* \
I
Evaluation of TRIM.FaTE
Volume I: Approach and Initial Findings
Risk \*
Vhmiiucmcnl J
x Decision /
-------
EPA-453/R-02-012
September 2002
Evaluation of TRIM.FaTE
Volume I: Approach and Initial Findings
BY:
Randy Maddalena, Deborah Hall Bennett, and Thomas E. McKone
Lawrence Berkeley National Laboratory, Berkeley, California
Interagency Agreement #DW89786601
Bradford F. Lyon, Rebecca A. Efroymson, and Daniel S. Jones
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Interagency Agreement #DW89876501
Alison Eyth
MCNC Environmental Modeling Center, Research Triangle Park, North Carolina
Contract #GS-35F-0067K
Mark Lee, Margaret E. McVey, David Burch, Josh Cleland, and Baxter Jones
ICF Consulting, Fairfax, Virginia
Contract #s 68-D6-0064, 68-D-01-052
Prepared for:
Terri Hollingsworth, EPA Project Officer & Work Assignment Manager
Deirdre Murphy, Technical Lead
Emissions Standards Division
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Emissions Standards & Air Quality Strategies and Standards Divisions
Research Triangle Park, North Carolina
-------
DISCLAIMER
DISCLAIMER
This document has been reviewed and approved for publication by the U.S.
Environmental Protection Agency. It does not constitute Agency policy. The opinions, findings,
and conclusions expressed are those of the authors and are not necessarily those of the
Environmental Protection Agency. Mention of trade names or commercial products is not
intended to constitute endorsement or recommendation for use.
SEPTEMBER 2002 i TRIM.FATE EVALUATION REPORT VOLUME!
-------
PREFACE
PREFACE
This document, Evaluation ofTRIM.FaTE, Volume 1: Approach and Initial Findings, is
part of a series of documentation for the overall Total Risk Integrated Methodology (TRIM)
modeling system. Subsequent evaluation analyses may be presented in subsequent volumes,
while the detailed documentation of TRIM's logic, assumptions, algorithms, and equations is
provided elsewhere in comprehensive Technical Support Documents (TSDs) and/or user's
guidance for each of the TRIM modules.
This report describes a set of evaluation analyses performed on the TRIM.FaTE model
primarily during 2000, with some spanning into 2002.
Comments should be addressed to Dr. Deirdre Murphy, U.S. EPA, Office of Air Quality
Planning and Standards, C404-01, Research Triangle Park, North Carolina, 27711;
murphy.deirdre@epa.gov.
SEPTEMBER 2002 iii TRIM.FATE EVALUATION REPORT VOLUME!
-------
TABLE OF CONTENTS
TABLE OF CONTENTS
Disclaimer i
Preface iii
Table of Contents v
1. Introduction 1-1
1.1 Background 1-1
1.2 Types of Model Evaluation 1-3
1.3 General Approach of This Evaluation 1-4
1.3.1 Chemical Selection 1-5
1.3.2 TRIM.FaTE Mercury Case Study 1-5
1.3.2.1 Case Study Site Selection 1-6
1.3.2.2 Overview of Evaluation Activities 1-8
2. Conceptual Model Evaluation 2-1
2.1 Initial Activities 2-1
2.2 Documentation 2-2
2.2.1 Status Reports 2-2
2.2.2 TRIM.FaTE Technical Support Document 2-2
2.3 Science Advisory Board Reviews 2-3
3. Mechanistic and Data Quality Evaluation 3-1
3.1 Background 3-1
3.2 Selection of Chemicals for Evaluation Runs 3-3
3.2.1 Methods 3-3
3.2.2 Results and Discussion 3-3
3.3 Computer/Software Evaluations 3-4
3.3.1 Evaluation of the Prototypes 3-4
3.3.2 Overall Evaluation of Versions 1 Through 2.5 3-6
3.3.3 Algorithm and Compartment Audit 3-7
3.3.3.1 TRIM.FaTE Libraries 3-7
3.3.3.2 Audit Scope 3-9
3.3.3.3 Audit Methods 3-9
3.3.3.4 Findings/Results 3-10
3.3.3.5 Audit Summary and Conclusions 3-13
3.4 Testing of Individual Process Models 3-14
3.5 Air Process Model Evaluation 3-14
3.5.1 Comparison with the Urban Airshed Model 3-14
3.5.1.1 Approach and Model Setup 3-14
3.5.1.2 Results 3-16
3.5.2 Comparison with the Industrial Source Complex Model 3-21
3.6 Evaluation of Mercury Speiciation In Air and Soil 3-22
3.6.1 Mercury Species and Transformations in Air and Soil 3-22
3.6.2 Methods 3-23
SEPTEMBER 2002 v TRIM.FATE EVALUATION REPORT VOLUME I
-------
TABLE OF CONTENTS
3.6.3 Results 3-24
3.6.4 Conclusions 3-25
3.7 Sediment and Surface Water 3-26
3.7.1 Purpose of Evaluations 3-27
3.7.2 Model Setup and Assumptions 3-27
3.7.3 Results 3-28
3.7.4 Conclusion 3-28
3.8 Evaluation of the TRIM.FaTE Plant Module 3-28
3.8.1 Compositional Audit 3-28
3.8.1.1 Conceptual Design of the PI ant Module 3-31
3.8.1.2 Reconciling the Conceptual Model and the Code 3-33
3.8.2 Algorithm Audit 3-34
3.8.3 Future Activities for Evaluation of the Plant Module 3-37
3.9 Concentrations and Flows Through Terrestrial Wildlife 3-37
3.9.1 Model Setup 3-37
3.9.2 Results 3-39
3.9.3 Conclusions 3-41
3.10 Concentrations and Flows Through Fish 3-42
3.10.1 Model Setup and Evaluation Methods 3-43
3.10.2 Evaluations and Results 3-44
3.10.2.1 Basic Relationships 3-45
3.10.2.2 Structural Problems 3-45
3.10.2.3 Comparison of Alternative Models 3-46
3.10.2.4 Comparison of TRIM.FaTE Outputs to Measured
Concentrations 3-47
3.10.2.5 Sensitivity of Models to Biomass of Higher Trophic-Level
Fish 3-50
3.10.2.6 Options for Addressing Impact of Fish Biomass on Fish
Mercury Concentrations in the Bioenergetic Model 3-51
3.10.3 Conclusions and Summary 3-52
4. Structural and Complexity Evaluation 4-1
4.1 Background and Approach 4-1
4.1.1 Introduction 4-1
4.1.2 General Structural Evaluation Approach for TRIM.FaTE 4-2
4.2 Air Compartment Evaluation 4-3
4.2.1 Regular Grid with Controlled Variation in Meteorology 4-5
4.2.1.1 Model Inputs and Grid Layout 4-5
4.2.1.2 Results and Observations 4-5
4.2.2 Variation of Compartment Sizes 4-7
4.2.2.1 Model Inputs and Grid Layout 4-7
4.2.2.2 Results and Observations 4-8
4.2.3 Variation of Overall Grid Area 4-10
4.2.3.1 Model Inputs and Grid Layout 4-10
4.2.3.2 Results and Observations 4-10
4.2.4 Variation of Compartment Shape for a Constant Grid Area 4-13
4.2.4.1 Model Inputs and Grid Layout 4-13
SEPTEMBER 2002 vi TRIM.FATE EVALUATION REPORT VOLUME I
-------
TABLE OF CONTENTS
4.2.4.2 Results and Observations 4-13
4.3 Biotic Complexity Evaluation 4-16
4.3.1 Benzo(a)pyrene 4-16
4.3.1.1 Modeling Scenarios for Benzo(a)pyrene 4-16
4.3.1.2 Results for Benzo(a)pyrene 4-19
4.3.2 Mercury 4-20
4.3.2.1 Modeling Scenarios for Mercury 4-20
4.3.2.2 Results for Mercury 4-23
4.4 Temporal Complexity Evaluation 4-28
4.4.1 Benzo(a)pyrene 4-28
4.4.1.1 Modeling Scenarios for Benzo(a)pyrene 4-28
4.4.1.2 Results for Benzo(a)pyrene 4-30
4.4.2 Mercury 4-35
4.4.2.1 Model Setup for Mercury 4-35
4.4.2.2 Results for Mercury 4-35
4.5 Spatial Complexity Evaluation 4-38
4.5.1 Effect of Distance from Source 4-39
4.5.2 Effect of Horizontal Compartment Dimensions 4-40
4.5.3 Effect of External Boundary Compartments 4-42
4.5.4 Effect of Source Compartment Size ane Configuration 4-45
5. References for Volume I 5-1
Appendices
I-A TRIM.FaTE Algorithm Pairing Tables I-A-1
I-B Biomass of Fish I-B-1
SEPTEMBER 2002 vii TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
1. INTRODUCTION
TRIM.FaTE is a predictive environmental fate and transport model designed to support
decisions on programmatic policy and regulation for multimedia air pollutants. These decisions
can have far-reaching human health, environmental, and economic implications. It is important
that an assessment of how well the model is expected to perform the tasks for which it was
designed is incorporated within the model development process. In other words, the
trustworthiness of models used to determine policy or to attest to public safety should be
ascertained (Oreskes et al. 1994). This report describes the model evaluation activities
performed to date to assess TRIM.FaTE's quality and acceptability. In short, it describes the
progress made to date in fulfilling the Evaluation Plan laid out in Chapter 6 of the November
1999 TRIM Status Report (U.S. EPA 1999a).
The Evaluation Report is composed of two volumes. This volume, Volume I, presents
conceptual, mechanistic, and structural complexity evaluations of various aspects of the model
(e.g., inputs, process models). Volume II (bound separately) presents performance evaluation of
the model as a whole, focusing on initial case study application.
This first chapter of Volume I provides background information on model evaluation and
describes the general approach of the TRIM.FaTE evaluation. Chapter 2 describes conceptual
model evaluation activities for TRIM.FaTE. Chapter 3 describes mechanistic evaluation of
individual TRIM.FaTE process models and algorithms. Chapter 4 describes structural and
complexity evaluation of TRIM.FaTE. Chapter 5 identifies literature references cited in this
report. Background information related to evaluations is provided in two appendices to this
volume.
1.1 BACKGROUND
Most of the early efforts to establish the quality of models used in supporting policy
decisions focused on model validation. The term validation does not necessarily denote an
establishment of truth, but rather the establishment of legitimacy (Oreskes et al. 1994).
However, common usage is not consistent with this restricted sense of the term, and the term
validation has been commonly used in at least two ways: (1) to indicate that model predictions
are consistent with observational data, and (2) to indicate that the model is an accurate
representation of physical reality (Konikow and Bredehoeft 1992). The ideal of achieving - or
even approximating - truth in predicting the behavior of natural systems is unattainable (Beck et
al. 1997). As a result, the scientific community no longer accepts that models can be validated
using American Society for Testing and Materials (ASTM) standard E 978-84 (i.e., comparison
of model results with numerical data independently derived from experience or observation of
the environment) and, therefore, that modeling results can be considered "true" (U.S. EPA
1998f). It is unreasonable to equate model validity with the model's ability to correctly predict
the actual (unknowable) future behavior of the system. Instead, a judgment about the validity of
a model is a judgment on whether the model can perform its designated task reliably (i.e.,
minimize the risk of an undesirable outcome (Beck et al. 1997)).
SEPTEMBER 2002 1 -1 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
The current approach used by EPA is to replace model validation, as though it were an
endpoint that a model could achieve, with model evaluation, a process that examines each of the
different elements of theory, mathematical construction, software construction, calibration, and
testing with data (U.S. EPA 1998f). Therefore, the term evaluation is used throughout this
report to describe the broad range of review, analysis, and testing activities designed to examine
and build consensus about TRIM.FaTE's performance.
Over the last 10 years, the Agency has been considering model acceptance or model use
acceptability criteria for selection of environmental models for regulatory activities. The
Agency's efforts in this area are a result of EPA's Science Advisory Board (SAB)
recommendations in 1989 that "EPA establish a general model validation protocol and provide
sufficient resources to test and confirm models with appropriate field and laboratory data" and
that "an Agency-wide task group to assess and guide model use by EPA should be formed" (U.S.
EPA 1989). In response, EPA formed the Agency Task Force on Environmental Regulatory
Modeling (ATFERM). This cross-agency task force was charged to make "a recommendation to
the Agency on specific actions that should be taken to satisfy the needs for improvement in the
way that models are developed and used in policy and regulatory assessment and decision-
making" (Habicht 1992). In its March 1994 report, ATFERM recommended the development of
"a comprehensive set of criteria for model selection (that) could reduce inconsistency in model
selection and ease the burden on the regions and states applying the models in their programs,"
and they drafted a set of "model use acceptability criteria" (U.S. EPA 1994a).
More recently, an Agency white paper work group was formed to re-evaluate the
recommendations in the 1994 ATFERM report. As a result, EPA drafted the White Paper on the
Nature and Scope of Issues on Adoption of Model Use Acceptability Guidance (U.S. EPA
1998f), which recommends the use of updated general guidelines on model acceptance criteria
(to maintain consistency across the Agency) and incorporation of the criteria into an Agency-
wide strategy for model evaluation that can accommodate differences between model types and
their uses. The work group also recommended the initial use of a protocol developed by the
Agency's Risk Assessment Forum to provide a consistent basis for evaluation of a model's
ability to perform its designated task reliably. The White Paper was reviewed by SAB in
February 1999. The approach followed for evaluation of TRIM.FaTE, as described in this
document, is intended to be consistent with the Agency's current thinking on approaches for
gaining model acceptability.
In their May 1998 review of TRIM.FaTE, SAB recognized the challenge in developing a
methodological framework for evaluating a model such as TRIM.FaTE. Further, SAB suggested
that "novel methodologies may become available for quantitatively assuring the quality of
models as tools for fulfilling specified predictive tasks" (U.S. EPA 1998d). Comments regarding
the complexity of mercury environmental chemistry, made by SAB in their December 1999
review (U.S. EPA 2000), led to additional evaluation activities that include focus on organic
chemicals (e.g., benzo[a]pyrene). At that time, SAB also commented on the need for continual
evaluation. In developing and implementing the evaluation plan for TRIM.FaTE, the Agency
has attempted to incorporate the essential ingredients for judging the acceptability of
TRIM.FaTE for its intended uses, while retaining the flexibility to accommodate and evaluate
new methods that become available or changes in direction indicated by knowledge gained
through the evaluation process.
SEPTEMBER 2002 1 -2 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
1.2 TYPES OF MODEL EVALUATION
Model evaluation is necessary to increase the acceptance of a model. Furthermore,
evaluation is not a one-time exercise but a continuing and critical part of model development and
application. Several model evaluation methods that have emerged in recent years can be
categorized as: (1) those that focus on the performance or output from the model
(Dennis et al. 1990, Hodges and Dewar 1992, U.S. EPA 1994b, Cohn and Dennis 1994, Spear
1997, Schatzmann et al. 1997, Arnold et al. 1998), and (2) those that test the internal consistency
(Beck et al. 1997, Beck and Chen 1999) or scientific credibility (Eisenberg et al. 1995) of the
model. All of these methods can be placed into one of two basic categories: (l)These methods
range from objectively matching model output with measurement data to more subjective and
abstract quality measures (e.g., expert judgment, peer review).
Model evaluation can be viewed as a consensus building process (Figure 1-1) including
three aspects as identified by Beck et al. (1997): (1) model composition, (2) model performance,
and (3) task specification. This process was recognized in the Agency's December 1998 White
Paper (U.S. EPA 1998f).
Figure 1-1
Schematic Representation of the Model Evaluation Process
Increasing Acceptability
Model
Composition
Model
Performance
Task
Specification
Conceptual model development
and review
Code verification
Model documentation
Performance evaluation
through a wide range of
applications and analyses
Continued structural and
sensitivity analysis
Round robin analysis
Peer review
Sensitivity analysis
Hypothetical case studies
Model-to-model comparison
SEPTEMBER 2002
1-3
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
The evaluation activities performed to date for TRIM.FaTE correspond to different (but
overlapping) types of model evaluation activities:
Conceptual model evaluation;
Mechanistic and data quality evaluation;
Structural evaluation; and
Performance evaluation.
The first three evaluation activities primarily focus on the information that goes into the model
(e.g., theory and data); how this information is synthesized (e.g., process models, algorithms, and
assumptions); and how the finished model is set up (e.g., appropriate level of complexity). The
fourth evaluation activity focuses mainly on the information that comes out of the model (e.g.,
comparing overall model outputs to various kinds of benchmarks). Detailed methods and results
of the first three evaluation types are presented in subsequent chapters of this volume. The
initial performance evaluation activities for TRIM.FaTE are presented in Volume II.
The model evaluation plan for TRIM.FaTE was designed at the output to be flexible.
Results from the TRIM.FaTE evaluation
efforts have posed new questions and led
to additional review, analysis, and testing,
not all of which is described here. The
various evaluation activities performed on
TRIM.FaTE increase the experience and
understanding that will ultimately lead to
a judgment about its quality, reliability,
relevance, and acceptability. The
activities that are currently part of the
consensus building process for
TRIM.FaTE are described in the
following sections. At this time, there has
been substantial progress on a number of
these activities (e.g., code verification,
model documentation, peer review,
mechanistic evaluation of individual
process models and algorithms). Other
evaluation activities (e.g., complexity analyses, overall performance evaluation) are continuing
and will continue with new applications and various analyses.
1.3 GENERAL APPROACH OF THIS EVALUATION
The evaluation of TRIM.FaTE is an iterative process, starting with simpler analyses and
proceeding to more-complex studies. Early model analyses, especially those mechanistic
evaluations focusing on one fate and transport aspect of TRIM.FaTE, have used limited time
periods, simplified modeling layouts (e.g., a single, square modeling compartment composed of
surface soil, surface water, and air volume elements), and simplified assumptions (e.g., fixed
concentrations in abiotic media). More-advanced layouts were constructed for some
intermediate evaluation activities (e.g., temporal and spatial complexity analyses). In order to
EVALUATION THROUGHOUT MODEL
DEVELOPMENT
As noted in the text, model evaluation is being
performed in conjunction with model development.
Earlier evaluation activities were performed using
the most current Prototype (i.e., I through V) of
TRIM.FaTE available at the time. These evaluation
activities are fully applicable to TRIM.FaTE Version
1.0, which was built from the same simulation
algorithms and data as Prototype V. Version 1.0 is
also the focus of model evaluation activities
described here, and Version 2.0 is the focus of
much of the mercury test case (see Volume II of the
Evaluation Report); later versions of TRIM.FaTE
will be used as evaluation activities and model
development continue.
SEPTEMBER 2002
1-4
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
test the whole model in a realistic setting, the mercury case study described in this section (and
in more detail in Volume II) was used. Various parts of this case study setup were used in most
of the evaluation activities described in both volumes of the Evaluation Report.
1.3.1 Chemical Selection
As part of the evaluation process for
TRIM, EPA must test TRIM.FaTE with both
organic and inorganic pollutants because of
their distinctly different multimedia fate and
transport properties. The EPA selected
PAHs for an initial organic chemical test
case, and the methodology and results of
that testing were reported in the 1998 TRIM
Status Report (U.S. EPA 1998b). For the
evaluations of the current version of
TRIM.FaTE, benzo(a)pyrene and mercury
were selected. Benzo(a)pyrene was used in
some analyses in order to test an organic
compound. The Agency selected mercury as
an inorganic chemical for testing
TRIM.FaTE because of its fate and transport
properties (e.g., transformation to multiple
chemical species), the concern for
multipathway exposure (particularly through
associated with exposure. In some instances
mercury was used with the assumption of no
1.3.2 TRIM.FaTE Mercury Case Study
MERCURY
Mercury is one of the 188 HAPs listed under
section 112(b) of the CAA, is one of 33 HAPs
being addressed by the Integrated Urban Air
Toxics Strategy under section 112(k) (U.S. EPA
1999e), is a pollutant of concern under the
section 112(m) Great Waters program (U.S.
EPA 1999b), and is one of the seven specific
pollutants listed for source identification under
section 112(c)(6). In addition, the findings of the
Mercury Study Report to Congress (U.S. EPA
1997) indicate that mercury air emissions may
be deposited to water bodies, resulting in
mercury uptake by fish. According to that report,
ingestion of mercury-containing fish is a critical
environmental pathway of concern for mercury-
related health effects in humans, particularly
developmental effects in children.
ingestion offish), and the potential health effects
(e.g., some air evaluation activities), elemental
transformation.
As a part of the evaluation activities for TRIM.FaTE, OAQPS has developed a case study
data set for mercury at a chlor-alkali plant in the U.S. This case study data set has been used in
sensitivity analyses and in mechanistic and structural evaluations, which have improved
understanding of the most important model processes and inputs and of the effects of varying the
model's spatial and temporal resolution. After gaining an understanding of and confidence in
the model's structure and performance, OAQPS will proceed to fuller spatial and temporal case
study simulations. The TRIM.FaTE case study outputs will be compared with outputs from
other models applied to the site, as well as biotic and abiotic mercury measurements available for
the case study area. The case study site and conditions also have served as the basis for
extensive testing and troubleshooting of TRIM.FaTE. This section provides summary
information on the mercury case study, including information on selection of the test site and an
overview of the evaluation activities. In the future, EPA may perform additional case studies
and apply TRIM.FaTE to other chemicals (e.g., dioxins) and other locations.
SEPTEMBER 2002
1-5
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
1.3.2.1 Case Study Site Selection
After selecting mercury for this case study, the Agency evaluated different stationary
sources of mercury that are significant on a national basis. The four types of stationary sources
with the highest total national air emissions of mercury, based on the findings of the Mercury
Study Report to Congress (U.S. EPA 1997), are - in order of highest to lowest mercury
emissions - electric utility plants, municipal waste combustors, medical waste incinerators, and
chlor-alkali plants. Electric utility plants, which are addressed in section 112(n) of the CAA, are
still undergoing evaluation by EPA for possible regulation of mercury air emissions. For
municipal waste combustors and medical waste incinerators, national air emission standards
have been promulgated under section 129 of the CAA, and these standards are expected to result
in large reductions of mercury air emissions.
Chlor-alkali plants were selected for further assessment in the TREVI.FaTE case study
because they are a substantial source of mercury air emissions and are not yet regulated for HAP
emissions. In addition, these plants are more likely than other major mercury emission sources
to pose localized health concerns as a result of their lower stack heights and relatively high
estimated level of fugitive emissions.
The Agency selected a single chlor-alkali plant for the mercury case study after
evaluating data availability for several sites. At the time of the site selection, 14 chlor-alkali
plants were in operation in the United States. Mercury air emission estimates were available for
all 14 plants; however, data on mercury levels in environmental media and biota were available
for only two of the plants. Fish tissue, water quality, and air quality data had been collected for
one of the two plants, but ultra-clean techniques were not used for collecting and analyzing the
water samples. For the second plant, air quality, soil, fish tissue, sediment, and additional biotic
data had been collected and analyzed. In addition, accumulation of mercury in environmental
media and biota near the second plant was possible because the plant has been in operation since
1967. Because the data set for the second plant was more complete, of higher quality, and
readily available for use, that chlor-alkali plant was selected for the mercury case study. A
schematic map of the site area showing delineation of the simple set of parcels used for many of
the evaluation activities is provided in Figure 1-2. (For a general discussion of the process of
defining parcels, volume elements, and compartments for a TREVI.FaTE application, see Chapter
5 of TRIM.FaTE TSD Volume I).1 Maps showing the full, more complex parcel layouts used for
the overall performance evaluation are included in Volume II.
1 While the case study site is a real facility and site-specific data are being used to the extent available, the
name and location of the site are being kept confidential.
SEPTEMBER 2002 1 -6 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
Figure 1-2
Simplified Parcel Layout for TRIM.FaTE Mercury Case Study Site2
t
2 This diagram shows the initial set of surface water (i.e., river, pond) and soil (i.e., all other) parcels for the
TRIM.FaTE mercury case study site; the air parcels are slightly different.
SEPTEMBER 2002
1-7
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
1.3.2.2 Overview of Evaluation Activities
As part of the TRIM.FaTE model evaluation, several different types of analyses that
correspond with different types of evaluations (i.e., mechanistic and data quality, structural,
performance) are being performed using the case study data set. These analyses are described in
general below and in more detail in subsequent chapters and in Volume II. The model input
values developed for the TRIM.FaTE mercury study are documented in an appendix of Volume
II.
Evaluating the quality of the input data for a given model application is an iterative
process. A literature search is completed to determine the value and identify any available
information on the predicted uncertainty or variability associated with that value. The current
values resulting from our search are listed in an appendix to Volume II. Then, a sensitivity
analysis is performed for all of the parameters to evaluate how varying an input value influences
the model output. If a model input is very uncertain and significantly influences the model
output, more research may be completed to refine that input value.
Evaluating the model's internal mechanisms (i.e., mechanistic evaluation) involves
assessing selected chemical fate and transport algorithms used in the model. In addition to
assessing selected components of the model, intermediate processes, such as flows between
compartments, are assessed to ensure that the model accurately represents the current
understanding of physical and chemical processes. It also must be confirmed that the algorithms
work effectively together within the model. Because of the number of compartment types and
links included in TRIM.FaTE, this is a complex process.
For example, one mechanistic evaluation performed was a comparison of the
TRIM.FaTE air component with a commonly used air dispersion model, the Urban Airshed
Model (UAM) available at http://www.epa.gov/scram001/. Specifically, the air concentrations
from UAM were compared to the concentrations estimated for the air compartments in
TRIM.FaTE to provide insight into how the methodology for modeling transport and fate in
TRIM.FaTE compares to a grid model with a track record of application and acceptance.
Another type of evaluation being performed using the TRIM.FaTE mercury case study
data set is an assessment of the influence of the structural representation of the system being
modeled. Some of the key assumptions in any TRIM.FaTE application involve determination of
the time step for input data averaging, the background and boundary concentrations of chemicals
of interest, the spatial representation (i.e., grid layout) of the modeled system, and the
compartment types selected for modeling. Examples of structural evaluation, some of which
have been performed and are reported here, include the following:
Understanding the effect of temporal variability, by assessing the impact of the
temporal resolution of the meteorological and source emissions data on model
outputs;
Understanding the effect of spatial configuration, by comparing results
obtained using spatial layouts of varying complexity and resolution; and
SEPTEMBER 2002 1 -8 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 1
INTRODUCTION
Determining the effect of external boundaries on internal compartments, by
assessing, for example, whether wind direction changes result in elevated
concentrations in the air advected back into the system.
Model performance evaluation can include comparisons of model outputs to outputs from
other models and to available measurement data for a specific site. Both of these types of
performance evaluations are being or have been performed as part of the TRIM.FaTE mercury
case study discussed in Volume II.
SEPTEMBER 2002 1 -9 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 2
CONCEPTUAL MODEL EVALUATION
2. CONCEPTUAL MODEL EVALUATION
Conceptual model evaluation activities
focus on the theory and assumptions
underlying the model. These activities
seek to determine if the model is
conceptually sound.
Conceptual model evaluation is initiated in
the early stages of model development. During the
process of framing the problem and designing the
conceptual model, the appropriate level of
modeling complexity (e.g., what to include and
what to exclude), the availability and quality of
information that will be used to run the model (i.e.,
input data), and the theoretical basis for the model should be evaluated. A literature review
should be undertaken to identify and evaluate the state-of-the-science for processes to be
included in the model, as well as to compile and document the initial set of values that will be
used as model inputs.
Examples of conceptual model evaluation activities include:
Literature review;
Development and review of model documentation; and
Peer review of problem definition and modeling concepts and approaches.
2.1 INITIAL ACTIVITIES
Considerable progress has been made in developing, documenting, evaluating, and
refining TRIM.FaTE, including the following.
An initial literature review identifying the state-of-the-science and the rationale
for development of TRIM.FaTE has been completed (U.S. EPA 1997b; U.S. EPA
1997c), and the problem and design objective have been clearly defined (U.S.
EPA 1998c).
Extensive model documentation has been presented:
TRIM Status Reports have been published in 1998 (U.S. EPA 1998b) and
1999 (U.S. EPA 1999a);
Presentations have been made at scientific meetings including the Society
of Environmental Toxicology and Chemistry (SET AC) annual meetings in
1997 (McKone et al. 1997a; Zimmer et al. 1997; Efroymson et al. 1997),
1998 (Vasu et al. 1998), 1999 (Efroymson et al. 1999), and 2000 (Murphy
et al. 2000; Lyon et al. 2000; Maddalena et al. 2000; Efroymson et al.
2000; Bennett et al. 2000; Burch et al. 2000a; Hetes and Langstaff 2000;
Fine et al. 2000; Bennett et al. 2000b); the Society for Risk Analysis
(SRA) in 1997 (Vasu et al. 1997; Guha et al. 1997; Lyon et al. 1997;
Bennett et al. 1997; McKone et al. 1997b; Johnson et al. 1997); and the
International Societies of Exposure Analysis and Environmental
Epidemiology (ISEA/ISEE) in 2002 (Murphy et al. 2002).
SEPTEMBER 2002 2-1 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 2
CONCEPTUAL MODEL EVALUATION
- A detailed Technical Support Document for TRIM.FaTE is available (U.S.
EPA 2002a and U.S. EPA 2002b).
Aspects of TRIM have been published in peer-reviewed j ournals (Palma et
al. 1999; Efroymson and Murphy 2001).
Two reviews by the SAB have been published (U.S. EPA 1998a; U.S. EPA 2000).
As refinements to TRIM.FaTE are made and as new applications are performed, conceptual
model evaluation will continue. These evaluations will continue to be reported in peer reviewed
journals and will be subject to additional SAB consultation and review.
2.2 DOCUMENTATION
Previously published TRIM.FaTE documentation include Status Reports and the
TRIM.FaTE Technical Support Document.
2.2.1 Status Reports
The first TRIM Status Report was published in March 1998 (U.S. EPA 1998b). This
report focused on the first developmental phase of TRIM, including the conceptualization of
TRIM and the implementation of the conceptual approach through the development of
TRIM.FaTE. Many aspects of the conceptual evaluation are described in the first Status Report,
including the initial goals and objectives of the TRIM project, the conceptual framework for
TRIM.FaTE (including a review of currently available models and tools), development of the
first prototype versions of TRIM.FaTE, and the limited testing and model evaluation analyses
that were completed on these prototypes.
A second TRIM Status Report was published in November 1999 (U.S. EPA 1999a). This
report summarized work performed on TRIM during the second developmental phase, including
the refinement of the initial TRIM.FaTE module following the 1998 SAB review of TRIM.
Details regarding TRIM.FaTE capabilities and the algorithms implemented in TRIM.FaTE, as
well as the plan for the evaluation described in the current document, were included.
Descriptions of the exposure and risk characterization modules of TRIM (i.e., TRIM.Expo and
TRIM.Risk) were also included in the 1999 Status Report.
2.2.2 TRIM.FaTE Technical Support Document
The TRIM.FaTE Technical Support Document is composed of two volumes (U.S. EPA
2002a; U.S. EPA 2002b). The first volume provides a description of the terminology, model
framework, and functionality of TRIM.FaTE. Volume II presents detailed descriptions of the
algorithms used in the TRIM.FaTE module.
In addition to SAB review (see Section 2.3), an internal draft of the Technical Support
Document, along with the Status Reports, were subjected to formal review by representatives
from the major program offices at EPA and an EPA Models 2000 review team.
SEPTEMBER 2002 2-2 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 2
CONCEPTUAL MODEL EVALUATION
2.3 SCIENCE ADVISORY BOARD REVIEWS
To date, two reviews of TRIM have been completed by the Environmental Models
Subcommittee of the Executive Committee of SAB. The first review, undertaken in May 1998
(U.S. EPA 1998a), focused on the conceptual approach for TRIM and the prototype of
TRIM.FaTE that was available at that time. Six charge questions related to TRIM and
TRIM.FaTE were posed to SAB. Responses to each question and recommendations to EPA for
improvements in the next versions of TRIM modules and TRIM.FaTE in particular were
summarized in a December 1998 report by SAB (U.S. EPA 1998a).
Overall, in its first review, SAB found the development of TRIM and the TRIM.FaTE
module to be conceptually sound and scientifically based. The SAB recommended that the
TRIM team (1) seek input from users before and after the methodology is developed to
maximize its utility; (2) understand the potential uses of TRIM to guard against inappropriate
uses; (3) provide documentation of recommended and inappropriate applications; (4) provide
training for users; (5) test the model and its subcomponents against current data and models to
evaluate its ability to provide realistic results; and (6) apply terminology consistently.
The second SAB review of TRIM took place in December 1999 and focused on the
TRIM Status Report (U.S. EPA 2000), the review draft of the two volumes of the TRIM.FaTE
Technical Support Document (U.S. EPA 1999c; U.S. EPA 1999d), and a separate draft
Technical Support Document developed for the TRIM.Expo module (U.S. EPA 1999b). Three
charge questions related to the overall TRIM system and three questions regarding the
TRIM.FaTE module in particular were posed to SAB, along with several questions regarding
other TRIM modules. The SAB's responses and comments are summarized in a final report
dated May 2000 U.S. EPA 2000. In this report, SAB described EPA's TRIM development
efforts as being "innovative and effective, given the significant challenges and the relatively new
and rapidly evolving state of science for multimedia fate, transport, exposure, and risk models."
Specific recommendations were proposed for the charge questions.
SEPTEMBER 2002 2-3 TRIM.FATE EVALUATION REPORT VOLUME I
-------
[This page intentionally left blank.]
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3. MECHANISTIC AND DATA QUALITY EVALUATION
This chapter presents the results of the mechanistic and data quality evaluation of
TRIM.FaTE during 1999 and 2000. Summaries of the initial shakedown evaluation of
TRIM.FaTE and the computer and software evaluations are included. This chapter also includes
descriptions of the evaluations of a number of the individual process models that comprise
TRIM.FaTE.
3.1 BACKGROUND
Mechanistic and data quality
evaluation activities focus on the
specific algorithms and assumptions
used in the model. These activities seek
to determine if the individual process
models and input data used in the model
are scientifically sound, and if they
properly "fit together."
Multimedia fate models are built around a
series of process models (i.e., algorithms or groups
of algorithms) that make up the mechanics of the
model. In some cases, individual process models
are taken directly from the literature and have been
tested previously for performance and peer
reviewed. The prior testing and review provides a
degree of confidence that the process model
correctly captures the behavior of the processes it is
intended to model. New process models and assumptions are often introduced during model
development; these new components need to be evaluated individually to ensure that they are
working properly.
Mechanistic and data quality evaluations help to elucidate the internal workings of the
model and, when necessary, provide a basis to refine process models and assumptions that play a
critical role in the calculations. Sensitivity analysis methods are used to identify important
model inputs during mechanistic evaluations and to identify the process models having the
greatest influence on the model output. For example, alternative algorithms for the same process
can be modeled and the results compared. Similarly, each time the model is used for a new kind
of application, a sensitivity analysis may be appropriate to identify inputs, algorithms, and
assumptions that have the greatest influence on the model outcome in that application. The
quality and reliability of these influential factors directly affect the quality and reliability of the
outcome from the analysis (Maddalena et al. 1999; Taylor 1993). When feasible, these
influential factors should be refined to provide the best inputs to the analysis or, at the very least,
identified as a potential source of uncertainty in the outcome.
