005
202-639-4130
Fax:202-347-6109
Lou Kerestesy
Office of Regional Operations &
State/Local Relations
U.S. Environmental Protection Agency
401 M Street, SW (MC-1502)
Washington, DC 20460
202-260-2294
Fax:202-260-9365
Holty Kimball
Weinberg Consulting Group, Inc.
1220 19th Street, NW
Suite 300
Washington, DC 20036-2400
202-833-8077
Fax:202-833-7057
Jonathan Kiser
Director, Waste Services Programs
Integrated Waste Services Association
1133 21st Street, SW - Suite 205
Washington, DC 20036
202-467-6240
Fax: 202-467-6225
Stephen Kroner
Environmental Scientist
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202-260-5219
Fax: 202-260-0225
Maryann Lamont
Vice President and Counsel
Don Clay Associates
1701 Pennsylvania Avenue
Washington, DC 20006
202-879-2693
Fax: 202-879-2697
David Layland
Environmental Engineer
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202-260-4796
Fax: 202-260-0225
Gary Liberson
Environmental Consultant
ENVIRON Risk Sciences
1000 Thomas Jefferson Street, NW
Suite 506
Washington, DC 20007
202-965-5188
Fax: 202-965-5187
Matthew Lorber
Environmental Engineer
Office of Research and Development
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202-260-8924
Fax: 202-260-1722
Ranjit Machado
ENVIRON
4350 North Fairfax Drive - Suite 300
Arlington, VA 22203
703-516-2300
Fax: 703-516-2345
-------
Mario Mangino
Toxicologist
U.S. Environmental Protection Agency
77 West Jackson Boulevard (HRP-8J)
Chicago, IL 60604
312-886-2589
Fax: 312-353-4788
Frank McAlister
Chief, Assistance Branch
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (MC-5302W)
Washington, DC 20460
703-308-8612
Fax: 703-308-8617
Alec McBride
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (MC-5302W)
Washington^ DC 20460
202-260-4761
Mark Mercer
Environmental Engineer
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (MC-5302W)
Washington, DC 20460
703-308-8648
Fax: 703-308-8617
Gary Metcalf
President
LWD, Inc.
P.O. Box 327
Calvert City, KY 42029
502-395-8313
Fax: 502-395-8153
James Michael
Chief, Disposal Technology Section
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (MC-5302W)
Washington, DC 20460
703-308-8610
Fax: 703-308-8617
Ian Moar
Executive Director
Coalition for Responsible
Waste Incineration
1133 Connecticut Avenue, NW
Suite 1200
Washington, DC 20036
202-775-9869
Fax: 202-775-2395
Connie Nakahara
Utah Bureau of Solid &
Hazardous Waste
P.O. Box 16700
Salt Lake City, UT 84116
801-538-6170
Fax: 801-538-6715
Beth Newman
Midwest Campaigner
Greenpeace
1017 West Jackson Boulevard
Chicago, IL 60607
312-666-3305
Fax: 312-226-2714
Deborah Ng
Utah Bureau of Solid &
Hazardous Waste
P.O. Box 16700
Salt Lake City, UT 84116
801-538-6170
Fax: 801-538-6715
William Omohundro
RCRA Public Involvement Coordinator
Office of Public Affairs
U.S. Environmental Protection Agency
77 West Jackson Boulevard (P-19J)
Chicago, IL 60604
312-353-8254
Fax: 312-353-1155
Marilyn Parkes
2487 Woodbine Avenue
East Liverpool, OH 43920
216-385-2527
Dorothy Patton
Executive Director
Risk Assessment Forum
U.S. Environmental Protection Agency
401 M Street, SW (MC-8101)
Washington, DC 20460
202-260-6743
Fax: 202-260-3955
Delta Pereira
Environmental Protection Specialist
Office of Environmental Equity
U.S. Environmental Protection Agency
401 M Street, SW (MC-2103)
Washington, DC 20460
202-260-3565
Fax: 202-260-0852
Bonnie Piper
Deputy Director, Press Division
Office of Communications,
Education & Public Affairs
U.S. Environmental Protection Agency
401 M Street, SW (Room 311W)
Washington, DC 20460
202-260-4366
Suellen Pirages
Executive Director
Karen and Associates
1701 K Street, NW - Suite 1000
Washington, DC 20006
202-463-0400
Fax: 202-463-0502
Mick Pompelia
Vice President
DMG Environmental, Inc.
21 Yost Boulevard - Suite 202
Pittsburgh, PA 15221
412-824-2355
Fax: 412-824-0131
Amy Porter
Daily Environment Report
Bureau of National Affairs
1231 25th Street, NW (S-361)
Washington, DC 20037
202-452-4106
Fax: 202-452-4105
Terry Quill
Attorney
Beveridge & Diamond, P.C.
1350 I Street, NW
Washington, DC 20005
202-789-6061
Fax:202-789-6190
Ranama Rao
Senior Engineer
Department of Environmental Protection
Montgomery County Government
101 Monroe Street - Sixth Floor
Rockville.MD 20850
301-217-6708
Fax: 301-217-6718
William Raub
Science Advisor to the Administrator
U.S. Environmental Protection Agency
401 M Street, SW (A-100)
Washington, DC 20460
202-260-7960
Fax: 202-260-3684
-------
Ed Repa
Director, Technical & Research Program
National Solid Waste
Management Associates
1730 Rhode Island Avenue, NW
Suite 1000
Washington, DC 20036
202-659-4613
Fax:202-775-5917
Glenn Rice
Facilities - Room 190
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, OH 45268
513-569-7813
Fax:513-569-7916
Sam Rondbcrg
Program Analyst
Science Advisory Board
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
202-260-2559
Fax:202-260-7118
Alan Rubin
Office of Water
U.S. Environmental Protection Agency
401 M Street, SW(430A)
Washington, DC 20460
202-260-1311
Fax:202-260-1026
Sonya Sasseville
Chief, Alternative Technologies Section
Office of Solid Waste
U.S. Environmental Protection Agency
401 M Street, SW (MC-5303W)
Washington, DC 20460
703-308-8648
Fax: 703-308-8617
John Schaum
Clucf, Exposure Assessment
Methods Branch
Office of Research and Development
U.S. Environmental Protection Agency
401 M Street, SW (MC-8603)
Washington, DC 20460
202-260-5988
Fax:202-260-1722
Herbert Schumann
Assistant Manager, North America
STEAGAG
1050 Thomas Jefferson Street, NW
Washington, DC
202-338-8767
Fax:202-338-7351
Joseph Scire
Vice President
Sigma Research Corporation
196 Baker Avenue
Concord, MA 01742
508-371-4270
Fax: 508-371-2468
Karl Sheaffer
Director, Hazardous Waste Siting
Pennsylvania Department of
Environmental Resources
P.O. Box 2063
Harrisburg, PA 17105-2063
717-783-8977
Fax: 717-772-3314
FredSigg
Director of Technology
von Roll, Inc.
3080 Northwoods Circle
Norcross, GA 30071
404-729-0500
Fax: 404-729-0403
Damu Smith
Campaigner
Greenpeace
1436 U Street, NW
Washington, DC 20009
202-232-5690
Fax: 202-462-4507
Greg Smith
7521 Weatherby Drive
Durwood,MD 20855
301-258-8479
Ann Southwick
Associate
Van Ness, Feldmont, & Curtis
1050 Thomas Jefferson Street, NW
Washington, DC 20007
202-298-1846
Fax:202-338-2416
Edward Sowinski
Advisor to East Liverpool
Board of Health
6758 St. Regis Boulevard
Hudson, OH 44236
216-655-2288
Fax: 216-653-3366
Alonzo Spencer
President
Save Our County, Inc.
P.O. Box 1242
East Liverpool, OH 43920
216-385-4584
Clare Stine
Risk Assessment Forum
U.S. Environmental Protection Agency
401 M Street, SW (MC-8101)
Washington, DC 20460
202-260-6743
Fax: 202-260-3955
Terri Swearingen
Tri-State Environmental Council
RD1 Box 365
Chester, WV 26034
304-387-0574
Fax: 304-387-0574
Robert Taylor
Attorney '
Swidler and Berlin
3000 K Street, NW
Washington, DC 20007
202-424-7520
Fax: 202-424-7643
Shirley Thomas
Risk Assessment Forum
U.S. Environmental Protection Agency
401 M Street, SW (MC-8101)
Washington, DC 20460
202-260-6743
Fax: 202-260-3955
Joseph Thornton
Researcher
Greenpeace Toxics Campaign
1436 U Street, NW
Washington, DC 20009
202-462-1177
Fax: 202-462-4507
Lynn Thorp
Director
Greenpeace Toxics Campaign
1436 U Street, NW
Washington, DC 20009
202-462-1177
Fax: 202^62^507
Eric Tiemeyer
Senior Environmental Engineer
Texas Industries, Inc.
7610 Stemmons Freeway
Dallas, TX 75247
214-647-3879
Fax: 214-647-3$77
Jim Turpin
Director, Government Affairs
A.M. Nukem
3444 South Wakefield Street
Arlington, VA 22206
703-379-7593
Fax: 703-379-6612
-------
Stephen Washburn
Principal
ENVIRON Corporation
210 Carnegie Center - Suite 201
Princeton, NJ 08540
609-452-9000
Fax: 609-452-0848
Michael Wigmore
Attorney
Swidler and Berlin
3000 K Street, NW
Washington, DC 20007
202-424-7792
Fax:202-424-7643
William Wood
Associate Director
Risk Assessment Forum
U.S. Environmental Protection Agency
401 M Street, SW (8101)
Washington, DC 20460
202-260-6743
Fax: 202-260-3955
Robin Woods
Press Officer
U.S. Environmental Protection Agency
401 M Street, SW (MC-1703)
Washington, DC 20460
202-260-4377
Fax:202-260-0186
Gary Worley
Air Quality Team Leader
Martin Marietta Energy Systems, Inc.
P.O. Box 2003
(K-303-8, MS-7314)
Oak Ridge, TN 37831
615-241-2591
Fax: 615-576-5380
Stephen Zemba
Senior Engineer
Cambridge Environmental, Inc.
58 Charles Street
Cambridge, MA 02141
617-225-0810
Fax: 617-225-0813
-------
-------
APPENDIX H
OBSERVER TECHNICAL QUESTIONS
H-l
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTT Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon. !
Please circle the topic that your question pertains to:
• Combustion Engineering * Meteorology/Air-Dispersion Modeling
• Exposure Assessment » Toxicology
NAME:
QUESTION://.
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WIT Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
Combustion Engineering " - • Meteorology/Air-Dispersion Modeling
• Exposure Assessment • Toxicology
£/ . / / i /7 / s-
A
QUESTION:
//
?
(W.\rt
~
.flg i '/'
-------
I
OBSERVER TECHNICAL QUESTIONS
I
Technical Workshop on WIT Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circIeJhetopic thatyour question pertains to:
Combustion Engineering ^x» Meteorology/Air-Dispersion Modeling
Exposure Assessment • Toxicology
f pvL
i
' X,/
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WIT Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
» Combustion Engineering • Meteorology/Air-Dispersion Modeling
Exposure Assessment • Toxicology
NAME:
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993 '
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
i
Please circle the topic that your question pertains to:
M Combustion Engineering • Meteorology/Air-Dispersion Modeling
Exposure Assessment ^\ • Toxicology
NAME:
QUESTION:
d
v^zXgv ek: c_&^.(ct^
d
Z>
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
^ — •-
Please circle the topic that your question pertains to:
i)_mbustionEngineering • Meteorology/Air-Dispersion Modeling
Exposure Assessment"-. • Toxicology
NAME:
QUESTION:
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WIT Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
» Combustion Engineering • Meteorology/Air-Dispersion Modeling
.—•— ' "•-., ' , , , , i
Exposure Assessment \ • Toxicology
NAME:
QUESTION:.
&T
££ i <*i&£r & «.*•}
itrf- W
.V
Jr.
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
Exposure Assessment 8 • Toxicology
NAME: /mm \i r\ iflr kg.
c/>n<>ipraT-Ati
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WIT Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
• Toxicology
Exposure Assessment \
NAME:
QUESTION:.
r f\n
/A
ir
tffe
cf-Ttf£-
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• ^jCombustion Engineering • Meteorology/Air-Dispersion Modeling
• Toxicology
Exposure Assessment
NAME:
QUESTION:
,- u
7
*~ loo/me
~7
Tt-rd
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
Exposure Assessment ^ • Toxicology
NAME:
QUESTION:.
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
•_ ....... ..^QnLbjusjtionJEngineering • Meteorology/Air-Dispersion. Modeling
'•Vrx..
Exposure Assessment ' • Toxicology
NAME:
..- 5-rgr/teo
QUESTION:
/
Tt
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
* • Exposure Assessment
*-^-^ _^^-^--
NAME: .flwV^ ff/n.
QUESTION:.
^-/ __ '
• -" '": j>-
,'Lt £'5
y
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
» Combustion Engineering • Meteorology/Air-Dispersion Modeling
• Toxicology
» Exposure Assessment
NAME:
g.
QUESTION: H' <>*&& I '?
? r
'^f-
y
dt^c
^mf~j
T
-------
OBSERVER TECHNICAL QUESTIONS
r
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993 ,
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
» Exposure Assessment / »
Toxicology .
