United States
Environmental Protection
Agency
EPA/630/R-92/005
May 1993
A Review of
Ecological Assessment
Case Studies from a
Risk Assessment
Perspective
RISK ASSESSMENT FORUM
-------
EPA/630/R-92/005
May 1993
A REVIEW OF ECOLOGICAL ASSESSMENT CASE STUDIES
FROM A RISK ASSESSMENT PERSPECTIVE
Assembled by:
Eastern Research Group, Inc.
110 Hartwell Avenue
Lexington, MA 02173-3198
EPA Contract No. 68-C1-0030
for the
Risk Assessment Forum
U.S. Environmental Protection Agency
Washington, DC 20460
Printed on Recycled Paper
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection : ,
Agency policy and approved for publication. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use. Case study data and interpretations were ..,•
current as of the peer review workshops held in the spring of 1991.
11
-------
CONTENTS
Foreword v
Report Authors and Case Study Work Group Leaders vi
Summary viii
PART I. CASE STUDIES OVERVIEW 1
1. Introduction 1
2. Guide to the Case Studies 2
2.1. Background 2
2.2. Case Study Highlights 2
2,2.1. Problem Formulation 6
2.2.2. Analysis 9
2.2.3. Risk Characterization 14
3. Key Terms 18
4. References 19
PART II. THE CASE STUDIES 20
1. Effects of Acidic Deposition on Aquatic Ecosystems—A Regional
Problem (Short Title: Acidic Deposition Case Study) 1-1
2. The Bay Drums, Peak Oil, and Reeves Southeastern Areawide
Wetland Impact Study (Bay Drums Case Study) 2-1
3. Special Review of the Granular Formulations of Carbofuran
Based on Adverse Effects on Birds (Carbofuran Case Study) 3-1
4. Ecological Evaluation of a Freshwater Stream and Wetlands Near
an Inactive Coke Production Plant (Coke Plant Case Study) 4-1
5. Commencement Bay Tidelands Assessment (Commencement Bay
Case Study) 5-1
6. The National Crop Loss Assessment Network (Crop Loss Case
Study) 6-1
m
-------
CONTENTS (continued)
7. Comparative Analysis of Mining Tailing Disposal for the Quartz
Hill Molybdenum Mining Project (Quartz Hill Case Study) 7-1
8. Assessing Ecological Risk at Rocky Mountain Arsenal (Rocky
Mountain Arsenal Case Study) 8-1
9. Selenium Effects at Kesterson Reservoir (Kesterson Case Study) 9-1
10. Risk Characterization Methods Used in Determining the Effects
of Synthetic Pyrethroids on Terrestrial and Aquatic Organisms
(Synthetic Pyrethroids Case Study) 10-1
11. Toxic Discharges to Surface Waters: Assessing the Risk to Aquatic
Life Using National and Site-Specific Water Quality Criteria
(Water Quality Criteria Case Study) 11-1
12. Modeling Future Losses of Bottomland Forest Wetlands and
Changes in Wildlife Habitat Within a Louisiana Basin (Wetlands
Loss Case Study) 12-1
IV
-------
FOREWORD
To gain insight into the process of ecological risk assessment, scientists from the
U.S. Environmental Protection Agency (EPA) have analyzed a cross-section of case studies
representing the "state-of-the-practice" in ecological assessment. Each case was evaluated by
scientific experts in a series of EPA-sponsored workshops between May 29 and June 20, 1991
(56 Federal Register 22869, 17 May 1991). These peer review workshops were chaired by
Dr. Charles Menzie and included reviewers from universities, private organizations, and other
federal agencies.
The case study workshops, along with two other workshops held during the spring of 1991
(57 Federal Register 22236, 27 May 1992), were part of a new EPA program to develop
guidelines for ecological risk assessment. Each workshop was designed to open a dialogue among
experts on issues pertaining to the development of such guidelines. This report brings together 12
case studies that illustrate important ecological risk assessment practices.
The case studies are wide-ranging in scope, representing a variety of ecosystems, ecological
endpoints, chemical and nonchemical stressors, and programmatic requirements within the Agency.
As a result, workshop participants were presented with a broad diversity of risk assessment and
scientific issues, and many useful principles emerged from the resulting discussions.
The case studies report provides a useful first look at some common approaches to
ecological assessment in relationship to a general ecological risk process. The cases selected were
evaluated at the workshops as to whether they (1) effectively addressed generally accepted
components of an ecological risk assessment, or (2) addressed some but not all of these
components or, instead, (3) provided an alternative approach to assessing ecological effects. The
analyses and discussions in this report provide useful information about ecological risk processes.
Dorothy E. Patton, Ph.D.
Chair
Risk Assessment Forum
-------
REPORT AUTHORS AND CASE STUDY WORK GROUP LEADERS
This report was prepared by Dr. William van der Schalie and Dr. Ronald Landy of the U.S.
Environmental Protection Agency, and Dr. Charles Menzie, a consultant for this activity. The case
studies were initiated by EPA work groups chaired by the individuals listed below, with Dr. Landy
as overall chair. The workshops were organized by Dr. Landy and Dr. van der Schalie, with the '
assistance of Dr. Menzie. Eastern Research Group, Inc. (ERG), EPA's contractor for this activity,
provided organizational support and assembled this report. Case study authors and peer reviewers
are listed at the beginning of each case study (part U).
REPORT AUTHORS
William van der Schalie
Risk Assessment Forum Staff
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
Ronald Landy
Office of Technology Transfer
and Regulatory Support
U.S. Environmental Protection Agency
Washington, DC
CASE STUDY WORK GROUP LEADERS
Steven Broderius
Environmental Research
Laboratory—Duluth
U.S. Environmental Protection Agency
Duluth, MN
Randall Bruins
Region 10
U.S. Environmental Protection Agency
Seattle, WA
Patricia Cirone
Region 10
U.S. Environmental Protection Agency
Seattle, WA
James Clark
Environmental Research
Laboratory—Gulf Breeze
U.S. Environmental Protection Agency
Gulf Breeze, FL
Charles Menzie
Menzie-Cura & Associates, Inc.
Chelmford, MA
Milton Clark
Region 5
U.S. Environmental Protection Agency
Chicago, IL
Clyde Houseknecht
Office of Water
U.S. Environmental Protection Agency
Washington, DC
Jay Messer
Atmospheric Research and
Exposure Assessment Laboratory
Research Triangle Park, NC
Ronald Preston
Region 3
U.S. Environmental Protection Agency
Wheeling, WV
VI
-------
CASE STUDY WORK GROUP LEADERS (continued)
Jon Rauscher
Region 6
U.S. Environmental Protection Agency
Dallas, TX
Harvey Simon
Region 2
U.S. Environmental Protection Agency
New York, NY
Quentin Stober
Region 4
U.S. Environmental Protection Agency
Atlanta, GA
Greg Susanke
Office of Pesticide Programs
U.S. Environmental Protection Agency
Washington, DC
Daniel Vallero
Atmospheric Research and
Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
Bill Williams
Ecological Planning and Toxicology, Inc.
Corvallis, OR
Vll
-------
SUMMARY
This report uses case studies to explore the relationship between the process of ecological
risk assessment and common approaches used by EPA (and others) to evaluate adverse ecological
effects. The case studies are wide-ranging in scope, representing a variety of ecosystems,
ecological endpoints, chemical and nonchemical stressors, and programmatic requirements within
the Agency. The case studies were evaluated at peer review workshops as to whether they
(1) effectively addressed general components of an ecological risk assessment—problem
formulation, analysis, and risk characterization—or (2) addressed some but not all of these
components or, instead, (3) provided an alternative approach to assessing ecological effects. Case
study strengths and limitations noted by the reviewers and authors are included in the comment
boxes contained in each case study.
Some of the themes that emerged from these diverse case studies are highlighted below.
• Discussions between the risk assessor, risk manager, and relevant experts are
critical both at the beginning and end of an ecological risk assessment. Discussion
at the beginning will help ensure that the final assessment will contribute to a
management decision as well as address important ecological concerns. When the
risk assessment has been completed, the conclusions, assumptions, and uncertainties
of the risk assessment must be clearly conveyed.
• Some difficulties encountered in the case studies might have been avoided had more
attention been paid to the initial planning stages of the assessment (problem
formulation). Some case studies were too narrowly focused and did not consider all
relevant stressors (both chemical and nonchemical). The ecological values to be
protected should be carefully considered and clearly identified at the outset of an
assessment.
« While ecological exposure and effects models were useful, sensitivity analyses and
validation studies were frequently insufficient to evaluate the relevance of the
models to "real world" situations.
• Field studies provided a level of realism not readily attainable in laboratory studies,
but multiple stressors frequently made it difficult to identify a particular stressor as
the cause of observed ecological effects. Finding a suitable reference site for
comparison against a potentially affected area also was difficult.
• The case studies varied widely in their approaches to presenting the results of an
assessment, although for chemical stressors relatively simple comparisons between
point estimates of exposure and effect levels were common.
The varied approaches to ecological risk assessment used in the case studies are generally
consistent with some, but not all, of the principles in EPA's recently published Framework for
Ecological Risk Assessment (Framework Report) (U.S. EPA, 1992). However, EPA notes that
the cases and peer review comments were developed using "pre-firamework" terminology and
Vlll
-------
concepts. Thus it is important to understand that while the cases are useful examples of the
state of the practice, they should not be regarded as models to be followed. EPA will
continue to study concepts found in the Framework Report in future case studies. These case
studies and others now being prepared will be used along with the Framework Report to provide a
foundation for future Agency-wide guidelines for ecological risk assessment.
This report has two main parts. Part I includes general information on the development
and the use of the case studies and listings of key terms and references. Part II includes the 12
case studies that were developed and reviewed for this report.
IX
-------
-------
PARTI. CASE STUDIES OVERVIEW
1. INTRODUCTION
The U.S. Environmental Protection Agency's (EPA) Risk Assessment Forum is developing
Agency-wide guidance for conducting ecological risk assessments. The Forum initiated this
activity in 1990 with a series of meetings to explore significant issues with experts in the field and
to meet with individuals from state and other federal agencies to discuss their approaches to
ecological risk assessment (U.S. EPA, 1991). Based on these and other discussions, the Forum's
initial goals were to: (1) develop the Framework for Ecological Risk Assessment (Framework
Report) that describes basic principles of ecological risk assessment (U.S. EPA, 1992); (2) prepare
a long-range plan for developing future ecological risk assessment guidelines; and (3) develop a set
of ecological assessment case studies. The impetus for developing the case studies came from EPA
scientists and managers, who felt that such studies could be useful "real world" examples of the
ecological risk assessment process.
This case studies report should be useful to EPA regional, laboratory, and headquarters
personnel conducting ecological risk assessments, as well as to interested individuals from other
federal and state agencies and the general public. In addition, this document is a useful companion
volume to EPA's Framework Report.
Case studies are integral to the development and evaluation of ecological risk assessment
guidelines. While this report examines other cases from a general risk assessment perspective,
future activities will evaluate existing case studies using principles described in the Framework
Report. In this way, case studies can serve both as examples of ecological risk assessment and as
a means to evaluate recommended procedures.
-------
2. GUIDE TO THE CASE STUDIES
2.1. Background
The case studies presented in part II of this report illustrate several types of ecological
assessments. As summarized in table 1, these cases involve:
• studies done under several different federal environmental laws;
• spatial scales ranging from local impacts to regional effects to national impacts;
• different types of stressors, including single chemicals and chemical mixtures, as
well as physical stressors such as hydrologic change or sedimentation;
• a variety of ecosystems, including aquatic (freshwater and marine), wetlands, and
terrestrial; and
• measurement endpoints reflecting different levels of biological organization, ranging
from effects on individual organisms up to and including effects on communities.
(See part I, section 3, for definitions of measurement and assessment endpoints.)
This set of case studies is not comprehensive. In particular, it does not include ecological
assessments done with very limited time or dollar resources. On the contrary, most of the studies
covered hi this report cost hundreds of thousands to millions of dollars and took months or years to
complete. This point is significant because many evaluations are done quickly based on limited
data. The reviewers of these case studies, therefore, recommended that future case studies include
examples of ecological risk assessments done with minimal resources.
2.2. Case Study Highlights
This section highlights some common themes and principles gleaned through development
and review of these case studies. The section is organized according to the framework for
ecological risk assessment provided in the recently published Framework Report (U.S. EPA, 1992)
(see figure 1):
" Problem formulation, which is a preliminary scoping process;
• Analysis, which includes characterization of both ecological effects and
exposure; and
• Risk characterization, which highlights qualitative and quantitative conclusions,
with special emphasis on data limitations and other uncertainties.
-------
I
o
a
(M To ®
"3 §"«H
5 "3 3
*ug
3
ft
i*
&
Spatial
Scale of
Assessment
Relevant
Federal
Legislationb
V
1
I
a
e
|
g
3
0
O
1
O
00
National
NAPAP
§
1
Q
.a
3
'3
|
U
*
*
1
ERCLA/SARA
U
e
Q
cs
1
•c;
*8
t"H
H
&
National
E
Carbofuran
'e
i
£
•>
a
5
s
i
CERCLA/
SARA
•*-»
1
o
-
B
e
g
g
-fc
0
U
|
s
1
CERCLA/
SARA
OQ
*k
ommencement
O
«
1
"3
&c
AH
H
H
National
3
u
.^
c
i
"»
a
1
CLn
y
13
$
U
1
1
-
1
g
g
•>
O
U
H
j^t
<
s
«
13
CERCLA/
SARA
.3
Rocky Mounta
Arsenal
oo
1
"3
&i
fc
*
K
13
FMBTA
Kesterson
|
g
g
0
O
.?.
M
National
I-H
CO
•o
e
i
i
tn
o
1
'>
•3
a
1— H
Si
»
13
3
1
•g
-S
U
I
5
tH
^
T— 4
1
g
§
0
O
*
OH
•|
1
i
z
CO
Wetlands Los
2
J;
I
a
3
1
-------
bLegislation
CAA: Clean Air Act (1970)
CERCLA/
SARA: Comprehensive Environmental Response, Compensation, and Liability Act (1980)/Superfund Amendments and
Reauthorization Act (1987)
CWA: Clean Water Act (1977)
FIFRA: Federal Insecticide, Fungicide, and Rodenticide Act (1972)
FMBTA: Federal Migratory Bird Treaty Act
NAPAP: National Acid Precipitation Assessment Program (under the Acid Precipitation Act of 1980)
NEPA: National Environmental Policy Act (1969)
°Stressor types
MC: Mixture of chemicals
P: Physical stressor (e.g., suspended solids deposition, hydrologic change)
SC: Single chemical (case study 10 addresses a group of closely related chemicals)
dEcosystem types
A/F: Aquatic—freshwater
A/M: Aquatic—marine or estuarine
T: Terrestrial
W: Wetlands
eHighest level of biological organization for the measurement endpoints used.
-------
Discussion
Between the
Risk Assessor
and
Risk Manager
(Planning)
Ecological Risk Assessment
PRD£LJEM FoSMULATtON -
A.,
S
l
s
Characterization ' Characterization
of | of
Exposure . Ecological
Effects
.
1
Data Acquisition; Verifi
on and Monitoring
Discussion Between the
Risk Assessor and Risk Manager
(Results)
I
Figure 1. The framework for ecological risk assessment (U.S. EPA, 1992). The ecological
risk assessment framework is the product of a series of workshops and reviews
that involved both EPA and outside scientists. While the Framework Report has
been a critical first step in developing ecological risk assessment concepts,
evolution of the framework concepts is expected and encouraged. One current
topic of discussion is the use of the term "exposure." Some scientists feel that
"exposure" is associated primarily with chemical stressors and, therefore, does not
adequately encompass physical and biological stressors that have great ecological
significance. This and other framework issues will be addressed in future
substantive guidance.
-------
2.2.1. Problem Formulation
Problem formulation is an initial planning and scoping process to define the feasibility,
scope, and objectives for the ecological risk assessment. This process includes preliminary
evaluation of exposure and effects, as well as examination of scientific data and data needs,
regulatory issues, and site-specific factors. Problem formulation defines the ecosystems potentially
at risk, the stressors, and the measurement and assessment endpoints. This information may then
be summarized in a conceptual model, which hypothesizes how the stressor might affect the
ecological components (i.e., the individuals, populations, communities, or ecosystems of concern).
Although the conceptual model idea was not available to the writers or reviewers of the case
studies in this document, elements of the conceptual model were present in most of the case
studies.
The Risk Assessment
Framework Is
Applicable to Physical
Stressors
Thorough Formulation
of the Problem and
Development of the
Scope Are Essential
First Steps for a
Successful Risk
Assessment
It Is Important to
Clearly Articulate
Management Issues at
the Beginning of an
Assessment
Although most of the case studies deal with exposures to chemicals, a
few involve assessments of physical modifications to habitats. These
include the change in water elevations in the Wetlands Loss case
study and the disposal of mine tailings in the Quartz Hill case study.
The stressors, ecological components, and endpoints in these
assessments can be discussed within the ecological risk assessment
framework.
The case studies illustrate the importance of clearly defining the goals
of the assessment and of developing a scope that is appropriate for
achieving those goals within the constraints of available resources and
the overall uncertainties of the analyses. Reviewers of the case
studies generally indicated that a good assessment is one that provides
the information needed to address the hypothesis, question, or
management decision at a level appropriate to the decision. To
accomplish this, the problem formulation should ensure that the
assessment focuses on the stressors, ecological components, and
endpoints that are most appropriate for the problem and for making
ultimate management decisions. Reviewers observed that this was
especially critical when resources are limited by fiscal constraints.
The strengths and weaknesses of the case studies seem to originate, in
large part, from decisions made during the preliminary planning
stages.
The Crop Loss case study is a good example of an assessment where
the ultimate management issue was clear from the onset; where the
stressor, ecological components, and endpoints were clearly defined;
and where the design of the study was structured around a clear set of
hypotheses amenable to scientific inquiry. This level of clarity was
achieved, in part, through meetings and interactions among
researchers and others involved with the risk assessment/risk
management process. The author and reviewers of this case study
-------
stressed the importance of this type of communication for clarifying
goals.
The Risk Assessor
Should State the
Hypotheses Being
Evaluated in the Risk
Assessment
The Possibility That
Multiple Stressors May
Confound the
Interpretation of Risks
Should Be Considered
Many reviewers noted that the problem formulation stage could
benefit from clearly stated hypotheses. Such an approach is consistent
with the scientific method and would ensure that the risk assessors
and managers understand the intent of the analysis. Most of the case
studies did not have explicit hypotheses.
Multiple stressors, including combinations of chemical and physical
stressors, required consideration in a number of the case studies.
Reviewers suggested two important questions to consider when
identifying stressor(s), as follows:
• Have all the relevant (or at least the most important)
anthropogenic stressors been identified? Were naturally
occurring stressors considered?
• Were the criteria used to identify the stressors appropriate and
defensible?
Typically, it is easiest to identify appropriate stressor(s) for risk
assessments leading to an explicit management decision associated
with an individual stressor. Examples include the Synthetic
Pyrethroids (pesticides), Acidic Deposition (hydrogen ions), and Crop
Loss (ozone) case studies. However, even in such cases the
relationship between observed effects in the field and the explicit
cause of these effects may be incompletely understood, suggesting that
other stressors may have influenced the observed effect.
It is more challenging to select the most relevant or important
stressors for assessments involving multiple stressors. The kinds of
information typically available to the assessor may include
observational data on effects (e.g., fish or birds have died or
chemicals are present in environmental media). These observations
do not always lead to clear cause-and-effect relationships. In aquatic
systems, for example, hypoxic events can lead to fish kills or
alterations of benthic habitat unrelated to the presence of toxic
chemicals in the sediments. Such effects might also occur in sediment
bioassays if elevated levels of naturally occurring ammonia are
present. These complications may have been present in the Bay
Drums case study but were not explicitly considered as stressors.
In some cases, stressors known to be present may be ignored because
of lack of information about the effects of the stressor. For example,
the Kesterson case study correctly focused on the element selenium
-------
Selection of Ecological
Components Should Be
Based on Ecological
Principles and Human
Values
The Possibility of
Indirect Effects Should
Be Considered When
Ecological Components
Are Selected
but chose not to look at boron because information on effects was
lacking. The reviewers felt that boron could have been important in
the overall management decision and that because the scope of this
case study was extensive, it would have been appropriate to develop
the information necessary to include boron in the assessment.
The case studies illustrate that several different ecological components
may be selected, depending on the focus of the study. These
components could range from one or more species (e.g., Rocky
Mountain Arsenal case study) to different communities (e.g.,
Wetlands Loss case study) or ecosystems. Selection of components is
typically based on several criteria, although these criteria were not
often explicitly stated in the case studies. Factors that influenced the
selection of ecological components include:
• the nature of the stressor and the potential for the stressor to
interact with the ecological component (as illustrated by use of
both aquatic biota and waterfowl in the Synthetic Pyrethroids
case study);
• the value of the ecological component from an ecological or
ecosystem perspective (as illustrated by the focus on trees in
the Wetlands Loss case study); and
• the value of the ecological component from a human
perspective. Examples from the case studies include:
(1) rare, threatened, or endangered species such as the bald
eagle in the Rocky Mountain Arsenal case study and the San
Joaquin Kit Fox in the Kesterson case study;
(2) species of commercial importance such as crop species in
the Crop Loss case study, English sole in the Commencement
Bay case study, and salmon in the Quartz Hill case study; and
(3) species of recreational importance such as freshwater fish
hi the Acidic Deposition and Water Quality Criteria case
studies and waterfowl in the Synthetic Pyrethroids case study.
The Synthetic Pyrethroids case study illustrates the importance of
considering ecological components that are indirectly affected by the
stressor. In this case study, waterfowl were included as an ecological
component, not because of any direct risk but because they might be
affected indirectly by a reduced prey base. Indirect or secondary
effects also were considered in the Wetlands Loss case study. This
case study evaluated the direct effects of an increase in water level
-------
(the stressor) on trees and also included wildlife species dependent on
the vegetation as ecological components of concern.
Defining Assessment
and Measurement
Endpoints and Their
Interrelationships Is
Essential
Some ecological components are considered especially important
because other components depend on them for habitat or food. These
include foundation or keystone species. In some cases, the presence
of such components is easily recognized. In others, however, the
relationships among species may not be understood. Reviewers of the
case studies felt it was important to identify these key ecological
components. Examples of such components include specific trees in
the Wetlands Loss case study, wetlands in the Bay Drums case study,
and benthic invertebrate communities in several of the case studies.
Reviewers of the Acidic Deposition case study felt that this
assessment may have focused too narrowly on survival of fish species;
the prey base (e.g., zooplankton) was not specifically identified as a
receptor.
There was some confusion among the case study authors (and
reviewers) over the meaning of the terms "assessment endpoint" and
"measurement endpoint" (see part I, section 3 of this report for
definitions). The case studies illustrated that defining the assessment
endpoint and selecting appropriate measurement endpoints are critical.
Selection of assessment endpoints is another area that requires
discussions with the risk manager and others. The Acidic Deposition
and Crop Loss case studies provide good examples of well-defined
assessment endpoints that are clearly related to measurement
endpoints.
Multiple Measurement
Endpoints Help
Evaluate "Overall
System Integrity"
When the assessment endpoint was the overall health or integrity of
the system, reviewers found that case studies using multiple
measurement endpoints were better able to assess risks than those
using only single measurement endpoints. One example is the
Commencement Bay case study, which includes field and laboratory
observations of fish and invertebrate species as well as chemical
measurements. In addition, this case study uses comparisons between
reference and affected areas. The use of several laboratory toxicity
test and field observation methods was critical to providing a more
nearly complete picture of the nature of exposure and effects in this
case study.
2.2.2. Analysis
Analysis includes the technical evaluation of data on both potential exposure to stressors
(characterization of exposure) and the effects of stressors (characterization of ecological effects).
-------
Characterization of exposure involves predicting or measuring the spatial and temporal distribution
of a stressor and its co-occurrence or contact with the ecological components of concern, while
characterization of ecological effects involves identifying and quantifying the effects elicited by a
stressor and, to the extent possible, evaluating cause-and-effect relationships.
Characterization of Exposure
Selection of Exposure
Models Depends on the
Purpose of the Risk
Assessment and
Available Resources
Validation of Models Is
Important for Reducing
the Uncertainty of
Exposure Estimates
Estimating or representing the exposure regime appears to be one of
the most technically challenging tasks in an ecological risk
assessment. Most of the case studies rely on simple models or
measurements, and the reviewers frequently commented on the
apparent oversimplifications in characterizing exposure.
A few of the case studies use simulation or complex fate-and-transport
models in an effort to represent the complex processes in real-world
exposure regimes. The selection of appropriate models is a matter of
scientific judgment combined with an appreciation of the overall
question to be addressed, and a recognition of the available resources.
Fate-and-transport models range from simple representations to
complex numerical models. For example, in the Synthetic
Pyrethroids case study, simple algebraic models are used to calculate
migration of the pesticide in runoff and in drift. In contrast, the
Acidic Deposition case study uses a combination of sophisticated fate-
and-transport models along with field measurements to estimate
exposure. Reviewers observed that fate-and-transport models should
be appropriate for the goals and scope of the study.
Simulation models have been developed to represent exposure regimes
and to relate these to certain systemwide effects. The FORFLO
model used in the Wetlands Loss case study is an example of a
simulation model that attempts to represent an exposure regime for a
physical stressor (change in water level). Hydrologic characteristics
of the environment are incorporated into the model.
Reviewers of the case studies that used models noted the importance
of verification and "reality checks" on the exposure estimates. The
Synthetic Pyrethroids case study uses simple models to project the
transport of the chemicals and resultant concentrations in surface
waters. However, no field data are provided to support these
estimates. Additional field studies are being planned so these data
may be generated hi the future.
The Acidic Deposition case study includes verification of model
estimates. The procedures include making hindcasts of historical
changes in surface water chemistry as a function of historical changes
10
-------
When Available, Life
History and Behavioral
Information Can Be
Useful in
Characterizing
Exposure
Food Chain and
Pathway Analyses
Can Be Useful for
Evaluating Indirect
Effects
in acidic deposition rates. These hindcasts are compared with
paleolimnological reconstructions using algal fossils. A watershed
chemistry model is calibrated using several watersheds, and then
confirmed through blind simulations using only input data. Analyses
are also conducted to identify those input variables and parameters to
which the models are sensitive. Reviewers felt that such sensitivity
analyses should be more widely used.
The Kesterson case study illustrates how life history and behavioral
information can be utilized as part of exposure characterization. For
each species in the Kesterson case study, the scientific literature was
reviewed to quantify food habits, and dietary preferences were
summarized by life stage, sex, and season, as appropriate. The home
range of each species was estimated from literature values. Diet
factors were used to model the fraction of the whole diet of each
organism contributed by each compartment of the simplified selenium
transfer diagram. Diet factors were based on species' food
preferences as described in the literature, on their home range relative
to the size of the Kesterson Reservoir area, and on the relative
abundance of different types of prey species. For example, for
species that consumed aquatic invertebrates, the relative abundance of
herbivorous and carnivorous aquatic invertebrates at the Kesterson
Reservoir was used to specify the composition of these organisms in
the diet.
Good examples of food chain and pathway analyses may be found in
the Rocky Mountain Arsenal and Kesterson case studies. In addition
to providing exposure-relevant information, evaluating food chain
relationships can reveal potential indirect or secondary effects on
higher trophic levels due to reduction in the prey base. The Synthetic
Pyrethroids case study considers such secondary effects.
Food chain and pathway analyses have been used in case studies
where wildlife species (including endangered species) have been
selected as the ecological components of concern. For example, a
major feature of the Rocky Mountain Arsenal risk assessment is the
development and use of a pathways model to establish a quantitative
relationship between concentrations of bioaccumulative contaminants
in abiotic media and concentrations at different trophic levels in
aquatic and terrestrial food webs.
For the Kesterson case study, seven representative wildlife species in
the area were selected. Food chain exposure was considered the most
important pathway for exposure of fish and wildlife to selenium, and
detailed food chain exposure diagrams for each of the selected species
were developed into simplified selenium transfer models. These
11
-------
models were used with transfer factors derived from chemical
measurements in media and tissues and a Monte Carlo simulation
technique, to estimate the probability distribution of selenium
concentrations in the diets of the key species under each of the three
remedial alternatives.
Similar Exposure The Wetlands Loss case study illustrates that the exposure associated
Principles May Be with a physical stressor (changes in water elevation) can be
Applicable to Physical characterized using principles similar to those employed in more
and Chemical Stressors familiar cases involving chemical stressors. In this case study,
exposure-effects relationships were developed between the physical
stressor and habitat alterations associated with changes in forest
vegetation.
Characterization of Ecological Effects
Information Needs The case studies illustrate that ecological effects may be characterized
Should Be Determined using the various kinds of information given below.
by the Purpose and
Scope of the Risk • Literature Values. Criteria, or Guidelines. A few case studies
Assessment (e.g., Quartz Hill) rely almost exclusively on existing
literature values, criteria, or guidelines, while others
(Kesterson, Bay Drums, and Wetlands Loss) incorporate such
information into the assessment along with the results of
laboratory and field tests.
« Laboratory and In-Field Exposure-Response Studies.
Laboratory studies include acute and chronic toxicity tests in
which organisms are exposed to individual or multiple
stressors (e.g., complex mixtures). Data from both laboratory
and field studies are included in the Bay Drums and Synthetic
Pyrethroids case studies. The Crop Loss case study uses field
enclosures of crop species to develop concentration-response
data for ozone exposures.
• Field Studies and Surveys. The Coke Plant and
Commencement Bay case studies include examples of these
approaches.
The reviewers commented on the advantages and disadvantages of
using different methodologies and information sources. A key point
is that the information or methods used should be appropriate for the
assessment. In general, the reviewers found that using a suite of
methods (literature values, bioassays, field studies) provided a more
complete characterization of ecological effects than relying on a single
measure or literature value.
12
-------
Site-Specific Criteria
Are Useful Tools
Criteria Should Be
Established for
Selecting Appropriate
Reference Areas for
Field Studies
Establishing Causality
Is Complicated by
Multiple Stressors
Many case studies used criteria or benchmarks, i.e., chemical
concentrations in environmental media (food, water, soil, sediment),
below which minor or no effects are anticipated for a particular
measurement endpoint. These measures can be improved by taking
into account important site-specific factors related to the toxicity or
other effects of the stressor upon the selected ecological components.
Site-specific factors may not be accounted for in national criteria.
The Water Quality Criteria case study explains how site-specific
cadmium criteria for protecting aquatic organisms in the St. Louis
River were derived. Another example is the Apparent Effects
Threshold (AET) values developed for chemicals in sediments for the
Commencement Bay case study. AET values are site-specific
concentration limits below which no effects are expected. They are
derived by combining information from field studies and laboratory
sediment bioassays.
Field studies sometimes compare potentially affected areas with
reference sites as a basis for characterizing effects. This approach is
used in the Bay Drums and Coke Plant case studies, but the
comparison between reference and test sites was difficult to interpret
because the reference areas themselves had potential stressor impacts.
An additional reference area had to be included in the Commencement
Bay case study because of difficulties with the originally selected
reference site.
Reviewers of case studies that included comparisons between affected
and reference areas noted the importance of developing synoptic
information on the observed ecological effects and on the presence
and magnitude of the stressors. A clear set of selection criteria
should be developed for reference areas, and selection should be
based on these criteria. When available reference areas have
limitations, these limitations should be recognized at the onset. In
such cases, it may be necessary to use more than one reference area.
In some case studies, it is difficult to link specific stressors with
observed ecological effects because of confounding factors such as the
presence of mixtures of chemicals, or a combination of chemical and
nonchemical stressors. For example, it was difficult to link stressors
and effects in the Commencement Bay case study because the
sediments contained numerous chemicals. In the Bay Drums case
study, physical factors (wetland alterations) and natural biochemical
factors (anaerobic sediments) confounded the interpretation of some of
the field and laboratory effects data.
13
-------
Ecological Effects Several case studies illustrate how models can be used effectively to
Models, if Applied relate ecological effects and stressors.
Correctly, Are Useful
for Relating Effects to • The Wetlands Loss case study links the FORFLO model,
Stressors which predicts changes in forest vegetation caused by changes
hi water level, to Habitat Suitability Indices (HSI), which
relate changes in wildlife to alterations in forest vegetation.
• The Crop Loss case study develops a model that relates plant
yield to ozone exposure.
• The Acidic Deposition case study models the relationship
between water chemistry variables (i.e., pH and
concentrations of aluminum and calcium ions) and the
probability of the presence or absence of fish for selected
species.
2.2.3. Risk Characterization
Risk characterization uses the results of the exposure and ecological effects analyses to
evaluate the likelihood that adverse ecological effects are occurring or will occur in association
with exposure to a stressor. A risk characterization highlights summaries of the assumptions,
scientific uncertainties, and strengths and weaknesses of the analyses. Finally, a risk
characterization discusses the ecological significance of the risks with consideration of the types
and magnitudes of the effects, their spatial and temporal patterns, and the likelihood of recovery.
Risks Can Be The Framework Report (U.S. EPA, 1992) notes that risk
Characterized Both characterization may be qualitative or quantitative and that it
Qualitatively and frequently relies heavily on scientific judgment. Opinions of the case
Quantitatively study reviewers about what risk characterization should be varied.
Some felt that risk characterization applies only to situations where
predictive, probabilistic statements can be made about future events.
Others held a broader view, in which risk characterization includes
either qualitative or quantitative statements of risk and involves
evaluating the causal relationship between stressors and effects for
existing situations as well as predicting the risk of future events.
Qualitative analyses are used in the Quartz Hill and Synthetic
Pyrethroids case studies. The Quartz Hill case study compares the
potential effects of mine tailings disposal in two fjords to determine
which fjord would be more at risk. Reviewers were concerned that
the absolute risks of disposal were not considered, but did agree that
the comparative analysis was useful for assessing the relative risk of
the mine tailings to the two fjords.
14
-------
The Toxicity Quotient
Method Is Frequently
Used
Exposure-Response
Models Offer Certain
Advantages
The Synthetic Pyrethroids case study includes a qualitative assessment
of the potential secondary effects of a reduced prey base on
waterfowl. Synthetic pyrethroids are known to be toxic to aquatic life
that waterfowl rely upon for food, and the reviewers agreed that
secondary effects on waterfowl were possible. However, a
quantitative analysis was not conducted, and the reviewers felt that the
qualitative analysis of the effects related to prey base was one of the
weaker parts of the case study.
The Toxicity Quotient Method is a simple method used in several case
studies to compare exposure levels with benchmark or criteria effect
levels. If the ratio exceeds "1," some potential for risk is presumed.
In addition to the results of Quotient Method comparisons, a complete
risk characterization includes the scientific uncertainties and
assumptions, underlying the assessment.
The Bay Drums, Coke Plant, and Water Quality Criteria case studies
all compare chemical concentrations in environmental media with
benchmark concentrations. For water samples, the benchmark
concentrations were often the ambient water quality criteria (AWQC)
or a screening value obtained from the relevant EPA AWQC
document. In the Bay Drums case study, chemical concentrations in
sediments are compared with toxic sediment concentrations reported
in the literature.
The Carboruran case study uses the Quotient Method in a somewhat
different manner. Risks are evaluated, in part, by comparing the
estimated levels of carbofuran granules per square foot of surface soil
with the granule dose estimated to kill 50 percent of exposed birds
The Rocky Mountain Arsenal case study also uses the Quotient
Method to evaluate risks associated with dietary exposure. This case
study uses an exposure pathway analysis to estimate numerical criteria
for potential exposure pathways, including dietary exposure
(bioaccumulation) and surface water ingestion. (A calibration and
validation process is being developed to reduce uncertainties in this
approach.) The Kesterson case study uses a similar pathway analysis
to determine levels of selenium in sediments that would cause adverse
effects in selected fish and wildlife species.
The major advantage of exposure-response models is that risks and
associated uncertainties can be quantified and the results presented to
risk managers in a form that allows comparison of alternatives.
Forecasted conditions can give more meaningful information to
managers than the simple numerical comparisons of the Quotient
15
-------
Risks May Be
Characterized at
Various Levels of
Ecological Organization
Method. For example, in the Crop Loss case study, reductions
(losses) in commercial crop yields can be related to specific ozone
levels which, hi turn, can be considered in formulating air quality
standards. The major disadvantage of using exposure-response
models is that adequate validation and sensitivity analyses are
frequently not available.
; • . I
The complexity of risk characterization increases with increasing
levels of ecological organization. For individuals of a species, it is
possible to determine if some will be at risk as a result of a particular
exposure. To estimate population-level risks, some combination of
field studies or population models is needed. Although population
models are available, they were generally not used in the case studies.
Instead,, most of the case studies inferred that population-level effects
could occur if there were risks to individuals. In the Carboftiran case
study, for example, evidence of mortality of individual birds was
considered sufficient to demonstrate adverse effects even without a
formal assessment of population-level.risks.
Risks to communities were assessed by considering species
representative of various trophic groups, taxa, or habitats. While
several case studies (e.g., Bay Drums, Commencement Bay, Coke
Plant) evaluate changes in the structure of benthic macroinvertebrate
assemblages, interactions among the organisms are not determined as
frequently. The Synthetic Pyrethroids case study does consider
reduction of the prey base as a possible risk to waterfowl, and the
Wetlands Loss case study uses models to relate predicted changes in
vegetation to risks to birds and mammals.
Ecosystem-level risks, which might involve changes in functional
processes such as productivity, nutrient cycling, or decomposition, are
seldom considered. The Acidic Deposition case study comes closest
to ecosystem-level risk characterization. The ecological components
for this case study include freshwater streams and lakes.
Some reviewers felt that certain case studies did not consider all the
interactions that may be present within an ecosystem and that the case
studies were too narrowly focused. Discussion of this point led to the
question "What is important?" It was recognized that this question
must be addressed from two sides:
1. What is important from a management standpoint? What
information is needed to reach a sound decision?
16
-------
Risk Assessors Should
Provide a Complete
Picture of the Risk
Assessment
2. What is important from an ecological standpoint? What
information must be brought to the attention of the risk manager
to aid in an informed decision?
It was clear from the discussions that ecological risk assessment does
not always mean ecosystem risk assessment. The selected ecological
components and methodologies must be appropriate to the ultimate
risk management decision.
The manner in which results are presented to the risk manager or to
those affected by the risk management decision can affect the
perception of the risk. For example, the Acidic Deposition case study
presents the percentages of lakes and streams at risk. Reviewers
noted that this has a different impact on the reader than if the absolute
number of affected lakes and streams are presented. The risk
assessor should know who the users of the risk assessment will be and
should present the results in a manner that is appropriate for them.
To provide a complete picture and accommodate users with varying
perspectives, it may be necessary to present results in a variety of
formats.
17
-------
3. KEY TERMS (U.S. EPA, 1992)
assessment endpoint—An explicit expression of the environmental value that is to be protected.
characterization of ecological effects—A portion of the analysis phase of ecological risk assessment
that evaluates the ability of a stressor to cause adverse effects under a particular set of
circumstances.
characterization of exposure—A portion of the analysis phase of ecological risk assessment that
evaluates the interaction of the stressor with one or more ecological components. Exposure
can be expressed as co-occurrence, or contact, depending on the stressor and ecological
component involved.
conceptual model—The conceptual model describes a series of working hypotheses of how the
stressor might affect ecological components. The conceptual model also describes the
ecosystem potentially at risk, the relationship between measurement and assessment
endpoints, and exposure scenarios.
ecological component—Any part of an ecological system, including individuals, populations,
communities, and the ecosystem itself.
ecological risk assessment—The process that evaluates the likelihood that adverse ecological effects
may occur or are occurring as a result of exposure to one or more stressors.
exposure—Co-occurrence of or contact between a stressor and an ecological component.
measurement endpoint—A measurable ecological characteristic that is related to the valued
characteristic chosen as the assessment endpoint. Measurement endpoints are often
expressed as the statistical or arithmetic summaries of the observations that comprise the
measurement.
risk characterization—A phase of ecological risk assessment that integrates the results of the
exposure and ecological effects analyses to evaluate the likelihood of adverse ecological
effects associated with exposure to a stressor. The ecological significance of the adverse
effects is discussed, including consideration of the types and magnitudes of the effects, their
spatial and temporal patterns, and the likelihood of recovery.
stressor—Any physical, chemical, or biological entity that can induce an adverse response.
18
-------
4. REFERENCES
U.S. Environmental Protection Agency. (1991) Summary report on issues in ecological risk
assessment. Risk Assessment Forum, Washington, DC. EPA 625/3-91/018.
U.S. Environmental Protection Agency. (1992) Framework for ecological risk assessment. Risk
Assessment Forum, Washington, DC. EPA 630/R-92/001.
19
-------
PARTEL THE CASE STUDIES
The case studies included in
this section follow the format shown
hi the box on the right. When reading
the case studies, it is important to
keep several points in mind:
• The original case studies
were not developed as risk
assessments as defined in the
Framework Report. The
Framework Report was not
available when the case studies
were conducted, written, or
reviewed. Fortunately, the
overall concepts of ecological
risk assessment applied by the
authors and reviewers were
compatible with the broad
principles (if not the details)
described in the Framework
Report. EPA notes that the
case studies are often partial
risk assessments that focus on
available information without
discussing other relevant
considerations such as the
uncertainties defined by a
limited data base.
At the workshops, each case
study was evaluated as to
whether it (1) effectively addressed the generally accepted components of an ecological risk
assessment, or (2) addressed some but not all of these components or, instead, (3) provided
an alternative approach to assessing ecological effects.
• The strengths and limitations of each case study are highlighted in the comment boxes.
Comments made by both the peer reviewers and the case study authors are included.
• The authors who compiled the case studies did not necessarily conduct the research
upon which the case studies are based. References to the original research are provided
in each case study.
Case Study Format
Abstract. The abstract summarizes the major
conclusions, strengths, and limitations of the
case study.
Risk Assessment Approach. This section
clarifies any differences between the
ecological risk assessment approach used in
the case study and the general process
described in the Framework Report.
Statutory and Regulatory Background. The
statutory requirements for the study are
described along with any pertinent regulatory
background information.
Case Study Description. This contains the
technical description of the case study,
organized according to the phases of
ecological risk assessment described in the
Framework Report: problem formulation,
analysis (characterization of exposure and
characterization of ecological effects), and
risk characterization. A comment box is
included at the end of each major section.
References.
20
-------
General characteristics of the case studies are summarized in table 1 (in part I), and a list
of selected ecological risk assessment techniques that were used in the case studies is provided in
table 2. The list is not meant to be comprehensive. Rather, it provides examples of different
methods and models that were used in the various phases of the ecological risk assessment process.
The application of these techniques may be reviewed by referring to the case studies in which they
are used. Case studies are referenced by the section of this report in which they appear. (The
corresponding titles of the case studies are given in table 1.)
21
-------
Table 2. Selected Case Study Methods and Models
METHOD
CHARACTERIZATION OF EXPOSURE
Estimating strcssor levels
in the environment
Aquatic
SWRRB
EXAMS
Steady-State
Oceanographie Model
MAGIC, ILWAS
to
10 Mass Balance Study
Hydrological Assessment
Acid Deposition Models
RADM
NADP/NTN
Steady-State Models
Empirical Acidification
Models
Sediment Chemistry
Terrestrial
Kriging
DESCRIPTION
Simulator for Water Resources in Rural Basins
Exposure Analysis Modelling System
Based on distribution of natural and anthropogenic
conditions using Monte Carlo simulations
Dynamic acidification models
Prediction of sediment concentrations in relation to
source loading, sedimentation rates and mixing,
biodegradationand diffusion
Surface and ground-water flow models
Regional Acid Deposition Model
National Acid Deposition Program/
National Trends Network
Relate deposition to lake acid neutralizing capacity and pH
Relate deposition to lake acid neutralizing capacity and pH
National Ozone Data Base-300 sites (SAROAD)
Deposition surface for loading to individual sites
CASE STUDY
REPORT SECTION
REFERENCE fPART HI
Arnold et al., 1990 10
Burns, 1990 10
7
Thornton et al., 1990 1
5
Standard EPA Protocols 2
Thornton etal., 1990 1
1
Thornton etal., 1990 1
Thornton etal., 1990 1
5
Lefohn et al., 1987 6
NAPAP, 1990 1
Nomograph
Estimation of pesticide residues found in wildlife
Urban and Cook, 1986
10
-------
Table 2. Selected Case Study Methods and Models (continued)
METHOD
DESCRIPTION
REFERENCE
CASE STUDY
REPORT SECTION
(PARTIT)
Multimedia
Exposure Algorithms
Estimating strestar levels
in biota
Aquatic
Bioaccumulation
Terrestrial
Normograph
Multimedia
Food Chain Models
CHARACTERIZATION OF EFFECTS
Aquatic
Organism/Population Level
Laboratory Toxioity
Tests
Determination of estimated environmental concentrations
(EEC) due to pesticide runoff and drift
Organics
Metals from mine tailings
Selenium
Organics and inorganics in fish and invertebrates
Calculation of pesticide residues in terrestrial food webs
Calculated selenium transfer factors from experimental
data and Monte Carlo simulations
Pesticide assessment protocols
Sediment bioassays with amphipods,
oyster larvae, and Microtox
Chemical mixtures
Clam burrowing behavior
Phytoplankton growth
Critical pH levels
Survival, growth, and reproduction of resident fish
species
Unpublished
U.S. EPA, 1985
U.S. EPA, 1988a
Saiki, 1986
Ohlendorfetal., 1986b
Standard EPA Protocols
Urban and Cook, 1986
USBR, 1986a
Preston and Hitch, 1982
Williams etal., 1986
Standard EPA Protocols
U.S. EPA, 1988a
U.S. EPA, 1988a
Baker etal., 1990
Turner etal., 1990
Speharetal., 1985
Carlson et al., 1984
10
10
10
5
2
7
7
1
11
-------
Table 2. Selected Case Study Methods and Models (continued)
METHOD
Condition Factors
Fish Hiitopathology
Fish Population Model
Community/Ecosystem Level
Aquatic Mesocosms
Pond Studies
WET Model
Macroinvertebrate
Community
Fish, Macroinvertebrate,
and Plankton Communities
Recolonization
Terrestrial
Organism/Population Level
Avian Laboratory
Toxicity Tests
Pathological Examinations
Field Observations
Open-Top Field
Chambers
Acetylcholinesterase
Activity
DESCRIPTION
Based on weight and length of individual fish
Examination of liven
Examination for tumors, lesions, and other anomalies
Logistic regression equation for presence/absence of fish
Impacts on invertebrate communities .
Impacts on fish and invertebrates
Wetland Evaluation Technique for assessing function and
values of wetlands
Species diversity index: Sorrenson's quotient and similarity
Abundance, species richness, Bray-Curtis similarity
Effects of mine tailings
Effects of catbofuran
Pesticide assessment protocols
Hatchability, deformities, and mortality
Rare and endangered species
Foliar injury, growth, reproduction, yield,
and physiological responses
REFERENCE
U.S. EPA, 1985
Baker etal., 1990
U.S. EPA, 1990
U.S. EPA, 1990, 1991
Adamus et al., 1987
U.S. EPA, 1985
Davis and Lathrop, 1991
U.S. EPA, 1988a
Urban and Cook, 1986
Preston and Hitch, 1982
Ohlendorfetal., 1986
Not included
Heck et al., 1991
Heagle and Heck, 1980
Robinson et al., 1988
CASE STUDY
REPORT SECTION
fPARTID
4
5
4
1
10
10
2
2
5
4
7
3
10
9
9
2
6
8
-------
Table 2. Selected Case Study Methods and Models (continued)
METHOD
Samples of Chance
Earthworm Population Studies
Avian Reproductive
Success
Population Studies
Multimedia
Organism/Population Level
Habitat Suitability
Index
Community/Ecosystem Level
FORFLO
RISK CHARACTERIZATION
Quotient Method
Ambient Water
Quality Criteria
Reference Indices
Apparent Effects
Threshold
DESCRIPTION
Necropsies on dead or dying organisms
Percent nests hatched and fledged, egg weight, volume,
dimensions, and thickness
Species occurrence, population density, and age-class
structure: snails, earthworms, vegetation, ducks and
coots, and prairie dogs
HSI model for present and future capability of a site to
provide basic habitat requirements for wildlife
indicator species
Bottomland forest succession model that simulates the
growth, reproduction, and competition of a mixed-tree
species forest stand
Effects concentration/expectedexposure concentration
Defining extent of area exceeding AWQC
Recalculation, indicator species, and resident species
procedures were used to modify national cadmium criteria
Elevation above reference indices
Based on amphipod mortality, oyster larvae abnormalities,
and bentbic macroinvertebratetaxa abundance
CASE STUDY
REPORT SECTION
REFERENCE (PART m
8
8
ESE, 1989 8
8
USFWS, 1981 12
Pearlstineetal., 1985 12
Dewitt, 1966 10
3
4
7
Spehar and Carlson, 1984a, b 11
U.S. EPA, 1989b 5
USER, 1986 9
U.S. EPA, 1989b 5
USBR, 1988 9
Weibull Model
Predicted yield losses based on ozone exposure
Somerville et al., 1989, 1990
-------
Table 2. Selected Case Study Methods and Models (continued)
CASE STUDY
REPORT SECTION
METHOD DESCRIPTION REFERENCE (PARTm
Pathways Analysis Criteria developed based on characterization of effects Fordham and Reagan, 1991 8
and observed environmental concentrations using a
modified food chain model
Wasteload Allocations Setting limits on wastewater loads and nonpoint source U.S. EPA, 1991 8
allocations
Indicator Species Takes into account factors that affect the bioavailability 8
Procedure and/or toxicity of stressors in characterizing risk
Weight-of-Evidence Multiplicity of evidence supporting hypothesis at a site 2
Approach where many potential stressors existed
Process of elimination of alternative stressors and NAPAP, 1990 1
corroborative support for acid deposition
K> : ; —
-------
SECTION ONE
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
EFFECTS OF ACIDIC DEPOSITION ON AQUATIC ECOSYSTEMS—
A REGIONAL PROBLEM
-------
AUTHOR AND REVIEWERS
AUTHOR
Kent W. Thornton
FTN Associates, Ltd.