Some mechanistic and data quality evaluation activities consider the model in its entirety.
Process models are typically developed and tested in controlled or simplified systems.
Therefore, how well these individual process models will perform when combined with other
models in a fully coupled system is unknown. Mechanistic and data quality evaluations are
designed and used to measure certain bounded indices of performance (e.g., mass balance,
appropriate and realistic mass transfer rates, relative concentrations within reasonable bounds).
In addition, algorithms or routines that are used in a model to manipulate the data or to solve a
system of equations (e.g., LSODE, the differential equation solver used in TRIM.FaTE) need to
be tested during the mechanistic evaluation to ensure proper performance.
SEPTEMBER 2002 3 -1 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Examples of mechanistic and data quality evaluation activities performed on TREVI.FaTE
include:
Computer code verification;
Verification of generic algorithms adapted for and used within a model;
Literature review to determine the extent of prior process model testing;
Peer review of model components;
Mass or molar balance checks;
Performance evaluation of new and existing individual process models and of
multiple process models in a linked system (e.g., compare with existing models or
with measurements, when available);
Comparison of alternative process models (e.g., equilibrium versus bioenergetic
model for fish bioaccumulation of mercury);
Data acquisition and evaluation (e.g., data quality or reliability relative to the
other inputs and assumptions), and development and documentation of default
input data;
Distribution development for input data to support probabilistic analysis; and
Generic sensitivity analysis to help identify parameters that are most influential
on model results, as well as potential data limitations (i.e., model inputs that need
further refinement or that are potential sources of uncertainty in the analysis).
One of the features of TRIM.FaTE that aids in mechanistic and data quality evaluation
(as well as in other types of evaluation) is its web-based output functions. There is an option to
create a "full-recursive output," which documents the mass flow, as well as the associated
transfer factors, to and from each compartment. Mass and molar balance checks are
incorporated in the model for non-transforming organic compounds and mercury to allow for the
quick assessment of model performance under a range of conditions. The equation for each
transfer factor can be viewed on a separate web page, and any calculated quantities used in that
equation can then be viewed on additional pages. In this manner, checks can be made to ensure
that the equations are input properly, and that the computer code is correctly calculating
intermediate values. Analyses have been conducted on various parts of the code using this
feature.
SEPTEMBER 2002 3 -2 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.2 SELECTION OF CHEMICALS FOR EVALUATION RUNS
Prior to conducting detailed evaluations of the process models within TRIM.FaTE,
numerous preliminary model runs were performed in a "debugging" mode. Given the amount of
information produced in a full run, a new approach to evaluating performance was adopted in
order to evaluate whether the model was producing results that were logical, internally
consistent, and reasonable. Thus, a screening set of hypothetical chemicals was developed and
used to conduct a systematic probe of the model across the range of applicable fate scenarios.
3.2.1 Methods
The environment in its simplest form can be divided into solid, aqueous, and gaseous
phases. The relative solubility of a chemical in each of these phases is indicative of how a
chemical will partition when released to the environment. The octanol/water partition
coefficient (Kow) and the non-dimensionalized Henry's Law constant (Kaw) provide a general
means to characterize the relative solubility of a chemical in the three primary environmental
phases (Cole and Mackay 2000; Cousins and Mackay 2000).
Crystal Ball software (Decisioneering 1996) and existing data on several hundred
chemicals were used to construct correlated probability distributions for the primary physical-
chemical properties used in TRIM.FaTE. A simple Monte Carlo sampling scheme was then used
to draw 500 random combinations from the correlated distributions. The 500 candidate
chemicals were then run through an existing mass balance model (McKone 1993a,b,c) to
evaluate their partitioning behavior. The results are plotted in Figure 3-1, where each physical-
chemical property combination is defined by a unique pair of Kow and Kaw values.
Single-medium chemicals are arbitrarily defined as those that have more than 90 percent
of their mass in a single medium. Multimedia chemicals are defined as those having not more
than 80 percent of their mass in a single medium. Chemicals with between 80 percent and 90
percent of total mass in a single medium were excluded from selection. The 500 candidate
chemicals were classified according to their partitioning behavior. Two chemicals were then
selected at random from each single-medium pollutant class, and three chemicals were selected
from the multimedia pollutant class. The resulting nine test chemicals made up the initial
shakedown evaluation set for TRIM.FaTE.
3.2.2 Results and Discussion
The test set was particularly useful during diagnostic evaluations. Having a general
understanding of the expected fate of a chemical provides insight into possible reasons for
unexpected model outcomes. For example, intermedia transfer for the single-medium gas-phase
pollutants occurs typically by diffusion and, to a lesser extent, washout from the atmosphere
during rain events. If TRIM.FaTE results during an initial evaluation are suspect for the single-
medium gas-phase chemicals, then the focus of the diagnostic evaluation can be placed on a
relatively small number of algorithms. This approach was used with TRIM.FaTE by running the
model with only a subset of the available compartment types to focus on a particular algorithm
or set of algorithms. For the TRIM.FaTE evaluation phase, the test set of chemicals was used to
SEPTEMBER 2002 3-3 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
evaluate the soil algorithms, the plant algorithms, and the general biotic algorithms. The general
test set will continue to be used for both initial evaluations and diagnostic evaluations for the
process models in TRIM.FaTE and for the model as a whole.
Figure 3-1
Single-Medium and Multimedia Chemical Regions for 500 Hypothetical Chemicals
ra
X
O)
O
» > 90% Air
a > 90% Aqueous
A > 90% Solid
x Multimedia*
Pseudo Test Chemicals
-15
-20
-2
* Multimedia is defined as having not more than 80 percent of total mass in any single medium. Chemicals with between 80 percent and 90
percent of total mass in any single medium were excluded from selection.
3.3 COMPUTER/SOFTWARE EVALUATIONS
This section provides an overview of computer/software evaluations for the successive
developmental versions of TRIM.FaTE. Section 3.3.1 describes evaluations of prototypes I
through V. Section 3.3.2 describes evaluations of TRIM.FaTE Version 1 through 2.5, the first
production versions of the model. Section 3.3 describes audits of the algorithm and
compartment sections of the TRIM.FaTE library.
3.3.1 Evaluation of the Prototypes
The TRIM Computer Framework started as a series of prototypes. Prototypes I-IV were
implemented using a combination of Microsoft Visual Basic, FORTRAN, and Microsoft
Excel software. Prototype V, the final prototype, was written in Visual Basic 5 and used an
Access database to store information about the model configuration and input parameters.
Prototype V was used for many of the initial mechanistic evaluations of TRIM.FaTE.
SEPTEMBER 2002
3-4
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Each of the prototypes used the Livermore Solver for Ordinary Differential Equations
(LSODE) (Radhakrishnan and Hindmarsh 1993) to solve the system of linear ordinary
differential equations that represent the transfer of chemical mass between compartments and
transformation of chemical mass within compartments. LSODE is a FORTAN program freely
available via several online numerical algorithm repositories. Early on, the use of LSODE in
TRIM.FaTE was evaluated by comparing its results to solutions of some small systems for which
the exact solution was known. Additional checks regarding mass balance preservation indicated
that the numerical solutions had the proper properties. Tests with matrices as large as 1000 x
1000 were successful. Thus, LSODE was determined to be an appropriate tool for use in
TRIM.FaTE.
Prior to specific process model evaluations, the user interface of Prototype V was
evaluated to determine possible modifications that could make it easier to use. Changes made to
Prototype V as a result of this evaluation included:
More efficient methods to set up the program and view results (in addition to
HTML results) under the "results" tab of the "setup run/view results" menu
selection. This included HTML output with an extra page of summary biotic
masses and population sizes for each compartment in order to check implications
of the input biota population densities. A page showing how each chemical is
distributed between biotic and abiotic compartments and within the biota was
added;
A new option to send the resulting concentrations in user-selected compartments
within volume elements to Excel plots; and
Changes/additions to the biotic import sheet, including the ability to perform an
"all biotic" and "all abiotic" run. These changes/additions led to automatic
creation of biotic sinks, which improved run time efficiency.
Performance improvements were also made in Prototype V. For example, multiple calls
within the same run were adjusted to decrease run-time (i.e., optimization passes). These
optimization passes included the ability to re-use the previously generated transition matrix
structure (as long as compartments are the same as in the previous run). Prototype V was set up
to generate links automatically (such that fewer output excretion links would occur for animals
and cross-composite container links would not occur which, for example, would prevent roots of
coniferous forest from hooking up to the stem of grasses/herbs). Further, a run option was added
to use a multiplier to preserve the mercury molar mass.
To facilitate the various model comparisons being considered, run options were made
available to omit certain types of links (e.g., soil to soil, soil to water, abiotic to abiotic diffusion,
biotic to abiotic diffusion). One-step disable/re-enable "all biota" was created. Extra result
pages (to make the model more amenable to post-processing) were made available to show
concentrations, average concentrations, and compartment-averaged cumulative fluxes. Because
certain links could be omitted, it was possible to analyze the net flux to the soil from the air and
thus evaluate the change in chemical mass due only to exchange with the atmosphere.
SEPTEMBER 2002 3-5 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.3.2 Overall Evaluation of Versions 1 Through 2.5
TRIM Version 1, which was written in Java, was the first production version of the
computer framework after Prototype V. TRIM Version 1 contained an implementation of
TRIM.FaTE that provided similar capabilities to Prototype V. TRIM Version 2 included
TRIM.FaTE and interfaces to the TRIM Expo inhalation programs APEX and HAPEM. In
versions after TRIM 2.5, the APEX and HAPEM interfaces will instead by provided by the
Multimedia Integrated Modeling System (MIMS). MIMS will provide a means of connecting
and running all the TRIM modules: FaTE, Expo, and Risk.
The implementation of TRIM.FaTE in TRIM Version 1 was evaluated by comparing its
results to those from Prototype V for a variety of scenarios. The first evaluations used simple
systems of abiotic compartments (e.g., air only, air/soil/water). Later evaluations used systems
with hundreds of biotic and abiotic compartments. At the conclusion of the evaluation, the
results of Prototype V and Version 1 were indistinguishable when the models used the same
configuration and input parameters. Like the prototypes, Version 1 also used LSODE to solve
systems of linear ordinary differential equations. Many of the optimizations that were added to
Prototype V during the process evaluation were implemented in Version 1. Optimizations
specific to the Java implementation were also added. In addition to benefitting from specifically
coded optimizations, the performance of the TRIM computer framework will continue to
improve with advances in the Java programming language and the availability of faster computer
hardware.
Initial applications demonstrated that Version 1 had some difficulties running on
Windows 95 and Windows 98 computer systems. These difficulties were primarily a result of
the fact that TRIM.FaTE was developed using Java on Windows NT and was not designed to
address the memory management limitations associated with Windows 95 and Windows 98. On
Windows NT, overall usage of memory by TRIM was reasonable for the size of the application
and stayed relatively constant while the program was running. However, on Windows 95/98 the
management of memory was far less efficient. That is, instead of memory usage remaining fairly
constant, it continued to grow while the program was used until it exceeded the available amount
of RAM, at which time the computer began to slow down significantly until it ultimately halted
execution of TRIM. It was not possible to determine whether the memory problems were due to
Java in particular or Windows 95/98 in general. As a result of these memory issues, Version 1 of
TRIM.FaTE can only be used for small scenarios on Windows 95/98. To run Version 1, at least
a 400 MHz processor with 256 MB of RAM running Windows NT or 2000 is recommended.
The Spring 2000 release of the TRIM computer framework (Version 1.1) included some
restrictions that made long-term (e.g. 30 year) studies difficult to run. The model required that
all time-varying data such as meteorology be read in to memory, and that all output results fit in
memory. This restricted the size and duration of scenarios that could be run with Version 1.1.
These restrictions are removed in TRIM Versions 1.3 and beyond. In these later versions, time-
varying data are read in from files on an as-needed basis and multiple variables can be read in
from the same file. In addition, outputs are written to disk as they are produced instead of being
stored in memory and then exported to disk. The Spring 2001 release of the TRIM computer
framework (Version 2.0) has a sensitivity analysis feature included in it. The results of the
SEPTEMBER 2002 3 -6 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
sensitivity analysis produced by TRIM Version 2.0 will be compared qualitatively and
quantitatively to the results obtained using Prototype V. TRIM Version 2.5, released in July,
2002, includes support for Monte Carlo simulations, and visualization tools (e.g. a food web
viewer and a results viewer that shows compartment concentrations in their volume elements
with other geographic-oriented data).
3.3.3 Algorithm and Compartment Audit
A comprehensive audit of the master TRIM.FaTE library was performed, with follow-up
investigation and resolution of discrepancies encountered. This follow-up has continued into
2002. The purpose of this audit was to verify that the equations, constants, and units in the
master library are accurate and consistent with the TRIM.FaTE documentation, and that the
current set of input parameters are implemented as intended.
3.3.3.1 TRIM.FaTE Libraries
The majority of the information describing chemical transport and transformation in
TRIM.FaTE is contained within files referred to as libraries. A library, which can be viewed by
the user and customized as needed or particular scenarios, is similar to a database in that it
contains data for a number of related objects in a single file. In particular, libraries consist of
properties and their associated values grouped by compartments (e.g., air, surface water, water
column herbivore, white-tailed deer). These properties can be defined as constants, booleans
(i.e., true/false values), or as formulas.
For a typical compartment, these properties include both constants (e.g., depth of a soil
compartment) and formulas (e.g., a function to calculate the wet deposition rate of particles in
air). The algorithms contain the equations that describe (1) how pollutant mass is transported
between compartments and (2) how pollutants are transformed within compartments over time.
TRIM.FaTE is unique among models in that the algorithms that define the transport and
transformation of pollutants in the environment are stored in a data file that can be edited by the
user using the TRIM graphical user interface (GUI). Thus, the user has direct access to these
algorithms, and would not need programming skills to make adjustments/corrections to them.
Examples of the properties associated with transport and transformation algorithms are provided
in Tables 3-1 and 3-2, respectively.
Chemicals and sources are similar to compartments in that they also consist of a set of
constant and formula properties that describe their characteristics. For a more detailed
description of the TRIM computer architecture, refer to the TRIM.FaTE TSD Volume I (U.S.
EPA 2002a).
SEPTEMBER 2002 3 -7 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-1
Example of Transport Algorithm Properties
(Ingestion of Arthropod by Mouse)
Property Name
Category
Chemic al C ate gory
DoesTransformChemical
DoesTransportChemical
Enabled
IsDefaultForCategory
ReceivingCompartmentCategory
SendingCompartmentCategory
TransferF actor
Value
Ingestion
All
False
True
True
True
Mammal/Mouse
Insect/ Arthropod
ReceivingCompartment.PopulationSize * ReceivingCompartment.BW *
TheLink.FractionSpecificcompartmentDiet *
ReceivingCompartment.FractionDietSoilArthropod *
ReceivingCompartment.FoodlngestionRate *
ReceivingCompartment. Chemical. AssimilationEfficiencyFromArthropods /
SendingCompartment. TotalMass
Table 3-2
Example of Transformation Algorithm Properties
(Methylation by Birds)
Property Name
Category
Chemic al C ate gory
DoesTransformChemical
DoesTransportChemical
Enabled
IsDefaultForCategory
ReceivingChemicalName
ReceivingCompartmentCategory
SendingChemicalName
SendingCompartmentCategory
TransferF actor
Value
Transformation
Same
True
False
True
True
Methylmercury
Bird
Divalent Mercury
Bird
SendingCompartment. Chemical. Methyl ationRate
SEPTEMBER 2002
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.3.3.2 Audit Scope
The audit of the TRIM.FaTE library compared the formulas, constants, and units
contained within the compartments and algorithms to the TRIM.FaTE documentation to confirm
that:
All formulas and constants in the library are included in the TRIM.FaTE
documentation (i.e., no omissions);
All formulas and constants in the library are consistent with the documentation
(i.e., no typos or translation errors);
The units used in the library are internally consistent (i.e., each formula using a
particular variable assumed the same units for the variable) and consistent with
the documentation; and
The most recent updates to the formulas and constants are reflected in the library
and in the TRIM.FaTE documentation.
The formulas and their units were compared with the TRIM.FaTE documentation presented in
the 1999 draft TRIM.FaTE TSD Volume II: Description of Chemical Transport and
Transformation Algorithms. The constants were compared with the data tables provided in
Appendix C of the 1999 TRIM.FaTE Status Report.
Note that this audit only reviewed the TRIM.FaTE documentation to confirm formulas,
constants, and units used in the library. It did not include an exhaustive review of the
documentation. In many cases, the TSD Volume II presents detailed derivations of formulas and
alternative algorithms and methods for characterizing compartments and mass transfer between
compartments that were not presented in the library. This audit only reviewed those parts of the
documentation that were required to verify a part of the library. Furthermore, there are several
"built-in" constants (e.g., TT) used by the TRIM.FaTE library that are not actually included in the
library, but in a separate constants file. These constants were not verified as part of this audit.
This audit did not include review of the main TRIM.FaTE code responsible for calling
the algorithms and identifying the sending and receiving compartments to which an algorithm is
to be applied, which is in part specified by the user. This audit focused solely on the content of
the TRIM.FaTE library and did not review the application of the library. A review of this
implementation is the subject of the mercury test case (see Volume II of this report).
3.3.3.3 Audit Methods
TRIM.FaTE is designed to estimate the transport and transformation of chemical mass
between and within both abiotic and biotic media, and a large number of constants and formulas
are required to describe these processes. The version of the TRIM.FaTE library assessed in this
audit contained 177 named algorithms describing the transfer of contaminant mass between
compartments and 47 compartment types (7 abiotic and 40 biotic). Many of the named
SEPTEMBER 2002 3-9 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
algorithms related to food ingestion are repeated many times, differing only with respect to the
identities of the pollutant and the receiving compartment (i.e., the species named as the
consumer). For example, the "Ingestion of Arthropod" algorithm was included 20 times in the
algorithm library to represent four species of animals that consume soil arthropods and five
different pollutants (i.e., 4x5 = 20). The algorithms for transfers between abiotic media can be
repeated to represent different chemicals (e.g., applied to two organic compounds, applied to
three mercury species, or applied to all chemicals). Most algorithms include one or more
formulas, but some simply refer to variables named within one of the compartments. Within
each compartment type, there are a number of different constants and formulas.
The audit began with a review of the chemical transport and transformation algorithms
and their associated properties as compared with Volume II of the 1999 draft TSD (U.S. EPA
1999d). Because many of the formulas presented in the TRIM.FaTE documentation are divided
between the algorithm and compartment sections of the library, however, a systematic audit of
the compartments was also needed.
When comparing formulas in Volume II of the TSD to the library, the units for each
variable described in the TSD and the units assigned to the variable in the library were examined
to ensure that the units were consistent and that all unit conversions (e.g., grams to kilograms)
that might be needed in the library were present. Because the units for the library variables were
not defined for all variables in the "Variable Definition" file, an audit of the units for each
variable was also conducted. After that audit, the audit of the algorithms was completed.
The comparisons were documented and discrepancies explained in some detail. These
comments were then sent to the experts responsible for developing, refining, or implementing
different components (e.g., air, biota, surface water and sediments) of the TRIM.FaTE model for
their review and recommendations on how to address the discrepancies. Their comments and
recommendations were incorporated into the detailed audit documentation, which was then
distributed again to the experts for review. For some of the discrepancies, several rounds of
review were required to establish a solution (i.e., a change in the library, a change in the
documentation, or changes in both). Finally, as some of the recommended library changes were
actually implemented, a few new discrepancies were uncovered. These were also resolved by
the expert(s) familiar with the compartments or transfers at issue. The resolution and its basis
were documented and reflected in the final TSD Volume II (U.S. EPA 2002b).
3.3.3.4 Findings/Results
This audit of the algorithms and compartments resulted in a number of changes being
made to both the library and TRIM.FaTE documentation. A summary of the changes is
presented in Appendix I-A and the TRIM.FaTE Algorithm and Compartment Audit (ICF
Consulting 2002); refer to the audit report for detailed review comments on each algorithm and
compartment.
SEPTEMBER 2002 3-10 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Algorithm Audit
To facilitate the evaluations and discussions, the named algorithms contained in the
library were numbered consecutively from 1 to 177. The chemicals to which each algorithm was
applied (e.g., two organic compounds) was noted, as well as how many different receiving
compartments were represented (e.g., three species of mammals). Thus, if a named algorithm
applied to two chemicals and three species of mammals, there were six instances of that named
algorithm in the algorithm library. The first instance of each named algorithm was evaluated.
The audit of the algorithm library revealed several types of discrepancies between the
TRIM.FaTE library and documentation that required changes to one or the other or both (ICF
Consulting 2002). Several general types of changes are described below.
The majority of the changes required did not change the output of the library.
Draft TSD Volume II indicated that the food ingestion algorithms should include
an assimilation efficiency of the contaminant by the receiving animal
compartment. Assimilation efficiencies were included for fish, but not for birds
and mammals. Assimilation efficiencies for contaminants from general "food"
(i.e., fish, birds, or mammals), arthropods, worms, and plants have been added to
the appropriate ingestion algorithms and compartments (the assimilation
efficiency is a function of both the food type and the consumer species).
Assimilation efficiencies of the contaminant from soil and from water also were
added to those algorithms and to the bird and mammal compartments. As a
default, the values of all assimilation efficiencies have all been set to one (1) in
the animal compartments. As stated in Volume II of the 1999 draft TSD, "if rate
constants for excretion and chemical transformation are determined with respect
to the mass of a contaminant that is taken up in the diet rather than the mass that
is assimilated, the dietary assimilation efficiencies may be ignored." The rate
constants used to-date were determined on the former basis (i.e., mass of
contaminant taken up in the diet, not mass assimilated). However, to allow for
future runs using different rate constants and to match the documentation of the
ingestion algorithms, the assimilation efficiencies were added.
Some of the changes involved changing the algorithm properties to ensure that the
algorithm name described the transfer in the algorithm in the correct direction (or
vise versa). For example, the demethylation algorithms called for the methylation
rate instead of the demethylation rate (Algorithm numbers 26 through 30). The
demethylation algorithms now call for the demethylation rate from the
appropriate compartments.
Some of the changes involved restricting the chemicals to which some of the
algorithms applied. Some algorithms were intended for use with only divalent or
both divalent and elemental mercury, yet were applied to all mercury species,
including methylmercury (e.g., Algorithms 46, 47, 173, 174, 175). The way these
algorithms were restricted to specific chemicals prior to the audit was by setting
rate constants to 0 for those chemicals to which the algorithm should not apply.
SEPTEMBER 2002 3-11 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
To streamline the TRIM.FaTE library and reduce runtime, the new method
removed the algorithms that did not apply.
Some of the algorithms reflected an older library that had not been updated
according to the mass transfer formulas developed for the draft TSD Volume II
(e.g., Algorithms 32, 38, 54, 55, 147). These algorithms have been updated.
Some of the documentation in the draft TSD Volume II reflected older versions of
a mass transfer formula that had not been updated, although the library had been
updated (e.g., Algorithms 40, 46, 47, 149). The final TSD has been updated to be
consistent with the formula.
Some of the mass transfer formulas in the algorithm library apparently had not
been documented in the draft TSD. Very few of the ingestion algorithms were
specified in the documentation; instead, the documentation provided one very
long formula describing all transfers into and out of an animal (e.g., Algorithms
63 to 65 and many of the remaining ingestion algorithms). Other mass transfer
formulas also were missing from the draft TSD (e.g., Algorithm 2, 150, 134). The
mass transfer formulas have been added to the final TSD.
Some of the documentation of mass transfer equations in the draft TSD, Volume
II, were in error through typographical error, omission of variables, or other
mistakes (e.g., TSD Equation 5-36 for Algorithm 148, TSD Equation 7-41 for
Algorithm 155, and TSD Equations 2-39 and 2-40, which incorrectly represented
the single equation that should have been included). These errors and omissions
have been corrected in the final TSD.
A few of the changes related to mistakes in unit conversions in the library (e.g.,
Algorithms 33 and 159).
Wet and dry depositions of particles from air to the surface of plant leaves were
missing from the algorithm library (e.g., new Algorithms 48b, 177b), even though
the mass that would have been deposited to the plant surfaces was removed from
the mass that fell to the soil. Also, the ingestion of particles on the leaf surface by
herbivores consuming the leaves was not represented in the library or in the draft
TSD. The algorithm library and documentation have been updated based on these
findings.
As a consequence of discovering algorithms that were missing altogether from the algorithm
library, we assembled the algorithm names in a table, pairing those that described a mass transfer
in one direction and in the reverse direction. Those results are provided in Appendix I-A. Using
that table, it was possible to identify additional algorithms that perhaps should be included and to
confirm with the model developers that those algorithms were not included in the model for a
reason. The final TSD Volume II has been prepared with the intent to provide an exact
crosswalk between equations in the library and the TSD.
SEPTEMBER 2002 3-12 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Compartment Audit
The audit of the compartment properties section of the TRIM.FaTE library revealed
several types of discrepancies with the TRIM.FaTE documentation that required changes to one
or the other or both (ICF Consulting 2002). Several general types of changes are described
below.
The majority of the changes resulting from the audit of the compartments
involved documenting properties (i.e., formulas and constants) for PAHs that are
contained in the library. The summary of input data included in Appendix C of
the 1999 TRIM.FaTE Status Report had only included values for mercury species.
There were a number of errors identified in the compartments resulting from
improper translation of equations from the documentation into the library. Some
of these mistakes were as simple as incorrect unit conversions impacting a single
property, while others were logic errors impacting a number of properties. These
logic errors were typically introduced by representing a single equation from the
documentation with several simpler equations in the library. All such errors have
been corrected.
As for the algorithm library, we also identified and corrected compartment
properties that had been updated in the documentation but not in the library, as
well as formulas that had been updated in the library but not in the
documentation.
There were a few properties in the compartments that were no longer being used,
having been replaced with newer formulas. We removed these properties to
improve computational efficiency and to prevent future confusion for users
reviewing the library.
Finally, as a result of our review, we identified a few formulas that were incorrect
in both the library and documentation. In these cases, the expert team reviewed
the environmental processes in question and developed new formulas.
Corresponding library changes were then implemented.
3.3.3.5 Audit Summary and Conclusions
The audit of the algorithm and compartment sections of the TRIM.FaTE library identified
some mistakes in both the library and the documentation of the library. Although difficult, the
exercise was extremely important to ensure that the model library is coded correctly and that the
documentation of TRIM.FaTE clearly and accurately represents the formulas in the model.
Moreover, many suggestions for clarification of the mass transfer formulas in TSD Volume II
were provided by the experts. This evaluation strengthened both the TRIM.FaTE model library
and documentation.
SEPTEMBER 2002 3-13 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.4 TESTING OF INDIVIDUAL PROCESS MODELS
Tests were designed and performed to evaluate individual process models, as well as for
approaches and algorithms developed specifically for TRIM.FaTE. Examples of process models
identified for evaluation include the particle/plant leaf algorithms, the soil flux model, and the air
transport algorithms.
When different models are available for the same process (e.g., bioaccumulation in fish),
model-to-model evaluations may be performed at a process model level to test the overall
performance of TRIM.FaTE using different input algorithms. As one example of this, EPA has
evaluated both an equilibrium-based approach and a bioenergetics approach to model
bioaccumulation in fish.
Input data acquisition and the careful evaluation of model inputs are ongoing. To date,
the majority of effort has focused on compiling an initial set of model inputs for a small set of
test chemicals (i.e., phenanthrene, benzo(a)pyrene, various mercury species) and environmental
settings (U.S. EPA 1998e).
3.5 AIR PROCESS MODEL EVALUATION
For the air process model evaluation, the TRIM.FaTE air transport module was compared
to the Urban Airshed Model (UAM) and the Industrial Source Complex (ISC) Model. These
comparisons are presented in Sections 3.5.1 and 3.5.2, respectively.
3.5.1 Comparison with the Urban Airshed Model
This section presents the approach and results of runs to compare the TRIM.FaTE air
transport module with the UAM transport module.
3.5.1.1 Approach and Model Setup
As part of the evaluation of the air transport module of TRIM.FaTE, a simple "air-only"
TRIM.FaTE run was set up and quantitatively compared with the transport module of the UAM,
an existing grid model used by EPA and others to estimate dispersion of air pollutants.3 The
source included in all runs emitted gaseous mercury at a constant rate of 1.514E-2 g/s. Mercury
emissions were 100 percent elemental, and no chemical transformation/degradation was
included. No background mercury concentrations were included and only dispersion (i.e., no
deposition) was considered.
Two grid layouts comprised of equal-sized square parcels were selected for this
comparison (modeled region 27.76 km on a side; total area approximately 770 km2):
3 As this comparison was based on a simple scenario (i.e., simple terrain, no chemical transformation, etc.),
the results are not necessarily indicative of all of the differences between the air module of TRIM.FaTE and UAM.
It is possible that more complex scenarios would lead to more marked differences in model output.
SEPTEMBER 2002 3-14 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
5x5 (25 square parcels)
9 x 9 (81 square parcels)
Figure 3-2 presents configurations for both of these grid layouts. For both grids, two vertical
layers were included. Layer 1 was defined as the lower layer, extending vertically from ground
level to the mixing height. Layer 2 extended from the mixing height to 1,864 m above ground
level (or 1,000 m above the highest mixing height for the meteorology data used in this analysis).
Scenario descriptions for each model and key model characteristics are described below.
One UAM run was completed for each grid layout. Surface meteorology data for UAM
were obtained from the National Weather Service (NWS) Hourly U.S. Weather Observation
(HUSWO) database; values for January 1990 for the chlor-alkali test site were used. Upper air
meteorology data were obtained from the NWS Radiosonde database for the same site and time
period. Before a run is initiated, UAM preprocesses these two sets of data to account for vertical
heterogeneity in wind speed and wind direction values. This preprocessing results in different
input data sets for the lower and upper layers (i.e., the layer 1 data are not identical to the
HUSWO data, and the layer 2 data are not identical to the radiosonde data); however, as a result
they best represent the average winds for both layers. The vertical wind speed (i.e., between
layers 1 and 2) is estimated internally by UAM based on mass continuity and changes in
compartment size that result from the changing mixing height. Average hourly concentrations
were calculated by UAM for 1 month for each compartment of each layer. These results were
averaged to obtain the monthly average for each compartment.
Figure 3-2
Grid Configurations
(source parcels are shaded)
5x5
9x9
1
6
11
16
21
2
7
12
17
22
3
8
13
18
23
4
9
14
19
24
5
10
15
20
25
i
10
19
28
37
46
55
64
73
2
11
20
29
38
47
56
65
74
3
12
21
30
39
48
57
66
75
4
13
22
31
40
49
58
67
76
5
14
23
32
41
50
59
68
77
6
15
24
33
42
51
60
69
78
7
16
25
34
43
52
61
70
79
8
17
26
35
44
53
62
71
80
9
18
27
36
45
54
63
72
81
For TRIM.FaTE, a set of four runs was completed for each grid layout resulting in eight
total TRIM.FaTE runs. These four runs were obtained by using two sets of meteorology data
and two sets of vertical wind speed data as follows:
SEPTEMBER 2002
3-15
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Horizontal meteorology data sets:
(1) NWS HUSWO database, values for January 1990 for the chlor-alkali test site
(referred to in the context of these analyses as the "original meteorology data").
(2) Data for January 1990 for the chlor-alkali test site after preprocessing by UAM to
account for vertical heterogeneity (referred to below as "UAM meteorology
data").
Vertical wind speed data sets:
(1) Vertical wind speed equal to zero (i.e., no transfer from lower layer to upper
layer).
(2) Vertical wind speed calculated by UAM (extracted from UAM and then input into
TRIM.FaTE).
These four sets of TRIM.FaTE runs were completed to evaluate the effects of the various
meteorology data and vertical wind speed input combinations. However, the TRIM.FaTE run
completed using UAM meteorology data and vertical wind speed from UAM was considered to
be the most similar to UAM with regard to input data. Therefore, results from this run were used
for the model-to-model comparisons.
3.5.1.2 Results
Several key UAM and TRIM.FaTE model characteristics should be considered when
comparing the results of these two models:
UAM calculates mass transport in the horizontal and vertical directions via three
mechanisms: advection, dispersion, and diffusion.
TRJM.FaTE explicitly models advective transport, only models diffusion by
assuming equal mixing of pollutant concentrations within a compartment, and
does not model dispersion.
TRJM.FaTE does not have a vertical sink and thus does not lose system mass
across the vertical boundaries. For the UAM runs in this analysis, vertical
transport out of the top of the modeled domain was not allowed.
Effect of Meteorological Data and Vertical Velocity Input Data for TRIM.FaTE
Overall, concentration patterns were similar for all TRIM.FaTE runs for each grid layout.
The following effects were observed:
SEPTEMBER 2002 3-16 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
As expected, a vertical velocity of 0 resulted in higher concentrations for layer 1
because no mass was transported to layer 2. Layer 1 concentrations were about
1.5 to 3 times larger relative to runs with a vertical velocity.4
Using the "original meteorology data" resulted in slightly higher concentrations
for some compartments in layer 1 but did not appear to introduce major
differences into the overall pattern of concentrations.
The general patterns for a given model were similar for both the 5x5 grid and the 9x9 grid,
with the denser grid providing greater resolution. Comparison of the different grid results did
not appear to provide any additional information regarding the model comparison. Observations
listed below for layer 1 and layer 2 are generalized (i.e., not specific to one grid layout);
quantitative results provided with this summary reflect the results for the 5 x 5 grid.
Layer 1 Comparison (UAMvs. TRIM.FaTE with UAMMeteorology Data)
TRIM.FaTE and UAM layer 1 results followed the same basic pattern across all
compartments (see Figure 3-3). Table 3-3 presents the percent differences between the average
results from the two models. On average, the concentrations estimated using UAM were
approximately 22 percent higher than those estimated using TRIM.FaTE. For the southwest
quadrant of the grid, the TRIM.FaTE concentration was noticeably less than the UAM
concentration. For compartments in the other three quadrants (NW, NE, and SE), the
TRIM.FaTE concentrations were generally equal to or higher than the UAM concentrations (see
Figure 3-4). Mass bias toward the axes of the grid (i.e., a "pipeline effect") was observed in both
TRIM.FaTE and UAM layer 1 results; however, the bias was slightly less pronounced in the
UAM results. For more details regarding this effect, see Section 4.2.
4 To verify that the vertical transport algorithms in TRIM.FaTE were operating properly, an additional run
was made for each grid layout with the vertical velocity set to zero. We compared the results of these scenarios to
the results of similar scenarios with only one vertical layer. As expected the results for the one vertical layer
scenarios were identical to the results for the two vertical layer/zero vertical velocity scenarios.