NAME:
QUESTION:
C
in ,c -^ /< s&'H -
-
r€-
\/# n_
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below'please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
• Exposure Assessment (^ • Toxicology
NAME:
QUESTION: THiL
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon. | ~
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
i
• Exposure Assessment djf Toxicology
NAME:
4
QUESTION
: C\\ '<-f tA \X &** --
\V
JO
vT
C . 4>t/u.
. -7
—^—
•J
.|)clv\
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12-00
noon.
Please circle the topic that your question pertains to:
» Combustion Engineering • • Meteorology/Air-Dispersion Modeling
^-
• Exposure Assessment
NAME:
QUESTION:
h////
a < &. it
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9, 1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
I
« Combustion Engineering • Meteorology/Air-Dispersion Modeling
• Exposure Assessment £ • Toxicology ;
NAME:
QUESTION:.
5 tujO'' jPt-/"Vv'f-w <" /"v&Cr U' I J-— t'},•«'<*// is*-+L.
1_. ::—^ a . - • j- »
'lV-;y b&J-'i..' b«
/
LI/ rr^
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon. <-
Please circle the topic that your question pertains to:
• Combustion Engineering • ^^-Meteorolpgy/Air-Dispersion Modeling
• Exposure Assessment C •
Toxicology
NAME:
QUESTION:
V
£-1
A . '
/ 5
-------
OBSERVER TECHNICAL QUESTIONS
e.^
Technical Workshop on WTI Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
Meteorology/Air-Dispersion Modeling
"Exposure ^Assessment / • Toxicology ^"^\
NAME:
(R- ^ C
W
QUESTION:
^p/^ SV1'
? /'
5 .
tive
a.* Cc'^Ptu^fz C-Jltr
C(/
-------
OBSERVER TECHNICAL QUESTIONS
Technical Workshop on WTT Incinerator Risk Issues
DECEMBER 8-9,1993
Below please state your question concisely. Drop off this form at the registration desk by 12:00
noon.
Please circle the topic that your question pertains to:
• Combustion Engineering • Meteorology/Air-Dispersion Modeling
j, -'
• Exposure Assessment • /Tbxicology
NAME: A JO KT- c
QUESTION:
*/-
-------
-------
APPENDIX I
OBSERVER COMMENTS
1-1
-------
OBSERVER COMMENTS
PLEASE SIGN-UP BELOW
Technical Workshop on WTI Incinerator Risk Issues
4:15PM, WEDNESDAY, DECEMBER 8, 1993
AFFILIATION
/ AlV A
3
4-
lan \\jf\
5-
/
k
v /'
< ^
12'
14.
15.
16.
-------
APPENDIX J
CITY OF EAST LIVERPOOL, OHIO, DEPARTMENT OF PUBLIC HEALTH
LETTER TO EPA REGIONAL ADMINISTRATOR
J-l
-------
V »"•
EAST
LIVERPOOL
OHIO
DEPARTMENT OF PUBLIC HEALTH
OH-HO,
43920
Phone 385-7900
Mr. Valdas V. Adamkus
EPA Regional Administrator
Environmental Protection Agency Region 5
77 West Jackson Blvd
Chicago, Illinois 60604-3507
December 6, 1993
Dear Mr. Adamkus:
Re: Comments on "WTI Phase II Risk Assessment Project Plan"
EPA ID # OHD980613541; November 1993
This letter outlines general technical comments concerning the
proposed "WTI Phase II Risk Assessment Project Plan". These
comments are offered in view of the "Technical Workshop on WTI
Incinerator Risk Issues" to be held in Washington DC on
December 8 and 9, 1993.
The East Liverpool Department of Health is concerned with
protecting the health and well being of the citizens of East
Liverpool. Acting under the direction of the East Liverpool
Board of Health and the Ohio Department of Health, the East
Liverpool Health Department has a long familiarity and concern
for the potential impact of the WTI incinerator on the health
of the community. We appreciate the opportunity to offer
comments on WTI Risk Assessment issues.
A brief description of programs and approaches which we have
advocated, some of which are underway, may be helpful to gain
perspective on our position and involvement in this issue. In
November of 1992 The Board of Health passed a resolution, based
on public input, which summarized the need for various
community monitoring programs (Attachment). The resolution was
aimed at encouraging affected parties, including WTI, to
develop background data on pollutants of concern within the
community before operation of the incinerator. It was further
recommended that comprehensive ongoing monitoring for specific
process and fugitive emissions, including metals and organics,
should occur and ongoing baseline evaluations of pollutants of
concern should continue within the community on an ongoing
-------
- 2 -
basis. Additionally, monitoring programs for lead in blood for
school age children and mercury in urine were advanced. The
community was involved in selection of environmental monitoring
sites for lead, encouraging lead screening and other aspects of
the program. In order to gain public acceptance, any technical
program must represent public input. The thrust of our
activity has been to develop specific localized data, based on
community input, which can indicate whether or riot
environmental exposures are associated with WTI, and if
exposures are found to be increasing, have the capacity to
bring about corrective action before harm can occur.
Unfortunately, environmental monitoring programs have developed
slowly and in a piecemeal fashion. This is due to a variety of
nonscientific reasons although good progress has been made in
some areas. For example, extensive lead in blood data have
been collected in the school age population through an Ohio
Department of Health program, baseline air monitoring for
metals at several sites has progressed under the direction of
the North Ohio Valley Air Authority (NOVAA), some data for
dioxin in soil have been collected by Federal EPA and
playground soil samples have been analyzed for lead with the
cooperation Of the Ohio EPA. Also, a community respiratory
study is planned by the Ohio Department of Health. However,
valid concerns within the community regarding the significance
of ongoing exposures, including any trends which may be
occuring, cannot be addressed satisfactorily with the limited
environmental and emission data collection programs currently
in place. Much more information relating to actual ongoing
exposures based on environmental monitoring and emission
monitoring is needed.
It has been the premise of EPA that constraints on the
operation of the WTI facility are adequate to limit
environmental impacts to acceptable levels. This premise is
not accepted by many individuals in the public sector since
they have not been a party in determining acceptable risk or
acceptable exposure levels. In the public's opinion the only
"acceptable level" is "no level of exposure" or at least no
increased level of risk. The community outrage factor in East
Liverpool has been justifiably high. The risk communication
aspects of this overall issue need to be re-examined and
re-done in light of currently accepted theories used in the
chemical industry. The public needs to be involved and
consulted to a far greater degree.
The EPA Phase II Risk Assessment for the WTI hazardous waste
incineration facility involves the evaluation of potential
human health risks associated with direct and indirect exposure
to stack emissions associated with routine operation of the
-------
- 3 -
facility. This purpose is incomplete. It does not address the
reasonable probability for fugitive emissions and accidental
occurances to also contribute to potential risk to health. The
public is very aware of these aspects and these exposure
sources need to be addressed in order for any study to be
considered comprehensive. The actual risk (or absence thereof)
needs to be addressed rather than only the probabality of
potential hazards if public acceptance is expected. This can
only be realistically achieved through the comprehensive
development of actual monitoring data which demonstrates that
no contaminant elevations are occuring within the community.
The only real data which EPA proposes to use in the Phase II
Risk Assessment involve trail burn emission data. This is not
sufficient to gain public confidence in East Liverpool. Much
more frequent monitoring for specific pollutants of concern,
ie. metals, organic products of incomplete combustion (PIC's),
PCDD's, PCDF's, volatile organics, etc. needs to be addressed
in stack and fugitive emissions. Also, community monitoring
programs for similar pollutants need to be expanded in scope
and frequency. This must include monitoring of air, soil and
crops, (ie. foodchain) at numerous sites within the community.
Relying solely on estimated concentrations of pollutants within
the community is not acceptable. Real community data must be
developed in an expanded manner and utilized in risk
assessment.
There can be considerable error in estimating community
exposures and health risk only from trial burn data as proposed
in the EPA Phase II Risk Assessment. No shift in public
confidence can be expected unless this aspect can be modified
to incorporate much more actual community environmental
monitoring data.
In summary, in order to address this complex issue effectively
we feel it is essential to incorporate public participation in
technical program development at the local level. Actual
environmental monitoring and expanded emission monitoring plus
community health surveillance needs to be emphasized rather
than estimations of theoretical risk based on model exposure
assumptions. Any assessment that involves only theoretical
assumptions on community exposures stands little chance of
gaining public support. The East Liverpool Board of Health
advocates the development of an expanded and accelerated
emission and environmental monitoring program rather than a
quantitative risk assessment based on little real data.
-------
DEC — S —
MON 16:3T CITY OF E (=1 S T LIVERPOOL.
- 4 -
The East Liverpool Board of Health encourages cooperation with
Federal EPA in further development of specific data collection
programs which can provide a realistic basis for protecting
public health, We would welcome further opportunity to discuss
this issue in more detail with the Agency.
Thank you.
Sin
Edward J.^sowinski, Ph.D., DA&T, CIH
EHMS Inc.
Advisor to the East Liverpool
Department of Health
Gary Ryan
Health Commissioner
City of Bast Liverpool
cc,
Mayor James Scafide
East Liverpool Board of Health
Dr. Dorothy Patton « EPA
Mr. Robert Sussman - EPA ,
Dr. Carol Braverman - EPA Region 5
-------
RERSOLUTION 92-3
A RESOLUTION, IN VIEW OF INCOMPLETE HEALTH AND ENVIRONMENTAL BASELINE
EVALUATIONS AND THE ABSENCE OF COMPREHENSIVE ONGOING EMISSION MONITORING
PLANS; DECLARING AS CONTRARY TO THE PUBLIC HEALTH AND COMFORT, THE IMMINENT
INCINERATION OF OR TRANSPORTATION FOR THE PURPOSE OF INCINERATION OF HAZARDOUS
CHEMICAL WASTE AT THE WTI FACILITY WITHIN THE EAST LIVERPOOL CITY HEALTH DISTRICT;
WHEREAS; The East Liverpool Board of Health is the body charged by the State
of Ohio with protecting the health of the citizens of the city, preventing
diseases and ordering the abatement of any unhealthful conditions, and,
WHEREAS; The Governor of the State of Ohio, under the authority of the Ohio
Department of Health and the Ohio EPA, has appropriated and delegated for use
a sum in excess of $600,000 to implement baseline studies to determine the
potential Health and Environmental impact of the WTI Hazardous Waste Incerator.
(These funds are being administered by the Ohio Department of Health and Ohio
EPA based on consultation with the East Liverpool Board of Health), and,
WHEREAS; The Health and Environmental baseline studies essential to aid in
determining the future potential impact of the WTI Incinerator are underway,
but incomplete and require approximately six to ten months for completion;
These studies include:
1. Lead and mercury screening in children including control populations,
2. Adult health screening including pulmonary status evaluations,:
3. Background levels of lead in soil,
4. Community background air sampling for metals adjusting for seasonal
variation and, \
5. Community background air sampling for organic chemicals, also adjusting
for seasonal variation; and,
WHEREAS; No specific plans are known to be in place for ongoing monitoring
after incinerator startup for specific organic products of incomplete combustion
(PIC's); (most recently, correspondence from EPA Region 5 dated October 30, 1992
and East Liverpool Board of Health letter'dated" November 12, 1992); and,
WHEREAS; No specific plans are known to be in place for ongoing monitoring of
fugitive emissions of organic chemicals after incinerator startup; and,
WHEREAS; No plans are known to be in place commensurate with chemical industry
standards of Responsible Community Awareness and Emergency Response (CAER)
associated with waste transport and/or incinerator accidental releases or
spills; and,
WHEREAS; The Liverpool Board of Health has requested (most recently in
November 12, 1992 letter to EPA Region 5),to no avail, the establishment of
a high level interagency technical steering committee or "ombudsman" for the
purpose of coordinating this complex issue among Federal, State, and
Local authorities, therefore;
-------
BE IT RESOLVED;
That the East Liverpool Board of Health opposes the imminent startup of
incineration of hazardous chemical waste and the transportation of hazardous
chemical waste within the Health District. The Board will pursue all legal
avenues open to it to prevent the imminent handling of Hazardous Chemical
Waste at the WTI Incinerator in view of the deficiencies outlined above. The
Board orders the Health Commissioner to notify concerned parties including
Federal, State and WTI officials of its opposition to imminent incinerator
operation and to endeavor to resolve the deficiencies outlined above. The
Board further orders that this action be published in accord with Ohio Revised
Code 3709.20.
Passed this .'> '•/ """ Day of November, 1992
*>y s~-('i,^
/Board of Health Members
East Liverpool, Ohio
n
(\ U I \ \CAJ(
L. R^ran \
Gary A. Ryan
Health Commissoner
City
'Presidentx6f
th
-------
-------
APPENDIX K
ADDENDUM TO PETITION REQUESTING REVOCATION OF TRIAL BURN PERMIT,
OPPOSING ISSUANCE OF PERMIT, REQUESTING PUBLIC HEARING, AND
REQUESTING THE PROMULGATION OF REGULATIONS TO PROTECT PUBLIC HEALTH
IN THE MATTER OF WTI
K-l
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY-
• • BEFORE THE ADMINISTRATOR
IN THE MATTER OF:
WASTE'TECHNOLOGIES INDUSTRIES/VON ROLL (OHIO)
HAZARDOUS WASTE INCINERATION FACILITY '
LOCATED IN EAST LIVERPOOL, OHIO .