Little Rock, AR
REVIEWERS
Thomas M. Frost (Lead Reviewer)
Center for Limnology
University of Wisconsin
Madison, WI
Nancy A. Bryant
ENSR Consulting & Engineering
Acton, MA
Kenneth L. Dickson
Institute of Applied Sciences
University of North Texas
Denton, TX
Judy L. Meyer
Institute of Ecology
University of Georgia
Athens, GA
Douglas P. Ormrod
Office of Graduate Studies
University of Guelph
Guelph, Ontario, Canada
Randall S. Wentsel
Toxicology Division
U.S. Army Chemical Research,
Development, and Engineering Center
Aberdeen Proving Grounds, MD
1-2
-------
CONTENTS
ABSTRACT 1-6
1.1. RISK ASSESSMENT APPROACH 1-7
1.2. STATUTORY AND REGULATORY BACKGROUND 1-7
1.3. CASE STUDY DESCRIPTION 1-7
1.3.1. Problem Formulation 1-7
1.3.2. Analysis: Characterization of Ecological Effects 1-12
1.3.3. Analysis: Characterization of Exposure 1-15
1.3.4. Risk Characterization 1-21
1.4. REFERENCES 1-30
APPENDIX A—SCIENTIFIC CONTRIBUTORS TO THE 1990
NAPAP AQUATIC ASSESSMENT 1-A1
1-3
-------
LIST OF FIGURES
Figure 1-1. Structure of analysis for effects of acid deposition 1-8
Figure 1-2. Steps in a regional ecological risk assessment 1-9
Figure 1-3. Critical pH levels for selected organisms 1-13
Figure 1-4. Relationship between median wet sulfate deposition
and median surface-water sulfate concentrations
in NSWS subregions 1-18
Figure 1-5. Median change in projected ANC for 50-year
MAGIC simulations versus median change in
sulfur deposition for each deposition scenario
and subregion 1-19
Figure 1-6. Emissions of sulfur dioxide from electric utilities 1-20
Figure 1-7. Percentage of Mid-Atlantic Highland streams with
(a) ANC <0 /ieq/L, (b) pH <6, and (c) chemistry
unsuitable for brook trout based on MAGIC
projections for 50 years under illustrative
deposition scenarios 1-24
Figure 1-8. Median and range of change in ANC over 50 years for
MAGIC projections for 35 lakes hi the Adirondack^
and for Cheney Pond, NY 1-25
LIST OF TABLES
Table 1-1. Percentage Estimates of Number of Lakes and
Streams With ANC and pH Below Three Arbitrary
Reference Values for Each Measure : . . . . 1-16
Table 1-2. Average Percentage of Lakes and Streams With
Chemistry Unsuitable for Sensitive Fish Species
Over 50-Year Projections Under Illustrative Deposition
Scenarios for Each Region 1-22
1-4
-------
LIST OF ACRONYMS
AERP Aquatic Effects Research Program
ANC acid-neutralizing capacity
ASI acidic stress index
CAAA Clean Air Act Amendments
DOC dissolved organic carbon
EPA U.S. Environmental Protection Agency
ILWAS Integrated Lake/Watershed Acidification Study
MAGIC Model of Acidification of Ground Water in Catchments
NADP/NTN National Acid Deposition Program/National Trends Network
NAPAP National Acid Precipitation Assessment Program
NSWS National Surface Water Survey
RADM Regional Acid Deposition Model
SBRP Southern Blue Ridge Province
SOS/T State-of-Science/Technology
USGS U.S. Geological Survey
1-5
-------
ABSTRACT
The National Acid Precipitation Assessment Program (NAPAP) is a 10-year congressionally
mandated research program to assess the effects of acidic deposition on the environment. The U.S.
Environmental Protection Agency (EPA) directed the Aquatic Effects Research Program (AERP) to
assess the regional effects of acidic deposition on lakes and streams in the United States. The AERP
designed and implemented a strategic assessment plan that implicitly followed the steps in risk
assessment. A National Surface Water Survey (NSWS) determined the percentage and extent of
lakes and streams that were acidic or potentially susceptible to acidic deposition (problem
formulation). It indicated that not all aquatic systems hi the surveyed regions were susceptible to
acidic deposition. For regions with subpopulations of aquatic systems at risk, total acidic deposition
estimates were determined from field measurements and model simulations (e.g., wet deposition
estimates from the National Acid Deposition Program [NADP] monitoring network and dry
deposition estimates from the Regional Acid Deposition Model [RADM]) (characterization of
exposure). Watershed chemistry models simulated the effects of acidic deposition on watershed-
lake/stream chemistry, and fish-response models determined the effects of surface-water acidification
on fish loss (characterization of ecological effects). The final step was to assess risk to aquatic
systems (both changes in water chemistry and in fish response) projected to occur under alternative
sulfur dioxide emission-control and no-emission-control scenarios (risk characterization). These
scenarios involved coupling emission models, the RADM, watershed-lake/stream chemistry models,
and fish-response models with associated error bounds. This effort represents one approach to
addressing a regional problem in an ecological risk assessment framework.
1-6
-------
1.1. RISK ASSESSMENT APPROACH
Although the risk assessment framework was not used in designing the National Acid
Precipitation Assessment Program (NAPAP), all framework elements are included in the program's
various phases (figure 1-1). Therefore, this case study can serve as a fairly complete example of the
framework's application. In this study, the ecological components are lakes and streams that may be
acidified. The stressor is acidic deposition.
1.2. STATUTORY AND REGULATORY BACKGROUND
Congress authorized the NAPAP under the Acid Precipitation Act of 1980 (P.L. 96-294,
Title VII) from concern that acidic deposition might have adverse effects on aquatic systems, forests,
agricultural crops, construction materials, cultural resources, atmospheric visibility, and human
health. NAPAP was given the statutory responsibility to prepare comprehensive scientific,
technological, and economic information to assist legislators and other decision-makers in developing
policies to control acidic deposition. NAPAP, charged with conducting a 10-year program of
research and assessment, investigated the causes and effects of acidic deposition and analyzed
alternative strategies to control or mitigate these effects. The program was responsible for
coordinating and collaborating with other relevant foreign and domestic research activities. The act
required the NAPAP to provide the President, Congress, and the public with annual reports and a
final 1990 Integrated Assessment. NAPAP was not to recommend specific emission-control levels
or targets; rather, it was to provide Congress, the President, and federal and state policy officials
with relevant information to use in formulating policy, legislation, and regulations.
1.3. CASE STUDY DESCRIPTION
This case study is based largely on analyses performed in conjunction with the NAPAP 1990
Integrated Assessment to evaluate the effects of acidic deposition on aquatic systems. The research
was conducted by the Aquatic Effects Research Program (AERP). See appendix A for a list of
scientific contributors. The U.S. Environmental Protection Agency (EPA) was the lead agency
responsible for the Aquatic Effects Task Group hi NAPAP. Four primary policy questions guided
the research and assessment efforts: (1) How extensive is the damage to aquatic resources due to
current and historical levels of acidic deposition; (2) what is the anticipated future damage to these
resources; (3) what levels of damage to sensitive surface waters are associated with various rates of
acidic deposition; and (4) what is the rate of change or recovery of affected systems if acidic
deposition rates decrease?
The general approach used in this assessment is shown schematically in figures 1-1 and 1-2
and described in the following sections.
1.3.1. Problem Formulation
In the 1960s and 1970s, Scandinavian scientists collected evidence of acidity in precipitation
and watershed runoff that was contributing to the acidification of lakes and streams (Aimer et al.,
1974; Gjessing et al., 1976; Hultberg, 1977; Henriksen, 1979; Okland, 1979; Drablos and Tollan,
1980). Parallel studies hi the United States, Canada, the United Kingdom, and Central Europe
1-7
-------
Figure 1-1. Structure of Analysis for
Effects of Acid Deposition
PROBLEM FORMULATION
Stressors: hydrogen ion derived from nitric and
sulfuric acids.
Ecoloaical Components: sensitive aauatic ecosvstems
A
including lakes and streams.
Endpoints: assessment endpoint is fish survival.
Measurement endpoint includes water chemistry (e.g.,
ANC, pH, DOC, Ca, Al).
II
1
|
Characterization of Characterization of
Exposure Ecological Effects
Models of acid deposition Laboratory and field
and watershed chemistry studies were used to
were used to predict examine fish response to
surface water conditions acidity. An acidic stress
for alternative emission index (ASI) was developed.
scenarios.
JL X
^ V
RISK CHARACTERIZATION
Emission, atmospheric, watershed, and fish response
models were coupled to provide estimates of the
proportion of water bodies that may be adversely affected
by acidic deposition under various scenarios.
Uncertainty was quantitatively evaluated using Monte Carlo
and other methods.
1-8
-------
Effects of Acidic Deposition on Aquatic Ecosystems
PROBLEM FORMULATION
All U.S. Lakes and Streams
Assessment Endpoint:
Fish, Amphibians,
Waterfowl-Survival
Measurement Endpoints:
ANC, pH, S04, Ca. Al
ASI
Receptors:
Lakes/streams
HAZARD
IDENTIFICATION
Target Population—
NSWS Lakes, Streams
ANALYSIS
ANC Screening
ANC <400 //eq/L
•Susceptible Target Population
ANC <200 //eq/L
Screening Sources of Acidity:
Acidic Deposition
Organic Acids
Land Use
Acid Mine Drainage
Internal Sources
•STRESSOR-
RESPONSE:
Paleolimnology
Statistical Analyses
Empirical Models
\Process-oriented Models
Target Population
Deposition
Dominated
Acidic Lakes,
Streams
Site Characteristics:
High Elevation
Small Watersheds
Granite Bedrock
Shallow Soils
Acidic Soils
Low % BS Soils
Low ANC
Low Interest Population
ANC >400 j/eq/L
or
Low Deposition Rate
Sensitive
Lakes & Streams
0< ANC <50 //eq/L
_|_ EXPOSURE
ASSESSMENT:
RADM
NADP/NTN
Paleolimnology
RISK CHARACTERIZATION
Future Projections
- Atmospheric Deposition Scenarios
- Chemistry Models
- Fish-Response Models
Resources at Risk
Figure 1-2. Steps in a regional ecological risk assessment
1-9
-------
during this period confirmed the existence of acidic deposition and its potential link to surface-water
acidification (Beamish and Harvey, 1972; Cogbill and Likens, 1974; Beamish et al., 1975; Davis
et al., 1980; Schindler et al., 1980; Burns et al., 1981; Altshuller and Linthurst, 1984). The
apparent loss of sport fish populations hi lakes hi areas of high acidic deposition, such as the
Adirondacks, added to mounting concerns (Schofield, 1976a, b).
Stressors. The primary components of acidic deposition that affect surface-water
acidification are sulfuric and nitric acids. The largest sources of sulfur compounds linked to acidic
deposition are coal and oil combustion in electric power-generating and industrial facilities. Sulfate
is the dominant acidic anion in wet deposition and, therefore, its effects received the greatest study
(National Academy of Sciences, 1984; National Research Council, 1986). Sulfuric and nitric acids
are strong acids that can cause changes in the hydrogen ion concentration (pH) hi surface waters,
with direct and indirect deleterious effects on aquatic organisms. Low pH (i.e., high hydrogen ion)
can be directly toxic to aquatic species, but also can indirectly affect the ecosystem by causing shifts
hi the food web (e.g., the loss of a prey species) (Schindler, 1988). Low pH also can mobilize
metals, such as aluminum, that are toxic to aquatic species, especially fish.
Ecological Components. The ecological components examined in this study include
"sensitive aquatic ecosystems." A measure used in the study to identify and characterize such
systems is acid-neutralizing capacity (ANC), the capacity of a system to buffer itself against changes
hi pH. The term alkalinity refers primarily to the neutralization of acids by the carbonate/
bicarbonate system hi freshwaters and was the common measure prior to NAPAP. ANC is the more
appropriate measure because it includes other proton acceptors besides carbonate/bicarbonate. In
general, alkalinity was measured before NAPAP, and ANC was measured during NAPAP. Acidic
lakes or streams, by definition, have ANC chronically less than or equal to zero. Chronically acidic
means that the annual average ANC concentration is less than zero; the term does not include
aquatic systems that have acidic episodes during storm or snow-melt events. Acidification refers to
the loss of ANC or the ability to decrease pH, and acidified refers to surface waters with previously
higher ANC or pH values that declined because of acidic inputs. A lake or stream might be
acidified but not be acidic. The effects of acidic deposition occur over decades rather than years and
over large geographic areas such as New England.
General attributes of sensitive aquatic ecosystems include granitic or noncalcareous bedrock,
shallow acidic soils with low base saturation, seepage lakes and small lakes or streams located in the
upper portions of the watershed, and systems with low ANC. These characteristics are common to
the European, Scandinavian, and North American lakes and streams cited earlier in the literature.
Various geologic and soil indices used these characteristics in describing regions potentially sensitive
to acidic deposition hi the United States (Norton, 1980; McFee, 1980).
Regions with low ANC/alkalinity lakes and streams were mapped based on existing surface
water alkalinity data (Omernik and Powers, 1983). The boundaries between regions were drawn
using measured alkalinity and receptor characteristics such as geologic and soil sensitivity indices.
These alkalinity maps identified regions with low alkalinity lakes and streams that might be
susceptible to acidic deposition. Alkalinity was selected to characterize aquatic systems because (1)
it is, by definition, a measure of the capacity of the system to neutralize acids; (2) it is an integrated
measure of many processes occurring in the watershed/aquatic system controlling surface water
1-10
-------
acidification; and (3) alkalinity measurements are available for many U.S. lakes and streams. The
alkalinity maps served as the basic framework for designing the National Surface Water Survey
(NSWS).
The first step in characterizing the ecological components was to determine the proportion
and extent of aquatic resources that were acidic or potentially susceptible to acidic deposition. In
1984, EPA initiated the NSWS to quantify the extent and distribution of such aquatic systems. The
NSWS, conducted between 1984 and 1986, was a statistically designed survey that provided
unbiased estimates of the number, length, area, and location of acidic and low-pH lakes and streams
in the United States, based on samples from areas of the country known to contain surface waters
with little capacity for neutralizing acids. For quantitative estimates, a specific target population of
lakes and streams was defined.
Target Population. The NSWS lake target population of interest was composed of lakes with
surface areas >4 ha (1 acre = 0.40 ha) in the East or > 1 ha and < 2,000 ha in the West. Lakes
identified on l:250,000-scale U.S. Geological Survey (USGS) topographic maps were given a
number, the lakes were stratified by alkalinity or ANC class (e.g., < 100, 100-200, >200
microequivalents/liter, or jieq/L), and a systematic random sampling process was used to select
sample lakes. The NSWS target stream population contained stream reach segments with drainage
areas less than 150 km2 that were large enough to be represented as blue lines on l:250,000-scale
USGS topographic maps. For streams, a regularly spaced dot-grid, with about 13 km between dots,
was randomly overlaid on the regions of interest, and an association rule was used to select the
appropriate stream associated with each dot for inclusion in the first-stage sample. These sample
reaches were characterized by site name, watershed area, and other geographic information. A
probability subsample of these stream reaches was selected because there were too many first-stage
streams to field sample for water chemistry. The target populations for evaluation were lakes in the
Northeast, Upper Midwest, Southeast, Florida, and mountainous West, and streams in the
Mid-Atlantic Highlands and Coastal Plain, Southeastern Highlands, and Florida Regions. Complete
descriptions of the National Lake Survey and National Stream Survey can be found in Linthurst et
al. (1986), Landers et al. (1987), and Kaufmann et al. (1988).
Endpoints. The assessment endpoint was the vulnerability of lakes and streams to
acidification at levels that would endanger the survival of fish species (i.e., based on laboratory
mortality data or field data of fish presence or absence as a function of water chemistry). The
measurement endpoints were chemical measures related to survival of fish, definition of acidic
versus nonacidic systems, or the source of acidity. These included ANC, pH, sulfate, dissolved
organic carbon (DOC), calcium, and aluminum. Secondary measurements were important in
refining the sources of acidity, assessing potential impacts on aquatic organisms, and ensuring the
quality of the data.
1-11
-------
Comments on Problem Formulation
Strengths of the case study include:
•The types and extent of aquatic systems at risk received a comprehensive examination
in this program. The study established criteria for screening lakes and streams
throughout the country to identify those with low resistance to changes in their acid
status. These criteria were then applied to information on the distribution of aquatic
systems throughout the country to identify the distribution of those at risk.
•This program benefited from the measurement of alkalinity as an integrative
Umnological variable by early ttmnologists. The results from this program also provided
an extensive background of information on which decisions could be based. The
availability of this information illustrates the value of long-term data-gathering efforts.
Limitations include:
•Identification of ecological components at risk should be treated more rigorously. The
case study focuses on fish survival as indicated from the acid stress index. Other
components of the system may have been at risk but are not evaluated (e.g.,
zooplankton). As part of the problem formulation stage of a risk assessment, it would be
helpful to have a clear rationale for selecting particular components and endpoints.
1.3.2. Analysis: Characterization of Ecological Effects
Developing the Relationship Between Acidity and Effects on Fish. The relationship between
surface-water acidity and acute effects on aquatic life could be demonstrated readily in laboratory
and field experiments. (Figure 1-3 presents critical pH levels for selected aquatic organisms.) This
information was synthesized and integrated in a series of State-of-Science/Technology (SOS/T)
reports that supported the NAPAP 1990 Integrated Assessment (J.P. Baker et al., 1990; Turner
etal., 1990).
Changes hi surface-water chemistry were related to changes in fish response through a
generic acidic stress index (ASI) and through logistic regression models based on statistical analyses
of historical fish records in susceptible regions with documented fish loss. The ASI was developed
to reflect the combined effects of pH, aluminum, and calcium on selected fish species (J.P. Baker
et al., 1990). The model output (i.e., ASI) ranges from 0 to 100, corresponding to between 0 and
100 percent fish mortality in the laboratory experiments used to develop the model. The specific
ASI threshold value above which effects might occur, however, varied by species. For example, an
ASI> 10 might be appropriate for sensitive species, while an ASI> 30 might be representative for
more tolerant species. Similar reference values were established for pH, with reference values of
6.0, 5.5, and 5.0 established for possible effects on biotic populations. The exposure-effects
1-12
-------
Critical pH Levels for Selected Aquatic Organisms
6.5 6.0 5.5 . 5.0 4.5
4.0
Yellow Perch
Brook Trout
Lake Trout
Smallmouth Bass
Rainbow Trout
Common Shiner
American Toad*
Wood Frog*
Leopard Frog*
Spotted Salamander*
Crayfish**
Mayfly"
Clam"
Snail"
*Embiyonic life stages.
**Selected species.
Figure 1-3. Critical pH levels for selected organisms (J.P. Baker et al., 1990)a
"Solid symbols reflect favorable pH ranges. Shaded symbols indicate less-favorable pH ranges.
pH ranges that generally do not support populations of a particular organism have no
symbol.
1-13
-------
relationships, therefore, were established between water chemistry variables (i.e., pH, Al, and Ca)
and the probability of fish presence/absence for selected species.
By comparing the ASI with other sources of information on fish response, approximate
reference levels were defined., above which fish populations might be lost as a result of the high
levels of acid stress (J.P. Baker et al., 1990). Surface waters with ASI values exceeding these
response thresholds are identified as unsuitable due to acidic stresses. These threshold values differ
between lakes and streams because fish responses differ in streams and lakes. Significant
uncertainty can be associated with extrapolating laboratory results to field conditions, however,
because of uncontrolled, natural variability in field systems, matrix effects on chemical reactions,
biotic interactions among species, and other factors.
In the Adirondack subregion, sufficient historical field data were available to formulate
logistic regressions for estimating the probability of fish presence/absence based on pH (J.P. Baker
et al., 1990). These formulations were used to refine estimates of unsuitable fish habitat in the
Adirondack subregion; they could not be developed for other subregions because of insufficient
historical data.
Developing the Relationship Between Acid Deposition and Acidification. Multiple analyses
were performed to quantify the relationship between acidic deposition and acidification of freshwater
lakes and streams. These analyses included statistical associations and correlations between stressor
variables and physical/chemical response variables or measurement endpoints (Church et al., 1989);
empirical relationships between sulfur deposition and changes hi surface-water chemistry (Church
et al., 1989; Thornton et al., 1990); process-oriented models describing the quantitative relationships
between acidic deposition and watershed processes and interactions with the receiving .aquatic system
to result in changes in surface-water chemistry (Church et al., 1989; Thornton et al., 1990);
paleolimnological analyses to quantify the relationship between historical changes in surface-water
chemistry and changes hi aquatic life (Sullivan et al., 1990); field manipulation studies to quantify
the exposure-effects relation between deposition rate and surface-water acidification (Church et al.,
1989); and studies documenting the change in surface-water chemistry following reduction in acidic
deposition and quantifying this exposure-effects relationship (Dillon et al., 1987).
The Environmental and Social Systems Analysts/Department of Fisheries and Oceans model
is an empirical steady-state model used to assess changes in surface-water chemistry as a function of
alternative deposition scenarios. A steady-state model predicts surface-water chemistry values that
eventually will be achieved if the system approaches equilibrium with a constant set of inputs
(Thornton et al., 1990). Three process-oriented, dynamic models were used to determine the change
hi water chemistry through time (Model of Acidification of Ground Water in Catchments [MAGIC]),
Regional MAGIC, and Integrated Lake/Watershed Acidification Study [ILWAS] [Thornton et al.,
1990]). Dynamic model projections are critical for systems that slowly approach steady-state
conditions. MAGIC, Regional MAGIC, and ILWAS were used for projections of future changes in
surface-water chemistry as a function of alternative deposition scenarios. MAGIC also was used for
hindcasts of historical changes hi surface-water chemistry as a function of historical changes in
acidic deposition rates. These hindcasts were compared with paleolimnological reconstructions using
algae fossils. Evidence from paleolimnological studies, which estimate historic ANC based on algae
1-14
-------
fossils in lake sediments, suggests that lakes with ANC less than 50 /ieq/L are most responsive to
changes in acidic deposition (Sullivan et al., 1990).
Calibration of Predictive Exposure-Effects Models. The watershed chemistry models were
calibrated on three Northeast watersheds and two watersheds in the Southern Blue Ridge Province
(SBRP) and then confirmed through blind simulations using only input data, without further
calibration, and comparing the model output with additional stream and lake data for the
confirmation period. The fish-response models were evaluated using field bioassay data to compare
projected with observed responses. Sensitivity analyses also were conducted to identify those input
variables and parameters to which the models were sensitive. These variables and parameters
received additional attention during the calibration process. These confirmation exercises, however,
were conducted on short periods of record (i.e., 3-5 years) while the model simulations made
projections for 50 years. Long-term data were not available for model confirmation.
Additional information on the relationship between exposure and effects was obtained by
examining paleolimnological data on historical changes in chemical (e.g., pH, aluminum levels)
conditions and biotic species assemblages (Sullivan et al., 1990). These were judged to be
particularly useful inasmuch as the effects of acidic deposition on aquatic ecosystems occur over long
time periods. These data were obtained for a target population of lakes in the Adirondack subregion
of New York and provided additional evidence not only of historical exposure levels, but also that
lakes in regions receiving acidic deposition had been acidified, and some had become acidic.
1.3.3. Analysis: Characterization of Exposure
Identifying the Target Population of Ecological Components HLakes and Streams'). Of an
NSWS target population of about 28,000 lakes and 60,000 streams, 4 percent and 8 percent,
respectively, were acidic (table 1-1). Almost no lakes were acidic in the West; virtually all acidic
lakes and streams were in the eastern United States. About half of these lakes (~ 16,000) and
streams (~ 30,000) had ANC <200 /ieq/L, indicating they were potentially susceptible to acidic
deposition, and 20 percent of the aquatic resources (—5,000 lakes and 12,000 streams) were
considered sensitive to acidic deposition (ANC <50 /ieq/L, table 1-1).
While there were a significant number of acidic systems in selected regions of the country,
not all were acidic because of atmospheric deposition. To identify those lakes and streams that
might be affected by acidic deposition, diagnostic procedures using NSWS data and other
information were developed to eliminate systems that were acidic because of nonatmospheric
sources. The screening determined that the source of acidity in about 75 percent of the acidic lakes
(890 lakes) and 50 percent of the acidic streams (2,200 streams) was dominated by acidic deposition.
During this screening process, the estimated target population of lakes and streams affected by, or
potentially susceptible to, acidic deposition was reduced from 28,300 lakes and 59,600 streams to
15,500 lakes and 27,900 streams, respectively.
Because acidic deposition is a regional problem, the analyses focused on population attributes
rather than on individual lakes or streams. Some relationships emerge at the population or regional
scale that are not apparent at individual sites. For example, the linear relationship between median
1-15
-------
Table 1-1. Percentage Estimates of Number of Lakes and Streams With ANC and pH
Below Three Arbitrary Reference Values for Each Measure8
Percentage of Lakes and Streams
Region
New England
Adirondacks
Mid-Atlantic
Highlands
Mid-Atlantic
Coastal Plain
Southeastern
Highlands
Florida
Upper Midwest
West
All NSWS
Lake or
Streamb
L
L
L
S
S
L
S
L
S
L
L
L
S
Total
Number
4,330
1,290
1,480
27,700
11,300
258
18,900
2,100
1,730
8,500
10,400
28,300
59,600
£0
4
14
6
8
12
<1
1
23
39
3
<1
4
8
ANC
£50
20
38
14
22
30
1
8
40
70
16
16
19
20
£200
64
73
41
48
56
34
52
50
78
41
66
56
51
<5
2
10
1
7
12
<1
1
12
31
2
<1
2
7
pH
£5.5
6
20
6
11
24
<1
2
21
50
4
<1
5
12
£6
11
27
8
17
49
<1
9
33
72
10
1
9
22
a The chemical definition of an acidic system is ANC < 0. Most scientists agree that acid-
sensitive fish are stressed at pH <5.0. However, some scientists believe that acidic episodes
can occur if ANC <200 and that fish can be stressed if pH < 6. Therefore, three reference
values are given for ANC and pH (NAPAP, 1990).
b The stream estimates in this table are based on the chemistry measured at the upstream end of
each surveyed stream reach. Because ANC and pH almost always increased with distance
downstream, the percentage estimates of acidic streams in this table are higher than estimates
based on downstream measurements (SOS/T 9:3). In low-ANC streams, the median
downstream change was +5 /neq ANC per km of stream length (+0.06 pH units per km).
On a length basis, 7,900 km (4%) of the 211,000 km of NSWS streams were acidic, and
26,400 km (13%) had ANC £50 /xeq/L.
Estimates were also made on the basis of lake surface area (SOS/T 9:8). As acidic lakes
tended to be smaller than nonacidic lakes, the percentage of acidic lake area in the NSWS was
a factor of 2 smaller than the percentage of acidic lakes based on numbers. Overall, 263 km2
(2%) of the 12,000 km2 of lake area in the NSWS were acidic, and 1,310 km2 (11%) had
ANC £50 /teq/L.
1-16
-------
wet sulfate deposition and median surface-water sulfate concentrations in NSWS subregions is
apparent at a regional scale (figure 1-4) even though there was no apparent relationship in individual
lakes or streams. Mid-Atlantic and Southeastern Highland streams did not follow this relationship
because these regions are still retaining sulfur and are not in sulfate steady-state between atmospheric
inputs and watershed outputs. Model predictions of chemical-response variables were also analyzed
for the target population as a whole. Sensitivity analyses were performed to evaluate the relative
changes in ANC and pH as a function of different changes in deposition (e.g., from 50 percent
decrease to 30 percent increase in 10 percent increments). These sensitivity analyses indicate that
there is a linear relation between the median change in sulfur deposition and the median change in
ANC in 50 years (figure 1-5).
Estimating Current and Future Levels of Acid Deposition. Aquatic systems are exposed to
both wet and dry acidic deposition. The wet deposition was monitored in the National Acid
Deposition Program/National Trends Network (NADP/NTN), but the dry component was
exceedingly difficult to measure even at research monitoring sites. Therefore, the Regional Acid
Deposition Model (RADM) was used to estimate the dry deposition and to project the fate, transport,
and transformation of sulfur and nitrogen emissions in various regions in the East. These estimates
were modified based on watershed physiography and vegetative cover (e.g., coniferous or deciduous
forest, open meadows, etc.). The dry deposition estimates were combined with NADP/NTN wet
deposition estimates to project the total acidic deposition to which the watershed was exposed.
These total acidic deposition estimates were used as inputs to the watershed chemistry models, which
projected changes in water chemistry over 50 years and, subsequently, exposure of selected fish
populations to hydrogen ion and toxic aluminum concentrations.
Field studies were conducted for RADM to confirm projections. Episodic, atmospheric
measurements made over the Ohio River Valley, the Mid-Appalachian area, and the Northeast were
compared with RADM projections of dry and wet acidic deposition during these same episodes.
Sensitivity analyses also were conducted to identify those input variables and parameters to which
the model was most sensitive. These variables and parameters were given additional attention during
the calibration process. Different acidic deposition scenarios were projected over the next 50 years.
Exposure estimates obtained from RADM and the NADP/NTN were used as inputs to watershed
chemistry models, and outputs from these chemical models were used as inputs to the fish-response
models described in section 1.3.2. The different acidic deposition scenarios that were selected did
not correspond directly to legislative bills but did bound the range of emission control scenarios
incorporated in the bills (figure 1-6). One scenario (SI) assumed no controls on sulfur emissions
from electrical utilities beyond those already legislated. This no-new-control scenario (SI) estimated
reductions in emissions over the 50-year period as new technology/equipment, more efficient and
cost-effective in removing sulfur, replaced older equipment. The other scenarios assumed reductions
in sulfur emissions (from 1980 levels) of 12 million tons (S3), 10 million tons (S4), and 8 million
tons (S5) by the year 2000. For each scenario, estimates of the proportion of currently acidic
systems that might recover and the additional proportion of systems that might become acidic were
made for systems in the Northeast, Mid-Atlantic, and a portion of the Southeast, the Southern Blue
Ridge Province.
1-17
-------
300-
lif 200
I
U)
o-
NSS-f Subregfons MLS Subregions
4b
2c|
T
2e
fed
3b
1c
3 A
Id
1o
Ib
38
2Bn
2Cr>
2X
2As
LEGEND
A HSS-I SUBRECIONS
• NLS SUBRECIONS
10 15 20 25
Wet SO42' deposition (kg/ha/yr)
30
35
Figure 1-4. Relationship between median wet sulfate deposition and median surface-water
sulfate concentrations in NSWS subregions (Sullivan et al., 1988)
1-18
-------
Median Change in ANC Over 50 Years (/xeq/L)
ou
20
10
0
-10
20
30
40
c;n
v Mid-Atlantic Highland
_ \ Lakes
\
\ \\N ^ew En9land
\ >XN
* ^V ^
\ ^^V \
X-
i — -»> Adirondacks
\;..
\.*'x Southern
~ \'''-.
Mid-Atlantic Highland \ '*• .
_ Streams *\
\
f I I f N
Blue Ridge
I
-12 -8-4 0 4 8 • 12
Median Change in Sulfur Deposition (kg/ha/y)
Figure 1-5. Median change in projected ANC (jieq/L) for 50-year MAGIC simulations
versus median change in sulfur deposition (kg/ha/yr) for each deposition
scenario and subregjon. (Points on each line correspond to -50%, -30%,
-20%, 0%, +20%, +30% change from current deposition.) The range of
absolute changes in sulfur deposition varies from region to region due to
differences in current deposition (NAPAP, 1990).
1-19
-------
SO2 Emissions (million tons) ,
20
15
10
, 5
Q
S1 •
- ~"
S8
S5
~ S2.S4
oo
oo
^ v
\
\ V
V , ^^
\ x
\ V
\ x
\ .^
\ ^s
'"~~""~V ^
— ^^c^\
•
1990 2000 2010 2020 2030
Year , •
Note: These emission patterns result from the use of uniform input
values for both the reference case baseline, S1, and the control
scenarios. There are many other plausible baselines which show future
emissions without new controls which are much higher or much lower
than those in this baseline. Use of other baselines would yield different
changes between the control and baseline cases but would not
significantly affect the emissions of the contra! scenarios.
Figure 1-6. Emissions of sulfur dioxide from electric utilities assuming no additional
sulfur controls beyond those already legislated (SI) and alternative sulfur
reductions (from 1980 levels) of 12 million tons (S3), 10 million tons (S4), and
8 million tons (S5) by the year 2000 (NAPAP, 1990)
1-20
-------
1.3.4. Risk Characterization
The final step in an ecological risk assessment is to integrate information on exposure,
exposure-effects relationships, and the refined target population of lakes and streams to characterize
the risk from acidic deposition to that population. Coupling emission, atmospheric, watershed, and
fish-response models provided estimates of the proportion of lakes and streams that might be exposed
to acidic deposition at rates sufficient to affect the region's fish. These estimates were based on
multiple independent criteria for assessing the effects of acidic deposition on aquatic systems. These
independent criteria included the proportion of lakes and streams with pH less than 6.0, 5.5, and
5.0; ANC less than 50 and 0 /teq/L; and ASI values greater than 30 and 10. The values included in
the following tables and figures are for pH <6.0, ANC <0 /teq/L, and ASI > 10 for sensitive fish
species to illustrate the approach. (Note: These three criteria are independent, i.e., an ANC <0
does not correspond to an ASI < 10 or a pH <6. See note a, table 1-1.) Multiple criteria were
used because of the uncertainty associated with the effects of acidity on different biotic species,
including fish. The regions and aquatic systems projected to show the greatest change in unsuitable
fish habitat were Adirondack lakes and Mid- Atlantic Highland streams (table 1-2). The Adirondack
region has a large proportion of lakes with low ANC. In NSWS, 38 percent of Adirondack lakes
had ANC less than 50 jieq/L and 14 percent had ANC less than 0
With no-emission controls, this estimate of acidic lakes was projected to increase to
22 percent and then decrease to 8 percent by the year 2030 (assuming emission reductions occur as
new technology replaces older equipment). The percentage of lakes unsuitable for sensitive fish
species is projected to follow the same pattern as lake acidity, increasing from 12 percent to 15
percent and then decreasing to about 6 percent by 2050.
There was little difference in projected effects with or without controls at the end of
50 years, but the control-reduction scenarios did prevent additional acidification of lakes and loss of
habitat for sensitive fish species that was projected to occur under the no-control scenario between
1990 and 2030. From the year 2000 to 2030, the average percentage of lakes with chemistry
unsuitable for sensitive fish species was projected to be 12 percent without controls and 4 to 6
percent with controls.
Under the control scenarios, 6 percent of the lakes were projected to be acidic in 2030. This
figure, however, was twice the number of Adirondack lakes that were likely to have been acidic in
preindustrial times based on paleolimnological evidence (Sullivan et al., 1990). Delays in emission
reductions under the no-control scenario relative to the reduction scenarios resulted in delays in
biological recovery. The longer the period of adverse chemical conditions, the greater the
probability of fish loss from lakes or streams. For this reason, extended periods of unsuitable
chemistry or delays in deposition reductions might have a greater detrimental effect on fish
communities than was apparent from the ASI.
Of the modeled subregions, the Mid- Atlantic Highlands received the highest levels of sulfur
deposition. Median total sulfur depositions were 17.1 and 21.8 kg/ha/yr for modeled lakes and
streams, respectively, hi this region. Surface water would be expected to respond to changes in
deposition based on the high levels of sulfur deposition and large proportion of low-ANC systems.
1-21
-------
Table 1-2. Average Percentage (Calculated for Years 2000-2030) of Lakes and Streams
With Chemistry Unsuitable for Sensitive Fish Species Over 50-Year Projections
Under Illustrative Deposition Scenarios for Each Region (NAPAP, 1990)
Scenario
Regions SI S5 S4 S3
Adirondacks
New England
12
0.4
6
0
4
0
4
0
Mid-Atlantic
Highland Streams
Sensitive Species 25 22 22 20
Brook Trout 16 6 5 5
Mid-Atlantic
Highland Lakes 2 ,0 0 0
1-22
-------
(Twenty-two percent of die streams,and 14 percent of the lakes have ANC <50 ^eq/L based on the
NSWS; 8 percent of the streams and 6 percent of the lakes have ANC < 0 /ieq/L.)
The percentage of acidic streams was projected to increase from the current 5 percent to
10 percent by the year 2000 with no new controls (SI) (figure 1-7). The proportion of acidic
streams is then projected to decline to 3 percent by 2030. The number of streams with pH less than
6 was projected to increase from the current 5 percent to 12 percent in the year 2000 and then to
decline to 9 percent by 2030 (figure 1-7). The proportion of streams with projected chemistry
unsuitable for sensitive fish species would increase from the current 22 percent to 25 percent hi 1990
through 2020, then decline to 22 percent by 2030.
Under the deposition-reduction or control scenarios, the proportion of acidic streams (i.e.,
ANC <0) was projected to decline and range from 0 to 3 percent by the year 2000 and hold nearly
constant until 2030 (figure 1-7). The number of systems with pH less than 6 would first increase
from 5 percent to 11 percent in 1990, then decrease from 4 percent to 6 percent in 2010 through
2020, and increase again from 7 percent to 9 percent by 2030. This projected increase reflects
continuing, long-term acidification hi some Mid-Atlantic Highland streams. The percentage of
systems with chemistry unsuitable for sensitive fish species would first increase from 22 percent to
25 percent in 1990, then decrease to fluctuate around 17 to 22 percent from 2000 to 2030. The
emission-control scenarios (S3-S5) prevented the doubling of the number of acidic streams that had
been projected for the no-new-control scenario in the Mid-Atlantic Highlands.
The proportion of streams unsuitable for sensitive fish species would be held roughly
constant by the S3-S5 scenarios at about 3 to 5 percent less than the proportion under SI. The
average percentage of streams with chemistry unsuitable for sensitive fish species would be
25 percent under SI and 20 to 22 percent under the S3-S5 scenarios.
Uncertainty was quantitatively evaluated for sampling variability, model input, and
calibration error. Sample variability is common to all data-collection efforts and depends on the
proportion of watersheds sampled. For example, the NSWS determined that 14 percent of the lakes
in the Adirondacks were acidic, based on the results of sampling 155 of the estimated 1,290
Adirondack lakes in the population of interest. In this case, one can calculate that if the NSWS were
repeated many times, randomly choosing a set of lakes each time, the estimated percentage of acidic
Adirondack lakes would fall between 9 percent and 19 percent in 95 of 100 surveys. Only about
one-fifth as many Adirondack lakes (35) were surveyed in the Direct/ Delayed Response Project, a
soil survey and modeling study; this smaller set of lakes was both simulated by MAGIC, a
process-oriented model, and used for paleolimnological studies. An estimate of the percentage of
acidic lakes in the Adirondacks made from the smaller sample is slightly different (16 percent
instead of 14 percent) and has somewhat greater uncertainty (e.g., range from 3 to 29 percent).
Although sampling uncertainty affects the starting point of simulations (i.e., what proportion of lakes
in the region are currently acidic), it does not affect the projected changes in the chemical
composition of these lakes.
Another view of sample uncertainty is illustrated in figure 1-8, which shows the variability of
projected ANC change from watershed to watershed within the Adirondacks. Individual watersheds
may behave differently, and there is often large variance about the median. The median ANC
1-23
-------
(a) Streams with ANC < 0 ^eq/L (percent)
30
20
0 A D O
S1 S3 S4 S5
1980
1990
2000
2010
2020
2030
(b) Streams with pH < 6 (percent)
40
1980
1990
2000
2010
2020
2030
(c) Streams Unsuitable for Brook Trout (percent)
30
1980 1990 2000 2010
Year
2020
2030
Figure 1-7. Percentage of Mid-Atlantic Highland streams with (a) ANC <0 peq/L, (b) pH
<6, and (c) chemistry unsuitable for brook trout based on MAGIC
projections for 50 years under illustrative deposition scenarios (NAPAP, 1990)
1-24
-------
Because the true uncertainty of these assessment techniques is
unknown, qualitative comparisons must suffice. The numerical model
estimates remain our only means of comparing effects of different
deposition scenarios. Model results are presented in this assessment as
median changes and percentage of systems falling within given criteria.
These presentations are useful ways to summarize the data so that
differences among scenarios and regions can be seen. These estimates
have a great deal of uncertainty about them and are intended for relative
comparisons only. Based on all lines of evidence, we have confidence in
the direction and relative amounts of projected change. Projections of
absolute future conditions and timing of changes are highly uncertain.
The numerical model results may represent an upper range for median
change in systems with ANC less than 50 /*eq/L.
Change in ANC frieq/L)
(a) Adirondack Region
Change in ANC (^eq/L)
20
15
10
5
0
-5
(b) Cheney Pond, New York
I
10
20 .," 30
Years
40
50
Figure 1-8. Median and range of change in ANC over 50 years for MAGIC projections
for (a) 35 lakes in the Adirondacks illustrating variability among watersheds
and (b) 10 MAGIC projections for Cheney Pond, NY, illustrating uncertainty
in model inputs and calibration (NAPAP, 1990)
1-25
-------
change for the Adirondack lakes under the 30 percent reduction scenario is about 5 /teq/L, but the
ANC change over 50 years ranges from 1 to 19 fieq/L, (figure 1-8).
The uncertainty generated by possible error in model inputs is generally smaller than the
watershed-to-watershed sampling variability. The effects of this second source of uncertainty can be
quantitatively estimated by various methods, such as using Monte Carlo simulations or running a
model many times with different input values that reflect the range of uncertainty in each input. The
Monte Carlo approach was used in the MAGIC projections. Each watershed modeled with MAGIC
was calibrated up to 10 times with slightly different sets of input parameters. Each input parameter
was randomly chosen from within the uncertainty range for that parameter. The range of model
projections resulting from the set of calibrations for each watershed represents the parameter and
calibration uncertainty in the model. The model input uncertainty for watersheds in the Adirondacks
ranges from 2 to 14 /aeq/L with a median of 7 /teq/L.
This regional ecological risk assessment provides a quantitative estimate of the target
population and quantitative estimates of the proportion of the population that might change at
different levels of acidic deposition. The state of science, however, does not currently permit
quantitative estimates of the probability that these changes will occur.
Comments on Risk Characterization
Strengths of the case study include:
•This case study illustrates how models may be used as pan of an integrated approach
for Unking source, fate, and effects of a stressor. The models employed include (a)
deposition models that estimated rates of add deposition within a region or at a site;
(b) acidification models that estimated levels of acidification associated with deposition,
taldng into account hydrologic regimes, in-watershed, and in-lake processes,- and (c)
population-response models employing stressor-response relationships.
• Overall, this case study provides a good example of a risk assessment. The
assessment achieved its main goals of (1) quantifying the extent to which lakes and
streams have suffered deleterious effects in response to acid deposition and (2)
projecting how the numbers of affected lakes and streams will change under different
regulatory scenarios.
Limitations include:
• The extensive geographic scope of this case study can be viewed as both a strength
and a limitation. It illustrates how a problem can be approached from a broad
regional perspective. However, it is also possible that too much emphasis is placed
1-26
-------
Comments on Risk Characterization (continued)
on the detail of this regional analysis. The percentage, of systems at risk nationwide
provides one perspective from which to view an environmental problem, but it is not
the only perspective that will influence management decisions. Too much emphasis
can be placed on such regionatization. For example, contrast the need to determine
how many lakes and streams are at risk in terms of percentages with the identification
of a single Superfimd site. Some general guidelines need to be developed.
* The format for presenting and comparing risk can introduce subjective biases. For
example, 38 percent of the number of Adirondack lakes have an ANC <50 peq/L; 12
percent of the lake area in the Adirondacks has an ANC <50 fieq/L. Most of the
sensitive Adirondack lakes are small, but there are many of them. The underlying
comparative base can influence the perception of risk. Risk communication
considerations must be initiated at the onset of the study, not at its completion. The
indicators selected, for example, contribute significantly to the comparison and
communication of risk. Most decision-makers focus on fish even though zooplankton
might be a more sensitive indicator. Indicators should be interpretable in terms of
assessment endpoints and/or societal values. Both acidic systems (e.g., ANC <0
ueq/L) and the probability of span fish loss are two indicators that are understood by
decision-makers and the public.
+ This study focuses on the current number and future estimates ofacidic systems (i.e.,
ANC <0) rather than acidified systems (i.e., decreased ANC because of acidic input).
This approach probably underestimates the number of systems affected by acidic
deposition. However, long-term monitoring records are required to estimate the
number of acidified systems, and these records do not exist (National Research
Council, 1986).
General comments:
Several key lessons learned in the NAPAP experience include:
• Policy-relevant questions should be formulated early in the process and used to guide
the research, analyses, and assessment.
^Long-term monitoring records are essential in assessing regional or large-scale
problems because of the long time scales associated with ecological effects.
•Selection of interpretable indicators is essential when assessing endpoints.
^Regional ecological risk assessments require a weight-of-evidence approach and the
integration of multiple levels of analyses, including:
1-27
-------
Comments on Risk Characterization (continued)
- Site-specific process research;
- Regional surveys to evaluate extent, magnitude, and distribution of stressor and
effects;
- Quality assurance programs that are an integral part of the analyses to ensure that
the quality of the data is known;
- Integrated modeling approaches to project future conditions that incorporate
regional differences; and
- Qualitative and quantitative estimates of error or uncertainty and clear
presentations of these estimates.
+Multiagency, multidisciplinary efforts are crucial for regional assessments, both in
reducing bias and in providing a broad-scale, policy-relevant perspective.
*SOS/T documents are useful for providing the synthesis of technical material and
serving as a reference for the assessment document. The 1990 NAPAP Integrated
Assessment is generally aimed at a different audience and should integrate and present
the technical results without extensive discussion of the details that are included in the
SOS/T documents.
Some of the remaining problems that need to be addressed include:
• Quantitative estimates of all components of uncertainty or error must be presented
clearly to decision-makers.
^Better procedures are required for presenting uncertainties to decision-makers and
policy analysts and for communicating environmental risk.
• Techniques for determining the likelihood or probability of impacts need to be
developed. Currently, it is possible to estimate the effects of acidic deposition on
aquatic systems qualitatively, but it is not possible to quantify these probabilities.
•Ecological monitoring networks, rather than chemical or specific resources, must be
implemented nationally and maintained for long periods of time.
*A set of decision models or tools is needed in addition to the scientific models or
tools for ecological risk assessment.
•* Technical information from the regional perspective should be transferred to regional
and program offices and states for their use.
The aquatic effects portion of the NAPAP involved substantial resources (probably
more than $100M over 10 years). Few assessments are likely to have such resources
1-28
-------
Comments on Risk Characterization (continued)
at their disposal. Thus, the extent to which this case study can serve as a model for
other assessments must be scaled to differences in resources available for those
assessments. At the same time, it should also be possible to evaluate where cuts in
the resources available for the NAPAP could have been made with a minimal effect
on the final product.
A major criticism of NAPAP is that it did not contribute to legislation on acidic
deposition because its results were (1) presented in scientific reports and (2)
published after the Clean Air Act Amendments (CAAA) were passed. The NAPAP
results were used during the CAAA debates in Congress and within the federal
agencies but were presented as briefing packages, congressional testimony, and fact
sheets. Better procedures, however, need to be developed to peer review results
quickly during congressional debates and to present information so it can contribute
to the decision-making process.
1-29
-------
1.4. REFERENCES
Aimer, B.; Dickson, W.; Ekstrom, C.; Hornstrom, E.; Miller, U. (1974) Effects of acidification on
Swedish lakes. Ambio 3:30-36.
Altshuller, A.P.; Linthurst, R.A. (1984) The acidic deposition phenomenon and its effects. In:
Critical assessment review papers. Vol. II, effects sciences. U.S. Environmental Protection
Agency, Washington, DC. EPA 600/8-83-016 BF.
Baker, J.P.; Bernard, D.P.; Christensen, S.W.; Sale, M.J.; Freda, J.; Heltcher, K.; Marmorek, D.;
Rowe, L.; Scanlon, P.; Suter, G.; Warren-Hicks, W.; Welbourn, P. (1990) Biological
effects of changes in surface water acid-base chemistry. NAPAP Report No. 13. In:
National Add Precipitation Assessment Program. Acidic deposition: state of science/
technology, volume II. Aquatic processes and effects. Washington, DC.
Beamish, R.J.; Harvey, H.H. (1972) Acidification of the La Cloche Mountain Lakes, Ontario and
resulting fish mortalities. /. Fish. Res. Board Can. 29:1131-1143.
