EPA-SAB-EPEC/DWC-COM-95-006
September 29, 1995
Honorable Carol M. Browner
Administrator
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC. 20460
Subject: Commentary on Bioaccumulation Modeling Issues
Dear Ms. Browner:
On April 28-29, 1994, a joint Bioaccumulation Subcommittee with representatives from
the Ecological Processes and Effects Committee and the Drinking Water Committee of the
Science Advisory Board met to engage in a consultation on approaches to estimating
bioaccumulation potential of chemicals and to discuss various mass balance/food web models. At
the meeting, Agency staff indicated that the ultimate goal of the Agency is to develop a uniform
approach to bioaccumulation assessment for use in a number of regulatory efforts (e.g.,
development of aquatic life, human health, sediment, and wildlife criteria).
The validity and utility of any model are largely dependent on its underlying assumptions,
the input parameters used, and the level of uncertainty acceptable in model outputs. Thus, rather
than discussing pros/cons of any specific mass balance/food web (MB/FW) model, the
Subcommittee prepared this commentary to provide more general advice on how and when the
Agency should use MB/FW models to estimate bioaccumulation and what research is needed to
improve model predictions.
Use of Mass Balance/Food Web Models to Predict Bioaccumulation
Bioconcentration factors (BCFs), based on laboratory studies of organisms exposed to
water containing a chemical of concern, have been shown in some instances to underestimate
bioaccumulation potential (specifically for hydrophobic compounds with log Kow > 4.5 which are
resistant to metabolism and degradation) since they do not account for biomagnification of
chemicals in the food web as predators consume prey containing lipophilic compounds.
Conversely, BCFs have been shown to overestimate bioaccumulation potential when a chemical is
bound or tightly sorbed to sediment, i.e., not bioavailable. Thus, the development and application
of models to predict bioaccumulation factors (BAFs) and biota-sediment accumulation factors
(BSAFs), a measure of the uptake and accumulation by organisms of chemicals in sediments, are
-------
important in improving our ability to evaluate the fate of chemicals in the environment. While
available MB/FW models have a number of serious limitations (as noted below), they can be quite
useful as tools for providing insights into environmental processes and interrelationships, for
hypothesis testing, and for providing order-of-magnitude estimates of bioaccumulation potential
for the classes of compounds for which they were developed.
Bioaccumulation in aquatic organisms is seen in several studies found in the open
literature. Available information focuses on a few chemicals and trophic levels in an aquatic
ecosystem. Compartmental models utilizing these data have been developed and have appeared in
peer-reviewed publications. The Thomann and Gobas models are among the ones that represent
this subject area. At the onset, therefore, it should be recognized that model development,
enhancement, and validation efforts are ongoing in the ecological research arena. It should also
be recognized that empirical information forms the basis of the Thomann and Gobas models.
Conceptually, however, the models can be extended to include mechanistic knowledge of the
processes involved in bioaccumulation. Given the state of scientific knowledge, it is natural to
recommend that the Agency continue developing/improving modeling approaches for estimating
(precisely and accurately) bioaccumulation of chemicals in aquatic organisms that are of
regulatory interest. In making this recommendation, the Subcommittee observes that existing
data and models are applicable for deriving an order-of-magnitude estimate of bioaccumulation
for a class of chemicals with a log K^ of 3.5 to 6.0 and for chemicals that do not degrade or
transform. Scientists in the Subcommittee have two views on this recommendation: one that
subscribes to the idea of conducting research and utilizing the information to enhance these other
models to the extent that the scientific results allow; and the second that subscribes to the idea
that extensive research to enhance these models would not prove useful for compounds other than
those already studied and published in the literature. The following comments elaborate on the
first view.
To improve the utility of BAFs and BSAFs for environmental regulation, including the
calculation and application of water and sediment quality criteria, the Subcommittee highlighted
several important limitations to currently available bioaccumulation models and made
recommendations for how these limitations might be addressed.
a) MB/FW models such as the Thomann model, which have been developed for
persistent, halogenated organic compounds, do not accurately predict
bioaccumulation potential for chemicals that are significantly metabolized by food
web organisms, degraded in the environment (including microbiological
degradation), or not bioavailable. Similarly, because model development has been
focused on a specific class of chemicals, the potential for adapting MB/FW models
to other classes of chemicals—particularly those whose partitioning may be driven
by mechanisms not represented by octanol/water partitioning—has not yet been
addressed.
-------
Recommendation 1: Significant uncertainties exist in the ability of MB/FW models
(Thomann's and others) to predict accurately the extent of biomagnification or
bioaccumulation of many important chemical classes. Thus, validation of these models is
necessary prior to their use. The Agency should conduct studies to determine whether
reliable MB/FW models can be developed for other classes of chemical compounds and
broaden validation of bioaccumulation models using data on compounds with different
chemical properties and in different environmental settings (e.g., other than large lakes).