SEPTEMBER 2002 3-17 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.50E-03
3.00E-03
2.50E-03
2.00E-03
1.50E-03
o
o
1.00E-03
5.00E-04
O.OOE+00
Figure 3-3
TRIM.FaTE and UAM Concentrations, Layer 1,5x5 Grid
TRIM vs. UAM, 5x5 Layer 1
UAM layer 1
TRIM layerl, UAM met, vWS=UAM
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
1234567
Table 3-3
Percent Difference ((UAM-TRIM.FaTE)/UAM), Layer 1,5x5 Grid
Parcel
1
2
3
4
5
% Diff
38%
55%
18%
-6.5%
-24%
Parcel
6
7
8
9
10
% Diff
29%
57%
35%
13%
-18%
Parcel
11
12
13
14
15
% Diff
3.9%
34%
19%
-10%
-52%
Parcel
16
17
18
19
20
% Diff
-7.1%
5.2%
-12%
-17%
-43%
Parcel
21
22
23
24
25
% Diff
-43%
-36%
-50%
-48%
-74%
SEPTEMBER 2002
3-18
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Figure 3-4
Quadrant Averages, Layer 1,5x5 Grid
Quadrant and overall averages, 5x5 grid, layer 1
Source parcel not included in quadrant averages
4.00E-04
3.50E-04
3.00E-04
2.50E-04
B 2.00E-04
1.50E-04
1 .OOE-04
5.00E-05
O.OOE+00
UAM data
DTRIM UAM met, vWS = UAM
NW
SW
SE
overall avg. (no overall avg. (with
source parcel) source parcel)
Overall, TRIM.FaTE and UAM seemed to agree more closely in the quadrants that are
predominantly upwind (NW) and downwind (SE) of the prevailing winds. The largest
compartment concentration differences between TRIM.FaTE and UAM were primarily located
in the compartments comprising the boundary of the grid; at these locations, TRIM.FaTE
concentrations were larger than UAM results. This effect may have resulted in part from
differences in how the two models treat mass flow beyond the grid boundaries. The UAM has a
set of boundary cells encompassing the entire perimeter of the modeling domain. Mass from the
modeling domain can flow into these cells or mass from these cells can flow back into the
modeling domain, depending on the direction of the wind. TRIM.FaTE, on the other hand, was
run without boundary cells and thus mass leaving the modeling domain was deposited in sinks
and was not available to be transported back into the modeling domain. The impact of adding
boundary cells to a TRIM.FaTE run is discussed in Section 4.2.1.3.
If the compartment results for the 5x5 grid for each model are ranked by concentration,
the rankings are close (within two ranks) for most of the compartments and match for 8 of the 25
grid compartments. The largest rank difference is six (e.g., 12th highest concentration vs. 18th),
calculated for two compartments.
SEPTEMBER 2002
3-19
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
UAM results were regressed on TRIM.FaTE results to obtain a rough estimate of
correlation. For the 5x5 grid, the R-square value for this regression was 0.9697. For the 9x9
grid, the R-square value was 0.9643. Thus, increasing grid resolution did not appear to result in
a higher correlation between the two models.
Layer 2 Comparison (UAM vs. TRIM.FaTE with UAM Meteorology Data)
For layer 2, the concentration patterns for the two models were relatively similar overall
but there were noticeable differences for some compartments. UAM results were larger than
TRIM.FaTE results for all compartments but one (see Figure 3-5 and Table 3-4). Concentration
gradients across these compartments were generally somewhat different for UAM and
TRIM.FaTE, especially in the eastern half of the grid. Regression of UAM results on
TRIM.FaTE resulted in an R-square value of 0.8219, which is smaller than the corresponding
value obtained for layer 1. Note that for both models, the level 2 results are roughly an order of
magnitude lower than the level 1 results.
Figure 3-5
TREVLFaTE and UAM Concentrations, Layer 2,5x5 Grid
Layer 2, 5x5: All parcels, UAM vs TRIM
4.50E-05
TRIM Iayer2, UAM met, vWS=UAM
UAM layer 2
O.OOE+00
1234567
SEPTEMBER 2002
3-20
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-4
Percent Difference ((UAM-TRIM.FaTE)/UAM), Layer 2, 5x5 Grid
Parcel
1
2
3
4
5
% Diff
99.9%
99.0%
-1.3%
38%
63%
Parcel
6
7
8
9
10
% Diff
99.2%
97%
22%
52%
68%
Parcel
11
12
13
14
15
% Diff
90%
86%
33%
59%
68%
Parcel
16
17
18
19
20
% Diff
73%
65%
46%
64%
70%
Parcel
21
22
23
24
25
% Diff
45%
41%
38%
60%
66%
The observed differences between TRIM.FaTE and UAM for layer 2 could be the result
of lower net mass transport to layer 2 in TRIM.FaTE compared to UAM. This effect may be due
to differences in the advection algorithms. In addition, quicker mass transport out of the grid for
layer 2 could occur in TRIM.FaTE (i.e., mass more quickly advected to the sinks). The
differences could also be due to UAM's inclusion of both vertical dispersion and diffusion.
TRIM.FaTE does not include vertical dispersion and only includes vertical diffusion in that mass
transported vertically by advection is assumed to spread evenly throughout the receiving
compartment.
Summary
For the regular grids examined in these analyses, UAM and TRIM.FaTE predicted
similar average monthly air concentrations across the entire grid for layer 1. Patterns for layer 2
were less similar, but concentrations for UAM and TRIM.FaTE were generally within an order
of magnitude and often within a factor of 2 or 3 of each other (UAM values were consistently
higher). It is important to note that these results were obtained using identical meteorology data
that had been preprocessed by UAM. Additionally, it should be stressed that this evaluation was
carried out with a focus on using identical input data for UAM and TRIM.FaTE (where possible)
in order to compare model operation. Also, these results may be different for other seasons, as
this evaluation was done in January when advection strongly dominates over diffusion. For
other locations and times diffusion may be as important as advection and the differences between
the models may be different. Overall, UAM and the air component of TRIM.FaTE seem to
produce comparable results for simple, single-vertical-layer scenarios. However, based on the
differences between the estimated concentrations in layer 2, TRIM.FaTE may not be appropriate
for modeling pollutants that are transported significant distances by aloft winds and dispersion
processes.
3.5.2 Comparison with the Industrial Source Complex Model
TRIM.FaTE also was compared to an EPA dispersion model widely used in regulatory
applications, the Industrial Source Complex Model (Short Term), version 3 (ISCST-3). The fate
and transport algorithms in ISC are based on Gaussian dispersion equations that are solved for a
given set of temporal and spatial circumstances. ISC is not a mass-balanced model, whereas
SEPTEMBER 2002
3-21
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
TRIM.FaTE is mass-balanced, as are grid models in general. Thus, this evaluation was carried
out to obtain a rough comparison with an air model currently used by EPA for regulatory
purposes. As ISC is used as part of EPA's Multimedia, Multi-pathway, Multi-receptor Exposure
and Risk Assessment (3MRA) model (http://www.epa.gov/athens/research/projects/3mra/), the
model-model comparison aspect of the mercury test case (Volume II) will include some
comparison of ISC output to TRIM.FaTE's air output.
3.6 EVALUATION OF MERCURY SPECIATION IN AIR AND SOIL
This section evaluates the mercury parameter values (e.g., reduction and oxidation rates
in air and soil) that affect mercury speciation in air and soil. Specifically, the parameter values
chosen for relevant air and soil processes are presented, and simulations are performed to
determine whether predicted speciation in soil and air is consistent with observed background
levels.
Section 3.6.1 describes the properties and behavior of mercury relevant to the air and soil
processes model. Section 3.6.2 presents the methods of the evaluation, including the chosen
parameter values. The results and conclusions of the evaluation are presented in Section 3.6.3
and Section 3.6.4, respectively.
3.6.1 Mercury Species and Transformations in Air and Soil
Mercury is emitted to the atmosphere both in elemental (Hg°) and divalent (Hg2+or
Hg(II)) forms and is found in both species in the atmosphere. Once emitted into the
environment, these species can be transformed to another species or deposited to other
environmental media.
Background concentrations of mercury in air are 1 to 3 ng/m3 in the Southern
Hemisphere and 2 to 4 ng/m3 in Northern Hemisphere (Slemr and Langer cited in Lin and
Pehkonen 1999). More than 90 percent of atmospheric mercury is elemental, and some
estimates are as high as 95 to 97 percent. The predominant oxidation reactions transforming
gaseous elemental mercury to divalent mercury are with O3 and H2O2. Because of such
transformations, the half-life of elemental mercury in the atmosphere is between 0.5 and 2 years
(U.S. EPA 1997). Although significant transformation of elemental to divalent mercury may
occur in aqueous media, all transformation of elemental to divalent mercury in TRIM.FaTE is
included in the general atmospheric transformation rate.
Most divalent mercury released to air is removed from the environment on the local or
regional scale (i.e., relatively close to the source). Divalent mercury is extremely reactive and
may be transformed back to the elemental form or other species. It may be gaseous, but often is
associated with particles. Some of the gaseous forms of divalent mercury in the atmosphere may
originate from evaporation from clouds. Divalent mercury often is found as HgCl2, Hg(OH)2, or
compounds of other halides, which collectively are referred to as reactive gaseous mercury
(RGM). All these forms are very water soluble, often 105 times more soluble than elemental
mercury (Lindberg and Stratton 1998). RGM can be rapidly reduced back to Hg°, at least in the
aqueous phase (Pleijel and Munthe 1995a and 1995b). All of these transformations make it
SEPTEMBER 2002 3-22 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
difficult to quantify source distribution (Lindberg and Stratton 1998). Because of its reactivity,
ROM is thought to exist in ambient air at very low concentrations. There are consistent dial and
diurnal cycles of ROM, with the peak concentrations occurring at midday and sharply dropping
off at night (Lindberg and Stratton 1998).
In ambient air, concentrations of particulate mercury are on the order of picograms per
cubic meter, but concentrations as high as a few nanograms per cubic meter have been measured
in urban/industrial areas (Lu and Schroeder 1999). Although particulate mercury usually is a
very small fraction of total atmospheric mercury, it plays an important role in the mercury
deposition patterns (Lu and Schroeder 1999). Munthe (1993) found that a large fraction of
mercury in rainwater was bound to particles, indicating the importance of understanding the
dynamics involved with the particulate phase.
Mercury deposition rates generally are much higher for the divalent species than for the
elemental species. Because divalent mercury is more water soluble and more likely to partition
to particles than elemental mercury, what little elemental mercury is transformed to divalent
mercury is quickly deposited to soil, particularly during rain events. In addition, atmospheric
divalent mercury near moist soil will partition out of the atmosphere and into the moist soil (U.S.
EPA 1997).
Once deposited to surface soil, each of the mercury species can diffuse or be advected to
lower soil zones, react into another species, or diffuse back into the air. Mercury speciation in
soil is highly variable, and it is difficult. However, mercury in soil generally would be expected
to be predominantly in the divalent form (e.g., Biester and Scholz 1997).
There is limited information on how deep mercury is found in soils, and even less
information is available on mercury speciation in subsurface soil. One study (DiGiulio and Ryan
1987) on peat soils found little change in mercury at soil depths up to one meter. Another study
(EPRI 1998) found similar mercury concentrations in the 0- to 2-cm and 5- to 10-cm horizons.
3.6.2 Methods
Information presented in Section 3.6.1, as well as other available information on the
environmental chemistry of mercury, was used to select preliminary default values for
parameters included in the TRIM.FaTE air and soil process models. Model simulations using a
simple hypothetical scenario then were performed to evaluate whether the preliminary
parameter values produced realistic predictions of mercury speciation in soil. Specifically, the
goal of the evaluation was to identify a set of parameter values that resulted in 90 to 98 percent
divalent mercury in the surface soil compartment. An iterative series of simulations was
performed with varied parameter values until this goal was achieved. Because mercury
deposition is substantially affected by precipitation, all simulations were run both with and
without precipitation included in the input meteorological data.
The model configuration for this evaluation consisted of air, surface soil, root zone soil,
and vadose zone soil compartments. Mercury was introduced into the system via a constant
background air concentration blowing into the system (i.e., no discrete source was included).
SEPTEMBER 2002 3-23 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Mercury speciation in background air was assumed to be 98 percent elemental and 2 percent
divalent. Model runs were completed for a constant precipitation case and a no precipitation
case.
3.6.3 Results
The final parameter values are presented in Table 3-5. For the scenario used in this
evaluation, the final parameter values result in 98.2 percent divalent mercury in surface soil with
and without precipitation. These results are slightly higher than the goal of this evaluation,
which was 90 to 98 percent divalent mercury in surface soil.
Table 3-5
Concluding Values of Particular Air and Soil Parameters
Parameter Name
Default Value
Comments
Percent of divalent mercury
in background air
2%
Based on mercury speciation observed
in various studies.
Oxidation rate in air
0.0038/day
Corresponds to a half-life of 6 months.
Reduction rate in air
0/day
None.
Oxidation rate in soil
0/day
Although transformation rates in soil
are unknown, they are thought to be
small. Thus, the default value was set
to zero.
Reduction rate in soil
0/day
Although transformation rates in soil
are unknown, they are thought to be
small. Thus, the default value was set
to zero.
Damkohler depth in soil
8 cm
The Damkohler depth is based on the
expected steady state soil concentration
profile and is used to help define the
transfer factors between the various
soil layers. The default value was
based on constant measurements down
to this depth. However, it may be too
small.
Predicted mercury concentrations in the three soil compartments are presented in Table
3-6, and mercury speciation results are presented in Table 3-7. When simulations with the final
parameters were run without precipitation, no divalent mercury was predicted in the root zone
and vadose zone. This finding is reasonable, because without precipitation the divalent mercury
deposited from the air would remain tightly bound to soil particles in the surface soil
compartment. The lack of moisture, as well as the relatively rapid transformation rates from
SEPTEMBER 2002
3-24
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
divalent to methyl and elemental mercury, would make diffusion of divalent mercury from
surface soil to deeper soil compartments unlikely. Under conditions of constant average
precipitation, mercury speciation in the root zone soil appears to be reasonable (see Table 3-7).
Mercury concentrations in surface soil are more important to the overall model results
than mercury concentrations in root and vadose zone soils, because human and ecological risks
are much more likely to result from exposure to surface soil than exposure to subsurface soil. In
this evaluation, very little elemental mercury was predicted in surface soil either with or without
precipitation. This result is reasonably consistent with field data (Biester and Scholz 1997),
which indicate that one to two percent elemental mercury generally would be expected in surface
soil. However, it is possible that the field data are based on experimental samples that included
an amount of root zone soil where considerable elemental mercury is found.
3.6.4 Conclusions
This evaluation developed default air and soil process parameter values for mercury. For
the scenarios used in this evaluation, TRIM.FaTE simulations with these parameters resulted in
predicted mercury concentrations and speciation in surface and subsurface soils that are
consistent with field data and other available information on the fate of mercury in soils. No
refinements to the air or soil process models were made based on this evaluation.
Table 3-6
Predicted Total Mercury Concentrations in Soil Compartments,
With and Without Precipitation
Soil Compartment
Surface
Root Zone
Vadose Zone
Mercury Concentration (ng/g bulk)
No Precipitation
23.3
0.0017
0.0019
Constant Average
Precipitation (0.001 m/day)
42.6
0.12
0.03
SEPTEMBER 2002
3-25
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-7
Predicted Mercury Speciation in Soil Compartments, With and Without Precipitation
Soil
Compartment
Surface
Root Zone
Vadose Zone
Mercury
Species
Elemental
Divalent
Methyl
Elemental
Divalent
Methyl
Elemental
Divalent
Methyl
Percent of Total Mercury
No Precipitation
0.0%
98.2%
1.8%
100.0%
0.0%
0.0%
100.0%
0.0%
0.0%
Constant Average
Precipitation
(0.001 m/day)
0.1%
98.2%
1.8%
24.6%
74.1%
1.3%
99.7%
0.3%
0.0%
3.7 SEDIMENT AND SURFACE WATER
The sediment module in TRIM.FaTE is treated as a well-mixed compartment of specified
depth that exchanges mass (by both physical and chemical processes) with the overlying water.
The compartment is composed of water and solids. Chemical fate processes include diffusive
and advective exchange in both directions across the water-sediment interface and loss from the
sediment due to "burial." Burial is the process of freshly deposited material accumulating at the
surface and preventing the deeper layer of sediment from interacting with the overlying water.
The structure provides a well-mixed cap over a sink. Chemical transformation is also included
in the sediment compartment.
The current model structure does not specifically include bioturbation, although it is
implicitly included by specifying the depth of the "surface mixed sediment layer"
(Schwarzenbach et al. 1993). Bioturbation may lead to increased mixing of both sorbed and
dissolved contaminants in the sediment, but not enough is currently known about the process to
include it as a separate input to the model at this point. Several other processes (e.g., ground
water infiltration, nepheloid layer interactions, relationships or correlation among sedimentation,
erosion, and water current) are not explicitly modeled in the sediment module as currently
implemented in TRIM.FaTE. Users are advised to evaluate whether the lack of these processes
limits the utility of TRIM.FaTE to their particular applications.
SEPTEMBER 2002
3-26
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
For mercury species, the algorithm for air/surface water diffusion/volatilization is
equivalent to the "Volatilization Option 4" that is used in the Water Quality Analysis Simulation
Program (WASP) (Ambrose et al. 1995). Sediment/surface water diffusion is modeled using the
same two-resistance model used in CalTOX (CalTOX 1993). The TRIM.FaTE TSD Volume II
(Chapter 4) provides more detailed background information on the sediment and surface water
algorithms included in TRIM.FaTE.
3.7.1 Purpose of Evaluations
Initial model runs resulted in predicted mercury speciations in surface water that were
inconsistent with the literature. In particular, the fraction that was elemental mercury was on the
order of 75 percent; by comparison, the results obtained in the 1997 Mercury Study Report to
Congress indicated that the fraction should be less than 10 percent (U.S. EPA 1997). Thus, a
more in-depth evaluation of the surface water and sediment components of the currently
implemented algorithms was conducted. The two primary focus areas of this evaluation were:
(1) identification of modifications to mercury-dependent parameters in order to achieve
predicted concentrations and speciation consistent with that achieved in the 1997 Mercury Study
Report to Congress, and (2) evaluation of the general surface water and sediment dynamics.
3.7.2 Model Setup and Assumptions
The basic modeling configuration for the sediment and surface water evaluations was a
small rectangular watershed adjacent to a surface water body (depth 3 m), with a single air
compartment over the soil and water. A single sediment layer (depth 0.02 m) was located below
the surface water. Both a flush sink (based on a flush rate of 0.5/yr) and a sediment burial sink
were included in this scenario. The burial rate is calculated from the resuspension velocity,
deposition velocity, and suspended sediment concentration in the water body so as to result in
zero net deposition of sediment to the sediment bed. Note that the outputs of interest in this run
were the predicted steady-state concentrations in surface water and sediment. The wind speed
was constant for a given run; for various runs, wind speeds were set to 2 m/s, 4 m/s, and 6 m/s.
In order to simplify the focus of this evaluation, the concentrations were fixed in the soil
and air compartments for the model runs at approximate "background" values and speciations.
In air, a concentration of 1 ng/m3 (98 percent elemental, 2 percent divalent) was assumed. The
air concentration is the midpoint reported for rural areas in the 1997 Mercury Study Report to
Congress. A total mercury soil concentration of 102 ng/g was assumed, with 98 percent
considered to be divalent, 2 percent considered to be methylmercury (to be consistent with the
Mercury Study Report to Congress) and 0.01 ng/g considered to be elemental. The value
assumed for elemental mercury was based on the assumption of only a small fraction of total
mercury present in the elemental form. A precise estimate of what this fraction should be was
considered unnecessary for the current effort.
Erosion and runoff were calculated based on an input areal soil erosion rate of
0.000289 kg/m2/day, from an average for soils in the test case state, and a runoff rate of
0.001009 m3[water]/m2[area]-day.
SEPTEMBER 2002 3-27 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.7.3 Results
Table 3-8 summarizes the changes made to default input parameters and modifications to
algorithms themselves as a result of this evaluation. The modeling results obtained using these
changes are summarized in the subsequent table (Table 3-8).
The predicted concentrations are within the range of measured values as indicated in the
1997 Mercury Study Report to Congress. Similarly, the predicted speciations of divalent and
methylmercury are also consistent with those obtained in the Mercury Study Report to Congress
(see Table 3-9). There are few data available regarding the fraction that is elemental mercury,
but the values here are consistent with the result obtained in the 1997 Mercury Study Report to
Congress using the IEM-2M model.
3.7.4 Conclusion
By making appropriate changes to various inputs and algorithms (including fixing a units
conversion error in one algorithm), realistic concentrations of elemental, divalent, and
methylmercury in surface water and sediment were obtained using fixed values for soil and air
concentrations that are typical of uncontaminated areas. It was also determined in this
evaluation that the wind speed assumed over the water body is an influential factor in the
predicted elemental mercury concentration in surface water; indeed, the predicted elemental
mercury concentration in the surface water is essentially inversely proportional to the wind speed
assumed. It is noted, however, that usually very little of the mercury in surface water is present
as elemental mercury (U.S. EPA 1997).
3.8 EVALUATION OF THE TRIM.FaTE PLANT MODULE
Two evaluation phases have been completed for the plant module in TRIM.FaTE:
A compositional audit, where the overall plant model was evaluated against the
original conceptual design; and
An algorithm audit, where the transfer factors (T-factors) for each of the
interfacial transfer processes were evaluated independently and in concert with
the other equations in the plant module.
3.8.1 Compositional Audit
The evaluation of the conceptual model described in Section 2 of this report is extended
here to provide a more detailed look at the plant module. This section presents results of a
compositional audit where the original conceptual design of the plant module is reviewed and the
code in Prototype V is evaluated to verify that each component of the conceptual model is
included.
SEPTEMBER 2002 3-28 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-8
Changes Made to Sediment and Surface Water Values as a Result of Sediment/Surface Water Runs
Term Changed
Dynamic Water
Viscosity; calculated
property of surface water
Default Suspended
Sediment Settling
Velocity; input property
of surface water
Default Suspended
Sediment Concentration;
input property of sediment
Default Sediment
Resuspension Velocity;
property of sediment
phi (Sediment Porosity);
property of sediment
Units
kg/m-s
m/day
kg [sediment]/m3 [water
column]
m/day (note that this is
the velocity of the moving
solids, not a volumetric
velocity)
m3 [pore space]/m3
[compartment]
Change Made
Added additional
units conversion term
to equation
Changed default value
to 0.5 m/day from 13
m/day
Changed default value
to 0.015 from 0.8
kg/m3
Changed default value
to 6.28E-6 m/day
from 1 .37E-05 m/day
Changed default value
to 0.7 from 0.2
Comment
The "dynamic water viscosity" was an intermediate term that contained a missing units
conversion term. The equation used is from IEM (and EXAMS and WASP)
water viscosity = 10A(-3.0233 + 13017(998. 333+8. 155*(Domain.Temperature_C -20) +
0.00585*(Domain.Temperature_C-20)A2.0))
The units in IEM (and EXAMS) are g/cm-sec. In TRIM, it is being treated as kg/m-sec.
The conversion factor was originally input as 1 but should be (converting from g/cm-s to
kg/m-s) 0.1 [ = (100 cm/m) * (1 kg / 1000 g)].
Previous water viscosity = 10A(-3.0233 + 1301/(998.333+8.155*(Domain.Temperature_C
-20) + 0.00585*(Domain.Temperature_C -20)A2.0))
New water viscosity = Constants. Convert g per cm to kg per m* 10A(-3.0233 +
1301/(998.333+8.155*(Domain.Temperature C -20) + 0.00585*(Domain.Temperature C
-20)A2.0))
where "Convert g per cm to kg per m" is a new constant = 0.1.
The deposition velocity was originally set to 13 m/d (this is a default value from CalTOX
for California). Typical values in the literature are around 0.5 m/d depending on particle
size and wind/turbulence. Note: Subsequent evaluation resulted in further change to this
value to match that from U.S. EPA 1997 (2m/day).
The suspended sediment concentration was originally set to 0.8 kg/m3 which converts to
800 mg/L (default value from CalTOX). Typical values for lakes and streams are around
15 mg/L. Much lower values may be relevant for many lakes.
Information on resuspension rates was unavailable, but using the input of solid from soil
(erosion), neglecting dry deposition from air and assuming the mass of particles flowing
into and out of the pond are about equal, the accumulation rate in the pond was estimated
to be around 0.0026 kg(solid)/m2/d, which was consistent with reported values in the
literature. With this value, the resuspension velocity was estimated to be 6.28E-6 m/d
with a benthic solids concentration of 780 kg/m3 (based on 30 percent solids in bulk
sediment). A value of 0.65 is used in U.S. EPA 1997.
Based on professional judgment under the assumption that the fraction of sediment that is
water will be much larger than 0.2.
SEPTEMBER 2002
3-29
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Term Changed
Benthic Solids
Concentration of
Sediment; property of
sediment
Default Kd in Soil,
elemental mercury;
property of soil
Kd in surface water,
elemental mercury;
property of surface water
Units
kg[sediment]/m3
[compartment]
ml/g
ml/g
Change Made
Changed to a
calculated parameter
instead of being an
input parameter
Changed default value
to 10 from 0.1 ml/g
Changed default value
to 10 from 0.1 ml/g
Comment
The benthic solids concentration is a function of the sediment porosity and the sediment
density; the previous default input was actually calculated in this way outside of the
model, and hence was consistent with the default porosity and sediment density.
However, it would not remain consistent if either had changed. The new equation
implemented is:
Benthic Solids Concentration = Compartment. rho * (1 - Compartment. phi)
where Compartment. rho is the sediment density (kg[sediment]/m3[sediment]), and
Compartment. phi is the porosity of the sediment compartment.
The previous value was based on previous test runs performed. Evaluation runs indicated
that an increase to the value was appropriate, as the lower value resulted in a substantial
flux of elemental mercury to the water body. The value used in the 1997 Mercury Study
Report to Congress (1,000) has also been used subsequently.
This value was changed in order to be consistent with change made in surface soil; runs
indicated that this did not impact the results in surface water significantly. The value
used in the 1991 Mercury Study Report to Congress (1,000) has also been used
subsequently.
SEPTEMBER 2002
5 ^1 C\
3-30
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-9
Summary of Predicted Steady-State Concentrations in Surface Water and Sediment
Species
Elemental
Divalent
Methyl
Surface Water Concentration
(ng/L)
0.45
5.38
0.11
Sediment Concentration, ng/g (ng/g dry
weight; see note)
0.33 (0.6)
116(219)
0.27 (0.5)
Notes:
Assumed fixed concentrations in air (1 ng/m3 total mercury, 98 percent elemental and 2 percent divalent). Assumed mercury concentration in soil
of 102 ng/g, with 98 percent divalent, 2 percent methylmercury, and a O.Olpercent elemental mercury. Assumed constant wind speed over surface
water of 4m/s.
Sediment concentration is calculated as (mass of sorbed chemical in sediment compartment)/(mass of dry sediment in sediment compartment).
For each species, over 99 percent is sorbed in the sediment compartment. The mass of dry sediment was calculated as :
Benthic Solids Concentration (kg[sediment]/m3[compartment]) * Compartment Volume (m3[compartment])
The benthic solids concentration was calculated as:
Benthic Solids Concentration (kg[sediment]/m3[compartment]) = Sediment Density (kg[sediment]/m3[sediment]) * (1 - Porosity)
where the default sediment density of 2,600 kg/m3 and porosity of 0.7 were used in the run.
3.8.1.1 Conceptual Design of the Plant Module
The TRIM.FaTE plant module includes four homogeneous compartment types: root,
stem, leaf and particles-on-leaf. These compartment types interact, as illustrated in Figure 3-6,
with three environmental compartment types: air, surface soil, and root-zone soil. Vegetation
also interacts with various animal compartment types through food-chain transfers but these are
described and evaluated elsewhere (see Sections 3.9 and 4.3). The linked structure and exchange
processes for the vegetation compartments are summarized in Table 3-10 along with descriptions
of each transfer factor.
The mass transfer processes (gains and losses) in the plant module include both diffusion
and advection. Mass can also be gained or lost through chemical reaction or metabolism in each
compartment. Although the compartments are fully coupled, some of the interfaces are limited
to unidirectional transfers (litter fall, soil-to-stem transfer). In addition, several of the transfer
processes are toggled (on/off) during a simulation depending on rain (wet vs. dry deposition),
season (litter fall to soil over a predefined period) or time of day (temporal response of stomata).
During rain events, the transfer from air to vegetation occurs by rain scavenging of both
gas-phase chemical and particle-bound chemical followed by interception and retention of the
advecting phase (raindrops) by the leaf. Particles and gases are also scavenged by vegetation
through dry deposition processes when it is not raining. Particle-bound chemical is washed from
the particle-on-leaf compartment to surface soil during rain and is blown into the air when it is
not raining.
SEPTEMBER 2002 3-31 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Figure 3-6
Schematic of the Plant Module
Air
Leaf particle
V
Leaf
TL-
T
J-
Stem
gg(during litter fall)
Surface soil
T
1 Sr -> St
Root soil
Advection
Diffusion
Root
Diffusive exchange of gas-phase chemical is included for transfer between air and leaf.
The boundary layer approach is used with air-side and vegetation-side resistance modeled in
series (Mackay 1991; Schwarzenbach et al. 1993). The vegetation-side resistance combines
cuticular and stomatal uptake in parallel and the stomata is modeled in series with the mesophyll
for certain mercury species based on evidence reported by Shi-Hua (1982).
Transfer of chemical between the leaf and the particles on leaf is modeled with a simple
first-order rate constant that is specified by the user. Litter fall is also modeled as a first-order
process that transfers a specified percentage of leaf mass, along with particles on the leaf, to
surface soil over a specified period of time (30 days for deciduous trees, several hundred days
for grasses and herbs, several years for conifers).
Transfer through the root and stem is modeled in TRIM.FaTE as an advective process
where the moving phase is either pore-water, xylem, or phloem fluid. Measurements of the
uptake of chemicals into vegetation from soil have been described by Stem-Concentration-
Factors (SCFs) and Transpiration-Stream-Concentration-Factors (TSCFs) for the stem/soil
relationship, and by Root-Concentration-Factors (RCFs) for the root/soil relationship. The
TSCF is used to predict the concentration in the advecting phase (xylem fluid), which is then
used along with an estimated SCF to model the dynamic uptake in stem. The leaf is then
SEPTEMBER 2002
3-32
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
modeled in series with the stem. The RCF and the time required to reach 95 percent of the
root/soil equilibrium value are used to model dynamic uptake into roots (Burken 1998; Briggs et
al. 1982; Briggs etal. 1983).
Table 3-10
Intended Gain/Loss Structure of TREVLFaTE Vegetation Compartments
Compartment
Type
%/ r
Particle-on-leaf
Leaf
Stem
Root
Gains from
Name
T
XL^LP
T
XA^LP
T
XLP^L
T
J^St^L
TA^L
TL^SI
T
xSr^St
T
LSi->R
Source/description
Leaf(F)
pseudo diffusion
Air (A)
dry particle deposition
wet particle
deposition
Leaf particle (F)
pseudo diffusion
Stem (A)
xylem flow
Air (D,A)
wet gas deposition
dry gas deposition
Leaf (A)
phloem flow
Root soil (A)
Root soil (A)
water uptake
Losses to
Name
T
J^LP^Ss
T
XLP^L
T
XLP^A
T^2
TL.SS
T
XL^LP
T
J^L^St
TL.A
Tl.2
TSI^L
Tl.2
T
LSt->Si
T,.,
Target/description
Surface soil (F,A)
wash-off and litter
fall
Leaf(F)
pseudo diffusion
Air (A)
blow-off of particles
Reaction (F)
Surface soil (A)
litter fall
Leaf particle (F)
psuedo diffusion
Stem (A)
phloem flow
Air (D,A)
diffusion
Reaction (F)
Leaf (A)
xylem flow
Reaction (F)
Root soil (A)
senescence
Reaction (F)
Notes: D indicates diffusive transfer, A indicates advective transfer, and F indicates user supplied first-order rate
constant.
3.8.1.2 Reconciling the Conceptual Model and the Code
The evaluation of the code in Prototype V of TRIM.FaTE identified a number of
discrepancies between the intended and actual model code. These discrepancies and the steps
taken to reconcile the differences are listed in Table 3-11.
SEPTEMBER 2002
3-33
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-11
Summary of Results from the Compositional Audit of Prototype V
Identified Discrepancy
Dry particle deposition from air to vegetation
missing from code
Wet particle deposition from air to vegetation
missing from code. Algorithm in TSD for wet
deposition of particle (Eq. 7-9) is same as that
for gas (Eq. 7-7)
Transfer from leaf to air and leaf to soil for
parti cle-on-leaf a results in "double-counting"
blow-off
Mesophyll resistance missing from stomatal
diffusion pathway
Limited basis for mass transfer rate between
parti cles-on-leaf and leaf
Not clear whether litter fall loss should be
linear or exponential
Current method for soil-root-stem continuum
may impact dynamic mass balance
predictions'3
Reconciliation
Implemented appropriate algorithm from
TSD
Developed algorithm for wet deposition of
particles for inclusion in code and TSD
Removed transfer to soil of blow-off
Developed and implemented algorithm for
mesophyll resistance in air-leaf diffusion
Used best engineering judgment to specify
value
Future analysis of different litter fall models
Further evaluation (e.g., via use of alternative
model)
a TRIM.FaTE includes a particle-on-leaf compartment to provide more relevant exposure estimates for chemicals that are predominantly
particle bound. However, little information is available for characterizing the particle-on-leaf compartment. Best engineering judgment
is used to estimate mass transfer rates for the particle-on-leaf compartment.
b TRIM.FaTE is a dynamic model, but the state of the science for vegetation root uptake decouples the soil-to-root transfers from the
soil-to-stem transfers. Structurally, the soil-root-stem-leaf continuum allows stem and root to simultaneously approach equilibrium
with the root-zone soil. Theoretically this could allow the stem (and subsequently the leaf) to approach equilibrium faster than it would
if modeled in series with the root. It is unclear how this model assumption impacts the dynamic mass balance.
3.8.2 Algorithm Audit
In the algorithm audit each T-factor in the plant module was audited for its technical
quality (correct derivation) and its relevance (appropriate for modeling tasks that TRIM.FaTE is
designed to address). The T-factors were grouped by compartment interface across which they
transfer mass. The TRIM.FaTE plant module includes eight interfaces:
1. Particle-on-leaf <->air;
2. Particle-on-leaf-^surface soil;
3. Parti cle-on-leaf<->leaf;
4. Leaf<->air;
5. Leaf>surface soil;
SEPTEMBER 2002
3-34
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
6. Leaf<->stem;
7. Root-zone soil>stem; and
8. Root-zone soil < root.
The arrows in the interface names indicate the directions of mass transfer allowed by the
transfer factors. Table 3-12 presents the findings of the algorithm audit and the changes made or
planned for each finding.