Beamish, R.J.; Lockhart, W.L.; Van Loon, J.C.; Harvey, H.H. (1975) Long-term acidification of a
lake and resulting effects on fishes. Ambio 4:98-102.
Burns, D.A.; Galloway, J.N.; Hendrey, G.R. (1981) Acidification of surface waters in two areas of
the eastern United States. J. Water Air Soil Pollut. 16:277-285.
Church, M.R.; Thornton, K.W.; Shaffer, P.W.; Stevens, D.L.; Rochelle, B.P.; Holden, G.R.;
Johnson, M.G.; Lee, J.J.; Turner, R.S.; Cassell, D.L.; Lammers, D.A.; Campbell, W.G.;
Liff, C.I.; Brandt, C.C.; Liegel, L.H.; Bishop, G.D.; Mortenson, D.C.; Pierson, S.M.;
Schmoyer, D.D. (1989) Direct/delayed response project: future effects of long-term
sulfur deposition on surface water chemistry in the northeast and southern Blue Ridge
Province. Office of Environmental Monitoring and Quality Assurance, U.S.
Environmental Protection Agency, Washington, DC. EPA 600/3-89/06Id.
Cogbill, C.V.; Likens, G.E. (1974) Acid precipitation in the northeastern United States. Water
Resources Res. 10:1133-1137.
Davis, R.B.; Norton, S.A.; Brakke, D.F.; Berge, F.; Hess, C.T. (1980) Atmospheric deposition in
Norway during the last 300 years as recorded in SNSF lake sediments. In: Drablos, D.;
Tollan, A., eds. Ecological impact of acid precipitation, IV. Synthesis and comparison with
New England, pp. 274-275. Proc. Int. Conf. Ecol. Impact Acid Precip., SNSF Project,
Oslo, Norway.
Dillon, P.J.; Reid, R.A.; de Grosbois, E. (1987) The rate of acidification of aquatic ecosystems in
Ontario, Canada. Nature 329:399^13. .
Drablos, D.; Tollan, A., eds. (1980) Ecological impact of acid precipitation. Proceedings of an
international conference. Sanderfjord, Norway, March 11-14, 1980.
1-30
-------
Gjessing, E.T.; Henriksen, A.; Johannessen, M.; Wright, R.F. (1976) Effects of acid precipitation
on freshwater chemistry. In: Braekke, P., ed. Impact cf acid precipitation on forest and
freshwater ecosystems in Norway, pp. 64-85. Sur Nedbors Virkning Pa Skog og Fisk Res.
Rep. 6/76.
Henriksen, A. (1979) A simple approach for identifying and measuring acidification in freshwater.
Nature 278:542-544.
Hultberg, H. (1977) Thermally stratified acid water in late winter—a key factor inducing
self-accelerating processes which increase acidification. /. Water Air Soil Pollut. 7:279-294.
Kaufmann, P.R.; Herlihy, A.T.; Elwood, J.W.; Mitch, M.E.; Overton, W.S.; Sale, M.J.; Messer,
J.J.; Cougan, K.A.; Peck, D.V.; Reckhow, K.H.; Kinney, A.J.; Christie, S.J.; Brown,
D.D.; Hagley, C.A.; Jager, H.I. (1988) Chemical characteristics of streams in the
mid-Atlantic and southeastern United States. Volume I. Population descriptions and
physico-chemical relationships. U.S. Environmental Protection Agency, Washington, DC.
EPA 600/3-88/021a.
Landers, D.H.; Eilers, J.M.; Brakke, D.F.; Overton, W.S.; Kellar, P.E.; Silverstein, M.E.;
Schonbrod, R.D.; Crowe, R.E.; Linthurst, R.A.; Omernik, J.M.; Teague, S.A.; Meier,
E.P. (1987) Characteristics of lakes in the western United States. Volume I. Population
descriptions and physico-chemical relationships, U.S. Environmental Protection Agency,
Washington, DC. EPA/600/3-86/054a.
Linthurst, R.A.; Landers, D.H.; Eilers, J.M.; Brakke, D.F.; Overton, W.S.; Meier, E.P.; Crowe,
R.E. (1986) Characteristics of lakes in the eastern United States. Volume I. Population
descriptions and physico-chemical relationships. U.S. Environmental Protection Agency,
Washington, DC. EPA/600/4-86/OQ7a.
McFee, W.W. (1980) Sensitivity of soil regions to acid precipitation. In: Shriner, D.S.; Richmond,
C.R.; Lindberg, S.E., eds. Atmospheric sulfur deposition: environmental impact and health
effects, pp. 495-506. Ann Arbor, MI: Ann Arbor Science.
National Academy of Sciences. (1984) Acid deposition: processes of lake acidification. Washington,
DC: National Academy Press.
National Acid Precipitation Assessment Program (NAPAP). (1990) 1990 Integrated assessment
report. NAPAP. Washington, DC. 520 pp.
National Research Council. (1986) Acid deposition: long-term trends. Washington, DC: National
Academy Press.
Norton, S.A. (1980) Geologic factors controlling the sensitivity of aquatic ecosystems to acidic
precipitation. In: Shriner, D.S.; Richmond, C.R.; Lindberg, S.E., eds. Atmospheric sulfur
deposition: environmental impact and health effects. Ann Arbor, MI: Ann Arbor Science.
1-31
-------
Okland, J. (1979) Environment and snails (Gastropoda): studies of 1,000 lakes in Norway. In:
Drablos, D.; Tollan, A., eds. Ecological impact of add precipitation, pp. 322-323. SNSF
Project, Sanderfjord, Norway.
Omernik, J.M.; Powers, C.F. (1983) Total alkalinity of surface waters—a national map. Ann. Assoc.
Amer. Geographers 73:133-136.
Schindler, D.W. (1988) Effects of acid-rain on freshwater eco-systems. Science 239:149-157.
Schindler, D.W.; Wagemann, R.; Cook, R.B.; Ruszcczynski, T.; Prokopowich, J. (1980)
Experimental acidification on Lake 223, experimental lakes area: background data and the
first three years of acidification. Can. J. Fish. Aquat. Sci. 37:342-354.
Schofield, C.L. (1976a) Acid precipitation: effects on fish. Ambio 5:228-230.
Schofield, C.L. (1976b) Lake acidification in the Adirondack Mountains of New York: causes and
consequences. In: Dochinger, L.S.; Seliga, T.A., eds. Proceedings of the First International
Symposium on Acid Precipitation and the Forest Ecosystem. USDA Forest Service Tech.
Rept. NE-2. p. 477.
Sullivan, T.J.; Eilers, J.M.; Church, M.R.; Blick, D.J.; Eshleman, K.N.; Landers, D.H.; Deltaan,
M.S. (1988) Atmospheric wet sulphate deposition and lakewater chemistry. Nature 331:607-
609.
Sullivan, T.J.; Small, M.J.; Kingston, J.C.; Gumming, B.F.; Dixit, S.S.; Smol, J.P.; Bernert, J.A.;
Thomas, D.R.; Vutala, A.J. (1990) Historical changes in surface water acid-base chemistry
in response to acidic deposition. NAPAP SOS/T Report No. 11. In: National Add
Precipitation Assessment Program. Acidic deposition: state of science/technology, volume II.
Aquatic processes and effects. Washington, DC.
Thornton, K.W.; Mannorek, D.; Ryan, P.P.; Heltcher, K.; Robinson, D. (1990) Methods for
projecting future changes in surface water acid-base chemistry. NAPAP SOS/T Report No.
14. In: National Acid Precipitation Assessment Program. Acidic deposition: state of
science/technology, volume II. Aquatic processes and effects. Washington, DC.
Turner, R.S.; Cook, R.B.; Van Miegrot, H. (1990) Watershed and lake processes affecting chronic
surface water acid-base chemistry. NAPAP SOS/T Report No. 10. In: National Acid Precipi-
tation Assessment Program. Addic deposition: state of sdence/technology, Volume II.
Aquatic processes and effects. Washington, DC.
1-32
-------
APPENDIX A
SCIENTIFIC CONTRIBUTORS TO THE 1990
NAPAP AQUATIC ASSESSMENT
1-A1
-------
CONTRIBUTORS
Dr. Joan Baker, Western Aquatics, Inc., Raleigh, NC
Dr. Lawrence Baker, U.S. Environmental Protection Agency, Corvallis, OR
Dr. David Brown, NAPAP, Washington, DC
t - , : .,
Dr. Sigurd Christensen, Oak Ridge National Laboratories, Oak Ridge, TN
Dr. Alan Herlihy, U.S.* Environmental Protection Agency, Corvallis, OR •
Dr. Philip Kaufmann, Utah State University, Logan, UT
Dr. Dixon Landers, U.S. Environmental Protection Agency, Corvallis, OR
Mr. David Mamorek, Environmental and Social System Analysts, Vancouver, British Columbia,
Canada
Dr. Mike Sale, Oak Ridge National Laboratories, Oak Ridge, TN
Dr. Timothy Sullivan, E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. Kent Thornton, FTN Associates, Ltd., Little Rock, AR
Dr. Robert Turner, Oak Ridge National Laboratories, Oak Ridge, TN
Dr. Parker Wigington, U.S. Environmental Protection Agency, Corvallis, OR
1-A2
-------
SECTION TWO
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
THE BAY DRUMS, PEAK OIL, AND REEVES SOUTHEASTERN
AREAWIDE WETLAND IMPACT STUDY
-------
AUTHORS AND REVIEWERS
AUTHORS
Quentin J. Stober
Region 4
U.S. Environmental Protection Agency
Athens, GA
REVIEWERS
Judy L. Meyer (Lead Reviewer)
Institute of Ecology
University of Georgia
Athens, GA
Nancy A. Bryant
ENSR Consulting & Engineering
Acton, MA
Kenneth L. Dickson
Institute of Applied Sciences
University of North Texas
Denton, TX
Thomas M. Frost
Center for Limnology
University of Wisconsin
Madison, WI
Alan G. Auwarter
Region 4
U.S. Environmental Protection Agency
Athens, GA
Douglas P. Ormrod
Office of Graduate Studies
University of Guelph
Guelph, Ontario, Canada
Randall S. Wentsel
Toxicology Division
U.S. Army Chemical Research,
Development, and Engineering
Center
Aberdeen Proving Grounds, MD
2-2
-------
CONTENTS
ABSTRACT 2-9
2.1. RISK ASSESSMENT APPROACH 2-11
2.2. STATUTORY AND REGULATORY BACKGROUND 2-11
2.3. CASE STUDY DESCRIPTION 2-11
2.3.1. Problem Formulation 2-11
2.3.2. Analysis: Characterization of Ecological Effects 2-27
2.3.3. Analysis: Characterization of Exposure . . 2-33
2.3.4. Risk Characterization 2-36
2.4. REFERENCES 2-58
2-3
-------
LIST OF FIGURES
Figure 2-1. Structure of analysis for Bay Drums areawide impact study 2-12
Figure 2-2. Wetland sample stations .2-14
Figure 2-3. Surficial aquifer zinc concentrations, Bay Drums, Peak Oil,
and Reeves Superfund sites, Tampa, Florida 2-16
Figure 2-4. Surficial aquifer toluene concentrations, Bay Drums, Peak Oil,
and Reeves Superfund sites, Tampa, Florida 2-17
Figure 2-5. Toxicity tests flowchart 2-28
2-4
-------
LIST OF TABLES
Table 2-1. Type and Number of Contaminants in
Each Group by Site and Media .2-15
Table 2-2. Vegetation List: North (NOW), Central
(CLW), South (SOW), Spray Field (SFW),
and Cypress Pond (CPW) Wetlands ...... '. . .2-18
Table 2-3. Checklist of Benfhic Macroinvertebrates '
' (Bay Drums Wetlands Study, January 1989). ..i......:........... 2-21
Table 2-4. Wetland Evaluation Technique (WET) Probability
Rating Comparison: North (NOW), Central (CLW),
South (SOW), Spray Field (SFW),
and Cypress Pond (CPW) Wetlands 2-26
Table 2-5. Statistical Analysis of Data From Toxicity
Tests Performed on Bay Drums Samples 2-29
Table 2-6. Wetland Bioassay Data 2-31
Table 2-7. Composite Whole Fish Analytical Data Summary,
Bay Drums, Peak Oil, and Reeves SE Areawide
Wetland Impact Study, Tampa, Florida 2-35
Table 2-8. Composite Whole Crayfish Analytical Data Summary,
Bay Drums, Peak Oil, and Reeves SE Areawide
Wetland Impact Study, Tampa, Florida .2-37
Table 2-9. Concentration of Inorganic Elements in Wetland Tissue
Samples as a Multiple of Low Background Wetland
Analysis, Bay Drums, Peak Oil, and Reeves SE
Areawide Wetland Impact Study, Tampa, Florida 2-38
Table 2-10. Wetland Surface Water Analytical Data Summary,
Bay Drums, Peak Oil, and Reeves SE Areawide
Wetland Impact Study, Tampa, Florida 2-39
Table 2-11. Wetland Sediment Analytical Data Summary,
Bay Drums, Peak Oil, and Reeves SE Areawide
Wetland Impact Study, Tampa, Florida 2-42
2-5
-------
Table 2-12.
Table 2-13.
Table 2-14.
Table 2-15.
LIST OF TABLES (continued)
Summary of EPA Ambient Water Quality Criteria and
Screening Concentrations for the Protection of
Freshwater Biota
Summary of ER-L, ER-M, and Overall Apparent
Effects Thresholds Concentrations for Selected
Chemicals in Sediment (Dry Weight)
Ratio of Analyte Concentration in Wetland Sediments
to a Biologically Effective Concentration
Ratio of Areawide Hydrologic Study Surface Water Constituents
to EPA Ambient Water Quality Criteria and Screening Values .
.2-46
.2-52
.2-53
2-55
2-6
-------
LIST OF ACRONYMS
AET apparent effects threshold
AWIS areawide wetland impact study
AWQC ambient water quality criteria
AWRI/FS areawide remedial investigation/feasibility study
BDRI Bay Drums remedial investigation
CERCLA Comprehensive Environmental Response, Compensation, and Liability Act
of 1980
CLW Central Wetland
CPW Cypress Pond Wetland
EPA U.S. Environmental Protection Agency
ER-L Effects Range-Low
ER-M Effects Range-Median
MSL mean sea level
NOAA National Oceanic and Atmospheric Administration
NOW North Wetland
NPL National Priority List
OCP organochlorine pesticides
PAH polycyclic aromatic hydrocarbon
PCB polychlorinated biphenyl
PRP potentially responsible party
SEG Southeastern Galvanizing
SEW Southeastern Wire
SFW Spray Field Wetland
2-7
-------
LIST OF ACRONYMS (continued)
SOW South Wetland
SVOC semivolatile organic compound
TCL target compound list
UNC Unnamed Creek
VOC volatile organic compound
WET Wetland Evaluation Technique
2-8
-------
ABSTRACT
The areawide wetland impact study that is the basis of this case study had two objectives.
The first was to evaluate the ecological status of wetlands associated with the Bay Drums, Peak
Oil, and Reeves Southeastern hazardous waste sites. Field surveys of the wetlands and laboratory
investigations of sampled material evaluated the existing flora and fauna, ecological functions,
water quality, bioaccumulation, and toxicity of the surface waters and sediments. The second
objective was to test the possible sources of toxicity; the soil, surface water, and sediments from
each of four industrial sites; and contaminant pathways to the wetlands. This provided input to
feasibility study design options and baseline information to evaluate the effectiveness of specific
remedial actions.
Three wetlands—identified as the North, Central, and South Wetlands—adjacent to the
hazardous waste sites were the subject of an impact investigation. Each wetland was potentially
subject to receiving ground- or surface-water flow contaminated with materials originating from the
hazardous waste sites. The South and Central Wetlands provided hydrologic connections between
surface and ground water. The same was assumed for the North Wetland.
In ecological functions, all wetlands received a moderate to high rating. With a diversity
of aquatic habitat, all wetlands supported a balanced community of aquatic life: wading birds, fish,
aquatic invertebrates, and reptiles.
The sediments and surface water of the three wetlands were found to be relatively
contaminated with an array of inorganic and organic chemicals. Florida water quality standards for
Class III waters and U.S. Environmental Protection Agency ambient water quality criteria for
aquatic life were exceeded by orders of magnitude for aluminum, arsenic, barium, cadmium,
copper, iron, and zinc. The excesses occurred primarily at stations on the Unnamed Creek and at
a single station in the South Wetland.
Surface waters of the three wetlands were free of acute or chronic toxicity with the
exception of a single station. Toxicity, however, was common to the sediments collected from the
wetlands, including the reference wetlands. Toxicity was most pronounced in the sediments from
the Central Wetland and one of the two reference wetlands, where virtually all test species were
affected. The source of the sediment toxicity remains to be further investigated.
The apparent toxicity of the sediment does not appear to have impaired wetland
functions—balanced communities of plants and animals remain. However, the Unnamed Creek, an
outlet from the North Wetland but also associated with direct drainage from the Reeves
Galvanizing facility, was severely affected by heavy metal contamination. The diversity of benthic
macroinvertebrates associated with the creek was severely reduced. Sediments and surface water
were significantly toxic to virtually all species bioassayed. Fish and benthic macroinvertebrates
collected from the creek revealed excessive tissue concentrations of zinc and iron. The levels of
zinc found in the biological tissues could constitute an environmental threat to predators of these
species.
2-9
-------
A large proportion of site source materials tested toxic. The toxicity site source evaluation
helped localize the areas of each site where environmental hazards are the most severe. Site
source data should also guide remedial decisions in order to correct existing or potential hazards
most efficiently. Background lexicological data are now in place from which to monitor later
remedial success. Some test organisms were found to be more sensitive than others to source
materials; this information can be used to select a subset of tests for future monitoring.
2-10
-------
2.1. RISK ASSESSMENT APPROACH
This case study was not initially designed as a risk assessment as defined in the framework
(figure 2-1). The Bay Drums, Peak Oil, and Reeves Southeastern areawide wetland impact study
(AWIS) is an impact assessment rewritten as a risk assessment. The AWIS, as designed, places
the wetlands associated with four Superfund sites at the center of concern. The wetlands are fully
characterized as habitats; in effect, they and their viability as habitats, rather than individual
species, are treated as the ecological components of concern. Sampling of surface water,
sediments, and benthic invertebrates in the study wetlands and in two reference sites is described at
length, but no adverse effects are named until they are empirically identified through macrobenthic
analyses, tissue analyses, and media-based toxicological testing.
This case study offers a number of valuable lessons on the application of risk assessment:
(1) the necessity for multiple measurement endpoints at several levels of biological organization,
(2) the difficulty in choosing clean reference sites in a heavily industrialized area, (3) the need for
better benchmarks for risk assessments in wetlands, and (4) the value of identifying measurement
endpoints to use in assessing the efficacy of remediation efforts.
2.2. STATUTORY AND REGULATORY BACKGROUND
The Comprehensive Environmental Response, Compensation, and Liability Act of 1980
(CERCLA or Superfund) directs that for designated hazardous waste sites, "The President shall
select a remedial action that is protective of human health and the environment ... at a minimum
[shall] take into account: . . . (c) the persistence, toxicity, mobility and the propensity to
bioaccumulate of such hazardous materials and their constituents . . . and conduct an assessment of
permanent solutions . . . that will result in a permanent and significant decrease in the toxicity,
mobility or volume of the hazardous substance, pollutant or contaminant" (U.S. Senate Committee
on Environment and Public Works, 1987, Sec. 121[b][l]).
This assessment focuses on the environmental component of that directive and attempts to
provide the basis for understanding the environmental impact of four National Priority List (NPL)
Superfund sites on the wetlands that surround them.
Section 404 of the Clean Water Act requires protection of wetlands. Under the authority of
the Clean Water Act, the U.S. Environmental Protection Agency (EPA) has developed federal
water quality criteria, which are used by the states in establishing water quality standards for
surface water. Florida state water quality standards thus make up an additional regulatory
justification for the AWIS.
2.3. CASE STUDY DESCRIPTION
2.3.1. Problem Formulation
Site Description. The wetlands impact study area is located on State Road 574, a quarter
mile west of Faulkenburg Road, Hillsborough County, near Tampa, Florida. The area under
assessment includes four Superfund sites—Peak Oil, Bay Drums, Reeves Southeastern Galvanizing
2-11
-------
Figure 2-1. Structure of Analysis for
Bay Drums Areawide Impact Study
PROBLEM FORMULATION
Stressors; chemical contamination of soil and sediments;
physical alteration of wetland habitat.
Ecological Components: wetlands, invertebrates, birds,
"fish, and endangered species.
Endpoints; assessment endpoint was ecological integrity
of wetlands. Measurement endpoints include surveys of
wetland function, sediment toxicity, and tissue residues.
ANALYSIS
Characterization of
Exposure
This study did not
focus on characterizing
exposure. Data
are available on levels of
contaminants but are not
used quantitatively.
Characterization of
Ecological Effects
A suite of measures
including toxicity tests,
field surveys, and residue
levels were used to compare
"impacted and reference"
areas.
RISK CHARACTERIZATION
Risks were characterized by:
comparisons of concentrations to benchmarks
(e.g., water quality criteria) using the Quotient Method.
field and laboratory studies of biological conditions.
This involved comparisons between "impacted and
reference" areas.
The use of a suite of measures provided a basis for
identifying "impacts," but the causative agents were
not easily discerned.
2-12
-------
(Reeves SEG), and Reeves Southeastern Wire (Reeves SEW)—and three freshwater
wetlands—designated North, Central, and South Wetlands—that surround them (see figure 2-2).
The Spray Field and Cypress Pond Wetlands are offsite and up gradient. These reference wetlands
were selected from a wide array of wetlands for comparison based on their similarities in
hydrology, vegetation, and sediment to the three site-related wetlands.
Stressors. Existence of multiple contaminants at these sites was determined by numerous
physical studies. A wide array of volatile organic compounds (VOCs), semivolatile organic
compounds (SVOCs), organochlorine pesticides (OCPs), and polychlorinated biphenyls (PCBs)
were found in the soils around Peak Oil, Bay Drums, and Reeves SEW (table 2-1). Metals were
found at all sites, but the Reeves sites were primarily contaminated with iron and zinc. Table 2-1
summarizes the number of contaminants in each group by media. The sites and the surrounding
media are contaminated with a mixture of inorganic elements and organic compounds.
Representative sampling results for zinc and toluene from the surficial aquifer are presented in
figures 2-3 and 2-4. The isopleths indicate contamination gradients onsite, and similar data from
emergency response and reconnaissance information for remedial work onsite were used in
planning the ecological studies.
In addition to the chemicals, physical effects also have stressed the habitat. Alteration of
physical wetland habitat is obvious because of intense industrial activities resulting from drum
recycling and storage. During the early stages of the investigation, 70,000 cubic yards of shingles
deposited on the site were removed before field site investigations were initiated. Construction of
drainage control berms, backfilling of wetlands, construction of wastewater holding ponds, and
other activities have reduced or eliminated wetland habitat at the site.
These investigations focused on offsite wetlands and the movement of contaminants to those
areas. In addition to contamination originating from the study sites, the Central and South
Wetlands receive surface-water runoff from the land application of treated domestic wastewater
from a nearby wastewater treatment plant. One comparison site, the Spray Field Wetland, also
located in the land application spray fields, was available to address the issue of water quality
impact from the treated waste. Comparison sites are limited in this area because of other
surrounding industrial activity with similar associated contaminants and widespread residential
development in this north Tampa area.
Ecological Components. Wetlands. The focal environmental resources in the area are
three wetlands associated with Superfund sites. A classification method developed by the U.S.
Fish and Wildlife Service (Cowardin et al., 1979) was used to evaluate all wetlands that had the
following vegetative classification similarities. AH wetlands were classified as having palustrine
vegetative systems in the emergent/aquatic bed class. In general, all five wetlands are dominated
with emergent vascular vegetation, such as sedges, rushes, cattails, and Peruvian seedbox.
Secondary vegetation comprises mainly floating plants including duckweed, water ferns, and water
hyacinth (table 2-2). Distinctions among the study wetlands are:
• The North Wetland (1.75 acres) is the only one of the five wetlands having a
surface-water inlet and outlet, the former through a ditch north of the Bay Drums
and Peak Oil sites and the latter an unnamed creek through which surface water
2-13
-------
( REEVES S.E.
GALV. PLANT
01-NOW
NORTH
WETLAND
REEVES S.E.
WIRE PLANT
OI-CLW
OE-CLW
03-CLW
03A-CLW
04-CLW
CENTRAL
WETLAND
SOUTH
WETLAND
POWER; I
LINES I
FAULXEN3URG
INTERIM
WWTP
•\^"- i
! FAULXEN6URG
•SPRAY FIELO RCAO
'.ACCSSS KOAO •
• I
OI-CPW,
'CYPffESS POND
WETLAND I
(COMPARISON^
RAY FIELD
WETLAND
(COMPARISON)
-OI-SOW
-02-SOW
-OS-SOW
-04-SOW
PERMANENT SAMPLING
STATIONS
D • 8IOACCUMULATION
* RAPID BIOASSESSMENT
. WATER LEVEL
RECORDER
0 • TOX1C1TY TESTING
Figure 2-2. Wetland sample stations (U.S. EPA, 1990c)
2-14
-------
Table 2-1. Type and Number of Contaminants in Each Group by Site and Media
Contaminants8
VOCs
SVOCs
OCPs
PCBs
Inorganics and
Metals
Contaminants
VOCs
SVOCs
OCPs
PCBs
Inorganics and
Metals
Reeves
Peak Oil Bay Drums SEG
X X
X X
X X
X X
X XX
Upper
Surficial Floridian Surface
Aquifer Aquifer Water
23 14 4
17 10 3
4 1
1
23 13 20
Reeves
SEW
X
X
X
X
Sediments
7
17
1
1
23
aVOC = volatile organic compound
SVOC = semivolatile organic compound
OCP = organochlorine pesticides
PCB = polychlorinated biphenyl
2-15
-------
H I.iJ3.000 -
ON
LEGEND:
~1—'—"~ RAILROAD
FENCE
------ UNIMPROVED ROADWAY
^——_^«- DRAINAGE DITCH
-». DIRECTION OF SURFACE WATER FLOW
STANDING WATER
'» SURFICIAL AQUIFER MONITORINa WELL
•* SURFICIAL AQUIFER PIEZOMETER
^-10— ZINC CONCENTRATION CONTOUR
(PPM)
0024'* . 1Q— ESTIMATED ZINC CONCENTRATION
CONTOUR (PPM)
B o 01, _ ZINC CONCENTRATION (PPM)
L ZINC DETECTED AT CONCENTRATION
ABOVE 1DL BUT BELOW ERDL
NOTES
1 COORDINATE GRID IS REFERENCED TO THE
FLORIDA STATE PLANE COORDINATE SYSTEM
2 MONITORING WELL P-2 IS FIELD LABELED MW-2
3 WELL LOCATION SURVEYED BY TAI1AMY,
VAN KUREN, GERTIS & ASSOCIATES,
MARCH 1990
4 IF A SAMPLE DUPLICATE WAS ANALYZED,
THE HIGHER DETECTED CONCENTRATION IS SHOWN
4. 5 ESTIMATED CONCENTRATIONS ARE SHOWN AS
POSITIVE DETECTIONS
SCALE
oOO
bOU KEE!
SURFICIAL AQUIFER
ZINC CONCENTRATIONS
BAY DRUMS, PEAK OIL. AND REEVES
SUPERFUND SITES
TAMPA, FLORIDA
(•i-ifAKiO •ilk
POTENTIALLY
RESPONSIBLE PARTIES
Canonie
- «BI) MlIM trUJialMtll,. »«.
PNOUCCt MllMtHM
Figure 2-3. Surficial aquifer zinc concentrations, Bay Drums, Peak Oil, and Reeves Superfund sites, Tampa, Florida
(Canonie Environmental, 1991)
-------
M I.J22.000 -
N 1.J21.000 -
LEGEND:
-•- '-'- RAILROAD
------ — PENCE
r = = = = ~ UNIMPROVED ROADWAY
-- ""'" DRAINAGE DITCH
-*' DIRECTION OF SURFACE WATER FLOW
O STANDING WATER
a SURFICIAL AQUIFER MONITORING WELL
« SURHCIAL AQUIFER PIEZOMETER
—- / 0 —~ TOLUENE CONCENTRATION CONTOUR
(PPM)
__ 7 /?.. ESTIMATED TOLUENE CONCENTRATION
CONTOUR (PPM)
p-ir
NO QUUS'*
TOLUENE CONCENTRAtlON (PPM)
DETECTION LIMIT
NONE DETECTED
w o cm- DETECTION LIMIT
' -
»->- -»«-Q v \\rg
-------
Table 2-2. Vegetation List: North (NOW), Central (CLW), South (SOW), Spray Field
(SFW), and Cypress Pond (CPW) Wetlands (U.S. EPA, 1990c)
Common Name
North Wetland:
Peruvina seedbox
Mosquito fern
Mud-midget
Duckweed
Water-hyssops
Soft rush
Central Wetland:
Peruvian seedbox
Pickerelweed
Duckweed
Soft rush
South Wetland:
Cattail
Water spangles
Water hyacinth
Sedge
Water penny
Smartweed
Spray Field Wetland:
Peruvian seedbox
Soft rush
Cattail
Cypress Pond Wetland:
Cattail
Soft rush
Arrowroot
Bald cypress
Pickerelweed
Peruvian seedbox
Species
Ludwigia peruviana
Azolla caroliniana
Wolffiellafloridana
Lemna spp.
Bacopa spp.
Juncus ffiisus
Ludwigia peruviana
Pontederia cordata
Lemna spp.
Juncus effusus
Typha .latifolia
Salvinia rotundifolia
Eichhornia crassipes
Cyperus haspan
Hydrocotyle utnbellata
Polygonum spp.
Ludwigia peruviana
Juncus effusus
Typha latifolia
Typha latifolia
Juncus effusus
Thalia geniculata
Taxodium distichum
Pontederia cordata
Ludwigia peruviana
Relative
Abundance8
t i
Dominant
Frequent
Infrequent
Frequent
Frequent
Frequent
Dominant
Frequent
Frequent
Frequent
Dom. emergent
Dom. aqua, bed
Frequent
Frequent
Frequent
Frequent
Dominant
Frequent
Frequent ,
Codominant
Codominant
Infrequent
Infrequent
Frequent
Dominant
Indicator
Status5
OBL
OBL
OBL
OBL
OBL
FACW+
OBL
OBL
OBL
FACW+
OBL
OBL
OBL
OBL
OBL
OBL
FACW+
OBL
OBL
FACW+
OBL
OBL
OBL
OBL
Common to All Study Wetlands:
Peruvian seedbox Ludwigia peruviana
Cattail Typha latifolia
Soft rush Juncus effusus
aRelative abundance by estimated coverage.
°OBL = obligate wetland species, FACW+ = facultative wetland species.
2-18
-------
leaves the area after crossing a corner of Reeves SEG. The North Wetland
receives, or may have received, runoff from any of the four Superfund sites.
• The Central Wetland (6.25 acres) is at present connected above ground with the
Bay Drums pond on the Bay Drums property, but it has no defined surface-water
inlet or outlet. Surficial water inflows are from direct rainfall or runoff. Its
catchment basin includes the Bay Drums and Peak Oil sites and land application
(spray) of treated domestic wastewater. The northern boundary of the Central
Wetland has been physically altered by activities such as stacking roofing shingles
on the Bay Drums site. The remainder is maintained pasture for land application of
wastewater, powerline right-of-way, and a rail spur.
• The South Wetland (9.7 acres) has no defined surface-water inlet or outlet. Its
catchment basin includes the Reeves SEW and Peak Oil sites and the land
application (spray) of treated domestic wastewater.
• The Spray Field Wetland (reference site, 0.8 acres) is located within the land
application spray field. It is surrounded by grassed fields maintained by routine
mowing. The water quality and hydrology of the wetland are influenced by the
discharge of treated domestic wastewater. The wetland has no defined surface-
water inlet, but a small outlet could be active during high water after a storm.
• The Cypress Pond Wetland (reference site, 1.5 acres) is located southwest of the
Spray Field Wetland. Like the Central and South Wetlands, it has no defined
surface-water inlets or outlets. Stands of bald cypress are found along the edges
and several trees are located close to the middle of the wetland. Contamination of
both reference sites is likely from adjacent industrial and urban activities.
The seasonal quantity and quality of surface water are critical hi the formation and
maintenance of wetland habitats. Water quantity was monitored via surface-water elevations in all
wetlands except the North from April 1988 to November 1989 by simultaneous staff gauge
readings. Seasonal variation was approximately 2 ft. Comparison of water levels at the South
Wetland with official rainfall records indicates a direct response. Since the site has insignificant
surface-water relief, water-level recession rates appear to be attributable to evapotranspiration.
Comparison of the four wetland systems shows that the seasonal shift in the water level of the
Central Wetland is significantly greater than that of the others.
Surficial flow patterns for the period 1948 to 1985 were provided by a photohistory from
EPA's Environmental Monitoring Systems Laboratory (U.S. EPA, 1985). Four of the five
wetlands have no permanent surface-water inlets and depend on rainfall and potential ground-water
recharge for surface-water inflow. These surficial flow patterns have been physically altered by
industrial activities at the site.
Monitoring of classical water quality parameters focused on dissolved oxygen. Low
concentrations of dissolved oxygen in the wetlands are illustrated by diel monitoring records. For
the Spray Field and Central Wetlands, dissolved oxygen peaked during the day with minimum
2-19
-------
concentrations of 0 to 2 mg/L at night. Dissolved oxygen levels in the Cypress Pond and South
Wetland were consistently near zero.
The potential for interaction of ground and surface waters hi these wetlands is pronounced.
Due to the shallow soil lithology, no significant confining layer was found that would prevent
vertical migration of site contaminants, representing a direct pathway for exchange of contaminants
between surface and ground water.
The North, Central, and South Wetlands and their associated ecotones provide food and
cover for several species of birds, reptiles, fish, and mammals. The study wetlands provide a
short-hydroperiod flooding regime that concentrates fish and macroinvertebrates for easy food
gathering by wading birds.
Macroinvertebrate communities. Benthic macroinvertebrate communities were sampled
qualitatively with a multihabitat rapid bioassessment protocol. The benthic macroinvertebrate
community was identified as an important measurement endpoint to evaluate one aspect of the
wetlands. The Cypress Pond Wetland was used as a reference site to compare benthic
macroinvertebrate communities in the other wetland sites. All sampling was conducted in January
1989 when 53 taxa were found (table 2-3).
Birds and fish. In April 1988, several wading birds, including white ibis and snowy egret,
were observed actively feeding within receding surface waters of the North Wetland. Great blue
herons, great egrets, roseate spoonbills, white ibis, and black skimmers have all been sighted in the
Central Wetlands. Several year-round residents such as the common marsh hen and red-winged
blackbirds have been sighted several times hi the study wetlands. Brown thrashers, mockingbirds,
and quail are prevalent hi the thickets, trees, and open grassed fields within the ecotones adjacent
to the study wetlands.
Four species of fish common to Florida wetlands were collected for identification or for
tissue analysis: Gambusia qffinis (mosquito fish), Mollienisia latipinna (sailfin molly), Fundulus
sp. (killifish), and Jordanella floridae (flagfish). They can survive low levels of dissolved oxygen
and are common forage fish for birds and other predators.
Endangered species. Several species of endangered plants and animals have a range that
includes Hillsborough County. The American alligator, listed as threatened (Federal Register, 4
June 1987), was sighted on several occasions. The South Wetland, relatively large, secluded, and
semipermanently to permanently flooded, probably supports their year-round presence, but given
the proximity and seclusion of the study wetlands, alligators are likely to migrate freely among
them to feed.
The endangered wood stork (Federal Register, 28 February 1984) also may feed in the
study wetlands, especially when fish and macroinvertebrates are concentrated in small isolated
pools. A wood stork was observed near the site. Suitable nesting habitat is not available,
however, except for the tree communities that fringe the South Wetland. No signs of previous or
current nesting sites have been observed in the South Wetland.
2-20
-------
Table 2-3. Checklist of Benthic Macroinvertebrates (Bay Drums Wetlands Study, January 1989) (U.S. EPA, 1989)
Sta. Sta. Sta. Sta.
01-CPW 01-SFW 02-SOW 03-SOW
Sta. Sta. Sta. Sta. Sta.
02-CLW 04-CLW 01-UNC 03-UNC 01-NOWa
to
tb
DIPTERA
Culicidae
Anopheles sp.
Hansonia sp.
Culex sp.
Uranotaenia sp.
Ceratopogonidae
und. sp. x
Chaoboridae
Chaoborus sp.
Tabanidae
Chrysops sp.
Tipulidae
Limonia sp.
Stratiomyidae
Eulalia sp.
Chironomidae
Chironomus sp.
C. sp. 1
C. sp. 2
Goeldichironomus sp. x
Kiefferullus sp. x
Parachironomus sp.
Polypedilum sp. x
P. trigonum
x
x
x
x
x
x
x
X
-------
Table 2-3. Checklist of Benthic Macroinvertebrates (continued)
Is)
to
EPHEMEROPTERA
Callibaetis sp.
Caenis sp.
ODONATA
Argia sp.
Ischnura sp.
Enallagma sp.
Nehalennia sp.
Anaxsp.
Coryphaeschna sp.
Pachydiplax
longipennis
Erythemis sp.
Sta. Sta. Sta. Sta.
01-CPW 01-SFW 02-SOW 03-SOW
X
X X
X X
X XX
X X
X
X X
X
X XX
X X
Sta.
02-CLW
xb
X
X
X
X
X
Sta. Sta.
04-CLW 01-UNC
X
X
X
X
X
X
Sta. Sta.
03-UNC 01-NOW*
X
X X
X X
LEPIDOPTERA
Nymphula sp.
HEMPTERA
Nepidae
Ranatra sp.
Naucoridae
Pelocoris sp.
Corixidae
Sigara sp.
Mesoveliidae
Mesovelia sp.
Belostomatidae
Belostoma sp.
x
x
x
X
X
X
X
X
-------
Table 2-3. Checklist of Benthic Macroinvertebrates (continued)
Gerridae
Gems sp.
Notonectidae
Notonecta sp.
COLEOPTERA
Enochrus sp.
Hydrocanthus sp.
Hygrotus sp.
Peitodytes sp.
Berosus sp.
Tropisternus sp.
w Coptotomus sp.
£J Hydrophilus sp.
Graphoderus sp.
.4/?/0rt sp.
Helodidae
CRUSTACEA
Hyalella azteca
Astacidae
Ostracoda
Sta. Sta.
01-CPW 01-SFW
x
x
X
X X
X
X X
X
X X
X X
X
Sta. Sta. Sta. Sta. Sta.
02-SOW 03-SOW 02-CLW 04-CLW 01-UNC
x
x
X XX
XXX
X X
XX XX
X X
X
X
X
XX XX
XX XX
Sta. Sta.
03-UNC 01-NOW11
x
X
X
X X
X X
X
X
X X
X X
OLIGOCHAETA
Tubificidae
Lumbriculidae
Dero sp.
Peloscolex sp.
X
X
-------
Table 2-3. Checklist of Benthic Macroinvertebrates (continued)
HIRUDINEA
Heldbdella lineata
H. stagnate
GASTROPODA
Lymnaea sp.
Pseudosuccinea sp.
Physella sp.
Loevapex sp.
TOTAL TAXA
aQuotient of Similarity
Sta, Sta. Sta, Sta.
01-CPW 01-SFW 02-SOW 03-SOW
X
x
22 23 19 11
0.35 0.44 0.42
Sta. Sta. Sta.
02-CLW 04-CLW 01-UNC
X
X
X
X
22 18 7
0.45 0.65
Sta. Sta.
03-UNC 01-NOW
X
X X
X
14 21
0.46
"Comparison to Sta. 01-CPW.
bAberrant
-------
The range of the bald eagle, Florida scrub jay, eastern indigo snake, and Florida golden
aster includes Hillsborough County, but adequate feeding or breeding opportunities do not exist
within the study wetlands and their adjacent ecotones.
Endpoints. The ecological integrity of the wetlands was chosen as the assessment endpoint
supported by a suite of measurement endpoints (techniques) used to measure integrity. The
interactions of a suite of specific wetland measurement endpoints with environmental contaminants
were assessed to determine the impacts to the wetlands. Toxicity testing of contaminant mixtures
in several media was used to evaluate the extent to which toxic conditions occurred at and around
the sites and to provide insight into the possible migration of toxic-causing materials from the site
into adjacent wetlands.
Wetland attributes were analyzed with the Wetland Evaluation Technique (WET), Version
2.0 (Adamus et al., 1987). WET assesses functions and values of a wetland in three categories:
Social significance assesses the value of a wetland to society in terms of its special designations,
potential economic value, and strategic location; effectiveness rates the capability of a wetland to
perform a function based on its physical, chemical, or biological characteristics; and opportunity
rates the ability of a wetland to perform a function to its level of capability. The wetland functions
and values that are assessed include ground-water recharge, ground-water discharge, floodflow
alteration, sediment stabilization, sediment/toxicant retention, nutrient removal/transformation,
production export, wildlife diversity/abundance, aquatic diversity/abundance, uniqueness/heritage,
and recreation. Only three functions and values are rated for opportunity: (1) floodflow alteration;
(2) sediment/toxicant retention; and (3) nutrient removal/transformation. A rating of high,
moderate, or low is assigned to each function and value hi the three categories (table 2-4). The
wetlands were rated as moderate to high in ecological functions and values.
Comments on Problem Formulation
^Because the AWIS was not designed as an ecological risk assessment, risk hypotheses
were not explicitly developed in this case study, although choosing endpoints is
implicitly based on hypotheses. Possible hypotheses include: (a) There is an inverse
relationship between level of contamination and biological integrity as measured by
various endpoints; and (b) by decreasing contaminant levels, there would be
improvement in biological integrity. Had these been the risk hypotheses at the outset,
the sampling design might have been different. Contaminant levels and known effects
were not used to establish a causal relationship between contaminants and their adverse
effects.
• At this site the stressors include habitat alteration, hydrologic changes, and chemical
contaminants. Synergistic effects are Ukely to be important, although they were not
assessed. The problem formulation step seems especially important at Superftmd sites,
where there is such a wide range of potential stressors.
2-25
-------
Table 2-4. Wetland Evaluation Technique (WET) Probability Rating Comparison: North
(NOW), Central (CLW), South (SOW), Spray Field (SFW), and Cypress Pond
(CPW) Wetlands (U.S. EPA, 19POc)
II ' II ., 1 ...... II. »i • 1 " .1.1..
Wetland2 ,
Functions/Values1 NOW
Social Significance:
Ground Water
Recharge
Discharge
Floodflow Alteration
Sediment Stabilization
Sed. /Toxicant Retention
Nut. Removal/Transformation
Wildlife D/A
Aquatic D/A
Uniqueness/Heritage
Recreation
Effectiveness :
Gound Water
Recharge
Discharge
Floodflow Alteration
Sediment Stabilization
Sed. /Toxicant Retention
Nut. Removal/Transformation
Production Export
Wildlife D/A
Breeding
Migration
Wintering"
Aquatic D/A
Opportunity:
Floodflow Alteration
Sed. /Toxicant Retention
Nut. Removal/Transformation
M
H
M
M
H
M
M
M
H
L
L
M
M
H
H
L
M
L
L
L
L
H
H
L
CLW
M
H
L
M
H
M
M
M
H
L
U
M
H
M
H
H
L
L
H
H
L
H
H
L
SOW
M
H
M
M
H
M
M
M
H
L
U
M
H
L
H
H
L
L
H
H
L
H
H
L
SFW
M
H
M
M
H
M
M
M
H
L
U
M
H
M
H
H
L
L
L
L
L
H
H
H
CPW
M
H
M
M
H
M
M
M
H
L
U
L
H
L
H
H
L
L
L
L
L
H
H
L
*D/A = diversity/abundance.
^ = high, M = moderate, L = low, U = uncertain.
2-26
-------
2.3.2. Analysis: Characterization of Ecological Effects
Ecological effects were characterized as part of the wetland assessment and to provide
toxicity data for samples collected at the industrial sites. The toxicity data were obtained: (a) to
understand and evaluate effects seen hi the wetlands, (b) to help define appropriate remedial
scenarios, and (c) to provide benchmarks from which to judge remedial effectiveness.
Environmental Toxicity Testing. The toxicity assessment involved the evaluation of the
three wetland areas under study, two reference wetland areas, and four industrial sites.
Implementation of the assessment required cooperation and coordination among several sections
within EPA Region 4, several contractors, subcontractors, and oversight contractors for both EPA
and the potentially responsible parties (PRPs).
More than 400 samples of surface water, sediments, soil, and ground water were taken
from the industrial sites and wetlands for chemical analysis. One hundred of these samples
(excluding ground-water samples) were selected to be split for toxicity testing, targeting those
samples most likely to test toxic (site-collected samples), or to establish gradients of effect (wetland
samples). No attempt was made to randomize locations of samples taken for toxicity tests.
Samples were collected, handled, and distributed for analysis in accordance with the EPA
and EPA-approved work plans that governed several remedial investigations, feasibility studies,
and site-source characterization efforts that were under way simultaneously at the study sites
(Canonie Environmental, 1988a, b; Pace Laboratories, Inc., 1988; U.S. EPA, 1989).
Prior to using soils and sediments for most tests, elutions were prepared with laboratory-
pure water (80:20, Milli-Q :Perrier ). For sediments and saturated soils this process diluted each
sample with four times its volume of water. For drier soils, the ratio was the same, but its basis
was wet weight to dry weight.
Water, sediment, soil samples, and eluates were divided according to the needs of
individual toxicity tests (figure 2-5). Water samples were tested with the Microtox bacterial assay
(Photobacterium nr. phosphoreum); a freshwater alga (Selenastrium capricornututri); a freshwater
cladoceran (Ceriodaphrda dubia); a freshwater fish (Pimephales promelas); and a terrestrial plant,
a species of lettuce, Lactuca sativa. Soils and sediments were eluted, and the eluates were tested
with the same suite of organisms. Table 2-5 identifies the statistical analyses used in working up
test data.
The toxicity tests used were those for which standard protocols existed and for which
individual organisms could be cultured. Several species from various phyla were chosen to
increase the confidence of correctly diagnosing the toxicity in an array of media. Although
laboratory-to-field extrapolation is always a concern, the observed toxic responses correlated with
other measurement endpoints at the most heavily affected sites.
Results of the toxicity testing are summarized below for each area tested.
2-27
-------
WATER
SAMPLE
SEDIMENT OR
SOIL SAMPLE"
150 ml TESTED
FOR DO, pH, HARD.,
ALK., COND.
STORED AT 4 C
IN DARK
50 gTESTED
FOR pH AND
MOISTURE
WATER SAMPLE
ADJUST pH IF
NECESSARY
SOIL / SEDIMENT
1200 ml
FATHEAD^"*-
MINNOW
600ml
CERIO-
DAPHNIA
240ml
ROOT
ELONGATION
200 ml
SELEN-
ASTRUM
2.5 ml
MICROTOX
600 g LETTUCE
SEED GERMINATION
ELUATE PREPARED
FROM.SOIL / SEDIMENT
ADJUST pH IF
NECESSARY
Figure 2-5. Toxicity tests flowchart (U.S. EPA, 1990c)
2-28
-------
Table 2-5. Statistical Analysis of Data From Toxicity Tests Performed on Bay Drums
Samples
Toxicity Test Statistical Analysis'1
Fathead Minnow Chronic Test Dunnett's procedure
(survival and growth)
Ceriodaphnia Chronic Test Student's t-test
(survival and reproduction)
Ceriodaphnia Acute Test LGso obtained by Probit
(survival) analysis or graphical method
(line intercept)
Lettuce Seed Germination Student's t-test
Lettuce Root Elongation Dunnett's procedure
(germination and root length)
Selenastrum Dunnett's procedure
Microtox EC50 obtained by
line intercept method
aData generated by the toxicity test were analyzed by either hypothesis testing (Dunnett's or
Student's t-test) or endpoint estimates (Probit analysis or line intercept methods).
Detailed procedures for these statistical methods can be found in the EPA manuals
EPA 600/4-85/013 and EPA 600/4-89/01, and in a computer software program supplied
by the Microbics Corporation (developer of the Microtox test).
2-29
-------
Bay Drums. The site was generally contaminated with a spectrum of inorganic elements
and organic compounds, many of which were present at concentrations well above those known to
cause adverse biological effects. Surface soils were contaminated with a larger number of
chemicals in higher concentrations than the majority of samples. Some sample stations on the
western and southern property borders appeared to be much less contaminated than the majority of
samples. Water from the Bay Drums Pond was essentially nontoxic to any of the test organisms
exposed to it. Water from the northern drainage ditch and the backfill pond was highly toxic to
daphnids (water fleas) but not to other species (table 2-6). All four water samples stimulated algal
growth.
Sediments from the northern drainage ditch and both ponds were highly toxic to daphnids
and algae, and inhibited germination of lettuce seeds (table 2-6). The fish (P. promelas) test was
sensitive to contaminants found on this, site, but Daphnia tests were more sensitive for almost every
sample. The daphnids would therefore be prime candidates for future toxicity testing at this
site, both as a monitor and to evaluate the success of remediation. Tests using bacteria
(P. phosphoreuni) and algae (S. capricomutwri) were also sensitive and may be useful in the
future. Lettuce (L. sativa) tests were largely insensitive: development of root length was not
hampered by any site sample, and seed germination was affected only by sediment samples and two
soil samples.