These studies will require collection of additional field and laboratory data to test how
well various classes of compounds can be modeled. Although Quantitative Structure-
Activity Relationships (QSAR) may offer insights into the bioaccumulation potential of
chemicals, these relationships should not be viewed as a substitute for such data.
b) The models contain many sources of uncertainty, and this uncertainty is often not
adequately characterized.
Recommendation 2: The Agency should attempt to quantify the uncertainties in model
outputs (i.e., place confidence limits on model predictions), including uncertainties
resulting from the stochastic nature of natural systems and from the natural variability
among different types of aquatic ecosystems.
Recommendation 3: The Agency should define the desired or acceptable range of
uncertainty in the prediction and minimum criteria for use of MB/FW models for different
applications of the methodology. For example, greater uncertainty would be tolerable in
screening tests intended to identify chemicals for further testing than for models used in
the promulgation of major regulations such as the Great Lakes Water Quality Initiative.
Recommendation 4: The Agency should focus field and laboratory data collection toward
reducing uncertainties in existing bioaccumulation models. Research in this area can be
prioritized by conducting sensitivity and uncertainty analyses to identify those processes to
which model output is most sensitive.
c) Model outputs or predictions are only as good as the data upon which they are
based.
Recommendation 5: To improve the quality of the data available for modeling efforts, the
Agency should clarify quality assurance requirements for collection of field data as well as
for screening of existing data.
Recommendation 6: The Agency should develop or identify standard analytical methods,
particularly for measuring bioavailable fractions of organic chemicals in water, sediment,
and biota, and for estimating rates of metabolism.
-------
Recommendation 7: Only robust extant field data of acceptable quality, i.e., data elements
with acceptable precision, as defined by the Agency, should be used for model validation.
Acceptable variances for data elements are determined by conducting sensitivity tests on
the input parameters for a prospective model. If extant data are of unacceptable quality,
then additional field data should be collected with the Data Quality Objectives set to yield
acceptable precision for each data element. Subsequently, the predicted biomagnification
or bioaccumulation of pollutants should be compared to field measurements to assess the
bias of the model.
Recommendation 8: Because of the significant analytical difficulties associated with
measuring the concentration of super-hydrophobic compounds in water, bioaccumulation
modeling results for these compounds are highly uncertain. Until better approaches are
developed for estimating water concentrations for compounds with log K^ greater than
about 6, decisionmakers must be particularly aware of the increased scientific uncertainties
associated with attempts to model super-hydrophobic compounds and especially wary of
the use of such results to support policy and regulatory decisions.
Recommendation 9: Since the process of model development and validation is iterative,
MB/FW models used by the Agency to predict bioaccumulation should be updated at
regular intervals using the best currently available empirical data.
Integration of Mechanistic and Empirical Models
The Subcommittee agrees with the Agency that both mechanistic and empirical modeling
approaches are needed to improve bioaccumulation predictions. Although models of natural
systems can never truly be verified or validated completely, high quality field data can be used to
calibrate and confirm model predictions. In turn, model predictions should identify uncertainties
in the field data and provide insight on what to measure, and when and where to sample.
For regulatory applications requiring the highest degree of accuracy or precision, BAFs or
BSAFs should be based on field data and tested to determine the relation between the measured
value and the specific driving variables known to affect the value (e.g., temperature, pH,
nutritional factors). This procedure will allow extrapolation within reason to sites within the
range of the tested variables. This approach is preferable to using inadequately validated models
(mechanistic or empirical) and will ultimately result in a database that can provide for a good
empirical model.
Following appropriate validation studies, including field data, model application can then
be made on the basis of Kow's with adjustment for effects of biomagnification, metabolism, and
other factors such as microbial alterations. However, the uncertainty associated with these
predictions may not be acceptable for all applications. The Subcommittee agrees with the Agency
that when site-specific BAFs or BSAFs cannot be measured but a high degree of accuracy and
precision is desired, site-specific bioaccumulation models may be required to account for
-------
differences in bioavailability and sorption dynamics, food web structure, selectivity of predators
with multiple prey choices, and other factors that affect a chemical's behavior in a given
ecosystem. Mechanistic models also can be used for screening new chemicals by applying them
for a range of prototypical environments (e.g., stream, river, estuary, large temperate lake,
shallow warm water fishery, system dominated by a benthic food web) to evaluate the
bioaccumulation potential of the same chemical in different possible ecosystems.