Table 3-12
Summary of Results from the Algorithm Audit
Finding
Recommendation
Particle-on-Leaf <-> Air Interface
Dry deposition not normalized to bulk air
compartment
Dry particle deposition rate specific for
certain chemical class
Particle blow-off from leaf to air forces
concentration in moving phase to equal
concentration in receiving compartment
particles
Wet particle deposition T-factor to
vegetation missing
Included appropriate partition coefficient to
normalize T-factor to lower air compartment
type
Applied a more general default value (500 m/d)
and placed an explanatory note in users'
manual
Developed estimate for volume of particles on
leaf and related T-factor to concentration in
sending compartment
Developed and included algorithm based on
wet deposition to soil of particles
Particle-on-Leaf > Surface Soil
Default value used for wash-off rate is
suspect (much too rapid)
Best judgment for litter fall rate seems
appropriate, but should be evaluated on
case-by-case basis
Used a 14-day half-life (wash-off rate of
0.05/day) for removal of particles-on-leaf to
soil
Included comment in users' manual and TSD
Vol II suggesting evaluation of default
assumptions using sensitivity analysis
Particle-on-Leaf <-> Leaf Interface
Data lacking for specified rate constant
Included comment in users' manual and TSD
Vol II suggesting evaluation of default
assumptions using sensitivity analysis
SEPTEMBER 2002
3-35
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Finding
Recommendation
Leaf <-» Air Interface
Incorrect washout ratio for gasses
Wet interception fraction model
significantly under-predicts interception
Mesophyll resistance missing from the
stomatal uptake pathway for certain
mercury species
Air-side boundary layer thickness model
difficult to parameterize
Natural log (In) of Kow used to predict
cuticle conductance but original model was
log base 10 (log)
Model for air-stomata transfer over-predicts
transfer when relative humidity (RH) is
greater than 80 percent
Changed Zpure air/ZTotal dr in the wet deposition of
gas algorithm to Zpure wate/ZTotal air
Further evaluation of cumulative rain (rather
than rain rate) for estimating interception
Developed and parameterized mesophyll
resistance model to add in series to air-stomata
pathway
Replaced equation with lognormal distribution
having a mean of 5.0E-4 and CV of 1
Changed natural log to log base 10 in estimate
of cuticular conductance
Placed upper limit on conductance based on
empirical evidence or use alternate algorithm
from Reiderer (1995) or Nobel (1999)
Leaf > Surface Soil
Best judgment for litter fall rate seems
appropriate, but should be evaluated on
case-by-case basis
See particle-on-leaf > surface soil (above)
Leaf <-» Stem Interface
Model for estimating xylem flux over-
predicts relative to reported values
Kow correction exponent for leaf and stem
are suspect
Model for estimation of stem/xylem
partition coefficient does not agree with
published model (Briggs et al. 1983)
Evaluate options of applying an upper limit or
replacing algorithm
Additional literature review to evaluate default
value
Evaluate calibration of prototype model or
replacement with Briggs et al. (1983) model
Root-Zone Soil » Stem Interface
Estimated transpiration stream flux in
excess of reported values
Alternate models exist for transpiration
stream concentration factor (TSCF)
See Leaf <-> Air Interface
Evaluate combining existing models or
incorporating uncertainty into prediction
SEPTEMBER 2002
3-36
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Finding
Recommendation
Root-Zone Soil < Root Interface
Default
rapid
value for time to equilibrium very
Evaluate incorporating uncertainty
in the input
3.8.3 Future Activities for Evaluation of the Plant Module
With the conceptual and mechanistic evaluations of the plant module completed, the next
steps will be to evaluate the actual performance of the module. Recent data for a class of organic
chemicals will be used to evaluate the deposition of pollutants to above-ground vegetation and to
compare the relative importance of particle-bound and gas-phase deposition processes.
3.9 CONCENTRATIONS AND FLOWS THROUGH TERRESTRIAL
WILDLIFE
This evaluation examines the reasonableness of predicted concentrations and flows of
contaminants in terrestrial wildlife compartments. The evaluation focuses on four aspects of the
TREVLFaTE's design and performance:
Algorithm coding (e.g., are all links included, are units correct, are biomasses of
different trophic levels reasonable);
Differences in Hg concentrations among trophic levels;
Relative input and output fluxes; and
Reasonableness of concentrations compared to soil and plant concentrations and
relative to values found in the literature.
As a result of preliminary simulations which illustrated the impact of unrealistically high
deer and vole biomass input values, biomass estimates typical for forested areas in the
northeastern United States were used in this evaluation
Section 3.9.1 describes the model setup for this evaluation. Section 3.9.2 presents the
results of the evaluation. Conclusions are presented in Section 3.9.3.
3.9.1 Model Setup
The model setup for this evaluation included single soil (with surface and root zone
components), air, and surface water (with sediment) volume elements. The configuration and
dimensions of the abiotic compartments are shown in Figure 3-7. Groundwater and vadose zone
compartments were omitted. All wildlife biotic compartment types available in the TRIM.FaTE
library were included in the model setup; the grasses/herbs vegetation type was the only plant
SEPTEMBER 2002 3-37 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
compartment included. Constant meteorology data were used. The temperature was set at 25
degrees Celsius, wind speed at two m/s, and there was no precipitation.
Figure 3-7
Modeling Configuration for Evaluation of Concentrations and
Flows through Terrestrial Wildlife
Horizontal Compartment Dimensions
N 9km
t
1km
10km
Soil Compartments (Surface, Root Zone, and
Vadose Zone Soil)
Water and Sediment Compartments
Note: Soil and water/sediment
compartments are overlain by a 10 km
by 10 km air compartment.
Vertical Compartment Dimensions
Surface Soil
Root Zone Soil
001m
055m
Air
A
\
k
r
w
100m
3.0m Surface Water
0.02 m Sediment
Note: Not to scale
The model setup did not include a source term. Instead, static concentrations in the
abiotic compartments, which are shown in Table 3-13, were the source of mercury to the system.
The simulation period for the evaluation was five years.
SEPTEMBER 2002
3-38
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-13
Mercury Concentrations in Abiotic Compartments in the Terrestrial Wildlife Evaluation
Mercury
Species
Elemental
Divalent
Methyl
Fixed Concentration in Environmental Media
Surface
Water (ntg/L)
l.OE-8
9.0E-7
l.OE-7
Sediment
(g/m3 or
mg/L)
2.3E-8
2.3E-4
4.6E-6
Root Zone
Soil (g/m3 or
mg/L)
l.OE-8
l.OE-4
2.0E-6
Surface Soil
(g/m3 or mg/L)
2.6E-5
2.6E-1
5.2E-3
Lower Air
(mg/L)
l.OE-9
2.0E-11
7.4E-93
a Concentration not fixed.
3.9.2 Results
The reasonableness of the terrestrial wildlife algorithm coding can be evaluated based on
the distribution of mercury mass between abiotic and biotic compartments at the end of the
simulation period, the final concentrations of mercury in the various biotic compartments, and
fluxes of mercury among compartments.
Table 3-14 shows the distribution of mass of various species of mercury in a few biotic
and abiotic compartments at the end of the simulation period. The total mass of mercury is
dependent on the size of the compartment as well as the chemical transfer algorithms (note that
the surface water volume element is smaller than the soil volume element in this setup). As
expected, a very small fraction of the mass of total mercury is in mammals and birds (i.e., less
than 0.001 percent of the mass in the system). Most of the mercury in terrestrial wildlife is in the
form of divalent and methylmercury (especially the former), with little mercury in elemental
form.
Table 3-14
Distribution of Mass Across Compartments at the End of a Five-Year Simulation Period
Compartment
Total mass, g
(not including sinks)
Abiotic
compartments
Biotic compartments
Soil
Air
Elemental Hg
6.6 E+4
99%
0.001%
98%
0.76%
Divalent Hg
6.6E+8
99%
0.052%
99%
0%
Methyl Hg
1.3E+7
99%
0.35%
99%
0.018%
SEPTEMBER 2002
3-39
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Surface water
Sediment
Fish
Macrophyte
Mammal
Bird
Plant
Insect
Worm
0.023%
0.34%
0%
0%
0%
0%
0%
0%
0.001%
0%
0.35%
0%
0%
0%
0%
0.052%
0%
0.001%
0.001%
0.35%
0.001%
0.002%
0%
0%
0.34%
0.001%
0.001%
Figure 3-8 presents concentrations of divalent and methylmercury in soil and key biotic
compartments at the end of the five-year simulation period. The highest concentrations of
divalent mercury in biota are predicted in the arthropod, shrew, racoon, and earthworm. The
highest concentrations of methylmercury in biota are predicted in the kingfisher, arthropod,
shrew, and mallard. The concentration of divalent mercury in soil is higher than the
concentrations of divalent mercury in biota. However, the concentrations of methylmercury in
the four biotic compartments noted above are higher than those in soil.
Figure 3-8
Mercury in Mammals and Birds After a Five-Year TRIM.FaTE Run
0 0.1 n
1
1 ~ o.oi -
o "St
U jf 0.001 -
1 0.0001 -
i
n nnnm
-
~
n
r-i
1
PI
I
m
DHg(2)
~
ifl
~
-
1
m
-
1
MeHg
fh
Compartment
Figure 3-8 shows that the shrew has the highest concentration of total mercury of the
vertebrates, which may be due to its high feeding rate, as well as its diet of earthworms and
arthropods, both of which have high levels of mercury. In general, the final concentrations of
SEPTEMBER 2002
3-40
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
mercury in the terrestrial wildlife compartments suggest that the input and output fluxes are
reasonable for the wildlife species included.
Further evaluation of the mercury fluxes showed the primary route of exposure for
terrestrial wildlife to be via the diet (including ingestion of soil), rather than via air, as expected
(Gnamus et al. 2000). For example, the flux of divalent mercury to white-tailed deer from the
diet during the growing season was 4.1E-2 g/d, from water was 1.7E-2 g/d, and from air was
5.7E-7 g/d. (Deer also received 5.3E-5 g/d from inhalation of elemental mercury.) An
unexpected observation of this evaluation was that the meadow voles ingested as much divalent
mercury from consumption of soil (5.5E-3 g/d) as from consumption of plants (6.1E-3 g/d )
during the growing season. Although this result was surprising, no evidence was readily
available that contradicts this relationship.
Algorithms, including dietary links, are correct for the select wildlife species that were
examined, with one exception. The wildlife were not linked to the leaf surface, which contains
deposited particles that are a component of their diets. Version 1 of the model addressed this
problem through a smart-linking feature that will automatically create basic links, including this
one, during set-up of a modeling scenario.
Although TRIM.FaTE estimates whole-animal mercury concentrations, mercury
concentration data for terrestrial mammals and birds generally are reported only for specific
organs (e.g., liver) in the scientific literature. Thus, very limited empirical concentration data are
available for comparison with the estimated biotic compartment mercury concentrations. Bull et
al. (1977) and Talmage and Walton (1993) measure concentrations of mercury in mammalian
organs only. The Mercury Study Report to Congress, Vol. VII (U.S. EPA 1997), emphasized
risks to piscivorous wildlife, so data on body burdens in terrestrial wildlife were not compiled. It
is reasonable to assume that piscivorous wildlife would have higher exposures to mercury than
most terrestrial wildlife, and that is not evident in this model evaluation run; loons have among
the lowest body burdens of mercury. However, concentrations of mercury in compartments in
the aquatic food chain and those in the terrestrial chain in this simulation should not be compared
for two reasons: (1) the fixed concentrations of mercury in soil and water may not have been
consistent (e.g., the assumed background concentrations of mercury in soil may have been higher
than mercury concentrations in most background soils and higher than would have been
expected to be observed adjacent to water bodies containing 1E-6 mg/L Hg in water) and (2)
fixing concentrations in abiotic compartments did not allow for exchange that may have changed
the relative concentrations of mercury in the soil, water, air and sediment compartments. For
example, if sediment is a long-term sink for mercury, that would not have been observed in this
evaluation.
3.9.3 Conclusions
Based on this evaluation, a single model alteration was made to reflect the fact that
herbivorous vertebrates eat particles on plant leaves along with the leaves themselves. Future
versions of TRIM.FaTE could consider the addition of a change in biomass feature (e.g.,
incorporate growth, birth, death, migration) for terrestrial wildlife.
SEPTEMBER 2002 3-41 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.10 CONCENTRATIONS AND FLOWS THROUGH FISH
Two sets of algorithms for the uptake of mercury by fish are available in TRIM.FaTE:
(1) A bioenergetic model (i.e., all uptake occurs via the diet)5; and
(2) An equilibrium model (i.e., concentrations at one trophic level are assumed to be
3.5 times the concentration at the next lower level, when equilibrium is reached).
The equilibrium model is based on empirical data for mercury from Lindqvst et al. (1991). Thus,
a comparison of the bioenergetic and equilibrium models is also a comparison of the
bioenergetic model with empirical bioconcentration factors (BCFs) and bioaccumulation factors
(BAFs).
At this time, trophic levels are designated as herbivore, omnivore, and carnivore. The
current terminology may be confusing and potentially misleading to some users. For example,
the term "omnivore" is currently applied to fish that eat only herbivores, and the term
"carnivore" is applied to fish that eat only herbivorous fish. The current nomenclature is
explained clearly in the TSD. In future versions of TRIM.FaTE, this nomenclature may be
changed to trophic level X, X+l, and so on.
The evaluation of concentrations and flows through fish was designed to examine and
compare the two sets of algorithms, with particular attention given to:
Algorithm coding (e.g., whether all links are included, all units are correct, and
the biomasses of different trophic levels are reasonable);
Differences between dynamic and equilibrium results;
The relationship between mercury concentrations in different trophic levels;
The sensitivity of mercury concentrations in fish to biomass at each trophic level;
Reasonableness of mercury concentrations in fish compared to surface water and
sediment concentrations (at background concentrations, fixed abiotic
concentrations, and concentrations representing chlor-alkali contamination); and
Reasonableness of BCFs and BAFs relative to literature.
5 Gill uptake of the chemical directly from water also is included in the bioenergetic model for use with
nonionic organic chemicals and potentially other inorganic compounds. This is not included for mercury, as uptake
of mercury directly from water via gills is considered negligible relative to dietary uptake.
SEPTEMBER 2002 3-42 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.10.1 Model Setup and Evaluation Methods
Compartment types included in this evaluation were surface water, sediment, three
trophic levels of water column fish, two trophic levels of benthic feeding fish, benthic
invertebrates, and loons. Algae were included as part of the surface water compartment (i.e., a
phase). Air and surface soil were also included in runs but they are not the focus of this
evaluation. The spatial configuration for this evaluation was similar to that shown in Figure 3-7.
Two scenarios were included in the evaluation, a background scenario and a
contaminated scenario. Fixed abiotic concentrations of mercury in air, surface soil, and sediment
were used in both scenarios; aquatic concentrations are presented in Tables 3-15 and 3-16. The
background scenario was used for most evaluative runs. The contaminated scenario was
generally used for comparisons with measured data, because some BAFs were derived using
contaminated conditions and may not be valid under background conditions. However, because
of the linear structure of the aquatic food chain in TREVI.FaTE, the model generated the same
trophic-level specific BAFs (i.e., ratio of the chemical concentration in fish to the chemical
concentration in water) regardless of the scenario used.
Table 3-15
Environmental Mercury Concentrations for the Background Scenario
Mercury
Species
Elemental
Divalent
Methyl
Fixed Concentration in Environmental Media
Surface Water
(mg/L)
l.OE-8
9.0E-7
l.OE-7
Sediment
(mg/L)
2.3E-5
4.6E-1
4.6E-3
(mg/kg)
l.OE-5
2.0E-1
2.0E-3
Table 3-16
Environmental Mercury Concentrations for the Contaminated Scenario
Mercury
Species
Elemental
Divalent
Methyl
Fixed Concentration in Environmental Media
Surface Water
(mg/L)
l.OE-7
9.0E-6
l.OE-6
Sediment
(mg/L)
2.3E-4
4.6E+0
4.6E-2
(mg/kg)
l.OE-4
2.0E+0
2.0E-2
SEPTEMBER 2002
3-43
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Background concentrations of mercury species were obtained from the Mercury Study
Report to Congress (U.S. EPA 1997). Ratios of species of mercury in the contaminated scenario
were assumed to be the same as those in the background scenario. The 1E-5 mg/L (10 ng/L)
concentration of total mercury in surface water in the contaminated scenario is typical of chlor-
alkali-contaminated lakes in New York (Onondaga Lake) and Ontario (Clay Lake) (Wang and
Driscoll 1995). The 2 mg/kg concentration of total mercury in sediment represents a high-end
value (the maximum concentration reported by Gonzalez (1991) for lakes impacted by a chlor-
alkali plant). In retrospect, 0.2 mg/kg would have been a better concentration to choose (a
reasonable estimate in Gonzalez and the average of the sediment concentrations measured near
the test case site), but because of the linear scaling of the model and the fact that BAF results
were not affected, there was no need to rerun the model. Overall results after modifying the
model confirmed biomagnification across trophic levels.
The ratio of methylmercury to divalent mercury (some of which is also taken up by algae,
passed on to fish, and then converted to methylmercury) in surface water that was fixed in the
model runs was realistic in comparison to literature values. For example, Watras et al. (1995)
measured the median concentration of total mercury in Wisconsin lakes to be 0.96 ng/L, and
methylmercury in Wisconsin lakes is 0.07 ng/L, giving a methylmercury to total mercury ratio of
7.3E-2. This ratio is close to the fixed ratio used in TREVLFaTE of 9.9E-2 methylmercury to
total mercury.
For most runs, the following biomasses of water column and benthic fish were used:
Water-Column Food Chain Benthic Food Chain
Herbivore: 1.65E-3 kg/m2 Mayfly/invertebrate: 3.73E-2 kg/m2
Omnivore: 5.85E-4kg/m2 Omnivore: 1.89E-3 kg/m2
Carnivore: 1.79E-4 kg/m2 Carnivore: 2.14E-4 kg/m2
As described in the next two sections, initial runs involved erroneous biomass values for water
column herbivores and omnivores. Additionally, the biomasses of water column fish were
altered for the evaluation described in Section 3.10.2.5. In all simulations, loons were assigned a
biomass of 2.0E-7 kg/m2, and an ingestion rate of 0.23 kg/kg body weight, with herbivores being
their only food item.
3.10.2 Evaluations and Results
The discussion of evaluations and results includes basic relationships (Section 3.10.2.1),
structural problems (Section 3.10.2.2), comparison of alternative models (Section 3.10.2.3),
comparison of model outputs to measured concentrations (Section 3.10.2.4), sensitivity of the
model to the biomass of higher trophic-level fish (Section 3.10.2.5), and options for addressing
impact offish biomass on fish mercury concentrations in the bioenergetic model (Section
3.10.2.6).
SEPTEMBER 2002 3-44 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
3.10.2.1 Basic Relationships
Initial results from the fish model were evaluated to confirm that two basic relationships
were present. First, mercury speciation in fish was checked to ensure that methylmercury was
the predominant species; this was found to be true. Second, model results were analyzed to
determine if biomagnification of methylmercury occurs as it should in fish. Two situations were
identified where biomagnification did not occur as expected:
Methylmercury concentrations in water-column herbivores were higher than those
in omnivores in the bioenergetic runs. This was rectified by correcting relative
biomasses of trophic levels (see Section 3.10.2.2). The problem arose again due
to an error in the ingestion rate equation.
Methylmercury concentrations in benthic omnivores were higher than those in
carnivores in the equilibrium model. This was rectified by correcting errors in the
algorithms.
3.10.2.2 Structural Problems
The fish model evaluation revealed several structural problems with the model. The
problems and their remedies are described below:
Biomass input values for water column fish of different trophic levels did not
reflect the reality that aquatic systems always have a lower biomass offish at each
higher trophic level. It was determined that some of the initial values had errors,
and the biomass input values were modified based on the methodology described
in Appendix I-B.
There was no link in the model from the benthic invertebrate to the benthic
omnivore in the bioenergetic scenario. The link was added.
The elimination rate constant for methylmercury in fish in the bioenergetic model
contained a term with incorrect units. The units were corrected.
An exponent in the feeding rate equation was off by a factor of 10. In addition,
the fish ingestion rate calculation did not include division by fish biomass. These
algorithms were corrected.
In initial equilibrium model runs, methylmercury concentrations in the benthic
omnivore were higher than in the benthic carnivore. The concentration of
methylmercury in the benthic carnivore was identical to that in the benthic
invertebrate. A close inspection of the algorithm revealed that the equations for
transfer between the omnivore and carnivore compartment types had been
reversed. The partition coefficient was used in the transfer from carnivore to
omnivore rather than in the opposite direction, which would have been correct
and parallel to other transfers up the food chain. These algorithms were corrected.
SEPTEMBER 2002
3-45
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
In addition to checking and correcting algorithms, three additional tasks were carried out:
(1) a comparison of results from the bioenergetic and equilibrium models; (2) a comparison of
TRIM.FaTE outputs to measured values; and (3) a qualitative sensitivity analysis of changes in
fish biomass. These tasks were initially completed when model algorithms contained errors but
were updated to reflect the updated version of the TRIM.FaTE library.
3.10.2.3 Comparison of Alternative Models
Concentrations of methylmercury predicted in aquatic biota by the bioenergetic and
equilibrium models were compared for both the background (Table 3-17) and contaminated
(Table 3-18) scenarios. Except in the case of the benthic carnivore, methylmercury
concentration differences between the two models were within a factor of three. While
methylmercury concentrations in the water-column herbivore, water-column omnivore, and loon
compartment types were slightly higher using the equilibrium model, concentrations in the
water-column carnivore, benthic carnivore, and benthic omnivore were slightly lower. However,
the concentration of mercury in benthic carnivores estimated by the bioenergetic model was 30
times the concentration estimated by the equilibrium model. It is not clear why this difference is
this large. Future evaluations will consider this.
It is notable that the concentrations in fish in both models scale up perfectly with the
concentrations in surface water or sediment. In the contaminated scenario, water-column and
benthic fish concentrations were exactly 10 times those in the background scenario. These
factors represent the change in surface water and sediment concentrations as well. Thus, to
answer questions related to BAFs, only one of the two scenarios needs to be run.
Table 3-17
Comparison of Concentrations of Methylmercury (mg/kg) in Aquatic Organisms,
As Calculated by Bioenergetic Model and Equilibrium Model (Five-Year Simulation)
With Background Levels of Mercury in Water and Sediment
Aquatic Organism
Water-column
herbivore
Water-column
omnivore
Water-column
carnivore
Benthic invertebrate
Benthic omnivore
Benthic carnivore
Bioenergetic
Model
2E-4
6E-4
1E-2
1E-2
8E-2
2E+0
Equilibrium
Model
5E-4
1E-3
3E-3
1E-2
3E-2
7E-2
Bioenergetic/
Equilibrium Ratio
4E-1
5E-1
3E+0
1E+0
2E+0
3E+1
SEPTEMBER 2002
3-46
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Aquatic Organism
Loon
Bioenergetic
Model
4E-4a
Equilibrium
Model
9E-4a
Bioenergetic/
Equilibrium Ratio
4E-1
total mercury
Table 3-18
Comparison of Concentrations of Methylmercury (mg/kg) in Aquatic Organisms,
As Calculated by Bioenergetic Model and Equilibrium Model (Five-Year Simulation) with
Fixed, Contaminated Levels of Mercury in Water and Sediment
Aquatic Organism
Water-column
herbivore
Water-column
omnivore
Water-column
carnivore
Benthic invertebrate
Benthic omnivore
Benthic carnivore
Loon
Bioenergetic
Model
2E-3
6E-3
1E-1
1E-1
8E-1
2E+1
4E-3a
Equilibrium
Model
5E-3
1E-2
3E-2
1E-1
3E-1
7E-1
9E-3a
Bioenergetic/
Equilibrium Ratio
4E-1
6E-1
3E+0
1E+0
2E+0
3E+1
4E-1
a total mercury
3.10.2.4 Comparison of TRLVLFaTE Outputs to Measured Concentrations
To calculate BCFs and BAFs from TRIM.FaTE outputs to compare to the Mercury Study
Report to Congress values, one must first calculate the dissolved concentration of
methylmercury in surface water. The outputs from TRIM.FaTE for these runs indicate 98.7
percent of the mass of methylmercury in the surface water compartment was sorbed to suspended
sediment, and 0.11 percent is in algae. By subtraction, 1.2 percent of the mass of methylmercury
in the surface water compartment is in the dissolved phase. Multiplying 1.2 percent by the
surface water concentrations yields 1.2E-9 mg/L as the appropriate denominator for calculating
the BAF in the background runs and 1.2E-8 mg/L for the contaminated scenario runs. As is
shown in Table 3-19, the BAFs derived from these TRIM.FaTE runs fall within the range
reported in the Mercury Study Report to Congress (U.S. EPA 1997).
SEPTEMBER 2002
3-47
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-19
Comparison of TREVLFaTE Bioconcentration Factors (BCFs) and Bioaccumulation
Factors (BAFs) to Related Factors from the Mercury Study Report to Congress
Factor
Mercury Study Report to
Congress
TRUVLFaTE
Phytoplankton
BCF
3,400 - 133,000
36,300
Trophic Level 3
Fish BAF
1.6E+6
Background and Contaminated
Scenario Runs:
-Herbivore: 1.47E+5
-Omnivore: 5.01E+5
Equilibrium Runs:
-Herbivore: 4.11E+5
-Omnivore: 1.14E+5
Trophic Level 4
Fish BAF
6.8E+6
Background and Contaminated
Scenario Runs:
-Omnivore: 5.01E+5
-Carnivore: 1.19E+7
Equilibrium Runs:
-Omnivore: 1.14E+5
-Carnivore: 2.12E+6
Foodchain
Multipliers
Phytoplankton - Zooplankton: 6.3
Zooplankton - Forage Fish: 6.2
Forage Fish - Piscivores: 4.9
i.5
Biota/Sediment
Accumulation
Factors (BSAFs)
(total Hg)
0.4-50
Bioenergetic Model
- Benthic Omnivore: 1.9
- Benthic Carnivore: 45
Equilibrium Model
-Benthic Omnivore: 0.78
- Benthic Carnivore: 1.6
The BCFs calculated for algae were compared to those in the Mercury Study Report to
Congress (U.S. EPA 1997). BCFs for methylmercury in phytoplankton, which relate a
concentration in algae (wet mass) to the dissolved concentration in water, have been calculated
at 3,400 for Lake Michigan, 38,400 for East Fork Poplar Creek in Tennessee, 107,000 for
Onondaga Lake in New York, 90,000 for a northern Wisconsin lake, and 133,000 for Little Rock
Lake in Wisconsin (U.S. EPA 1997). The value calculated for all final test runs of TRIM.FaTE
was 36,300, which is within this empirical range. Note that unlike other equilibrium
relationships in TRIM.FaTE that have constant BAFs or BCFs, the BCF for phytoplankton is a
SEPTEMBER 2002
3-48
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
calculated parameter, based on factors such as pH, chloride concentration in water, and algal
radius.
The Mercury Study Report to Congress (U.S. EPA 1997) presents BAFs in L/kg (wet
weight concentration divided by dissolved concentration) for fish at trophic levels 3 and 4, where
trophic level 3 fish are forage fish feeding on zooplankton and trophic level 4 fish are
piscivorous fish feeding on forage fish. It is difficult to equate these with the trophic levels
currently used in TRIM.FaTE, but trophic level 3 fish are probably comparable to herbivores or
omnivores in TRIM.FaTE, and trophic level 4 fish are probably comparable to omnivores or
carnivores in TRIM.FaTE. As summarized in Table 3-19, the median BAF for level 3 fish is
1.6E+6 (5th percentile, 4.6E+5; 95th percentile, 5.4E+6), and for level 4 fish is 6.8E+6 (5th
percentile, 3.3E+6; 95th percentile, 1.4E+7) (U.S. EPA 1997). These BAFs from the Mercury
Study Report to Congress are based on measurements from field studies. In the background and
contaminated scenario runs for TRIM.FaTE, BAFs for herbivorous, omnivorous, and
carnivorous fish in the bioenergetic model were calculated to be 1.47E+5, 5.01E+5, and
1.19E+7, respectively. In equilibrium runs, the same BAFs were calculated to be 4.11E+5,
1.14E+6, and 2.12E+6, respectively for these three types of fish. These values are in line with
the values in the Mercury Study Report to Congress.
The Mercury Study Report to Congress presents food-chain multipliers of 6.3, 6.2, and
4.9 for phytoplankton-to-zooplankton, zooplankton-to-forage fish, and forage fish-to-piscivore,
respectively. These values are somewhat higher than the food chain multiplier of 3.5, which
serves as the basis for the equilibrium model. However, 3.5 is well within the range of values
from which the multipliers in the Mercury Study Report to Congress were derived. A separate
evaluation will focus on the effect of consumers on concentrations of mercury in their prey,
including effects on mercury accumulation across the food chains (e.g., on the resulting "food
chain multipliers").
Biota/sediment accumulation factors (BSAFs) for total mercury in aquatic biota (dry
weight basis) range from 0.4 to about 50 (U.S. EPA 1997). Within a system, BSAFs usually
increase with increasing trophic level (U.S. EPA 1997). In these TRIM.FaTE simulations,
sediment is 20 percent pore space and has a density of 2.6E-3 kg/m3; thus, given the density of
water (1,000 kg/m3), sediment is approximately 92 percent solids by weight, and wet weight
concentrations may be multiplied by 1.08. Fish are assumed to be 80 percent water, on average,
so wet weight concentrations may be multiplied by 5. Therefore, on a dry weight basis and
under the scenarios with fixed abiotic concentrations, the bioenergetic model gives BSAFs of 1.9
for the benthic omnivore and 45 for the benthic carnivore. The equilibrium model gives BSAFs
of 0.78 for the benthic omnivore and 1.6 for the benthic carnivore. These BSAFs are within the
range of values in the Mercury Study Report to Congress.
Most measurements of concentrations of mercury in loon tissues are in feathers and
blood; it is unclear how feather and blood concentrations relate to whole animal concentrations.
However, EPA (1997a) cites two studies in which total mercury concentrations in breast muscle
tissue were measured. These measurements should be more representative of concentrations in
whole animals than the measurements taken in feathers or blood. In one of these studies (Wren
et al. 1983), a single loon had a muscle tissue concentration of 1.5 mg/kg (total mercury, wet
weight), which was higher than the total mercury concentration found in any fish in the study.
SEPTEMBER 2002 3-49 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Concentrations of mercury in bluntnose minnow, rainbow smelt, smallmouth bass, northern pike,
and lake charr ranged from 0.1 to 1 mg/kg. This small line of evidence suggests that
TRIM.FaTE may be underestimating the mercury levels in the loon, because the modeled
concentrations (for both models) of total mercury in the loon at five years of exposure (of an
eight-year loon lifespan) are lower than those in the omnivore and carnivore. However, this
measurement is only of a single loon. Also, the assumption was made in TRIM.FaTE that loons
eat only herbivores, which may not be correct. The Mercury Study Report to Congress (U.S.
EPA 1997) cites Barr (1996) in describing the diet of loons to be almost exclusively trophic level
3 (i.e., foraging fish). In addition, inhalation was not included as an exposure pathway in the
fixed-abiotic concentration model runs.
3.10.2.5 Sensitivity of Models to Biomass of Higher Trophic-Level Fish
A comparison was performed of TRIM.FaTE runs using realistic biomass information
(calculated from Kelso and Johnson 1991, Appendix I-B), which was also used in the analyses
described above and artificial biomass values as shown in Table 3-20. This comparison was
designed to determine the sensitivity of concentrations of methylmercury in fish at lower trophic
levels to the biomass offish in higher trophic levels for both the bioenergetic and the equilibrium
models. The biomass values in upper trophic-level fish in the artificial scenario were lowered.
With the bioenergetic model, using a lower biomass of water-column omnivore and
carnivore fish results in substantially higher methylmercury concentrations in all three water-
column trophic levels (Table 3-21). The increase in methylmercury concentration in the
carnivore was greater than an order of magnitude. With the equilibrium model, small differences
were observed in methylmercury concentrations in fish between the two biomass scenarios
(Table 3-22). Concentrations of total mercury in the loon were dependent on the fish biomass
for both the bioenergetic and equilibrium runs. Thus, biomass offish has an impact on the
concentration of methylmercury in fish at lower trophic levels, as well as on piscivorous wildlife,
and those impacts appear to be greater when the bioenergetic algorithm is used compared to the
equilibrium model.
Table 3-20
Biomass of Water-Column Fish in Realistic and Altered Scenarios
Biotic Compartment
Water-Column Herbivore
Water-Column Omnivore
Water-Column Carnivore
Realistic Biomass (kg/m2)
1.7E-3
5.9E-4
1.8E-4
Altered Biomassa (kg/m2)
1.7E-3
1.7E-5
1.7E-7
a Altered to reduce potential impact of dietary uptake of higher trophic-level fish on lower trophic-level fish.
SEPTEMBER 2002 3-50 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Table 3-21
Sensitivity of Methylmercury Concentrations to Biomass of Fish in Bioenergetic
Model with Background Mercury Levels
Biotic Compartment
Water-Column Herbivore
Water-Column Omnivore
Water-Column Carnivore
Semi-aquatic Piscivore (Loon)
Concentration of Methyl Mercury (mg/kg)
Realistic Biomass Scenario
1.8E-4
6.0E-4
1.4E-2
3.7E-43
Altered Biomass
Scenario
8.4E-4
1.6E-2
3.8E-1
1.4E-33
a total Hg
Table 3-22
Sensitivity of Methylmercury Concentrations to Biomass of Fish in Equilibrium
Model with Background Mercury Levels
Biotic Compartment
Water-Column Herbivore
Water-Column Omnivore
Water-Column Carnivore
Semi-aquatic Piscivore (Loon)
Concentration of Methyl Mercury (kg/m2)
Realistic Biomass Scenario
4.9E-4
1.4E-3
2.6E-3
3.7E-43
Altered Biomass
Scenario
5.4E-4
1.8E-3
3.5E-3
9.3E-43
a total Hg
3.10.2.6 Options for Addressing Impact of Fish Biomass on Fish Mercury
Concentrations in the Bioenergetic Model
An option for altering biomass offish is to adjust the biomass of a compartment as a fish
consumes the biomass of the compartment and the mercury contained in the biomass. However,
consumption without replacement is probably not advisable, and concentrations in all fish should
be diluted by growth, birth, death, and other natural processes.
The 1998 draft version of Aquatox (Park 1998 draft) includes a derivative of biomass
(g/m3/d) which equals the sum of:
Load (usually from upstream)
Consumption of food
SEPTEMBER 2002
3-51
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 3
MECHANISTIC AND DATA QUALITY EVALUATION
Migration
Promotion (into size class or trophic level)
minus:
Defecation
Respiration
Excretion
Death
Predation
Gamete loss
Washout downstream
Migration
Promotion (into next size class or trophic level)
The most important factors to include are likely consumption of food, predation, and death.