Peak Oil. All surface waters analyzed contained many inorganic elements. Aluminum,
iron, and lead exceeded the ambient water quality criteria (AWQC) most frequently. Water from
the southernmost of the three onsite ponds was considerably more toxic than water from the other
two. Almost all sediments and soils from there were highly toxic to most species. Soil samples
split for bioassay were logged as being brown or dark brown in color and oily. The Daphnia,
minnow, bacterial, and algal tests were all sensitive to most site samples and would be useful as
future remedial monitoring tools.
Reeves Southeastern Wire. Pond water and sediments were highly toxic. Pond sediments
contained both inorganics and organics, including arsenic at biologically high concentrations. Pond
water exceeded AWQC for 10 inorganic chemicals: the highest rates were for copper, lead, and
zinc. Site soils were generally contaminated with lead and zinc above toxic concentrations and
tested chronically toxic to living organisms. Cyanide was found in soils on the west side of the
site. Daphnia would be a preferred test species for monitoring remedial progress at this site since
they were sensitive to more samples than were other species.
Reeves Southeastern Galvanizing. Surface waters from the inactive ponds were nontoxic.
Their sediments were highly toxic, with high concentrations of several metals. Soil samples from
the extreme eastern side of the site were essentially nontoxic. Other soils were generally toxic,
contaminated with several inorganic elements. This is especially true for the former drum storage
area and the drainage south of it, the area northwest of the Reeves Galvanizing property, and the
high conductance area. Daphnia and algae were most sensitive to samples from this site and would
be useful in future remedial monitoring.
2-30
-------
Table 2-6. Wetland Bio-assay Data (U.S. EPA, 1990c)
3AY DRUMS, PEAK OIL AND REEVES SE AREAUIOE
UETLAND IMPACT STUDY, TAMPA, FLORIDA
SAMPLE
TYPE
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SURFACE WATER
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SEDIMENT
SAMPLE
LOCATION
NORTH WETLAND
UNNAMED CREEK
CENTRAL WETLAND
CENTRAL WETLAND
CENTRAL WETLAND
CENTRAL UETLAND
SOUTH UETLAND
SOUTH UETLAND
SOUTH UETLAND
SOUTH UETLAND
REFERENCE-CYPRESS POND
REFERENCE-SPRAY FIELD
NORTH UETLAND
UNNAMED CREEK
CENTRAL UETLAND
CENTRAL UETLAND
CENTRAL WETLAND
CENTRAL UETLANO
SOUTH UETLAND
SOUTH UETLAND
SOUTH WETLAND
SOUTH UETLAND
REFERENCE -CYPRESS POND
REFERENCE -SPRAY FIELD
STATION
ID tt
.UVK.jBtA',*..
01 -HOW
01-UNC
01-CLU
02-CLU
03-CLU
04-CLU
01-SOU
OZ-SOU
03 -SOU
04 -SOU
01 -CPU
01-SFU
01 -NOW
01-UNC
01-CLU
OJ-CLU
03-CLU
04-CLU
01-SOU
02 -SOU
03 -SOU
04 -SOU
01 -CPU
01-SFU
DATE
TAKEN
^JIJtJXA&JM^
11/30/89
11/30/89
11/30/89
11/30/89
11/30/89
11/30/89
11/28/89
11/28/89
11/28/89
11/28/89
11/28/89
11/28/89
11/30/89
11 /SO/89
11/30/89
11 /SO/89
11/30/89
11/30/89
11/28/89
11/28/89
11/28/89
11/28/89
11/28/89
11/28/89
CERIOOAPHNIA DUB I A
CHRONIC (c)
MEAN HO
OFFSPRING
^W*-JW\J«,*jkJ«-Tn
SO. 5
0 *
26.9
32.4
27.8
JO. 4
2.4 «
24.2
21.6
20.7
24.8
23.0
0 •
0 *
0 *
0 *
0 *
0 *
0 *
0 *
0 •
0.3 *
0 •
11.8 •
* - SIGNIFICANTLY MORE TOXIC THAN CO TROL ON SOIL AND SEDIMENT SAMP
JtAIED SAMPLES UtRE NO! ANAL
7. GERM-
INATION
f
f
f
f
f
f
f
f
f
f
f
(
90
33 «
63 *
58 «
23 "
28 •
40 *
48 «
58 "
40 «
45 •
67 *
PHOTOBACTERIUM
nr PHOSPHOREUM
5 KIM,
EC50
X CONC.
a
21.56
a
a
a
a
a
a
30.44
a
a
a
4.21
a
6.96
4.02
6.35
8.13
6.54
6.35
10.29
63.79
7.24
66.13
15 HIM.
EC50
% CONC.
a
4.94
a
a
a
a
a
a
36.53
a
a
a
3.30
a
3.80
3.35
8.27
10.28
5.79
5.45
11.24
72.07
6.14
69.68
S. CAPRI -
CORNUTUM
X
CHANGE
* 259
+ 28
+ 88
- 100 *
+ 456
» 40.3
- 100 »
- 100 *
- 100 *
- 100 *
+ 167
« 4
- 100 *
- 80. 4«
- 100 *
- 100 *
- 100 "
- 100 "
- 100 *
- 100 *
- 94 *
- 94 *
- 1.00 •
- 79 "
EST ORGANISMS DIE UITHIH^B HOURS
.ES
I7ED FOR SIGNIFICANT 10XICIIY
-------
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
^Multiple endpoints are used. None of the measurement endpoints used would have
been effective as a single indicator of response. Hence, a suite of metrics is necessary,
not because we do not know the right one to use in a particular situation, but because
no single metric will permit a risk characterization. The level of confidence increases
with the number of different metrics used.
•A thorough description of the habitat was developed (by necessity abbreviated in this
version); a thorough assessment was made of the toxicity of environmental media. By
applying the full suite of toxicity tests, the study is able to identify sensitive species and
fulfill its practical (and clearly stated) purpose: to provide baseline data to track
remediation. Field details and documentation of methods are sufficient to make further
samph'ng and testing replicable.
+In a component-centered study, reference sites were chosen that are comparable in
hydrology, sediment, and vegetation to the study wetlands.
Limitations include:
+Little attention is given to comparing the wetlands to the designated reference sites,
and the reference sites do not appear to have been fortunately chosen. One reference
site (along with two study sites) is described as receiving runoff from the land
application of treated domestic wastewater, a potentially confounding factor. The other
reference site, Cypress Pond, chosen to be at some distance (see figure 2-2) from the
Superfimd sites, nevertheless shares with one study site the lowest level of dissolved
oxygen and its sediment was found to be highly toxic. Further investigation of Cypress
Pond oxygen levels and sediment toxicity is beyond the scope of the AWIS, to be sure,t .
but these factors severely limit its use as a reference site.
•/» this case study, contaminant levels and known effects are not used to establish a
causal relationship between contaminants and their adverse effects.
*A possible limitation of the toxicity testing is the use of eluates of soil and sediments.
Eluates primarily remove the water-soluble contaminants from soil and sediment,
leaving potentially toxic insoluble or particle-bound chemicals in the solid phase. If the
resulting eluate is toxic, that is significant, but a negative result cannot be considered
conclusive. Even a positive result may not fully characterize the toxicity of the
sediment, since other sublethal effects may be associated with the solid-phase
contaminants.
2-32
-------
Comments on Analysis: Characterization of Ecological Effects (continued)
General comment:
•In extrapolating from laboratory toxidty analyses to field situations, several issues
need to be considered: (a) How do the sensitivities of native species differ from those of
the test organisms? (b) Does a laboratory study offer a conservative estimate of the
effect (because of differences in the exposure regime in the field) or would afield assay
provide a more conservative estimate (because of the incorporation of indirect and
synergistic effects)?
2.3.3. Analysis: Characterization of Exposure
Hydrologic Assessment/Transport of Chemicals. The potential for interaction of surficial
ground and surface waters in these wetlands is pronounced. The minimum bottom elevations of
the Central and South Wetlands are approximately 35.0 and 33.2 ft mean sea level (MSL),
respectively. The ground-water elevations in the vicinity of the Central and South Wetlands range
from 33.59 to 36.80 ft MSL and 36.57 to 37.72 ft MSL, respectively (Canonie Environmental,
1990). From these data, the potential for ground water to be intercepted by the wetlands is clearly
evident.
Furthermore, in the shallow soil lithology assessment (U.S. EPA, 1990a) no significant
confining layer was found that would preclude vertical migration of site contaminants. Assuming
the shallow soil lithology remains similar beneath the wetlands, particularly the South and Central
areas, the wetlands represent a direct pathway for the exchange of contaminants between the
ground- and surface-water regimes. Thus, a probable pathway is established for the movement of
contaminants originating onsite to the wetlands. Equally apparent is that the same hydrologic
connection provides a pathway for contaminated water to enter the ground-water system.
A ditch from Bay Drums and Peak Oil connects to the North Wetland. Surface and
subsurface flow from these sites to the Central Wetland is also probable. The proximity of
Reeves SEG to the North Wetland makes direct surface runoff appear probable, but some surface
water from this site drains to the Unnamed Creek, which also drains the North Wetland.
Another potential source of contamination to the Central and South Wetlands is surface
runoff from spray irrigation of treated effluent from a central sewage treatment facility. Surface-
water drainage from this operation appears to influence the southern region of the South Wetland
and the Spray Field Wetland.
Spatial Distribution of Contamination. Surface-water and sediment samples were 'collected
from all five wetlands and the Unnamed Creek associated with the North Wetland. Each sample
coincided with the location selected for toxicity testing (figure 2-2). Only grab samples were
collected. Sediment samples were collected with a hand auger. All sample collection and
2-33
-------
processing in the field were conducted according to ESD standard operating procedures. The
Spray Field and Cypress Wetlands served as reference sites for determining the nature and extent
of chemical contamination hi the North, Central, and South Wetlands. Surface-water and sediment
analyses indicate that the chemical contamination of these three wetlands appears primarily related
to the hazardous waste sites.
For metals, the surface water associated with the South Wetland and the Unnamed Creek
was the most contaminated. Reeves SEG appears to be the principal source of metal
contamination.
Maximum total phthalate ester, concentrations in the surface water were associated with the
four stations located hi the South Wetland. The maximum concentrations were reported as
potentially toxic to phytoplankton.
For aquatic sediments, the North Wetland and the Unnamed Creek were the most
contaminated with metals. Lead and zinc concentrations at the Unnamed Creek were the highest
levels reported and probably reflected the effects of surface drainage from Reeves SEG.
Organic contamination hi the sediments was most pronounced in the Central Wetland,
probably because the Central Wetland extended into the Bay Drums facility, in recent years
actually connecting with onsite waste disposal ponds. Drainage from the Peak Oil site is also a
probable source of contaminants.
Bioaccumulation of Contaminants. Whole-body contaminant levels in aquatic organisms
living hi or near the study sites were measured to provide an indication of exposure and to evaluate
possible effects on higher levels of the food chain. This analysis may also reveal bioaccumulative '
chemicals of concern that exist in surface waters or sediments in concentrations too low to measure
in the water itself. Chemicals found hi elevated concentrations can be tracked during and after
remedial efforts onsite as a measure of the success of those efforts.
Aquatic animals were sampled on January 18 and 19, 1989, from all wetland areas near the
four industrial sites that constitute the areawide remedial investigation and from the two reference
areas; samples were not taken within the property lines of the industrial sites. High water at the
time of sampling made the Central Wetland extend well into the Bay Drums property, connecting
the sources of toxicants; two sampling stations were located within the Central Wetland. An effort
was made to collect at least 20 g of tissue for a fish and invertebrate species from each sampling
area.
Samples of fish were collected at each sampling site with a small seine or a dip net. From
each collection the largest individuals were retained for analysis. The fish species found
throughout the wetlands were almost exclusively topwater, live-bearing fish that are not dependent
on oxygenated water above the sediments for reproduction. In every case, samples for tissue
analysis consisted exclusively of mosquito fish (Gambusia qffinis). Samples were analyzed for all
target compound list (TCL) metals and for TCL organics other than volatiles (table 2-7). Volatiles
were not analyzed since the grinding step in sample preparation would be expected to drive off
these compounds.
2-34
-------
Table 2-7. Composite Whole Fish Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide
Wetlands Impact Study, Tampa, Florida (U.S. EPA, 1990c)
INORGANIC ELEMENTS
ALUMINUM
BARIUM
BORON
CALCIUM
COPPER
IRON
MAGNESIUM
MANGANESE
S> MERCURY
Ai POTASSIUM
^ SODIUM
STRONTIUM
TITANIUM
ZINC
ZIRCONIUM
01-CPW
CYPRESS
POND
01/18/89
0930
(MG/KG)
8.9
1.7
N/A
10000
O.B8
32
380
4.7
0.12
2500
910
19
1.3
40
N/A
01-SFW
SPRAY
FIELD
01/18/89
1300
(MG/KG)
12
2.4
N/A
14000
1.2
20
490
8.8
0.04
3000
1000
35 .
1.7
26
N/A
01-HOW
NORTH
WETLAND
01/19/89
1815
(MG/KG)
35
3.4
N/A
10000
0.88
55
400
9.7
N/A
2500
940
10
1.3
50
H/A
01-UNC
UNNAMED
CREEK
01/19/89
1630
(MG/KG)
59
4.2
N/A
12000
1.2
390
370
7.8
,0.08
2400
1100
12
1.6
370
N/A
03-UNC
UNNAMED
CREEK
01/19/89
1715
(MG/KG)
31
5.9
N/A
10000
1.1
150
370
7.3
0.05
2500
960
11
1.4
180
N/A
02-CLW
CENTRAL
WETLAND
01/19/89
1300
(MG/KG)
-
8.4
5.3
N/A
7300
1.1
77
300
20
0.02
2000
620
9.7
1.0
38
N/A
04-CLW
CENTRAL
WETLAND
01/19/89
0930
(MG/KG)
11
8.8
II/A
10000
1.4
28
390
11
--
2600
840
13
1.3
60
H/A
02-SOW
SOUTH
WETLAND
01/18/89
1430
(MG/KG)
20
3.2
N/A
9300
1.5
77
360
7.1
0.05
2400
860
10
1.3
81
H/A
EXTRACTABLE ORGANIC COMPOUNDS
BENZOIC ACID
0.97"
"Estimated value
- MATERIAL WAS ANALYZED FOR BUT NOT DETECTED.
-------
Crayfish were collected at all but two sampling locations—one relatively deep-water station
and a second just outside the Reeves Galvanizing site where no aquatic or benthic invertebrates
were found (table 2-8).
Overall, fish and crayfish sampled from the various wetland areas did not reveal a wide
spectrum of contaminants at concentrations grossly over background. The primary exceptions were
very high concentrations of iron and zinc found in samples from the Unnamed Creek, which
receives runoff from the Reeves Galvanizing operation.
Several inorganics were widely present at concentrations moderately elevated over
background. These include aluminum, barium, copper, iron, manganese, titanium, and zinc
(table 2-9).
Similar overage tables were prepared for data from samples collected from each implicated
industrial site to document potential hazard from these sources.
Comments on Analysis: Characterization of Exposure
*The links between stressors (chemicals) and ecological components via specific
exposure routes are not clear. The assessment would have been strengthened by
illustrating how the spatial and temporal distribution of the stressors interface with the
spatial distribution and life history patterns of the components. More information on
the spatial distribution of contaminants and their fate and transport would have been
helpful. This information was developed as pan of the Remedial Investigation Study for
the site but should be viewed as pan of the overall risk assessment.
2.3.4. Risk Characterization
The AWIS was conducted to document the contamination in the surface waters and
sediments of the wetland areas, toxicity to plants and animals, and ecological functioning in the
wetlands as a whole. This information establishes existing conditions and can be used to evaluate
the efficacy of remedial actions hi reducing toxicity and overall biological effects. As such, the
work conducted as part of the AWIS serves as the first (i.e., baseline) step in an overall
monitoring program. With regard to the characterization of risks under existing conditions,
analytical data for each study site were compared with relevant criteria or other available effects-
related information to link specific contaminants with adverse environmental effects (tables 2-10
and 2-11). A small portion of the original investigation results has been selected to illustrate the
approach used.
Risks Based on Comparisons to Benchmarks. To characterize ecological risk, the
concentration of each chemical identified in samples of environmental media was compared with a
benchmark concentration known to produce an adverse biological effect. For water samples, the
effective concentration chosen was the AWQC (table 2-12). In the absence of an AWQC, a
2-36
-------
Table 2-8. Composite Whole Crayfish Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland
Impact Study, Tampa, Florida (U.S. EPA, 1990c)
to
01-CPW 01-SFW 01-NOW 03-UNC OA-CLH 02-SOW
CYPRESS SPRAY NORTH UNNAMED CENTRAL SOUTH
POND FIELD WETLAND CREEK WETLAND WETLAND
01/18/89 01/18/89 01/19/89 01/19/89 01/19/89 01/18/89
1000 1330 1800 1730 1030 1500
(MG/KG) (MG/KG) (MG/KG) (MG/KG) (MG/KG) (MG/KG)
INORGANIC ELEMENTS
ALUMINUM
BARIUM
BORON
CALCIUM
COPPER
IRON
MAGNESIUM
MANGANESE
MERCURY
NICKEL
POTASSIUM
SODIUM
STRONTIUM
TIN
TITANIUM
ZINC
ZIRCONIUM
EXTRACTABLE ORGANIC COMPOUNDS
BENZOIC ACID
10
3.1
N/A
9600
4.1
48
2*0
2.4
0.12
1.6
1700
1700
37
2.4 .
1.2
11
N/A
1.1*
14
4.2
N/A-
16000
7.8
24
360
3.3'
0.17*
--
1800
2000
80
--
1.7
20
N/A
--
45
7.2
N/A
12000
5.4
140
260
8.9
0.04
—
2000
1600
23
--
1.6
22
N/A
--
68
7.5
N/A
25000
6.8
700
390
20
0.02
--
1900
1700
42
--
3.0
180
N/A
--
15
9.2
N/A
17000
7.9
25
370
4.9
N/A
--
2100
1300
41
--
1.9
18
N/A
--
31
6.5
N/A
16000
8.4
150
340
5.8
0.03
—
2200
1600
33
.--
2.0
21
N/A
--
•Estimated value
- MATERIAL WAS ANALYZED FOR BUT NOT DETECTED •
-------
Table 2-9. Concentration of Inorganic Elements in Wetland Tissue Samples as a
Multiple of the Low Background Wetland Analysis,8 Bay Drums, Peak
Oil, and Reeves SE Areawide Wetland Impact Study, Tampa, Florida
(U.S. EPA, 1990c)
ALUMINUM
BARIUM
CALCIUM
COPPER
IRON
MAGNESIUM
MANGANESE
MERCURY
NICKEL
POTASSIUM
SODIUM
STRONTIUM
TIN
TITANIUM
ZINC
01 -NOW
NORTH UETLAND
FISH CRAYFISH
3.9 4.5
2.0 2.3
1.3
1.3
2.8 5.8
1.1 1.1
2.1 3.7
H/A
-
1.2
-
-
-
1.3
1.9 2.0
01-UHC
UNNAMED
CREEK
FISH
6.6
2.5
1.2
1.4
19.5
-
1.7
2.0
-
-
1.2
-
-
1.2
14.2
02-CLU
03-UNC CENTRAL 04-CLU 02-SOU
UNNAMED CREEK WETLAND CENTRAL UETLAND SOUTH UETLAND
FISH CRAYFISH FISH FISH CRAYFISH FISH CRAYFISH
3.5 6.8 1.2 2.2 1.5 - 3.1
3.4 2.4 1.9 5.2 3.0 1.9 2.1
2.6 - 1.3 I./
1.3 1.7 1.3 1.6 1.9 1.7 2.0
7.5 29.2 3.9 1.4 - 3.9 6.3
1.6 - 1.5 1.4
1.6 8.3 4.3 2.3 2.0 1.5 2.4
1.3 1.3
.
1.1 - - 1.2 - 1.3
1.1 ----- -
1.1 - - 1.1 - -
.
1.1 2.5 - - 1.o - \.'t
6.9 16.4 1.5 2.3 1.6 3.1 1.9
aFor reference values, see U.S. EPA, 1990c.
- = Concentration equal to or lower than the lower background station concentration.
2-38
-------
Table 2-10. Wetland Surface Water Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide
Wetland Impact Study, Tampa, Florida (U.S. EPA, 1990c)
UIS STATION ID
BORI STATION 10
AWftl/FS STATION ID
INORGANIC ELEMENTS
ALUMINUM
ARSENIC
BARIUM
BERYLLIUM
CADMIUM
CALCIUM
CHROMIUM
COBALT
COPPER
IRON
LEAD
MAGNESIUM
MANGANESE
MOLYBOEHIUM
NICKEL
POTASSIUM
SILVER
SODIUM
VANADIUM
ZINC
01-CPW
01-RFW
CYPRESS
POND
11/28/89
1115
UG/L
1800
--
--
•-
•-
32000
••
••
--
2100
-•
8100
--
HA
--
17000
--
180000
--
--
01-SFW
02-RFW
SPRAY
FIELD
11/28/89
1355
UG/L
210
--
--
--
--
58000
--
--
••
--
•-
13000
65
NA
--
39000
--
250000
•-
••
01 -NOW
A
NORTH
WETLAND
11/30/89
UG/L
422
B 3
B 22.8
--
--
34800
••
--
--
841
-•
B 32B
49.7
--
B 4.1
B 3260
--
29100
--
48.7
01-UNC
' 6
UNNAMED
CREEK
11/30/89
UG/L
77600
S 41.2
268
B 3.1
9.8
185000
135
18.6
60.9
678000
352
1 78QOO
1710
--
155
aaoo
N 10.3
373000
59.9
172000
01-CLW
5R
CENTRAL
WETLAND
1/1/90
UG/L
625
--
8 34.4
--
--
491000
B 5.3
--
••
4640
165
i 5520
72.6
--
--
B 4360
--
288000
8 15.1
E 410
02-CLW
6
CENTRAL
WETLAND
11/30/89
UG/L
B 111
BW 3.2
8 41.8
--
--
E 61700
--
--
--
743
W 4.6
8760
79.1
--
B .7
6680
--
74500
--
49.1
03-CLW
7
CENTRAL
WETLAND
11/30/89
UG/L
206
--
B 13.7
--
--
E 12300
--
--
--
2U
BW 2.5
B 3270
31.3
--
-- •
6670
--
57500
--
B 19
04-CLW
a
CENTRAL
WETLAND
11/30/89
UG/L
390
--
a 16.6
--
-.
E 12900
--
--
-.
260
W 4.6
B 3390
43.2
..
--
6870
-.
59100
--
34.5
01 -SOU
D
SOUTH
WETLAND
11/28/89
UG/L
11200
B 6.2
B 145
8 1.5
B 4.0
50100
27.7
B 4.5
56.4
11000
248
5710
105
B 22.4
B 1170
-.
111000
B 30.5
3980
02 -SOW
E
SOUTH
WETLAND
11/28/89
UG/L
572
..
B 8.3
_.
__
E 9550
B 3.0
..
..
898
W 4.9
B 1150
23.7
B 5.5
8 1490
,.
35600
..
63.2
03 -SOW
f
SOUTH
-WETLAND
11/28/89
UG/L
B 128
,.
B 23
..
„.
f. 55500
..
..
..
..
BW 2.6
8180
57.5
* .
B 7.2
14400
__
171000
T
67.7
04 -SOW
G
SOUTH
WETLAND
11/28/89
UG/L
232
__
B 19.5
..
E 51000
..
..
..
169
W 3.4
7440
41.9
B 4.2
12200
185000
.-
45.9
02-CLW
C
CENTRAL
WETLAND
11/30/89
UG/L
364
._
B 20.4
__
„ .
E 14500
27.7
366
20.1
B 3550
38.4
B 4.6
6660
..
59700
..
32.8
01-UNC
01 -SAP
UNNAMED
CREEK
1/9/90
UG/L
5500
...
.j
80000
16
_.
__
25000
15
6400
370
..
9200
94000
__
11000
GENERAL INORGANIC PARAMETERS
CYANIDE
HG/L
0.01
MG/L
0.01
MG/L
0.01
MG/L
0,04
MG/L
0.01
HG/L
0.016
MG/L
0.019
MG/L
0.01
MG/L
0.015
MG/L
0.01
MG/L
0.019
MG/L
0.01
MG/L
0.015
HG/L
-.
-------
Table 2-10. Wetland Surface Water Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland
Impact Study, Tampa, Florida (continued)
UtS STATION ID
80RI STATION ID
AWfU/FS STATION 10
01-CPW
01-RFU
01-SFU
02-RFU
01 -HOW
A
01-UHC
B
01-CLW
5R
02-CLU
6
03/CLH
7
04-CLW
8
01 -SOU
0
02- SOW
E
03 -SOW
f
04-SOW 02-CLW
G C
01-UHC
01 -SAP
CYPRESS SPRAY
POND FIELD
NORTH
UNNAMED CENTRAL CENTRAL CENTRAL CENTRAL SOUTH
SOUTH
SOUTH
SOUTH
CENTRAL UHNABED
WETLAND CREEK WETLAND WETLAND WETLAND UETIANO WETLAND WETLAND WETLAND WETLAND WETLAND CREEK
11/28/69 11/26/89 11/30/89 "/30/89 1/4/90 11/30/89 11/30/89 11/30/69 11/28/89 11/26/89 11/28/89 11/28/89 11/30/89 1/9/90
HG/L HG/L MG/L HG/L HG/L HG/L HG/L MG/L HG/L HG/L HG/L HG/L HG/L UG/L
• '
PUROEABLE ORGANIC COMPOUNDS
ACETONE
CARBON
DISULFIDE
ETHYLTRIAZOLH
METHYL ETHYL
KETONE
METHYLENE-
CHLORIDE
B 0.0100 B 0.0100
.039
.200 JH
.010 UR
.042
.010 UR
BJ 0.0090 --
-- 0.0580
BJ 0.0810 --
0.0280 J 0.0340 0.0190
B 0.0100
k
EXTRACTABLE ORGANIC COMPOUNDS
BIS --
(2-ETHYLHEXYL)
PHTHALATE
BUTYL BENZYL —
PHTHALATE
DI-N-BUTYL -•
PHTHALATE
OI-H-OCTYL --
PHTHALATE
IOEHO •-
(1.2.3-CD)
PYRENE
4-HETHYL PHENOL —
BJ 0.0040 -- 8J 0.0030 BJ 0.0040 BJ 0.0030 BJ 0.0080 B 0.0120 BJ 0.0070 BJ 0.0080 BJ 0.0030 --
J 0.0020 --
J 0.0020
J 0.0090 J 0.0030 --
B 0.0920 B 0.0910 B 0.0470 B 0.0180
-- B 0.0110 --
-- B 0.0110 --
-- -- J 0.0030 --
ORGANOCHLOR1NE PESTICIDE ANALYSIS
PCB-1260 --
0.0010
-------
Table 2-10. Wetland Surface Water Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland
Impact Study, Tampa, Florida (continued)
The Mowing are AWRI/FS qualifiers:
B - This flag is used when the analyte is found in the associated blank as well as in the sample.
E - This flag identifies compounds whose concentrations exceed the calibration range of the GC/MS instrument for that specific analysis.
J - Indicates an estimated value.
The following are AWRI/FS qualifiers for inorganic analysis:
N - The spiked sample recovery was not within control limits.
S - The value reported was determined by the Method of Standard Additions (MSA).
W - The post-digestion spike for furnace AA analysis is outside of the 85-115% control limits while sample absorbance is less than 50% of the spike
absorbance.
BDRI Station qualifier:
' '
NA - Not analyzed.
General Data Qualifier:
— The analyte was analyzed for but not detected.
WIS: U.S. EPA, Region 4, Environmental Services Division, Ecological Services Branch. Ausust 1990. Wetland Impact Study and Environmental Assessment.
Bay Drums, Peak Oil, and Reeves Southeastern Superfund Sites, Tampa, Florida.
BDRI: U.S. EPA, Region 4, Environmental Services Division, Environmental Compliance Branch, June 25, 1990. Bay Drums First Draft Working Document.
AWRI/FS: Canonie Environmental. February 1990. Area-Wide RI/FS. Bay Drums, Peak Oil, and Reeves Southeastern Sites, Tampa, Florida.
Sources: Data for samples 01-CPW and 01-SFW from BDRI.
Data for all other samples for AWRI/FS.
-------
Table 2-11. Wetland Sediment Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland Impact
Study, Tampa Florida (U.S. EPA, 1990c)
UIS STATION ID
BOD I STATION 10
AWRI/FS STATION 10
INORGANIC ELEMENTS
ALUMINUM
ANTIMONY
ARSENIC
BARIUM
CADMIUM
CALCIUM
CHROMIUM
COBALT
COPPER
IRON
tO LEAD
JS. MAGNESIUM
bJ MANGANESE
MERCURY
NICKEL
POTASSIUM
SELENIUM
SODIUM
VANADIUM
ZINC
01-CPU
01-RFW
CYPRESS
POND
11/28/89
1115
MG/KG
4900
-.
..
22
--
2200
8.1
--
..
720
43J
170
..
.-
3.5J
550
13
29
01-SFU
02-RFU
SPRAY
FIELD
11/28/89
1355
HG/KG
440
.-
..
• -
-.
--
--
—
..
89
2.9.1
..
..
--
-.
..
--
--
01-UOU
A
NORTH
WETLAND
11/30/89
MG/KG
B 4070
-.
12.70
57.70
B 1.10
E 5910
22.30
8 1.70
12.90
1980
266
B 217
21.60
M.10
8 4.50
B 250
..
B 255
B 6.60
EN 355
01-UHC
B-
UNNAMED
CREEK
'11/30/89
HG/KG
E 3790
.-
.-
B 26.70
B 0.800
E 8210
21.9
3 1.70
11.10
22600
70.80
B 401
62.60
"0.22
9.70
a 226
-.
1280
B 6.80
EH 11200
01-CLUa
5R
CENTRAL
WETLAND
1/4/90
MG/XG
*0.70
--
B 0.0014
B 0.0164
--
B 1.220
0.0044
--
B 0.0057
* 0.414
N 0.4150
B 0.0366
0.0092
—
• -
—
—
—
B 0.0015
H 0.5410
02-CLU
6
CENTRAL
WETLAND
11/30/89
•HG/KG
E 1030
--
..
B 13.10
B 0:45
E 3340
8.60
B 0.91
11.600
1520
65.300
8 240
10.40
•0.24
8 2.40
B 261
—
--
B 3.30
EH 402
03-CLW
7
CENTRAL
WETLAND
11/30/89
MG/KG
E 1120
B 3.70
B 0.80
B 8.50
--
E 1060
2.600
--
B 1.10
308
11.30
B 153
2.90
*0.14
B 1.30
B 220
.-
B 332
B 1.90
EH 42.60
04-CLU
8
CENTRAL
WETLAND
11/30/89
MG/KG
E 794
--
..
B 4.90
--
BE 590
2.10
.-
B 0.72
161
8.40
B 83.90
B 1.70
*0.11
B 0.64
B 0.198
--
B 286
B 1.50
EN 32.4
01 -SOU
D
SOUTH
WETLAND
11/30/89
HG/KG
E 1640
.-
..
B 4.0
B 0.36
BE 737
3.300
..
B 0.63
208
7.90
B 47.10
B 1.60
--
..
B 156
.-
B 192
B 1.60
EH 140
02-SOU
E
"SOUTH
WETLAND
11/28/89
HG/KG
B 2160
..
B 1.0
B 7.0
.-
i 1850
3.50
.-
B 1.60
486
11.10
8 70.60
3.10
•0.08
B 2.20
B 152
B 0.75
8 161
b 2.80
EN 206
03-SOU
F
SOUTH
WETLAND
11/28/89
HG/KG
E 1360
..
8 1.40
B 7.40
--
E 1550
3.30
--
B 1.60
279
6.90
8 73.10
B 2.20
•0.08
B 1.60
8 187
--
B 421
B 2.50
EN 234
04-SOU
G
SOW
WETLAND
11/28/89
HG/KG
E 839
..
--
B 4.80
--
E 1590
2.50
--
8 0.88
232
7.10
B 66.90
B 1.30
«0.06
--
B 148
--
B 341
B 1.80
EN 109
02-CLU
C
CENTRAL
WETLAND
11/30/89
HG/KG
E 629
..
--
B 11.30
--
BE 597
2.0
--
--
150
38.10
8 83.10
B 1.80
•0.09
--
B 157
--
B 200
B 1.10
EN 25.60
01-UHC
01-SAP
UNHAHEO
CREEK
1/9/90
MG/KG
220J
--
• —
--
--
550
--
—
—
390J
18
--
--
—
--
--
--
--
1.3
--
:NERAL INORGANIC PARAMETERS
CYANIDE
N 0.54
N 5.70
-------
Table 2-11. Wetland Sediment Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland Impact
Study, Tampa Florida (continued)
UIS STATION 10
BORI STATION ID
AWRI/FS STATION 10
PURGEABLE
ORGANIC
COMPOUNDS
ACETONE
ENZENE
1-BUTANONE
HETHYLENE
CHLORIDE
TOLUENE
XYLENE(S)
EXTRACTABLE ORGANIC
(3-AMO/OR 4-)
HETHYLPHENOL
BENZOIC ACtO
BENZOCA)-
ANTHRACENE
BENZOC8)-
FLUOR-
AHIHEWE
BEHZOCK)
FLUOR -
ANTHENE
BENZO-
CG.H.I.)
PERYLENE
BENZOCA)
PYRENE
BIS
C2-ETHYL-
HEXYL)
PHTHALATE
BUTYL
BENZYL
PHTHALATE
CHRYSENE
01-CPW
01-RFW
CYPRESS
POND
11/28/89
"
HG/KG
R
R
R
R
R
R
COMPOUNDS
290J
--
--
--
--
--
—
--
—
--
01-SFU 01-NOW
02-RFU
A
SPRAY NORTH
MELO WETLAND
11/28/89 11/30/89
HG/KG HG/KG
R B 0.1300
R J 0.003
R 0.033
R B 0.1300
R 3 0.0140
R J 0.0070
..
J 0.1900
..
..
..
..
..
J 0.1400
..
.-
01 -UNCa
B
UHHAMED
CREEK
11/30/89
HG/KG
B 0.0440
--
J 0.0110
B 0.0000
BJ 0.007
-•
-.
--
--
JX 0.120
« 0.1200
--
J 0.0560
J 0.2300
--
J 0.0600
01-CLW 02-CLW
5R 6
CENTRAL CENTRAL
WETLAND WETLAND
1/4/90 11/30/89
MG/KG HG/KG
B 0.0800 B 0.060
J 0.0020
....
"B 0.0210 B 0.0290
0.0340 B 0.0520
--
..
J 0.0870 J 0.2600
J 0.0820 --
JX 0.2700 --
JX0.2700 --
--
J 0.0910 —
J 0.0490 J 0.2300
—
J 0,1200 -
03-CLW 04-CLU 01 -SOU
7 ' 8 D
CENTRAL CENTRAL SOUTH
WETLAND WETLAND WETLAND
11/30/89 11/30/89 11/30/89
HG/KG HG/KG HG/KG
B 0.3800 B 0.130 B 0.0160
J 0.0110 J 0.0050
0.0140
B 0.1900 B 0.1300 3 0.0330
B 0.9300 B 0.5800 B 0.0810
J 0.0140 --
..
J 0.5400 J 0.0440
J 0.3600 -
J 0.2000 --
J 0.1700
a 0.2200 --
J 0.2400
J 0.4300 J 0.0510 --
J 0.2200 -
J 0.2000 —
02-SOU 03-SOW 04-SOU 02-CLM 01-UWC
01 -SAP
E F - G C
SOUTH SOUTH SOUTH CEMTRAt U!WA«£»
WETLAND WETLAND WETLAND WETLAND CREEK
11/28/89 11/28/89 11/28/89 11/30/89 1/9/90
HG/KG HG/KG HG/KG HG/KG HG/KG
B 0.0440 B 0.0500 8 0.0250 --
..
0.0320
B 0.0150 B 0.0130 3 0.0140 -
B 0.1500 B 0.0140 B 0.0380 --
--
,.
—
..
--
—
—
—
J 0.0530 J 0.0750 J 0.0850 --
.. .- -- --
"
-------
Table 2-11. Wetland Sediment Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland Impact
Study, Tampa Florida (continued)
to
UIS STATION ID
BORI STATIC!' ID
AURI/FS STATION ID
_
DIBENZO (A.H)
ANTHRACENE
OI-N-BUTYL-
PHTHALATE
OI-N-OCTYL
PHTHALATE
FLUORANTHENE
HEXADECENOIC ACID
INDENO-
(1.2.3-CD)
PYRENE
NITROBENZENE
4-NITRO
PHENOL
PHENOL
PYRENE
01 -CPU 01-SFU
01-RFU 02-RFW
CYPRESS SPRAY
POMD FIELD
11/28/89 11/28/89
KG/KG MG/KG
..
•-
—
..
3000JN
..
1800UR 920UR
..
..
"
01-HOU
A
WORTH
WETLAND
11/30/89
HG/KG
..
--
--
--
••
--
-.
--
--
"
01-UHC 01-CLU
B 5R
UNNAMED CENTRAL
CREEK UETLAHD
11/30/89 1/4/90
KG/KG HG/KG
—
1.3000
---.
J 0.0750
..
..
.-
..
J 0.0910 -- '
J 0.0610 J 0.0800
OZ-CLV
6
CENTRAL
UETLAND
11/30/89
HG/KG
..
J 0.1100
"
"
"
--
--
"
03-CLU 04-CLU
7 8
CENTRAL CENTRAL
UETLAUO WETLAND
11/30/89 11/30/89
HG/KG MG/KG
J 0.2000 --
J 0.1500 --
J 0.1500 --
..
—
J 0.2200 --
.. ..
J 0.2200 •-
J 0.3700 --
01-SOW 02-SOW 03-SOW 04-SOU 02-CLW 01-UHC
01-SAP
D E F G C
SOUTH SOUTH SOUTH SOUTH CENTRAL UHNAHED
UETLANO yETLAHD UETLAND UETLANO UETLAND CREEK
11/30/89 11/28/89 11/28/89 11/28/89 11/30/89 1/9/90
•HG/KG HG/KG HG/KG Ho/kc HG/Kfi Htf>kd '
.-
J 0.3900 J 0.0740 --
--
..
..
--
..
.. .. -. -. -- _ .
J 0.0540 •- J 0.0760 J 0.3600 J 0.0450 —
iGAWOCHLORlWE PESTICIDES ANALYSIS
GAMMA
CHLORDANE
4.4'-ODD
4.4' -DDE
PCB-1260
.. ..
..
..
..
J 0.0970
J 0.0320
0.1200
0.0810
—
--
0.2600
j O.TOOO
-•
0.1200
J 0.4800
,_
--
..
..
..
--
..
J 0.4900 --
-------
Table 2-11. Wetland Sediment Analytical Data Summary, Bay Drums, Peak Oil, and Reeves SE Areawide Wetland Impact
Study, Tampa Florida (continued)
The following are AWRI/FS qualifiers:
B - This flag is used when the analyte is found in the associated blank as well as in the sample.
J - Indicates an estimated value.
X - Other specific flags and footnotes may be required to properly define the results.
The following are AWRI/FS qualifiers for inorganic analysis:
B - The reported value is less than the contract required detection limit (CRDL) but greater than the instrument detection limit (IDL).
E - The reported value is estimated because of interference.
N - The spiked sample recovery was not within control limits.
* - Refer to the original lab report for the Form 1, Sample Data Summary, Case Narrative.
Also the "X" flag definition is specific for each result; refer to the original lab report to find out what combination of flags "X" stands for.
The following are BDRI station qualifiers:
J - Estimated value.
N - Presumptive evidence of presence of material.
R - Quality control indicates that data are unusable, compound may or may not be present, resampling and reanalysis are necessary for verification, the
value is that reported by the laboratory. All purgeable organic compounds were reported as "R" for sediment samples 01-CPW and 01-SFW.
U - Material was analyzed for but not detected. The number shown is the minimum quantitation limit.
General data qualifier:
— The analyte was analyzed for but not detected.
WIS: U.S. EPA, Region 4, Environmental Services Division, Ecological Services Branch. August 1990.
Wetland Impact Study and Environmental Assessment. Bay Drums, Peak Oil, and Reeves Southeastern Superfund Sites, Tampa, Florida.
BDRI: U.S. EPA, Region 4, Environmental Services Division, Environmental Compliance Branch, June 25, 1990.
Bay Drums First Draft Working Document.
AWRI/FS: Canonic Environmental. February 1990. Area-Wide RI/FS. Bay Drums, Peak Oil, and Reeves Southeastern Superfund Sites,
Tampa, Florida.
alt appears that metals and cyanide values reported in the AWRI/FS as originating with the Compuchem subcontractor are reported about 3 orders of
magnitude lower than is probable. We have interpreted these values for the purpose of this WIS to be g/kg rather than tag/kg.
Sources: Data for samples 01-CPW and 01-SFW from BDRI.
Data for all other samples from AWRI/FS.
-------
Table 2-12. Summary of EPA Ambient Water Quality Criteria and Screening Concentrations for the Protection of
Freshwater Biota (U.S. EPA, 1990c, adapted from table 4.28)
TSS790
UPDATE:
EPA REG
304(a)
FOR TOX
SCREENING LIST
JANUARY 1991
IV - WATER MANAGEMENT DIVISION D
SCREENING VALUES AND RELATED INFORMATION
tr oni i MTAUTC
E T E C
[40
DATE COMPOUND
REVISED
4/89
7/89
7/90
7/90
12/89
7/90
1/91
4/89
6/89
4/89
7/90
7/90
11/89
1/91
7/90
4/89
7/89
4/89
4/89
4/89
1/91
6/89
1/91
1/91
4/89
7/90
4/89
1/91
4/89
4/89
4/89
4/89
11/89
11/89
4/89
7/90
PRIORITY POLLUTANTS
1 m Antimony (B)
2 in Arsenic (c)
3 m Beryllium (c)
4 ra Cadmium (H)
5 m Chromium (III) (H)
5 m Chromium (VI)
6 m Copper (H)
7 m Lead (H)
8 m Mercury
9 m Nickel (H)
10 m Selenium
11 m Silver (H, 8)
12 m Thallium
13 m Zinc (H)
14 Cyanide
Asbestos (c)
2.3,7,8-TCOD-Oioxin (c).
1 v Acrolein
2 v Acrylonitrile (c)
3 v Benzene (c)
5 v Bromoform (c)
6 v Carbon Tetrachloride (c)
7 v Chlorobenzene
8 v Chlorodibromomethane (c)
9 v Chloroethane
10 v 2-Chloroethylvinyl Ether (c)
11 v Chloroform (HM, c)
12 v Dichlorobromomethane (c)
14 v 1,1-Di chloroethane
15 v 1,2-Oichloroethane (c)
16 v 1,1-Dichloroethylene (c)
17 v 1,2-Oichloropropane
18 v 1,3-Oichloropropylene (Cis)
v 1,3-Dichloropropylene (Trans)
19 v Ethylbenzene
20 v Methyl Bromide
Cjig/W
200
2
5
5
50
5
20
100
0.2
40
2
10
100
5
5
0.00001
nr
nr
4.4
4.7
2.8
6
3.1
nr
nr
1.6
2.2
4.7
2.8
2.8
6
5
nr
7.2
nr
EPA
T I 0 N
CFR 1361
Ref.
Method
204.1
•206.3
210.1
213.1
218.1
218.4
220.1
239.1
245.1
249.1
270.3
272.1
279.1
289.1
335.3
hrms
624x
624x
624
624
624
624
624
624
624
624
624
624
624
624
624
624
624
624
LEV
Oig/U
3
1
0.2
0.1
1
1
1
~~
1
2
0.2
1
0.05
--
0:002
0.7
0.5
0.2
0.2
0.12
0.25
0.09
0.52
0.13
0.05
0.1
0.07
0.03
0.13
0.04
Q.34
0.2
0.2
1.18
E L
Ref.
Method
204.2
206.2
210.2
213.2
218.2
220.2
239.2
249.2
270.2
272.2
279.2
289.2
613
603
603
602
601
601
601
601
601
601
601
601
601
601
601
601
601
602
601
F R
Screening
l/a I fjo
Value
(Maximum)
(Jjg/U
Hardnessdng/l
1300 2s
360 Mil
16 6s
1.79 »
984.32 *
16 *
9.22 «
33.78 *
2.40 *
789.00 *
20.00 *
1.23 *
140.00 3s
65.04 *
22'*
..
0.1
6.8 3s
755 4s
530 7s
2930 2s
3520 3s
1950 5s
--
--
35400 Is
2890 3s
--
—
11800 3s
3030 3s
5250 3s
606 2s
606 2s
4530 5s
1100 1s
E S H U A T E
Screening
Value
(Continuous)
Oig/H
as CaC03):
pt):
160 2s
190 Mil
0.53 1s
0.66 *
117.32 *
11 *
6.54 *
1.32 *
0.012 *T
87.71 *
5.00 *
0.012 2s
4.00 2s
58.91 *
5.2 *
..
0.00001 T
2.1 2s
75.5
' 53
293
352
195
--
--
3540
289
--
--
2000 Is
303
525
24.4 Is
24.4 1s
453
110
R
95X
LC50
Value
(>ig/U
50.0
6
1300
720 Mil
16 6s
3.59 *
1968.63 *
32 *
18.45 *
67.57 *
4.8 *
1578.01 *
40 *
1.23 *
140 '
130.09 *
44 *
__
0.1
6.8
755
530
2930
3520
1950
--
--
35400
2890
--
--
11800
3030
5250
606
606
4530
1100
CRITERIA
DATES
10/80, 1/87:RfD 0.0004
1/85:aq life, 6/21/88:q1* 1.75
10/80, 1/90:q1* 4.3
10/80, 1/85:aq life, 10/89:RfO
0.0005(water) O.OOKfood)
10/80, 1/85:aq life, 3/88:RfD 1
10/80, 1/85:aq life, 3/88:RfD 0
10/80, 1/85:aq life
10/80, 1/85:aq life
10/80, 1/85:aq life, 2/89:RfD 0
10/80, 9/86:aq life, 3/88:RfD 0
10/80, 9/87:aq life
10/80, 6/88:RfD 0.003
10/80
10/80, 2/87:aq life
10/80, 1/85:aq life, 3/88:RfD 0
10/80
2/84 :q1* 156000
10/80
10/80, 2/89:q1* 0.54
10/80, 12/88:q1» 0.029
10/80, 6/88:q1» CHC13, 9/90:q1*
10/80, 3/88:q1* 0.13
10/80, 11/90:RfO 0.02
10/80, 6/88:q1* CHC13, 11/90:q1
10/80
10/80
10/80, 6/88:q1* 0.0061
10/80, 6/88:ql* CHC13, 10/90:q1
10/80
10/80, 3/88:q1* 0.091
10/80, 12/88:q1* 0.6
10/80
10/80, 3/B8-.RfO 0.0003
10/80, 3/88:RfD 0.0003
10/80, 3/88:RfD 0.1
10/80, 8/90:RfD 0.0014
Admin, memo
.005
.0003
.02
.02
0.0079
* 0.084
* 0.13
-------
Table 2-12. Summary of EPA Ambient Water Quality Criteria and Screening Concentrations for the Protection of Freshwater
Biota (continued)
TSS790
UPDATE:
EPA REC
304(a)
SCREENING LIST
JANUARY 1991
tW LlATCO MAUAfCIUCUT (\f Vf Cf HU n
tv " WA i cK nftnnutnt N 1 u i V i a I UN u
SCREENING VALUES AND RELATED INFORMATION
ET p C
ICu
[40
EPA
T r f\ u
I i >J n
CFR 136]
I P V
u c *
FOR TOX"" Dnt IIIT4HTS
DATE
REVISED
E'/tff
4/89
6/89
4/89
1/91
7/90
7/90
4/89
7/89
4/89
1/91
1/91
1/91
4/89
4/89
4/89
JO 4/89
.L. V91
S V91
1/91
7/90
1/91
1/91
1/91
4/89
4/89
4/89
4/89
1/91
4/89
4/89
V89
/89
19
,1/89
1/91
4/89
1/91
4/89
11/89
4/89
7/90
1/91
COMPOUND
21 v Methyl Chloride (HM, c)
22 v Methylene Chloride (c)
23 v 1,1,2,2-Tetrachloroethane (c)
24 v Tetrachloroethylene (c)
25 v Toluene
26 v 1,2-Trans-Dichloroethylene
27 v 1,1,1-Trichloroethane
28 v 1,1,2-Trichloroethane (c)
29 v Trichloroethytene (c)
31 v Vinyl Chloride (c)
1 a 2-Chlorophenol
2 a 2,4-Dichlorophenol
3 a 2,4-Dimethylphenol
4 a 2-Methyl-4,6-Dinitrophenol
5 a 2,4-Oinitrophenol
"6 a 2-Nitrophenol
7 a 4-Nitrophenol
8 a 3-Methyl-4-Chlorophenol
9 a Pentachlorophenol (pH)
10 a Phenol
11 a 2,4,6-Trichlorophenal (c)
1 bn Acenaphthene
2 bn Acenaphthylene
3 bn Anthracene
4 bn Benzidine (c)
5 bn Benzo(a)Anthracene (PAH, c)
& bn Benzo(a)Pyrene (PAH, c)
7 bn 3,4-Benzofluoranthene (PAH, c)
8 bn Benzo(ghi)perylene
9 bn Benzo(lc)Fluoranthene (PAH.c)
10 bn Bis(2-Chloroethoxy)Hethane
11 bn Bis(2-Chloroethyl)Ether (c)
12 bn Bis(2-Chioroisopropyl)Ether
13 bn 8is(2-Ethylhexyl)Phthatate (c, 8
14 bn 4-8roraopheny(Phenyl Ether
15 bn Butylbenzyl Phthalate
16 bn 2-Chloronaphthalene
17 bn 4-Chlorophenyl Phenyl Ether
18 bn Chrysene (PAH, c)
19 bn Dibenz(a,h)Anthracene (PAH, c)
20 bn 1,2-Oichlorobenzene
21 bn 1,3-Dichlorobenzene
22 bn 1,4-Dichlorobenzene
23 bn 3,3'-Oichlorobenzidine (c)
(yg/U
nr
2.8
6.9
4.1
6
1.6
3.8
5
1.9
nr
3.3
2.7
2-7
24
42
3.6
2.4
3
3.6
1.5
2.7
1.9
3.5
1.9
44
7.8
2.5
2.5
4.1
2.5
5.3
5.7
5.7
2.5
1.9
2.5
1.9
4.2
2.5
2.5
nr
nr
nr
16.5
Ref.