ADDENDUM TO - . /
PETITION REQUESTING REVOCATION .OF TRIAL BURN PERMIT,.
OPPOSING ISSUANCE OF PERMIT, REQUESTING PUBLIC. HEARING,
AND REQUESTING THE PROMULGATION OF REGULATIONS '
TO PROTECT PUBLIC HEALTH . ' -
' f
The enclosed materials,- based, on Tiew information, serve as.;
an addendum to the administrative petition submitted to the ;*•
Administrator by plaintiffs on. February 5, 1993. The persons and
organizations submitting the petition and addendum are:
Greenpeace, Inc., Frank. C. Smith, M. Joy Allison, Becky Ammon,.
Beth Knapp, Becky Tobin, Robert, and Margery McKirinon, Virgil'•
Reynolds-, Lois and Keith Sevy, Alonzo Spencer, and Teresa .
Swearingen. • ' .'"••'"' '
The petition was brought under the authority of the Resource
Conservation arid Recovery Act,. 42 U'..S.C. sec. 6974, National .
Environmental Policy Act (NEPA)', . and the First: Amendment to the . .
U.S. Constitution. The relief requested included immediate
revocation of the trial burn permit and denial of any . '
authorization-to permanently operate- for. the WTI hazardous\waste
facility, as well as promulgation .of regulations to address ' . ;
numerous . issues in EPA's policy concerning hazardous waste .
incineration..' • ' , : ' .. - •• " ' ' ' .
• •'!>-: -. . ' ' ' ' ''•:..'•] ''''''' '''• '• '
The fpllowing new information, which has come to light in
connect ions "with proceedings in U,S. District: Court in Cleveland,
Ohio, adds powerful new support to the factual., and scientific '
basis* for .the. petition as- submitted. This new. information bears
directly upon section 6 of the. earlier petition, "Halogianated .
dioxins, furans, and related compounds." . '. ;
• NEW nJFORMATION .- ' ' • . ' . •' -:"' ' .; . :-'-. .";'.. '
1. EPA's-site-specific preliminary.risk assessment.'
-------
Materials leaked to Greenpeace indicated that EPA'S Office
of Research and Development had conducted a preliminary, site-
specific food-chain risk assessment for dioxin emissions from
'WTI. EPA's attorneys attempted to prevent that.assessment from.
coming to light, arguing that the materials were work-products
-protected by attorney-client privilege. On'February 9, however,
at the judge's request, EPA did bring the study before the court.
The study, "WTI Screening Analysis," is attached as exhibit C-l.
Petitioners believe, this new- study, by itself, must justify
EPA action to grant the relief requested. This site-specific
assessment demonstrates that, using EPA's best information,
modeling techniques, and risk characterization methods, food- '
chain exposures to .dioxin 'emissions from WTI. would pose
extraordinary risks, far above the levels considered acceptable
by EPA or.from a public health standpoint.... . .
The study is a preliminary risk assessment- of exposures' to
emissions .of dioxins and furans emitted from WTI under four
scenarios directly relevant to WTI: subsistence beef farmer,
partial-subsistence beef farmer/ nearby resident, and nearby t
schoolyard.. The emissions data were the same used in EPA Region..
V's Phase I Risk Assessment, based on average (not worst-case) .4
emissions measured, at VonRoll-'s similar -aic aerator in -i ,
Biebesheim, Germany. Dispersion models and. air concentrations
were also taken from the EPA Phase.I Assessment. The
incinerator/food-chain, exposure models were, applied" as developed.
by ORD in its 1992 document,- "Estimating Exposures to Dioxin-Like
Compounds." The exposure scenarios were chosen because of their.
site-specific relevance to .WTI. •" For instance, petitioner Joy "
Allison raises beef a few miles from WTI, her cattle consume all
locally-grown feed, and she consumes 100% locally-grown beef.
2. Results.of the. risk assessment. . .
a. The study found that food^chain exposures to dioxin and
furans for the subsistence beef-farmer would be 40,000 times
greater than by inhalation alone — the only, exposures considered
in EPA's Phase I risk .assessment. . . : '
b. Cancer "risks to the subsistence beef farmer were
estimated at 42 per-million due to the one^-year trial burn period
alone. '.-...... - . ... . .-
This, risk level from cine year of operation ia far above ' '.'-••
EPA's acceptable LIFETIME total risk standard for incinerators of
10 per million. . . ' . . . . .;
We are aware that EPA has not yet adopted an of f icial risk-
standard for foodchain risks due to short-term exposures to v
incinerator emissions. However, these risks exceed any risks
-------
historically considered acceptable by,the agency for incinerators
via any other"routes of exposure, or for food-chain dioxin risks
from other sources (i.e., paper mills).. Historically, these .
acceptable risks have ranged from'1 to 10 per million. We do not
believe there is any sound, basis on which EPA can justify
allowing greater public health risks through one route of
exposure than another.
c. Based on the methods and.results of this one-year
assessment, risks from permanent operation can be estimated. The
subsistence-beef farmer's risk of 42 per million can be
multiplied by 70 years and then by- 2 to account for the reaching
of steady-state soil concentrations. Cancer risks-from dioxin
exposure due to permanent operation at WTI would thus total 5,880
per million. Addition of a second kiln at'the site would bring
this risk to 11,760 per million. • :
•Such extraordinary risks, from either one kiln or two,
represent a major threat to public health and exceed any standard
EPA might possibly choose. Knowledge of such serious health
risks is in itself grounds for EPA to deny WTI authorization for
permanent operation and for trial burn.
3. Omissions in the risk assessment.
Although petitioners believe the methods used- in the ORD risk
assessment are essentially sound as far as they go, we note that
a number of important factors,were excluded. Inclusion of these
factors would raise the risk estimates. These omissions are
discussed in'the testimony of Dr. Barry Commoner (exhibit C-2)
and are partially, listed below.
a. Dioxin-like compounds. ,
The assessment did not consider emissions of other dioxin-
like compounds, such-as coplanar PCBs,. brominated dibxins and
furans, halogenated biphenylenes, etc.• These compounds, which
are known to be emitted from hazardous waste incinerators, could
add significantly add to the risks posed by the chlorinated
dioxins and furans.' ' .. ••' ,
b. Actual emissions. . . •- ,
The assessment did not consider increased emissions during
routine operation. The emissions estimates were taken from a
carefully-supervised trial burn at VonRoll^.s HIM incinerator in
Germany. Trial burn results, as discussed in /the earlier
petition, are known to underestimate, routine, emissions due to the
failure to burn real-world wastes, the maintenance of idealized,.
steady-state combustion conditions, and the elimination .of human
error and equipment'malfunction. . •" . •
In fact, recent data from the HIM incinerator — with- an ESP
installed — show that dioxin and furan emissions may in fact be
-------
significantly higher than EPA .assumed. Preliminary data from an
ambitious sampling program at that incinerator in late 1992 show
dioxin emissions up to 4_.6 times higher than_the 1989 values used
in 'EPA's' assessment., /{exhibits C-2, C-3, and~C-4) ~— "~
If this 4.6-fold increase in emissions were included, l-year
trial-burn risks would be calculated at 193 per million, .and
permanent risks at more than 25,OOO per million for each kiln.
Ci Other food-chain exposure routes.
The EPA analysis does not include exposures via consumption
of contaminated milk, eggs, fish and other meats. EPA's analysis
notes that milk may result in exposures comparable to those via
beef. For children, whose milk consumption- rates are often
relatively'high, .dioxin exposure from milk may be even higher
than via beef. ' • . ,
At least some residents of the.area near WTI eat both local
beef and local dairy. Foodchain risks may thus be at least
double those found in EPA's analysis.
d.
mother7
Exposures to developing children in utero and via
s milk.
According to Dr. Commoner's testimony, such- exposures to
children may be 10 times greater, than, those-, to the mother.
Consideration of this factor alone would bring the one-year risks
to 420 per million- and the permanent risks to more than- 50,000
per million for each kiln.
e. Failure to use plausible^ conservative values for each.
term in the assessment. .
When specific data are lacking, accepted risk assessment
methods require the use of the. most conservative reasonable
assumptions in order to protect the entire exposed population.
• . * -. * - * ' " * •
While conservative assumptions were used in. some cases in
the EPA assessment, they were not uniformly so. For instance, an
average beef consumption rate was used. . In fact,- a subsistence
beef farmer-.would be likely to consume greater-than-average
quantities^ of beef, and such is the case for petitioner .Joy
Allison. .Non-conservative values were also used for vegetable"
consumption and soi± ingestion, rates.
f. Failure to consider non-cancer effects. . . .
EPA's assessment is based on the cancer endpoint alone.
.However, recent scientific data and findings of the EPA dioxin
reassessment have noted that reproductive,, .developmental, and .
immunological effects. appear to be the most, sensitive endpoints
for dioxin exposure, [ex. C-5, C-2, A-13] '•'.'••• ' .• .
-------
• In the- range of exposures for both 1-year and 70-year opera
tion, these non-cancer effects, must be considered a serious risk,
especially for sensitive persons in the population, such, as
developing children.
Recent findings indicate that there is no threshold dioxin
dose below which such effects can. be assumed not to occur, and
that current background exposures to .the general. population are
in the range at- which such effects are known/to occur, [ex. C-5,
C-2, A-13] on this basis, it must.be expected, that additional
dioxin exposures caused by WTI will cause reproductive, develop
mental, and/ or immunological effects.
Petitioners believe this information, .in^ itself, is adequate
to show that WTI will pose unacceptable and illegal health risks
to the public. Based on EPA's mandatory- .duty- to protect | pub lie
Health and the other legal grounds discussed in , the original :
petition, we request EPA to grant- the relief requested.
EXHIBITS . • . . . • ,
C-l. U.S. EPA Office of Research' and Development. WT±'Screening
Level Analysis. February 5,; 1993- -
C-2. Testimony of.Barry Commoner, in Greenpeace et al vs. WTI
et. al. U.S. District Court, Cleveland, Ohio, February.16, 1993
s^ . • •' ••'•-.: ",-. • ' ^'
C-3. O. Wasserman (Christian-Albrechts Universitat).
nary results from 1992 .sampling program f or - dioxin and
«* . ... * » . <• .*_j_ *— - %* . -^ - "
emissions
1993.
B
at'HIM incinerator. Letter to B. Commoner
Commoner and. T. Webster. Analysis of preliminary re.
suits froa-:i992. sampling program at.HIM incinerator. 'February
12, 1993
v-
C-5. Bretthauer, E. (Assistant Administrator, EPA Of f ice of
Research and Development) . Update on Dioxin Reassessment
. Activities. Memorandum to the. Administrator. October 9, 1992.
-------
2
3.
4
5
6
7
8
9 .
10
11
12
13
14
i ^
16
17
18 .
19
20
21
22
23
24 '
25
Commoner - Direct/Condit
Monday, February 15, '1993
• ; :' BARRY COMMONER
being first duly sworn, was examined and testified as
follows:•
657
DIRECT EXAMINATION
BY.MR. CONDTT: " '
Q. • Good, morning, Doctor, . . . .
l Good morning. , ~ . •
Q.'' Could you please state your full name for the record
and tell us where you -work, please?
A". 3arry Commoner, director for the Center or diology '
of National Systems at Queens College City'University of
New York. ' ' • . '''-.-•_
C- Dr. Commoner. I'm going to show you a document that
I'm going to mark as Plaintiffs' Exhibit 30'. It's a copy
~ *" " * "
of your CY. . •'."""'
• . . i . .
Dr. Comaoner, do you have a copy of that before you?
A. : I don'tJ . ' . . . :
Q. I*» sorry, correction, we're going to make this
Exhibit 31, r an told. You do have a copy? '
A. I don't. ' ' ; . . - •'.•'•.'. -..- • " .
Q. You do not. Sorry. •
Q. Dr. Coamoner, if you would, could you please
suBunarize your educational background for the Court? ... .
Marian E. Spehar -Official Court Reporter--- 566-0655
-------
Commoner - Direct
686
All right.
if I may lay some foundation?
Go ahead-
1 THE COURT: ' ' Well, I with, have to check the
2 transcript from last .Wednesday.. My recollection is that she
3 did use some numbers that she got with, respect to the German
4 facility, but perhaps you. can find it.
5 " MR. HARRISON:. 'We-will try to find that for,
6 your Honor. .
7 ^ THE COURT:• • '
8 MR. CQNDIT;'
9 . THE COURT:
10 BY MR. CONDIT: • ' • . ,
11 Q. Dr.' Commoner, can you tell us who Dr. Wassermann is,
12 first of "all?
13 A. Yes, sir, Dr. Wassermann is a well-known expert in the
14 area of incinerator emissions. He is a colleague of.mine in
IS the sense that we have met on a number of occasions, and we
16 customarily exchange data. • :
17 I sent him copies of our reports, he sends me
/ *
18 copies of his reports. . ' ' .
19 He knew of the hearing here/ and as it happened,
20 ha had been asked by the. Regional Government of Hessen to do
21 a series of emission tests at the HIM ihcineratoir. And when
22 he got his -first result he thought it would be usieful for me
23 to know about them, and sent mo a latter with tho data
24 reported on the back.' .. ''.'.-.'•'.
25 . He is, to my knowledge, .someone who is very activ*
TERRY D. GIMMELLIE, RPR-CM!