Birth is apparently not included because it does not bring new biomass into the population.
Migration is also not included, perhaps because biomass gains and losses may be balanced.
3.10.3 Conclusions and Summary
Based on this evaluation, the following model alterations have been implemented:
Altered default biomasses of fish at different trophic levels;
Corrected units of biomass in the elimination rate constant equation in the
bioenergetic model;
Added the link from benthic invertebrates to benthic omnivores in the
bioenergetic equation;
Corrected the algorithms for the bidirectional exchange between benthic
omnivores and benthic carnivores in the equilibrium model;
Corrected units in equation used to calculate mass fraction of algae in water;
Corrected mass fraction of carbon in algae;
Increased elimination rate constant for divalent mercury in fish;
Corrected exponent in fish feeding rate equation; and
Corrected fish ingestion rate to include division by the biomass of the fish
consumer.
Future enhancements to TRIM.FaTE may consider dynamic modeling offish biomass to
simulate biomass losses and gains (e.g., via predation, etc.).
SEPTEMBER 2002 3-52 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
4. STRUCTURAL AND COMPLEXITY EVALUATION
This chapter presents the results of structural and complexity evaluations of TRIM.FaTE.
The first section provides background on the types of analysis that are included in this evaluation
and describes the general approach of the evaluation of TRIM.FaTE. The subsequent sections
describe the specific evaluation activities that have been completed to date, including the
analysis of air compartments only, evaluation of the effect of the biotic system on the overall
mass balance, and analyses of how temporal and spatial characteristics affect TRIM.FaTE.
4.1 BACKGROUND AND APPROACH
This section provides an overview of the structural complexity evaluation for
TRIM.FaTE. Section 4.4.1 is an introduction to model structural evaluation, and Section 4.1.2
explains the general structural evaluation approach for TRIM.FaTE.
4.1.1 Introduction
Judging the reliability of a model requires an _ ... . ..
, 1-^1 1 11 11 Structural evaluation activities focus
understanding or how the model responds to changes
on how changes in modeling complexity
affect model performance. These
activities seek to determine how the
model will respond when modeling
in complexity (i.e., changes in the modeling
structure). Both temporal and spatial changes can be
made to the model structure. Structural evaluation
addresses these kinds of changes and provides
valuable information about the performance and
behavior of the model under a range of conditions,
improving the ability to judge the model's quality
and reliability. Ideally, these evaluations can help determine the optimal model structure to
balance the level of complexity needed to create reliable outputs with the simplifications that can
make the model easier and more practical to use. If the model is less complex, it is easier to
perform additional analyses, such as uncertainty and sensitivity analyses, and the model is more
practical to apply to specific sites and situations. Structural evaluation can provide insight and
guidance for future model applications, and it is a valuable input to developing user guidance.
A large number of well-designed runs is necessary to evaluate model structure and
complexity. These structural evaluations combine sensitivity analysis methodology with model-
to-model comparisons. For a structural evaluation, the model is set up for an application using
either real or hypothetical data. Changes are then made to the structure (e.g., spatial elements are
split or recombined; time steps are changed; parcel shapes, sizes, and locations are altered), and
the model outcomes are compared (i.e., the model is compared to itself under various setup
conditions).
Structural and complexity evaluations encompass a series of comparisons designed to
measure the model's response to various changes, which can include:
SEPTEMBER 2002 4-1 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Different run duration and/or time step values;
Varying spatial configurations;
Changes in initial and boundary chemical concentrations;
Changes in the source and/or target locations; and
Other changes in complexity (e.g., including/excluding biota, using average
precipitation versus discrete rain events).
4.1.2 General Structural Evaluation Approach for TREVLFaTE
TRIM.FaTE is intended to be used in a wide range of modeling applications (e.g., various
chemicals, environmental settings, exposure conditions). Because TRIM.FaTE can be used at
various levels of complexity, it is important to understand the level of complexity needed for a
particular analysis and the stability of model outputs when the system structure is changed.
Given the complexity of the "real world" and the large number of inputs used in TRIM.FaTE, a
complete set of structural evaluations cannot be identified and performed. The focus to date of
structural evaluation activities for TRIM.FaTE has been the responsiveness to changes in model
complexity with respect to both temporal and spatial scales and the types of compartments
included.
Several structural evaluation activities have been performed for TRIM.FaTE, including
the following.
Response to changes in the size, shape, and location of parcels. Using the
mercury case study data set, EPA examined the effect of varying spatial
configurations on TRIM.FaTE results. This included changing the size of parcels
in multiple dimensions to determine the most appropriate way to set up the model
layout, as well as adding parcels at the edges of the model region to examine the
boundary effects around the model system (i.e., flux of chemical mass into or out
of the system).
Response of abiotic compartments to the exclusion/inclusion of biota. It has
typically been assumed that the mass of a chemical in biota compared to the mass
in abiotic compartments (e.g., soil, water, air) is not large enough to influence the
overall chemical mass balance. However, if both the flux into the biotic
compartment and the reaction rates within the compartment are rapid enough, the
biota could potentially influence the mass balance even when a relatively small
volume of biota is present (Maddalena 1998). Testing was performed to measure
the model response to biota inclusion to determine when and to what extent biota
need to be included in mass balance calculations.
SEPTEMBER 2002 4-2 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Response to temporal scales of analysis and to aggregate inputs. Detailed
meteorological data (and in some cases other time-varying input data, such as
emission rates) are available and have been used in a simplified scenario to test
the model's response to aggregation of input data over various time periods. By
running the model with varying degrees of input aggregation, (e.g., hourly vs.
daily vs. seasonal average wind or temperature data), the level of input detail
required to achieve a specified level of detail in the output can be determined.
4.2 AIR COMPARTMENT EVALUATION
All aspects of TRIM.FaTE, including the air transport module, comprise a mass-balanced
grid model. For the air module, advection and diffusion are the processes currently included to
transport chemical mass between different air parcels. Although TRIM.FaTE is not designed to
replace currently existing air transport or dispersion models for inhalation exposure and risk
assessments, the air module serves as the key link in transporting mass from an emission source
to other media (e.g., soil, water, sediment) and biota compartments.
Four separate initial structural and complexity evaluations of the TRIM.FaTE air module6
have focused on evaluating the effect of parcel size and shape, grid layout, and grid orientation
on air concentration results:
Regular grid with controlled variation in meteorology. A regular grid (i.e., a
"checkerboard" with square parcels) was used with "controlled" meteorology
input data constructed to test particular aspects of TRIM.FaTE;
Variation of parcel sizes. Actual hourly meteorological data for a location in the
northeastern United States was used with a series of regular grids with parcels of
various sizes;
Variation of overall grid area. Actual hourly meteorological data for a location
in the northeastern United States was used with a series of regular grids covering
different total grid areas; and
Variation of parcel shape. Actual hourly meteorological data for a location in
the northeastern United States was used with an "approximated polar" grid
(parcels were regularly shaped, non-square polygons).
The orientations of these layouts are presented in Figures 4-1 and 4-2. For each of these
analyses, TRIM.FaTE was set up with air compartments only (i.e., no other media were included
in these runs). The model setup, inputs, and results and observations for each analysis are
described in the subsequent sections.
6 Only the air module has been tested separately as part of the structural and complexity evaluation; other
modules of TRIM.FaTE have been evaluated collectively as part of the biotic, temporal, and spatial complexity
analyses described in subsequent sections of this chapter.
SEPTEMBER 2002 4-3 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-1
Example of Regular Grid Layout
(the shaded parcel is the source location)
1
6
11
16
21
2
7
12
17
22
O
8
13
18
23
4
9
14
19
24
5
10
15
20
25
Figure 4-2
Example of Approximated Polar Grid Layout
(the shaded parcel is the source location)
12
11
17
10
SEPTEMBER 2002
4-4
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
4.2.1 Regular Grid with Controlled Variation in Meteorology
For this evaluation, TRIM.FaTE was run under controlled meteorological conditions to
evaluate the impact of using a regular grid parcel layout on mass transfer in the air transport
component of TRIM.FaTE.
4.2.1.1 Model Inputs and Grid Layout
Two regular grids comprised of a single layer of square parcels centered on a single
source were tested:
5x5 parcel grid (770.6 km2 total area, 5.55 km x 5.55 km parcels)
13x13 parcel grid (770.6 km2 total area, 2.14 km x 2.14 km parcels)
The simulation was run for 30 days with background concentrations set to zero and constant
source emissions. The output concentrations (i.e., average air concentrations for each parcel)
were obtained at hourly intervals. In order to focus on the effects of a regular grid layout, a
controlled meteorological data set was used consisting of the following values:
Constant height of air compartment: 30m.
Constant wind speed: 3 m/s.
Constant vertical wind velocity: 0 m/s.
Constant stability class: 4 (unitless).
Uniformly changing wind direction: an initial southerly wind was used (in the
meteorological sense, hence, blowing toward due north); the value of the wind
direction was changed 0.5 degree/hour in the clockwise direction (i.e., one
clockwise sweep of the compass over a 30-day period).
4.2.1.2 Results and Observations
For a single time step (i.e., one hour), mass emitted by the source is always transported
from the source compartment into one quadrant only, where a quadrant is defined as the area
between the axes of the four compass directions (i.e., due north, south, east, and west). These
results were observed for both sizes included in this analysis. The following related results were
also observed:
For wind directions directly along the axis lines (e.g., due east or 90°), mass was
transported into volume elements along the axis only.
For wind directions (angles) close to the axis lines defining the quadrants, mass
transport was biased toward the axis lines, with no crossing of the boundary to the
adjacent quadrant. For wind directions more centrally located between axes (e.g.,
SEPTEMBER 2002 4-5 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
45°, or northeast), mass was transported to a greater number of volume elements,
resulting in greater distribution of mass.
As a result, greater mass occurred in parcels along the four axes than in other parcels in this grid
layout over the 30-day period. Increasing the square grid resolution for the fixed total area (i.e.,
from 5 x 5 to 13 x 13) resulted in larger variation in results across parcels within the quadrants.
In both layouts, somewhat higher average concentrations were observed along the axes.
Monthly average air concentrations for the 13x13 grid are shown in Figure 4-3. Each
linear intersection on this plot represents an average concentration for an individual volume
element (i.e., parcels would actually be centered on the intersections of the x-y grid in this chart).
For this figure, the concentrations for each volume element were used to derive a three-
dimensional surface with contour lines. The different shades on this plot represent ranges of
concentrations (contour interval = 2E-9). Note that the large peak in the center of the plot
represents the concentration in the source parcel.
Figure 4-3
Monthly Air Concentrations (g/m3), Controlled Meteorology Data
2.4E-08-2.6E-08
D2.2E-08-2.4E-08
D2E-08-2.2E-08
1.8E-08-2E-08
1.6E-08-1.8E-08
D1.4E-08-1.6E-08
1.2E-08-1.4E-08
D1E-08-1.2E-08
8E-09-1E-08
D6E-09-8E-09
D4E-09-6E-09
2E-09-4E-09
DO-2E-09
In a separate analysis (results not shown), wind direction was varied by 15 degrees per hour (i.e.,
one clockwise sweep of the compass each day). The overall monthly average concentrations for
parcels were comparable to the results (shown here) obtained when varying the wind direction
by only 0.5 degree/hour.
SEPTEMBER 2002
4-6
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Two additional observations were noted. First, the amount of mass and the resulting air
concentrations in the source parcel changed with wind direction in relationship to grid
orientation. Concentrations were higher when the wind blew along the axes than when the wind
blew between the axes, with the greatest difference being about 30 percent (for 13 x 13 grid).
Secondly, total mass in the modeling domain also changed with wind direction. The total mass
and resulting average air concentration in the modeling domain was lower when the wind blew
along the axes than when it blew at angles to the axes, with the greatest difference being less
than 15 percent (for 13x13 grid). These effects appear to result from changes in cross-section
area (i.e., the area perpendicular to the wind direction) of the source parcel or the whole grid as
the wind changes direction.
Because wind direction was the only varying parameter for all runs performed in this
analysis, the effects noted here (i.e., bias along the axes and the fluctuating source parcel and
overall concentration patterns) are concluded to result from the geometric relationship between
wind direction and source-parcel orientation (e.g., the trigonometric resolution of wind vectors
relative to the volume element surfaces). The significance of these air compartment design
findings will need to be considered in the context of other uncertainties in the simulation and
their impact on the prediction of mass distribution and concentrations in non-air media.
Investigation of this effect on the alternate parcel layout design is presented in Section 4.2.4.
4.2.2 Variation of Parcel Sizes
This evaluation was focused on the effect of grid resolution on mass transfer in the air
component of TRIM.FaTE. In particular, this evaluation examined the impact of dividing a
given modeling area into different-sized square parcels.
4.2.2.1 Model Inputs and Grid Layout
This analysis used five regular grids comprised of square parcels centered on a single
source:
One-parcel grid (770.6 km2 total area, 27.76 km x 27.76 km parcel)
3x3 parcel grid (770.6 km2 total area, 9.25 km x 9.25 km parcels)
5x5 parcel grid (770.6 km2 total area, 5.55 km x 5.55 km parcels)
9x9 parcel grid (770.6 km2 total area, 3.08 km x 3.08 km parcels)
13x13 parcel grid (770.6 km2 total area, 2.14 km x 2.14 km parcels)
A single source with constant emissions of elemental mercury only was located in the
center of the grid. A simulation period of one month (744 hours) was used, and hourly results
were obtained and averaged to calculate an overall monthly average. Actual hourly
meteorological data from January 1990 for a location in the northeastern U.S. were used for the
runs. One vertical layer was defined with the height equal to the mixing height. Vertical wind
velocity data were not relevant for this evaluation because only one vertical layer was used.
Note that the conclusions from this analysis are dependent upon the ratio of advection to
dispersion for the location and season. Analyses for sites where mean wind speeds are low but
SEPTEMBER 2002 4-7 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
considerable direction changes occur (e.g., wintertime in the San Joaquin Valley of California)
may be significantly different.
4.2.2.2 Results and Observations
As expected, the highest average monthly concentration for each of the scenarios
occurred in the air parcel containing the source. For the most part, the results for all of the
scenarios seemed qualitatively consistent with each other and the modeled meteorological data
(i.e., the higher concentrations occurred in the directions the wind blew most frequently).
Monthly average concentrations for the entire grid were very consistent across scenarios
(average concentration of 6.2E-10 g/m3, standard deviation of 3.6E-11), especially when the
one-parcel scenario was omitted (see Table 4-1). Thus, the amount of grid resolution did not
have significant effects on the total mass in the system (if the one-parcel grid is ignored). It did,
however, have an impact on the spatial distribution of mass in the system. For example, the
effect of the regular grid (described in Section 4.2.3) of producing somewhat higher
concentrations along the axes than between the axes is also observed here, although it is less
pronounced here than with the constant meteorological data scenarios described in Section 4.2.3.
Additionally, this effect seems to be lessened further with increased spatial resolution.
Table 4-1
Average Monthly Concentrations Across Modeled Area
Regular Grid Scenarios
One-parcel grid
3x3 parcel grid
5x5 parcel grid
9x9 parcel grid
13x13 parcel grid
Average
Standard Deviation
Average (without one-parcel grid)
Standard Deviation (without one-parcel grid)
Average Concentration Over Regular
Grid Area (g/m3)
6.8E-10
6.1E-10
6.0E-10
6.0E-10
6.0E-10
6.2E-10
3.6E-11
6.0E-10
4.3E-12
As expected, additional grid resolution provided greater spatial variation in the results.
For example, the ratios of maximum:minimum values, going from the 3x3 grid to the 13x13
grid, are 42, 96, 236, and 414. To compare the concentration patterns between scenarios, the
results were processed in two ways:
First, the results of the 9x9 Regular Grid were projected into the grid layout of
the 3x3 Regular Grid. Based on these results, it appears that increased grid
resolution results in higher concentration estimates in the grid corners, especially
SEPTEMBER 2002
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
in the predominant wind direction. It appears that as the grid resolution increases,
the movement of mass is less biased to the north-south and east-west axes and
thus more mass is able to reach the corners before exiting the system to the sinks.
It also appears that added resolution around the source parcel and/or decreased
source parcel size will result in higher concentrations in the source parcel but
lower overall concentrations in the area surrounding the source. These
differences were less than 20 percent for the two grids evaluated.
Second, the concentration estimates for the 3 x 3, 5 x 5, 9 x 9, and 13 x 13
Regular Grid scenarios were averaged into quadrants. This was done by dividing
the modeling region of each scenario into four pieces, defined by axes running
through the middle of the source parcel in the north-south and east-west
directions. For parcels that were divided between two quadrants (i.e., those along
the axes), their concentrations were weighted half as much as those that fit
completely into the quadrant. Furthermore, the source parcel was divided equally
between the quadrants and its concentration was weighted one-fourth as much as
the parcels contained completely within the quadrant. The results are presented in
Table 4-2.
The estimated quadrant concentrations decrease with increased resolution in the
NW, NE, and SW quadrants, while increasing with increased resolution in the SE
quadrant. It also should be noted that the SE quadrant had the highest average
concentration among the quadrants. In general, it appears that increased
resolution will result in lower concentrations upwind and higher concentrations
downwind. This may be attributable to the lessening of "artificial" dispersion7
with increased resolution. The concentration difference from the grid layout with
the least resolution (3 x 3) to that with the greatest resolution (13 x 13) was less
than a factor of two.
Table 4-2
Quadrant Average Concentrations (g/m3) for Grids in the Parcel Size Evaluation
Grid Layout
3x3 Regular Grid
5x5 Regular Grid
9x9 Regular Grid
13 x 13 Regular Grid
NW
4.1E-10
3.3E-10
2.7E-10
2.4E-10
SW
5.5E-10
5.1E-10
4.7E-10
4.4E-10
NE
6.4E-10
6.2E-10
6.0E-10
5.9E-10
SE
8.4E-10
9.4E-10
1.1E-09
1.1E-09
Average
6.1E-10
6.0E-10
6.0E-10
6.0E-10
7 "Artificial" dispersion refers to instantaneous dispersion of chemical mass throughout a given air
compartment, regardless of size or shape. A larger compartment will spread chemical mass evenly throughout a
larger volume in the same time period relative to a smaller compartment.
SEPTEMBER 2002
4-9
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
4.2.3 Variation of Overall Grid Area
This evaluation considered the effect of varying total grid area on mass transfer in the air
component of TRIM.FaTE. In particular, this evaluation looked at the impact of changing the
modeled area for different grid resolutions.
4.2.3.1 Model Inputs and Grid Layout
For this evaluation, the total grid area (i.e., the modeling domain) was changed by either
adding or subtracting parcels around the edge of a regular, square grid. Two scenarios were
designed to test this effect:
Increased area scenario. An additional ring of square parcels was added to the 5
x 5, 9 x 9, and 13 x 13 Regular Grid layouts described in Section 4.2.1.2, thus
creating three new scenarios (7 x 7, 11 x 11, and 15 x 15) comprised of identical
square parcels and larger total areas than the Regular Grid scenarios from which
they were derived. The new scenarios consisted of the following grids:
* 7x7 parcel grid (1510.4 km2 total area, 5.55 km x 5.55 km parcels)
> 11x11 parcel grid (1151.2 km2 total area, 3.08 km x 3.08 km parcels)
> 15x15 parcel grid (1026 km2 total area, 2.14 km x 2.14 km parcels)
Decreased area scenario. The outer ring of square parcels was removed from
the 5 x 5 Regular Grid scenario to create a 3 x 3 grid with a smaller total area.
(277.4 km2 total area, 5.55 km x 5.55 km parcels)
A single source with constant emissions of elemental mercury only was located in the
center of the grid. A simulation period of one month (744 hours) was used, and hourly results
were obtained and averaged to derive an overall monthly average. Actual hourly meteorological
data for a location in the northeastern U.S. were used for the runs. One vertical layer was
defined with the height set equivalent to the mixing height. Vertical wind velocity data were not
relevant for this evaluation because only one vertical layer was used.
4.2.3.2 Results and Observations
Overall, the trends in the results for the Increased Area scenarios were relatively
consistent with those from the Regular Grid scenarios described in Section 4.2.2 (i.e., the highest
average monthly concentration for each of the scenarios occurred in the parcel containing the
source and the results for all of the scenarios seemed consistent with the wind direction data).
To assess the impact of increasing the modeled area, the estimated concentrations in the
common parcels between the original grids (i.e., 5x5,9x9, and 11x11 Regular Grids) and the
corresponding grids with increased area (i.e., 7x7, 11x11, and 15x15 Grids) were compared.
The results for the 7x7 and 11x11 grids are presented in Figures 4-4 and 4-5. The estimated
concentrations from the Increased Area scenarios were higher than the estimates for the
corresponding Regular Grid parcels across the entire grid. In general, the differences were larger
farther from the source parcel (i.e., nearer the edges). The differences were smaller when grid
SEPTEMBER 2002 4-10 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
resolution was increased. However, in both comparisons, the difference was less than 25
percent. This effect could be minimized in future model applications by increasing the distance
of the boundary of the modeling domain from the area of interest.
Figure 4-4
Percentage Difference in Common Parcels Between
7X7 Increased Area Grid and 5X5 Regular Grid
21%
15%
6.3%
9.7%
13.%
16%
12%
3.6%
5.3%
7.3%
6.1%
3.2%
0.41%
1.2%
2.3%
11%
6.4%
1.6%
2.7%
4.6%
15%
9.2%
3.4%
5.2%
8.2%
N
The value in each cell represents (7x7-5x5)75x5.
Note that the source parcel is shaded.
Figure 4-5
Percentage Difference in Common Parcels Between
11X11 Increased Area Grid and 9X9 Regular Grid
12%
11%
9.0%
6.9%
4.1%
7.0%
8.1%
10%
13%
11%
9.8%
7.9%
5.9%
2.9%
4.5%
5.1%
6.8%
9.2%
9.8%
8.9%
7.0%
5.0%
2.0%
2.8%
3.2%
4.3%
6.2%
8.5%
7.7%
6.0%
4.2%
1.2%
1.8%
2.1%
2.9%
4.3%
3.2%
2.6%
1.7%
0.99%
0.11%
0.31%
0.48%
0.88%
1.7%
5.4%
4.6%
3.2%
2.0%
0.46%
0.65%
0.85%
1.4%
2.3%
7.4%
6.4%
4.4%
2.9%
0.90%
1.2%
1.4%
2.2%
3.4%
9.6%
8.1%
5.6%
4.0%
1.6%
2.0%
2.5%
3.6%
5.2%
11%
9.4%
6.6%
4.9%
2.3%
3.0%
3.7%
5.3%
7.4%
N
t
The value in each cell represents (11x11 - 9x9)79x9.
Note that the source parcel is shaded.
SEPTEMBER 2002
4-11
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Monthly average concentrations across the entire Regular Grid area region of the
Increased Area scenarios were very similar. That is, the average of the interior 5 x 5, 9 x 9, and
13x13 areas of the 7x7, 11x11, and 15x15 grids, respectively, was 6.2E-10 g/m3, with a
standard deviation of 6.7E-12. However, across the three Increased Area scenarios, the average
concentration for the Regular Grid regions decreased slightly as the grid resolution increased
(e.g., the average concentration of the regular grid area for the 7x7 scenario is greater than the
11x11 average concentration). This result (shown in Table 4-3) was expected, considering that
the lower resolution scenarios had larger total grid areas providing more area over which the
mass could distribute prior to irretrievably reaching the sinks.
Table 4-3
Average Monthly Concentrations in the Increased Area Scenarios
Increased Area Scenarios
7x7 grid
11x11 grid
15 x 15 grid
Average
Standard Deviation
Average Concentration Over Regular
Grid Area* (g/m3)
6.2E-10
6.1E-10
6.1E-10
6.2E-10
6.7E-10
* Regular grid scenarios (5 x 5, 9 x 9, and 13 x 13) averaged 6.0E-10 g/m
To further examine the impact of changing the number of identically sized parcels on
concentration estimates, the common parcels between the 3x3 Decreased Area Grid, the 5x5
Regular Grid, and the 7x7 Increased Area Grid were compared. These results are presented in
Table 4-4. As the number of parcels (and thus the modeled area) increases, the concentrations in
all of the common parcels increases. This increase is greatest in the directions of the
predominant winds (i.e., NE, E, and SE). The largest increase was on the order of 20 percent.
These differences are likely a result of increasing the area of the modeled system, thus
preventing mass from leaving the system via the sinks as quickly.
Table 4-4
Comparison of Common Parcels Between
3X3 Decreased Area Grid, 5X5 Regular Grid, and 7X7 Increased Grid
Scenario
3x3
5x5
7x7
Concentrations in Air Parcels (g/m3)
Source
4.2E-09
4.3E-09
4.3E-09
N
5.5E-10
5.9E-10
6.1E-10
NE
3.5E-10
3.9E-10
4.2E-10
E
1.4E-09
1.5E-09
1.5E-09
SE
8.0E-10
8.5E-10
8.8E-10
S
1.4E-09
1.5E-09
1.5E-09
SW
2.2E-10
2.4E-10
2.5E-10
W
3.2E-10
3.4E-10
3.6E-10
NW
9.1E-11
1.1E-10
1.2E-10
Average
6.1E-10
9.7E-09
9.9E-09
SEPTEMBER 2002
4-12
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
4.2.4 Variation of Parcel Shape for a Constant Grid Area
This evaluation considered the effect of varying grid resolution and parcel shape on mass
transfer in the air component of TRIM.FaTE. In particular, this evaluation looks at the impact of
dividing a given area into different sized polygons.
4.2.4.1 Model Inputs and Grid Layout
For this evaluation, three grid layouts were designed in an attempt to replicate a radial
grid using polygons. The resulting grids subdivided the total grid area referenced in Sections
4.2.1. and 4.2.2 (770.6 km2) into four-sided parcels centered on a square-shaped source (see
Figure 4-2 for the layout of the 17-parcel grid as an example). The source parcel is the shaded
square parcel in the center of the grid.
17-parcel grid (770.6 km2 total area, same source size and shape as 5 x 5 Regular
Grid)
33-parcel grid (770.6 km2 total area, same source size and shape as 9 x 9 Regular
Grid)
49-parcel grid (770.6 km2 total area, same source size and shape as 13 x 13
Regular Grid)
As in the analyses described previously, a single source with constant emissions of
elemental mercury only was located in the center of the grid. A simulation period of one-month
(744 hours) was used; hourly results were obtained and averaged to derive an overall monthly
average. Actual hourly meteorological data for a location in the northeastern United States were
used for the runs. Vertical wind speeds were derived from these actual data using a simplified
method based on stability class and horizontal wind speed. One vertical layer was defined with
the height set equivalent to the mixing height.
4.2.4.2 Results and Observations
The general trend in the estimated monthly average concentrations was that the larger the
number of parcels, the higher the average concentration across the full grid area. This was the
opposite of the trend observed with the Regular Grid scenarios, where increasing the number of
parcels resulted in lower average concentrations across the full grid area. As with the Regular
Grid findings (Section 4.2.1), this difference is small (e.g., less than a factor of 2). Furthermore,
the estimated concentrations from the Irregular Grid scenarios (area-weighted average
concentration of 5.6E-10 g/m3, standard deviation of 1.1E-11; full results presented in Table 4-5)
varied slightly less than the results from the Regular Grid and generally predicted lower overall
concentrations. Variation was greater when the one-parcel Regular Grid scenario was omitted.
SEPTEMBER 2002 4-13 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-5
Area-weighted Average Monthly Concentrations Across Modeled Area
Irregular Scenarios
17 Parcel Grid
33 Parcel Grid
49 Parcel Grid
Average
Standard Deviation
Average Concentration
(g/m3)
5.5E-10
5.6E-10
5.7E-10
5.6E-10
1.1E-11
The estimated concentration patterns for the Irregular Grid scenarios were compared by
averaging the concentrations into quadrants. To account for the varying parcel sizes within the
scenarios, the concentrations were weighted based on the area of the parcels. Overall, the
concentrations (presented in Table 4-6) in the Irregular Grid scenarios decreased with increased
numbers of parcels in the upwind quadrants (i.e., NW and SW quadrants) and increased with
increased numbers of parcels in the downwind quadrants (NE and SE).
The area-weighted quadrant concentrations from the Irregular Grid scenarios were also
compared to the estimated quadrants concentrations from the Regular Grid scenarios. Overall,
the Irregular Grid scenarios seem to estimate similar concentrations to the Regular Grid
scenarios, but with fewer modeled parcels (i.e., the results for the 49-Parcel Irregular Grid are
similar to the results for the 13 x 13 Regular Grid and the results from the 33-Parcel Irregular
Grid are similar to the results from the 9 x 9 Regular Grid). Furthermore, the concentrations in
both the Regular and Irregular scenarios decreased with increased numbers of parcels in the
upwind quadrants (i.e., NW and SW quadrants) and increased with increased numbers of parcels
in the SE quadrant. However, whereas concentrations in the NE quadrant decreased with
increased numbers of parcels in the Regular Grid scenarios, concentrations in this quadrant of
the Irregular Grid increased as the grid resolution increased. In addition, the concentrations in
the downwind directions (i.e., NE and SE) were generally lower in the Irregular Grid scenarios,
possibly indicating that this grid layout causes mass to leave the system via the downwind sinks
more rapidly than the Regular Grid layouts.
SEPTEMBER 2002 4-14 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-6
Quadrant Average Concentrations (g/ni3)
Scenario
3x3 Regular Grid
5x5 Regular Grid
9x9 Regular Grid
13x13 Regular Grid
17-Parcel Grid
3 3 -Parcel Grid
49-Parcel Grid
NW
4.1E-10
3.3E-10
2.7E-10
2.4E-10
3.7E-10
2.7E-10
2.5E-10
SW
5.5E-10
5.1E-10
4.7E-10
4.4E-10
4.7E-10
4.5E-10
4.4E-10
NE
6.4E-10
6.2E-10
6.0E-10
5.9E-10
5.6E-10
5.7E-10
5.7E-10
SE
8.4E-10
9.4E-10
1.1E-09
1.1E-09
8.4E-10
9.5E-10
l.OE-09
Overall
Average
6.1E-10
6.0E-10
6.0E-10
6.0E-10
5.5E-10
5.6E-10
5.7E-10
As opposed to the square parcel layouts, the Irregular Grid layouts did not seem to
produce the finding noted in Section 4.2.1 for the Regular Grid of somewhat higher
concentrations along the axes. The geometry of these layouts allowed mass emitted from the
source to be transported over half of the full grid area (180 degrees) for a given wind direction
unless the wind blew directly in either the north, south, east, or west direction, in which case
mass would be transported over one-quarter of the full grid area (90 degrees). However, this
feature seemed to allow transport of some mass in directions that did not seem reasonable (e.g., a
wind blowing one degree east of north would result in a small amount of mass in the southeast
corner of the grid).
In general, the concentration estimates for all three Irregular Grid scenarios were more
evenly spread over the full grid area than were the Regular Grid scenarios. This is likely due to
differences in the range of directions mass can be transported and increased "artificial
dispersion" caused by the shape of the parcels.
Several advantages of the Irregular Grid layout versus the Regular Grid layout were
noted. Specifically, implementing the Irregular Grid layout used in these evaluations:
Avoided or minimized the occurrence of somewhat higher concentrations along
the north-south and east-west axes than between the axes;
Produced similar results to the Regular Grid with far fewer parcels, thus reducing
computational burden;
Provided more grid resolution near the source (where concentrations are typically
higher) and less resolution farther from the source (where concentrations are
typically lower); and
Appeared to be potentially less sensitive to small fluctuations in wind direction.
SEPTEMBER 2002
4-15
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
The following disadvantages of the Irregular Grid layout versus the Regular Grid layout were
also observed:
Some, albeit relatively small, mass can be transported in directions that do not
seem reasonable for certain wind directions; and
Mass seems to exit the system via the sinks more quickly in the downwind
direction.
4.3 BIOTIC COMPLEXITY EVALUATION
This section of the complexity evaluation examines howbiotic compartments (including
terrestrial and aquatic plant and animal compartments) affect the mass balance of modeled
compounds. It has typically been assumed that the chemical mass in biota is small compared to
the chemical mass in abiotic compartments and is not large enough to influence the overall
chemical mass balance. The biotic complexity analysis tests this assumption by successively
adding biotic compartments in a series of simulations. Comparison of the results for the
simulations helps to identify TRIM.FaTE's response to biota and to determine which biotic
compartments need to be included to effectively model chemical concentration in the
compartments of concern. The results of this evaluation have been used to select biotic
compartments to include in further complexity evaluations reported in Sections 4.4 and 4.5.
Biotic complexity evaluations were performed using benzo(a)pyrene and mercury in
order to compare the effects of biotic systems for both organic and inorganic pollutants. Section
4.3.1 and Section 4.3.2 present the biotic complexity evaluations using benzo(a)pyrene and
mercury, respectively.
4.3.1 Benzo(a)pyrene
Section 4.3.1.1 presents the modeling scenarios for the biotic complexity evaluation
using benzo(a)pyrene. Results of the evaluation are presented in Section 4.3.1.2.
4.3.1.1 Modeling Scenarios for Benzo(a)pyrene
The biotic complexity evaluation for benzo(a)pyrene included simulations with four
scenarios. Figure 4-6 shows the configuration of abiotic compartments used in the four
scenarios. This modeling configuration included a single air parcel above adjacent surface soil
and surface water parcels. Root zone and vadose zone soil parcels were included beneath the
surface soil parcel, and a sediment parcel was included beneath the surface water parcel. The
horizontal and vertical dimensions of all parcels are shown in Figure 4-6.
SEPTEMBER 2002 4-16 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-6
Modeling Configuration for the Biotic Complexity Analysis with Benzo(a)pyrene
Horizontal Dimensions
N 9km
t
1km
10km
Soil (Surface, Root Zone, and
Vadose Zone Soil)
Surface Water and Sediment
Note: Soil and surface water/sediment
are overlain by a 10 km by 10 km air
parcel
Vertical Dimensions
Surface Soil 0.01 m
Root Zone Soil 0.55 m
Vadose Zone Soil 0.8 m
Air
100m
3.0m Surface Water
0.02 m Sediment
Note: Not to scale
Biotic compartments included in the four scenarios are shown in Table 4-7. Scenario 1
included only abiotic compartments.8 Scenario 2 included abiotic compartments plus
herb/grassland terrestrial vegetation compartments. Scenario 3 included abiotic compartments,
terrestrial vegetation compartments, and terrestrial and semi-aquatic animal compartments.
Scenario 4 equaled Scenario 3 with the addition of aquatic plant and animal compartments.
8 As algae are represented as a phase in surface water, they are included in all scenarios with inclusion of
the surface water compartment.