Method
624
624
624
624
624
624
. 624
624
624
624
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
624
624
624
625
Ciig/U
0.08
0.25
0.03
0,03
0.2
0.1
0.03
0.02
0.12
0.18
0.31
0.39
0.32
16
13
0.45
2.8
0.36
7.4
0.14
0.64
1.8
2.3
0.66
--
0.013
0.023
0.018
0.076
0.017
0.5
0.3
0.8
2
2.3
0.34
0.94
3.9
0.15
0.03
1.9
1.9
4.4
0.13
Ref.
Method
601
601
601
601
602
601
601
601
601
601
604
604
604
604
604
604
604
604
604
604
604
610
610
610
610
610
610
610
610
' 611
611
611
606
611
606
612
611
610
610
625
625
625
605
Screening
Value
Maximum)
'(ug/U
"55000
19300
932
528
1750
13500
5280
3600
-
438
202
212
23
62
"
828
3
3.32
1020
32
170
--
--
250
--
--
--
--
--
--
23800
1110
36
330
--
--
~-
158
502
112
--
F R
1s
3s
3s
5s
5s
1s
2s
3s
5s
3s
3s
4s
3s
3s
ESHWATER
Screening
Value
(Continuous)
C>J9/U
5500
1930
240 1s
84 1s
175
1350
528
940 1s
"~
43.8
36.5 1s
21.2
2.3
6.2
3500
82.8
1s 0.3
* 2.10 *
16s 256 Is
3s
2s
4s
1s
2s
2s
4s
4s
3s
5s
3.2
17
--
--
25
--
--
. .,
-"_
--
--
2380
<0.3 2s
12.2 1s
22 2s
-•
--
-•*
15.8 3s
50.2
11.2
--
95%
LCSO
Value
(jjg/D
55000 ,
19300
932
528
1750
13500
5280
3600
•""
438
202
212
23
62
--
828
3
29.98 *
1020
32
170
--
--
250
--
--
--
--
--
-•
23800
1110
36
330
-•
--
-*•
158
502
112
--
CRITERIA
DATES
10/80,
10/80,
10/80,
10/80
10/80,
10/80,
10/80,
10/80,
10/80
10/80
10/80,
10/80,
10/80,
10/80
10/80,
10/80
10/80
10/80
10/80,
10/80,
10.80,
10/80,
10/80
10/80,
10/80,
10/80
10/80
10/80
10/80
10/80
10/80
10/80,
10/80,
10/80,
10/80
10/80,
10/80.
10/80
10/80
10/80
10/80,
10/80
10/80
10/80,
6/88:q1* CHC13
1/89:q1* 0.0075
3/88:q1* 0.2
3/88:RfD 0.3, 8/90:RfD 0.2
1/89:RfD 0.02
6/88:RfO 0.09
3/88:q1* 0.057
8/88:RfD 0.005
6/8S:RfD 0.003
11/90:RfD 0.02
3/88:RfD 0.002
9/86:aq life, 6/88:RfD 0.03
6/89:RfD 0.6
6/90:q1* 0.011
11/90:RfO 0.06
9/90:RfD 0.3
3/88:q1* 230
3/85 :q1* 1.1
10/89:RfD 0.04
2/89:q1* 0.014
9/89 :RfD 0.2
11/90:RfO 0.08
8/89:RfD 0.09
8/90: ql* 0.45
-------
Table 2-12. Summary of EPA Ambient Water Quality Criteria and Screening Concentrations for the Protection of
Freshwater Biota (continued)
I
TSS790
UPDATE:
EPA REG
SCREENING LIST
JANUARY 1991
tu . UATPD Uftuirtpupur niurcinu n
304(a) SCREENING VALUES AND RELATED INFORMATION
E T E C
C40
EPA
T I 0 H
CFR 136]
LEV
E L
FOR TOX'r Pni I IITAMTS
DATE
REVISED
11/6*
4/89
4/89
4/89
4/89
4/89
4/89
1/91
1/91
4/89
7/89
1/91
4/89
4/89
1/91
4/89
12/89
10/90
1/91
4/89
1/91
1/91
11/89
7/90
4/89
4/89
4/89
4/89
4/89
6/89
7/90
7/90
4/89
4/89
4/89
4/89
7/90
7/90
4/89
12/89
6/89
6/89
6/89
6/89
6/89
COMPOUND
24 bn Di ethyl Phthalate
25 bn Dimethyl Phthalate
26 bn Di-n-Butyl Phthalate
27 bn 2,4-Oinitrotoluene (c)
28 bn 2,6-Oinitrotoluene
29 bn Di-n-Octyl Phthalate
30 bn 1,2-Diphenylhydrazine (c)
31 bn Fluoranthene
32 bn Fluorene
33 bn Hexachlorobenzene (c, B)
34 bn Hexachlorobutadiene (c)
35 bn Hexachlorocyclopentadiene
36 bn Hexachloroethane (c)
37 bn Indeno(1,2.3-cd)Pyrene (PAH, c)
38 bn Isophorone (c)
39 bn Naphthalene
40 bn Nitrobenzene
41 bn N-Nitrosodimethylamine (c)
42 bn H-Nitrosodi-n-Propylamine (c)
43 bn N-Nitrosodiphenylamine (c)
44 bn Phenanthrene (B)
45 bn Pyrene
46 bn 1,2,4-Trichlorobenzene
1 p Aldrin (c)
2 p a-BHC (c)
3 p b-BHC (c)
4 p g-BHC (c)
5 p d-BHC .(c)
6 p Chlordane (c)
7 p 4-4' -DOT (c>
8 p 4, 4' -ODE (c)
9 p 4,4'-ODO (c)
10 p Oieldrin (c)
11 p a-Endosulfan
12 p b-Endosulfan
13 p Endosulfan Sulfate
14 p Endrin
15 p Endrin Aldehyde
16 p Heptachlor (c)
17 p Heptachlor Epoxide (c)
18 p PCB-1242 (PCB, c)
19 p PCB-1254 (PCB, c)
20 p PCB-1221 (PCB, c)
21 p PCB-1232 (PCB, c)
22 p PCB-1248 (PCB, c)
C/jg/O
" 1.9
1.6
2.5
5.7
1.9
2.5
20
2.2
1.9
1.9
0.9
nr
1.6
3.7
2.2
1.6
1.9
nr
nr
1.9
5.4
1.9
1.9
1.9
nr
4.2
nr
3.1
nr
4.7
5.6
2.8
2.5
--a
-b
5.6
nr
nr
1.9
2.2
nr
36
30
nr
nr
Ref.
Method
625
625
625
625
. 625
625
1625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
(jjg/U
0.49
0.29
0.36
0.02
0.01
3
0.21
0.21
0.05
0.34
0.4
0.03
0.043
15.7
1.8
13.7
0.15
0.46
0.81
0.64
0.27
0.05
0.004
0.003
0
0
0.009
0.014
0.012
0.004
0.011
0.002
0.014
0.004
0.066
0.006
0.023
0.003
0.083
0.065
nr
nr
nr
nr
Ref.
Method
606
606
606
609-EC
609-EC
606
610
610
612
612
612
612
610
609-EC
610
609-EC
607
607
607
610
610
612
60S
603
608
608
608
608
60S
608
608
608
608
608
608
608
608
608
608
608
608
608
608
608
F R
Screening
Value
(Maximum)
(ug/U
5216 2s
3300 2s
94 6s
3100 2s
—
—
27 2s
398 2s
--
--
9 5s
0.7 4s
98 5s
--
11700 2s
230 4s
2700 2s
--
585 2s
• -
• -
150 4s
3 *
..
--
2 *
--
2.4 *
1.1 *
105 1s
0.064 8s
2.5 *
0.22 *
0.22 *
--
0.18 *
—
0.52 *
0.52 *
0.2 7s
0.2 7s
0.2 7s
0.2 7s
0.2 7s
ESHUATER
Screening
Value
(Continuous)
OJg/U
521
330
9.4
310
—
—
2.7
39.8
--
--
0.93 1s
0.07
9.8
--
1170
62 1s
270
--
58.5
--
--
44.9 1s
0.3
500 p
5000 p
0.08 *
--
0.0043 *T
0.001 *U
10.5
0.0064
0.0019 *T
0.056 *
0.056 *
--
0.0023 *T
--
0.0038 *T
0.0038 *T
0.014 *U
0.014 *U
0.014 *W
0.014 *W
0.014 *U
95X
LC50
Value
(Jjg/U
5210
3300
94
3100
--
--
27
398
--
--
9
0.7
98
--
11700
230
2700
--
585
--
--
150
3 *
--
—
2 *
--
2.4 *
1.1 *
105
0.064
2.5 *
0.22 *
0.22 *
--
0.18 *
"
0.52 *
0.52 *
0.2
0.2
0.2
0.2
0.2
CR
ITPfflA
DATES
10/80,
10/80
10/80,
10/80
10/80
10/80
10/80,
10/80,
10/80,
10/80
10/80,
10/80,
10/80,
10/80
10/80,
10/80
10/80,
10/80,
10/80,
10/80,
10/80
10/80,
10/80
10/80,
10/80,
10/80,
10/80
10/80
10/80,
10/80,
10/80,
io/ao.
10/80,
10/80,
10/80,
10/80
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
9/87:RfO
1/87:RfD
3/88:q1*
9/90:RfD
9/90:RfD
3/88:q1«,
3/88:RfD
3/88:q1*
9/89:RfD
5/88:RfD
3/88:q1*
3/88:ql*
3/88:q1*
9/90:RfD
12/88: ql
3/88:q1*
9/87:q1*
3/88:q1*
8/88:q1«
8/88:q1*
8/88:qI*
9/88:q1*
3/88:RfO
3/88:RfD
9/88:RfD
9/88:RfD
3/88:q1*
3/88:q1*
5/89:q1*
5/89:q1*
5/89:q1*
5/89:q1*
5/89 :q1*
0.8
0.1
0.8
0.04
0.04
6/89:q1* 0.078
0.007
0.014
0.2, 8/90:q1* 0.0041
0.0005
51
7.0
0.0049
0.03
• 17
6.3
1.8
1.3
0.34
0.34
0.24
16
O.jQOOOS
0.00005
0.0003
0.0003
4.5
9.1
7.7
7.7
7.7
7.7
7.7
oo
-------
Table 2-12. Summary of EPA Ambient Water Quality Criteria and Screening Concentrations for the Protection of
Freshwater Biota (continued)
TSS790
UPDATE:
PDA npr;
urn NCli
304(8)
FOR TOX
DATE
REVISED
11/69
A/89
4/89
4/89
4/89
4/89
4/89
1/91
1/91
4/89
7/89
1/91
4/89
4/89
1/91
4/89
12/89
10/90
1/91
4/89
1/91
1/91
11/89
7/90
4/89
4/89
4/89
4/89
4/89
6/89
7/90
7/90
4/89
4/89
4/89
4/89
7/90
7/90
4/89
12/89
6/89
6/89
6/89
6/89
6/89
SCREENING LIST
JANUARY 1991
IV " WATER MANAGEMENT DIVISION
SCREENING VALUES AND RELATED INFORMATION
1C POLLUTANTS
COMPOUND
24 bo Di ethyl Phthalate
25 bn Dimethyl Phthalate
26 bn Di-n-Butyl Phthalate
27 bn 2,4-Oinitrotoluene (c)
28 bn 2,6-Oinitrotoluene
29 bn Di-n-Octyl Phthalate
30 bn 1,2-Oiphenylhydrazine (c)
31 bn Fluoranthene
32 bn Fluorene
33 bn Hexachlorobenzene (c, B)
34 bn Hexachlorobutadiene (c)
35 bn Hexachlorocyclopentadiene
36 bn Hexachloroethane (c)
37 bn Indeno(1,2,3-cd)Pyrene (PAH, c)
38 bn Isophorone (c)
39 bn Naphthalene
40 bn Nitrobenzene
41 bn H-Hitrosodimethylaimne (c)
42 bn N'Nitrosodi-n-Propylamine (c)
43 bn N-Nitrosodiphenylamtne (c)
44 bn Phenanthrene (B)
45 bn Pyrene
46 bn 1,2,4-Trichlorobenzene
1 p Aldrin (c)
2 p a-BHC (c)
3 p b-BHC (c)
4 p g-BHC (c)
5 pd-BHC.
9 p 4,4'-ODD (C>
10 p Oieldrin (c)
• 11 p a-Endosulfan
12 p b-Endosulfan
13 p Endosulfan Sulfate
14 p Endrin
15 p Endrin Aldehyde
16 p Heptachlor (c)
17 p Heptachlor Epoxide (c)
18 p PCB-1242 (PCB, c>
19 p PCB- 1254 (PCB, c)
20 p PCB-1221 (PCB, c)
21 p PCB-1232 (PCB, c)
22 p PCB-1248 (PCB, c)
D E T E C
(40
(>ig/O
1.9
1.6
2.5
5.7
1.9
2.5
20
2.2
1.9
1.9
0.9
nr
1.6
3.7
2.2
1.6
1.9
nr
nr
1.9
5.4
1.9
1.9
1.9
nr
4.2
nr
3.1
nr
4.7
5.6
2.8
2.5
--a
-b
5.6
nr
nr
1.9
2.2
nr
36
30
nr
nr
EPA
T I 0 N
CFR 136]
Ref.
Method
625
625
625
625
. 625
625
1625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
. 625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
625
LEV
(ug/u
0.49
0.29
0.36
0.02
0.01
3
0.21
0.21
0.05
0.34
0.4
0.03
0.043
15.7
1.8
13.7
0.15
0.46
0.81
0.64
0.27
0.05
0.004
0.003
0
0
0.009
0.014
0.012
0.004
0.011
0.002
0.014
0.004
0.066
0.006
0.023
0.003
0.083
0.065
nr
nr
nr
nr
E L
Ref.
Method
606
606
606
609-EC
609-EC
606
610
610
612
612
612
612
610
609-EC
610
609-EC
607
607
607
610
610
612
608
608
608
608
608
608
60B
608
608
608
608
608
608
608
60S
608
608
608
608
608
608
608
f R
Screening
Value
(Maximum)-
(jjg/U
" 5216 "2s
3300 2s
94 6s
3100 2s
--
--
27 2s
398 2s
—
--
9 5s
0.7 4s
98 5s
--
11700 2s
230 4s
2700 2s
-.
--
585 2s
--
--
150 4s
3 *
--
--
2 *
--
2.4 *
I.I *
105 1s
0.064 8s
2.5 *
0.22 *
0.22 *
--
0.18 *
--
0.52 *
0.52 *
0.2 7s
0.2 7s
0.2 7s
0.2 7s
0.2 7s
ESHVIATER
Screening
Value
(Continuous)
(>ig/U)
521
330
9.4
310
--
--
2.7
39.8
—
--
0.93 1s
0.07
9.8
--
1170
62 1s
270
--
--
58.5
--
--
44.9 1s
0.3
500 p
5000 p
0.08 *
--
0.0043 *T
0.001 *W
10.5
0.0064
0.0019 *T
0.056 *
0.056 *
.,
0.0023 *T
--
0.0038 *T
0.0038 *T
0.014 *VI
0.014 *U
0.014 *U
0.014 *«
0.014 *U
95%
LC50
Value
(Hg/U)
5215
3300
94
3100
--
--
27
398
-.
--
9
0.7
98
.-
11700
230
2700
--
--
585
.-
--
150
3 *
--
—
2 *
--
2.4 *
1.1 *
105
0.064
2.5 *
0.22 *
0.22 *
--
0.18 *
--
0.52 *
0.52*
0.2
0.2
0.2
0.2
0.2
CRITERIA
DATES
10/80,
10/80
10/80,
10/80
10/80
10/80
10/80,
10/80,
10/80,
10/80
10/80.,
10/80,
10/80,
10/80
10/80,
10/80
10/80,
10/80,
10/80,
10/80,
10/80
10/80,
10/80
10/80,
10/80,
10/80,
10/80
10/80
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
10/80,
9/87:RfD
var:Rfo
3/88:q1*
9/90:RfD
9/90:RfD
3/88:q1*,
3/88:RfD
3/88:q1*
9/89:RfO
5/88:Rfl>
3/88:q1*
3/88:q1*
3/88:q1«
9/90:RfD
12/88: q1
3/88:q1*
9/87:q1*
3/88:q1*
8/88:q1*
8/88:q1*
8/88:q1*
9/88:q1*
3/88:RfD
3/88:RfO
9/88:RfD
9/88:RfD
3/88:q1*
3/88:q1*
5/89:q1*
5/89:q1*
5/89:q1*
5/89:q1*
5/89:q1»
o.a
0.1
0.8
0.04
0.04
6/89:q1* 0.078
0.007
0.014
0.2, 8/90:q1* 0.0041
0.0005
51
7.0
0.0049
0.03
* 17
6.3
1.8
1.3
0.34
0.34
0.24
16
0.00005
0.00005
0.0003
0.0003
4.5
9.1
7.7
7.7
7.7
7.7
7.7
to
fe
-------
Table 2-12. Summary of EPA Ambient Water Quality Criteria and Screening Concentrations
for the Protection of Freshwater Biota (continued)
KET
m: metal
ci carcinogen, 10-6 risk level
Os based on orgoholeptle data
HCLs SDUA volue
Us Final Residue Value based on wildlife feeding study
It bosed on marketability of Mob ' '•
Xs t\ot recwrroended If conpound knovm to be present In Benple
nr» not reported • •
hrmsj high resolution mass spectroscopy , ,
lilt! halomethane, human health criteria apply to total halomethanes
PAIl! polynuclear aromatic hydrocarbon, human health criteria apply to total PAHs
V! volatile cottpounds
as acidic cwtpounds
ECs electron capture detector
fit flame lonttatlon detector
PCBf polychlorlnated blplienyl criteria apply to total PCBs
IRCs meosured as total residual chlorine
q1*s Cancer Potency Factor
*s criterion
lit; trlvaUnt fcpacles
VI! hexavalent species
ss iiutfaer of species
Irs for long term Irrigation of sensitive crops (mlhlmum standard)
ps lowest plant value reported ••' ' •
BCFi bloconcentratlon factor •» tissue concentration divided by Hater concentration
d! eee table" Ambient Water buatlty Criteria for Atmwnla-1984 EPA 440/5-84-001
Cllcl3t Based on chloroform criteria
RIDs verified Reference Dose for Hone^arelnogens
es aee table Anblent Ualer Quollty Criteria for Aimwnla (Saltwater) 6PA 440/5-88-004
f/ls mwfcer of flbera per liter of water - based on consumption of Mater only
Us based on hardness equation
pit: based on pi I equation
bm bose neutral conpounds'
f: freshuater organisms
e/cs estuarlne/coastal organisms
oo! op«n ocean Cmarlne) organisms
Hardness Equations:
COHPOOHD CHC CCC 95X IC50
Cotinlun e(1.1Z8(lnll)-3.82B) el0.7852(lr\!l)-3.47> 2e(1.128(lnll)-3.828)
Chromium III e(O.Bl9(lnll)*3.688) e(0.819(lnll)+1.561) 2e(0.819(lnH)+3.688)
^PP^1" e(0.9422anll)-1.464) e(0.8S45(lnll)-1.465)
••end «(1.8?3< lull).-1.46) e(1.2ntlnll)-4.703)
IIIck81 'etO.B^ednlDtS.SSIZ) e(O.B46(lnll)t1.1645)
Silver e(1.r2(lnll)-6.52) ' e(1.72(lnH)-6.52)
Zinc e(0.8«73(ln!l)»0.8604) e(
2-50
-------
screening value, also obtained from the relevant EPA ambient water quality criteria document, was
used when available.
For sediments, analytical values were compared with biologically effective sediment
concentrations from a recent publication of the National Oceanic and Atmospheric Administration
(NOAA): "Technical Memorandum NOS OMA 52: The Potential for Biological Effects of
Sediment-Sorbed Contaminants Tested in the National Status and Trends Program" (Long and
Morgan, 1990). This memorandum compiles data from existing studies that link sediment
concentrations of specific trace elements, PCBs, pesticides, and polycyclic aromatic hydrocarbons,
(PAHs) with predicted or observed biological effects (table 2-13). Data were screened and ordered
to select concentrations that represent the lower 10th percentile of the screened available data (ER-
L: Effects Range-Low), the 50th percentile of the screened available data (ER-M: Effects Range-
Median), and the overall apparent effects threshold (AET).
Sediment concentrations of chemicals that exceeded the ER-L concentration were identified
for each study site. The ER-L is not a particularly conservative value, since 10 percent of the
studies conducted have demonstrated or predicted actual biological effects at concentrations at or
below that level. No safety factor was applied to the ER-L hi generating overage tables.
Therefore, the ER-L may not always be adequately protective of the aquatic environment and
measured concentrations that exceed the ER-L should be viewed as potential environmental
hazards. (Although the NOAA document concentrates on estuarine sediments and sediments
sampled for the AWIS are essentially freshwater, often the difference between effective
concentrations for specific chemicals in freshwater and saltwater habitats is fairly narrow. No
similar compilation and synthesis of sediment toxicity information was found or is believed to exist
for freshwater systems.)
The benchmark concentration used for comparison with chemical concentrations in soil
samples collected from the Superfund sites was also the ER-L. Although based on sediment rather
than soil literature, the NOAA sediment benchmarks represent the only consolidated source of
guidance for toxic effects in solid media at this point. Furthermore, there is often little
difference in the physical characteristics of soils and sediments here: both are primarily sand, the
water table is within inches of the soil surface over most of the site, and precise definition of
whether a specific sample represents soil or sediment may, in many cases, be a function of recent
rain events and water table level at the time of sampling. Nevertheless, the level of uncertainty
associated with using the ER-L as a benchmark for the concentration of a chemical in soil will be
greater than its use as a benchmark for the concentration of a chemical in sediments.
From tabulated environmental concentration/benchmark ratios it was apparent that the
greatest potential for environmental effects associated with contaminants in wetland sediments was
concentrated in a few specific areas. The highest ratio of sample concentration to benchmark
(table 2-14) was for chlordane in the North Wetland: 194 times benchmark (194X). That same
area contained levels of 4,4'-DDE (dichlorodiphenyldichloroethylene), 4,4'-DDD
(tetrachlorodiphenylethane), total lead, and mercury at 60X, 16X, 8X, and 7X their benchmark
values, respectively. The North Wetland area historically has received drainage from Bay Drums,
Peak Oil, and Reeves Southeastern.
2-51
-------
Table 2-13. Summary of ER-L, ER-M, and Overall Apparent Effects Thresholds
Concentrations for Selected Chemicals hi Sediment (Dry Weight)
Chemical
Analyto
ER-L
Concentration
ER-H
Concentration
ER-L:ER-H
Ratio
Overall
Effects
Apparent
Threshold
Subjective Degree
of Confidence in
ER-L/ER-M Values
Trace Elementa (ppm)
Antimony 2 25
Atsanic 33 85
•Cndmium 5 g
Chromium BO 145
Coppor 70 390
Lond 35 110
Mercury 0.15 1.3
Hicfcol 30 50
Silver 1 2.2
Tin HA NA
Zinc 120 270
Polyclilorinatod Biphonyls (ppb)
Total FCBs 50 400
DDT and Metabolites (ppb)
DDT 1 7
ODD Z 20
DDE 2 15
Total DDT 3 350
Othor Fosbicidos (ppb)
Lindano HA HA
Chlordans 0.5 6
Hoptachlor NA HA
Dloldrln • . '0.02 8
Aldrin NA HA
Endrln 0.02 45
Mirax HA HA
Polynuclcar Aromatic Hydrocarbons (ppb)
12.5
2.6
1.8
1.8
5.6
3.1
8.7
1.7
2.2
NA
2.2
7.6
7
10
7.5
117
HA
12
NA
400
NA
2250
NA
25
50
5
NO
300
300
1
USD
1.7
NA
260
370
6
USD
NSD
NO
NA
2
NSD
NO
NSD
USD
NSD
Moderate/moderate
Low/moderate
High/high
Moderate/moderate
High/high
Moderate/high
Moderate/high
Moderate/moderate
Moderate/moderate
NA
High/high
Moderate/moderate
Low/low
Moderate/low
Low/low
Moderate/moderate
NA
Low/low
NA
Low/low
NA
Low/low
NA
Acenaphthene 150
Anthracene 85
Benzo(a)anthracene 230
Bonzo(a)pyrono 400
Benzo(a)pyrene HA
Biphenyl NA
Chryseno 400
Dibonz(a,h)anthracene 60
2, 6-dimethylnaphthylene HA
Fluoronthono 600
Pluorene 35
1-methylnaphthalene HA
2-methylnaphthalene 65
l-mothylphenanthrene NA
tlaptholono 340
Perylcne NA
Phananthrene 225
Pyrtme 350
2,3,5-trimothylnaphthalene HA
Total PAH 4000
650
960
1600
2500
NA
NA
2800
260
HA
3600
640
NA
670
NA
2100
NA
1380
2200
NA
35000
4.3
11.3
7
6.2
NA
NA
7
4.3
NA
6
18.3
HA
10.3
HA
6.2
NA
6.1
6.3
NA
8.8
150
300
550
700
HSD
HSD
900
• 100
NSD
1000
350
NSD
300
HSD
500
NSD
260
1000
NSD
22000
Low/ low
Low/moderate
Low/moderate
Moderate/moderate
NA
NA
Moderate/moderate
Moderate/moderate
NA
High/high
Low/ low
NA
Low/moderate
NA
Moderate/high
NA
Moderate/moderate
Moderate/moderate
HA
Low/low
ER-L " Effects Range-Low. That concentration equivalent to the lower 10 percentile of the screened
available data. Indicates the low end of the range of concentrations in which effects were observed
or predicted.
ER-H » Effects Range-Median. That concentration equivalent to the 50 percentile point in the screened
available data.
HSD " Hot Sufficient Data
HA " Hot Available
Modified from: Long, E.R. and L.G. Morgan. March 1990. The Potential for Biological Effects of
Sadiment-Sorbed Contaminants Tested in the National Stations and Trends Program.
NOAA Technical Memorandum NOS OMA 52. Office of Oceanography and Marine
Assessment. National Oceanographic and Atmospheric Administration. Seattle,
Washington.
2-52
-------
Table 2-14. Ratio of Analyte Concentration in Wetland Sediments to a Biologically Effective Concentration (U.S. EPA, 1990c)a
K>
.'IS STATION ID 01-CPU 01-SFU
&DRI STATION ID 01-RFU 02-RFU
AURI/FS STATION ID
SPRAY
POND FIELD
11/28/89 11/28/89
1115 1355
INORGANIC ELEMENTS
ANTIMONY
LEAD 1.22
MERCURY
ZINC
iXTRACTABLE ORGANIC COMPOUNDS
BENZO(A>-
ANTHRACENE -'-
DIBEMZO (A.H)'-
ANTHRACENE
ORGANOCKLORIME PESTICIDES ANALYSIS
GAMMA
CHLORDANE
4.4' -ODD
4.4'-DDE
PCB-1260
01-NOU 01-UNC 01-CLU
A B 5R
NORTH UNNAMED CENTRAL
UETLAND CREEK UETLAND
11/30/89 11/30/89 1/4/90
--
r.6o 2.02
?.33 1.46
2.95 93.3
—
~
194.00
16.00
60.00
1.62 - 5.20
02-CLU 03-CLU 04-CLW
678
CENTRAL CENTRAL CENTRAL
WETLAND WETLAND WETLAND
11/30/89 11/30/89 11/30/89
1.85
1.86
1.60
3.35
1.56
3.33
--
--
60.00
9.60
01 -SOU 02-SOW 03-SOW 04-SOU 02-CLW 01-UNC
01-SAP
0 E F G c
SOUTH SOUTH SOUTH SOW CENTRAL UNNAMED
WETLAND WETLAND WETLAND WETLAND WETLAND CREEK
11/30/89 11/28/89 11/28/89 11/28/89 11/30/89 1/9/90
..
1.09
1.16 \.n 1.95
--
-. '.'. '.'.
9.8
•This table compares sample concentration. Table 4.11, with the ER-L, Table 4.13, for each element or compound listed in both tables.
-------
Another unnamed wetland undergoing an environmental risk is a small wet area that serves,
along with the North Wetland, as the headwaters of the Unnamed Creek. This wet area receives
drainage from Reeves Galvanizing, and sediments were found to contain 93X the benchmark value
for zinc. Sediments here also contained very high concentrations of iron (22,600 ppm), aluminum
(3,790 ppm), and other metals that contribute to the potential environmental hazard posed by this
site. Because there are no benchmark values for these additional metals, they do not appear on the
overage tables.
One of four Central Wetland stations exceeded sediment benchmarks for 4,4'-DDE and
PCB-1260 by 60X and 10X, respectively. One of four South Wetland stations exceeded the
benchmark by 10X. No other wetland stations exceeded sediment benchmarks by more than 5X.
Risks Based on Biological Observations. Field observations on the number of taxa (species
diversity) of benthic macroinvertebrates were used as a method to evaluate risks associated with
existing conditions. A total of 53 kinds of benthic macroinvertebrates were collected from the five
wetlands. The taxa checklist reveals similar species richness (number of taxa) at all wetland sites
with the exception of a station in the South Wetland. At this station, the diversity of benthic
macroinvertebrates was significantly reduced to 11 taxa compared to a range of 19 to 23 taxa for
the other wetlands sampled. Elevated conductivity of the surface waters coincided with the
reduced numbers of taxa. The Spray Field Wetland and the Cypress Pond Wetland also shared a
similar conductivity regime, possibly indicating ground-water inflow. The elevated conductivity
values, however, did not appear to be a factor affecting the diversity of macroinvertebrates in the
other wetlands under study; the Cypress Pond and Spray Field Wetlands featured the most diverse
community of benthic macroinvertebrates sampled.
Mercury concentrations in tissues were typically lower than national mean values, but three
of four samples for fish and crayfish taken from the reference wetlands exceeded criteria proposed
for the protection of birds that may prey on them.
Bioassay results and macrobenthic community analysis supported the potential for risk
identified for the Unnamed Creek headwater area (table 2-15). High toxicity to several test species
was demonstrated and the area was essentially devoid of benthic life. The relationship among
toxicity, benthic community, and benchmark excesses was not as clear for the North Wetland
station.
To summarize the results in biota, waters of the North, Central, and South Wetlands
showed little toxicity to the organisms tested. The sediments of each wetland were at least
chronically toxic to Daphnia.
The water and sediments of the Unnamed Creek were moderately to highly toxic to almost
all organisms tested. Of all wetland sample locations tested, this is by far the one most needing
further attention as evidenced by toxicity test results, benthic community structure, bioaccumulative
contaminants, and water and sediment chemistry. Use of multiple measurement endpoints provides
a basis for assessing the efficacy of future remediation efforts at these sites.
2-54
-------
Table 2-15. Ratio of Areawide Hydrologic Study Surface Water Constituents to EPA Ambient
Water Quality Criteria and Screening Values (U.S. EPA, 1990c)
Sample
Designation
4W, Drainage to
N Wetland5
4W (R) Drainage
to N Wetland
9W (01-UNC),
Creek H of
SE Gal5
10W, Creek N-
of SE GalB
Analyte1
Aluminum
Iron
Lead (H)
Zinc (H)
Hardness Adjusted
CMC
750
125.30
155.63
Methylene chloride 19300"1'
Aluminum
Chromium III
Chromium VI
Chromium (T)
Iron
Lead (H)
Zinc (H)
Aluminum (H)
Copper (H)
Iron
Lead (H)
Zinc (H)
750
(H) 3410.57
16
-
-
233.12
235.26
750
24.66
—
127.58
157.50
AHQC or SV+ (ug/L)
CCC
87
1000
4.88
140.96
1930+
'87
406.52
11
-
1000
9.08
213.08
87
15.95
1000
4.97
142.66
HC2 .
257
5630
5.1
37.9
2.0
5500
-
16
25000
15
11000
372
Q.7B
•JOBQ
5. OH
1410
Ratio3
MC/CHC MC/CCC
3.0
5.6
1.0
7.3 63.2
1.4
25
1.7
46.8 51.6
4.3
3.1
1.0
9.0 9.9
Includes analytes found both in AWQC listings (Table 4.28) and in surface water samples (Table
4.26).
Total chromium measured in samples; criteria calculated for trivalent and hexavalent chromium.
Calculated ratios for chromium hold only if all measured chromium is present in that form.
n
Measured concentration (Tables 4.26).
3 Ratios calculated only for those measured concentrations .that exceed CMC or CCC.
Water hardness - 140 mg/L as CaCOj.
Water hardness = 228 mg/L as CaCOj.
6 Water hardness = 142 mg/L as CaC03.
CMC = Criterion Maximum Concentration. Tl a 1-hour average concentration not be exceeded more than
once every 3 years on the average.
CCC - Criterion Continuous Concentration. The 4-day average concentration not be exceeded more
than once every 3 years on the average.
(H) = CMC and CCC have been adjusted for water hardness.
(T) - Total
(R) ~ Resampled for volatiles. - . .
+ = Screening value.
B " Analyte found in associated blank as well as in sample.
W = The posot-digestion spike for furnace AA analysis is outside of the 85 to 115X control limits,
while sample absorbance is less than SOX of the spike absorbance.
2-55
-------
The sediments of the Cypress Pond Wetland were highly toxic to fish, daphnids, algae, and
bacteria, and the source of this toxicity should be further explored. The contamination of reference
sites is highly likely in industrial areas, indicating the need for a suite of measurement endpoints.
The apparent toxicity of the sediment does not appear to have impaired wetland functions-
balanced communities of plants and animals remain.
Comments on Risk Characterization
•Evaluating risks by ratio to benchmarks (AWQCfor water, NOAA Effects Range-Low for
sediments) is valid, but limited: it does not account for possible synergistic effects of exposure
to multiple contaminants and does not consider site-specific conditions that could affect the
bioavailability and toxicity of the chemicals.
•The case study presents information on risks but these are not brought together to provide
an overview of risk. The information is not clearly related to the assessment of wetland
integrity. If the data on contaminant concentration, toxicity assays, and bioaccumulation
•were put into a spatial context, that could be used to assess risk and to determine how much
of the resource was at risk as a result of contamination.
•A useful addition to the risk characterization would be a summary of the measurement
endpoints and statements about what these mean in terms of assessment endpoints.
•This case study offers a number of valuable lessons on the application of risk assessment.
One is the necessity for multiple measurement endpoints. In situations such as this, where
one is dealing with multiple and in many cases unknown stressors, it is essential to have more
than one measurement endpoint. For example, had the sole endpoint in this assessment been
a comparison of affected wetlands to the chosen reference wetland, this analysis would have
failed because the reference was found to be contaminated. A suite of metrics also provides
the assessor with one measure of uncertainty in the assessment, i.e., a clearer picture of the
weight of evidence behind a particular conclusion.
9 A second lesson in this study is the clear need for: (a) care in selecting reference sites, (b)
caution in interpreting the data from those sites, and (c) the inappropriateness of relying on
reference sites as the sole standard for assessing risk. It is noted that for Superfimd sites,
good reference sites may be hard to identify because there are frequently other industries in
the area. Given the natural range of variability in ecological systems, wisdom (or statistics)
would dictate using more than one reference site.
•A third lesson from this study is the need for: (a) better benchmarks for freshwater
sediments and soils; (b) test organisms for detecting direct sediment toxicity; (c) rapid
bioassessment metrics for wetland species; and (d) models useful in assessing risk at
Superfimd sites. In developing metrics, consideration should be given to using measures other
than (or in addition to) species abundance or presence/absence data; for example, species-
specific biomass information may be a more sensitive metric.
2-56
-------
Comments on Risk Characterization (continued)
^Several questions that should be addressed in an ideal risk assessment at a Superfimd site
include: (1) What are the ecological resources in the area of the site? (2) What are the risks
associated with the range of human impacts in this area (e.g., habitat alteration, chemical
contaminants, hydrologic change)? (3) What are the risks caused by the site to adjacent
ecological resources? What is the spatial extent of these risks?
2-57
-------
2.4. REFERENCES
Adamus, P.R.; Clairain, E.J., Jr.; Smith, R.D.; Young, R.E. (1987) Wetland evaluation technique
(WET); Volume II: methodology. Operational Draft Technical Report Y-87-10. U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Canonie Environmental. (1988a) Work plan, areawide hydrological RI/FS, Peak Oil, Reeves,
Southeastern, and Bay Drums Superfund sites, Tampa, FL. March 1988.
Canonie Environmental. (1988b) Site source characterization, Peak Oil Superfund site, Tampa, FL.
November 30, 1988.
Canonie Environmental. (1990) Draft summary results. Bay Drums, Peak Oil, and Reeves
Southeastern sites areawide RI/FS.
Canonie Environmental. (1991) Phase I areawide hydrologic remedial investigation of Bay
Drums, Peak Oil, and Reeves Superfund sites. Draft report. January 1991.
Cowardin, L.M.; Carter, V.; Golet, F.C.; LaRoe, E.T. (1979) Classification of wetlands and
deepwater habitats of the United States. U.S. Fish and Wildlife Service, Washington, DC.
FWS/OBS-79/31.
Long, E.R.; Morgan, L.G. (1990) The potential for biological effects of sediment-sorbed
contaminants tested in the National Status and Trends Program. National Oceanic and
Atmospheric Administration Technical Memorandum NOS OMA 52. Office of Oceanography
and Marine Assessment. National Oceanic and Atmospheric Administration, Seattle, WA.
March 1990.
Pace Laboratories, Inc. (1988) Reeves Southeastern Corporation, site source characterization RI/FS,
work plan. December 1, 1988.
U.S. Environmental Protection Agency, Environmental Photographic Interpretation Center. (1985)
Site analysis. Reeves Southeastern Corporation; Peak Oil; Bay Drum. Brandon, FL. (Two
volumes.)
U.S. Environmental Protection Agency, Region 4. (1989) Bay Drums site remedial investigation
project operations plan. June 19, 1989.
U.S. Environmental Protection Agency. (1990a) First draft working document, Bay Drums Remedial
Investigation. Bay Drum site. Tampa, FL. BSD Project No. 88E-212.
U.S. Environmental Protection Agency, Region 4. (1990b) Environmental Services Division,
Environmental Compliance Branch. Bay Drums Remedial Investigation first draft working
document. June 25, 1990.
2-58
-------
U.S. Environmental Protection Agency, Region 4. (1990c) Bay Drums, Peak Oil, and Reeves
Southeastern areawide wetland impact study. Environmental Services Division. Draft report.
Athens, GA.
U.S. Senate Committee on Environment and Public Works. (1987) The Comprehensive
Environmental Response, Compensation, and Liability Act of 1980 (CERCLA: Superfund)
(P.L. 96-510) as amended by the Superfund Amendments and Reauthorization Act of 1986
(P.L. 99-499). U.S. Government Printing Office, Washington, DC.
2-59
-------
-------
SECTION THREE
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
SPECIAL REVIEW OF THE GRANULAR FORMULATIONS OF CARBOFURAN
BASED ON ADVERSE EFFECTS ON BIRDS
-------
AUTHOR AND REVIEWERS
AUTHOR
Clyde R. Houseknecht
Office of Water
U.S. Environmental Protection Agency
Washington, DC
REVIEWERS
Christian E. Grue (Lead Reviewer)
Washington Cooperative Fish
and Wildlife Research Unit
U.S. Fish and Wildlife Service
Seattle, WA
Joel Brown
Department of Biological Sciences
University of Illinois
Chicago, IL
Arthur L. Buikema
Biology Department
Virginia Polytechnic Institute and State
University
Blacksburg, VA
Carolyn L. Fordham
Terra Technologies
Golden, CO
Clifford Hupp
Water Resources Division
U.S. Geological Survey
Reston, VA
Ronald J. Kendall
Institute of Wildlife and
Environmental Toxicology
Clemson University
Clemson, SC
3-2
-------
CONTENTS
ABSTRACT 3-6
3.1. RISK ASSESSMENT APPROACH 3-7
3.2. STATUTORY AND REGULATORY BACKGROUND 3-7
3.3. CASE STUDY DESCRIPTION 3-9
3.3.1. Problem Formulation 3-9
3.3.2. Analysis: Characterization of Ecological Effects • • • 3-13
3.3.3. Analysis: Characterization of Exposure . ". 3-17
3.3.4. Risk Characterization 3-18
3.4. REFERENCES . .3-24
3-3
-------
LIST OF FIGURES
Figure 3-1. Structure of analysis for carbofuran effects on birds
3-8
LIST OF TABLES
Table 3-1. Representative Bird Species Likely to Be Exposed to Carbofuran 3-10
Table 3-2. Summary of Bird Kills Due to Poisoning by Direct Consumption of
Carbofuran Granules (1973-1987) 3-11
Table 3-3. Summary of Bird Kills Due to Secondary Poisoning From Carbofuran
Granules (1983-1986) 3-12
Table 3-4. Acute Oral Toxicity (LD50) Values and Associated 95 Percent
Confidence Intervals (CI) of Carbofuran to Birds 3-14
Table 3-5. Summary of Field Studies for Granular Carbofuran 3-16
Table 3-6. Number of Exposed Carbofuran Granules After Band and
In-Furrow Application 3-19
Table 3-7. Minimum and Maximum Values for Avian LD50s/Ft2 for
10 Crops 3-21
Table 3-8. Annual Estimated Acreage Treated With Granular Carbofuran 3-22
3-4
-------
LIST OF ACRONYMS
AI active ingredient
EPA U.S. Environmental Protection Agency
FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
LC50 lethal concentration to 50 percent of organisms tested
LD50 lethal dose to 50 percent of organisms tested
OPP Office of Pesticide Programs
3-5
-------
ABSTRACT
The Federal Insecticide, Fungicide, and Rodenticide Act authorizes the cancellation of
registration of a pesticide that produces an unreasonable risk to humans or the environment. The
U.S. Environmental Protection Agency's (EPA's) Office of Pesticide Programs initiated a special
review of the granular formulations of the broad-spectrum insecticide/nematicide carbofuran in
light of the possible risks it may pose to birds. The special review process utilized data from
laboratory toxicity studies, field studies, and reported incidents of bird kills to assess the potential
for adverse impacts to avian species. Based on this information, EPA proposed to cancel
registration for the use of granular carbofuran, concluding that granular carbofuran generally poses
unreasonable risks to birds through direct and secondary poisoning.
3-6
-------
3.1. RISK ASSESSMENT APPROACH
An overview of the risk assessment approach is shown in figure 3-1. The risk
characterization method used in this case study is a combination of the quotient and the weight of
evidence methods (U.S. EPA, 1990). This study describes the ecological risk analysis that formed
the basis for the issuance of the carbofuran Position Document 2/3 (U.S. EPA, 1989). It follows
the standard evaluation procedures used by the Office of Pesticide Programs (OPP) to determine
the risks of pesticides to nontarget species (Urban and Cook, 1986).
A major strength of this case study is the quantity of data brought to bear on the risk
characterization. Laboratory studies demonstrated that granular carbofuran is acutely toxic to
birds, and the presumption of risk was confirmed by field studies and numerous reports of bird
kills. The analysis did not deal with information about effects at the population or ecosystem
levels, nor did it evaluate possible impacts on other organisms, such as small mammals.
The use of this case study alone as a model for the evaluation of pesticides in general may
be limited, since other pesticides may differ from carbofuran in their mode of action, toxicity,
environmental persistence, and the relative importance of direct and indirect effects. Critical to
future pesticide studies is an assessment of contamination within habitat matrices, routes of
exposure, and bioavailability.
3.2. STATUTORY AND REGULATORY BACKGROUND
A pesticide product may be sold or distributed in the United States only if it is registered or
exempt from registration under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA),
as amended (7 U.S.C. §132 et seq.). Before a product can be registered unconditionally, it must
be shown that it can be used without "unreasonable adverse effects on the environment" (FIFRA
section 3[c][5]); that is, without causing "any unreasonable risk to man or the environment, taking
into account the economic, social, and environmental costs, and benefits of the use of the
pesticide" (FIFRA section 2[bb]). The burden of proving that a pesticide meets this standard for
registration is at all times on the proponent of initial or continued registration. If at any time the
Agency determines that a pesticide no longer meets this standard for registration, the Administrator
may initiate proceedings to cancel or suspend the registration under FIFRA section 6.
The special review process provides a mechanism through which the Agency gathers
information about pesticides that appear to pose risks of adverse effects to human health or the
environment. Evidence of risk submitted to and/or gathered by the Agency must be evaluated and
considered in light of benefit information. If the Agency determines that the risks appear to
outweigh the benefits, the Agency can initiate action under FIFRA to cancel, suspend, and/or
require modification of the terms and conditions of registration.
In 1985, the Agency determined that pesticide products containing granular carbofuran
exceeded the existing risk criterion for avian toxicity set forth in 40 CFR 162.11(a)(3)(i)(B) and in
40 CFR 162.11(a)(3)(ii)(c). The Agency also determined that granular carbofuran met or
exceeded the proposed risk criterion (50 Federal Register 12195; March 27, 1985). The proposed
risk criteria are how in effect as set forth in 40 CFR 154.7.
3-7
-------
Figure 3-1. Structure of Analysis for
Carbofuran Effects on Birds
PROBLEM FORMULATION
Stressors: the granular formulation of the insecticide
and nematicide carbofuran.
Ecoloaical Components: birds (kills involving at least
A
30 species have been reported).
Endpoints: assessment endpoint is survival of birds
that forage in agricultural areas. This was
evaluated using laboratory and field measurement
endpoints where lethality was documented.
Ii
1
|
Characterization of Characterization of
Exposure Ecological Effects
Estimates were developed babor,at,?ry tO)"'cological
^~^,aoe »^^,d
soils for various b'rd,s were °btamed: ,F'eld
applications studies on bird mortalities
applications. following applications
were performed.
1 _l_
4r *
RISK CHARACTERIZATION
Risk was evaluated using a form of the Quotient Method
in which estimated exposure (granules/sq. ft.) was divided
by the number of granules associated with the LDso value.
(The larger the number, the greater the acute risk.) An
estimate was made of potential bird mortality for acreage
treated with carbofuran. Field data were used to verify
predictions.
3-8
-------
3.3. CASE STUDY DESCRIPTION
3.3.1. Problem Formulation
Ecological Components. This assessment focuses on birds. Birds,may be exposed by
ingesting carbofuran granules as they forage for seeds or grit on or below the surface of the soil.
They also may be exposed by feeding on birds or other animals that contain carbofuran granules or
residues and that are moribund or have already died from carbofuran poisoning. Residues of
carbofuran in birds confirm that exposure to carbofuran has occurred.
OPP evaluated whether or not birds would be present during and immediately after the
application of carbofuran. This analysis was conducted for 10 of the crops for which use is
registered (corn, sorghum, soybeans, rice, peanuts, tobacco, cotton, cranberries, pineseed orchards
and seedlings, and sunflowers). These 10 crops account for 95 percent of the annual application of
granular carbofuran. Representative bird species that were expected to be present in these crops
are summarized in table 3-1. Many of these species were among those found killed by direct or
secondary carbofuran poisoning.
Stressors. Carbofuran is the common name for 2,3-dihydro-2,2-dimethyl-7-benzofuranyl
methylcarbamate. It is a broad-spectrum insecticide and nematicide registered for control of pests
on 27 agricultural crops and for certain forest and pineseed orchard uses. Approximately 7 to 10
million Ib of active ingredient (AI) of all carbofuran formulations are applied annually, with
approximately 6 to 9 million Ib AI of the annual usage accounted for by the granular formulations.
FURADAN is the only trade name currently used.