-------
Commoner - Direct
687
1 in this field, and I greatly respect any information that he
2 provides- for me. This was clearly an analysis done under
3 official circumstances, interests on the HIM incinerator,
4 and I think very interesting and important information.
5 Q. And, Dr. Commoner, would-this be the type of information
6 with respect to Dr. Wassermann or anyone else that you
7 respect in the various fields that you work in; would this
8 be the type of information that you might rely upon in
9 forming an opinion about certain issues with respect to a
10 facility like this or other environmental issues?
11 A. Yes.. This goes back to what I was saying before about
12 the way in which you determine what emissions you expect
13 from an incinerator such as.WTI.
14 What Wassermann has given us is additional data on
15 the behavior of the HIM incinerator that are added to the
16 data reported in the Region 5 assessment. The more data you
17 have, the more accurate is your understanding of the -
r- ' ' * f • ' •'•.-
18 behavior of that incinerator.
19 Q. ..All right.
20 A. So, that this is important, as far as I'm concerned,
21 important information" tha€HLmproves our understanding of the
22 emissions to b« expected, from the WTI incinerator. .
.23 Q. Doctor Commoner/ did you and your staff perform analysis
24 of that data upon receiving it?-. ; . - -•_•'..
25 A. Yes.: I discussed it with ay colleague, Tom Webstar, and
TERRY D. GIMMELLIE, RPR-CM
-------
Commoner - Direct
638
1 we worked out an approach for 'analyzing it which took into '
2 'account the fact that.what Dr. Wassermann has analyzed thus
3 far are the total groups, rather than the individual
4 isomers.
5 Q. Is that a dioxin?
6 A. Yes, of dioxin and furans.
7 Q. Can I stop you for a second there?
8 I would like to mark Plaintiffs' .Exhibit next
9 exhibit. Dr. Commoner, do you have this exhibit in front of
Id you, the analysis that you performed?
11 A. Yes. "'
12 Q. £xnibit 33 is Dr. Commoner's analysis. Please go on
13 with your description. . .
14 THE COURT: " I'm -sorry, Exhibit 33 'is what?
IS MR. .CONDIT: Is Dr. Commoner's analysis of
16 the HIM data received from Dr. Wasserman. Is that correct,
17 - Dr. Commoner? .... ' .
18 THE WITNESS: ' . That's right. • There are three
19 tables on the back side of that exhibit.
20 " The first one is the set of data that Region 5
21 used from the «arii«r-tests of th« HIM incinerator.
22 Q. Dr. Commoner, if I can stop you for a second. Can you
23' lay out for us, b«for« getting to th« details/what this
24 table that, you have- created, that.your staff and you
25 'cr*atad, was intended to represent? ' .
TERRY D. GIMMELLIE, RPR-CM
-------
Commoner - Direct
639.
, 1 A. It represent a comparison between the. earlier test
2 result reports in the Region 5 risk analysis, with two
3 samples reported by Dr. Wasserman. One taken with a sample
4 means that you pass the gas through an absorbing agent, and
5 then analyze what's been picked up.
6 The customary agent, which I think must have been
7 used in the Region 5 data is a resin called XAD-2. And what
8 such analysis that Wassermann did was a XAD-2 analysis.
. 9 He did a second one with a material called nonane.
10 He expects to do seven.more, as a matter of fact. But he
11 has done these two.
. 12 '.. So, what we did was to compare the total toxicity
13 equivalence yielded.'by the original test as reported in the
14 Region 5 risk assessment, the analysis that Dr. Wasseraann
15 did just recently based on XAD-2, and the analysis based on
16 nonane. . . "
17 Wfca't. tne results show is that in the XAD-2
IS analysis, although, fewer dioxins were picked up in the new
• $'~ '" ' " '
19 analysis/ than in the old one, that was overwhelmed by a
20 sharp increase in the. furan analysis.
21 The net result, is that in comparison, just looking
22 at the mean toxic equivalent, the old data are .91 nanbgrams
23 per meter cubed. • :
24 .. .The XAD data, the new data are 4.24.
•'."•• . ' • '"".-.
25 " Q. In summary — ;
-------
Commoner - Direct
690
1 A.. In other words, about Tour times higher.
2 It's also -interesting to note that the nonane
3 test, which was done subsequent to the XAD-2 test, is less
4 ' than the old -test..
5 • , Now, what that reflects is, again, the whole
6 business of variability. There are two things that you have
7 to consider in comparing XAD-2 and nonane.
8 One is, they are different in their ability to
9 pick up compound. But, they are also subsequent tests.
10 They are not done at the same time.
11 So, that part of that variability, may well
12 reflect' a time difference in" rh'e behavior of the
13 incinerator. .
14 At. any rate,' what, this shows .is that if you repeat
15 the kind of analysis that 'the risk five, or Region 5 risk
16 assessment included, which I believe was XAD-2, the new
17 results are four times- higher, approximately, than the old
18 results, and that is simply an .indication of the
19 variability. .'.•'. .. -
20 •"• • I'm sure that if more results were gotten on the
21 HIM incinerator, you would see more variability, and if you
22 wanted to waited long enough you could take an average and
23 that might give you the good prediction as to what WTI would
24 '• dp. . . . •
25 THE COURT: ..You are still suggesting this
TERRY D. GIMMELLIE, RPR-CM
-------
Commoner - Direct
691
1
2
3
4
5
6
7
B
9
10
11
^.^
13
is
17
13
19
20
21
22
23
24
25
is the best way to — • • •
THE WITNESS: You're dealing with inherently
very variable phenomenon in both respects and this is well
established, the communication in the combustion efficiency
- from minute to minute.
Secondly,- the variation in the input of the
hazardous waste.
In fact, my understanding is the fuel that goes
into a hazardous waste incinerator is never analyzed to any
great detail, and you are really very often are .at a loss as
r
to what's in it. It would be enormously expensive to do a
detailed analysis which means that the output is going to
vary a good deal as well.
•i-Mfe; COURT: '. • I. know there is an assumption
we don't have a foundation for.
MS. GOLDMAN:. . He has not been qualified as an
expert on hazardous waste incineration. .
THE COURT: . Right.
BY MR. CONDITr - .
Q. Dr. Commoner, what adjustments, if any, would you make
on the risk analysis numbers that we were_ discuss ing a .
little while ago if you considered .this emissions data from
Dr. Was
A. Well, clearly, if Dr. Farland were to use the XAD-2,
knew XAD-2 analyses, bis values would come out four times
TERRY D. GIMMELLIE, RPR-CM
-------
Commoner — Direct
692
'• 1 -higher than they sure.
2 In other words, instead of 4.2 times ten to the
3 minus five risk, he would get a little over — -around 17.-
4 If he used the nonane, it would be just a little
5 * less than his 4.2 figure.
6 Q. Thank you. Turning the direction of your examination to
7 a little different, area on public health issues, can you -
•
8 tail us, Doctor, have you had an opportunity to consider
9 what the lifetime risk from the operation of the WTI
10 facility might been given the EPA risk assessment figures?
11 A. Yes> I have done that. . ;
12 -Q. All right. Can you tell us —•'
13 ' MR. WISEMAN: This is beyond the scope of
14 .anything that's gone forward in this hearing. We have been
IS testifying here as to the test burn and posr trial burn
16 period. . .."."- . • .
17 .HR. CONDIT: Your Honor, Dr. Green testified
18 specifically that the lifetime risk was not significant,
19 analyzing it, on, I. believe, a 70 basis. ± believe that was
i™"1**. , , ' ..*!
20 the same assessments made by Dr. Far land as well, in his
21 testimony. 'Certainly, Dr. Green has raised it, and I
22 believe it's appropriate area of rebuttal given her raising
23 " that issue-. The lifetime risk also speaks directly to our
24 .' ability to prevail on the merits as we go through the course
25 of this proceeding'. .
TERRY D. GIMMELLIE, RPR-CM
-------
EXHIBIT C-3
"T
-------
.
KLINIK'UM DER 2HKTITIAN-A L3 X
A5TEILUNG
r.' Pnrf. r.
Sr.
£335,
t>*i»«;4cjnta.''Jfta«.i« 3W. A rw ic* i
*
C Or'aTc; .oner
. I
Bat'ardou* Vlt8tt»
''1 "
• J
rrl«nt ai
Cr.
Bu*'to th« in"fi*i««St:,It» of the atAta fovemment:ef*Hd**«& -
oos«ra:tisn w.Uh tfc - w* .got «csa3* ;tc 't»* st^ck. ef
sitbaBh«ia csctntlyy wh«;r»"» J^eciJl laboratory dre
aiaang tSi« aiastt flopniaticat« h*«4
7h* •ch«ai«al analysis of "*M« »aapl.«a' ii In
* aa a iirsCr- pralimiaar^ result 1 an *ncio»ihg' rorj«tjmj)«-
th* dftti' foe. XAC-2 and nonah«, r«ip,, to fhcw b«th th«
fCiarption Att'<3/o.* ioI\iMlity '• »fli«ttncy. of ab
diff«r«wi«a cf.-tha -SCXO/r ccnc«ntr*t^o^f «mit-
in, th* J3MJ '
dia-
thi» ictu«i' d»t* «r« h*ipful also in
In th* U.3.' ^n..ozd«r to
and r**lity.
I3r; ot-nar tfasaermasn
-------
-2 / ?rofc« 2
TottMCW
. Total TCCO .
Total Ps£0i:
Tc^ai ?*COO
Tct«l .KxOJF
Total H*CDO
Total HjpCOF
Tot*! HpCDO
OCOF
OCOO '-..
. - 14.4
°4^
63.2
3.3 " .
3S.S
3'.S
55. J!
19.3
5.Z!
• 15.4-
4rl.
1.3
3,3"
0.3
u.a"
2,3
5.2
1.3.
1.5
3.8 :
. . ' •
i
; ..
• •
.' "
• .
-------
0
»*;•
»-> • • T\_;*I I -••«».'»--••».••-<•-•••. • «- «' • «•- ' * • •• * -»ft « ••^•y- •vjx-'-w.
•v* g^C 100 .CHO <=X>0«ai f>Ti 83 On t- •n^\^J,<*°r^i|3-^-s;l9i^'i;.-?^
-------
February 23, 1993
Mr. D ick Morgenstern
Deputy Administrator- . . • ,
U.S. EPA • i \<-r-^ fL^t * , l s*>t- ^WUu
4oi:M st.;.sw ;•• fi^jK^i^»eir*V9\r
Washington, DC . . • . :
RE: Petition requesting revocation of trial burn permit at WTI
Dear Mr. Morgenstern, . _ . •
Pursuant to Resource, Conservation and Recovery Act, 42 U.S.C, sec.
6974, National Environmental Policy Act (NEPA) , and the First
Amendment to the. UiS. Constitution, the following organizations,
submitted an addendum to the petition which was. filed with the
Office of .the Administrator on February 5, 1993. - Events oh that
day require your attention to this matter, since the Administrator,
Carol Browner, has recused herself ' of any matters involving the
Waste Technologies Industries (WTI) hazardous waste incinerator in
East Liverpool, Ohio. The -addendum was delivered to you yesterday
'afternoon, February 22, 1993. . .- .• ' . ;
For the. reasons fully detailed- in the original petition, arid based
on new information discovered during the . Preliminary Injunction
proceedings in Cleveland, Ohio, .the petitioners request the same
relief detailed in the February 5, 1993, document.. • " ;;
Respectfully submitted, ";
.
Richard Condit, Esq.. ,': . " . '
Mick Harrison,. Esq. . .; - • . .
Government Accountability Project . "
810 First St., HE #630 :.. . • . • .
Washington, DC 20002 - • • -
202-408--0034 . '. •• -: '-• - • .
.for petitioners . , . . " . . ,
iSreenpeace, _Inc. ,, Frank C, Sipith^ M. Joy Allison,. Becky^ Ammon,_ Betix-
Knapp, Becky Tobin, Robert and Margery McKinnon, Virgil Reynoids,
Lois and Keith. Sevy, Alonzo Spencer, and Teresa. Swearingen. .
-------
EXHIBIT C-4
Atel,
-------
•CENTER FOR THE BIOLOGY OF NATURAL SYSTEMS
CBNS : • . "
CUEEMS COU-ECS; CUNY" -
FLUSHING, NEW Y^RK U36?
. 0? TCDD-ZQUXVALSjjTS OF 1TBW UTI EGSSI6N3 •
*•.'•- •• . . • •- .
4
l.-In a latta$r, Dr, .ctsar Waaaarmann provi^td'
> ' ' —
iat* rroa th«| Bi«35«sh*ia ha5ardou» wasta incin«ratpr Sasplea
froa tvo oi aisa adficr&unta aethoda have iscen" supplied.
2 ,
new
sr specific, fcut givan oV tha haais of
r • • ; •
vith tha saaia r
- ? •:''. •
Isoaei- da-ia ars nseded Cor ^CDD-aqoivaiehtV (5SC} "
* • * . ^*c
•> f ,-. " • *
j«5^r 2i943ol'55"i-3, cpro'up free t£a aarlidr
data (US2FA ^gr'^.lainagv Rigfc^aaagmmsfc ^
yXgoa"'-Tg'3 cc5 sfeac:.^ "S^jgnicns ^rcm tha ffTT Xncines'atar1' "fahTa a ^i
^ ^i^^^»^^^^i«^*^h^i •i"*^^^^w«*^«^ta^»*»BB»«,«ai«»^«B^ii^^fcpi^L«!S,SfcJ^^ *• ^*A^ J^ ^* ^\ + *J 1
Tr.a isomer frac?i.=«s from tie earlier data. ar4 multiplied
iha-homolegaa^data of tha new «7ai4siona result*. ' r
1 * i
. .5 i
' '' }
iaccar-fipocific «ai«aion* data ara"
tha appropriata TCDO-«qiii7alancy fAotcra (T2J*) and siinsad
• hack coneantration of TCDD-aquivaianta in tha
• '•***. -i •
i-t oi«i>«3h«ia," t^sti "sampled using XAD-2 adsorbant "ias 4.2. '-.I.