SEPTEMBER 2002
4-17
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-7
Biotic Compartments Included in the Biotic Complexity Evaluation with Benzo(a)pyrene
Compartment Type
(Trophic Functional Group)3
Algae
Aquatic macrophyte
Water column herbivore
Water column omnivore
Water column carnivore
Benthic invertebrate (herbivore)
Benthic omnivore
Benthic carnivore
Terrestrial omnivore
Semi-aquatic piscivore
Semi-aquatic predator/scavenger
Terrestrial insectivore
Semi-aquatic herbivore
Terrestrial predator/scavenger
Semi-aquatic insectivore
Terrestrial herbivore
Semi-aquatic omnivore
Terrestrial ground-invertebrate feeder
Flying insect
Soil detritivore
Plant-leaf
Plant-particle on leaf
Plant-stem
Plant-root
Representative Population or
Subgroup
Generalized algal species
User input
NAb
NAb
NAb
Mayfly
NAb
NAb
White-footed mouse
Common loon; mink; belted kingfisher
Bald eagle
Black-capped chickadee
Mallard duck
Red-tailed hawk; long-tailed weasel
Tree swallow
White-tailed deer; mule deer; black-
tailed deer; meadow vole; long-tailed
vole
Raccoon
Short-tailed shrew; Trowbridge shrew
Mayfly
Earthworm; soil arthropod
Leaf (herb/grassland)
Particle on leaf (herb/grassland)
Stem (herb/grassland)
Root (herb/grassland)
Scenario
1
/
2
/
/
/
/
/
3
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
4
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
a Plant parts constitute different compartment types even though they are not different trophic groups.
b This compartment represents a mixed population of species fitting this trophic functional group.
SEPTEMBER 2002
4-18
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
All four scenarios were run with constant meteorological inputs, including a wind of 4
m/sec from due north and rainfall of l.OE-3 m/d. In all scenarios, boundary conditions and
initial benzo(a)pyrene concentrations were set equal to zero, and the benzo(a)pyrene source
emitted a constant 100 g/d (1.2E-3 g/s) into the air compartment. The simulation period was ten
years. The evaluation does not address how biotic complexity affects mass balance over time.
4.3.1.2 Results for Benzo(a)pyrene
As shown in Figure 4-7, the largest effect of the biotic compartment additions on
distribution of mass in the modeling system appears to be on the surface soil compartment. The
addition of plant compartments with Scenario 2 decreased the final concentration and mass of
benzo(a)pyrene in surface soil by about six percent (from approximately 880 g to 830 g). This
reduction may occur because a portion of the chemical mass is now contained in the plants and
thus is unavailable to the surface soil compartment. When present, the plant leaf compartment
accounted for just over ten percent of the total mass in the system. The macrophytes
accumulated the most mass among the aquatic biota compartments, although this represented
less than one percent of the total system mass in the simulation.
Figure 4-7
Benzo(a)pyrene Mass in Biotic and Abiotic Compartments
100 n
80 -
c§
S 60-
3
"8
H
5 40-
a
(U
0.
20-
n
u -~
-
Is!
D Abiotic
Biotic
1
1
Surface Soil
Air
Water
D Sediment
Plants
U Mammals
D Birds
Macrophytes
II DFish
D Insect
Abiotic Only Plus Plants Plus Terrestrial Biota Plus Aquatic Biota
Worm
Because inclusion of terrestrial vertebrates and invertebrates and aquatic biota did not
appear to significantly affect the masses and concentrations in the plants or abiotic media for
benzo(a)pyrene (see Table 4-8), they were not included in other complexity evaluations for
benzo(a)pyrene. Thus, only abiotic media and plants are included in other further complexity
evaluations for benzo(a)pyrene. The response of the modeling system to addition of biotic
compartments may vary for other chemicals.
SEPTEMBER 2002
4-19
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-8
Benzo(a)pyrene Concentrations in Biotic and Abiotic Compartments
Compartment
Air(g/m3)
Water (mg/L)
Sediment (g/m3)
Surface Soil (g/m3)
Leaf(g/kg)
Weasel (g/kg)
Loon (g/kg)
Eagle (g/kg)
Abiotic Only
2.1E-10
2.4E-09
5.8E-06
9.7E-04
Plus Plants
2.1E-10
2.4E-09
5.8E-06
9.7E-04
1.8E-06
Plus Terrestrial
Biota
2.1E-10
2.4E-09
5.8E-06
9.7E-04
1.8E-06
6.3E-09
3.8E-10
1.1E-09
Plus Aquatic
Biota
2.1E-10
2.4E-09
5.8E-06
9.7E-04
1.8E-06
6.3E-09
1.6E-08
5.6E-09
4.3.2 Mercury
The objectives of the biotic complexity evaluation with mercury were (1) to determine
the most important biotic compartment types to include in mercury fate and transport
assessments (where fish are a major focus), and (2) to determine which biotic compartment types
need to be included to effectively model chemical concentrations in the compartment types of
concern. In addition, the biotic complexity evaluation with mercury focused specifically on
whether terrestrial animal compartments affect the contaminant mass in fish. This latter
objective is intended to inform TRIM users about the importance of including terrestrial animals
in simulations focused on mercury accumulation in fish.
4.3.2.1 Modeling Scenarios for Mercury
Five modeling scenarios were developed for the biotic complexity analysis with mercury.
The configuration of abiotic compartments used in all five scenarios is shown in Figure 4-8.
This configuration is similar to the configuration of abiotic compartments used for the
benzo(a)pyrene biotic complexity evaluation.
SEPTEMBER 2002
4-20
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-8
Modeling Configuration for the Biotic Complexity Evaluation With Mercury
Horizontal Dimensions
N 9km
t
1km
10km
Soil (Surface, Root Zone, and
Vadose Zone Soil)
Surface Water and Sediment
Note: Soil and surface water/sediment
compartments are overlain by a 10 km
by 10km air parcel
Vertical Dimensions
Surface Soil 0.01 m
Root Zone Soil 0.56 m
Vadose Zone Soil 1.31 m
Air
100m
3.0m Surface Water
: 0.02 m Sediment
Note: Not to scale
The biotic compartments included in the five scenarios are presented in Table 4-9. All
scenarios included the minimum set of aquatic and semi-aquatic animals identified for Scenario
1. Scenario 1 includes all abiotic compartments, all fish compartments, and the benthic
invertebrate and loon compartments, as well as the short-tailed shrew and earthworm, both of
which are expected to accumulate mercury from the soil.
In Scenario 2, aquatic macrophytes were added to the baseline biotic compartments of
Scenario 1. The results for Scenarios 1 and 2 were compared to determine whether aquatic
macrophytes affect mercury concentration in fish. The addition of aquatic macrophytes in
Scenario 2 may decrease mercury concentration in fish if the plants absorb a significant mass of
mercury from the surface water compartment.
SEPTEMBER 2002
4-21
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-9
Biotic Compartments Included in the Biotic Complexity Evaluation with Mercury
Compartment Type
(Trophic Functional Group)3
Algae
Aquatic Macrophyte
Water column herbivore
Water column omnivore
Water column carnivore
Benthic invertebrate (herbivore)
Benthic omnivore
Benthic carnivore
Terrestrial omnivore
Semi-aquatic piscivore
Semi-aquatic predator/scavenger
Terrestrial insectivore
Semi-aquatic herbivore
Terrestrial predator/scavenger
Semi-aquatic insectivore
Terrestrial herbivore
Semi-aquatic omnivore
Terrestrial ground-invertebrate
feeder
Flying insect
Soil detritivore
Plant-leaf
Plant-particle on leaf
Plant-stem
Plant-root
Representative Population or
Subgroup
Generalized algal species
User input
NAb
NAb
NAb
Mayfly
NAb
NAb
White-footed mouse
Common loon
Mink
Belted kingfisher
Bald eagle
Black-capped chickadee
Mallard duck
Red-tailed hawk; long-tailed weasel
Tree swallow
White-tailed deer; mule deer; black-
tailed deer; meadow vole; long-tailed
vole
Raccoon
Short-tailed shrew; Trowbridge shrew
Mayfly
Earthworm; soil arthropod
Leaf (herb/grassland)
Particle on leaf (herb/grassland)
Stem (herb/grassland)
Root (herb/grassland)
Scenario
1
/
/
/
/
/
/
/
/
/
/
/
2
/
/
/
/
/
/
/
/
/
/
/
/
3
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
4
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
5
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
a Plant parts constitute different compartment types even though they are not different trophic groups.
b This compartment represents a mixed population of species fitting this trophic functional group.
SEPTEMBER 2002
4-22
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Scenario 3 was equal to Scenario 2 with the addition of terrestrial plant compartments. In
Scenario 4, a semi-aquatic piscivore (mink) and a semi-aquatic omnivore (raccoon) were added
to the biotic compartments included in Scenario 3. This scenario was developed to evaluate how
these semi-aquatic fish-eaters influence the mercury concentrations in fish. Because
TREVI.FaTE uses a fixed fish population (i.e., fish biomass in a surface water compartment stays
constant over time) rather than modeling population dynamics, fish predation by raccoons and
mink acts to remove mercury mass from the fish compartments.
Scenario 5 included everything in Scenario 4 plus all other terrestrial biota that were
being considered for the mercury test case (i.e., black-capped chickadee, kingfisher, bald eagle,
mallard, red-tailed hawk, tree swallow, meadow vole, long-tailed vole, long-tailed weasel, white-
footed mouse, white-tailed deer, black-tailed deer, mule deer).
The simulation period for all five scenarios was five years. Meteorological inputs were
annually averaged data of a controlled dataset. These data included a constant precipitation rate
of 3.088E-3 m/d and a wind of 4.862 m/sec from 112 degrees. The mercury emission rate into
the compartment was 1,308 g/d. Mercury emissions speciation was 95 percent elemental and 5
percent divalent. Boundary conditions and initial mercury concentrations were set to zero in all
compartments.
4.3.2.2 Results for Mercury
The presentation of results for the biotic complexity evaluation with mercury focuses on
selected biotic and abiotic compartments. All results presented are the concentrations predicted
in the selected compartments at the end of the five-year simulation period. Figures 4-9, 4-10,
and 4-11 present concentrations of elemental, divalent, and methylmercury, respectively, in
surface soil, a terrestrial ground-invertebrate feeder (i.e., short-tailed shrew), surface water, and a
semi-aquatic piscivore (i.e., common loon). Figure 4-11 also presents concentrations of
methylmercury in the water column carnivore compartment. Concentrations of elemental and
divalent mercury in the water column carnivore compartment were predicted to be zero in all
scenarios and are not included in the figures. Results presented in Figures 4-9 to 4-11 are also
presented in tabular form in Tables 4-10 to 4-12.
SEPTEMBER 2002 4-23 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-9
Elemental Mercury Concentrations in Selected Compartments After Five Years
(a) Elemental Mercury: Surface soil (forest)
2.8E-0
2.7E-08
c
o
o
23
Model Run
(c) Elemental Mercury: Surface water
(b) Elemental Mercury: Terrestrial ground-invertebrate
n nc nn feeder (short-tailed shrew)
o 8 6E 09 -
Of -Z ; Q OC nQ
"c -^
0) 7 RF nQ
C
O 7 4F nQ
7 np nQ
I |
n
~~|r
12345
Model Run
(d) Elemental Mercury: Semi-aquatic piscivore
(common loon)
2.5E-09
Figure 4-10
Divalent Mercury Concentrations in Selected Compartments After Five Years
(a) Divalent Mercury: Surface soil (forest)
2.0E-02
e 2.0E-02
2 1.9E-02
£ o1.9E-02
= i 1.8E-02
o SiSE-02
o 1.7E-02
0 1.7E-02
1.6E-02
234
Model Run
(c ) Divalent Mercury: Surface water
6 4E 04
"c "B) 5 8F-04 -
n
in rz
~LJ
12345
Model Run
(b) Divalent Mercury: Terrestrial ground-invertebrate
feeder (short-tailed shrew)
« -
8s
O
rn
12345
Model Run
(d) Divalent Mercury: Semi-aquatic piscivore
3 QC Q, (common loon)
SB
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-11
Methylmercury Concentrations in Selected Compartments After Five Years
(a) Methyl Mercury Surface soil (forest)
(e) Methyl Mercury Benthic carnivore
(b) Methyl Mercury: Terrestrial gxxincHnvertebrate feeder
(short-tailed shrew)
n8E08
o
2 9^WTR
C ^
g 0) ^4tHJo
2 1E08
1
1
1234
Model Rn
I I
r
r
5
c 1.1E^5
0
2IT 9.0506
4-1 ^
C O)
8j=
c = 7.CE-C6
O
° 5.0E-06
(c) Methyl Mercury: Surface water
~~|
: T
12345
Model Run
(d
o
it aoE^
gs 6CE07
0 2CE07
0.0&CO
) Methyl Mercury:
Seni-aquatic pisdvore (corrrron loon)
1 2
I I I I
r
_r
345
Model Rjn
5.0E-16 i
0 4.0E-16
,§
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-10
Concentrations of Elemental Mercury After Five Years
Compartment
Surface Water (mg/L)
Water Column Carnivore (g/kg)
Common Loon (g/kg)
Surface Soil (g/m3)
Short-tailed Shrew (g/kg)
Air (g/m3)
Scenario
1
5.8E-05
O.OE+00
2.2E-09
2.8E-08
8.4E-09
4.0E-10
2
5.5E-05
O.OE+00
2.1E-09
2.8E-08
8.0E-09
4.0E-10
3
5.5E-05
O.OE+00
2.1E-09
2.6E-08
7.9E-09
4.0E-10
4
5.5E-05
O.OE+00
2.1E-09
2.6E-08
7.9E-09
4.0E-10
5
5.4E-05
O.OE+00
2.1E-09
2.6E-08
7.7E-09
4.0E-10
Table 4-11
Concentrations of Divalent Mercury After Five Years
Compartment
Surface Water (mg/L)
Water Column Carnivore (g/kg)
Common Loon (g/kg)
Surface Soil (g/m3)
Short-tailed Shrew (g/kg)
Air (g/m3)
Scenario
1
6.2E-04
O.OE+00
2.7E-07
2.0E-02
1.2E-06
1.7E-11
2
5.9E-04
O.OE+00
2.3E-07
2.0E-02
1.2E-06
1.7E-11
3
5.8E-04
O.OE+00
2.2E-07
1.8E-02
1.1E-06
1.8E-11
4
5.8E-04
O.OE+00
1.4E-07
1.8E-02
1.1E-06
1.8E-11
5
5.72E-04
O.OE+00
1.4E-07
1.8E-02
l.OE-06
1.8E-11
Table 4-12
Concentrations of Methylmercury After Five Years
Compartment
Surface Water (mg/L)
Water Column Carnivore (g/kg)
Common Loon (g/kg)
Surface Soil (g/m3)
Short-tailed Shrew (g/kg)
Air (g/m3)
Scenario
1
9.8E-06
5.3E-05
1.2E-06
3.5E-04
2.7E-08
4.4E-16
2
7.6E-06
4.2E-05
l.OE-06
3.5E-04
2.6E-08
3.5E-16
3
7.2E-06
4.1E-05
9.9E-07
3.1E-04
2.4E-08
3.3E-16
4
7.2E-06
4.7E-06
5.2E-07
3.1E-04
2.4E-08
3.3E-16
5
7.1E-06
4.5E-06
5.1E-07
3.1E-04
2.4E-08
3.3E-16
SEPTEMBER 2002
4-26
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
In Scenario 2, aquatic macrophytes were added to the baseline biotic compartments of
Scenario 1 to determine whether aquatic macrophytes affect mercury concentrations in other
compartments, particularly the surface water and aquatic carnivore compartments. As shown by
Figure 4-1 l(c), the inclusion of aquatic macrophytes in Scenario 2 reduced the concentration of
methylmercury in the surface water by approximately 21 percent and, apparently consequently,
resulted in a 21 percent reduction in the concentration of methylmercury in the water column
carnivore compartment, as seen in Figure 4-1 l(e).
The addition of aquatic macrophytes affected the mass balances of elemental and divalent
mercury less than it affected the mass balance of methylmercury. As shown in Figure 4-9(c) and
Figure 4-10(c), concentrations of elemental and divalent mercury in surface water with aquatic
macrophytes (i.e., Scenario 2) were about five percent less than the concentrations of elemental
and divalent mercury in surface water without aquatic macrophytes (i.e., Scenario 1). The
addition of aquatic macrophytes did not affect concentrations of elemental and divalent mercury
in the air aquatic carnivore compartments.
Scenario 3 included all compartments included in Scenario 2, plus the terrestrial plant
compartments. Comparison of the results for Scenarios 2 and 3 in Figures 4-9(a), 4-10(a), and 4-
1 l(a) shows that the addition of terrestrial plant compartments to the modeling scenario reduced
the concentrations of all three mercury species in surface soil by approximately 7 to 11 percent.
This result is possibly due to the interception of atmospheric mercury deposition by the
terrestrial plants and is comparable to findings presented for benzo(a)pyrene (see Section 4.3.1).
Further, the addition of the terrestrial plant compartments decreased the concentrations of
divalent and methylmercury in the terrestrial ground-invertebrate feeder (i.e., short-tailed shrew)
by approximately 8 percent, as shown by Figures 4-10(b) and 4-1 l(b). This effect is likely due
to the fact that the shrew diet is comprised of soil-residing biota.
The addition of a semi-aquatic piscivore (mink) and a semi-aquatic omnivore (raccoon)
in Scenario 4 significantly affected methylmercury concentrations in fish. Specifically, inclusion
of these fish-eaters in Scenario 4 reduced methylmercury concentration in the water column
carnivore compartment by approximately 88 percent. This result can be seen by comparing
methylmercury concentrations for Scenarios 3 and 4 in Figures 4-1 l(e). Because TRIM.FaTE
uses a fixed population for the aquatic carnivore compartment rather than a dynamic population
model, predation of largemouth bass by mink and raccoon removes mercury mass from the
aquatic carnivore compartment without reducing the population size (i.e., the biomass). Thus,
the predicted average mercury concentration in the aquatic carnivores is reduced.
Figure 4-1 l(d) also shows that Scenario 4 reduced mercury concentrations in the semi-
aquatic piscivore (i.e., common loon) by about 48 percent. The decrease in the concentration of
mercury in the common loon in Scenario 4 is likely to have resulted from the greatly decreased
mercury concentrations in the fish compartment eaten by loons (e.g., herbivore). As expected,
Scenario 4 did not significantly affect mercury concentrations in surface soil, the terrestrial
ground-invertebrate feeder (i.e., short-tailed shrew), surface water, or air.
Scenario 5 included all remaining terrestrial biota not included in Scenario 4. The
inclusion of all other terrestrial biota compartments did not greatly affect mercury concentrations
SEPTEMBER 2002 4-27 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
(i.e., less than five percent change) in most compartments for any of the mercury species, as
shown by comparing Scenarios 4 and 5 in Figures 4-9 through 4-11.
In summary, it appears that the following biotic compartments should be included in
further complexity evaluations with mercury to enable effective prediction of contaminant mass
distribution among the abiotic compartments of interest:
Aquatic macrophytes (Elodea densa);
Water column herbivore;
Water column omnivore;
Water column carnivore;
Benthic invertebrate (mayfly);
Benthic omnivore;
Benthic carnivore;
Semi aquatic piscivore (common loon);
Semi-aquatic predator/scavenger (bald eagle);
Semi-aquatic carnivore (raccoon)
Terrestrial ground-invertebrate feeder (short-tailed shrew);
Soil detritovore (soil arthropods); and
Terrestrial plants.
These biota include those whose inclusion in the scenario significantly affected mercury fate in
the system, as well as in compartments of particular interest (e.g., fish).
4.4 TEMPORAL COMPLEXITY EVALUATION
The temporal complexity evaluation examines how predicted pollutant concentrations are
affected by changes in the simulation time step. In particular, the evaluation examines the
effects of aggregating meteorological data (and in some cases other time-varying input data, such
as emission rates) over various time periods. By comparing the results of simulations with
varying time steps, the level of input parameter detail required to achieve a specified level of
detail in the output can be evaluated. For example, longer input data averaging times (e.g.,
seasonal meteorological data) may be useful and sufficient for some analyses.
Temporal complexity evaluations were performed with benzo(a)pyrene and mercury.
Chemical transformation was not included in runs for benzo(a)pyrene.
4.4.1 Benzo(a)pyrene
This section presents the temporal complexity evaluation for benzo(a)pyrene.
4.4.1.1 Modeling Scenarios for Benzo(a)pyrene
The temporal complexity evaluation for benzo(a)pyrene was performed with a series of
scenarios based on two modeling configurations. Modeling configuration A is the spatial
configuration of abiotic parcels developed for the biotic complexity analysis with
benzo(a)pyrene (see Section 4.3.1). Modeling configuration B is slightly more complex than
SEPTEMBER 2002 4-28 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
modeling configuration A. The scenarios for each modeling configuration are described further
below.
Scenarios for Modeling Configuration A
The spatial configuration of abiotic parcels for modeling configuration A was described
in Section 4.3.1 of the biotic complexity evaluation and is shown in Figure 4-6. The modeling
configuration includes an air parcel above adjacent surface soil and surface water parcels. Root
zone and vadose zone soil parcels are beneath the surface soil parcel, and a sediment parcel is
beneath the surface water parcel. All scenarios for modeling configuration A include all
terrestrial plant compartments, but no other biotic compartments.
No source term was included. Instead, all benzo(a)pyrene was introduced as constant
boundary condition levels in air blowing into the system. This approach facilitated analysis of
the effect of rain patterns. Because the input of chemical was constant, temporal variation in
deposition was associated with precipitation events. Meteorological inputs were the
meteorological data described previously (e.g., Section 4.2.2). The key meteorological
conditions that varied in these data were wind direction, wind speed, rainfall, and temperature.
The temporal complexity runs performed with modeling configuration A featured various
combinations of input data averaging time (i.e., aggregation period for meteorological data),
simulation period, and reporting time step. The modeling runs are summarized in Table 4-13.
Table 4-13
Temporal Complexity Runs for Modeling Configuration A
Simulation Period
One year
One year
One year
Input Data Averaging Time
Hourly
Day/Night
Daily
Reporting Time Step
Every six hours
Every six hours
Every six hours
Scenarios for Modeling Configuration B
Modeling configuration B is shown in Figure 4-12. The modeling configuration
consisted of five adjacent air and surface soil parcels each measuring 2 km wide by 10 km long.
Root zone soil, vadose zone soil, and groundwater parcels are under each surface soil parcel, and
all terrestrial plant compartments were included. The vertical dimensions of the abiotic parcels
are the same as the vertical dimensions of the abiotic parcels in modeling configuration A (see
Figure 4-6).
SEPTEMBER 2002 4-29 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-12
Modeling Configuration B
10km
i
\
k
r
s
T
N
t
2km
A source of benzo(a)pyrene was emitted directly into the air parcel labeled with an "S" in
Figure 4-12, and media concentrations were evaluated for the target parcel marked with a "T."
For these scenarios, the initial concentration of benzo(a)pyrene was set equal to zero in all
parcels, and boundary conditions were set to zero. The source emitted a constant 1 g/s. The
meteorological data, which were used in scenarios for modeling configuration A, also were used
in scenarios for modeling configuration B. Simulations with modeling configuration B were run
using hourly, daily, and monthly input averaging times.
4.4.1.2 Results for Benzo(a)pyrene
Modeling Configuration A
Average concentrations over a one-year run in the air and surface water compartments
were obtained using meteorology inputs at all input averaging and reporting time steps shown in
Table 4-13. The overall average concentrations for the first year of the modeling run are
presented in Table 4-14. The averages for the entire year are compared (rather than results at a
particular time step) because the individual concentration values for these compartments
fluctuated widely throughout the year. For other compartments, the mass may be compared over
time because the fluctuations were not as great. The average benzo(a)pyrene mass over the first
year of each run is presented in Table 4-15. As Table 4-15 shows, lengthening the input
averaging time interval had only a small effect on average benzo(a)pyrene mass in the air
compartment for a one-year run. However, lengthened input averaging time tended to increase
the average benzo(a)pyrene mass in surface water and surface soil for one-year runs. For
example, the average mass in surface soil estimated with monthly input averaging and reporting
time steps is about 30 percent larger than the average mass estimated with hourly input
averaging time.
Cumulative estimates of benzo(a)pyrene mass in the surface soil compartment for hourly,
daily, and monthly input averaging times are shown in Figure 4-13. As can be seen in this
figure, mass accumulated over time in all scenarios, although scenarios with shorter input
averaging times exhibited greater fluctuations in mass. Concentration and mass were
proportional because concentration is equal to mass divided by volume, which is constant.
SEPTEMBER 2002
4-30
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-14
Average Benzo(a)pyrene Concentrations over One Year in Air and Surface Water for
Various Input Averaging Times
Compartment
Air (g/m3)
Surface Water (mg/L)
Length of Meteorological Data Averaging
Hourly
1.2E-09
1.8E-08
Day/Night
7.6E-10
2.0E-08
Daily
8.3E-10
2.0E-08
Monthly
8.1E-10
2.2E-08
Yearly
8.2E-10
2.2E-08
Table 4-15
Average Benzo(a)pyrene Mass (g) over One Year in Air, Surface Water,
and Surface Soil for Various Time Steps
Input Data
Averaging Time
Hourly
Day/Night
Daily
Daily
Monthly
Monthly
Annual
Reporting Times
Every six hours
Every six hours
Every six hours
Daily
Daily
Monthly
Monthly
Air
4.5E+1
4.5E+1
4.6E+1
4.6E+1
4.6E+1
4.5E+1
4.6E+1
Surface Water
5.3E-1
5.9E-1
6.1E-1
6.1E-1
6.5E-1
6.8E-1
6.6E-1
Surface Soil
2.5E+3
2.7E+3
2.8E+3
2.8E+3
3.0E+3
3.2E+3
3.2E+3
SEPTEMBER 2002
4-31
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-13
Benzo(a)pyrene Mass in Surface Soil Compartment for Various Input Averaging Times
6.0E+03
5.0E+03 -
« 4.0E+03 -
5 3.0E+03 -
S
c3 2.0E+03 -\
a.
S
r^ l.OE+03 H
O.OE+00
Surface soil - hourly
Surface soil - daily
Surface soil - monthly
50 100 150 200
Time (days)
250
300
350
Modeling Configuration B
Figure 4-14 shows how increasing input averaging times affected benzo(a)pyrene mass in
the surface soil compartment of the target parcel in modeling configuration B. Benzo(a)pyrene
mass estimates for surface soil were more affected by input averaging times with the five-parcel
modeling configuration B than with the simple, one-parcel modeling configuration A. This
difference between the results for the two modeling configurations can be explained by wind
direction considerations that apply to modeling configuration B, but not modeling configuration
A. In particular, wind had negligible effect on the results of modeling configuration A, and all
air changes were due to precipitation only.
The combined influence of spatial and temporal complexity is further illustrated in
Figures 4-15 and 4-16, which show the benzo(a)pyrene concentrations in surface soil in the
source and target parcels, respectively, of modeling configuration B. Concentration estimates
are plotted for both the daily- and monthly-average input averaging times.
Figure 4-15 shows that benzo(a)pyrene concentrations in surface soil near the source
were much higher using the daily averaging times than the monthly averaging times. Figure 4-
16 shows that the opposite was true in the target parcel (i.e., benzo(a)pyrene concentrations were
higher using the monthly averaging times than the daily averaging times). This difference may
result primarily from the effects of rain events. In the source parcel, large rain events, which are
shown with arrows in Figure 4-15, are generally associated with increases in the concentration in
the soil, regardless of the wind direction. The deposition due to distinct rain events is greater
than the deposition due to a steady rain in this parcel; thus, the concentration with daily
averaging time is greater than for monthly averaging time.
SEPTEMBER 2002
4-32
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
In contrast, large deposition of benzo(a)pyrene to the soil further from the source (i.e., in
the target parcel) seems to require that the wind blow in that direction while it is raining. In
Figure 4-16, the arrows point to times when there were both strong winds blowing from the
source to the target parcel and a strong rain, conditions which increased the deposition rate to
soil. Because rain events typically do not continue over extended time periods, using monthly
averaged meteorological data can impact where deposition occurs. In this case, benzo(a)pyrene
concentrations in the target parcel were higher using monthly averaged meteorological data
rather than daily averaged data because the wind direction changed less frequently, causing
deposition to occur only in the average monthly wind direction. If the daily averaging times are
used, the wind is often blowing a different direction during the distinct rain events resulting in
fewer large-deposition events.
This evaluation showed that the interaction of wind direction and rain events is important
for determining the benzo(a)pyrene concentration in the soil downwind from the site. These
results imply that it may be important to consider increased temporal resolution with increased
spatial resolution. However, it may be appropriate to aggregate input data if long-term, average
exposures are of interest for a simple, non-transforming (in this analysis) chemical such as
benzo(a)pyrene in a modeling configuration with limited spatial resolution.
These results (which suggest variation in soil concentration values on the order of 20
percent for parcels not adjacent to the source) should be compared to the variance resulting from
parameter uncertainty. It has not been determined whether differences due to time fluctuations
are greater or less than those derived from parameter uncertainty. If the differences are less,
there should be a focus on gathering more information on the other parameter values prior to
additional attention to details regarding increased temporal resolution.
Figure 4-14
Benzo(a)pyrene Mass in Target Surface Soil Parcel, Modeling Configuration B
20,000
4* __
«! S
t ~
4* f*
Ml b
- «
« a.
H S
VI
VI
15,000 -
10,000
5,000 -
Target, hourly
-Target, daily
Target, monthly
20 40 60 80 100 120
Time (days)
140
160
180
200
SEPTEMBER 2002
4-33
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-15
Benzo(a)pyrene Surface Soil Concentrations
for Modeling Configuration B-Source Location
1.5E-04
O.OE+00
Input time step - Monthly
- Input time step - Daily
20 40 60 80 100
Time (days)
120
140
4.0E-05
3.0E-05
2.0E-05
+j
V
u
=
o l.OE-05
U
O.OE+00
Figure 4-16
Benzo(a)pyrene Surface Soil Concentrations
for Modeling Configuration C-Target Location
Input time step -Monthly
Input time step - Daily
Big rain event &
wind from east
20 40 60 80
Time (days)
100
120
140
SEPTEMBER 2002
4-34
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
4.4.2 Mercury
This section presents the temporal complexity evaluation for mercury.
4.4.2.1 Model Setup for Mercury
The temporal complexity evaluation with mercury focused on methylmercury in surface
water and aquatic biota. The temporal evaluation for mercury used the same modeling
configuration used for the biotic complexity evaluation for mercury, which is shown in Figure 4-
8. Biotic compartments in the model setup included:
Terrestrial plants (leaves, particles on leaves, roots, and stems for grasses/herbs);
Terrestrial ground-invertebrate feeder (shrew);
Water column herbivore;
Water column omnivore;
Water column carnivore;
Benthic invertebrate (mayfly);
Benthic omnivore;
Benthic carnivore;
Semi-aquatic omnivore (raccoon); and
Semi-aquatic piscivore (mink).
The model setup also included a constant source term of 1,300 g/day of 95 percent elemental and
5 percent divalent mercury. All initial mercury concentrations and boundary conditions were set
equal to zero. Five-year simulations were run with a controlled meteorological data set. In
separate runs, meteorological data were averaged for daily, monthly, and annual input averaging
time steps.
4.4.2.2 Results
Results presented for this evaluation focus on surface water and fish, because exposure to
methylmercury through fish ingestion is a primary focus for mercury human health risk
assessment. Average mercury concentrations in the surface water compartment are presented in
Table 4-16 and average methylmercury concentrations in fish are shown in Figure 4-17. Table
4-16 shows that, for this simple spatial configuration, the average concentrations of all mercury
species in surface water increased slightly with increasing resolution of input averaging time
steps. This relationship between input averaging time and surface water mercury concentrations
appears to be magnified in the water column fish compartments, particularly in the water column
carnivore compartment. Average mercury concentrations in the water column herbivore and
carnivore fish compartments are presented in Table 4-17. Figures 4-17, 4-18, and 4-19 illustrate
this relationship further. Figure 4-17 presents the concentrations of elemental and divalent
mercury in surface water with daily, monthly and yearly input averaging times, and mercury
concentrations in fish are presented in Figures 4-18 and 4-19, respectively. These results
indicate that for a single air compartment, the mercury concentrations in underlying surface
water did not vary greatly as the input averaging time was changed. However, input averaging
time more significantly affected mercury concentrations in some fish compartments. In addition,
as suggested in the multi-parcel temporal complexity analysis with benzo(a)pyrene (Section
SEPTEMBER 2002 4-35 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
4.4.1.2), changes in averaging time for input data may have had a greater impact if the modeling
configuration was more spatially complex and/or a source compartment was included in the
evaluation.
Table 4-16
Average Mercury Concentrations in Surface Watera
Input Averaging
Time Step
Daily
Monthly
Annual
Mercury Concentration (g/L)
Elemental Mercury
6.4E-05
5.8E-05
5.2E-05
Divalent Mercury
5.7E-04
5.7E-04
5.5E-04
Methylmercury
7.0E-06
6.8E-06
6.6E-06
' Average concentrations are presented for the fifth year of the five-year simulation period.
Table 4-17
Average Methylmercury Concentrations in Water Column Fish3
Input Averaging Time Step
Daily
Monthly
Annual
Methylmercury Concentration (g/kg)
Water Column Herbivore
3.9E-06
3.8E-06
3.7E-06
Water Column Carnivore
6.8E-06
6.4E-06
4.1E-06
1 Average concentrations are presented for the fifth year of the five-year simulation period.
SEPTEMBER 2002
4-36
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
8.E-04 n
6.E-04 -
4.E-04 -
4*
o 2.E-04 -
U
O.E+00 -F
0
Figure 4-17
Mercury Concentrations in Surface Water
Daily Hg(0)
Monthly Hg(2+)
500 1,000
Time (days)
Daily Hg(2+)
"\7" 1 T i" //A\
i eany rig^uj
1,500
--MonthlyHg(0)
Yearly Hg(2+)
Figure 4-18
Divalent Mercury Concentrations in the Water Column Herbivore
Compartment
O.OE+00
Daily Hg(2+)
500 1,000
Time (days)
Monthly Hg(2+)
1,500
Yearly Hg(2+)
SEPTEMBER 2002
4-37
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-19
Methylmercury Concentrations in Water Column Herbivore and Carnivore
Fish
-M
=
=
o
-6.5E+00
-7.5E+00
Herbivore - Daily
Carnivore - Dairy
500 1,000
Time (days)
^^~ Herbivore - Monthly
^^~ Carnivore - Monthly
1,500
Herbivore - Yearly
Carnivore - Yearly
4.5 SPATIAL COMPLEXITY EVALUATION
The spatial complexity evaluation examines how variations in parcel sizes, shapes, and
configurations affect chemical concentrations estimated by TRIM.FaTE in surface soil and air.
The evaluation is performed only for benzo(a)pyrene, and builds on the air parcel evaluation
(Section 4.2) by considering transfers of this pollutant from the air parcel to surface soil.
Specifically, the spatial complexity evaluation examines how predicted chemical concentrations
are affected by four aspects of a hypothetical modeling configuration's design:
The distance between source and target parcels;
The number and size of parcels between the source and target parcels;
The presence of boundary parcels surrounding the source and target parcels; and
The size and configuration of the source parcel.