Carbofuran is an acute toxicant that inhibits cholinesterase and results in stimulation 'of the
central, parasympathetic, and somatic motor systems. It is generally applied as a prophylactic
treatment when seeds are planted at the beginning of the growing season.
OPP received documentation for more than 40 incidents of carbofuran-related bird kills
involving nearly 30 species of birds (tables 3-2 and 3-3). The most commonly affected birds were
waterfowl. However, Lapland longspurs, robins, several species of sparrows and other songbirds,
marsh and shore birds, and others also were affected. The number of birds involved in any single
incident ranged from 1 to more than 2,000 individuals. With two possible exceptions noted in
table 3-2, these kills were attributed to use of the chemical according to label instructions.
Many of these kills resulted from secondary poisoning of avian predators and scavengers
(table 3-3). The ability of granular carbofuran to cause secondary poisoning is an important factor
in the cumulative avian risk. This is especially true for raptors, because they occupy an important
niche in the food chain by controlling vertebrate populations. Because of the problems inherent in
studying the effects of carbofuran on raptor populations (i.e., to effectively monitor secondary
poisoning), OPP believes that these incidents constitute only a few of the secondary poisoning
deaths caused by granular carbofuran.
Incidents of avian mortality attributable to carbofuran poisoning have occurred throughout
the year (table 3-2), although most have occurred during the usual planting season of April through
3-9
-------
Table 3-1. Representative Bird Species Likely to Be Exposed to Carbofuran (U.S. EPA,
1989)
Wading Birds
Great Blue Heron
Snowy Egret
Little Blue Heron
Cattle Egret
White Ibis
Glossy Ibis
Waterfowl
Fulvous Whistling Duck
Brant
Canada Goose*
-Aleutian Canada Goose
Wood Duck"
Green-Winged Teal*
American Black Duck
Mottled Duck
Mallard1
Northern Pintail*
Blue-Winged Teal8
Cinnamon Teal*
Northern Shoveler
Gadwall*
American Wigeon*
Canvasback
Ring-Necked Duck
Raptors
Black Vulture
Turkey Vulture
Mississippi Kite
Bald Eagle8
Northern Harrier*
Sharp-Shinned Hawk
Cooper's Hawk
Harris' Hawk
Red-Shouldered Hawk*
Broad-Winged Hawk
Swainson's Hawk
Red-Tailed Hawk*
Rough-Legged Hawk
Golden Eagle
American Kestrel
Apolomado Falcon
Peregrine Falcon
Rails and Allies
Black Rail
Sora
Purple Gallinule
American Coot?
Sandhill Crane
-Mississippi
Sandhill Crane
Whooping Crane
Shore Birds
Semipalmated Plover
Piping Plover
Killdeer*
Spotted Sandpiper
Semipalmated Sandpiper
Pectoral Sandpiper
Stilt Sandpiper
Laughing Gull
Game Birds
Ring-Necked Pheasant8
Greater Prairie-Chicken
-Attwater's Greater
Prairie-Chicken
Sharp-Tailed Grouse
Ruffed Grouse
Northern Bobwhite
California Quail
Mountain Quail
Wild Turkey
White-Winged Dove
Mourning Dove8
Common Ground Dove
Common Snipe
American Woodcock
Owls
Common Barn Owl
Eastern Screech Owl
Great Horned Owl
Barred Owl
Long-Eared Owl
Great Gray Owl
Short-Eared Owl8
Northern Saw-Whet Owl
Woodpeckers
Red-Headed Woodpecker
Red-Bellied Woodpecker
Red-Cockaded Woodpecker
Northern Flicker
Eastern Kingbird
Horned Lark
Blue Jay8
Scrub Jay
American Crow
Chihuahuan Raven
Common Raven
Tufted Titmouse
White-Breasted Nuthatch
Eastern Bluebird
American Robin8
Brown Thrasher
Northern Mockingbird
Loggerhead Shrike
Northern Shrike
European Starling
Northern Cardinal
Pyrrhuloxia
Rose-Breasted Grosbeak
Blue Grosbeak
Indigo Bunting
Painted Bunting
Rufous-Sided Towhee
Field Sparrow
Vesper Sparrow
Lark Sparrow
Lark Bunting
Savannah Sparrow8
Grasshopper Sparrow
Henslow's Sparrow
Seaside Sparrow
Song Sparrow
Lincoln's Sparrow i
White-Throated Sparrow
Lapland Longspur3
Bobolink
Red-Winged Blackbird
Tricolored Blackbird
Eastern Meadowlark
Western Meadowlark
Yellow-Headed Blackbird
Rusty Blackbird
Brewer's Blackbird
Boat-Tailed Crackle8
Common Crackle"
Brown-Headed Cowbird
American Goldfinch
'Confirmed kill from carbofuran ingestion.
3-10
-------
Table 3-2. Summary of Bird Kills Due to Poisoning by Direct Consumption of Carbofuran
Granules (1973-1987) (U.S. EPA, 1989)
Site Formulation
Corn 15G
10G
Corn/Soybeans 15G
Potatoes 15G
10G
Rapeseeda 10CRb
Rice 5G
Turnips* 10G
Occurrence
August 1983
February 1984
June 1986
1987
1972
1973
May 1979
May & June
1983
June 1986
November-
December 1974
May 1984
April-
June 1984
April 1985
October 1985
April 1986
October 1986
April 1987
1975
1977
December 1973
June 1986
Location
Maryland
Illinois
Indiana
New York
Wisconsin
Wisconsin
New York
New York
New York
Canada
Canada
California
California
California
California
California
California
Canada
Canada
Canada
Canada
Number
of Birds
200
not known
12
3
11
3
10
25
20
80
> 2,000
39
95
100-135°
36
150C
4
1,100
50
50-60
500-1,000
"Carbofiiranis not registered for use on this crop in the United States.
°10 percent corncob granule.
These kills may have resulted from a misuse or from carbofuran applied earlier in the year.
3-11
-------
Table 3-3. Summary of Bird Kills Due to Secondary Poisoning From Carbofuran Granules (1983-1986) (U.S. EPA, 1989)
Location Pate Site
Utah 1983 Corn
Maryland 1983 Corn
Iowa and 1984 Corn
Illinois
Virginia 1985 Corn
Raven
Northern
Harrier
Red-
Shouldered
Hawk
Description
-Two ravens contained residues up to 8.1 ppm and 38 granules
-Another exhibited signs of poisoning but did not die
-Contained up to 21.8 ppm carbofuran and 23 granules
-Another exhibited signs of poisoning but did not die
-Female found intoxicated after feeding on small mammals and birds; bird was sacrificed;
gut contained 47 ug and gastrointestinal tissue 49.6 ug carbofuran
Bald Eagle —One adult male dead at base of active nest with 59 % brain acetyicholinesterase inhibition;
gastrointestinal tract contained 0.64 ppm carbofuran
—One dead eaglet in nest along with pigeon and grackle remains
—Cornfields nearby treated with carbofuran; dead and dying pigeons and other birds found
nearby
Virginia 1985 Com Bald Eagle —Brain exhibited 60% acetyicholinesterase inhibition; esophagus contained 82 ppm, stomach
0.067 ppm, and duodenum 1.1 ppm carbofuran
Virginia 1986 Com
California 1986 Rice
California 1986 Rice
California 1985 Rice
Red-Tailed
Hawk
Red-Tailed
Hawk
Northern
Harrier
Northern
Harrier
New York 1985 Unknown Red-Tailed
Hawk
—one found dead; contained 0.107 ppm carbofuran
—two found dead during larger waterfowl kill incident
—Found in incident above
-64 ppm carbofuran recovered from crop contents, including flies, duck viscera, and maggots;
84% brain cholinesterase inhibition
-Adult female found dead near nest with 0.06 ppm carbofuran in liver and remains of a starling
and voles in stomach
-------
June. In reviewing the information, OPP concluded that label-directed application of carbofuran
presents a hazard to birds throughout the United States.
Endpoint Selection. OPP examined the use of granular carbofuran because of numerous
kills related to its use. The assessment endpoint of concern was survival of birds that forage in
agricultural fields and of birds that may prey on or scavenge other organisms that have been
exposed. These were evaluated using laboratory measurements of acute toxicity as well as field
studies.
Comments on Problem Formulation
General comment:
• The present case study represents a unique situation and, as a model for future studies,
may be most applicable for other granular formulations. Granular carbofuran is acutely
toxic to birds, and poisoned animals either die or recover quickly. Environmental
persistence is relatively short. The endpoint (mortality) is relatively easy to quantify
because of the high acute toxicity of the chemical, and the habitat most frequently treated is
open, freshly plowed fields. Furthermore, indirect effects (e.g., pesticide-induced
reductions in food resources or habitat) were not important endpoints. Other pesticides will
vary in mode of action, toxicity, environmental persistence, and the relative importance of
direct and indirect effects. For these reasons, the use of this case study alone as a model
for evaluations of other pesticides may be limited.
3.3.2. Analysis: Characterization of Ecological Effects
To evaluate the acute toxicity of carbofuran to bird species, OPP reviewed several acute
toxicity studies using the technical grade of the active ingredient (Tucker and Crabtree, 1970;
Schafer et al., 1973; Schafer and Brunton, 1979; Hudson et al., 1984; Hill and Camaradese,
1984). These studies measured the single dose that kills 50 percent of the test organisms (LD50).
The results of these studies (table 3-4) indicate that carbofuran is very highly toxic to a variety of
avian species but that the degree of sensitivity of bird species varies.
In addition, OPP evaluated a study (Balcomb et al., 1984a) that measured the toxicity of
granules of FURADAN 10G. The study demonstrated that the consumption of a single granule of
the insecticide can be fatal to a small bird.
Data for the subacute dietary concentration that kills 50 percent of the test organisms (LC50)
were also evaluated (Shellenberger and Gough, 1972; Hill et al., 1975; Fink, 1974, 1976). These
studies indicated that the 5-day subacute LC50s range from 21 to 681 parts per million (ppm) for
various species of birds. These LC50 values confirm that carbofuran is highly toxic to birds via the
diet.
3-13
-------
Table 3-4. Acute Oral Toxfcity (LD50) Values and Associated 95 Percent Confidence
Intervals (CI) of Carbofuran to Birds (U.S. EPA, 1989)
Species
Fulvous Whistling Duck
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Northern Bobwhite
Northern Bobwhite
Japanese Quail
Japanese Quail
Ring-Necked Pheasant
Rock Dove
Red-Winged Blackbird
Brown-Headed Cowbird
Common Crackle
Starling
Quelea
House Sparrow
Age
3-4 mth
36 h
7d
30 d
6mth
3-4 rath
12mth
12mth
3mth
12mth
0.5 mth
0.5 mth
NR
NR
NR
NR
NR
NR
NR
NR
LD50
(mg Al/kg of
Sex body weight)
F
NR
NR
NR
NR
F
F
M
F
M/F
F
M
NR
NR
M
NR
NR
NR
NR
NR
0.24
0.37
0.63
0.51
0.41
0.40
0.51
0.48
5.04
12
1.7
1.9
4.15
1.33
0.42
1.33
1.33
5.62
0.42
1.33
CI
0.20-0.28
0.28-0.48
0.53-0.74
0.41-0.64
0.33-0.52
0.32-0.50
0.41-0.64
0.38-0.60
3.61-6.99
7-19
1.3-1.9
1.7-2.1
2.38-7.22
NR
NR
NR
NR
3.16-10.0
NR
NR
NR = not reported.
3-14
-------
Six field studies were evaluated by OPP, and the results are summarized in table 3-5.
These studies investigated the effects of carbofuran exposure resulting from label-directed,
soil-incorporated uses of FURADAN* 10G and 15G applied as band and in-furrow applications and
FURADAN 10G applied with specialized equipment. The conditions surrounding these field
studies and the hypotheses being tested varied, resulting in a lack of standardization among the
studies. The studies evaluating secondary mortality were even less standardized than those
evaluating direct mortality.
The field studies reported that birds were killed by direct poisoning with carbofuran. When
possible, bird mortality per acre was estimated for each study. These estimates were based on
(1) mortalities confirmed by residue analysis, (2) carcasses in which carbofuran was strongly
implicated in the cause of death (granules in the digestive system, cholinesterase depression, no
evidence of trauma), or (3) cases of probable sublethal poisoning in which a bird was judged
unable to recover prior to being subjected to other sources of mortality such as predation.
The mortalities reported in these studies are probably underestimates because of problems
associated with carcass search efficiency and removal of carcasses by predators and scavengers.
Birds may not have been found because predators carried carcasses away from the site.
Additionally, birds may have sought cover when dying and may not have been noticed.
Although the field studies were performed at different application rates with different
methods of application, all resulted in granules being left unincorporated and available to birds.
Also, each field study resulted in avian mortality and each had one or more factors (e.g., lack of
sufficient area searched, lack of sufficiently trained personnel, failure to assess carcass removal by
predators) indicating that the number of dead birds encountered in carcass searches was an
underestimate of the actual impact of granular carbofuran on birds. In interpreting such studies,
important considerations such as the habitat and ecology of the species exposed also must be
considered.
As previously mentioned, EPA also has information on a number of carbofuran-related bird
kills confirming that granular carbofuran is acutely toxic to birds. In assessing these bird kills,
OPP considered only those incidents in which carbofuran was clinically diagnosed as the cause of
mortality or was strongly implicated.
OPP has concluded that the number of reported bird kills underestimates the number of
birds that may actually be killed from exposure to granular carbofuran. Several factors support
this conclusion. First, no systematic or reliable mechanism exists for accurate monitoring and
reporting of wildlife kills. OPP relies heavily on incident monitoring by states, and state efforts
tend to be highly variable. Only a few states have trained and equipped personnel to respond to
kill reports and to conduct the thorough investigation necessary to determine the pesticide and
application rate used and whether label directions were followed. In addition, few states regularly
report bird kills to EPA.
Second, even if dead birds are found, the observer may not attribute the deaths to an
insecticide application. Field evidence indicates that carbofuran may cause bird mortality at least
3-15
-------
Table 3-5. Summary of Field Studies for Granular Carbofuran (U.S. EPA, 1989)
Study Use and Site Formulation Method
1 Corn 10G&15Gb Band
2 Corn 10G In-furrow
3 Corn 10G In-furrow
4C Kneseed 10G POWR-TIL
Florida
South
Carolina
Mississippi
Louisiana
•I* 5 Corn 10G&15Gb Band
ON
Illinois
Iowa
6 Corn 10G&15Gb In-furrow
Texas
7 Rice 3G Broadcast
RateffiAJ/
acre)
4
1
1
28.3
25.2
22.3
25.2
1
1
0.5
Acres
Searched
254
92
34
30
30
30
-27
171
307
214
NRd
Mortalities
877
10
23
12
26
19
39
92
32
58
5
Mortalities/
acre
3.6
0.1
0.7
0.4
0.9
0.6
1.4
0.5
0.1
0.3
—
Estimated LD^/ft2
in Treated Areas*
S G W
1,179
62
62
—
59
93
94
401
401
62
234
29
1
1
—
1
1
2
9
9
1
5
61
3
3
—
3
2
5
20
20
3
12
aS = songbirds, G = upland game birds, and W = waterfowl. Estimates are exclusive of turn-rows. Estimated LD50s/ft2 combine toxicity data from table 3-4 (LD^s) and
exposure data from table 3-6 (granules exposed/ft2). LD^s (mg Al/kg) were converted to LD^s (number of granules) assuming 0.6 mg Al/granule and average bird body
weight obtained from the literature.
bBecause the study plots were close together, birds may have moved between plots. Researchers determined that carbofuran poisoning was the cause of death, but they could
not determine whether the 10G or 15G was responsible. As the data for the 10G and 15G plots are not independent, EPA could not calculate separate mortality rates.
Reported rates of incorporation efficiency were, in order, 100%, 99.5%, and 99.2%.
dNR = not reported. Therefore, mortality/acre could not be calculated.
-------
60 days after it was first applied. Thus, a farmer or other observer not familiar with the site
history may not attribute the death to carbofuran application. If a person does suspect that a bird
may have been poisoned, the individual may not know to whom to report or may believe that there
is some liability associated with reporting. Finally, problems associated with the reporting of bird
kills are greater for small, more inconspicuous songbirds. Many small birds do not form large
flocks, and small carcasses disappear more quickly than large ones. As a result, small dead birds
are less likely to be noticed than large dead birds such as waterfowl.
Comments on Analysis: Characterization of Ecological Effects
Strengths of this case study include:
•There is a considerable amount of field and laboratory data on the toxicity of the
chemical.
Limitations include:
•This study does not include field data for mammals; such information would be helpful
for future studies.
•By today's standards, the design and conduct of the field studies used to support this
case study are inadequate. In the majority of these studies, there are no true control
plots or determinations of carcass-search efficiency. However, in this case, the evidence
for avion mortality following use of granular carbofuran is so great that deficiencies in
the field studies do not affect the conclusions.
3.3.3. Analysis: Characterization of Exposure
Birds may be exposed to carbofuran granules that are on the surface of the soil as well as
below it. Balcomb (1980) has shown that the size of carbofuran granules overlaps that of grit and
seeds consumed by birds. In addition, dead birds have been found with soil caked on their bills in
fields treated with granular carbofuran. This suggests that they had been probing for food or grit
shortly before death (I. Sunzenauer, personal communication).
Granules may be applied aerially or with ground equipment at the beginning of the growing
season when corn or other crops are planted. Granules may be left on the soil surface following
the use of band application methods or from incomplete incorporation following in-furrow
application. 'Granules also may be left on the soil surface when machinery is being loaded, when
planter shoes are lifted out of furrows to permit turning, and when planter shoes rise out of the
soils of irregularly contoured fields.
Several investigators have confirmed that both band and in-furrow application of carbofuran
or other granular pesticides using conventional commercial application equipment result in exposed
3-17
-------
granules on the soil surface (Beskid and Fink, 1981; Hummel, 1983; Dingledine, 1985). Erbach
and Tollefson (1983) reported that 5.8 to 40.2 percent of granules remained unincorporated after
band application. Hummel (1983) showed that in-furrow application results in approximately
99 percent incorporation.
Based on these reports, OPP conservatively estimated the number of granules that would
remain exposed on the soil surface as a result of incomplete incorporation at 15 percent of the
granules applied by band and 1 percent applied by in-furrow application. These percentages were
used to estimate the number of exposed granules. The results are given in table 3-6.
DeWitt (1966) analyzed available avian laboratory and field studies and reported a possible
relationship between the quantity of pesticide ingested by birds and the quantity of pesticide
deposited per unit area. Given the finding by Balcomb et al. (1984a) that the consumption of a
single granule can cause death hi a small bird, such large quantities of exposed granules represent a
significant risk to avian species.
Birds may be seen feeding in fields during spring planting operations, often following
behind planting equipment. Some birds are probably attracted to soil invertebrates, seeds, or old
crop remains, that may be brought to the surface by the planter. Birds foraging for seeds or grit
may be unable to avoid ingesting granular pesticides.
Granules also may be applied aerially later in the season. These granules are not
incorporated. Birds are exposed to granules falling on the field surfaces and into the leaf whorls of
plants. Moreover, aerial application may contaminate more of the field edge than ground
application at the beginning of the season because of the inaccuracy of the aerial placement of
granules.
In addition, birds may ingest carbofuran that has been applied directly to water or to fields
that are subsequently flooded. Numerous duck kills have resulted from the application of
carbofuran granules to rice fields in California.
Poisoning of avian predators and scavengers may occur from ingesting food items that have
been exposed to carbofuran. For example, Balcomb et al. (1984b) reported whole-body carbofuran
residues ranging from 0.3 to 670 ppm in 11 of 12 samples of earthworms. These worms were
found hi furrows after carbofuran application to corn. Worms were washed prior to chemical
analyses; thus, whole-body residues did not include additional carbofuran in granules attached to
the body. Secondary poisoning incidents have involved bald eagles, red-tailed hawks, northern
harriers, and other birds of prey. These species are attracted to dead and dying birds and small
mammals affected by granular carbofuran.
3.3.4. Risk Characterization
Risks were evaluated, hi part, by comparing estimated exposure levels of granules in
surface soils with that calculated as the amount needed to kill 50 percent of the organisms (LD50).
DeWitt (1966) proposed the square foot as the unit area for determining risks to birds from
3-18
-------
Table 3-6. Number of Exposed Carbofuran Granules After Band and In-Furrow Application
(U.S. EPA, 1989)
Formulation Application Rate Application Granules Exposed/ft2
(IbAI/a) Method
15G 3 7" band 596
3 In-farrow 93
1 T band 198
1 In-farrow 31
10G 3 7" band 837
3 In-farrow 130
1 7" band 279
1 In-farrow 43
3-19
-------
exposure to granular pesticides. The square foot is a useful dimension because (1) it is easy to
visualize the number of granules in a relatively small area, and (2) most birds, even a small
songbird, can readily forage over a square foot of'soil surface. In keeping with DeWitt's
suggestion, the following equation can be used to relate toxicity and exposure:
_ Exposure _ Granules/sq ft _ LDsos
Toxicity Granules/LD50 ft2
OPP estimated the number of avian LD50s/ft2 for each of the 10 crop uses listed in
table 3-7. Table 3-7 presents the minimum and maximum values for broadcast, band, in-furrow,
and turn-row areas, based on label-directed application rates. These estimates suggest that the uses
of granular carbofuran pose a risk to individuals of each of the representative avian groups.
OPP estimated the potential magnitude of avian mortalities from direct carbofuran poisoning
resulting from application to 9 of the 10 crops. This estimate was based on the number of acres
treated per year (table 3-8) and the mortality in the field studies conducted in corn. (As stated
earlier, mortality estimates are probably quite conservative.) If it is assumed that similar mortality
occurs in all crops, there is a potential for several million avian deaths each year resulting from the
use of granular carbofuran. Because of the difficulties involved in finding dead birds, these
estimates may be low.
From an analysis of the laboratory toxicity data, estimated exposure data, field studies, and
incident reports described above, OPP concluded that granular carbofuran posed a risk to birds
based on its acute toxicity. OPP determined that the risks associated with the continued use of
granular carbofuran outweigh possible benefits and that the granular formulations should be
canceled to prevent unreasonable adverse effects to the environment. See the carbofuran Position
Document 2/3 (U.S. EPA, 1989) for a comprehensive discussion of benefit analyses.
Ecological risk analyses always involve a degree of uncertainty with regard to
characterization of ecological effects and exposure. For example, laboratory toxicity data are
available for only a limited subset of the nearly 850 species of birds that breed in or pass through
the United States. It is unlikely that the most sensitive species was tested for its susceptibility to
carbofuran poisoning.
Similarly, exposure estimates were based on both the highest and lowest registered
application rates for each site. This resulted in wide variations in the estimates of the number of
granules exposed on the soil surface. It should be noted also that, in order to keep calculations
manageable, exposure estimates were limited to granules on the soil surface. Additionally, it is
assumed that birds can and do consume granules when they are available. There is little
information, however, on the extent to which bird anatomy and behavior influence granule
consumption.
Finally, in the LD50/ft2 calculations, it is assumed that the higher the number of LD50s/ft2,
the greater the risk. The actual relationship between the number of available granules and the
actual risk, however, is not known. These uncertainties are counterbalanced by the more than 40
actual incidents of avian mortality discussed above.
3-20
-------
Table 3-7. Minimum and Maximum Values for Avian LD50s/Ft2 for 10 Crops
(U.S. EPA, 1989)
Crop
Songbirds
Upland Game Birds
and Waterfowl
Minimum
Maximum
Corn
Cotton
Cranberries
Peanuts
Pineseed
Rice
Sorghum
Soybeans
Sunflowers
Tobacco
68
68
1,052
32
5,170
156
56
68
68
1,368
2,211
527
1,052
2,632
10,002
261
1,033
2,104
1,368
3,159
Minimum
6
6
89
3
438
13
5
6
6
114
Maximum
187
45
89
223
846
22
89
178
116
267
3-21
-------
Table 3-8. Annual Estimated Acreage Treated With Granular Carbofuran (U.S. EPA, 1989)
Crop
Corn
Sorghum
Soybeans
Peanuts
Tobacco
Cotton
Cranberries
Pineseed
orchards
Sunflowers
Acres Treated
4,500,000-5,500,000
640,000-2,040,000
210,000-280,000
136,000
20,000-40,000
30,000
185
400-1,250
18,400-105,900
3-22
-------
Comments on Risk Characterization
Strengths of the case study include:
*The data upon which the risk characterization is based are substantial. Both field
and laboratory data are employed. Even though there are inadequacies in the design
of field studies, the evidence for avion mortality following use of granular carbojuran
is so great that these inadequacies do not affect the conclusions. OPP believes that
only for a few pesticides will there be an equal amount of data upon which to base a
risk characterization for adverse effects on birds.
Limitations include:
• With respect to most other pesticides, this case study represents a rather unique
situation and, as a model for future case studies, may be most applicable to other
granular formulations. It may not be appropriate for other pesticide formulations,
however, especially those that have long-term or sublethal effects.
•Although LD^/fi2 appears to be a useful measure for assessing toxicity of
carbojuran, it may not be appropriate for other pesticides, including other granular
formulations. Determination of the amount of active ingredient of the pesticide of
concern within different matrices, routes of exposure, and bioavailability is essential
in future studies.
* Inadequacies identified in the field studies include: (a) control sites are lacking or
inappropriate, as are data on carcass-search efficiencies; (b) possible synergistic
effects resulting from the use of other pesticides are not examined; and (c) although
the case study focuses on birds, determinations of possible impacts on other
taxonomic groups, particularly small mammals, should have been incorporated.
3-23
-------
3.4. REFERENCES
Balcomb, R. (1980) Granular pesticides: restricted classification based on hazard to avian wildlife
(unpublished document). Ecological Effects Branch, U.S. EPA.
Balcomb, R.; Bowen, C.A., HI; Wright, D.; Law M. (1984a) Effects on wildlife of at-planting
corn applications of granular carbofuran. J. Wildl. Mgmt. 48(4): 1353-1359.
Balcomb, R.; Stevens, R.; Bowen, C.A., III (1984b) Toxicity of 16 granular insecticides to wild-
caught songbirds. Butt. Environ. Contam. Tox. 33:302-307.
Beskid, J.; Fink, R. (1981) Simulated field study—bobwhite quail: final report. Unpublished report
by Wildlife International, Ltd. Submitted to U-S. EPA by American Cyanamid, Princeton,
NJ.
DeWitt, J.B. (1966) Methodology for determining toxicity of pesticides to wild vertebrates. J.
Appl Ecol Suppl. 3:275-278.
Dingledine, J. V. (1985) An evaluation of the effects of Counter 15G to terrestrial species under
actual field use conditions. Unpublished report by Wildlife International, Ltd. Submitted to
U.S. EPA by American Cyanamid, Princeton, NJ.
Erbach, D.; Tollefson, J. (1983) Granular insecticide application for corn rootworm control.
Trans. Am. Soc. Agric. Eng. 26:696-699.
Fink, R. (1974) Eight-day dietary LC50—mallard duck: final report. Unpublished report by
Wildlife International, Ltd. Submitted to U.S. EPA by Mobay Chemical Corp., Kansas
City, MO.
Fink, R. (1976) Eight-day dietary LC50—bobwhite quail: final report. Unpublished report by
Wildlife International, Ltd. Submitted to U.S. EPA by Chevron Chemical Co., Richmond,
CA.
Hill, E.F.; Heath, R.G.; Spann, J.W.; Williams, J.D. (1975) Lethal dietary toxidties of
environmental pollutants to birds. U.S. Fish and Wildlife Serv. Spec. Scien. Rep. No. 191.
USFWS, Washington, DC.
Hill, E.F.; Camaradese, M. (1984) Toxicity of anticholinesterase insecticides to birds: technical
grade versus granular formulations. Ecotox. Environ. Safety 8:551-563.
Hudson, R.H.; Tucker, R.K.; Haegele, M.A. (1984) Handbook of toxicity of pesticides to wildlife.
2nd. ed. U.S. Fish and Wildlife Service, Resource Pub. No. 153.
Hummel, J.W. (1983) Incorporation of granular insecticides by com planters. Trans. ASAE-1983.
Am. Soc. Agric. Eng. Paper No. 83-1017/83-7008.
3-24
-------
Schafer, E.W., Jr.; Brunton, R.B.; Lockyer, N.F.; DeGrazio, J.W. (1973) Comparative toxicity
of seventeen pesticides to the quelea, house sparrow, and red-wingedblackbird. Toxicol.
Appl. Pharmacol 26(1): 154-157.
Schafer, E.W., Jr.; Brunton, R.B. (1979) Indicator bird species for toxicity determinations: Is the
technique usable in test method development? In: Beck, J.R., ed. Vertebrate pest control
and management. ASTM STP 680. Philadelphia, PA: Am. Soc. Test, and Materials, pp.
157-168.
Shellenberger, T.E.; Gough, B.J. (1972) Acute toxicity evaluations of carbofuran with bobwhite
quail chicks and young adults and mallard ducklings. Unpublished report by Gulf South
Res. Inst. Submitted to U.S. EPA by FMC Corp., Philadelphia, PA.
Tucker, R.K.; Crabtree, G.D. (1970) Handbook of toxicity of pesticides to wildlife. U.S. Fish and
Wildlife Service, Resource Pub. No. 84. 131 pp.
Urban, D.J.; Cook, NJ. (1986) Standard evaluation procedure for ecological risk assessment.
EPA 540/09-86/167.
U.S. Environmental Protection Agency. (1989) Carbofuran special review technical support
document. EPA 540/09-89/027.
U.S. Environmental Protection Agency. (1990) Office of Pesticides and Toxic Substances state of
the practice ecological risk assessment document. Ecological Effects Branch, Environmental
Fate and Effects Division, Office of Pesticide Programs. Washington, DC (unpublished).
3-25
-------
-------
SECTION FOUR
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
ECOLOGICAL EVALUATION OF A FRESHWATER STREAM AND WETLANDS
NEAR AN INACTIVE COKE PRODUCTION PLANT
-------
AUTHORS AND REVIEWERS
AUTHORS
Wayne S. Davis
Region 5
U.S. Environmental Protection Agency
Chicago, IL
REVIEWERS
G. Allen Burton (Lead Reviewer)
Biological Sciences Department
Wright State University
Dayton, OH
Robert Brooks
School of Forest Resources
Pennsylvania State University
University Park, PA
Rick D. Cardwell
Parametrix, Inc.
Bellevue, WA
Michael J. Dover
The Cadmus Group, Inc.
Peterborough, NH
Anthony F. Maciorowski
Fish Culture and Ecology Laboratory
U.S. Fish and Wildlife Service
KearneysviHe, WV
John Bascietto
U.S. Department of Energy
Washington, DC
Joseph Makarewicz
Department of Biological Sciences
State University of New York
at Brockport
Brockport, NY
Thomas P. O'Connor
National Status and Trends Program
National Oceanic and Atmospheric
Administration
Rockville, MD
Jerry R. Schubel
Marine Sciences Research Center
State University of New York
at Stony Brook
Stony Brook, NY
4-2
-------
CONTENTS
ABSTRACT 4-6
4.1. RISK ASSESSMENT APPROACH 4-7
4.2. STATUTORY AND REGULATORY BACKGROUND 4-7
4.3. CASE STUDY DESCRIPTION 4-9
4.3.1. Problem Formulation 4-9
4.3.2. Analysis: Characterization of Exposure 4-12
4.3.3. Analysis: Characterization of Ecological Effects 4-16
4.3.4. Risk Characterization . 4-18
4.4. REFERENCES 4-21
APPENDIX A—SUMMARY OF METHODS USED FOR IDENTIFICATION OF
STRESSORS 4-A1
APPENDIX B—COMPOUNDS AND PARAMETERS SAMPLED IN SURFACE
WATERS AND SEDIMENTS 4-B1
APPENDIX C—RESULTS OF BIOLOGICAL SURVEYS 4-C1
4-3
-------
LIST OF FIGURES
Figure 4-1. Structure of analysis for inactive coke production plant 4-8
Figure 4-2. Sampling location 4-13
LIST OF TABLES
Table 4-1. Sample Matrix 4-14
Table 4-B1. Surface Waters: Toxic Pollutants 4-B2
Table 4-B2. Surface Waters: Classical Parameters 4-B2
Table 4-B3. Sediments: Polycyclic Aromatic Hydrocarbons 4-B3
Table 4-C1. Benthic Macroinvertebrates Collected in Petite Ponar Grab Samples 4-C2
Table 4-C2. Numbers and Species of Fishes Collected . 4-C3
Table 4-C3. Condition Factors Calculated for Fish 4-C4
Table 4-C4. Plankton Collected in Lagoons 3 and 4 4-C5
4-4
-------
LIST OF ACRONYMS
ARAR applicable or relevant and appropriate requirement
CERCLA Comprehensive Environmental Response, Compensation, and Liability Act of 1980
DO dissolved oxygen
EPA U.S. Environmental Protection Agency
HRS Hazard Ranking System
NCP National Contingency Plan
PAH polycyclic aromatic hydrocarbon
RI/FS Remedial Investigation/Feasibility Study
SARA Superfund Amendments and Reauthorization Act of 1986
4-5
-------
ABSTRACT
A small stream and wetlands in southeastern Ohio received long-term waste discharges from
a coke production facility before the facility closed in the early 1980s. The area affected by these
discharges was determined to be eligible for cleanup action under Superfund. A Remedial
Investigation/Feasibility Study was initiated to evaluate the site's potential impacts on human health
and the environment and to develop a cost-effective remedial action plan.
Coke production facilities have been associated with high discharge levels of polycyclic
aromatic hydrocarbons and adverse effects on aquatic biota, particularly bottom-feeding fish. The
study included an examination of the surface water and sediment chemistry for a wide array of
inorganic and organic chemicals as well as the aquatic biota. Studies were conducted on fish,
benthic macroinvertebrates, phytoplankton, and zooplankton communities. Bottom-feeding fish
were examined for neoplasms.
Conclusions from the study were that:
" the lagoons are the most severely affected areas, being highly chemically
contaminated, supporting a benthic community indicative of polluted conditions, and
containing no fish life;
• nutrient enrichment and reduction in dissolved oxygen concentrations in the stream
may be attributable to untreated sewage discharges into the stream;
• the fish community found in the backwater marsh (slough) and at the mouth of the
stream consisted of species commonly found in the larger rivers of the area, reflecting
the influence of these larger rivers;
• the fish examined did not have a significant incidence of liver neoplasia, but this may
not have been a sensitive indicator; and
• an improved study plan might have provided a basis for identifying exposure-effects
relationships with the data generated.
4-6
-------
4.1. RISK ASSESSMENT APPROACH
This case study, which was originally prepared by Ballantyne et al. (1984) and IT
Corporation (1990), represents a typical impact assessment (ecological reconnaissance study) and
does not follow the ecological risk assessment process as defined in the framework (figure 4-1). It
can, however, serve as a good example of a baseline risk assessment with some modification (as
noted later in the Comments sections). Many contaminated sites are similar to this one and follow
similar assessment scenarios. The section-by-section comments identify the study's deficiencies
and recommend improvements. The study uses multiple measurement endpoints (chemical,
physical, and biological) that have been documented elsewhere as being effective and valid and
should be considered in future risk assessments.
4.2. STATUTORY AND REGULATORY BACKGROUND
Environmental assessment of a Superfund site is done in accordance with the U.S.
Environmental Protection Agency's (EPA's) responsibility to protect public health and the
environment at uncontrolled hazardous waste sites under the Comprehensive Environmental
Response, Compensation, and Liability Act of 1980 (CERCLA), as amended by the Superfund
Amendments and Reauthorization Act of 1986 (SARA). The regulation that enables EPA to carry
out its responsibilities under CERCLA/SARA is the National Contingency Plan (NCP).
Under the NCP, EPA must evaluate a site for eligibility for certain cleanup actions under
CERCLA/SARA authorities. Although it is not a risk assessment methodology, the Agency's
Hazard Ranking System (HRS) is used to determine national priorities for cleanup of hazardous
waste sites. The HRS scoring involves a detailed evaluation of exposure and hazardous potential
relative to the known and potential contaminants at the site, the exposure pathways (air, ground
water, surface water, direct contact, etc.), and the known or potential ecological component
population. The process of remedy development starts with a Remedial Investigation/Feasibility
Study (RI/FS), with the primary objectives of evaluating the site's potential impact on human
health and the environment and developing a cost-effective remedial action plan. The NCP calls
for the identification and mitigation of environmental impacts on these sites and the selection of
remedial actions that are "protective of environmental organisms and ecosystems." Federal and
state laws and regulations that aid in this process are potentially "applicable or relevant and
appropriate requirements" (ARARs). Compliance with these laws and regulations increasingly
requires that the site's ecological effects be evaluated and measures be taken to mitigate those
adverse effects.
The Clean Water Act, as amended by the 1987 Water Quality Act, is another ARAR and
major federal regulation that requires the maintenance and restoration of the chemical, physical,
and biological integrity of the Nation's waters. Most Superfund sites potentially affect surface
waters and need to be assessed for both onsite and offsite effects. Recently, EPA identified the
biological integrity of surface waters as indicators of both chemical and physical stressors and as
direct measurements of the aquatic life that is protected by federal and state regulations (U.S. EPA,
1991a). EPA recommends that an integrated approach be used for assessing aquatic resources
utilizing chemical, biological, and physical measurement and assessment endpoints (U.S. EPA,
1988a,b; 1990a). A detailed discussion of the legal and technical requirements for environmental
4-7
-------
Figure 4-1. Structure of Analysis for
Inactive Coke Production Plant
PROBLEM FORMULATION
Stressors: inorganic and organic chemicals
associated with coke production waste residuals; low
dissolved oxygen levels associated with sewage
discharges.
Eco!
zoop
oaici
>lank
al Components: benthic macroinvertebrates and
ton and fish of a stream and river.
Endpoints; assessment endpoint is biological integrity
in surface water; measurement endpoints include
structural properties of fish, benthic, and planktonic
communities as well as fish histopathology.
ANALYSIS
Characterization of
Exposure
Exposure was evaluated by
measuring the
chemical concentrations
at six stream stations,
two lagoon locations, and
one river station.
Characterization of
Ecological Effects
Effects were evaluated
based on the literature and
examination of community
structure in affected and
reference areas and by
examining fish for tumors.
RISK CHARACTERIZATION
Risks were characterized by:
- comparing data on chemical concentrations to criteria,
- comparing biological communities in potentially
affected and reference areas, -*> ^
- examining the occurrence of tumors.
4-8
-------
assessments at Superfund sites can be found in EPA's Risk Assessment Guidance for Superfund:
Environmental Evaluation Manual (1989a).
4.3. CASE STUDY DESCRIPTION
4.3.1. Problem Formulation
Site Description. The site examined in this case study is an inactive coke production facility
that operated for most of this century in southeastern Ohio. Products from the coking operation,
which ended in the early 1980s, included crude tar, coke, light oil, and ammonia. From
approximately 1920 through the late 1960s, wastewater and solid wastes generated in the coking
process were discharged into wetlands east of the plant, adjacent to Ice Creek, which traverses the
property. This creek is a tributary to the Ohio River, a major interstate river, and is located about
750 ft from the plant. The waste streams included process wastewater, coke and coke fines,
decanter tank car sludge, boiler ash, and weak ammonia liquor.
Stressors. Information collected during the RI/FS identified contaminants within the stream,
onsite lagoons, and wetlands including ammonia, benzene, cyanide, chlorides, metals, naphthalene,
phenol derivatives, polycyclic aromatic hydrocarbons (PAHs), and phthalate esters. Routine
parameter analyses to assist in identification of potential stressors included dissolved oxygen (DO),
pH, temperature, conductivity, CaCO3 hardness, and a subjective habitat review.
Methods used for the identification of stressors included chemical analyses of surface waters
and sediments, biological community assessments, and histopathological examination of bottom-
feeding fish. A brief summary of the specific methods used appears in appendix A.
Ecological Components. Ideally, because natural systems are composed of individual
organisms, local populations, and communities, the risk to each of these components should be
assessed. However, because of the large number of species present in natural systems and the
complexity of the interspecies relationships, it is necessary to select representative components of
the ecosystem and to develop appropriate endpoints. The Superfund program has addressed many
of these issues in recent technical guidance and review documents (U.S. EPA, 1987, 1989a,b,
c,d,e) and is in the process of developing supplemental guidelines to their Environmental
Evaluation Manual.
This case study examined selected ecological components of the biological communities in
the freshwater stream and associated wetlands and lagoons. These included benthic
macroinvertebrates, plankton, and fish. Sessile aquatic life, such as benthic macroinvertebrates,
are particularly useful indicators of local environmental effects due to their lack of mobility.
Planktonic organisms are the basis of the aquatic food chains, and fish represent a high-level
consumer in aquatic systems. Each of these groups served as indicators of the aquatic resources
and were assessed in this case study. Fish populations (golden redhorse sucker and freshwater
drum) were also used to indicate sublethal effects from sediment contaminant exposure through the
incidence of liver neoplasia, although this is acknowledged to be an extreme indicator.
4-9
-------
Endpoint Selection. The assessment endpoint in this study was the biological integrity in the
surface waters adjacent to the site. The measurement endpoints used to evaluate the assessment
endpoint included structural determinations of the fish, benthic macroinvertebrate, and planktonic
communities in the surface waters, as well as the incidence of liver neoplasia and other anomalies
in two fish populations. Condition factors for fish populations were also included in the analysis
based upon weight and length of the individuals.
Comments on Problem Formulation
Strengths of the case study include:
+The case study documents the history of the site and identifies the primary chemicals
(stressors) of concern.
*The identification of the various levels of components and key references is good and
often overlooked. The use ofbiotic metrics also is good; however, they should be
briefly explained showing their many components and their relationship to the
community structure and ecosystem functioning.
+The use of community-level endpoints as an in-field measure of condition complements
the chemical-specific methods that are used to predict ecological effects. This approach
is supported by EPA's recent Policy on the Use of Biological Assessments and Criteria
in the Water Quality Program (1991a).
limitations include:
• The test hypothesis is not stated. \
^Literature information that establishes the adverse effect relationship between key
ecological components (sensitivity, abundance, life history, and contaminant fate) and
the contaminants of concern is not incorporated. This would involve defining the scope
of the assessment and justifying why the study contaminants and components were
selected (essentially this represents a preliminary risk assessment).
*The site description is weak. It should allow the reader to evaluate the
appropriateness of the various ecological components, stressors, and endpoints. There
may be other migratory species (fish and wildlife) that could visit the site area and be
exposed to contaminated media.
*No mention is made of replicate sampling design or the possible importance of*PAH
photoactivation to highly toxic compounds at the site.
4-10
-------
Comments on Problem Formulation (continued)
• Key species (e.g., based on sensitivity, endangered status, abundance, ecosystem
niche) should be identified, with a discussion of relevant information on life history,
food, and habitat requirements that may affect exposure-specific stressors. This
information is available from preliminary site surveys, regional experts, data bases,
and the literature.
• A complete risk assessment for the site should have included waterfowl and mammal
species that make use of the lagoons and areas around them. The current study
presents information on onfy the aquatic component.
General comments:
• Many of the endpoints that were measured could have been supplemented by more
contemporary analytical techniques or field collection methods (Karr et al., 1986;
Klemm et al., 1990). Endpoints that were not examined include sediment toxicity and
habitat effects. Quantification of habitat conditions is important to determine the
ecological expectations, of the area, and whole-sediment toxicity testing using a variety
of species (e.g., amphipod, chironomid, crustacean) would have provided an
indication of potential and actual ecological risks due to sediment contamination.
*A muUimetric approach for assessing ecological community health is recommended
similar to EPA's Rapid Bioassessment Protocols (Plafldn et al., 1989). This study '
was conducted in Ohio, a state that has established methods and standard operating
procedures for assessing the biological integrity of aquatic systems; their approach
and techniques should have been employed (Ohio EPA, 1989, 1990). The Ohio EPA
uses ecoregion-based, multiassemblage (fish and benthos), and multimetric endpoints
(Index ofBiotic Integrity, Modified Index ofWell-Being, Invertebrate Community
Index, and Qualitative Habitat Evaluation Index). It is important to consult state
biologists and scientists to determine whether special or deliberate sampling methods
and interpretation of results are needed.
• Shannon's diversity index is not recommended as a primary method for assessing
benthic community health due to its inconsistent relationship to ecological health
(Hughs, 1978; Chadwick and Canton, 1984; Washington, 1984; Resh and Jackson,
1990; Davis and Lathrop, 1992). The use of Shannon's diversity index for assessing
the benthic macroinvertebrate community in this study is misleading because of the
'reference to Wilhm's (1970) pollution classification, which is calculated using base 2,
while those in this study are calculated using base 10.
4-11
-------
4.3.2. Analysis: Characterization of Exposure
Sample Locations. Six sampling stations were established for chemical and biological
community sampling along the linear section of Ice Creek that received discharges and runoff from
the plant area (figure 4-2). Stations IC-1 and IC-2 were reference stations in the creek, upstream
from the plant area. Station IC-3 was adjacent to the upper part of the plant site, and Station IC-4
was established near the east side of the slough area to observe the potential effects of an
unidentified effluent discharge (possibly an untreated sewage outfall). Station IC-S was at the
lower end of the plant site near the west side of the slough area, and Station IC-6 was downstream
of the plant site a few hundred feet away from the Ohio River. Two lagoon-wetland stations (LG1
and LG2) were established along lagoon transects. The mouth of Ice Creek at the Ohio River was
station OR-1, just downstream from Station IC-6.
The investigation of fish neoplasia occurrences was conducted in three reaches of the
creek—two reference reaches and one test reach. The upstream reference reach was located
starting from Station IC-2 and proceeding about 1,000 ft downstream, about halfway to Station
IC-3. The test reach was located starting just downstream of Station IC-3 covering a distance of
about 750 ft downstream. The second reference reach was located in the major interstate river
upstream from the confluence with the tributary extending a distance of about 1 mile.
Sample Matrix. Water sampling investigations were designed using the results of the
previous phases of the remedial investigation. Table 4-1 presents the matrix of parameters that
were sampled at each site. Acid and base/neutral priority pollutants, cyanide, benzene, chloride,
sulfate, arsenic, and metals such as lead, mercury, and cadmium were considered potential
contaminants for analysis screens (results are in appendix B). Standard physical water quality
parameters measured included DO, pH, temperature, specific conductance, alkalinity (mg/L
CaCO3), and total hardness (mg/L CaCO3). Ammonia (mg/L NH3-N) was also measured.
Sediments were analyzed for arsenic, heavy metals, and organic priority pollutants. Biological
exposures were assessed through community studies of the benthic macroinvertebrates, fish, and
plankton and examination of liver neoplasia in bottom-feeding fish.
4-12
-------
JC-1
ICE CREEK TEST STATION
ICE CREEK REFERENCE
STATION
iC-1 ' ICE CREEK SEDIMENT BORING.
• BATTELIE 1985
OR-2 OHIO RIVER SEDIMENT BORING,
• 8ATTEUE 1985
L,G LAGOON 1
SCALEOF FEET
1000
2000
Figure 4-2. Sampling location (IT Corporation, 1990)
4-13
-------
Table 4-1. Sample Matrix (Ballantyne et al., 1984)
Station Number
Parameter IC-1 IC-2 IC-3 KM IC-5 IC-6 LG1 LG2 OR-1
Water chem. XXXXXXXXX
Sediment chem. XXXXXXXXX
Benthos XX X XX X XXX
Zooplankton X X
Phytoplankton X X
Fish X X X X X X X
Pathology XX .X
4-14
-------
Comments on Analysis: Characterization of Exposure
Strengths of the case study include:
• Chemicals were measured in various environmental media.
Limitations include:
•Little information is provided allowing the user to evaluate exposure. The
hydrodynamics of surface water and ground water infiltration is not described and no
information is provided on the loading of stressors. The migration of contaminants
from the lagoon-depositional areas to other locations is unknown.
•Data on tissue residues are not reported, nor are key fish condition factors
including age, size, sex, spawning status, and lipid content.
9An attempt should have been made to locate reference (least affected) stations in a
nearby basin for comparison with the results of this study, since the upstream stations
of the creek appeared to be affected. In the absence of using reference locations from
a nearby watershed, more care should have been taken in selecting the stations that
were sampled. The most upstream station was only a few thousand feet away from
the potential influence of the facility and was subject to potential roadway influences.
Station IC-3 was adjacent to the facility but could have been a few hundred feet
further downstream, or another sample location should have been added just upstream
of the influence of the lagoons. Stations IC-4 and IC-5 were located on either side of
the wetland/backwater area adjacent to the lagoons and were not representative of the
creek habitat. A better station selection would have been in the middle of the main
creek channel that fed the backwater. Station IC-6 was properly located immediately
downstream of the backwater in a channel connecting the creek to the Ohio River.
The river station was not located in the river itself, but was a little further
downstream from Station IC-6 in the connecting channel. Most of the stations, with
the possible exception of the three upstream stations, are directly influenced by flow
and depth changes in the river during dry weather. In wet weather events, it is likely
that the creek upstream of the facility receives some contaminant contribution. Ohio
River stations directly upstream and downstream from the confluence with the creek
should have been included. The transitory nature offish is not recognized, nor are
habitat effects on biotic indices.
•Method detection limits are relatively high for most of the compounds, especially
mercury, and this makes it difficult to conduct a definitive assessment.