, ovar ^our tiaas tUa avaraga used in Pralinknary Risk •
J - - - • • ' . . , *..-•'
.Assessaant. fa ,tatk concentration oi TCDD-«qulvai«nta' maaaurad
using nonana wa.4,a.«4 ag/ltoa, about 70* of tb* 4vataga'udad In.
-------
-------
APPENDIX L
COMMENTS ON "WTI PHASE II RISK ASSESSMENT PROJECT PLAN,"
EPA ID# OHD980613541
L-l
-------
COMMENTS ON "WTI PHASE II RISK ASSESSMENT PROJECT PLAN,"
EPA ID#OHD980613541
Prepared by
Edmund A. C. Crouch, Ph.D.
Laura C. Green, Ph.D., D.A.B.T.
Katherine K. Perkins
and
Stephen G. Zemba, Ph.D.
Sponsored by
Waste Technologies Industries (WTI)
Decembers, 1993
©1993 Cambridge Environmental Inc.
All rights reserved
Cambridge Environmental Inc.
58 Charles Street Cambridge, Massachusetts 02141
6t7-225-0810 617-225-0813 FAX
-------
Contents
1. Introduction and General Remarks 1
2. Lack of sufficient detail for review, and over-reliance on anonymous, unreferenced,
and unreviewed sources ' 2
3. A source of more specific methods, data, and an outline of a Monte Carlo
approach to risk assessment for WTI 3
4. Air modeling 4
5. Cancer potency value for dioxin and its equivalents 7
6. Validation 8
7. Emission rate estimates 9
8. Mercury -\-\
9. Retention of washout by plants 12
10. Food consumption rates, Table A-9 13
11. Lead 13
12. Photodegradation of vapors 13
13. Vapor uptake by plants .- 15
14. Fugitive emissions and related issues 23
Cambridge Environmental Inc.
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
1.
Introduction and General Remarks
This document presents some technical comments on the "WTI PHASE II RISK ASSESSMENT
PROJECT PLAN," EPA ID# OHD980613541. Due to the limited time during which the Project
Plan has been available for review, our analysis has not been as detailed as we would have
wished. The following sections present our preliminary impressions, and suggest some
alternative approaches.
In the comments presented below, we wish to emphasize the following major points.
• The air dispersion modeling proposed in the Project Plan is quite inadequate. Since all of
the impacts, direct and indirect, to be estimated in the risk assessment depend upon air
dispersion modeling, this inadequacy cannot be overlooked.
• The risk assessment desired is not a screening analysis; it is instead something meant to
predict "actual risks," and something upon which specific permit decisions purportedly will
be based. As such, it cannot rest on simple, deterministic, "conservative" screening
models. It must instead address squarely and correctly the tremendous uncertainties and
sizable variabilities involved in all of the models. These uncertainties and variabilities
plague the air models, the exposure models, and the toxicity models alike. All can be
addressed best by use of probabilistic techniques, the most appropriate and well-studied
of which are the Monte Carlo techniques.
• The current U.S. EPA cancer potency value for dioxin and its equivalents is based on out-
moded readings of the slides from the critical bioassay, makes inappropriate use of
certain data, and ignores the large uncertainty in extrapolation from mice to men. Given
our ignorance as to the true potency in humans for low-level exposures to dioxin, careful
risk assessors must utilize a range of potency estimates. Ranges should be used also for
estimates of toxicity and/or carcinogenicity for all other major chemicals of concern.
• None of the models upon which the risk assessment is to be based has been validated.
• "High-end" estimates of emission rates should comport in at least some rough sense with
reality. At least three bounds on reality, suggested herein, should be used.
Overall, we believe that the Project Plan will not accomplish its major goal, which is (page 14 of
the Project Plan) "to provide a more complete understanding of the actual risks to the population
surrounding the WTI facility." It will not even supply a correct measure of the (page 20) "central
tendency" estimate of individual risk. A major reason for this failure is that actual risks (to health)
depend on two main factors — exposure (to a chemical or mixture) and toxic potency (of that
Cambridge Environmental me
1
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
chemical or mixture) — but the Project Plan envisions no attempt to generate or use any "central
tendencies" for estimates of toxic potency. Instead, the proposal is to combine (more or less)
best estimates of exposure with upper-bound estimates of toxic potency. Clearly, such
combination gives nothing like the "central tendency" of risk, it also misleads the risk manager or
other decision-maker, who might be fooled into thinking that what will be advertised as an
"actual" or "average" risk is just that — an unbiased, more or less actuarial prediction, little
different from predicted traffic deaths, for example, given so many miles of motor vehicle travel.
Fortunately, there are techniques that would allow the risk assessor to present to the risk
manager genuine central tendency estimates (as well as properly constructed upper-percentile
estimates, ranges, and so on). These better methods are mentioned in part below. They are
also elaborated upon in our Protocol for a multi-pathway risk assessment for the WTI facility in
East Liverpool, Ohio (Cambridge Environmental Inc., September, 1993 and December, 1993),
about which more below.
Our more detailed observations and suggestions are as follows.
2.
Lack of sufficient detail for review, and over-reliance on anonymous, unreferenced,
and unreviewed sources
Many parts of the draft Project Plan cannot be adequately reviewed, since they give little or no
indication of the methods that are to be used. An example is the section on "Development of
Chemical-Specific Emission Rates" on pages 34ff, in which there are such statements as (page
35):
"data ... will be reviewed, and if found applicable, will be used..."
"the data ... only if it is determined that the data can be extrapolated.."
There is also no indication of the methods to be used to determine applicability, or to extrapolate
the data. One cannot review methodology that is not given.
As another example, while there is a cursory discussion of air dispersion modeling methods
(about which more below), there is no mention of some critical input data that are to be used.
For example, the COMPDEP model requires wet scavenging coefficients for the various particle
size ranges, but none are suggested.
As another example, it is left undefined what water bodies are to be modeled to take account of
fishing. Discussion of reports on the Ohio River early in the draft Project Plan suggests that the
Ohio River may be included. However, this is never made explicit, and there is no indication of
Cambridge Environmental Inc.
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
how the watershed areas are to be defined or used, either for the Ohio River or other water
bodies. It is also not clear whether the same models are to be applied to all water bodies — it is
likely that different models would be required for flowing rivers than for stagnant ponds, for
example.
Similar difficulties face the reviewer throughout the document. In too many places, the selections
of data, methods, and values are left undefined; the reader is given no list of the possible data,
methods, or values that might be selected, or even sources for such data, methods, or values.
Moreover, in many tables in the appendix, values or whole models are attributed to anonymous,
unreferenced, undocumented, and apparently unreviewed "recommendations made by the
Exposure Assessment Working Group". Without some documentation of what these
recommendations are based upon, it is impossible to adequately review them.
We appreciate the difficulties that the EPA contractor faces in this situation — without their
having tried various options, they find it difficult to make concrete suggestions. However, the
EPA and its contractor cannot benefit from review if there is so little available for review. The
data, methods and values finally selected should indeed be reviewed, but it is not clear that such
review will then occur.
3. A source of more specific methods, data, and an outline of a Monte Carlo approach
to risk assessment for WTI
Given how many details are missing from the Project Plan, it appears that what is being
requested in many cases is not review, but the provision of ideas for subsequent development
and review. The accompanying document that we present (Revised Protocol for a multi-pathway
risk assessment for the WTI facility in East Liverpool, Ohio) provides many such ideas, giving
specific methods and data (or data sources) in many instances where the EPA draft Project Plan
is silent.
Our Protocol was originally circulated to risk assessors at U.S. EPA, its contractor, and many
other risk assessors on September 24,1993. It does not appear to have been reviewed or relied
on in any way by U.S. EPA or its contractor. Our Revised Protocol, dated December 2,1993, is
largely similar, differing only in that we have made the corrections and adopted some of the
suggestions offered by those peers who reviewed our September Protocol. Other changes are
contemplated, but have not been included at this stage for lack of time. Now that EPA's draft
Project Plan has been issued, we would hope that risk assessors could feel free to utilize our
document.
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 61.7-225"-0813. FAX
-------
4. Air modeling
The proposed air dispersion modeling is inadequate. This inadequacy stems from at least four
factors:
The proposed COMPDEP model is untested and inapplicable to most of the study
area;
The setting of the facility is too complex for straight-line gaussian plume models
such as COMPDEP;
Site-specific meteorological conditions are not properly considered; and
There are no stated plans to assess the considerable uncertainty that will be
inherent in air dispersion modeling.
Air dispersion modeling has been a part of EPA's regulatory framework for many years, and
many of the models have become commonplace.1 Almost ten years ago EPA (1984) noted the
shortcomings of models available to predict pollutant concentrations in complex terrain (which
terrain dominates the area surrounding WTI). Modeling difficulties in modeling plume dispersion
in valleys were discussed therein as a research objective.
Ten years later, no models have resulted that are appropriate for valley settings. Instead, the
result of many years of research on complex terrain modeling has been the development of the
CTDMPLUS model, which is now the only preferred model for complex terrain application (U.S.
EPA, 1993). As is true for all of EPA's complex terrain models, however, CTDMPLUS is only
valid for assessing plume impaction on the windward sides of hills, and is specifically not
recommended for application on the lee side of hills or on a series of hills (let alone a valley
wall). CTDMPLUS is adequate for most regulatory applications, however, since (1) maximum
airborne pollutant concentrations are usually of interest and (2) such maxima generally occur
upon plume impaction.
The application of a straight-line gaussian plume model such as COMPDEP to the WTI site is
inappropriate for risk assessment purposes in which the major route of exposure is expected to
be through indirect routes, especially the farm-related exposure routes that are typically the most
important contributors to dioxin and furan exposures. The stack of the WTI facility is located well
below the top of Ohio River Valley. From the land use maps contained in the draft Project Plan it
is evident that, (1) more than 98% of the study area, and (2) all farmlands lie outside of the river
11t is a popular fallacy that models and modeling assumptions are validated by repetitive use.
More frequently, models and assumptions are used over and over again because (1) regardless
of their adequacy, they are the only available tool and (2) references (sometimes with erroneous
information) perpetuate their use.
Cambridge Environmental Inc.
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
valley. At these locations, EPA's complex terrain models are inapplicable. Since all risk
estimates depend upon the pollutant concentrations estimated by air dispersion modeling, the
entire n'sk assessment is compromised by the lack of an appropriate air dispersion model.
COMPDEP is not recommended by the EPA as either a preferred or alternative model in their
detailed guidance on air dispersion modeling (U.S. EPA, 1993). The COMPDEP model was
released for a time in 1990, removed from the public domain abruptly, and modified and released
again in September 1993. Given the short time the model has been available, there has been
limited time for testing the model, and thus few people are experienced with its use.
The stated reason for using COMPDEP in the WTI draft Project Plan is its ability to generate
estimates of wet deposition. The capability for generating estimates is, of course, very different
from the capability to do so accurately or even adequately. Wet deposition is poorly understood;
the extreme uncertainties of the governing physical processes (and their subsequent effects in
the environment) are part of the reason that so few models exist. Moreover, the validation
criteria for new models as described in regulatory guidance (U.S. EPA, 1993) require
demonstration of the validity of the model, which typically includes a comparison with field data.
With respect to wet deposition, this aspect has not been demonstrated for COMPDEP or any
other model.
In addition, meteorological conditions in the WTI study area are far too complex for a straight-line
gaussian plume model such as COMPDEP. The narrow Ohio River Valley near East Liverpool is
extremely susceptible to valley flow regimes, as detailed.in our Protocol. In many cases, winds
measured at the East End Elementary School (Site 1 in the on-site WTI meteorological
monitoring program) blowing southward toward West Virginia represent gravity drainage flows.
Under such conditions, winds simultaneously measured at the Beaver Valley Power station,
located eight miles upriver from the WTI facility but on the southern shore of the river, blow
towards the north (opposite from the WTI facility), reflecting the gravity drainage from the
opposite valley wall. These down-slope winds will turn and channel along the direction of the
river, as is typical for valley flows, and as has been empirically observed near the Beaver Valley
site.
Under such conditions, atmospheric conditions within the valley can be extremely stable, and
may reflect inversions that serve to (1) decouple the air flows in the valley from the regional flow
pattern and (2) trap emissions within the valley. The application of the COMPDEP model,
however, will assume that the plume moves in a straight line from the WTI stack, impacts the
valley wall in West Virginia (resulting in an extremely but artefactually high estimated air pollutant
concentration because of the stable atmosphere), then continues to move up and out of the
valley on the same straight line trajectory. In reality, the plume may not reach the opposite wall
of the valley at all, nor may much of it leave the valley.
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
The draft Project Plan also does not offer much direction in the treatment of meteorological data.