Each of the spatial complexity evaluations in this section compares predicted
benzo(a)pyrene concentrations in the target parcels of various model scenarios. Except where
otherwise noted, the only difference between the scenarios evaluated are the designs of the
modeling configurations (e.g., the number and size of parcels between the source and target
parcels). All scenarios included the following compartment types: lower air, surface soil, root
soil, vadose soil, and grasses/herbs.
All scenarios were run with a controlled meteorological dataset. The meteorological
conditions that varied in these data were wind direction, wind speed, rainfall, and temperature.
SEPTEMBER 2002
4-38
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
All scenarios included a constant one gram per second source of benzo(a)pyrene. The
simulation period was six months for all scenarios.
4.5.1 Effect of Distance from Source
This section of the spatial complexity evaluation examines how predicted benzo(a)pyrene
concentrations in surface soil are affected by the distance between the source and target parcels
of the modeling configuration. The modeling configuration used for this section of the
evaluation, shown in Figure 4-20, consists of five adjoining parcels, each of which is ten
kilometers long by two kilometers wide. The benzo(a)pyrene source, which is represented by an
"S" in Figure 4-20, is located in the second parcel from the west (i.e., left). The evaluation
focuses on surface soil concentrations in a target parcel identified in Figure 4-20 with a "T."
Figure 4-20
Modeling Configuration for the Evaluation of Effects of Distance From Source
10km
I
\
k
r
S
T
N
t
2km
Predicted benzo(a)pyrene concentrations in surface soil and air compartments of the
target parcel are presented in Figure 4-21 and Figure 4-22, respectively. The concentrations are
plotted for selected time intervals during the six-month simulation period. Distances are from
the edge of the source parcel (i.e., distance zero).
Based on these results, the soil concentration accumulated over time with a fairly
consistent profile. The concentrations in air were more variable than concentrations in soil,
sometimes decreasing less rapidly with distance and at other times decreasing more rapidly.
These differences were expected because the processes affecting estimates of mass in the air
compartment are much more rapid, and therefore much more dynamic, than the processes
affecting estimates of mass in the soil compartment. The estimates of mass in the soil
compartment are less dependent than estimates of mass in the air compartment on current
meteorological conditions.
SEPTEMBER 2002
4-39
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-21
Benzo(a)pyrene Concentrations in Surface Soil
for Evaluation of Effects of Distance from Source"
10 days
- 60 days
120 days
180 days
O.OE+00
-20246
Location Relative to Source (km)
a Although concentrations were calculated for each 20-day interval throughout the 180-day
simulation period, only selected time intervals are plotted.
Figure 4-21 shows that estimated concentrations of benzo(a)pyrene in soil decreased by
roughly 65 percent to 80 percent over the first two kilometers from the source. This indicates
that concentrations in soil decreased more rapidly with distance from the source than was
estimated previously (Bennett et al. 2000) based on the 30-kilometer characteristic travel
distance (CTD) for benzo(a)pyrene. This difference was expected, however, based on
differences between the spatial complexity analysis and the methods used to estimate the CTD.
For example, the method used to develop the CTD assumed advection only in one direction, with
no dispersive or advective losses in other directions. The TRIM.FaTE air compartment included
advection in multiple directions, and the concentration was estimated to change with distance
differently at each time step due to varying meteorological conditions. Because the TRIM.FaTE
air compartment included advection in multiple directions, the rate of decrease in air
concentrations, and thus deposition to the soil compartment, was more rapid than would be
predicted based on the CTD's single-wind-direction method.
4.5.2 Effect of Horizontal Parcel Dimensions
This section of the spatial complexity evaluation examines how benzo(a)pyrene
concentrations at the target parcel were affected by the horizontal dimensions of parcels between
the source and target parcels. The distances between the source and target parcels are equal in
the two modeling configurations developed for this evaluation, which are shown in Figure 4-23.
In modeling configuration 4-23 A, the source and target parcels are separated by a single parcel
with a horizontal dimension of two kilometers. In modeling configuration 4-23B, the source and
target parcels are separated by four parcels, each having a horizontal dimension of 0.5
kilometers.
SEPTEMBER 2002
4-40
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
The predicted concentrations of benzo(a)pyrene in surface soil of the target parcels are
plotted in Figure 4-24. The benzo(a)pyrene concentration in soil of the target parcel of modeling
configuration 4-23 A, in which the source and target parcels are separated by a single two-
kilometer parcel, was slightly higher than the concentration in soil of the target parcel of
modeling configuration 4-23B. Consideration of these results relative to the variance resulting
from parameter uncertainty will inform next steps as to collection of additional parameter
information and/or further evaluation of the effects observed here.
Figure 4-23
Modeling Configurations for Evaluation of the
Effects of Horizontal Parcel Dimensions
(A)
10km
T
N
t
2km
(B)
0.5km
10km
2km
N
t
Note: Not to scale
SEPTEMBER 2002
4-41
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-24
Benzo(a)pyrene Concentration in Surface Soil at Target Parcels of Modeling
Configurations 4-23A and 4-23B
2.5E-05
g, 2.0E-05
^
g 1.5E-05
1
g l.OE-05
a
a
fj 5.0E-06
O.OE+00
Modeling
Configuration
4-23A
Modeling
Configuration
4-23B
0 20 40 60 80 100 120 140 160 180
Time (days)
4.5.3 Effect of External Boundary Parcels
This section of the spatial complexity evaluation for benzo(a)pyrene examines how
external boundary parcels surrounding internal parcels affect media concentrations in the internal
parcels. It was expected that when boundary parcels were added, changes in wind direction
could result in higher media concentrations in the internal parcels in some cases. For example, if
the wind shifted a full 180-degrees, there would be advection of chemical mass directly back into
the internal parcels. Shifts of less than 180 degrees but more than 90 degrees would result in
advection of a lesser amount of contaminant mass back into the internal parcel. This evaluation
compares the predicted benzo(a)pyrene concentrations in the target parcels of the modeling
configurations in Figure 4-25, which have progressively increasing numbers of boundary
parcels.
SEPTEMBER 2002
4-42
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-25
Modeling Configuration for Evaluation of Effects of External Boundary Parcels
(A)
10km
(B)
A
1
L
r
S
T
2km
N
t
N
t
(C)
N
t
(D)
\
^ ht
^ w
30km
/
i
\
i
\
\.
3
r
3
k
r
3km
30 kn
T
i
/
30km
\
N
t
SEPTEMBER 2002
4-43 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-18 presents the concentrations of benzo(a)pyrene predicted in air in the target
parcels of each modeling configuration. Concentration estimates are presented for each of the
first 22 days, as well as the daily averages for the first 22 days of the simulation period. Because
the results were fairly consistent throughout the simulation period, daily results after the 22nd day
are not presented. Ratios of the air concentrations predicted under various scenarios are included
in Table 4-18 to facilitate comparison of the effects of the scenarios.
In all four modeling configurations, the average predicted benzo(a)pyrene concentration
in air in the target parcel was approximately 3.1E-8 g/m3. Daily predicted concentrations in the
target parcel were equivalent in all four modeling configurations during 17 of the first 22 days of
the simulation period. On the five days when winds were from the east (i.e., the target parcel
was upwind from the source parcel) predicted concentrations in the target parcel differed as
much as 88 percent among the four modeling configurations. However, the concentrations in the
target parcel during these days were much lower than on days when winds blew from the west,
and did not appear to significantly affect the average concentrations for the full simulation
period.
Simulations for modeling configurations 4-25A and 4-25D were also compared without
degradation included in the air compartment. The average air concentration in the target parcel
of modeling configuration 4-25D (4.4E-8 g/m3) was nearly identical to the average air
concentration in the target parcel of modeling configuration 4-25 A (4.3E-8 g/m3). In this
evaluation, adding border parcels to the modeling configuration did not greatly affect the average
benzo(a)pyrene concentration predicted in the target parcel.
SEPTEMBER 2002 4-44 TRIM.FATE EVALUATION REPORT VOLUME!
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Table 4-18
Target Air Parcel Concentrations for Modeling Configurations 4-25A, B, C, and D
(With Degradation)3
Time
(Day)
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Averape
Predicted Concentration in Air (g/m3)
4-25A
O.OE+00
3.4e-08
5.1e-08
2.2E-08
4.5E-08
5.7E-08
4.1E-08
6.1E-08
8.9E-13
6.9E-08
4.3E-08
2.6E-08
6.6E-13
3.2E-08
9.5E-08
7.3E-12
1.5E-12
2.4E-08
1.7E-08
2.2E-08
l.OE-08
99E-13
3. IE-OS
4-25B
O.OE+00
3.4E-08
5.1E-08
2.2E-08
4.5E-08
5.7E-08
4.1E-08
6.1E-08
1.3E-12
6.9E-08
4.3E-08
2.6E-08
l.OE-12
3.2E-08
9.5E-08
9.6E-12
2.3E-12
2.4E-08
1.7E-08
2.2E-08
l.OE-08
1 4E-12
3.1E-08
4-25C
O.OE+00
3.4E-08
5.1E-08
2.2E-08
4.5E-08
5.7E-08
4.1E-08
6.1E-08
1.4E-12
6.9E-08
4.3E-08
2.6E-08
l.OE-12
3.2E-08
9.5E-08
1.1E-11
2.5E-12
2.4E-08
1.7E-08
2.2E-08
l.OE-08
1 4E-12
3.1E-08
4-25D
O.OE+00
3.4E-08
5.1E-08
2.2E-08
4.5E-08
5.7E-08
4.1E-08
6.1E-08
1.5E-12
6.9E-08
4.3E-08
2.6E-08
1.2E-12
3.2E-08
9.5E-08
1.1E-11
2.8E-12
2.4E-08
1.7E-08
2.2E-08
l.OE-08
1 6E-12
3.1E-08
Scenario Ratios
4-25B/
4-25A
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.4
1.0
1.0
1.0
1.5
1.0
1.0
1.3
1.5
1.0
1.0
1.0
1.0
1 4
1.0
4-25C/
4-25A
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.0
1.0
1.0
1.6
1.0
1.0
1.5
1.7
1.0
1.0
1.0
1.0
1 4
1.0
4-25D/
4-25A
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.6
1.0
1.0
1.0
1.8
1.0
1.0
1.6
1.9
1.0
1.0
1.0
1.0
1 6
1.0
a Results are presented for the first 22 days of the simulation period. Similar results were
observed after 22 days.
4.5.4 Effect of Source Parcel Size and Configuration
The effects of the size and configuration of the source parcel on benzo(a)pyrene
concentrations in the target parcel were evaluated with the four modeling configurations shown
in Figure 4-26. In modeling configuration 4-26A, the source is located in a 2 kilometer by 10
kilometer parcel. In modeling configuration 4-26B, the source is located in a smaller parcel
within a 2 kilometer by 10 kilometer parcel. This modeling configuration allows the model to
provide benzo(a)pyrene concentration estimates immediately adjacent to the source but in a
separate parcel. Modeling configuration 4-26C includes two concentric parcels within the 2
SEPTEMBER 2002
4-45
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
kilometer by 10 kilometer parcel. In modeling configuration 4-26D, the source parcel is
surrounded by four trapezoidal boundary parcels. Simulations for each modeling configuration
were run for a 180-day modeling period.
Among the four modeling configurations, there was very little difference between the
daily average benzo(a)pyrene concentrations in both the air and surface soil compartments. For
example, the 180-day average benzo(a)pyrene concentrations in air of the target parcels of
modeling configurations 4-26A and 4-26B were 2.4E-8 g/kg and 2.3E-8 g/kg, respectively. The
180-day average benzo(a)pyrene concentration surface soil of the target parcel was 1.3E-5 g/kg
for both modeling configurations 4-26A and 2-26B (see Figure 4-27). Concentrations in the air
and surface soil of the target parcels of modeling configurations 4-26C and 4-26D also were
slightly lower than for modeling configuration 4-26A. Thus, among the simple modeling
configurations used for this evaluation, the size and configuration of parcels did not appear to
significantly affect air and soil concentrations in a common target parcel.
Figure 4-26
Modeling Configuration Evaluation of Effects of Source Parcel Size
(A)
10km
I
\
i.
r
s
T
2km
(B)
(C)
N
t
SEPTEMBER 2002
4-46
TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 4
STRUCTURAL AND COMPLEXITY EVALUATION
Figure 4-27
Surface Soil Concentrations in the Target Parcels of
Modeling Configurations 4-26A and 4-26B
2.0E-05
- - 'Configuration 4-26B
9 Configuration 4-26A
O.OE+00
20 40 60
80 100
Time (days)
120
140
160
180
SEPTEMBER 2002
4-47
TRIM.FATE EVALUATION REPORT VOLUME I
-------
[This page intentionally left blank.]
-------
CHAPTER 5
REFERENCES FOR VOLUME I
5. REFERENCES FOR VOLUME I
Ambrose, R.A., Jr., T.A. Wool, and J.L. Martin. 1995. The Water Quality Analysis Simulation
Program, WASPS. Part A: Model Documentation. Athens, GA: U.S. EPA National Exposure
Research Laboratory, Ecosystems Division.
Arnold, J.R., R.L. Dennis, and G.S. Tonneson. 1998. Advanced techniques for evaluating
Eulerian air quality models: Background and methodology. Presented at the 10th Joint
Conference on the Applications of Air Pollution Meteorology with the A&WMA. Phoenix,
Arizona, January 11-16, 1998.
Barr, J.F. 1996. Aspects of common loon (Gavia immer) feeding biology on its breeding
ground. Hydrobiologia. 321:119-144.
Beck, M.B. and J. Chen. 2000. Assuring the quality of models designed for predictive tasks. In
Handbook of Sensitivity Analysis, eds. A. Saltelli, K. Chan, and E.M. Scott (401-420).
Chichester: John Wiley and Sons.
Beck, M.B. 1997. Water quality monitoring: A review of the analysis of uncertainty. Water
Resources Research. 23(8): 1393-1442.
Bennett, D.H., T.E. McKone, and R.G. Hetes. 2000a. Determining the appropriate level of
model complexity [abstract]. Presented at the 21st Annual Meeting of the Society of
Environmental Toxicology and Chemistry. Nashville, TN, November 12-16, 2000.
Bennett, D.H., D.F. Burch, R.M. Lee, B.F. Lyon, and D.L. Murphy. 2000b. Spatial resolution
of TRIM.FaTE, a fate, transport, and ecological exposure model for air pollutants [abstract].
Presented at the 21st Annual Meeting of the Society of Environmental Toxicology and
Chemistry. Nashville, TN, November 12-16, 2000.
Bennett, D.H., T.E. McKone, and W.E. Kastenberg. 2000. Characteristic time, characteristic
travel distance, and population based potential dose in a multimedia environment: a case study.
In Human and Ecological Risk Assessment: Theory and Practice, ed. D.J. Paustenbach. New
York, NY: John Wiley and Sons.
Bennett, D.H., T.E. McKone, M. Matthies, and W.E. Kastenberg. 1998. General formulation of
characteristic travel distance for semivolatile organic chemicals in a multimedia environment.
Environ. Sci. Technol. 32:4023-4030.
Bennett, D.H., T.E. McKone, and M.G. Dusetzina. 1997. Building uncertainty and sensitivity
analysis into the TRIM framework. Presented at the Annual Meeting of the Society of Risk
Analysis. Washington, DC, December 7-10, 1997.
Biester, H., and C. Scholz. 1997. Determination of mercury phases in contaminated soils:
mercury pyrolysis versus sequential extractions. Environ. Sci. Technol. 31:233-239.
SEPTEMBER 2002 5-1 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
Briggs, G.G., R.H. Bromilow, A.A. Evans, and M.R. Williams. 1983. Relationships between
lipophilicity and the distribution of non-ionized chemicals in barley shoots following uptake by
the roots. Pestic. Sci. 14:492-500.
Briggs, G.G., R.H. Bromilow, and A.A. Evans. 1982. Relationship between lipophilicity and
root uptake and translocation of non-ionized chemicals by barley. Pestic. Sci. 13:495-504.
Bull, K.R., R.D. Roberts, MJ. Inskip, and G.T. Goodman. 1977. Mercury concentrations in
soil, grass, earthworms and small mammals near an industrial emissions source. Environ. Pollut.
12:135-140.
Burch, D.F., R.M. Lee, B.F. Lyon, and D.L. Murphy. 2000. Model performance evaluation of
TRIM.FaTE, a fate, transport, and ecological exposure model for air pollutants [abstract].
Presented at the 21st Annual Meeting of the Society of Environmental Toxicology and
Chemistry. Nashville, TN, November 12-16, 2000.
Burken, J.G. and J.L. Schnoor. 1998. "Predictive Relationships for the Uptake of Organic
Contaminants by Hybrid Poplar Trees," Environmental Science and Technology, 32 (21)
3379-3385.
CalTOX. 1993. CalTOX, A Multimedia Total-Exposure Model for Hazardous-Waste Sites.
Background documentation prepared for California Environmental Protection Agency, the
Office of Scientific Affairs, Department of Toxic Substances Control. December 1993.
Chamberlain, A.C. 1970. Interception and retention of radioactive aerosols by vegetation.
Atmospheric Environment. 4:57-78.
Chen, J. and M.B. Beck. 1999. Quality Assurance of Multi-Media Model for Predictive
Screening Tasks. U.S. Environmental Protection Agency, Athens, GA. Publication No.
EPA/600/R-98/106.
Cohn, R.D. and R.L. Dennis. 1994. The evaluation of acid deposition models using principal
component spaces. Atmospheric Environment. 28(15):2513-2543.
Cole, J.G. and D. Mackay. 2000. Correlating environmental partitioning properties of organic
compounds: The three solubility approach. Environmental Toxicology and Chemistry. 19(2):
265-270.
Cousins, I. and D. Mackay. 2000. Correlating the physical-chemical properties of phthalate
esters using the 'three solubility' approach. Chemosphere. 41(9): 1389-1399.
Cowan, C.E., D. Mackay, T.C.J. Feijtel, D. van de Meent, A. DiGuardo, J. Davies, and N.
Mackay. 1995. The Multi-Media Fate Model: A Vital Tool for Predicting the Fate of
Chemicals. Pensacola, FL: SETAC Press.
Decisioneering. 1996. Crystal Ball Version 4.0 User Manual. Denver, CO: Decisioneering, Inc.
SEPTEMBER 2002 5-2 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
Dennis, R.L., W.R. Barchet, T.L. Clark, S.K. Seilkop, and P.M. Roth. 1990. Acid Deposition:
State of Science and Technology, Report 5, Evaluation of Regional Acidic Deposition Models
(Part I) and Selected Applications of RADM (Part II). National Acid Precipitation Assessment
Program.
DiGiulio, R.T., and E.A. Ryan. 1987. Mercury in soils, sediments and clams from a North
Carolina peatland. Water, Air, and Soil Pollution. 33:205-219.
Ebinghaus, R., S.G. Jennings, et al. 1999. International field intercomparison measurements of
atmospheric mercury species at Mace Head, Ireland. Atmospheric Environment. 33(18):
3063-3073.
Efroymson, R.A. and D.L. Murphy. 2001. Ecological risk assessment of multimedia hazardous
air pollutants: Estimating exposure and effects. Sci Total Environ. 274:219-230.
Efroymson, R.A., D.S. Jones, B.F. Lyon, D.H. Bennett, R.L. Maddalena, and D.L. Murphy.
2000. Biotic compartments in TRIM.FaTE, a fate, transport, and ecological exposure model for
air pollutants, [abstract] Presented at the 21st Annual Meeting of the Society of Environmental
Toxicology and Chemistry. Nashville, TN, November 12-16, 2000.
Efroymson, R.A., D.S. Jones, and A. Vasu. 1999. An ecological risk assessment methodology
for hazardous air pollutants [abstract]. Presented at the 20th Annual Meeting of the Society of
Environmental Toxicology and Chemistry. Philadelphia, PA, November 14-18, 1999.
Efroymson, R., B. Sample, C. Hunsaker, B. Lyon, A. Simcock, and G. Suter. 1997. A dynamic
model for terrestrial ecological exposure to toxic air pollutants. Presented at the 18th Annual
Meeting of the Society of Environmental Toxicology and Chemistry. San Francisco, CA,
November 16-20, 1997.
Eisenberg, N., M. Federline, B. Sagar, G. Wittmeyer, J. Andersson, and S. Wingefors. 1995.
Model validation from a regulatory perspective: A summary. In GEOVAL '94: Validation
through model testing, OECD documents, safety assessment of radioactive waste repositories,
Nuclear Energy Agency, Proceedings of an NEA/SKI Symposium. Paris, France, October 11-
14, 1994.
EPRI. 1998. Mercury flux measurements: an intercomparison and assessment: Nevada
mercury emissions project. Report TR-111346. Palo Alto, CA.
Fine, S.S., A.M. Eyth, B.F. Lyon, and T. Palma. 2000. The TRIM computer framework
[abstract]. Presented at the 21st Annual Meeting of the Society of Environmental Toxicology and
Chemistry. Nashville, TN, November 12-16, 2000.
Gnamus, A., A. R. Byrne, and M. Horvat. 2000. Mercury in the soil-plant-deer-predator food
chain of a temperate forest in Slovenia. Environ. Sci. Technol. 34:3337-3345.
Gonzalez, H. 1991. Mercury pollution caused by a chloralkali plant. Water, Air, Soil Pollut.
56:83-93.
SEPTEMBER 2002 5-3 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
Guha, S., S. Forbes, I.E. McKone, D.H. Bennett, and B.F. Lyon. 1997. A generalized, mass-
conserving multimedia transport and transformation model: Development of transfer factors.
Presented at the Annual Meeting of the Society of Risk Analysis, Washington, DC, December 7-
10, 1997.
Habicht, F.H. 1992. Memorandum from the Deputy Administrator of the U.S. EPA. U.S. EPA,
Office of the Administrator, Task Force on Environmental Regulatory Modeling. Washington,
DC, March 7.
Hetes, R.G. and I.E. Langstaff. 2000. Estimation of uncertainty and variability within the total
risk integrated methodology (TRIM) [abstract]. Presented at the 21st Annual Meeting of the
Society of Environmental Toxicology and Chemistry. Nashville, TN, November 12-16, 2000.
Hodges, J.S. and J.A. Dewar. Is It You or Your Model Talking? A Framework for Model
Validation. RAND Corporation, Santa Monica, CA, 1992. ISBN: 0-8330-1223-1.
Hoyer, M., J. Burke, et al. 1995. Atmospheric sources, transport and deposition of mercury in
Michigan: Two years of event precipitation. Water Air and Soil Pollution. 80(1-4): 199-208.
ICF Consulting. 2002. TRIM.FaTE Algorithm and Compartment Audit. Prepared by ICF
Consulting for U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, October 2002.
Johnson, T., G.W. Suter, and T. Palma. 1997. TRIM.FaTE: A TRIM module for linking
multimedia environmental systems with biotic domains. Presented at the Annual Meeting of the
Society of Risk Analysis. Washington, DC, December 7-10, 1997.
Kelso, J.R.M. and M.G. Johnson. 1991. Factors related to the biomass and production offish
communities in small, oligotrophic lakes vulnerable to acidification. Canadian Journal of
Fisheries and Aquatic Sciences. 48:2523-2532.
Konikow, L.F. and J.D. Bredehoeft. 1992. Ground-water models cannot be validated. Adv.
Water Resour. 15(l):75-83.
Lin, C.J. and S.O. Pehkonen. 1999. Aqueous phase reactions of mercury with free radicals and
chlorine: Implications for atmospheric mercury chemistry. Chemosphere. 38(6):1253-1263.
Lindberg, S.E. and W.J. Stratton. 1998. Atmospheric mercury speciation: Concentrations and
behavior of reactive gaseous mercury in ambient air. Environmental Science & Technology.
32(l):49-57.
Lindberg, S.E., T.P. Meyers, et al. 1992. Atmosphere-surface exchange of mercury in a forest:
Results of modeling and gradient approaches. Journal of Geophysical Research-Atmospheres.
97(D2):2519-2528.
SEPTEMBER 2002 5-4 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
Lindqvst, O., K. Johansson, M. Aastrup, A. Andersson, L. Bringmark, G. Hovsenius, L.
Hakanson, A. Iverfeldt, M. Meili, and B. Timm. 1991. Mercury in the Swedish environment:
recent research on causes, consequences and corrective methods. Water, Air, and Soil Pollution.
Lu, J.Y. and W.H. Schroeder. 1999. Sampling and determination of parti culate mercury in
ambient air: A review. Water, Air, and Soil Pollution. 112(3-4):279-295.
Lyon, B.F., T.E. McKone, and D.H. Bennett. 2000. Conceptual design and mass balance
framework for TRIM.FaTE [abstract]. Presented at the 21st Annual Meeting of the Society of
Environmental Toxicology and Chemistry. Nashville, TN, November 12-16, 2000.
Lyon, B.F., S. Guha, and T.E. McKone. 1997. TRIM: Mathematical and numerical aspects.
Presented at the Annual Meeting of the Society of Risk Analysis. Washington, DC, December 7-
10, 1997.
Mackay, K. 1991. Multimedia Environmental Models: The Fugacity Approach. Chelsea, MI:
Lewis Publishers.
Maddalena, R.L., T.E. McKone, D.P.H. Hsieh, and S. Geng. 2001. Influential input
classification in probabilistic multimedia models. Stochastic Environmental Research and Risk
Assessment. 15(1): 1-17.
Maddalena, R.L., D. H. Bennett, D. Murphy, and R.G. Hetes. 2000. Evaluation of a complex
environmental model: the TRIM.FaTE example [abstract]. Presented at the 21st Annual Meeting
of the Society of Environmental Toxicology and Chemistry. Nashville, TN, November 12-16,
2000.
McKone, T.E., S. Guha, T. Johnson, B. Lyon, G. Suter, and A. Vasu. 1997a. A multimedia
health and ecological risk assessment methodology for air pollutants. Presented at the 18th
Annual Meeting of the Society of Environmental Toxicology and Chemistry. San Francisco, CA,
November 16-20, 1997.
McKone, T.E., T. Johnson, G.W. Suter II. 1997b. Estimating multi-pathway human and
ecosystem exposures within the total risk integrated model (TRIM). Presented at the Annual
Meeting of the Society of Risk Analysis. Washington, DC, December 7-10, 1997.
McKone, T. E. 1993. CalTOX, A Multimedia Total -Exposure Model for Hazardous-Wastes
Sites Part II: The Dynamic Multimedia Transport and Transformation Model. Livermore, CA,
prepared for the State of California, Department of Toxic Substances Control, Lawrence
Livermore National Laboratory.
McKone, T.E. 1993a. CalTOX, A multimedia total-exposure model for hazardous-wastes sites
Parti: Executive Summary. Laboratory. UCRL-CR-111456PtI. Livermore, CA: Lawrence
Livermore National Laboratory.
SEPTEMBER 2002 5-5 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
McKone, I.E. 1993b. CalTOX, A multimedia total-exposure model for hazardous-wastes sites
Part II: The dynamic multimedia transport and transformation model. UCRL-CR-111456PtII.
Livermore, CA: Lawrence Livermore National Laboratory.
McKone, T.E. 1993c. CalTOX, A multimedia total-exposure model for hazardous-wastes sites
Part III: The multiple-pathway exposure model. UCRL-CR-111456PtIII. Livermore, CA:
Lawrence Livermore National Laboratory.
Muller, H., and G., Prohl. 1993. Ecosys-87: A dynamic model for assessing radiological
consequences of nuclear accidents. Health Phys. 64:232-252.
Murphy, D.L., R.L. Maddalena, J. Langstaff, D.S. Jones, L.M. Lee, A.M. Eyth, B.F. Lyon, M.
McVey, and G. Laniak. 2000. Evaluation of the TRIM.FaTE multi-media model: a mercury
case study [abstract]. Presented at the Joint Meeting of the Society of Exposure and Analysis and
the International Society of Environmental Epidemiology. Vancouver, Canada, August 11-15,
2002.
Murphy, D.L., T. Palma, R.G. Hetes, H.M. Richmond, and A.B. Vasu. 2000. Total risk
integrated methodology (TRIM): focus on the fate, transport and ecological exposure module
(TRIM.FaTE) [abstract]. Presented at the 21st Annual Meeting of the Society of Environmental
Toxicology and Chemistry. Nashville, TN, November 12-16, 2000.
Nobel, Park S. 1999. Physicochemical and Environmental Plant Physiology, 2nd edition,
Academic Press.
Oreskes, N., K. Shrader-Frechette, K. Belitz. 1994. Verification, validation, and confirmation of
numerical models in the earth sciences. Science. 263:641-646.
Pai, P., S. Heisler, et al. 1998. An emissions inventory for regional atmospheric modeling of
mercury. Water, Air, and Soil Pollution. 101(l-4):289-308.
Palma, T., A. Vasu, and R. Hetes. 1999. Total risk integrated methodology (TRIM). Air and
Waste Management Association - EM Magazine. March 1999,30-34.
Park, R. A. 1998. AquaTOX for Windows: A modular toxic effects model for aquatic
ecosystems. Draft. EcoModeling, Montgomery Village, MD.
Pleijel, K. and J. Munthe. 1995a. Modeling the atmospheric chemistry of mercury: The
importance of a detailed description of the chemistry of cloud water. Water, Air, and Soil
Pollution. 80(l-4):317-324.
Pleijel, K. and J. Munthe. 1995b. Modeling the atmospheric mercury cycle: Chemistry in fog
droplets. Atmospheric Environment. 29(12):1441-1457.
Radhakrishnan, K. and A.C. Hindmarsh. 1993. Description and Use of LSODE, the Livermore
Solver for Ordinary Differential Equations. LLNL report UCRL-ID-113855, December 1993.
SEPTEMBER 2002 5-6 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
Reiderer, M. 1995. Partitioning and transport of organic chemicals between the atmospheric
environment and leaves. In: Plant Contamination: Modeling and Simulation of Organic
Chemical Processes, eds. S. Trapp and J.C. McFarlane (153-190). Boca Raton: Lewis
Publishers.
Revis, N.W. et al. 1989. Distribution of mercury species. Water, Air and Soil Pollution.
45(1-2): 105-113.
Schatzmann, M., S. Rafailidis, and M. Pavageau. 1997. Some remarks on the validation of
small-scale dispersion models with field and laboratory data. Journal of Wind Engineering and
Industrial Aerodynamics. 67(8):885-893.
Schroeder, W.H., J. Munthe, et al. 1989. Cycling of mercury between water, air, and soil
compartments of the environment. Water, Air, and Soil Pollution. 48(3-4):337-347.
Schwarzenbach, R.P., P.M. Gschwend, and D.M. Imboden. 1993. Environmental Organic
Chemistry. New York, NY: John Wiley and Sons.
Spear, C.R. 1997. Large simulation models: Calibration, uniqueness and goodness of fit.
Environmental Modeling and Software. 12:219-228.
Talmage, S.S. and B. T. Walton. 1993. Food chain transfer and potential renal toxicity of
mercury to small mammals at a contaminated terrestrial field site. Ecotoxicol. 2:243-256.
Taylor, A.C. 1993. Using objective and subjective information to develop distributions for
probabilistic exposure assessment. Journal of Exposure Analysis and Environmental
Epidemiology. 3:285-298.
U.S. EPA. 2002a. U.S. Environmental Protection Agency. TREVI.FaTE Technical Support
Document Volume I: Description of Module. EPA-453/R-02-011a. Research Triangle Park,
NC: Office of Air Quality Planning and Standards. September.
U.S. EPA. 2002b. U.S. Environmental Protection Agency. TREVI.FaTE Technical Support
Document Volume II: Description of Chemical Transport and Transformation Algorithms. EPA-
453/R-02-01 Ib. Research Triangle Park, NC: Office of Air Quality Planning and Standards.
September.
U.S. EPA. 2000. U.S. Environmental Protection Agency. An SAB Advisory on the Agency's
"Total Risk Integrated Methodology (TRIM)." EPA-SAB-EC-ADV-00-004. Washington, DC:
Science Advisory Board. May.
U.S. EPA. 1999a. U.S. Environmental Protection Agency. Total Risk Integrated Methodology
Status Report. EPA-453/R-99-010. Research Triangle Park, NC: Office of Air Quality Planning
and Standards. November.
SEPTEMBER 2002 5-7 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
U.S. EPA. 1999b. U.S. Environmental Protection Agency. TREVI.Expo Technical Support
Document. External Review Draft. EPA-453/D-99-001. Research Triangle Park, NC: Office of
Air Quality Planning and Standards. November.
U.S. EPA. 1999c. U.S. Environmental Protection Agency. TRIM.FaTE Technical Support
Document Volume I: Description of Module. External Review Draft. EPA-453/D-99-002A.
Research Triangle Park, NC: Office of Air Quality Planning and Standards. November.
U.S. EPA. 1999d. U.S. Environmental Protection Agency. TREVI.FaTE Technical Support
Document Volume II: Description of Chemical Transport and Transformation Algorithms.
External Review Draft. EPA-453/D-99-002B. Research Triangle Park, NC: Office of Air Quality
Planning and Standards. November.
U.S. EPA. 1999e. U.S. Environmental Protection Agency. National air toxic program: the
integrated urban strategy. Federal register 64: 38705-38740. July 19.
U.S. EPA. 1998a. U.S. Environmental Protection Agency. An SAB Advisory on the
TREVl.FaTE Moesl of the Total Risk Integrated Methodology. EPA-SAB-EC-ADV-99-003.
Washington, DC: Science Advisory Board. December.
U.S. EPA. 1998b. U.S. Environmental Protection Agency. Total Risk Integrated Methodology,
Implementation of the TRIM Conceptual Design Through the TRIM.FaTE Module, A Status
Report. EPA-452/R-98-001. Research Triangle Park, NC: Office of Air Quality Planning and
Standards. March.
U.S. EPA. 1998c. U.S. Environmental Protection Agency. Study of hazardous air pollutants
from electric utility steam generating units-final report to congress. EPA 453/R-989-004a. Office
of Air Quality Planning and Standards. February.
U.S. EPA. 1998d. U.S. Environmental Protection Agency. The total risk integrated
methodology: technical support document for the TRIM.FaTE module. Draft. EPA-452/D-98-
001. Office of Air Quality Planning and Standards.
U.S. EPA. 1998e. U.S. Environmental Protection Agency. Methodology for assessing health
risks associated with multiple exposure pathways to combustor emissions. External Review
Draft. Update to EPA/600/6-90/003. NCEA-C-0238. National Center for Environmental
Assessment.
U.S. EPA. 1998f. White Paper on the Nature and Scope of Issues on Adoption of Model Use
and Acceptability Guidance. External Review Draft. Washington, DC: Science Policy Council.
U.S. EPA. 1997. U.S. Environmental Protection Agency. Mercury study report to congress
(Volumes I-VIII). EPA-452/R-97-005. Office of Air Quality Planning and Standards.
U.S. EPA. 1997b. U.S. Environmental Protection Agency. Mercury study report to congress.
Volume V: health effects of mercury and mercury compounds. U.S. EPA Office of Air Quality
Planning and Standards, and Office of Research and Development.