4-15
-------
4.3.3. Analysis: Characterization of Ecological Effects
Ecological effects were characterized by identifying state water quality standards, examining
biological community structure, and studying the histopathology of fish populations.
Water Quality Standards. Water quality standards are adopted by states and are legal
requirements for establishing the desired condition of a water body. Water quality standards are
ARARs that must be addressed at all Superfund sites. State standards consist of three parts:
(1) the beneficial use designated for that water body, (2) numerical or descriptive criteria that
measure specific conditions of the water body designed to protect the beneficial use, and (3) an
antidegradation statement to ensure that high-quality waters are not arbitrarily lowered to meet the
standards. The predominant criteria used to measure the condition of the water body have been the
numerical chemical criteria that are based on considerations of the magnitude, duration, and
frequency values for protection against both acute and chronic toxicity. Recently, direct measures
of the biological community structure and function have been shown to improve dramatically
EPA's ability to assess the attainment of a water body's use. As a result, EPA is requiring the
states to develop biological criteria and adopt them into their water quality standards by September
30, 1993 (U.S. EPA, 1990b; 1991b). Several states currently use biological criteria as a
regulatory tool, either alone or in combination with other ecological parameters (U.S. EPA,
1990b).
Biological Community Structure. Community structure can be measured at specific
locations in areas suspected of contamination and at reference locations either in the same body of
water in areas thought to be unaffected by contamination or in a nearby body of water that is
similar in characteristics. Selection of reference locations is critical in providing the desired, or
"least affected" condition, with which the community structure of the test location will be
compared.
Fish Histopathology. The morphological condition of the fish community can provide an
important indication of the biological integrity of a body of water. The incidence of tumors,
lesions, and other anomalies is generally recorded for the entire fish community. Histopathological
analyses, however, are usually limited to a few indicator species due to the time involved for a
complete analysis. As with the biological community structure, comparisons of tumors, lesions,
and anomalies are made between the affected locations and the reference locations. It is as critical
to choose a proper reference site as it is to make a proper selection of the indicator species
subjected to the histopathological evaluations. The results of the biological studies are presented in
appendix C.
4-16
-------
Comments on Analysis: Characterization. of Ecological Effects
Strengths of the case study include:
+A variety of good data are generated, namely water and sediment chemistry and
benthic macroinvertebrate and fish community data.
Limitations include:
•No quantification of the relationship between exposure and the probability of
adverse ecological response exists. Biological, chemical, and physical components all
vary and should be acknowledged. Biotic and chemical criteria address some of the
varying exposure issues. The biological data could have been analyzed to a much
greater extent to define relationships. The data could be compared to ecoregion
biocriteria reference stations such as those of the Ohio EPA that are incorporated into
state water quality standards (Ohio Administrative Code 3745-1, adopted in February
1990, effective May 1990).
•Water column and sediment toxicity testing of multiple trophic levels would
significantly improve the characterization of ecological effects and could address
gradient (concentration) effects via spatial sampling and/or sample dilution. Toxicity
testing might be considered only if sites are predicted to be toxic based on elevated
stressor levels or "Equilibrium Partitioning." Key indicator, endangered, target, or
important ecosystem species (ecological components) could be monitored based on
community indices, acute-chronic effects, and/or tissue residues.
• The selection of reference stations for the histopathological assessment is not
appropriate. The upstream reference station chosen was within the influence of the
facility. Also, no consideration is given to the transitory nature of the fish and their
relationship with the major river. Because there were no natural or artificial barriers
to fish movement, it is likely thai the fish collected in the reference area also
inhabited the test location at various times.
• The use of liver neoplasia incidence in freshwater drum and golden redhorse sucker
as an indicator of aquatic community effects from contaminated sediments may not be
a sufficiently sensitive indicator because the incidence is related to gross pollution by
PAHs; a more sensitive indicator is needed for an appropriate ecological assessment.
There is also no demonstrated susceptibility of the chosen populations for neoplasia
development due to chemical stressors.
4-17
-------
4.3.4. Risk Characterization
Risks Based on Comparisons to Criteria. The chemistrvdata (appendix B) showed no
measured excesses of the aquatic life standards, with the exception of ammonia-nitrogen, even
when compared with chronic water quality standards that are more stringent than acute standards.
Ammonia and specific conductances were an order of magnitude higher in the lagoons than at any
of the stations. Ammonia exceeded aquatic criteria at Stations LG1, LG2, IC-3, IC-4, IC-6, and
OR-1. The potential untreated sewage outfall at Station IC-6 could have contributed to the
increased contaminant levels at the downstream stations.
Priority organic pollutants were not detected in surface waters. Acid and base/neutral
priority pollutants were sampled in the sediments but were not found. PAHs were elevated in the
sediments hi both lagoons and Station IC-6, which received direct discharge from the lagoons.
However, the PAHs do not appear to be mobilized from the sediments or adversely affecting the
aquatic community, as evidenced by the absence of fish liver neoplasia. No established sediment
quality criteria were identified to compare with the sediment chemistry; therefore, the
environmental significance of the values of PAHs in the sediments from the stream and lagoons
could not be directly assessed from the exposure information. Toxicity testing was not conducted
with the sediments or surface waters.
Risks Based on Biological Community Surveys. The results of the biological surveys are
summarized hi appendix C. Benthic macroinvertebrate structure was measured by the number of
individuals, the number of taxa, and the Shannon diversity indices (table 4-C1), which were all
very low and indicative of polluted conditions (Wilhm, 1970). The dominant species were tubificid
worms, which can tolerate low levels of DO and high organic enrichment. The highest tubificid
concentrations occurred at Station IC-4, downstream of the untreated sewage outfall. Organisms
with higher DO requirements were found in the shoreline qualitative samples. Stream habitats at
the reference stations were not comparable to the test stations, due mainly to a sandy substrate
upstream and much sillier sediments downstream. Statistical tests such as Kruskal-Wallis test
(Hollander and Wolfe, 1973) and Shannon's diversity index were calculated and showed no
significant differences in species diversity among the reference and test stations.
Twenty-seven fish species and one amphibian were collected from the stream (table 4-C2).
The stations averaged 13 species with a range of 10 to 16 species. The proximity of Stations IC-6
and OR-1 resulted hi the reach between these two stations being sampled as one location, and the
data were combined. Emerald shiners, bluegill, and gizzard shad were the most numerous species,
but no significant differences occurred in the number of species found at any of the test stations
compared with the reference stations. Downstream stations generally produced larger specimens,
perhaps reflecting the use of these reaches by fish that are characteristic of the large river.
Upstream samples had greater numbers of first year class shad, minnow, and shiners.. Condition
factors were calculated for sport fish and dominant species (table 4-C3). Mean values were
comparable to the expected ranges of weight and length for the area (Carlander, 1969; Bennett,
1970). Although no fish were observed in the lagoons, many turtles and waterfowl were observed
there. Overall, the fish community appeared to be more diverse than the macroinvertebrate
community. The fish appeared to be robust and relatively free from disease. The backwater areas
of the stream may serve as refuge for fish from the larger river.
4-18
-------
The dominant phytoplankton species in the lagoons was the blue-green alga, Anabaena, and
both lagoons supported limited zoqp.lankton communities (table 4-C4). The density of
phytoplankton cells in Lagoon 1 was almost 80 times greater than hi Lagoon 2. Conversely, the
density of zooplankton was about three times greater hi Lagoon 2 than in Lagoon 1. The lagoons
collect surface water runoff from the site, as well as ground water percolating through the
surrounding area, and contain waters that are in direct contact with highly contaminated sediments.
This factor likely accounts for the limited phytoplankton and zooplankton communities, and
probably accounts for the total lack of fish life. It is reasonable to conclude that the lagoons could
contain fish because of their connection to the stream during flooding.
Risks Based on Fish Histopathology. The examination of fish liver neoplasia was designed
to determine whether there was an adverse effect on the aquatic fauna due to contaminated creek
sediments. No neoplastic lesions were observed in the liver of golden redhorse sucker or
freshwater drum from any of the reference or test station populations. Fin rot in freshwater drum
was 44 percent in both creek stations and 70 percent in the; major river. For the golden redhorse,
the incidence of fin rot was 51 percent in the creek test area and 19 percent in the test reference
ijv
Conclusions. Based on the three methods used to evaluate conditions, the principal
conclusion drawn from this study is that the Superfund sitejand the creek sediments do not have an
adverse effect on the aquatic fauna. f;
Comments on Risk Characterization
Strengths of the case study include:
• The use of biological community assessment and population measures greatly
complements the chemical-specific results. Such an integrated approach to evaluating
the quality of the water resources is strongly advocated. EPA has prepared a number
of guidance documents and critical review documents for use by EPA programs to
facilitate the utilization and application of ecological assessments, including many
documents specifically directed toward the Superfund process (U.S. EPA, 1987,
1989a,b,c,d,e, 1990a). EPA also supports regional efforts, such as the annual
Midwest Pollution Control Biologists Meeting in Chicago, to report on new and
promising ecological assessment methods (U.S. EPA, 1988b, 1989f, 1990c). These
documents, as well as others cited in this case study, should be obtained and carefully
reviewed for application to Superfund ecological assessments.
.r
Limitations include:
• This study was an impact assessment that did not consider the probability of effects
or uncertainty.
4-19
-------
Comments on Risk Characterization (continued)
*A better selection ofendpoints and sample locations was needed before the study
was initiated for a more definitive exposure assessment and subsequent risk
characterization. In this case study, a conclusion was reached that the contaminated
sediments did not adversely affect the aquatic life in the creek based on a study of
liver neoplasia from two nonindigenous species offish. This is a nonsensitive
measure that may require high contaminant levels to produce an effect. More
sensitive measures to characterize the risk could have included whole-sediment toxicity
testing coupled with a more extensive and frequent chemical analysis of the sediment
and a more rigorous analysis of the benthic community by using more useful
biological metrics and by using artificial substrates.
• Overall, the conclusion that no community impacts are associated with the
Superfund site or creek sediments is not adequately substantiated.
4-20
-------
4.4. REFERENCES
Ballantyne, M.A.; Clement, W.H.; Dena, J.H; Duke, K.M. (19&fjiSllal report on aquatic
ecological studies at Allied Chemical's Ironton, Ohio, coke site. Battelle Columbus
Laboratories, Columbus, OH.
Bennett, G.W. (1970) Management of lakes and ponds, 2nd. ed. New York: Van Nostrand
Reinhold Co. ;
Carlander, K.D. (1969) Handbook of freshwater fishery biology, Vol. 1. Ames, IA: Iowa State
University Press.
Chadwick, J.W.; Canton, S.P. (1984) Inadequacy of diversity indices in discerning metal mine
drainage effects on a stream invertebrate community. Water Air Soil Pollut. 22:217-223.
Davis, W.S.; Lathrop, I.E. (1992) Freshwater benthic macroihvertebrate community structure and
function. In: Sediment classification methods compendium, Chapter 8. EPA Office of Water,
Washington, DC. EPA 823/R-92/006. pp. 8-1-8-26. /;•_-
Hollander, M.; Wolfe, D.A. (1973) Nonparametric statistical methods. New York, NY: John
Wiley and Sons.
Hughs, B.D. (1978) The influence of factors other than pollution on the values of Shannon's
diversity index for benthic macroinvertebrates in streams. Water Res. 12:359-364.
IT Corporation. (1990) Draft feasibility study: coke plant/lagoon area, Allied-Signal/Ironton coke
site, Ironton, Ohio. Volume II. Prepared for Allied-Signal, Inc., Project No. 303816.
Karr, J.R.; Fausch, K.D.; Angermeier, P.L.; Yant, P.R.; Schlosser, I.J. (1986) Assessing
biological integrity in running waters: a method and its rationale. Illinois Natural History
Survey, Special Publication 5, Springfield, IL. 28 pp.
Klemm, D.J.; Lewis, P.A.; Fulk, F.; Lazorchak, J.M. (1990) Macroinvertebrate field and
laboratory methods for evaluating the biological integrity of surface waters. EPA Office of
Research and Development, Environmental Monitoring and Systems Laboratory, Cincinnati,
OH. EPA 600/4-90/030. 256 pp.
Ohio Environmental Protection Agency. (1989) Biological criteria for the protection of aquatic life:
volume 111. Standardized biological field sampling and laboratory methods for assessing fish
and macroinvertebrate communities. Division of Water Quality Planning and Assessment,
Ecological Assessment Section, Columbus, OH.
Ohio Environmental Protection Agency. (1990) The use ofbiocriteria in the Ohio EPA surface
water monitoring and assessment program. Division of Water Quality Planning and
Assessment, Ecological Assessment Section, Columbus, OH.
4-21
-------
Plafkin, J.L.; Barbour, M.T.; Porter, K.D.; Gross, S.K.; Hughs, R.M. (1989) Rapid
bioassessment protocols for use in streams and rivers: benthic macroinvertebrates and fish.
Office of Water Regulations and Standards, Washington, DC. EPA 444/4-89/001.
Resh, V.H.; Jackson, J.K. (1990) Rapid assessment approaches to biomonitoring using benthic
macroinvertebrates. In: Rosenberg, D.M.; Resh, V.H. Freshwater biomonitoring and
benthic macroinvertebrates. New York: Chapman and Hall.
U.S. Environmental Protection Agency. (1987) A compendium of Superfund field operations
methods. Section 12, Biology/Ecology. Office of Emergency and Remedial Response,
Washington, DC. EPA 540/P-87/001.
U.S. Environmental Protection Agency. (1988a) Report of the National Workshop on Instream
Biological Monitoring and Criteria. EPA Region 5 Instream Biological Criteria Committee,
EPA Office of Water, Washington, DC. 34 pp.
U.S. Environmental Protection Agency. (1988b) Proceedings of the First National Workshop on
Biological Criteria. Lincolnwood, IL, December 2-4, 1987. EPA Region 5 Instream
Biocriteria and Ecological Assessment Committee, Chicago, IL. EPA 905/9-89/003.
129 pp.
U.S. Environmental Protection Agency. (1989a) Risk assessment guidance for Superfund:
environmental evaluation manual. Interim Final. Office of Emergency and Remedial
Response, Washington, DC. EPA 540/1-89/001A.
U.S. Environmental Protection Agency. (1989b) Ecological assessment of hazardous waste sites.
Office of Research and Development, Corvallis, OR. EPA 600/3-89/013.
U.S. Environmental Protection Agency. (1989c) Summary of ecological risks, assessment methods,
and risk management decisions in Superfund and RCRA. Office of Policy Analysis,
Washington, DC. EPA 230/03-89/046.
U.S. Environmental Protection Agency. (1989d) The nature and extent of ecological risks at
Superfund sites and RCRA facilities. Office of Policy, Planning, and Evaluation,
Washington, DC. EPA 230/03-89/043.
U.S. Environmental Protection Agency. (1989e) Ecological risk assessment methods: a review and
evaluation of past practices in the Superfund and RCRA programs. Office of Policy,
Planning, and Evaluation, Washington, DC. EPA 230/03-89/044.
U.S. Environmental Protection Agency. (1989f) Proceedings of the 1989 Midwest Pollution
Control Biologists Meeting. Chicago, IL, February 14-17, 1989. Davis, W.S.; Simon, T.P.,
eds. EPA Region 5 Instream Biocriteria and Assessment Committee, Chicago, IL. EPA
905/9-89/007. 153 pp.
4-22
-------
U.S. Environmental Protection Agency. (1990a) Biological criteria: national program guidance for
surface waters. Office of Water, Washington, DC. EPA 440/5-90/004.
U.S. Environmental Protection Agency. (1990b) Development of biological criteria by the states.
Draft, Office of Water, Washington, DC.
U.S. Environmental Protection Agency. (1990c) Proceedings of the 1990 Midwest Pollution
Control Biologists Meeting. Chicago, IL, April 10-13, 1990. Davis, W.S., ed. EPA Region
5, Environmental Sciences Division, Chicago, IL. EPA 905/9-90/005. 142 pp.
U.S. Environmental Protection Agency. (1991a) Policy on the use of biological assessments
and criteria in the Water Quality Program. Office of Water, Office of Science and
Technology, Washington, DC.
U.S. Environmental Protection Agency. (1991b) A guide to the Office of Water accountability
system and regional evaluations. Office of Water, March 1991, Washington, DC.
Washington, H.G. (1984) Diversity, biotic, and similarity indices: a review with special relevance
to aquatic ecosystems. Water Res. 18(6):653-694.
Hi i
Wilhm, J.L. (1970) Range of diversity index inbenthic macroinvertebrate populations. Water
Pollut. Control Fed. 4(5):R221-R224. •;/!
4-23
-------
-------
APPENDIX A
SUMMARY OF METHODS USED FOR
IDENTIFICATION OF STRESSORS
4-A1
-------
Summary of Methods Used for Identification of Stressors
Biological Community. Benthic macroinvertebrates were sampled in the lagoons, stream,
and river and were assessed primarily by calculating Shannon-Weiner's diversity index for the
benthic macroinvertebrates (Ballanryne et al., 1984) and enumerating the organisms. Samples were
taken from representative habitat types at each station with petite ponar grab samplers.
Fish populations were sampled hi the stream and river by seining with a 6 by 8-foot,
1/4-inch mesh seme for 30 minutes per sampling effort. The community was assessed by
calculating Carlander's condition index for fish (Carlander, 1969). The lagoons showed no
evidence of use by fish life. The histopathological analysis was conducted by determining the
percentage of incidences of liver neoplasia hi golden redhorse suckers and freshwater drum.
Phytoplankton and zooplankton were sampled in the lagoons. The number of plankton cells
per liter of water sample for the phytoplankton and zooplankton were counted. The number of
taxa and individual abundances were also listed for each of the biological communities sampled.
Plants and animals were usually identified to genus.
Fish Neoplasia Occurrences. The occurrence of tumors in bottom-feeding fish was used as
an indicator of effects related to exposure to PAH. Freshwater drum and golden redhorse sucker
were examined; the brown bullhead—known to be susceptible to PAH-inducecl neoplasia—would
have been used but these animals were not found to occur in the study area. All of the fish
collected during the study were examined for external tumors, fin rot, hemorrhaging, and parasites.
Only target fish specimens (freshwater drum and golden redhorse sucker) not showing any
observable capture-related damage were retained for histopathological examination, so external
injuries related to capture would not be confused with environmental effects.
Chemical Analysis. Surface water samples were collected mid-channel into 1-L Nalgene
bottles for metal analysis and were adjusted to pH 2-3. Samples for cyanide were adjusted to pH
11-12. Surface waters analyzed for organic priority pollutants were collected in 2-L glass jars
with Teflon-lined lids and kept at 4°C until analyses were conducted. Surface water samples
collected for classical parameters were analyzed the same day.
Sediments were collected along a representative transect of the station using a petite ponar
grab sampler. Composites of the grabs for the transects were made by manual mixing. The
samples were stored at 4°C in 1-L Nalgene bottles for'metals and in 1-L glass jars with Teflon-
lined lids for organic priority pollutants. Analyses for metals and priority organic pollutants were
conducted using EPA-approved methods.
4-A2
-------
APPENDIX B
COMPOUNDS AND PARAMETERS SAMPLED IN
SURFACE WATERS AND SEDIMENTS
4-B1
-------
Table 4-B1. Surface Waters: Toxic Pollutants (BaHantyne et al., 1984)8
Station
Zinc Cadmium Lead Arsenic Selenium Mercury Benzene Cyanide
LG1
LG2
IC-1
IC-2
IC-3
IC-4
IC-5
IC-6
OR-1
Standardb
75
77
48
26
22
87
56
34
36
>99
<0.8
1.8
1.3
0.8
<0.8
<0.8
<0.8
<0.8
<0.8
0.8-3.1
6.9
7.5
11.4
9.4
11.6
8.0
6.7
4.5
6.9
30.0
5.4
5.0
<2.0
<2.0
<2.0
3.1
<2.0
4.0
3.6
190.0
6.0
0.6
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
24.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
<1.0 <0.1 < 100.0
0.2 — 8.1
aAH units are /tg/L.
bState water quality standards based upon a 30-day average for worst-case comparisons.
Note: A " <" sign indicates the method detection limits.
Table 4-B2. Surface Waters: Classical Parameters (Ballantyne et al., 1984)
Station
LG1
LG2
IC-1
IC-2
IC-3
IC-4
IC-5
IC-6
OR-1
DO
(mg/L) pH
9.1
9.3
8.8
8.9
6.7
3.6
9.8
9.4
8.8
8.6
8.3
7.3
7.6
7.3
7.4
7.3
7.0
7.2
Temp.
Conduct.
(°C) (/tmhos/cm)
20.3
21.0
19.0
19.1
19.6
18.9
25.4
25.2
25.6
1193
741
548
562
650
618
502
438
411
Hardness
Total Ammonia
(mg/L CaCO3) (mg/L NH3-N)
544
336
252
224
288
236
180
168
124
12.5
0.86
0.003
0.04
0.101
3.10
1.6
6.4
5.5
(0.4)a
(0.6)
(3.5)
(2.4)
(3.3)
(3.1)
(2.3)
(2.9)
(2.6)
Approximate state standard at pH and temperature of sample, 30-day average for
worst-case situations.
4-B2
-------
Table 4-B3. Sediments: Polycyclic Aromatic Hydrocarbons (Ballantyne et al., 1984)
Station Numbers
Parameter4 OR-1 LG1 LG2 IC-1 IC-2 IC-3 IC-4 IC-5 IC-6
Acenaphthene
Anthracene
Benzo(a)pyrene
Chrysene
Fluoranthene
Fluorene
Naphthalene
Phenanthrene
Pyrene
NDb
ND
4.02
6.0
8.75
ND
ND
5.65
7.01
4.0
6.0
9.9
13.1
31.2
11.2
52.6
22.1
25.1
ND
21.1
90.8
90.3
140
ND
44.2
83.4
100.0
ND
ND
ND
ND
0.56
ND
ND
ND
0.44
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.7
ND
ND
ND
ND
15
.4
9.3
49
44
57
.9
.2
.1
9.7
92
38
43
.5
.7
.5
aAll units are mg/kg.
bND = not detected.
4-B3
-------
-------
APPENDIX C
RESULTS OF BIOLOGICAL SURVEYS
4-C1
-------
Table 4-C1. Benthic Macroinvertebrates Collected in Petite Ponar Grab Samples (Ballantyne
et al., 1984)
Taxon
DIPTERA
Chironomidne
Chironomus
Crytochironomus
PoJypedilum
Tanytarsus
Tanypus
Parachironomus
Psectrotanypus
Procladlus
Chaoboridae
Chaoborus
EPHEMEROPTERA
Baetlsca
Caenis
Hexagenia
ODONATA
Progomphus
Argia
Plathemis
COLEOPTERA
Elmidae
Dubiraphia
Dytiscidae
Laccophilus
OLIGOCHAETE
Tubifex
Branchiura
Lumbriculus
Dero
HIRUDINEA
Illindodella
AMPHIPODA
Gammarus
GASTROPODA
Physa
BIVALVIA
Corblcula
Sphaerium
Number of Individuals
Number of Taxa
Shannon's H
Station
IC-1 IC-2 IC-3 IC-4 IC-5 IC-6
14 45 7 1
1 1 1
1
1
843
3 1 12 2
6 1
4
1 1
1 1
4
1
1
31 238 66 71
5 59 30 3
1
1
1 1
4 1
12 2 67 351 122 84
6 295 9 6
0.68 03 0.69 0.41 058 0.34
OR-1
, 1
4
3
16
3
1
2
1
176
28
2
1
11
2
1
252
15
05
LG1 LG2
50
2
2
4
11
1
55 15
4 2
0.1 0.25
4-C2
-------
Table 4-C2. Numbers and Species of Fishes Collected (Ballantyne et al., 1984)
Station
IC-1
IC-2 IC-3
IC-4
IC-5 OR-1
Cyprinidae
Carp (Cyprinus carpio)
Silverjaw minnow (Ericymba buccata)
Emerald shiner (Notropis antherinoides) .
Steelcolor shiner (Notropis whipplei)
Bluntnose minnow (Pimephales notatus)
Percidae
Fantail darter (Etheostoma flabellare)
Johnny darter (Etheostoma nigrum)
Yellow perch (Percaflavescens)
Centrarchidae
Northern rock bass (Ambhplites rupestris)
Warmouth sunfish (Chaenobryttus gulosus)
Green sunfish (Lepomis cyanettus)
Bluegill (Lepomis macrochirus)
Longear sunfish (Lepomis megalotis)
Longear sunfish x Bluegill hybrid
Longear sunfish x Green sunfish hybrid
Green sunfish x Bluegill hybrid
Northern largemouth black bass
(Micropteus sabnoides)
White crappie (Pomoxis annularis)
Black crappie (Pomoxis nigromaculatus)
Pumpkinseed sunfish (Lepomis gibbosus)
Catostomidae
Quillback carpsucker (Carpiodes cyprinus)
Creek chubsucker (Erimyzon oblongus)
Clupeidae
Gizzard shad (Dorosoma cepedianum)
Skipjack herring (Pomolobus chrysochloris)
Sciaenidae
Freshwater drum (Aplodinotus chrysochloris)
Ictaluridae
Channel catfish (Ictalurus punctatus)
Flathead catfish (Pilodictis oUyaris)
Amphibia
Central mudpuppy (Necturus maculosus)
26
30
1
27
2
11
1
4
52
6
1
1
62
38
34
13
15
11
290
25
3
12
31
10
23
21
2 4
16 45
6 18
1
1
45
7
5
2
1
13
4
10
no
15
26
34
2
34
1
1
2
7
1
112
24
10
24
3
9
4
115 15 80 100 31
2
Total Number of Species
Total Number of Individuals
13
168
12
308
10
395
15
209
16
365
14
293
4-C3
-------
Table 4-C3. Condition Factors Calculated for Fish (Carlander, 1969)
Station
Species
Condition
Factor
Normal
Range
Range
IC-2 Largemouth bass 5.2 — 4.6-5.5
IC-3 Bluegill 5.8 5.6-6.0 7.1-8.0
IC-5 Bluegill 7.1 6.4-8.2 7.1-8.0
Crappie 5.0 — 4.6-5.5
Carp 1.5 1.49-1.51 1.2-2.9
IC-6 Bluegill 7.0 6.0-7.7 7.1-8.0
Largemouth bass 5.0 — 4.6-5.5
Flathead catfish 1.0 — 0.9-1.1
Gizzard shad 1.6 1.2-2.0 0.9-2.2
OR-1 Bluegill 7.7 5.7-11.0 7.1-8.0
Largemouth bass 5.4 4.7-6.0 4.6-5.5
Channel catfish 0.98 — 0.8-1.2
4-C4
-------
Table 4-C4. Plankton Collected in Lagoons 3 and 4 (Ballantyne et al., 1984)
Station
Phytoplanktona
Chlorophyta
ChlorogoTuum
Rhizoclonium
Utothrix
Cyanophyta
Anabena
Microchaete
Nostoc
Spirulina
Chrysophyta
Dipolneis
Fragilaria
Gomphonema
Melosira
Oscillatoria
Euglenophyta
Euglena
LG1
1,460
2,230
S.VxlO6
—
208
3,300
208
208
625
—
2,910
208
LG2
1,670
3.12X104
416
6,250
—
416
625
1,040
830
1,250
416
Zooplanktona
Copepoda
Cyclops
Rotifera
Pleurotrocha
Pofyartha
Keratella
Asplanchna
Ciliophora
Bursaria
Station
LG1 LG2
52 34
170
32
4
6
46
aAll units are cells/L.
4-C5
-------
-------
SECTION FIVE
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
COMMENCEMENT BAY TIDELANDS ASSESSMENT
-------
AUTHORS AND REVIEWERS
AUTHORS
Patricia A. Cirone
Region 10
U.S. Environmental Protection Agency
Seattle, WA
REVIEWERS
Donald P. Weston (Lead Reviewer)
Horn Point Environmental Laboratory
University of Maryland
Cambridge, MD
James J. Anderson
School of Fisheries Research
University of Washington
Seattle, WA
Jay S. Jacobson
Boyce Thompson Institute
Cornell University
Ithaca, NY
Larry Kapustka
Ecological Planning and Toxicology, Inc.
CorvallSs, OR
Robert A. Pastorak
PTI Environmental Services
Bellevue, WA
Wayne G. Landis
Institute of Environmental Toxicology
and Chemistry
Western Washington University
Bellingham, WA
Thomas Sibley
School of Fisheries Research
University of Washington
Seattle, WA
Frieda B. Taub
School of Fisheries Research
University of Washington
Seattle, WA
5-2
-------
CONTENTS
ABSTRACT 5-6
5.1. RISK ASSESSMENT APPROACH 5-7
5.2. STATUTORY AND REGULATORY BACKGROUND 5-7
5.3. CASE STUDY DESCRIPTION 5-9
5.3.1. Problem Formulation 5-9
5.3.2. Analysis: Characterization of Ecological Effects 5-14
5.3.3. Analysis: Characterization of Exposure 5-21
5.3.4. Risk Characterization 5-22
5.4. REFERENCES 5-31
5-3
-------
LIST OF FIGURES
Figure 5-1. Structure of analysis for Commencement Bay, Washington 5-8
Figure 5-2. Commencement Bay nearshore/tidefiats study area 5-10
Figure 5-3. The AET approach applied to sediments tested for lead and
4-methylphenol concentrations and toxicity response during bioassays .... 5-16
Figure 5-4. Bioassay responses and abundances of benthic macroinvertebrate
taxa near the four major contaminant sources evaluated
in Commencement Bay 5-20
Figure 5-5. Definition and prioritization of Commencement Bay problem areas 5-28
LIST OF TABLES
Table 5-1. Summary of Chemicals With Sediment Elevations Above
Reference (EAR) Between 100 and l,000x and Greater Than
1,000 Averaged Over Commencement Bay 5-11
Table 5-2. Action Assessment Matrix of Sediment Contamination,
Sediment Toxicity, and Biological Effects Indices
for Commencement Bay Study Areas 5-23
Table 5-3. Action-Level Guidelines 5-24
Table 5-4. Sediment Cleanup Objectives Related to Environmental Risks 5-26
5-4
-------
LIST OF ACRONYMS
AET Apparent Effects Threshold
CERCLA Comprehensive Environmental Response, Conservation, and Liability Act of 1980
EP Equilibrium Partitioning
EPA U.S. Environmental Protection Agency
MLLW mean lower low water
NCP National Contingency Plan
PAH polycyclic aromatic hydrocarbon
PCB polychlorinated biphenyl
SARA Superfund Amendments and Reauthorization Act of 1986
5-5
-------
ABSTRACT
During an ecological assessment for the Commencement Bay, Washington, nearshore/
tideflats area, field studies were designed to document the extent of sediment contamination and
adverse biological effects, including sediment toxicity, alterations of benthic macroinvertebrate
assemblages, chemical residues in tissues of crab and English sole, and liver lesions in English
sole. During the Superfund remedial investigation for the site, indices of biological effects and
sediment quality values were used to identify and prioritize problem areas for possible source
control and/or sediment remedial action. The multi-indicator approach, based on chemical and
biological variables, provides a powerful weight of evidence for ecological assessment. In the
absence of biological data, sediment quality values may be used to interpret historical sediment
chemistry data to predict the occurrence of adverse biological effects. The approach could be
improved by defining assessment endpoints more explicitly, validating sediment quality values, and
incorporating a probabilistic approach to exposure and risk assessment.
5-6
-------
5.1. RISK ASSESSMENT APPROACH
The Commencement Bay (Puget Sound, Washington) case study combines a retrospective
assessment (i.e., determination of effects that have already occurred) based on indices of biological
effects with a predictive assessment of potential biological effects based on site-specific sediment,
quality criteria (figure 5-1). The study approach was based on three premises: (1) site-specific
field data were needed to establish cleanup goals, (2) no single chemical or biological indicator
could be used to define areas at risk, and (3) adverse biological effects were linked to sediment
contamination and chemical-biological relationships could be characterized empirically.
The primary approach to identifying and ranking problem areas in the Commencement Bay
ecosystem relied on direct measurements of sediment chemistry, sediment toxicity (i.e., bioassays),
benthic macroinvertebrate abundances, concentrations of contaminants in English sole and crab,
and prevalence of liver lesions in English sole. The amphipod mortality bioassay and oyster larvae
developmental abnormality bioassay were used to characterize sediment toxicity during the
remedial investigation. Microtox bioassays of sediments were performed to provide ancillary data,
but Microtox response was not one of the indices of biological effects used in the remedial
investigation.
Because recent historical data on sediment contamination were available at additional
stations throughout the study area where biological data had not been collected, the Apparent
Effects Threshold (AET) approach was developed to predict the presence or absence of specific
biological effects based on chemical data alone. The derivation of the AET is explained in section
5.3.2., Analysis: Characterization of Ecological Effects. However, these empirical thresholds for
chemical concentrations associated with biological effects do not establish cause and effect
relationships. Exceedances of AET values were used to identify problem sediments and to identify
and rank problem chemicals during source evaluation.
5.2. STATUTORY AND REGULATORY BACKGROUND
Concerns about the potential ecological and human health effects of hazardous substances in
sediments of the nearshore area of Commencement Bay led to the addition of the Commencement
Bay nearshore/tideflats area to the National Priorities List of Comprehensive Environmental
Response, Conservation, and Liability Act (CERCLA) sites on September 8, 1983. CERCLA, as
amended by the Superfund Amendments and Reauthorization Act (SARA) of 1986, requires the
U.S. Environmental Protection Agency (EPA) to ensure the environment is protected when (1)
remedial alternatives are selected and (2) the degree of cleanup needed is assessed. As mandated
under CERCLA and the National Contingency Plan (NCP), a Remedial Investigation and
Feasibility Study is required to define risks to public health and the environment. The major focus
of the remedial investigation for the Commencement Bay Superfund site was characterizing impacts
to aquatic organisms of exposure to contaminated marine sediments. The feasibility study screened
remediation alternatives for their effectiveness in reducing risks. In the absence of regulatory
standards or guidelines for establishing cleanup criteria for contaminated sediments, a decision-
making approach based on chemical and biological indicators and sediment quality objectives was
developed specifically for the Commencement Bay nearshore/tideflats investigations.
5-7
-------
Figure 5-1 Structure of Analysis for
Commencement Bay, Washington
PROBLEM FORMULATION
Stressors-. complex mixture of organic and
inorganic chemicals in marine sediments.
Ecological Components: benthic macroinvertebrates and
fish.
Endpoints: assessment endpoints are the health
and condition of selected ecological components. Multiple
measurement endpoints were used at different levels of
biological organization.
ANALYSIS
Characterization of
Exposure
Exposure was evaluated by
measuring concentrations
of chemicals in sediments.
A model was used to
predict natural recovery.
Characterization of
Ecological Effects
Effects were evaluated by
examining benthic
abundance, occurrence of
liver abnormalities in fish,
and various measures of
sediment toxicity.
RISK CHARACTERIZATION
Risks were estimated using two basic methods:
1. comparison of conditions at contaminated sites to
benchmarks or reference locations.
2. application of apparent effects threshold (AET) values
for chemical concentrations in sediments.
Study areas were ranked with regard to potential risk
or impact.
5-8
-------
In 1989, a Record of Decision was signed that presented the remediation actions for the
Commencement Bay nearshore/tideflats Superfund site. The site assessment in the Remedial
Investigation and Feasibility Study concluded that actual or threatened releases of hazardous
substances from this site, if not corrected by response actions, present an imminent and substantial
endangerment to public health, welfare, or the environment.
5.3. CASE STUDY DESCRIPTION
5.3.1. Problem Formulation
Site Description. The study area is described in U.S. EPA (1985). The Commencement
Bay nearshore/tideflats study site is located in a heavily industrialized area at the southern end of
the main basin of Puget Sound (figure 5-2). The tideflats area, formed by the Puyallup River
delta, comprises seven waterways, associated shoreline, and waters of depths less than 60 feet
below mean lower low water (MLLW). Various industrial and municipal sources are located on
filled areas of the tideflats, including a pulp mill, petroleum refineries and storage facilities,
chemical manufacturers, aluminum processors, a shipbuilding/repair yard, and numerous storm
drains. A municipal sewage treatment plant discharges into the Puyallup River immediately
upstream of the tideflats area. The nearshore portion of the site is northwest of the tideflats,
including waters of depths less than 60 feet below MLLW. The city of Tacoma and a major
copper smelter (now closed) are located within the nearshore area. Contaminants in the waterways
or sediments of the nearshore area also may have originated from drainage associated with creeks,
the Puyallup River, seeps, and open channels, or nonpoint sources such as spills and atmospheric
deposition.
Stressors. Toxic contaminants have been identified in sediments throughout the
Commencement Bay study area (Malins et al., 1980; Washington Department of Ecology/EPA,
1985; Becker et al., 1990). Chemical analyses were completed for over 190 samples of surface
and subsurface sediments collected from intertidal and subtidal areas of the Commencement Bay
Superfund site (summarized in U.S. EPA, 1985). Routine analyses were conducted for about 150
chemicals. Chemicals detected in more than two-thirds of the surface sediments include phenol,
4-methylphenol, polycyclic aromatic hydrocarbons (PAHs), 1,4-dichlorobenzene, polychlorinated
biphenyls (PCBs), dibenzofuran, and metals. The chemicals of concern based on sediment
chemistry include 8 metals and 18 organic compounds. Chemicals of concern were selected if
their concentrations in Commencement Bay Superfund site sediments exceeded the range of
reference concentrations for Puget Sound. Concentrations of several contaminants in the study area
sediments are substantially elevated above those characteristic of reference areas (table 5-1).
Bioaccumulation of toxic substances and associated abnormalities observed in indigenous
fish and crabs found in Commencement Bay (Malins et al., 1980; Washington Department of
Ecology/EPA, 1985) led to the conclusion that the contaminants were potentially hazardous to
biota. In particular, PCBs were detected in muscle and liver tissues of English sole (Parophys
vetulus) throughout the study area at concentrations substantially elevated above those found in
reference areas. Other contaminants, especially PAHs, are not frequently detected in tissue
samples because they are readily metabolized by fish and crabs, but PAH was suspected as a cause
of the observed liver abnormalities in English sole. Alterations of benthic macroinvertebrate
5-9
-------
I
^—
o
Commencement Bay
Nearshore/Tsdeflats
Problem Areas
\T JJTJ r '• I
W$fwr*s<(*8Z,> s JJ t> ;*V^T* ••'•.•>
Area Shown on
the Location Map
Problem Areas
Huston Shoreline
Mouth of Ctty
3J Head of Ctty
4) Wheeler-Osgood
Middle
B! StPats!
T) Sitcum
?) Mouth of Hylebos
Head of Hylebos
Figure 5-2. Commencement Bay nearshore/tideflats study area (U.S. EPA, 1989b)
-------
Table 5-1. Summary of Chemicals With Sediment Elevations Above Reference (EAR) Between
100 and l,000x and Greater Than 1,000 Averaged Over Commencement Bay
(Washington Department of EcologyAJ.S. EPA, 1985)
Chemicals >100x and <1.000x Reference
Aromatic hydrocarbons (4-6 rings)
Aromatic hydrocarbons (1-3 rings)
Bis(2-ethylhexy)phthalate
Isopirnaradiene
Kaur-16-ene (tentative identification)
l-Methyl-2-(l-methylethyl)benzene
2-Methylnaphthalene
1-Methylpyrene
2-Methylpyrene
2-Methoxyphenol
Total chlorinated butadienes
Total PCBs
Antimony
Arsenic
Copper
Lead
Mercury
Chemicals Exceeding l.OOOx Reference
Benzo(a)pyrene
4-Methylphenol
2-Methyoxyphenol
Phenanthrene
Trichlorobutadienes
Tetrachlorobutadienes
Antimony
Arsenic
Copper
Mercury
Waterway. Segment, or Station
Hylebos Waterway
City Waterway
Ruston Shore
City Waterway
Middle Waterway
Ruston Shore
City Waterway
Inner Hylebos Waterway
Sitcum Waterway
St. Paul Waterway
Inner City Waterway
Inner City Waterway
Hylebos Waterway
Sitcum Waterway
Milwaukee Waterway
St. Paul Waterway
Middle Waterway
City Waterway
St. Paul Waterway
City Waterway
Middle Waterway
Inner Hylebos Waterway
Inner Hylebos Waterway
Sitcum Waterway
Milwaukee Waterway
Middle Waterway
City Waterway
Hylebos Waterway
Hylebos Waterway
Ruston-Pt. Defiance.
Ruston-Pt. Defiance
Ruston-Pt. Defiance
Ruston-Pt. Defiance
Ruston-Pt. Defiance
Hylebos Waterway
St. Paul Waterway
St. Paul Waterway
Ruston Shore
Hylebos Waterway
Hylebos Waterway
Ruston Shore
Ruston Shore
Ruston Shore
Ruston Shore
5-11
-------
assemblages also were associated with sediment contamination in selected areas of the
nearshore/tideflats zone (Becker et al., 1990).
Ecological Components. Sediments of the study area support a diverse assemblage of
benthic organisms that can be directly influenced by sediment contamination. Toxicity may be
acute or chronic depending on the contaminant, its concentration, and the sensitivity of the
component. Many fish and crab species that live in close association with the sediment also feed
on benthic organisms and are exposed to contaminants through the food chain. Surrogates for
various groups of organisms including fish, crustaceans, and bivalves were represented in the
measurement endpoints as environmental indicators of the populations or communities at risk.
Benthic macroinvertebrates. Benthic macroinvertebrates are an integral part of the Puget
Sound estuarine ecosystem. Many benthic macroinvertebrate species are sedentary and consume
organic materials associated with sediments. Because of their direct interaction with sediments and
their sensitivity to organic enrichment and chemical contamination, they are excellent indicators of
the areal extent and magnitude of environmental stress.
In the Commencement Bay study, 407 species of benthic macroinvertebrates were collected
(Washington Department of Ecology/EPA, 1985). The major taxonomic groups were Polychaeta
(marine worms), Bivalvia (clams), Nematoda (round worms), Crustacea (e.g., amphipods and
cumaceans), Echinodermata (e.g., sea cucumbers and brittle stars), Oligochaeta (e.g., tubificid
worms), and Sipuncula (marine worms). Two species (the polychaete Tharyx multifilis and the
bivalve mollusc Axinopsida serricatd) accounted for 59 percent of the benthic macroinvertebrates
collected. These two species and an associated assemblage of species characterized much of the
waterway system. Nematodes were the third most abundant group overall because of their high
densities at a few stations. Crustaceans such as the ostracods Euphilomedes producta and E,
carcharodonta and the tanaid shrimp Leptochelia dubia were also abundant. Commercially
harvestable species of bivalve shellfish are not found in the Commencement Bay waterways.
Fishery resources. Commencement Bay supports important fishery resources. Four
salmonid species (Chinook, coho, chum, and pink) and steelhead inhabit Commencement Bay for
part of their life cycle. These anadromous fish have critical estuarine migratory and rearing habitat
requirements. Adults pass through the bay en route to their spawning grounds, and juveniles reside
in nearshore habitats. Recreational and commercial harvesting of these species occurs in the bay.
Inshore marine fish resources, which include flatfish such as English sole, rock sole, flathead sole,
c-o sole, sand sole, starry flounder, and speckled sanddab, are the most abundant within the
waterways. Rock sole, c-o sole, and special species of rockfish are most abundant along the outer
shoreline.
Endpoints. The assessment endpoint in this program was the health and condition of
selected components (benthic invertebrates and fish) with regard to contaminated sediments. Two
indices were used as a basis for this assessment: (1) biological effects and (2) AET values. Each
involves elements of both exposure and effects.
Multiple measurement endpoints at different levels of biological organization were
evaluated. These endpoints included organism-level responses in sediment toxicity bioassays,
5-12
-------
population abundances of benthic macroinvertebrate species, community indices (e.g., species
richness and community similarity), and biomarkers (i.e., tissue residues of contaminants and
histopathology).
Sediment toxicity bioassays. Whole sediment toxicity was measured in the laboratory
based on the amphipod, oyster larvae, and Microtox (saline extract) bioassays. The amphipod
bioassay is an important indicator because it measures acute lethality in a crustacean species
(Rhepoxynius dbronius) that resides hi the study area and is an important prey item for higher
trophic-level species, especially various fishes (Swartz et al., 1985; PSEP, 1986). Use of
R. abronius to determine the acute lethality of field-collected sediments has been documented by
Swartz et al. (1982, 1985), Chapman et al. (1982a, b), and Chapman and Fink (1984).
Amphipods are relatively sensitive to toxic chemicals and are highly likely to be exposed to
paniculate contaminants because they burrow hi and feed on sediment material. The oyster larvae
test measures the prevalence of developmental abnormalities in larvae exposed to sediments for 48
hours (Chapman et al., 1982b; PSEP, 1986). The oyster Crassostrea gigas resides in Puget
Sound, although it is not found in the study area. The life stages tested (embryo and larva) are
also very sensitive to toxic chemicals. The primary endpoint represents a sublethal effect that may
reduce survival of larvae or affect population recruitment. The Microtox bioassay is an acute test
that measures the reduction in luminescence of bacteria exposed to an extract of sediment (PSEP,
1986). Bacteria play key roles in ecosystems as decomposers and primary species in detrital-based
food webs. Moreover, the Microtox test is a sensitive indicator of the effects of toxic chemicals on
oxidative enzyme systems common among diverse taxa.
Benthic macroinvertebrates. The abundances of benthic macroinvertebrate species were
determined from field-collected samples. Benthic macroinvertebrates are valuable indicators
because they live in direct contact with sediments, are relatively stationary, and are important
components of estuarine ecosystems. If sediment-associated impacts are not detected by analyses
of benthic macroinvertebrates, then it is unlikely that similar population-level impacts are present in
other biotic groups such as fish or plankton.
Changes in benthic macroinvertebrate assemblages were based on community-level
endpoints. Community indices included the relative abundances of major taxa (i.e., polychaetes,
crustaceans, molluscs, total benthos); species richness; and Bray-Curtis similarity. Only decreases
in abundances of major taxa relative to reference area values were used to identify and rank
problem areas.
Bioaccumulation. Contaminant concentrations in muscle tissue of English sole and
Dungeness crab (Cancer magister) were measured as an indicator of exposure. Only the English
sole data were used to identify and rank problem areas because contaminants were detected
relatively infrequently in the crab muscle tissue.
Histopathology. Histopathological analyses were conducted on the livers of English sole
(summarized in U.S. EPA, 1985). The prevalence of all identifiable lesions was determined. The
prevalence of major lesions (i.e., preneoplastic nodules, megalocytic hepatosis, nuclear
polymorphisms, and neoplasms) was the primary indicator used to identify and rank problem areas.
These biomarkers have been associated with exposure to toxic chemicals, particularly PAH
5-13
-------
compounds, but causal relationships were not firmly established during the Commencement Bay
Superfund investigations. Liver lesions are not definitely known to result in adverse effects on
organism survival or reproduction.
Comments on Problem Formulation
Strengths of the case study include:
+The extensive sediment chemical analyses and care in their quality assurance
provide a high degree of confidence in this data set. Table 5-1 provides a good
summary of the chemicals that served as the focus for much of the remedial
investigation.
*A number of ecological components and endpoints are used to quantify impact.
These include several bioassay species, benthic community composition, and fish
histopathology. It is also noteworthy that all these ecological components can be
justified on ecological grounds. The investigators avoided limiting ecological
components of concern to commercially important species or to those selected for the
sake of political expediency.
5.3.2. Analysis: Characterization of Ecological Effects
Characterization of ecological effects for the Commencement Bay Superfund project relied
mainly on statistical comparisons of study sites with reference areas to define significant (P<0.05)
biological effects. Two reference areas (Carr Inlet and Blair Waterway) were used for
comparisons with contaminated sites in the Commencement Bay study area. The relationships
between exposure and effects for single chemicals could not be developed because complex
mixtures of chemicals were found in sediments of contaminated areas. The two methods developed
for characterizing effects—Indices of Biological Effects and AET—are described below followed by
an overview of the program design. The two methods incorporate elements of both exposure and
effects. As such, they also may be used directly within the risk characterization.
Indices of Biological Effects. A series of indices was developed based, on the magnitude of
observed contamination (i.e., sediment contamination) and biological effects as determined by the
measurement endpoints (sediment toxicity, benthic macroinvertebrate, bioaccumulation, and
histopathology variables). The indices have the general form of a ratio between the value of an
effect at a Commencement Bay site and the value at a reference site. The ratios are structured so
that the value of the index increases as the deviation from background conditions increases. Each
ratio is termed an "elevation above reference" index. These indices are not used in lieu of the
original data (e.g., contaminant concentrations), but are considered complementary forms of
information. The original data are used to evaluate whether there are statistically detectable
increases in contamination or effect variables and to evaluate quantitative relationships among these
5-14
-------
variables. The indices are used to reduce large data sets into interpretable numbers that reflect the
different levels of contamination and effects among subareas. A matrix is constructed to integrate
the individual indices in an overall evaluation and prioritization of problem areas.