The stated intention is to use data from the Greater Pittsburgh International Airport to predict
atmospheric stability and mixing heights; but this will incorrectly assess plume dispersion
characteristics in the valley and fail to account for inversions. We suggest at least two
improvements. First, consider merging the on-site meteorological data with certain parameters
measured at the Beaver Valley Power Station (such as vertical temperature gradient) to better
predict stability conditions in the valley. EPA's own guidance (U.S. EPA, 1987) for using on-site
meteorological data should be used, rather than the ad hoc approach taken in the draft Project
Plan. Second, under stable in-valley conditions, decouple in-valley and out-of-valley modeling.
One method to implement this suggestion is contained in our Protocol.
In any case, air dispersion estimates can be expected to be highly uncertain. In this application,
air dispersion modeling will be a large source of uncertainty to the risk assessment. This factor is
belittled by the oft-quoted reference (not appropriate here) to "a factor of two" that is contained in
the draft Project Plan (page 105). EPA's relevant regulatory guidance (U.S. EPA, 1993)
emphasizes the need to evaluate quantitatively uncertainties in air dispersion modeling, and this
recommendation is most relevant to the assessment of the WTI facility.
Finally, page 65 of the Draft Project Plan states EPA's intention to conduct three modeling
simulations to account for differences in the deposition of particle-bound pollutants (mass and
surface weighted) and vapors. While we agree that single model runs will be sufficient for
assessing the depositions of particle-bound pollutants, multiple runs will be necessary for vapors
since both the wet scavenging coefficients and dry deposition velocities of vapors differ between
pollutants (in some cases, by orders of magnitude). Thus, it may be necessary to perform a
separate model run for each pollutant considered in the vapor phase. Depending upon the
number of vapors considered, the scope of this problem may be prohibitive. Moreover, the Draft
Project Plan does not specify where scavenging coefficients and dry deposition velocities will be
obtained.
U.S. EPA (1984). Scientific assessment document on status of complex terrain dispersion
models for EPA regulatory applications. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Environmental Sciences Research Laboratory. EPA-600/3-84-103.
U.S. EPA (1987). On-site Meteorological Program Guidance for Regulatory Modeling
Applications. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards. EPA-450/4-87-013.
U.S. EPA (1993). Requirements for preparation, adoption, and submittal of implementation plans;
final rule. 40 CFR Part 51 et a/., Federal Register, July 20,1993.
Cambridge Environmental Inc
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
5. Cancer potency value for dioxin and its equivalents
There is substantial uncertainty about the cancer potency value for 2,3,7,8-TCDD, upon which all
the estimates of carcinogenic risk from all of the PCDDs/PCDFs are based. This uncertainty is
not reflected in the EPA's draft Project Plan.
The current U.S. EPA cancer potency value of 156,000 kg-d/mg for 2,3,7,8-TCDD is predicated
on outmoded (since 1986) readings of the slides from the critical bioassay (see e.g. Keenan et
a/., 1991). With the updated readings, and EPA's methodology, the cancer potency value would
be approximately 52,000 kg-d/mg. Furthermore, examination of EPA's original derivation of the
value of 156,000 kg-d/mg (EPA, 1985) shows that an error was made that, when corrected
according to standard EPA protocols (Anderson et a/., 1983, and the protocol desired in EPA,
1985), would give a higher value (approximately 263,000 kg-d/mg). Some of the uncertainty due
to various aspects of the EPA protocols was recognized (EPA, 1985, page 11-117, for example),
but these uncertainties have not apparently been recognized anywhere in the draft Project Plan.
In the original derivation, EPA (EPA, 1985, page 11-117) took a geometric average between two
values obtained in Tables 8-8A and 8-9A of EPA, 1985. Unfortunately, in both cases the
goodness-of-fit %2 value was incorrectly calculated, and the resulting p-value for goodness of fit is
incorrect in the two tables. The correct calculation shows that the results in Table 8-8A are
acceptable (p = 0.091), with a q* of 151,000 kg-d/mg, but for Table 8-9A the correct %2 value
corresponds to p = 0.0040, which is unacceptably bad according to the criterion stated in that
same table. The highest dose should thus be dropped, leading to an estimate for q* of 458,000
kg-d/mg using the three lowest doses (including zero). Taking the geometric average of these
two values (which was the procedure adopted by EPA) leads to the estimate of 263,000 kg-d/mg
just given.
More important for any hope of an accurate risk estimate, these values ignore the empirical
evidence indicating that unbiased interspecies extrapolation requires using a bodyweight-based
interspecies factor. The generally accepted understanding that unmetabolized TCDD is the
active agent also argues strongly for scaling on the basis of body weight, not surface area (a
surrogate measure of metabolic rate). The value advanced in the Project Plan also ignores the
huge (and empirically determinable) uncertainties involved. Appendix B of our Protocol
summarizes an approach that can correct these problems.
El. Anderson and the Carcinogen Assessment Group of the U.S. Environmental Protection
Agency, 1983. Dose-Response models in use for estimating cancer risk. Risk Analysis
3:277-295.
Cambridge Environmental Inc
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
EPA, 1985. Health Assessment Document for Polychlorinated Dibenzo-p-Dioxins. Office of
Health and Environmental Assessment, Washington, DC 20460. EPA/600/8-84/014F,
September 1985, Final Report.
R.E. Keenan, DJ. Paustenbach, Richard J. Wenning, A.H. Parsons, 1991. Pathology
reevaluation of the Kociba et al. (1978) bioassay of 2,3,7,8-TGDD; Implications for risk
assessment. J. Toxicol. Environ. Health 34:279-296.
6. Validation
Of the many sources that lend uncertainties to risk estimates, the appropriateness and validity of
the web of models used to simulate pollutant dispersal, movement, and accumulation is rarely
discussed, and never (to our knowledge) addressed rigorously in a quantitative manner. Since
any model is at best a representation of reality, model uncertainty is perhaps the most important
source of uncertainty,.and is especially critical to the assessment of the WTI facility, as
emphasized by our discussion of air dispersion modeling.
Based upon a reading of the single paragraph in the draft Project Plan devoted to model
uncertainty (page 105), we conclude that EPA's risk assessment will grant cursory attention to
model uncertainty. Such treatment is inadequate for an assessment of a setting as complicated
as the Ohio River Valley in which the WTI facility is located. In addition, as scientific research
has identified new processes important to pollutant fate and transport, the scope of a typical
multi-pathway risk assessment has expanded to envelop an ever-increasing complexity both in
the number of models used and in refinements and elaborations of models already developed.
Some of these models are based on state-of-the-art research findings, with unknown quality,
replicability, and general applicability.
Most of the models used in a risk assessment — especially those that trace pollutant movement
through various environmental media — are simple, first-order representations of the physical
processes that are believed to be most important. In many cases, these beliefs lack
confirmation. As a specific example, there are three models that predict the rates at which
vegetation can assimilate and bioaccumulate dioxin and furan congeners: direct uptake from
contaminated soil,2 deposition of particle-borne pollutant onto leaf surfaces, and direct uptake of
2
Even the model that is used to predict pollutant concentrations is inherently uncertain.
Within this model, a pollutant is assumed to mix uniformly to an arbitrary depth and be removed
from the soil layer by various physical processes. For organic chemicals such as dioxins and
furans, the removal rate is typically based upon limited empirical observations of half-lives for a
(continued...)
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
vapor by vegetation from the atmosphere. Based upon current knowledge, the second and third
processes are predicted to be more important. The draft Project Plan proposes that these
processes add in a linear fashion. It is equally possible, however, that any particle-borne dioxin
that deposits to plant surfaces may affect the vapor uptake process — for example, the dioxin
that is particle-borne and adhering to a leaf may be in equilibrium with the dioxin that adsorbs to
the leaf from the vapor phase. We have no knowledge of the importance of such competing
processes under field conditions, since no relevant experiments have been performed. This
uncertainty complicates any comparison of model predictions with the limited measurements of
dioxins and furans available in the literature, since field observations lack the detail to distinguish
such processes. For example, it is unclear whether the available experiments on vapor uptake
include or exclude (ambient) particle deposition.
Furthermore, most field data focus on a single homolog — Tetra-CDD. Thus, even if the models
are shown to be consistent with field observations of TCDD, the extrapolation of these models to
other dioxin and furan congeners is compromised because we cannot be certain the consistency
has resulted from the choice of the correct physical model. These considerations demand explicit
recognition and quantification of model uncertainty. As we discuss elsewhere, these and all
uncertainties can be addressed best by a probabilistic analysis.
7. Emission rate estimates
"High-end" estimates of emission rates should comport in at least some rough sense with reality.
At least three bounds on reality that should be recognized are: (1) conservation of mass (total
metal mass emitted cannot exceed total particulate mass); (2) Federal operating limits, which are
in general considerably more stringent than the Ohio limits referred to in the EPA draft Project
Plan as being (incorrectly) rate-limiting; and (3) demonstrated and routine removal efficiencies for
particulates and hence for the metals which bind thereto.
The Federal interim stack emission limits and feed rate limitations are imposed on the WTI facility
pursuant to the regulations on combustion of hazardous waste in boilers and industrial furnaces
promulgated by the EPA and published February 21,1991 (BIF rules). These limits are currently
2(...continued)
single congener (2,3,7,8-TCDD). For subsurface contamination, most estimates of the 2,3,7,8-
TCDD half-life are of the order of ten years. This half-life, however, has been shown to be much
shorter when the contamination is near the surface (possibly the result of photodegradation).
Since a finite amount of time is required to mix pollutants within a soil layer of any depth, and the
risk assessment models assume that pollutants deposit to the surface, it is not at all clear how to
choose a single half-life appropriate for an entire mixing depth.
Cambridge Environmental Inc
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX '
02141
-------
based on Reference Air Concentrations (RAG) and Risk Specific Doses (RSD) contained in the
BIF regulations, on the dispersion factor of 1.5 u,g/m3 per g/s developed as part of the U.S.
EPA's Preliminary Risk Assessment, and on the demonstrated destruction and removal
efficiencies achieved in the trial burns. The actual values are (Certified letter from W.E. Muno,
Director, Waste Management Division, EPA Region 5, to Waste Technologies Industries, c/o Mr.
D.J. Blake Marshall, dated October 20,1993):
Metal
Sb
Ba
Pb
Hg
Ag
Tl
As
Be
Cd
Cr
RAC (pg/m3)
0.30
50
0.09
0.08
3
0.50
RSD (ng/m3)
0.0023
0.0042
0.0056
0.00083
Interim
Maximum
Stack Emission
Level, Ib/hr
1.59
265
0.47
0.42*
15.9
2.64
0.0121
. 0.0222
' 0.0296
0.00439
Interim
Maximum Feed
Rate, Ib/hr
9.4
265
100
0.146
15.9
2.65
3.8
0.3
11.7
178
* The permitted stack emission limit for mercury in this table is in error; it exceeds the permitted feed rate.
According to this letter from EPA, WTI will be considered "to be in violation of its RCRA permit if
these emission limits and/or feed rates are not maintained". The EPA Project Plan should
certainly reflect the reality that the BIF rules will always be applied, so that emission rates will not
be permitted to exceed those allowed under BIF rules, independent of the Ohio limits.
In addition, there is the physical constraint on emission rates that the total metal emission rate
(excluding mercury) cannot exceed the total particulate emission rate, since all the metals are
emitted bound in or to particles. The stack temperature is 200°F, and the exhaust gases are
cooled below that before being reheated prior to being exhausted through the stack, so that it is
highly unlikely that any appreciable fraction of these metals (except mercury) will be in the vapor
phase in the stack emissions (only some chlorides and fluorides of arsenic and antimony, among
10
Cambridge Environmental Inc.
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
the compounds that might be formed in the oxidizing environment of the combustion system,
have any appreciable vapor pressures at such temperatures, but these compounds are soluble
and so will be removed by the wet scrubber system). The current limit on total particulate
emission rate corresponds to approximately 50 Ib/hour, but the measured particulate emission
rate has never exceeded approximately 11b/hour. This exceedingly low actual rate of particulate
emission is due largely to the operation of the electrostatic precipitator, any failure of which
automatically shuts down the incinerator. Overall, it must be recognized that the sum of the
metal emission rates (excluding mercury) cannot physically exceed the particulate emission rate,
no matter what the regulatory limits may be.
8. Mercury
The draft Project Plan acknowledges the complexity of modeling the fate and transport of
mercury, but it is clear that considerable work must be done. Given that (1) a worst-case hazard
index of 0.37 was found for mercury in the Phase I risk assessment and (2) mercury exposures
from contaminated fish are a perennial concern of multi-pathway risk assessments, the proposed
methodologies used to evaluate mercury are critical.
I
The current intent to model mercury as particle-bound — and evenly distributed in particles (page
45) — makes little sense. This statement is also contradicted on page 65, where the draft
Project Plan states that mercury will be treated as though condensed on particle surfaces (and so
not evenly distributed). The draft Project Plan states that most mercury will be emitted as vapor,
which agrees with a recent review of the literature (Campbell and Zankel, 1993). If mercury
vapor is to condense on particles, it is much more likely to condense on particles already in the
atmosphere rather than preferentially condensing evenly in new particles from the stack. Thus,
for consistency, mercury should be treated either as a vapor or a surface-weighted contaminant
on particles.3
As an additional note, the property values listed in Table 13 of the draft Project Plan also
demonstrate the cursory attention paid to mercury thus far. The diffusion coefficient of mercury in
air is listed as to be determined, yet it is one of the few property values available for mercury
(see, for example, Perry and Green, 1984). Moreover, there is no discussion as to whether the
properties listed are valid for the elemental, inorganic, or organic forms of mercury(or whether
such differentiation matters).