SEPTEMBER 2002 5^8 TRIM.FATE EVALUATION REPORT VOLUME I
-------
CHAPTER 5
REFERENCES FOR VOLUME I
U.S. EPA. 1994a. U.S. Environmental Protection Agency. Report of the agency task force on
environmental regulatory modeling. Guidance, support needs, draft criteria and charter. EPA
500-R-94-001. Washington, DC: Office of Solid Waste and Emergency Response.
U.S. EPA. 1994b. U.S. Environmental Protection Agency. Review of draft "addendum to the
methodology for assessing health risk associated with indirect exposure to combustor
emissions". EPA-SAB-1AQC-94-009B. Washington, DC: Science Advisory Board.
U.S. EPA. 1989. Resolution on the Use of Mathematical Models by EPA for Regulatory
Assessment and Decision Making. EPA-SAB-EEC-89-012. Washington, DC, Science Advisory
Board, Environmental Engineering Committee.
Vasu, A.B., R.G. Hetes, T. Palma, and I.E. McKone. 1998. TRIM: a multimedia, multipathway
framework for assessing human and ecological exposure and risk. Presented at the 19th Annual
Meeting of the Society of Environmental Toxicology and Chemistry. Charlotte, NC, November
15-19, 1998.
Vasu, A., M. Dusetzina, R. Hetes, T. Palma, H. Richmond, T. McKone, and T. Johnson. 1997.
Introduction to the total risk integrated methodology (TRIM). Presented at the Annual Meeting
of the Society of Risk Analysis. Washington, DC, December 7-10, 1997.
Wagrowski, D.M., and R.A. Kites. 1997. Polycyclic aromatic hydrocarbon accumulation in
urban, suburban, and rural vegetation. Environ. Sci. Technol, 31:279-282.
Wang, W. and C.T. Driscoll. 1995. Patterns of total mercury concentrations in Onondaga Lake,
New York. Environ. Sci. Technol. 29:2261-2266.
Watras, C.J., K.A. Morrison, J.S. Host, and N.S. Bloom. 1995. Concentration of mercury
species in relationship to other site-specific factors in the surface waters of northern Wisconsin
lakes. Limnol. Oceanogr. 40:556-565.
Wren, C.D., H.R. Maccrimmon, and B.R. Loescher. 1983. Examination of bioaccumulation and
biomagnification of metals in aPrecambrian shield lake. Water, Air Soil Pollut. 19:277-291.
Xiao, Z., J. Sommar, et al. 1998. Atmospheric mercury deposition to grass in southern Sweden.
Science of the Total Environment. 213(l-3):85-94.
Zimmer, R., J. Dee, D. Jones, B. Sample, and G. Suter. 1997. Use of EcoFaTE in modeling
pollutant mass transfers and subsequent risks within aquatic systems. Presented at the 18th
Annual Meeting of the Society of Environmental Toxicology and Chemistry. San Francisco, CA,
November 16-20, 1997.
SEPTEMBER 2002 5-9 TRIM.FATE EVALUATION REPORT VOLUME I
-------
[This page intentionally left blank.]
-------
APPENDIX I-A
TRIM.FaTE ALGORITHM PAIRING TABLES
-------
[This page intentionally left blank.]
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
APPENDIX I-A
TRIM.FaTE ALGORITHM PAIRING TABLES3
PAIRED ALGORITHMS
(TRANSFER FROM ONE COMPARTMENT TO ANOTHER AND THE REVERSE TRANSFER)
ALGORITHMS FOR EXCHANGES BETWEEN COMPARTMENTS OF THE SAME TYPE
1. Algorithm: Advection from Air to Air [all chemicals] [TF 3-l]b
148. Algorithm: Runoff from Surface Soil to Surface Soil [all
chemicals] (occurs in the down-gradient direction only) [TF 5-10a]
50. Algorithm: Erosion from Surface Soil to Surface Soil [all
chemicals] (occurs in the down-gradient direction only) [TF 5-1 la]
129. Algorithm: Percolation from Root Zone to Root Zone [all
chemicals] [TF 5-9]
133. Algorithm: Percolation from Vadose Zone to Vadose Zone [all
chemicals] [TF 5-9]
44. Algorithm: Dispersive Waterflow from Surface Water to Surface
Water, General [all chemicals] [TF 4-12]
169. Algorithm: Waterflow from Surface Water to Surface Water [all
chemicals] (occurs in the downgradient direction only) [TF 4-9]
134. Algorithm: Pore Water Diffusion from Sediment to Sediment [all
chemicals] [TF 4-13]
To reverse the direction of the flow, the receiving and sending compartments
are switched (with the exception of erosion of surface soil to surface soil, or
runoff from surface soil to surface soil, for which the processes occur in the
down-gradient direction only).
Diffusion of contaminants between soil compartments of the same type occurs
only in the root-zone soil and vadose-zone soil compartments.
SEPTEMBER 2002
I-A-1
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
TWO-WAY TRANSFERS
EXCHANGES BETWEEN AIR AND PLANT LEAVES, SURFACE SOIL, AND SURFACE WATER
Diffusive Transfers
31. Algorithm: Diffusion from Air to Plant Leaf [all chemicals except
divalent mercury] [TF 7-9a]
32. Algorithm: Diffusion from Air to Surface Soil [all chemicals except
divalent mercury] [TF 5-1]
33. Algorithm: Diffusion from Air to Surface Water, Two Film [all
chemicals, except divalent mercury] [TF 3-2]
34. Algorithm: Diffusion from Plant Leaf to Air, Default [all chemicals,
except divalent mercury] [TF 7-8a]
38. Algorithm: Diffusion from Surface Soil to Air [all chemicals, except
divalent mercury] [TF 5-2]
40. Algorithm: Diffusion from Surface Water to Air, Two Film [all
chemicals, except divalent mercury] [TF 3-3]
Air Deposition of Vapors and Particles and Resuspension/Washoff of Same
On and Off Plant Leaves
172. Algorithm: Wet deposition of vapor phase to Plant Leaf from Air,
Organics [TF 7-7b] (The remainder is wet deposited to the soil, #172b, a
new algorithm)
173. Algorithm: Wet deposition of vapor phase to Plant Leaf from Air
[divalent and elemental mercury only] [TF 7-7a] (The remainder is wet
deposited to the soil, #174)
46b. [NEW] Algorithm: Dry deposition of vapor from air to plant leaf
[= net deposition; divalent mercury only] [TF A-3]
48b. Algorithm: Dry Deposition of particles to plant leaf [, i.e., to leaf
particle compartment] [new] [all chemicals] [TF 7-1] (the remainder is
dry deposited on the soil, #48)
177b. Algorithm: Wet deposition of particles to plant leaf [i.e., to leaf
particle compartment] [new] [all chemicals] [TF 7-3] (the remainder is
wet deposited to the soil, #177)
164. Algorithm: Transfer from Leaf Particle on surface to leaf [all
chemicals] [TF 7-5]
Algorithm # 34 [TF 7-8] (diffusion algorithm, see above) is the only instance of
the reverse transfer, from plant leaf to air for the vapor phase.
126. Algorithm: Particles Blown off from Plant Leaf to Air (DRY) [all
chemicals] [TF 7-2]
128. Algorithm: Particles washed off leaf onto ground [all chemicals] [TF 7-
4]
165. Algorithm: Transfer from Leaf to Leaf Particle on surface [all
chemicals] [TF 7-6]
SEPTEMBER 2002
I-A-2
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
Between Air and Soil
172b. Algorithm: Wet deposition of vapor phase to soil for organics
[new] [TF 5-5a] (Some is intercepted by the plant leaves, #172.)
174. Algorithm: Wet deposition of vapor phase to Soil [elemental and
divalent mercury] [TF 5-5b] (Some is intercepted by the plant leaves,
#173.)
46. Algorithm: Dry deposition of vapor from air to surface soil
[divalent mercury] [TF A-2] (Some is intercepted by plant leaves, #46b)
48. Algorithm: Dry Deposition to soil of particles [all chemicals] [TF 5-
3] (Some of the dry deposition is intercepted by plants, #48b.)
177. Algorithm: Wet deposition to soil of particles [all chemicals] [TF
5-4] (Some of the wet deposition of particles is intercepted by plants, #
177b.)
Algorithm # 38 [TF 5-2] (diffusion algorithm, see above) is the only instance of
the reverse transfer, from soil to air for the vapor phase.
147. Algorithm: Resuspension from Surface Soil to Air, Set to Deposition
rate of particles [all chemicals] [TF 5-6]
Between Air and Surf ace Water
111. Algorithm: Wet Deposition of Vapor from Air to Surface Water
Organics [TF 4-3a]
175. Algorithm: Wet deposition of vapor phase to Surface water
[elemental and divalent mercury] [TF 4-3b]
47. Algorithm: Dry deposition of vapor from air to surface water [=
net deposition; divalent mercury only] [TF A-l]
45. Algorithm: Dry Deposition of Particles from Air to Surface Water
[all chemicals] [TF 4- Ib]
170. Algorithm: Wet Deposition of Particles from Air to Surface
Water [all chemicals] [TF 4-2b]
Algorithm # 40 (diffusion algorithm, see above) is the only instance of the
reverse transfer, from surface water to air for the vapor phase.
There is no "reverse" process. Particles in the surface water can be deposited to
sediments (#151) and resuspended into the surface water (#146).
SEPTEMBER 2002
I-A-3
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
EXCHANGES BETWEEN SOIL LAYERS AND SOIL BIOTA (ROOTS AND SOIL INVERTEBRATES)
Vertical Exchanges between Soil Compartments
43a. Algorithm: Diffusion downward from higher Vadose Zone to
lower Vadose Zone [all chemicals] [TF 5-7]
43c. Algorithm: Diffusion downward from higher Root-zone Soil to
lower Root-zone Soil [all chemicals] [TF 5-7]
43b. Algorithm: Diffusion upward from lower Vadose Zone to higher
Vadose Zone [all chemicals] [TF 5-8]
43d. Algorithm: Diffusion upward from lower Root-zone Soil to higher
Root-zone Soil [all chemicals] [TF 5-8]
Between Soil and Soil Biota
158. Algorithm: Time-dependent partition from root zone [soil] to
root. Interacts with bulk soil [for all three mercury species] [TF 7-10a]
159. Algorithm: Time-dependent partition from root zone [soil] to
root, Interacts with soil pore water [for organics only] [TF 7-1 Ob]
167. Algorithm: Transfer from root zone bulk soil to stem [for all
chemicals] [TF 7-12a]
167b. Algorithm: Transfer from root-zone soil pore water to stem, for
organic chemicals [TF 7-12bl] [not in current library]
157. Algorithm: Time-dependent partition from Root Zone to
arthropod [bulk soil] [all chemicals] [TF 7-19]
160. Algorithm: Time-dependent partition from Root Zone to Worm,
Interacts with bulk soil [all mercury species] [TF 7-17a]
160b. Algoirthm: Time-dependent partition from Root Zone to Worm,
Interacts with soil pore water [organics] [TF 7-17b]
155. Algorithm: Time-dependent partition from root to root zone [soil],
Interacts with bulk soil [for all three mercury species]
156. Algorithm: Time-dependent partition from root to root zone [soil],
Interacts with soil pore water [for organics only]
No match for transfer in the reverse direction.
167c. Algorithm: Transfer from stem root-zone soil pore water, for organic
chemicals] [TF 7-12b2] [not in current library]
152. Algorithm: Time-dependent partition from arthropod to Root Zone [
bulk soil] [all chemicals] [TF 7-20]
163. Algorithm: Time-dependent partition from worm to Root Zone,
Interacts with bulk soil [all mercury species] [TF 7-18a]
163b. Algoirthm: Time-dependent partition from Worm to Root Zone,
Interacts with soil pore water [organics] [TF 7-18b]
SEPTEMBER 2002
I-A-4
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
Between Soil Compartments, Diffusion, Percolation to Ground Water, and GW Recharge to Surface Water
39. Algorithm: Diffusion from Surface Soil to Root Zone [all
chemicals] [TF 5-7]
36. Algorithm: Diffusion from Root Zone to Vadose Zone [all
chemicals] [TF 5-7]
131. Algorithm: Percolation from Surface Soil to Root Zone [all
chemicals] [TF 5-9]
130. Algorithm: Percolation from Root Zone to Vadose Zone [all
chemicals] [TF 5-9]
132. Algorithm: Percolation from Vadose Zone to GW [all chemicals]
[TF 5-12]
35. Algorithm: Diffusion from Root Zone to Surface Soil [all chemicals]
[TF 5-8]
42. Algorithm: Diffusion from Vadose Zone to Root Zone [all chemicals]
[TF 5-8]
Percolation is a one-way process (down). Diffusion works both up and down
through the soil column.
135. Algorithm: Recharge from GW to Surface Water [TF 5-13]
EXCHANGES BETWEEN SURFACE WATER, SEDIMENTS, AND BIOTA
151. Algorithm: Sediment deposition from surface water to sediment
[all chemicals] [TF 4-4]
2. Algorithm: Algae deposition from surface water to sediment [all
chemicals] [TF 4-6]
37. Algorithm: Diffusion from Sediment to Surface Water,
Fugacity-based [all chemicals] [TF 4-10]
49. Algorithm: Elimination from fish to surface water [all three
mercury species] [TF A-4]
56. Algorithm: Exchange from fish to surface water, organics [TF 6-8]
53. Algorithm: Exchange from macrophyte to surface water [organics
only] [TF 6-2a]
153. Algorithm: Time-dependent Partition from macrophyte to
surface water [all three mercury species] [TF 6-2b]
54. Algorithm: Exchange from benthic invertebrate to sediment [pore
water] [organics only] [TF 6-4a]
146. Algorithm: Resuspension from Sediment to Surface Water [all
chemicals] [TF 4-5]
41. Algorithm: Diffusion from Surface Water to Sediment, Fugacity-based
[all chemicals] [TF4-11]
No significant uptake of mercury directly from surface water by fish.
52. Algorithm: Exchange from surface water to fish, organics [TF 6-5]
57. Algorithm: Exchange from surface water to macrophyte [organics only]
[TF6-la]
162. Algorithm: Time-dependent Partition from surface water to
macrophyte [all three mercury species] [TF 6-lb]
55. Algorithm: Exchange from sediment to benthic invertebrate [pore
water] [organics only][TF 6-3a]
SEPTEMBER 2002
I-A-5
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
154. Algorithm: Time-dependent Partition from benthic invertebrate
to sediment [all three mercury species only] (TSD Equation 7-106) [TF
6-4b]
161. Algorithm: Time-dependent Partition from sediment to benthic
invertebrate [all three mercury species only] (TSD Equation 7-107) [TF 6-3b]
DEMETHYLATION AND METHYLATION REACTIONS FOR MERCURY
21. Algorithm: Demethylation (MHg->Hg2) rate in air [TF A-8]
22. Algorithm: Demethylation (MHg->Hg2) in birds [TF A-8]
25. Algorithm: Demethylation (MHg->Hg2) in mammals [TF A-8]
20. Algorithm: Demethylation(Hg2 -> MHg) in plant [leaf] [TF A-8]
20. Algorithm: Demethylation(Hg2 -> MHg) in plant [stem] [TF A-8]
No apparent demethylation in plant roots.
28. Algorithm: Demethylation(MHg -> Hg2) in surface soil [TF A-8]
26. Algorithm: Demethylation(MHg -> Hg2) in root zone [TF A-8]
30. Algorithm: Demethylation(MHg -> Hg2) in vadose zone [TF A-8]
24. Algorithm: Demethylation(MHg -> Hg2) in GW [TF A-8]
29. Algorithm: Demethylation(MHg -> Hg2) in surface water [TF A-8]
23. Algorithm: Demethylation (MHg->Hg2) in fish [TF A-8]
27. Algorithm: Demethylation(MHg -> Hg2) in sediment [TF A-8]
Demethylation in benthic invertebrates not included owing to lack of data.
104. Algorithm: Methylation(Hg2 -> MHg) in air [TF A-7]
105. Algorithm: Methylation (Hg2->MHg) in birds [TF A-7]
108. Algorithm: Methylation (Hg2->MHg) in mammals [TF A-7]
109. Algorithm: Methylation(Hg2 -> MHg) in plant leaves [TF A-7]
110. Algorithm: Methylation(Hg2 -> MHg) in plant stem [TF A-7]
No apparent methylation in plant roots.
113. Algorithm: Methylation(Hg2 -> MHg) in surface soil [TF A-7]
111. Algorithm: Methylation(Hg2 -> MHg) in root zone [TF A-7]
115. Algorithm: Methylation(Hg2 -> MHg) in vadose zone [TF A-7]
107. Algorithm: Methylation(Hg2 -> MHg) in GW [TF A-7]
114. Algorithm: Methylation(Hg2 -> MHg) in surface water [TF A-7]
106. Algorithm: Methylation (Hg2->MHg) in fish [TF A-7]
112. Algorithm: Methylation(Hg2 -> MHg) in sediment [TF A-7]
No apparent methylation in benthic invertebrates.
OXIDATION AND REDUCTION REACTIONS FOR MERCURY
116. Algorithm: Oxidation(HgO -> Hg2) in air [TF A-6]
117. Algorithm: Oxidation(HgO -> Hg2) in birds [TF A-6]
120. Algorithm: Oxidation(HgO -> Hg2) in mammals [TF A-6]
No apparent oxidation in plant leaves, stems, or roots.
123. Algorithm: Oxidation(HgO -> Hg2) in surface soil [TF A-6]
136. Algorithm: Reduction(Hg2 -> HgO) in air [TF A-5]
137. Algorithm: Reduction(Hg2 -> HgO) in birds [TF A-5]
140. Algorithm: Reduction(Hg2 -> HgO) in mammals [TF A-5]
No apparent reduction in plant leaves, stems, or roots.
143. Algorithm: Reduction(Hg2 -> HgO) in surface soil [TF A-5]
SEPTEMBER 2002
I-A-6
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
121. Algorithm: Oxidation(HgO -> Hg2) in root zone [TF A-6]
125. Algorithm: Oxidation(HgO -> Hg2) in vadose zone [TF A-6]
119. Algorithm: Oxidation(HgO -> Hg2) in GW [TF A-6]
124. Algorithm: Oxidation(HgO -> Hg2) in surface water [TF A-6]
new. Algorithm: Oxidation(HgO -> Hg2) in macrophytes [TF A-6]
118. Algorithm: Oxidation(HgO -> Hg2) in fish [TF A-6]
122. Algorithm: Oxidation(HgO -> Hg2) in sediment [TF A-6]
Oxidation in benthic invertebrates not included due to lack of data.
141. Algorithm: Reduction(Hg2 -> HgO) in root zone [TF A-5]
145. Algorithm: Reduction(Hg2 -> HgO) in vadose zone [TF A-5]
139. Algorithm: Reduction(Hg2 -> HgO) in GW [TF A-5]
144. Algorithm: Reduction(Hg2 -> HgO) in surface water [TF A-5]
No apparent reduction in macrophytes.
138. Algorithm: Reduction(Hg2 -> HgO) in fish [TF A-5]
142. Algorithm: Reduction(Hg2 -> HgO) in sediment [TF A-5]
No apparent reduction in benthic invertebrates.
MOVEMENT WITHIN PLANTS
166. Algorithm: Transfer from leaf to stem [all chemicals] [TF 7-13]
No movement from root to stem, just from soil to stem as a short-cut.
168. Algorithm: Transfer from stem to leaf [all chemicals] [TF 7-14]
No movement from stem to root.
SEPTEMBER 2002
I-A-7
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
ONE-WAY TRANSFERS
DEGRADATION/REACTION SINKS
[These Transfer Factors are not numbered in TSD II]
5. Algorithm: Degradation/Reaction sink in Air [organic chemicals]
[TF 2-1]
6. Algorithm: Degradation/Reaction sink in Birds [organic chemicals]
[TF 2-1]
10. Algorithm: Degradation/Reaction sink in Mammals [organic
chemicals] [TF 2-1]
9. Algorithm: Degradation/Reaction sink in Leaf [organic chemicals]
[TF 2-1]
15. Algorithm: Degradation/Reaction sink in Stem [organic chemicals]
[TF 2-1]
12. Algorithm: Degradation/Reaction sink in Root [organic chemicals]
[TF 2-1]
19. Algorithm: Degradation/Reaction sink in Worm [organic
chemicals] [TF 2-1]
Note: No degradation/reaction sink in soil arthropods.
16. Algorithm: Degradation/Reaction sink in Surface Soil [organic
chemicals] [TF 2-1]
13. Algorithm: Degradation/Reaction sink in Root Zone [organic
chemicals] [TF 2-1]
18. Algorithm: Degradation/Reaction sink in Vadose Zone [organic
chemicals] [TF 2-1]
8. Algorithm: Degradation/Reaction sink in Groundwater [organic
chemicals] [TF 2-1]
17. Algorithm: Degradation/Reaction sink in Surface Water [organic
chemicals] [TF 2-1]
RUNOFF AND EROSION FROM SOIL TO SURFACE WATER
149. Algorithm: Runoff from Surface Soil to Surface Water [all chemicals]
[TF 5-10b]
51. Algorithm: Erosion from Surface Soil to Surface Water [all chemicals]
[TF5-llb]
TERRESTRIAL ANIMAL EXCRETION TO SOIL OR SW
58. Algorithm: First-order excretion to soil, birds [all chemicals] [TF 7-31]
59. Algorithm: First-order excretion to soil, mammals [all chemicals] [TF 7-
31]
60. Algorithm: First-order excretion to water (General), birds [all
chemicals] [TF 7-32]
61. Algorithm: First-order excretion to water (General), mammals [all
chemicals] [TF 7-32]
TERRESTRIAL ANIMAL DIRECT INGESTION OF SOIL OR SW
94. Algorithm: Ingestion of soil by birds, General [all chemicals]
[TF 7-22]
95. Algorithm: Ingestion of soil by mammals, General (surface soil) [all
chemicals] [TF 7-22]
97. Algorithm: Ingestion of water by birds, General (surface water) [all
chemicals] [TF 7-21]
98. Algorithm: Ingestion of water by mammals, General (surface water)
[all chemicals]] [TF 7-21]
84b. Ingestion of leaf particles by birds [new] - applies only to certain birds,
depending on their diet. [TF 7-24]
SEPTEMBER 2002
I-A-8
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-A
TRIM.FATE ALGORITHM PAIRING TABLES
7. Algorithm: Degradation/Reaction sink in Fish [organic chemicals]
[TF 2-1]
14. Algorithm: Degradation/Reaction sink in Sediment [organic
chemicals] [TF 2-1]
11. Algorithm: Degradation/Reaction sink in Benthic Invertebrates
[organic chemicals] [TF 2- 1 ]
BULK ADVECTION TO ADVECTION SINK
3. Algorithm: Bulk advection from air to advection sink [all chemicals]
[TF 3-lb]
4. Algorithm: Bulk advection from surface water to flush-rate
advection sink [all chemicals] [TF 4-7a,b]
OTHER SINKS
179. Algorithm: Runoff from Surface Soil to Surface Soil Sink [all
chemicals] [new] [TF 5-10c]
178. Algorithm: Erosion from Surface Soil to Surface Soil Sink [all
chemicals] [new] [TF 5- lie]
150. Algorithm: Sediment burial from sediment to sediment burial
sink, Zero net deposition, [all chemicals] [TF 4-8]
84c. Ingestion of leaf particles by mammals [new] - applies only to certain
mammals, depending on their diet. [TF 7-24]
TERRESTRIAL ANIMAL INHALATION
100. Algorithm: Inhalation by birds, General [all chemicals]
[TF 7-30]
101. Algorithm: Inhalation by mammals, General [all chemicals] [TF 7-30]
LITTER FALL
102. Algorithm: Litter fall to soil from leaf particle [all chemicals]
[TF 7-16]
103. Algorithm: Litter fall to soil from leaves [all chemicals] [TF 7-15]
aFood ingestion algorithms are not presented in this table.
b Numbers in brackets (e.g., [TF 3-1]) refer to the numbered transfer factor algorithms in the TRIM.FaTE Technical Support Document Volume II: Description of
Chemical Transport and Transformation Algorithms, September 2002 (EPA-453/R-02-01 Ib).
SEPTEMBER 2002
I-A-9
TRIM.FATE EVALUATION REPORT VOLUME I
-------
[This page intentionally left blank.]
-------
APPENDIX I-B
BIOMASS OF FISH
-------
[This page intentionally left blank.]
-------
APPENDIX I-B
BIOMASS OF FISH
APPENDIX I-B
BIOMASS OF FISH
The mass offish in each trophic level is derived from studies of the biomass of
individual species in various systems and studies of feeding strategies of those species. Kelso
and Johnson (1991) reported the biomass of individual species in 19 central Ontario lakes. Only
adult fish (>_1 year of age) were retained by the traps and nets used for this study. Population
biomass was estimated using the mark and recapture method. The biomass of rare species could
not be quantitatively estimated and are not included in the estimates of total biomass per lake or
per area. Thus, these methods likely underestimate the total and species-specific biomass.
The most significant impact of this bias on TREVI.FaTE is that the young-of-the-year
(YOY) of many species (e.g., perch and bass) are planktivorous. Thus, YOY biomass had to be
approximated in order to populate the water column herbivore (planktivore) domain. Robust
estimates of YOY biomass have not yet been found, though future efforts will focus on this task.
The ratio of YOY to adult biomass for rainbow and brown trout in the Tule River, California (H.
Yagger, personal communication) was used as a first approximation for the species
quantitatively surveyed by Kelso and Johnson (1991). The reproductive strategies of most
freshwater fish rely on the production of large numbers of eggs, with relatively few individuals
progressing to adulthood. The result is that the YOY biomass may equal or exceed the total
adult biomass. In the Tule River, the YOY:adult ratio for rainbow trout ranged from 1.0 to 1.66.
The brown trout population was less stable and the YOY:adult ratio ranged from 0.002 to 0.4.
A YOY:adult biomass ratio of 1.0 is used for the current version of TREVI.FaTE. Thus, the total
biomass estimates for each species were doubled.
The trophic status of each species was preliminarily assigned according to the
designations used by EPA for evaluating the biological integrity of surface waters (U.S.EPA
1993). Those designations are based on reviews of feeding studies, with each species assigned
to the one trophic level that best describes the feeding habits of adult fish. However, most fish
species occupy more than one trophic level. This is especially true when YOY fish are included.
But even adults of some species will feed opportunistically on a range of prey. Therefore,
dietary studies of selected species were used to refine and confirm the preliminary trophic level
designations (Etnier and Starnes 1993).
Exhibit I-B-1 presents the current TREVI.FaTE trophic level designations for the species
quantitatively surveyed by Kelso and Johnson (1991). These levels correspond to the current
fish domains in each of the two food chains (e.g., water column omnivore and benthic
omnivore). Species occupying more than one trophic level were fractionally assigned to
multiple domains. For example, the stomach contents of yellow perch have been found to consist
of approximately 70 percent invertebrates and 30 percent fish (Hayes et al. 1992). Therefore, 70
percent of the adult yellow perch population (based on biomass) is assigned to the benthic
omnivore domain and 30 percent of the population is assigned to the water column omnivore
domain. The water column omnivore domain was selected instead of the piscivore domains
because yellow perch consume relatively small fish, which are assigned to the water column
herbivore (planktivore) domain.
Biomass of the species categorized as carnivores was similarly modified to reflect
changes in diet with size. That is, older, larger carnivores (e.g., bass) are assumed to prey on
omnivores, whereas younger, smaller members of the population are assumed to prey on
SEPTEMBER 2002 I-B-1 TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-B
BIOMASS OF FISH
herbivores. Data have not yet been found with which to estimate the fraction of bass (or other
species identified as carnivores) to be reassigned to the omnivore categories As a first
approximation, the water column carnivore biomass was split as 30 percent carnivore and 70
percent omnivore. These are the values Hayes (1992) presented for yellow perch consumption
offish (30 percent) and invertebrates (70 percent). While these data are not for a species
included in the carnivore categories, the inherent assumptions are reasonably analogous.
Therefore, the final water column carnivore biomass is 0.3 times the initial, literature-derived
water column carnivore biomass. The new water column omnivore biomass is the sum of the
initial, literature-derived water column omnivore biomass and 0.7 times the initial,
literature-derived water column carnivore biomass. This appears to produce a reasonable
apportionment of biomass among the water column trophic levels (7.43 percent water column
carnivores; 24.26 percent water column omnivores; 68.31 percent water column herbivores; ratio
of 1:3:9)
Benthic carnivores were similarly assigned as 30 percent benthic carnivores and 70
percent benthic omnivores. Thus, the new benthic carnivore biomass is 0.3 times the initial,
literature-derived benthic carnivore biomass The new benthic omnivore biomass is the sum of
the initial, literature-derived benthic omnivore biomass and 0.7 times the initial,
literature-derived benthic carnivore biomass. Benthic herbivores are comprised of benthic
invertebrates rather than fish. The current default value for benthic invertebrates is a total
biomass per area (kg m"2) of 3.7E-03 (value for Brewer Lake, Maine, from D. Courtemanch).
These values appear to produce a fairly reasonable apportionment of biomass among the benthic
trophic levels (3.69 percent benthic carnivores; 32.53 percent benthic omnivores; 63.78 percent
benthic invertebrates; ratio of 1:3:17). Although four percent benthic carnivores seems low,
there is not a good rationale for apportioning the benthic carnivore biomass differently than the
water column carnivore biomass (i.e., 30/70).
Exhibit I-F-2 presents the current default estimates of biomass per area by trophic level
for TRJJVI.FaTE. Only data reported by Kelso and Johnson (1991) for lakes with pH > 6.0 in
which both food chains are represented are included.
SEPTEMBER 2002 I-B-2 TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-B
BlOMASS OF FISH
Exhibit I-B-1
Trophic Level Designations for Species Quantitatively Surveyed
Species
Creek chub
Creek chub
Pearl dace
Pearl dace
Northern redbelly dace
Northern redbelly dace
Brown bullhead
Brown bullhead
Brook trout
Brook trout
Common shiner
Common shiner
Pumpkinseed
Yellow perch
White sucker
White sucker
Largemouth bass
Northern pike
Northern pike
Rockbass
Smallmouth bass
Largemouth bass
Northern pike
Northern pike
Rockbass
Smallmouth bass
Largemouth bass
Pumpkinseed
Rockbass
Smallmouth bass
Yellow perch
Yellow perch
Largemouth bass
Northern pike
Northern pike
Rockbass
Smallmouth bass
Largemouth bass
Northern pike
Northern pike
Smallmouth bass
by
Species ID
CC
CC-yoy
PD
PD-yoy
RD
RD-yoy
BB
BB-yoy
BT
BT-yoy
CS
CS-yoy
PS
YP
WS
WS-yoy
LB
NP
NP-yoy
RB
SB
LB
NP
NP-yoy
RB
SB
LB-yoy
PS-yoy
RB-yoy
SB-yoy
YP-yoy
YP
LB
NP
NP-yoy
RB
SB
LB
NP
NP-yoy
SB
Kelso and Johnson (1991)
EPA Trophic Level TRIM
G
G
G
G
H
H
1
1
1
1
1
1
1
1
0
O
P
P
P
P
P
P
P
P
P
P
1
P
P
P
P
P
P
P
P
P
Trophic Level
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BO
BC
BC
BC
BC
BC
WCH
WCH
WCH
WCH
WCH
WCO
WCO
WCO
WCO
WCO
WCO
wcc
wcc
wcc
wcc
% of Biomass
100
100
100
100
100
100
100
100
100
100
100
100
100
70
100
100
35
35
35
35
35
15
15
15
15
15
100
100
100
100
100
30
35
35
35
50
35
15
15
15
15
SEPTEMBER 2002
I-B-3
TRIM.FATE EVALUATION REPORT VOLUME I
-------
APPENDIX I-B
BIOMASS OF FISH
Exhibit I-B-2
Biomass (kg/ha) by Trophic Level for All Lakes with pH > 6.0 in Which
Both Food Chains Are Represented
TRIM Trophic Level
Lake
04
20
33
A1
C1
P1
Grand Total
Mean
Min
Max
Ave Biomass (kg/m2)
Typical mass (kg)
Population (#/m2)
BO
13.27
39.47
5.95
14.80
22.43
17.32
320.31
18.87
5.95
39.47
1.89E-03
0.5
3.77E-03
BC
2.15
2.20
1.46
4.08
1.22
1.75
12.85
2.14
1.22
4.08
2.14E-04
2
1.07E-04
WCH
15.89
16.63
13.35
27.18
11.10
14.59
98.75
16.46
11.10
27.18
1.65E-03
0.025
6.58E-02
WCO
5.48
6.60
4.49
9.51
4.07
4.93
35.08
5.85
4.07
9.51
5.85E-04
0.5
1.17E-03
WCC
2.15
1.33
1.46
4.08
0.00
1.73
10.74
1.79
0.00
4.08
1.79E-04
2
8.95E-05
Grand
Total
38.95
66.22
26.70
59.65
38.82
40.32
477.73
45.11
26.70
66.22
4.51 E-03
References for Appendix I-B
Etnier, D. A., and W. C. Starnes. 1993. The fishes of Tennessee. University of Tennessee Press,
Knoxville, TN.
Hayes, D. B., W. W. Taylor, and J. C. Schneider. 1992. Response of yellow perch and the
benthic invertebrate community to a reduction in the abundance of white suckers. Transactions
of the American Fisheries Society 121:36-53.
Kelso, J.R.M. and M.G. Johnson. 1991. Factors related to the biomass and production offish
communities in small, oligotrophic lakes vulnerable to acidification. Canadian Journal of
Fisheries and Aquatic Sciences. 48:2523-2532.
U.S. EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors
Handbook, Volume I. EPA/600/R-93/187a. Washington, D.C.: Office of Water and Office of
Research and Development. December.
SEPTEMBER 2002
I-B-4
TRIM.FATE EVALUATION REPORT VOLUME I
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-453/R-02-012
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation of TRIM.FaTE. Volume I: Approach
and Initial Findings
5. REPORT DATE
September, 2002
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Emissions Standards &
Air Quality Strategies and Standards Divisions
Research Triangle Park, NC 27711
Technical Report
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report is part of a series of documentation for the Total Risk Integrated Methodology
(TRIM). TRIM is a time series modeling system, with multimedia capabilities, designed for
assessing human health and ecological risks from hazardous and criteria air pollutants. This
report describes a set of evaluation analyses performed on the Environmental Fate, Transport,
and Ecological Exposure module of TRIM (TRIM.FaTE), primarily during 2000, with some
spanning into 2002.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATT Field/Group
Risk Assessment
Multimedia Modeling
Exposure Assessment
Air Pollutants
Air Pollution
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
151
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
-------
P
.
=s "3
bo 3
"3 fa ^
3 -d bo
?! I
"3 t/2 r^
«
o g^
8'l §
g'a s
O W P^
S
s
PH
"3
"S
>,
o &1
.
.^ >
a S
<;
------- |