Research to Improve Model Predictions
The Subcommittee supports the research priorities identified by the Agency, including the
need to:
a) better characterize exposures of benthic organisms resulting from ingestion of
sediments and sediment-bound chemicals;
b) determine the effect of food web structure on BAFs or BSAFs for generic
ecosystems as a function of chemical class and Kow;
c) adapt food web models to incorporate residue-based analyses for ecological risk
assessments; and
d) develop methods to incorporate BAFs and BSAFs into complex chemical mixture
assessment procedures.
In addition, the Subcommittee urges the Agency to assess the effect on model predictions
of environmental factors influencing bioavailability of chemicals (e.g., microbiological
degradation, dissolved organic matter, matrix effects, water chemistry, sediment characteristics,
presence/absence of light).
In summary, while the Subcommittee agrees that mass balance/food web models such as
the Thomann model hold promise for predicting bioaccumulation of certain types of chemicals,
we urge the Agency to further field test the models for additional classes of compounds and for
additional environmental settings and assess the uncertainties in model predictions prior to their
wide-spread application in a regulatory context. Ongoing peer review should be an integral part
of this process. Finally, the use of models, no matter how refined, should be augmented by
appropriately designed laboratory and field experiments and monitoring.
-------
The Subcommittee appreciated the opportunity to meet with Agency staff to discuss
approaches to estimating bioaccumulation, including the application of models such as that of
Thomann and Gobas. We hope our comments and recommendations are helpful to the Agency
and we look forward to your response.
Sincerely,
Dr. Genevieve M. Matanoski, Chair
Executive Committee
Dri Mark A. Harwell, Chair
Ecological Processes and
Effects Committee
Dr. Verne Ray,
Drinking Water Com
Dr. Anne McElroy, Co-Chair
Bioaccumulation Subcommittee
Attachments
-------
U.S. Environmental Protection Agency
NOTICE
This report has been written as part of the activities of the Science Advisory Board, a
public advisory group providing extramural scientific information and advice to the Administrator
and other officials of the Environmental Protection Agency. The Board is structured to provide
balanced, expert assessment of scientific matters related to problems facing the Agency. This
report has not been reviewed for approval by the Agency and, hence, the contents of this report
do not necessarily represent the views and policies of the Environmental Protection Agency, nor
of other agencies in the Executive Branch of the Federal government, nor does mention of trade
names or commercial products constitute a recommendation for use.
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD
BIO ACCUMULATION SUBCOMMITTEE
OF THE
ECOLOGICAL PROCESSES AND EFFECTS COMMITTEE
AND
DRINKING WATER COMMITTEE
CO-CHAIRS
Dr. Anne McElroy, Director, NY Sea Grant, SUNY at Stony Brook, Stony Brook, NY
Dr. Richard H. Reitz1, McLaren/Hart, Midland, MI
MEMBERS
Dr. Lenore S. Clesceri, Rensselaer Polytechnic Institute, Materials Research Center, Troy, NY
Dr. Alan W. Maki, Exxon Company, USA, Houston, TX
Dr. Edo D. Pellizzari, Research Triangle Institute, Research Triangle Park, NC
Dr. Frederic K. Pfaender, Carolina Federation for Environmental Studies, University of North
Carolina, Chapel Hill, NC
Dr. William H. Smith, School of Forestry and Environmental Studies, Yale University, New
Haven, CT
Dr. Terry F. Young, Environmental Defense Fund, Oakland, CA
CONSULTANT
Dr. Anne Spacie, Department of Fisheries and Aquatic Science, Purdue University, West
Lafayette, IN
INVITED EXPERT
Dr. Joseph V. DePinto, Great Lakes Program, SUNY at Buffalo, Buffalo, NY
JDr. Reitz did not support all of the recommendations in the final report and resigned as co-chair
of the Subcommittee.
-------
SCIENCE ADVISORY BOARD STAFF
Ms. Stephanie Sanzone, Designated Federal Officer, US EPA, Science Advisory Board (1400F),
401 M Street, SW, Washington, DC 20460. Phone: (202) 260-6557
Dr. Manuel Gomez, Designated Federal Officer, US EPA, Science Advisory Board (1400F), 401
M Street, SW, Washington, DC 20460. (Dr. Gomez is no longer on the SAB Staff)
Ms. Mary Winston, Staff Secretary, US EPA, Science Advisory Board (1400F), 401 M Street,
SW, Washington, DC 20460. Phone: (202) 260-6552
-------
DISTRIBUTION LIST
Administrator
Deputy Administrator
Assistant Administrators
Deputy Assistant Administrator for Research and Development
Deputy Assistant Administrator for Water
EPA Regional Administrators
EPA Laboratory Directors
EPA Regional Libraries
EPA Laboratory Libraries
National Technical Information System
Congressional Research Service
Library of Congress
------- |