3 One might also expect other metals with relatively high vapor pressures (e.g., lead,
cadmium, arsenic, and possibly thallium) to volatilize during the combustion process and
subsequently condense. Thus, it may be appropriate to treat additional metallic compounds as
concentrated on the surfaces of particles. :
Cambridge Environmental Inc
11
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
Given its potential importance to the risk assessment, a detailed protocol for assessing mercury
should be developed and subject to review. If not already the case, we suggest that reviewers
with specific expertise in relevant aspects of mercury be consulted.
Campbell, S.A. and Zankel, K.L. (1993). A new method for estimating wet deposition of vapor-
phase mercury compounds from power plant and incinerator plumes. Presented at the
86th annual meeting of the Air & Waste Management Association (paper no. 93-WA-
73.06), Denver, Colorado, June 13-18,1993.
Perry, R.H, and Green, D. (eds.) (1984). Perry's Chemical Engineers'Handbook. 6th edition.
New York: McGraw-Hill Inc.
9. Retention of washout by plants
The EPA draft Project Plan suggests that the fraction of wet deposition that adheres to plant
surfaces (Fw in Table B-15) is a step function of the octanol-water coefficient, based on one of
the anonymous, unreferenced, undocumented, and apparently unreviewed "recommendations
made by the Exposure Assessment Working Group". There is no justification for such an
assumption, nor, apparently, any physical logic behind it. In the methodology of the draft Project
Plan, all the material washed out will be adsorbed or absorbed to particulate material. But
materials with large octanol-water coefficients will be strongly bound to such particulate material,
and it may be argued that they are the least likely.to be retained on the plant leaf, since they will
be washed off with the particles to which they are bound.
A possible logical and supportable estimate for this fraction may be derivable from the
methodology using local rainfall rates and the fraction of rain that adheres to leaves, suggested
by Brower et al., 1990 (appendix B, pages 2-4).
Brower, R., J. Gerritsen, K. Zankel, A. Muggins, N. Peters, S. Campbell, and R. Nilsson, 1990.
Risk Assessment Study of the Dickerson Site, Volume II — Appendices A-J. Prepared
for the Power Plant and Environmental Review Division, Maryland Department of Natural
Resources, Tawes State Office Building, Annapolis, Maryland. August, 1990.
12
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
10. Food consumption rates, Table A-9
This table is in error for at least the vegetables and fruits, since the "high-end" rates given are
identical to the "typical" rates. It is noteworthy that for both vegetables and fruits, the reference
given is to the Exposure Factors Handbook, but the Exposure Factors Handbook does not
contain all the values given in this table — there are no vegetable or fruit consumption entries
separated by age in that reference. Moreover, that reference gives an average consumption rate
of vegetables as 201 g/day, not the 228 g/day given (for "adult") in Table A-9 (the corresponding
value for fruit intake is 142 g/day, as reported in the Exposure Factors Handbook and Table A-9
under "adult").
11. Lead
The EPA draft Project Plan calls for the evaluation of lead by using the EPA U/BK model. There
are several problems with the discussion of this evaluation as given on page 54.
• Which version of the EPA lead U/BK model is to be used, and who is to peer-
review that model? Currently the EPA does not have an available U/BK model,
and will provide no details of previously released versions. The latest released
version is numbered 0.5, although others have apparently been circulated within
EPA. EPA will not currently provide details about any version of this model, and
recommends that it not be used.
• It is not stated what is to be the level of concern in the evaluation of lead. Is the
background level of lead to be included in the evaluation, and if so, how will that
background level be determined?
12. Photodegradation of vapors
Now that some of the dioxin/furan congeners of interest are assumed to be present in vapor
form, the photodegradation of these molecules in air must be modeled. Neither EPA's protocol
nor ours currently addresses this issue, but it should be addressed. One approach would be to
use available measurements on TCDD, with extrapolations to other congeners using measured
photo-quantum yields. Dr. William Lowenbach (personal communication) has provided us some
relevant ideas, which we share in the remainder of this section.
"^••P^-" —^^B™»—-
Cambridge Environmental Inc
13
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
The major experimental difficulty in studying the vapor phase photodegradation of PCDDs/PCDFs
is their low vapor pressure at room temperature. Although various studies have been purported
to estimate the rate of direct photolysis in the vapor phase, the best experimental study appears
to be that of pith ef a/. (1989).4
Orth et al. conducted photolysis experiments with vapor phase 2,3,7,8-TCDD under illumination
with a light source and filters to achieve radiation-simulating sunlight in the UV region from 250
nm to 340 nm. The rate constant for the photoloss in this experiment was 5.9 ± 0.5 x 10'3 sec"1.
The quantum yield for the process was determined to be 0.033 ± 0.046 (t1/2 = 2 minutes).
Podell, Jaber, and Mill (1986) estimated the photolysis rate of 2,3,7,8-TCDD vapors in the
atmosphere. Based on the quantum yield of TCDD in hexane (§ = 0.049), estimated error ±
42%, Dulin et al. 1985) and the atmospheric sunlight intensity at 40° latitude, a degradation rate
of 0.012 min"1 is obtained.
As observed by Podell et al. (1986), TCDD vapor will photolyze rapidly in the atmosphere
assuming the quantum yield is the same as in hexane. Direct photolysis may be an important
degradative pathway in the atmosphere that should be included in exposure analysis for
PCDDs/PCDFs. EPA has published a laboratory protocol for evaluating the fate of organic
chemicals in the atmosphere. Moreover, these results should be extended to other PCDDs and
PCDFs using the methods outlined in Tysklind et al. (1992).
Dulin, D., Drossman, H., and Mill, T. (1986). Products and quantum yields for photolysis of
chloroaromatics in water. Environ. Sci. Technol. 20:72.
Leifer, A., Brink, R., Thom, G., and Partymiller, K. (1983). Environmental transport and
transformation of polychlorinated biphenyls. U.S. Environmental Protection Agency, Office
of Toxic Substances, Washington, D.C. EPA-560/5-83-025.
Mill, T., Maby, W.R., Bomberger, D.C., Chou, T.W., Hendry, D.G., and Smith, D.H. (1982).
Laboratory Protocols for Evaluating the Fate of Organic Chemicals in Air and Water.
EPA/600/3-82-022.
4 Leifer et al. (1983) report a half-life of gaseous 2,3,7,8-TCDD in sunlight of 5 to 24 days.
Degradation may be the result of indirect photolysis through the attack of hydroxyl radicals on the
dioxin. Mill et al. report photolysis experiments with vapor phase TCDD; their preliminary results
suggest that the half life for vapor phase 2,3,7,8-TCDD in simulated sun is several hours.
Cambridge Environmental inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
Mill, T.f Rossi, M., McMillen, D., Coleville, M., Leung, D., and Spang, J. (1987). Photolysis of
tetrachlorodioxin and PCBs under atmospheric conditions. Internal report prepared by
STl International for U.S.EPA, Office of Health and Environmental Assessment. As cited
in: Estimating Exposure to Dioxin-Like Compounds, U.S. EPA, Office of Research and
Debelpment, Review Draft, EPA/600/6-88/005B, August, 1992.
Orth, R., Richie, C., and Hileman, F. (1989). Measurement of the photoinduced loss of vapor
phase TCDD. Chemosphere18-A275.
Podoli, T., Jaber, H., and Mill, T. (1986). Tetrachlorodibenzodioxin: rates of volatilization and
photolysis in the environment. Environ. Sci. Technol. 20:490.
Tysklind, M., Lundgren, K., Rappe, C., Eriksson, L, Jonsson, J., Sjostrom, M., and Ahlborg, U.
(1992). Multivariate characterization and modelling of polychlorinated dibenzo-p-dioxins
and dibenzofurans. Environ. Sci. Technol. 26:1023.
13. Vapor uptake by plants
The EPA draft Project Plan calls for an estimate of vapor uptake by vegetation that is based on
an unreferenced correlation with octanol-water partition coefficient, modified by a factor of 10
attributed to unreferenced, undocumented, and apparently unreviewed "recommendations made
fay the Exposure Assessment Working Group". From the coefficient of 1.14 for log(Kow), the
correlation appears to be based on Bacci's (1990) correlation for azalea leaves. This correlation
omits 1,2,3,4-TCDD, and Bacci published an update including this PCDD congener in 1991
(Patersonefa/., 1991).
However, for the most important vapor-phase uptake, that of the PCDD/PCDFs, there are direct
experimental measurements that should supersede such general correlations. These
measurements (McCrady and Maggard, 1993) were for 2,3,7,8-TCDD vapor-phase transfer to
grass. Since grass is exactly what is required for the major route of exposure (through cattle,
either via beef or via milk), this experiment is most apposite. In addition, McCrady et a/.'s
measurements allow separate evaluation of the photodegradation that occurs on plant leaves.
It appears that the fudge factor of to may be an attempt to modify the azalea model of Bacci to
obtain an estimate for grass. This has been proposed elsewhere, and we have commented
elsewhere on the attempted modification, and made suggested improvements. In any such
modification, care must be taken to estimate the correction factor correctly. What is required is a
Cambridge Environmental Inc.
15
58 Charles Street
617.225-0810
Cambridge, Massachusetts
617-225-0813 FAX
02141
-------
"McCrady" Bvpa5 that combines the Bacci model with the results of the McCrady and Maggard
study of 2,3,7,8-TCDD, essentially extrapolating McCrady and Maggard's data to other congeners
(and/or other chemicals). The "McCrady" Bvpa is estimated by applying a correction factor to the
Bvpa predicted by the Bacci model, where the correction factor is essentially the ratio of McCrady
and Maggard's measured Bvol6 for 2,3,7,8-TCDD to Bacci's predicted Bvol for 2,3,7,8-TCDD. In a
previous attempt at this correction factor, it was assumed that the ratio of the Bvols is equal to the
ratio of the Bvpas, but this is not a correct assumption.
The conversion factor from Bvpa to Bvol is dependent on the experimental situation. Specifically,
the conversion equation is of the following form:
v-air
where
B
B,
R.
Vol
vpa
V-plt
R,
volume/volume partition coefficient [(pg/L^, leaf)/(pg/Lair)],
mass/mass partition coefficient [(pg/gleafDwt)/(pg/gair)],
conversion from dry weight to fresh weight ratio (gDwt/gFwt),
conversion from fresh weight to liter of plant (gFwt/L^,,,), and
conversion from air mass to volume (gai/Lalr).
As shown, the conversion between Bvpa and Bvol depends on the fresh weight (Fwt) to dry weight
(Dwt) ratio and the grams (fresh weight) in a liter of fresh leaf. Both of these factors will vary for
different plants. Since the Bacci and the McCrady and Maggard experiments use different plants
— azaleas and grass, respectively — the conversion factors differ. The conversion factor for the
McCrady and Maggard experiment is given in Equation 9 of their paper. The conversion factor
for the Bacci experiment on azalea leaves is given in Equation 9 (oddly enough) of Bacci et al.
(1990). These equations are reproduced below:
5B,,
-vpa represents a mass/mass partition coefficient between plant and air with units similar to
(pg/gleafDWt)/(pg/gair). In the McCrady and Maggard and the Bacci article, the mass/mass partition
coefficient is referred to as BCFm.
6 Bvol represents a volume/volume partition coefficient between plant and air with units similar
'n tne McCrady and Maggard and the Bacci article, the volume/volume
to
partition coefficient is referred to as BCFV.
16
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
McCrady and Maggard:
Bacci ef a/.:
BCFV = 97.1 BCFm
BCFV = 224.37 BCFm
where
BCFV =
BCF =
and
Since the ratio of the Bvols is not the same as the ratio of the Bvpas, it is necessary to calculate the
correct ratio of the Bs. McCrady and Maggard report a log(Bvol) of 6.9. Using the above
conversion factor, the measured Bvpa for 2,3,7,8-TCDD vapor uptake by grass is 81 ,805
(1CP/97.1). The B^ for 2,3,7,8-TCDD predicted by Bacci's model is 1,728,192 (using Kow and H
values supplied by M. Lorber of the Office of Research and Development at EPA). Therefore,
the ratio of McCrady and Mag'gard's measured Bvpa to Bacci's predicted Bvpa for 2,3,7,8-TCDD is
0.047. This is about a factor of 2 smaller than the correction factor of 1/10 given in the EPA draft
Project Plan, if this is indeed the source of that correction factor.
If this is the objective of the approach taken in the EPA draft Project Plan, then it appears to be
reasonable as a method for extrapolating McCrady and Maggard's measurement of the 2,3,7,8-
TCDD air-to-leaf transfer factor to other congeners. However, slight modifications to this method
might improve the accuracy of the extrapolation. Bacci's model predicts the Bvpa for volatilization
only, not accounting for loss due to photodegradation. It relates the Bvpa to Kaw (or H) and Kow,
indicating that the rate of uptake over the rate of volatilization is proportional to Kaw and Kow. The
extrapolation method just described is to take the ratio of McCrady and Maggard's measured Bvpa
(volatility and photodegradation) for 2,3,7,8-TCDD to Bacci's predicted Bvpa (volatility only) for
2,3,7,8-TCDD and apply this ratio to Bacci's predicted Bvpas for the other congeners (and possibly
other chemicals). This method does not directly evaluate the rate of photodegradation; but it is
known that the photodegradation rate will vary with congener (and chemical).
i
The relationship between B^ and the uptake rate, the photodegradation rate, and the
volatilization rate is given by the following equation:
R =
ypa' X.
where
B,
V
vpa
mass/mass partition coefficient for air-to-leaf -transfer,
rate of uptake,
rate of volatilization, and
rate of photodegradation.
A summary of the experimental results from Bacci's studies of air-to-leaf transfer in azalea leaves
is presented in Table II. The Bacci model was developed by regressing log(BvpaxKaw) against
17
Cambridge Environmental inc
58 Charles Street
617-225-0810
Cambridge, Massachusetts
617 -225 -0813 FAX
02141
-------
l°g(K0J. This model assumes that the coefficient of log(Kaw) is -1. Alternatively, we suggest that
a multi-variable regression of logfBJ against log(Kow) and log(Kaw) might provide a better
prediction of Bvpa by not restricting the value of the coefficient of log(Kaw) — Bacci's model is
heuristic anyway since he does not provide substantial mechanistic detail for his expectation that
the coefficient is -1. With our multi-variable regression, the coefficient of !og(Kaw) is -0.427. It
should be noted that Bacci's regression excluded the data for Alachlor on the basis of a low Bvpa;
however, alachlor does not appear to be an outlier and it is included in our analysis.
Examining the rates of uptake and volatilization recorded in Table II, both log(ku) and log(kv)
correlate with log(Kow) and log(Kaw), but the correlation with log(Kaw) does not appear to be
significant in either case. Since k^, and \^ are independently correlated with Kow and Kaw, we can
develop a relation to predict !<„ and k,. Once predicted, \^ and k, can be used to estimate Bvpa,
the ratio of k,, to ^ (no photodegradation). Given the data in Table II, a multi-variable regression
of log(kj against log(Kow) and log(Kaw) leads to the following relationship:
log (ku) = -0.1399 log (KJ + 0.4396 log (Kow) - 1.3217
This same regression for \^ produces:
log (kv) = 0.2870 log (Kaw) - 0.3758 log (KJ - 0.1258
These equations can be used to predict'!^ and \^ in azalea leaves; what is wanted, though, is to
determine k,, and \^ in grass. McCrady and Maggard measured the rate of uptake and the rate of
volatilization for 2,3,7,8-TCDD in grass. By taking the ratio of the measured values of ku and l^ in
grass to the predicted ky and k, for 2,3,7,8-TCDD in azalea leaves, conversion factors can be
developed to estimate k^ and k,, in grass from the predicted values in azaleas. These factors are
shown in Table III.
By combining these ratios with the previous correlations, equations for estimating ku and \^ in
grass can be developed. These are as follows:
k = 15.91 X •JO(~al3"'°9(KJ * 0-«96 log (KJ - 1.3217)
k = 18 46 x io
(0'2870 '°9 (KJ " °'3758 109 (KJ " °'1258)
In order to predict the congener-specific B^, it is necessary to estimate kp, the rate of
photodegradation. McCrady and Maggard's experiment on vapor uptake of 2,3,7,8-TCDD in
grass measures the rate of photodegradation under natural outdoor conditions, that is, day and
night lighting conditions. The average rate of photodegradation in the McCrady and Maggard
experiment is 1.561 x 10~2 hr~1. We know of no measurements of the rate of photodegradation of
18
Cambridge Environmental Inc
58 Charles Street
617-225-0810
Cambridge, Massachusetts 02141
617-225-0813 FAX
-------
other PCDD/Fs on plants, but it may be possible to predict other kp's from some known
quantities.
One possibility is to use the photo-quantum yield (PQY). The degradation rate is proportional to
the PQY under the exposure conditions for which the PQY is measured. We do not know the
PQY under the actual exposure conditions, but might guess that it is proportional to the PQY
under some standard conditions, such as the common conditions of measurement in a water-
acetonitrile solution at 313 nm.
The photo-quantum yields for some specific isomers reported in the draft document of Estimating
Exposure (EPA, 1992) are listed in Table IV (we made no adjustments for differences in the
water-acetonitrile solutions and we assume that the units of the reported PQYs are consistent).
The ratio of the known kp to the photo-quantum yield for 2,3,7,8-TCDD can be applied to the
photo-quantum yields for the other congeners to estimate the corresponding kp as shown in
Table IV, so that the equation is:
kD = 7.10 (f> (7)
where
rate of photodegradation (hr1) and
-------
Table II Summary of data from Bacci's studies of azalea leaves
Compound
Trifluralin3
HCBa
Sulfotep3
Thionazin3
Mirexa
1,2,3,4-TCDDb
Dieldrin0
3A,3'A'-TCBC
p.p'-DDT"
p.p'-DDE"
a-HCHd
y-HCHd
PCB (60% Cl)d
Alachlor0
KU*
(hr1)
2.90
9.50
0.90
0.93
215.40
85.28
15.87
115.88
108.00
49.90
4.14
3.43
38.90
13.33
togflg
0.462
0.978
-0.046
-0.032
2.333
1.931
1.200
2.064
2.033
1.698
0.617
0.535
1.590
1.125
KV
(hr1)
6.2e-03
5.1e-03
9.06-03
7.7e-03
4.16-03
2.16-04
3.36-03
3.1e-04
5.66-04
3.56-04
9.06-04
1.06-03
4.66-04
1.16-02
logOO
-2.21
-2.29
-2.05
-2.11
-2.39
-3.68
-2.48
-3.51
-3.25
-3.46
-3.05
-3.00
-3.34
-1.97
Ka«
1.66-03
5.46-02
1.26-04
3.66-05
3.46-01
1.56-03
4.66-04
3.96-03
9.76-04
3.36-03
3.66-04
5.36-05
7.16-03
2.56-06
iog(Kaw)
-2.80
-1.27
-3.92
-4.44
-0.47
-2.82
-3.34
-2.41
-3.01
-2.48
-3.44
-4.28
-2.15
-5.60
iog(Kj'
3
6
3
1.2
6.9
6.6
3.7
6.7
6
5.7
3.8
3.8
6.9
2.8
BVpa
measured
465
1863
100
121
52537
406111
4808
373807
192857
142571
4600
3430
84565
1257
a Bacciefa/. (1990a)
b Bacciefa/. (1992)
0 Bacciefa/. (1990b)
d Bacci and Gaggi (1987)
8 In some cases the reported rate of uptake is a volume as opposed to a mass based rate.
cases, ky has been converted to a mass based rate by the conversion factor provided in the
' Provided in Bacci et al. (1990b)
In these
papers.
>ll
Cambridge Environmental Inc.
20
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
Table III Ratios of measured ku and kv to predicted ku and kv, respectively
H (atm-m3/mol)
Kaw
Iog(KJ
togflU
ka predicted (from the correlation with K^, and Kaw)
ky predicted (from the correlation with K,,w and Kaw)
k,, measured (from McCrady and Maggard (1993))
k, measured (from McCrady and Maggard (1993))
Ratio k,, measured to k,, predicted
Ratio k, measured to k, predicted
1.65e-05
6.75e-04
-3.171
6.64
109.86
2.94e-04
1747.5
5.435e-03
15.91
18.46
Table IV Estimating kp
Chemical
2,3,7,8-TCDD
1,2,3,4,7,8-HCDD
1,2,3,4,6,7,8-HCDD
1,2,3,4,6,7,8-OCOD
1,2,3,4,7,8-HCDF
Photo-quantum yield
(PQY)
2.20e-03
1.10e-04
1.536-05
2.26e-05
6.906-04
kp known
1.56e-02
Ratio of known kp
to PQY
7.10
kppred
1.56e-02
7.81e-04
1.096-04
1.606-04
4.906-03
Cambridge Environmental Inc
21
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
Table V Bacci and "McCrady" model vs. Proposed model
Chemical
2,3,7,8-
TCDD
1,2,3,4,
7,8-HCDD
1,2,3,4,6,
7,8-HCDD
1,2,3,4,6,7
8,9-OCDD
[/ a
•\>w
4.376-1-06
6.17e+07
1.58e+08
3.986+08
if a
^aw
6.756-04
1.836-04
5.366-05
2.766-04
"og(Kj
6.64
7.79
8.2
8.6
iog(Kaw)
-3.17
-3.74
-4.27
-3.56
Bvpa in Azaleas
(for volatility only-
W
Bacci
model
1.73e+06.
1.076+08
9.986+08
5.176+08
Proposed
model
3.736+05
5.64e+06
2.066+07
2.17e+07
Bvpa in Grass
(photodegradation and
volatility)
McCrady
B b
vpa
8.306+04
5.146+06
4.806+07
2.496+07
Proposed
model
8.306+04
3.116+06
1.536+07
1.55e+07
a Taken from August 1992 draft of Estimating Exposure to Dioxin-Like Compounds, (EPA, 1992)
b Derived using corrected ratio of 0.047.
Bacci, E., D. Calamari, C. Gaggi, and M. Vighi (1990a). Bioconcentration of Organic Chemical
Vapors in Plant Leaves: Experimental Measurements and Correlation. Environ. Sci.
Technol. 24:885-889.
Bacci, E., M. Cerejeira, C. Gaggi, G. Chemello, D. Calmari, and M. Vighi (1990b).
Bioconcentration of Organic Chemical Vapours in Plant Leaves: The Azalea Model.
Chemosphere 27:525-535.
Bacci, E., M. Cerejeira, C. Gaggi, G. Chemello, D. Calamari, etal. (1992). Chlorinated Dioxins:
Volatilization from Soils and Bioconcentration in Plant Leaves. Bull. Environ. Contam.
Toxicol. 48:401-408.
Bacci E., and C. Gaggi. Chlorinated Hydrocarbon Vapours and Plant Foliage: Kinetics and
Applications. Chemosphere 76:25:15-2522.
Dulin, D., H. Grossman, and T. Mill. Products and Quantum Yields for Photolysis of
Chloroaromatics in Water. Environ. Sci. Technol. 20:72-77.
22
Cambridge Environmental me
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
McCrady, J. and S. Maggard (1993). Uptake and Photodegradation of 2,3,7,8-
Tetrachlorodibenzo-p-dioxin Sorted to Grass Foliage. Environ. Sci. Technol. 27:343-350.
Paterson, S., D. Mackay, E. Bacci, and Davide Calamari, 1991. Correlation of the equilibrium
and kinetics of leaf-air exchange of hydrophobic organic chemicals. Environ. Sci.
Technol. 25:866-871.
U.S. EPA (1992). Estimating Exposure to Dioxin-like Compounds: Review Draft. Office of
Research and Development. EPA/600/6-88/005B.
14. Fugitive emissions and related issues
The Project Plan proposes (page 24):
a highly conservative screening level analysis [that] will be performed in the Phase
II Risk Assessment to provide an indication of the potential risks associated with
fugitive emissions from the facility. If this screening level analysis indicates that
fugitive emissions may be significant, a more detailed evaluation of facility
operations and design could be performed."
Such a proposal is too vague to review in detail, but we do wish to make the following points.
First, EPA and its contractor should work closely with WTI and Von Roll in developing relevant
information about fugitive emissions and failures. Emissions under such scenarios are heavily
dependent on the engineering design of the incineration train and ancillary systems. We
understand that considerable effort has been made to eliminate or minimize the effects caused by
malfunctions of equipment. Basing estimates of fugitive emissions on operations of existing
incinerators other than WTI could result in significant over-estimation.
, | .
Second, careful consideration should be given to assessing startup and shutdown, upsets, and
malfunctions. EPA's study of Dow's incinerator in Plaquemine, LA may contain useful procedures
for estimating fugitive emissions and emissions from upset conditions, but these must be adapted
to the WTI facility design. For example, WTI's incinerator does not include an emergency vent
stack, commonly used at some facilities to vent partially burned gases in the event of an air
pollution control system failure. The scrubbing system at WTI contains various features to keep it
operational during (partial) equipment failures. Estimation of the frequency of malfunctions
should be based on operating experience at WTI rather than on reports from other facilities.
Third, WTI conducts monthly assessments of fugitive emission sources. The reports of these
assessments should be used to estimate fugitive emissions from valves, flanges, pumps, and so
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX
-------
on. The engineering design for WTI's processing facilities (drum handling and storage, tank farm,
and bulk solid waste pits) has focused on minimizing fugitive emissions: vapors are collected at
the source and vented to a vapor recovery system that discharges either to the incinerator or to
an activated carbon adsorption unit. These details of design and operation should form the basis
for evaluating the potential for emissions from these sources.
Finally, data available at WTI should be used to help characterize the real effects of certain .
failure scenarios. The continuous emission monitoring system records emissions of acid gases
and other pollutants. These recordings continue during power failures and air pollution control
system upsets, so could be used to quantify the effects of such events. This data-based
approach is preferable to extrapolating from unrelated events at different industries.
24
Cambridge Environmental Inc
58 Charles Street Cambridge, Massachusetts 02141
617-225-0810 617-225-0813 FAX ^.u_s. GOVERNMENT PRINTING OFFICE: 1994 - 55o-o64/sooo6
-------
------- |