Apparent Effects Threshold. Because biological effects data were not available for all
portions of the study area where chemical data were available, a method was developed to estimate
threshold concentrations of contaminants above which biological effects would be expected. The
data base generated for each of three site-specific biological indicators (i.e., benthic
macroinvertebrate abundances, amphipod mortality bioassay, and oyster larvae abnormality
bioassay) was used to develop these AET values. These three indicators were selected because of
their sensitivity to sediment contamination, availability of standard protocols, and ecological
relevance. The AET also can be established for biological indicators that reflect areawide
conditions (i.e., over multiple sediment stations) such as bioaccumulation and histopathology in
fish, but the uncertainty in determining exposure area concentrations for areawide indicators is
relatively high. AET values are compared with measured concentrations of sediment contaminants
in a predictive method to determine the potential risk to aquatic organisms.
An AET for a chemical is defined as the concentration in sediments above which
statistically significant biological effects (relative to reference sediments) would always be expected
(Barrick, 1985; Chapman et al., 1982b). The AET approach uses matched (i.e., synoptically
collected) data on sediment chemistry, sediment toxicity bioassays, and benthic macroinvertebrate
effects (figure 5-3). To derive an AET value, sampling stations are arranged in a sequence
according to the concentration of the chemical for which the AET is being determined (figure 5-3).
Next, adverse effects are defined for a given biological endpoint as a statistically significant
difference (P<0.05) between conditions in a study area relative to conditions in an appropriate
reference area. Stations that exhibit adverse effects are identified. The AET value is set by the
no-effect station with the highest chemical concentration (i.e., all stations with chemical
concentrations above the AET showed significant biological effects for the given endpoint).
To apply a set of AET values, if any chemical exceeds its AET for a particular biological
indicator, then an adverse biological effect is predicted for that indicator. If all chemical
concentrations are below their respective AET for a particular biological indicator, then a lack of
impact is predicted for that indicator. Thus, the potential for adverse ecological effects is assessed
essentially by a quotient method. The AET method does not include a probabilistic estimate of
risk. Moreover, the AET approach does not prove cause-effect relationships between contaminants
and effects. Nevertheless, AET values may account for some interactive effects of chemicals and
for unmeasured chemicals that vary with quantifiable contaminants.
Since the development of the AET approach in the Commencement Bay Superfund project,
AET values have been developed for 64 organic chemicals and metals in Puget Sound and for 4
separate biological indicators (amphipod, oyster larvae, Microtox bioassays, and benthic
macroinvertebrate abundances). When applied as a set of sediment-quality screening criteria to
independent data, these AET values have displayed a high reliability in predicting biological
effects, while maintaining a low rate of false positives (Barrick and Seller, 1989). The AET
approach is most predictive when applied to a large data base with a wide diversity of chemical
5-15
-------
LEAD
NO SEDIMENT TOXICITY
t» • *»• • •• •
• •••{••••••OTMMM* • «0»« *•»•***•*
• • >>•« • • • • •
SEDIMENT TOXICITY OBSERVED
\
SP-15
SP-14
t
RS-19
700 ppm
HS-181
I
15300 ppm
I
10
1 M
100
t
1000
10,000
CONCENTRATION (mg/Kg DW)
THRESHOLD
MAXIMUM •
OBSERVED
LEVEL AT A
BIOLOGICAL
STATION
4-METHYLPHENOL
-NO SEDIMENT TOXICITY
-SEDIMENT TOXICITY OBSERVED
RS-19
RS-18
1200
t
SP-15
ppb
|!
SP-14J
I
I
U10
I
100
' I <_
1000
10.000
APPARENT '
CONCENTRATION (ng/Kg DW) TOXICITY
THRESHOLD
U - undetected al detection limit shown
98.000
MAXIMUM J
OBSERVED
LEVEL AT A
BIOLOGICAL
STATION
Figure 5-3. The AET approach applied to sediments tested for lead and 4-methylphenol
concentrations and toxicity response during biossays (U.S. EPA, 1985)
5-16
-------
contaminants, each represented by a broad range of concentrations. The use of 30 to 50 stations
for each contaminant of concern is recommended to reduce the uncertainty associated with tests for
statistical significance on small data bases.
Selection of Reference Areas. Carr Inlet was selected as a reference area because:
• a complete data set was available for Carr Inlet, including synoptic data for metals,
organics, grain size, organic carbon content, and other conventional
variables; and
• the lowest detection limits for most substances of concern in Puget
Sound embayments were available for Carr Inlet.
The range of sediment types represented by the Carr Inlet stations sampled did not,
however, encompass many of the fine-grained sediments characteristic of the Commencement Bay
waterways. Therefore, it was necessary to define Blair Waterway as an internal (or nearfield)
reference area. Blair Waterway was selected as a reference area because:
• it was the least chemically contaminated of the seven waterways;
• only one significant bioassay result (amphipod mortality) was found
from the 12 stations tested in Blair Waterway; and
• fine-grained sediments at Blair Waterway stations spanned a range
(37 to 84 percent) similar to that observed for all waterway stations
in Commencement Bay except Hylebos Waterway.
Sediment Toxicity. The amphipod mortality and oyster larvae developmental bioassays
were conducted on intact sediment samples (i.e., nondilution tests) and on sediment samples diluted
with clean sediments. In the present study, exposure to sediments from 18 of the 52 stations tested
induced statistically significant (P<0.05) acute lethality in the amphipod R. abronius as compared
with a reference area sediment. Statistically significantly elevated (P<0.05) oyster larvae
abnormalities were observed at 15 of the 52 stations when compared with the reference station.
Sediments from 24 of the 52 stations tested had statistically significant (P<0.05) toxicities
in either one or both of the amphipod and oyster larvae bioassays. Sediments from 10 stations
were toxic in both bioassays. In some areas, sediments were toxic to the extent that a 90 percent
dilution was needed to reduce amphipod toxicity responses to reference values. The level of
agreement (43 percent) between lethal toxic effects for an adult organism (amphipod test) and
sublethal toxic effects for a fertilized egg (oyster larvae test) enhances the weight of evidence
supporting toxicity of the sediment contaminants.
Ancillary data for Microtox bioassay response were also collected. Although the Microtox
data were not used as part of the ecological assessment approach for the Commencement Bay
remedial investigation (Washington Department of Ecology/U.S. EPA, 1985), Williams et al.
(1986) showed a significant overall concordance (Kendall's coefficient of concordance = 0.64,
5-17
-------
among the amphipod, oyster larvae, and Microtox bioassays. Comparisons of the
quantitative responses for the three bioassays, however, revealed substantial heterogeneity,
indicating the value of a diversity of toxicity tests in wide-scale surveys of sediment contamination.
Benthic Macroinvertebrates. To develop indices of benthic degradation as decision criteria,
abundances of major benthic invertebrate taxa at potentially affected sites were compared
statistically with invertebrate abundances at reference sites. A statistically significant decrease
(P<0.05) in the abundance of a major taxon was considered a benthic impact. At each station,
indices were based on the abundance of the total assemblage (i.e., total taxa) and the abundance of
polychaetes, molluscs, and crustaceans. Only depressions in abundance were defined as adverse
impacts because an increase in abundance of one or more major taxa without a corresponding
significant (P<0.05) decrease in other taxa was not considered to be an adverse response at the
community level.
Significant benthic effects were observed at 18 of 50 stations sampled for benthic
macroinvertebrates. Benthic macroinvertebrate assemblages in Commencement Bay were found to
be distinct from assemblages at reference stations, as evidenced by reduced numbers of species,
high dominance, and enhanced total abundances within waterways (summarized in U.S. EPA,
1985). Dominance was evaluated as the proportion of total abundance represented by the five
numerically dominant taxa in each area. Dominance ranged from 63 percent to 95 percent within
the waterways, but was only 36 percent along the northwestern nearshore area and only 44 percent
in Carr Inlet. The overall high abundances of a mixed polychaete-mollusc assemblage indicated
that effects to benthic communities were localized. Areas having depressed abundances of at least
two major taxonomic groups were limited to discrete areas in the waterways and one station near
the copper smelter located along the northwestern nearshore area. The organic content of the
sediments appeared to account for a considerable amount of faunal variation. In some areas (e.g.,
the head of City Waterway near a major storm drain), organic enrichment of the sediments was
attributable to anthropogenic sources.
Bioaccumulation. Concentrations of metals in English sole muscle tissues were relatively
homogeneous across locations in the study area (summarized in U.S. EPA, 1985). The maximum
concentrations of most metals in fish were less than two times the average reference
concentrations, but concentrations of copper in fish tissue were significantly elevated (3 to 9 times)
in fish from several stations. Concentrations of lead and mercury were elevated in Dungeness
crabs. Maximum concentrations of these metals were about 5 times the reference concentrations.
PCBs were detected in all fish and crabs sampled. Maximum concentrations of PCBs in English
sole muscle tissue exceeded reference concentrations by an order of magnitude.
Histopathology. The histopathological analyses indicated that the prevalence of liver
abnormalities (e.g., preneoplastic nodules, megalocytic hepatosis, and nuclear polymorphisms) was
significantly elevated (P<0.05) in English sole collected from the Commencement Bay Superfund
site compared with those from the Carr Inlet reference stations. The incidence of liver lesions was
greatest in fish from areas with the highest concentrations of sediment-associated contaminants.
The effects of these lesions on the fish are unknown. In this study, fish with serious liver lesions
did not exhibit reduced condition (weight at a given length) when compared with fish without
lesions.
5-18
-------
Exposure-Effects Relationships and AET. Single-chemical relationships between exposure
and effects could not be established from the field studies because organisms were exposed to
complex mixtures of chemicals. A field-based method such as the AET approach incorporates the
net effects of a variety of factors including interactive effects of chemicals, unmeasured chemicals
or stressors, matrix effects, and bioavailability. Presumably, the AET value represents a threshold
point on the concentration-response curve, above which significant effects (P<0.05) are observed,
but the quantification of a concentration-response relationship for a given chemical is confounded
by the effects of a complex mixture in the field.
While traditional exposure-response curves for individual chemicals cannot be easily
developed from field data, trends in the magnitude of biological response variables relative to
distance from a source may be used to derive an in situ relationship between exposure and effects.
In the Commencement Bay study, as the abundances of major benthic taxa increased with
increasing distance from the four major sources of contamination, the toxicity response variables in
the three sediment bioassays generally declined along the same spatial gradients (figure 5-4, from
Becker et al., 1990). In these source areas, spatial gradients of contamination that were defined
independently based on the distributions of contaminants corresponded to the gradients in biological
responses.
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
*The characterization of ecological effects uses multiple indicators of biological
impacts. Ecological effects are evaluated by three independent bioassays as well as
benthic invertebrate analysis. Concordance of all these measures provides the basis for
identifying and quantifying the potential effects.
Limitations include:
*The selection of an appropriate reference area is important to any risk assessment and
essential to the Commencement Bay study because all biological and chemical measures
are expressed relative to reference sites. Some limitations are associated with the
reference areas selected. The Carr Inlet sites differ in sediment type and the later
selection of Blair Waterway appears to be an a posteriori attempt to salvage the
reference area concept. Blair Waterway may be the least polluted area of those
studied, but it is hardly a pristine environment unaltered by the urbanization and
industrialization of the ndeflats.
^Degradation of benthic communities is characterized only by a decrease in the
abundance of total amphipods, molluscs, polychaetes, or total macrofauna. While some
species may decrease in abundance due to pollution, more pollution-tolerant species are
likely to increase, making changes in abundance at a major taxon level an insensitive
indicator. Defining degradation only by a decrease in abundance is particularly weak
for polychaetes and, by inference, total macrofauna.
5-19
-------
UJ
o
cc
UJ
UI
o
cc
o
cc
Ul
Ul
o
<
D
z
OQ
(3
O
—*— POLYCHAETES
•••a— MOLLUSCS
—S7— CRUSTACEANS
4 -i
2-
HY-22
HY-24
2-
LJ 'I
SPT14 SP-15
spiie
4-i
2-
CI-11
Cl-13 CI-17
STATION
INCREASING DISTANCE FF1OM SOURCE -
RS-18 RS-19
RS-20
figure 5-4. Bioassay responses and abundances of benthic macroinvertebrate taxa near the four
major contaminant sources evaluated in Commencement Bay (Becker et al., 1990)
5-20
-------
5.3.3. Analysis: Characterization of Exposure
The study of environmental effects in Commencement Bay focused on exposure to
contaminated sediments. The data on water column concentrations of contaminants were not used
to estimate risks to aquatic organisms because of the potential for high spatial and temporal
variability, the effects of environmental factors such as currents or salinity, and the association of
most of the contaminants of concern with particulate matter that was deposited in the sediments.
Station Locations. After a review of historical information and data from the preliminary
survey (summarized in U.S. EPA, 1985), sediment stations in the study area were selected for
sample collection and analysis. Stations were selected to:
• fill data gaps;
• define known areas of contamination more precisely; and
• determine gradients of contamination in relation to suspected
sources.
Magnitude. Frequency, and Duration of Exposure. The magnitude of exposure is
determined by the concentrations of contaminants in sediments. Sediments throughout the
Commencement Bay study area contain concentrations of one or more toxic contaminants that
exceed levels commonly found in reference areas. Table 5-1 lists chemicals or chemical groups
with sediment concentrations that were elevated between 100 and 1,000 times reference conditions
at one or more stations and chemicals that had concentrations that were 1,000 times greater than
reference area concentrations. Because the sediments represent a sink for contaminants, exposure
is essentially continuous for organisms that reside in or on the sediments. Water-column species
may be periodically exposed to contaminants by resuspension of bottom sediments. The frequency
of sediment resuspension events was not characterized in the Commencement Bay project.
Natural recovery of contaminated sediments is the process whereby the contamination in the
upper sediment layers is reduced due to elimination of sources, burial, biodegradation, biological
uptake, and diffusion (U.S. EPA, 1989b). A mass balance equation predicting sediment
concentration in relation to source loading, sedimentation rates, sediment mixing, biodegradation,
and diffusion was used to describe the recovery process. Results of this modeling suggested that
natural sedimentation (i.e., control of anthropogenic sources was assumed) is expected to cover the
study site with 10 centimeters of relatively clean sediments over a 10-year period.
5-21
-------
Comments on Analysis: Characterization of Exposure
*The characterization of exposure often takes the form of a predictive modeling exercise
to quantify the probability and magnitude of contaminant exposure given some
particular toxicant discharge or remedial action. Such an approach is not relevant to
the Commencement Bay investigation where the toxicant release has already occurred,
biological effects have been measured, and no remedial activities are addressed within
this case study. In this case, characterization of exposure is implicit within the elements
of both characterization of ecological effects and risk characterization.
5.3.4. Risk Characterization
Ideally, characterization of ecological risks due to exposure to sedimeints would be
supported by definitive cause-effect relationships between specific chemicals and biological
endpoints. To date, however, very little information of this type is available for mixtures of
chemicals in the environment. In lieu of definite risk estimates, relative measures of effect have
been developed. In the Commencement Bay project, a decision-making approach based on
empirical measures of risk (i.e., elevation above reference indices of biological effects) and a
predictive assessment (i.e., AET approach) were developed to identify and rank problem areas for
remediation.
Risks Based on Comparisons to Benchmarks. Evaluation of elevation above reference
indices for each indicator showed that chemical contamination of sediment was spatially
heterogeneous (table 5-1). The average values of elevation above reference indices are shown for
each of the Commencement Bay stations in table 5-2 (U.S. EPA, 1989b).
The average concentrations of several organic compounds exceeded all Puget Sound
reference conditions hi all study areas. Average elevation above reference values for selected
metals, sediment toxicity bioassay endpoints, and chemicals indicative of bioaccumulation were
significantly elevated above reference stations in six of the seven study areas evaluated.
Depressions of benthic macroinvertebrate taxa were statistically significant (P<0.05) in five of the
seven study areas; liver lesions in English sole were statistically significant (P<0.05) in four of the
seven study areas.
Action-level guidelines used to identify problem areas are summarized in table 5-3 (see
U.S. EPA, 1985, for details). For example, areas requiring further evaluation of contaminant
sources and remedial actions were identified when values for three or more indices were
significantly elevated. Other combinations of significant indices and the magnitude of the
elevations above reference triggered problem area definition as well (table 5-3). With the
guidelines in table 5-3, problem areas were identified within the Commencement Bay Superfund
site. Six segments within the large study areas had significant elevations above reference for all
three site-specific indicators (contamination, toxicity, and benthic effects).
5-22
-------
Table 5-2. Action Assessment Matrix of Sediment Contamination, Sediment Toxicity, and Biological Effects
Indices for Commencement Bay Study Areas (U.S. EPA, 1985)
Study Area Elevations'
Variable
Hylebos Blair Sitcum Milwaukee St. Paul Middle City Ruston
Reference
Valueb
Sediment Chemistry (ppb)
Sb
As
Cd
Cu+Pb+Zn
Hg
Ni
Phenol
PCP
LPAH
HPAH
Cl. benzene
Cl. butadiene
Phthalates
PCB
4-Methylphenol
Benzyl alcohol
Benzoic acid
Dibenzofuran
Nitrosodiphenylamine
Tetrachloroethene
Sediment Toxicity (%)
Amphipod bioassay
Oyster bioassay
Infauna6
Total benthos
Polychaetes
Molluscs
Crustaceans
Fish Pathology (%)
Lesion prevalence
Fish Bioaccumulation (ppb)
Copper
Mercury
Naphthalene
Phthalates
PCB6
DDE
8.0
4.2
2.2
1.7
5.5
5.1
0.8
9.3
0.7
7.0
9.6
2.8
18.0
26.0
7.5
5.5
22.0
10.0
1.4
< 1.0 U 0.1 „< 1.6
i 52.0 1
U 1.2
1 80.0 I
< 1.0
58.0
14.0
U 1.0
U 1.0
1.3
0.8
0.7
1.1
0.4
1.9
1.5
1.5
0.7
i 6.8 1
1.0
5.4
4.6
25113.51
2.7
1.7
5.6 1.0
T5~l 0.93
0.67 0.41
21.0
9.2
3.8
11.0
7.0
5.1
510.0
620.0
27.0
120.0
160.0
2.8
< 2.1 |
< 0.9
<60.0
<68.0
< 2.5.
< 1.2
< 0.66
13.0
3.4
12.0
U 1.9
<73.0
<27.0
< 1.8
< 1.3
< 0.56
<17.0
1,300.0
< 6.7
11.0
5.6
< 110.0
< 97.0
< 6.1
< 6.8
< 5.1
8.5
< 33.0
3.3,
9.4
1.9
120.0
< 140.0
< 9.0
< 1.9
< 7.1
< 12.0
30.0
4.7
4.5 I
< 1.0
< 87.0
< 85.0
< 3.3
1.7
4.5
19.0
< 10.0
< 1.2
0.6
0.5
1.2
0.7
2.1
110.0
3370.0
950.0
35,000.0
40.0
1,740.0
< 33.0
U33.0
< 41.0
< 79.0
U21.0
U62.0
< 280.0
< 6.0
< 13.0
U 10.0
< 140.0
U 3.7
U 4.1
U 10.0
9.3
13.0
6.7
U 38.0
U 55.0
< 54.0
< 74.0
< 36.0
< 1.8
"Boxed numbers represent elevations of chemical concentrations that exceed all Puget Sound reference area values, and statistically
significant toxicity and biological effects at the P<0.05 significance level compared with reference conditions. The "U" qualifier indicates
the chemical was undetected and the detection limit is shown. The " < " qualifier indicates the chemical was undetected at one or more
stations. The detection limit is used in the calculations. ,
Elevation above reference (EAR) values shown for each area are based on Carr Inlet reference values for each variable except for benthos
(see footnote d).
clnfauna EAR are based on the ratio of population abundances at the reference site to those at the Commencement Bay sites. For example,
the EAR for total benthos at Hylebos is 1.2, meaning that the abundance of benthic populations at the reference site is 20 percent greater
than at the Commencement Bay site. Since decreases in abundance of infauna are considered to be adverse effects, higher rates of
reference site to Commencement Bay site abundance reflect a greater likelihood of adverse biological effects.
"Different benthic reference values were used depending on sediment grain size.
"Locations where PCB concentrations are significantly elevated also poge a significant health risk to the exposed population.
5-23
-------
Table 5-3. Action-Level Guidelines (Washington Department of Ecology/U.S. EPA, 1985)
Condition Observed
Threshold Required for Action
I. Any THREE OR MORE significantly elevated indices?
n. TWO significantly elevated indices
1. Sediments contaminated, but below 80th
pcrcentile PLUS: Bioaccumulation without an
increased human health risk relative to that at
the reference area, OR Sediment toxicity with
less than 50 percent mortality or abnormality,
OR Major benthic invertebrate taxon depressed,
but by less than 95 percent
2. Sediments contaminated but below 80th
percentile PLUS elevated fish pathology
3. Any TWO significantly elevated indices, but
NO elevated sediment contamination
. SINGLE significantly elevated index
1. Sediment contamination
Threshold exceeded, continue with definition of problem area.
No immediate action. Recommended site for future monitoring.
Threshold for problem area definition exceeded if elevated
contaminants are considered to be biologically available. If not,
recommended site for future monitoring.
Conduct analysis of chemistry to distinguish site from adjacent
areas. If test fails, no immediate action warranted. Otherwise,
threshold exceeded for characterization of problem area. Re-
evaluate significance of chemical indicators.
If magnitude of contamination exceeds the 80th percentile for all
study areas, recommended area for potential source evaluation at
a low priority relative to areas exhibiting contamination and
effects.
2. Bioaccumulation
3. Sediment toxicity
4. Depressed benthic abundance
5. Fish pathology
Increased human health threat, defined as: Prediction of greater
than or equal to 1 additional cancer cases in the exposed
population for significantly elevated carcinogens, OR
For noncarcinogens, exceeding the acceptable daily intake value
is required.
Greater than 50 percent response (mortality or abnormality).
95 percent depression or greater of a major taxon (equals an
EAR of 20 or greater).
Insufficient as a single indicator. Recommended site for future
monitoring. Check adjacent areas for significant contamination,
toxicity, or biological effects.
"Combinations of significant indices are from independent data types (i.e., sediment chemistry, bioaccumulation, sediment toxicity, benthic
infauna, fish pathology). >
Significant indices are defined as follows: sediment chemistry = chemical concentration at study site exceeds highest values observed at
any Pugct Sound reference area.
Sediment toxicity, benthic abundance, bioaccumulation, and pathology = statistically significant (P<0.05) difference between study area
and reference area.
5-24
-------
Application of AET. During the Commencement Bay remedial investigation, AETs were
generated for three categories of biological effects variables (i.e., amphipod mortality, oyster
larvae abnormality, and benthic macroinvertebrate taxa abundance) for a data set of 50 to 60
stations. Following the remedial investigation, the AET data set was expanded to 334 stations,
including data from other areas of Puget Sound. Table 5-4 lists AET values used to define
sediment quality objectives for the cleanup of Commencement Bay sediments during the feasibility
study (from U.S. EPA, 1989b).
These values represent the lowest AET for the three biological effects indicators. Toxicity
or benthic AET (i.e., the concentration above which all sediments had significant toxicity or
benthic effects, respectively) was exceeded by several chemicals at most, but not all, of the 29
biological stations exhibiting significant effects. A detailed discussion of the AET for each
biological effect is presented in chapter 4 of the remedial investigation report (Washington
Department of Ecology/U.S. EPA, 1985) and summarized in U.S. EPA (1985).
Ranking of Study Areas. Prioritization of study areas was based on average and maximum
conditions in each area that exceeded the action-level guidelines. Results of the average ranking
method and the maximum ranking method were then compared. The spatial extent and general
priority of all problem areas for evaluation of sources and remedial action in the Commencement
Bay Superfund project are illustrated in figure 5-5. Problem areas defined only by mid-channel
stations in the current study were assumed to extend from shoreline to shoreline, unless historical
data indicated otherwise. Fourteen of the 21 problem areas identified were recommended for
priority source evaluation. Six areas were defined as the lowest priority areas for source
evaluation because they contained stations where contaminant concentrations exceeded AET, but no
confirming biological data were available. The seventh area not recommended for high-priority
source evaluation (Milwaukee Waterway) contained no chemicals measured above their AET.
Based on the ranking according to environmental indices, AET, spatial extent of
contamination, and confidence in source identification, nine discrete areas of sediment
contamination were identified (figure 5-5). These areas were designated as requiring future
evaluation and response under the Superfund program. Potential problem chemicals of varying
priorities for source identification were also identified in each of the 14 areas recommended for
priority source evaluation.
Uncertainty Analysis. Two measures of reliability of the AET approach were evaluated
with actual field data from 13 urban and nonurban embayments in Puget Sound (summarized in
U.S. EPA, 1988, based on Commencement Bay data; also see Barrick and Seller, 1989, for an
evaluation based on an expanded AET data base): (1) sensitivity in detecting environmental
problems (i.e., are all biologically affected sediments identified?) and (2) efficiency in screening
environmental problems (i.e., are only biologically affected sediments identified?).
These measures of reliability were applied to a range of sediment criteria generated by the
Equilibrium Partitioning (EP) and AET approaches. Overall reliability ranged from 44 to 64
percent for the EP approach and from 42 to 85 percent for the AET approach, depending on the
particular criterion and biological indicator tested. A higher percentage of correct predictions was
made using a combination of the two approaches than by using either approach alone.
5-25
-------
Table 5-4. Sediment Cleanup Objectives Related to Environmental Risks (U.S. EPA, 1989b)
Chemical
Sediment
Cleanup Objective"
Metals (mg/kg dry weight)
Antimony
Arsenic
Cadmium
Copper
Lead
Mercury
Nickel
Silver
Zinc
150B
57B
5.1B
390L
450B
0.59L
>140AB
6.1A
410B
Organic compounds (fig/kg dry weight)
Low-molecular-weight PAHs
Naphthalene
Acenaphthylene
Accnaphthene
Fluorcne
Phcnanlhrcnc
Anthracene
2-McthylnaphthaIene
High-molecular-weight PAHs
Fluoranthene
Pyrene
Bcnz(a)anthracene
Chrysene
Bcnzofluoranthenes
Benzo(a)pyrene
Indcno(l,2,3-c,d)pyrene
Benzo(a,h)anthracene
Benzo(g,h,i)pcrylene
Chlorinated organic compounds
1,3-Dichlorobenzene
1,4-Dichlorobcnzene
1,2-Dichlorobenzene
1,2,4-Trichlorobenzene
Hexachlorobenzene (HCB)
Total PCBs
5,200L
2,100L
1,300AB
500L
540L
1,500L
960J-
670L
17,000L
2,500L
3,300L
1,600L
2,800L
3,600L
1,600L
690L
230L
720L
170ALB
110B
50LB
51A
22B
1,000B
"Option 2 - Lowest AET among amphipod, oyster, and benthic infauna:
A - Amphipod mortality bioassay;
L - Oyster larvae abnormality bioassay;
B - Benthic infauna
5-26
-------
Table 5-4. Sediment Cleanup Objectives Related to Environmental Risks (continued)
Sediment
Chemical Cleanup Objective*
Phthalates
Dimethyl phthalate 160L
Diethyl phthalate 200B
Di-n-butyl phthalate 1,400AL
Butyl benzyl phthalate 900^
Bis(2-ethylhexyl)phthalate l.SOO8
Di-n-octyl phthalate 6,200B
Phenols
Phenol 420L
2-Methylphenol 63AL
4-Methylphenol 670L
2,4-Dimethylphenol 29L
Pentachlorophenol 360A
Miscellaneous extractables
Benzyl alcohol 73L
Benzoic acid 650LB
Dibenzofuran 540L
Hexachlorobutadiene 11B
N-nitrosodiphenylamine 28B
Volatile organics
Tetrachloroethene 57B
Ethylbenzene 10B
Total xylenes 40B
Pesticides
p.p'-DDE 9B
p,p'-DDD 16B
p.p'-DDT 34B
aOption 2 - Lowest AET among amphipod, oyster, and benthic infauna:
A - Amphipod mortality bioassay;
L - Oyster larvae abnormality bioassay;
B - Benthic infauna
5-27
-------
Ul
ib
oo
COMMENCEMENT
BAY
HIGHEST PRIORITY PROBLEM AREAS
SECOND PRIORITY PROBLEM AREAS
POTENTIAL PROBLEM AREAS
(NO CONFIRMING BIOLOGICAL
DATA AVAILABLE)
POTENTIAL PROBLEM AREA BY
HISTORICAL DATA ONLY
CHEMICALS EXCEED APPARENT
EFFECTS THRESHOLD
CHEMICALS BELOW APPARENT
EFFECTS THRESHOLD
1000
CITY
WAIIIIWAT
Figure 5-5. Definition and prioritization of Commencement Bay problem areas (U.S. EPA, 1989b)
-------
The EPA Science Advisory Board (U.S. EPA, 1989a) evaluated the AET approach for
assessing sediment quality. The conclusions of the Subcommittee on Sediment Criteria are:
The method has major strengths in its ability to determine biological effects and
assess interactive chemical effects. The method is considered by the Subcommittee
to contain sufficient scientific merit that, with appropriate validation of the AET
values, it could be used to establish sediment quality values for use at specific sites.
In the Subcommittee's opinion, the AET approach should not be used to develop
general, broadly applicable sediment quality criteria. Some major limitations drive
this opinion, including the site-specific nature of the approach, its inability to
describe cause-and-effect relationships, its lack of independent validation, and its
inability to describe differences in bioavailability of chemicals in different
sediments. The Subcommittee recommendations for strengthening the approach
include building in replicate sediment samples to assessments, devising criteria for
selecting reference sites, including considering physical factors and developing
measures of variance.
Comments on Risk Characterization
Strengths of the case study include:
• The use of data on multiple chemical measurements and biological endpoints (e.g.,
sediment chemistry, sediment toxicity, benthic macroinvertebrate assemblages, tissue
residues resulting from bioaccumulation, and fish liver histopathology) provides a
powerful weight-of-evidence approach to identify and rank problem areas.
•The AET is an innovative approach dealing with the problems created by contaminant
suites and uncertain cause-effect relationships. The approach is a noteworthy attempt
to use empirical relationships to sidestep currently intractable issues like bioavailability
and synergistic/antagonistic effects among toxicants. The combination of field-collected
sediment bioassays and the AET approach takes a step toward differentiating between
effects associated with different contaminants. The predictive AET approach relies on
objective statistical criteria for determining adverse effects for each biological indicator.
The spatial extent of areas of high risk may be delineated using the concordance
between biological response and chemical concentrations.
•Disparate chemical and biological data sets are integrated into the definition of
problem areas. By expressing all chemical and biological measures as elevations
relative to a reference site, comparisons among these measures and demonstration of
concordance becomes straightforward. It should be noted, however, that this requires
reducing biological endpoints to single values. This may be appropriate for some
endpoints, such as percentage of mortality in the bioassays, but it becomes more
problematic and results in substantial data loss when applied to a complex biological
response such as benthic community change.
5-29
-------
Comments on Risk Characterization (continued)
Imitations include:
9 The ecological assessment of Commencement Bay does not provide a probabilistic
approach to risk characterization. Also, the case study does not include a predictive
assessment of contaminant transport and fate in relation to the distribution of ecological
components.
» The ecological significance of some of the measurement endpoints is not explained,
particularly with respect to individual site characteristics. For example, it is not clear
how the adverse effects observed in the bioassay tests are representative of the insult to
the entire ecosystem of Commencement Bay.
General comments:
+There are some limitations in the use of the AET method. The definition of the AET
as the highest concentration at which no effect is observed (rather than the lowest
concentration at which any effect is observed) is the least protective of the possible
definitions for effects thresholds. Moreover, a correlation between contaminant
concentrations and biological effects does not demonstrate cause and effect. The AET
method assumes a consistently increasing biological response at increasing
concentrations of a chemical. Unmeasured chemicals or physical conditions may alter
this relationship. Species interactions and other community-level processes may also
alter the assumed dose-response relationship. A large data base of synoptic chemistry
and biological effects is needed to verify a particular AET value for a specific chemical.
*The Commencement Bay investigation was not originally conceived as a risk
assessment in the sense that it was neither predictive nor probabilistic. It was an
impact assessment intended to define the extent of contamination in marine sediments
and quantify the magnitude of existing biological damage. Consequently, application of
risk assessment terminology becomes confusing at times.
5-30
-------
5.4. REFERENCES
Barrick, R.C. (1985) Personal communication (comments to a draft report submitted by JRB
Associates, Bellevue, WA; comments sent under cover letter from Dr. Thomas C. Ginn).
Tetra Tech, Inc., Bellevue, WA, March 12.
Barrick, R.C.; Seller, H.R. (1989) Reliability of sediment quality assessment in Puget Sound. In:
Proceedings of Oceans '89. Institute of Electrical and Electronics Engineers, pp. 421-426.
Becker, D.S.; Bilyard, G.R.; Ginn, T.C. (1990) Comparisons between sediment bioassays and
alterations of benthic macroinvertebrate assemblages at a marine Superfund site:
Commencement Bay, Washington. Environ. Toxicol. Chem. 9;669-685.
Chapman, P.M.; Farrell, M.A.; Kocan, R.M.; Landolt, M. (1982a) Marine toxicity tests in
connection with toxicant pretreatment planning studies, METRO Seattle. E. V.S.
Consultants, Vancouver, B.C., Canada. 15 pp.
Chapman, P.M.; Vigers, G.A.; Farrell, M.A. (1982b) Survey of biological effects of toxicants
upon Puget Sound biota. I: broad-scale toxicity survey. NOAA Technical Memorandum
OMPA-25. 98 pp.
Chapman, P.M.; Fink, R.P. (1984) Effects of Puget Sound sediments and their elutriates on the
life cycle of Capitella capitata. Bull Environ. Contain. Toxicol. 33:451-459.
Malins, D.C.; McCain, B.B.; Brown, D.W.; Sparks, A.K.; Hodgins, H.O. (1980) Chemical
contaminants and biological abnormalities in central and southern Puget Sound. NOAA
Technical Memorandum OMPA-2. National Oceanic and Atmospheric
Administration, Boulder, CO. p. 295.
PSEP. (1986) Puget Sound protocols. Puget Sound Estuary Program. U.S. Environmental
Protection Agency, Region 10, Office of Puget Sound, Seattle, WA.
Swartz, R.C.; DeBen, W.A.; Sercu, K.A. (1982) Sediment toxicity and the distribution of
amphipods in Commencement Bay, Washington, USA. Mar. Pollut. Bull. 13:359-364.
Swartz, R.C.; DeBen, W.A.; Jones, J.K.P.; Lamberson, J.O.; Cole, F.A. (1985) Phoxocephalid
amphipod bioassay for marine sediment toxicity. In: Cardwell, R.D.; Purdy, R.; Banner,
R.C., eds. Aquatic toxicology and hazard assessment, seventh symposium. ASTM STP 854.
American Society of Testing and Materials, Philadelphia, PA. pp. 284-307.
U.S. Environmental Protection Agency. (1985) Summary report for the Commencement Bay
nearshore/tideflats remedial investigation. EPA 910/9-85/134a.
U.S. Environmental Protection Agency. (1988) Briefing report to the EPA Science Advisory
Board: the apparent effects threshold approach. U.S. Environmental Protection Agency,
Region 10, Seattle, WA.
5-31
-------
U.S. Environmental Protection Agency. (1989a) Report of the Sediment Criteria Subcommittee.
Evaluation of the apparent effects threshold approach for assessing sediment quality.
Science Advisory Board, Washington, DC. SAB-EETFC-89-027.
U.S. Environmental Protection Agency (1989b) Commencement Bay nearshore/tideflats. Record of
Decision. U.S. Environmental Protection Agency, Region 10, Seattle, WA.
Washington Department of Ecology/U.S. Environmental Protection Agency. (1985)
Commencement Bay nearshore/tideflats remedial investigation. Volume 1. U.S.
Environmental Protection Agency, Region 10, Seattle, WA.
Williams, L.G.; Chapman, P.M.; Ginn, T.C. (1986) A comparative evaluation of marine
sediment bioassays using bacterial luminescence, oyster embryo, and amphipod sediment
bioassays. Mar. Environ. Res. 19:225-249.
5-32
-------
SECTION SIX
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
THE NATIONAL CROP LOSS ASSESSMENT NETWORK
-------
AUTHOR AND REVIEWERS
AUTHOR
Walter W. Heck
Agricultural Research Service
U.S. Department of Agriculture
Raleigh, NC
REVIEWERS
Douglas P. Ormrod (Lead Reviewer)
Office of Graduate Studies
University of Guelph
Guelph, Ontario, Canada
Nancy A. Bryant
ENSR Consulting & Engineering
Acton, MA
Kenneth L. Dickson
Institute of Applied Sciences
University of North Texas
Denton, TX
Thomas M. Frost
Center for Limnology
University of Wisconsin
Madison, WI
Judy L. Meyer
Institute of Ecology
University of Georgia
Athens, GA
Randall S. Wentsel
Toxicology Division
U.S. Army Chemical Research,
Development, and Engineering
Center
Aberdeen Proving Grounds, MD
6-2
-------
CONTENTS
ABSTRACT 6-7
6.1. RISK ASSESSMENT APPROACH 6-8
6.2. STATUTORY AND REGULATORY BACKGROUND 6-8
6.3. CASE STUDY DESCRIPTION 6-8
6.3.1. Problem Formulation 6-8
6.3.2. Analysis: Characterization of Ecological Effects 6-12
6.3.3. Analysis: Characterization of Exposure 6-23
6.3.4. Risk Characterization 6-26
6.4. REFERENCES 6-31
6-3
-------
LIST OF FIGURES
Figure 6-1. Structure of analysis for the National Crop Loss Assessment Network 6-9
Figure 6-2. Field exposure design 6-14
Figure 6-3. Relative yield response of peanut to O3 using a 7-hr/day
seasonal mean O3 = 0.025 ppm 6-17
Figure 6-4. Relative yield response of tobacco to O3 value 6-17
Figure 6-5. Relative yield response of sorghum to O3 using a 7-hr/day
seasonal mean O3 value 6-18
Figure 6-6. Relative yield from two clover-fescue studies (Forage Model 223)
to O3 using a 10-hr/day seasonal mean O3 value 6-18
Figure 6-7. Relative yield response from three cotton studies (Model 323)
to O3 using a 12-hr/day seasonal mean O3 value 6-19
Figure 6-8. Relative yield response from six soybean studies (Model 123)
to O3 using a 12-hr/day seasonal mean O3 value 6-19
Figure 6-9. Category III models (heterogeneous) for all data from each
of three species 6-20
Figure 6-10. Wald and LRT 95 percent confidence interval estimates of RYL
for sorghum using the Category I concentration-response equation 6-22
Figure 6-11. Wald and LRT 95 percent confidence interval estimates of RYL
for peanut using the Category I concentration-response equation . 6-22
Figure 6-12. Location of monitoring sites used to estimate monthly maximum
7-hr O3 for July 1984 6-24
Figure 6-13. Map of kriged estimates (concentrations in ppb) of maximum 7-hr
O3 for July 1984 6-25
Figure 6-14. Estimated mean percent loss of cotton yield based on measured
O3 (SUM06) and a composite O3 exposure-response function 6-28
6-4
-------
LIST OF TABLES
Table 6-1. Summary of Crop Studies in NCLAN Program 6-11
Table 6-2. Estimated Relative Yield Losses (Percent) at
Four Seasonal (7-hr or 12-hr/day) Mean Ozone
Concentrations Using Homogeneous Weibull Models 6-16
Table 6-3. Wald Confidence Interval Estimates (95 Percent) of
Percentage Yield Loss From Ozone Relative to Yield
at O3 = 0.025 ppm Based on Category I or II
Weibull Response Equations 6-21
Table 6-4. Comparison of 1984 and 1988 Economic Surplus
Estimates in 1982 Dollars 6-27
6-5
-------
LIST OF ACRONYMS
EPA U.S. Environmental Protection Agency
LRT likelihood ratio test
NAAQS National Ambient Air Quality Standards
NCLAN National Crop Loss Assessment Network
NF nonfiltered
RYL relative yield losses
SAROAD Storage and Retrieval of Aerometric Data
USDA U.S. Department of Agriculture
6-6
-------
ABSTRACT
The National Crop Loss Assessment Network was initiated by the U.S. Environmental
Protection Agency in 1980 to develop an approach for assessing the impact of ozone on crop
production. Four primary and two secondary regional sites were established to conduct field
studies from 1980 through 1986. Forty-four field experiments, using regression designs with 17
crop species (including 38 cultivars and 3 corn crosses), were run to determine the impact of ozone
on growth and productivity. Physiological studies also were carried out in these designs. Plants
were grown under field conditions and exposed to different ozone concentrations in open-top field
chambers. Thirty-five experimental designs were factorial and included secondary stressors (i.e.,
low soil moisture, sulfur dioxide); cultivar testing; testing of dispensing methodology; or detailed
growth studies. Results from these studies helped investigators interpret results from the ozone
studies. Yield loss estimates for the economic analysis were obtained using the Weibull function
and were reported as relative yield losses. An ozone statistic was identified (a seasonal mean of
the 7-hr or 12-hr daily exposure period), and ozone monitoring data across the United States were
interpolated by a kriging technique to give spatial-level ozone statistics. Crop yields on the same
spatial basis were obtained from the U.S. Department of Agriculture. National yield losses,
estimated from the models, showed reductions from 0 to 22 percent at a seasonal mean value of
0.05 ppm ozone. Results were used in an economic model to derive estimated producer and
consumer benefits with increasing and decreasing ozone concentrations. The final economic
assessment using data from eight crops estimated that increased yields associated with a 25 percent
reduction in tropospheric ozone resulted in a $1.9 billion annual net benefit, while a 40 percent
reduction gave almost $3 billion in net annual benefits. The study was limited in size and only one
or two sites were used (total of six) in four regions of the country. The results of the effects of
soil moisture on plant yield response were not conclusive, and very little was done to determine the
effects of multiple abiotic and biotic stressors on crop yield response. Major scientific issues, in
addition to those mentioned above, include: (1) effects of the field methodology on plant response
to ozone; (2) the most applicable concentration statistic to use; (3) the accuracy of the kriging
approach; and (4) understanding plant processes in relation to the impact of ozone so that more
process-oriented models can be developed.
6-7
-------
6.1. RISK ASSESSMENT APPROACH
The National Crop Loss Assessment Network (NCLAN) study was designed as a risk
assessment and, with some modification, was fit to four parts of the ecological risk assessment
framework (figure 6-1): The ecological components were characterized and endpoints were
defined; the ozone exposure-plant response was studied experimentally in the field, and response
models were developed; exposure characteristics were described and documented; and risks were
characterized in both crop yield and economic terms. Primary limitations for a national risk
assessment are in the spatial representativeness of the test sites (only six for a national assessment)
and in the few experimental designs set up to assess the effects of interacting stressors. The
principal reference used in this case study was Heck et al. (1991).
-------
Figure 6-1. Structure of Analysis for
the National Crop Loss Assessment Network
PROBLEM FORMULATION
Stressors: ozone in ambient air.
Ecological Components: 17 crop species,
Endpoints: the primary assessment and measurement
endpoint was the impact of ozone on yield of the
crop part important for human use.
A
ANALYSIS |
Characterization of
Exposure
Monitoring data for
ambient ozone from over
300 sites were used to
calculate seasonal
concentrations using a
kriging model.
•
1
Characterization of
Ecological Effects
In-field chambers were
used to examine
relationships between ozone
exposure and plant yield.
The Weibull model was used
to describe dose-response
relationship.
RISK CHARACTERIZATION
Ozone exposure dose-crop yield response functions were
used to predict loss of yield in test species throughout
the country based on the calculated ozone values. The
estimated losses of yield were used as part of an economic
assessment.
6-9
-------
produced photochemically from gases associated with human activities probably accounts for
between one- and two-thirds of the tropospheric ozone found during the growing season in most
parts of the United States. Control of ozone requires control of nitrogen oxides or non-methane
hydrocarbons, or both (U.S. EPA, 1986).
The principal focus of the NCLAN program was to quantify the relationship between ozone
concentration and reduced yield hi economically important, agronomic crops (Heck et al., 1982).
The primary stressor was ozone. Research on crop response to ozone and extensive field
observations prior to the beginning of the NCLAN program gave clear evidence that all crop
species were sensitive to ozone (Heck et al., 1977; Heagle and Heck, 1980). Further research
documented yield losses due to ozone exposure for a number of crops (Heck et al., 1986).
Ozone injury was first identified by foliar symptoms, often described as necrotic flecking or
stipple on leaves. All early studies used foliar injury as an indication of the severity of plant
response. Ozone has been shown to reduce photosynthetic efficiency, alter carbon allocation to
various plant parts, affect many metabolic plant processes, reduce yield, and change food quality.
There are differences hi both species and cultivar sensitivities to ozone. Injury to many crop
species (i.e., bean, watermelon, cotton, peanut, soybean, oat, clover, potato) has been observed in
the field under ambient conditions of ozone (Heck et al., 1977; U.S. EPA, 1986).
Ecological Components. Because its focus was primarily economic, the NCLAN program
studied crops that represented the major production crops in the country. The degree of sensitivity
played no role hi crop selection. The growing season ranged from April through September in
most areas of the country, with a longer season available in the southern areas. Seed crops were
most sensitive during flowering and fruit development.
The crop species were chosen because they are economically valuable to the country
(together they represent about 85 percent of current crop acreage), they are of regional importance,
and they represent a number of crop types. The 17 species studied are shown in table 6-1. This
table also summarizes the number of studies, the number of cultivars, and the interacting factors in
different designs (Heck et al., 1991).
Endpoint Selection. Agroecosystems are relatively simple systems that are highly managed
to provide maximum crop yield under any given set of field conditions. The NCLAN program
•was designed to permit a national assessment of the impact of ozone on crop yield (Heck et al.,
1982, 1983). The yield information was then used to determine the economic losses related to
ozone impact on crop production systems in the United States.
The primary measurement and assessment endpoint was yield of the crop part important
directly or indirectly for human use (Heck et al., 1982, 1983). Although yield is one of the
primary assessment endpoints, the economic loss/gain in dollars was also considered an assessment
endpoint in this study. Additional measurement endpoints used in the study but not addressed in
this summary were a number of physiological (net photosynthesis, water use efficiency, water use)
and biochemical (oxidative enzymes, metabolic systems, and crop quality) parameters that could
eventually be used for process-level model development. From an ecological risk assessment
approach, the economic analysis was an extra step in the study. Additional measurement endpoints
6-10
-------
Table 6-1. Summary of Crop Studies in the NCLAN Program (Heck et al., 1991)s
Crop
Alfalfa
Barley
Corn
Cotton
Forage
Timothy/Red Clover
Fescue/Ladino Clover
Bean
Lettuce
Peanut
Sorghum
Soybean
Tobacco
Tomato
Turnip
Wheat
Total (17)
No. of Studies
1 yr. 2 yr.
1 1
2
2
5
1
1
2
2
1
1
14
1
2
1
4
38 3
No. of Cultivars
2
2
5
5
1/1
1/1
1
1
1
1
9
1
1
4
4
41
Ozone and
Other Factors
(No. of Studies)
Moisture (1), SO2
Moisture (1)
S02 (1)
Moisture (4), SO2
Moisture (1)
—
—
—
—
Moisture (7), SO2
— ,
SO2 (2)
. —
S02 (1)
(1)
(1)
(4)
1
Moisture (14), SO2 (10)
aCultivars were exposed to 4 to 6 ozone concentrations in each study.
6-11
-------
that could have affected crop yield response to ozone for which data were not gathered in the study
included susceptibility to pests (disease and insect) and other stressors.
The U.S. Department of Agriculture (USDA) crop inventory is done on a regular basis,
and the procedures are well established (USDA, 1981; Heck et al., 1983). The program did not
collect these data but used the data available from the crop inventory to estimate national yield
losses (USDA, 1981).
Comments on Problem Formulation
Strengths of the case study include:
•All reviewers felt that the problem formulation 'was thorough and well supported.
Limitations include:
+Agroecosystems are simple in comparison with natural ecosystems. Therefore,
caution should be exercised in extrapolating from this assessment to natural plant
communities.
6.3.2. Analysis: Characterization of Ecological Effects
Design of Field Studies on Effects of Ozone. EPA decided that long-term field studies
should be initiated to determine the impact of ozone on the yield of major agronomic crops.
Although the focus of NCLAN on agronomic crops restricted the understanding of effects on other
important plant species, it was decided that, for the funds available, an economic assessment would
give the most useful data to EPA for the next review of the ozone standard. Thus, EPA initiated
NCLAN in 1980 (Heck et al., 1982). Although the field test sites selected provided a limited basis
for a national assessment, the sites did include six major agricultural regions in the country with
varying soils and climatic conditions. Despite these variations, the study was able to compare
relative yield losses across regions, suggesting that the results obtained gave a reasonable estimate
of impact.
Research undertaken between 1958 and 1977 (Heck et al., 1977; U.S. EPA, 1978) using
controlled environmental or greenhouse exposure facilities clearly showed the cause-effect
relationship between ozone and injury or damage to plant species. Effects studied included foliar
injury, growth and reproduction, yield, and physiological responses. During this time, selected
field studies were undertaken, but only a few of these used exposure concentration-plant response
designs that would permit the development of response functions (Heagle and Heck, 1980; U.S.
EPA, 1986). These functions are necessary for extrapolation to other locations. These studies
clearly showed that the ambient concentrations of ozone present during the growing season were
sufficient to cause yield reductions in wheat, corn, soybean, and peanut.
6-12
-------
The NCLAN experimental designs used a charcoal-filtered air treatment (ca. 0.025 ppm
seasonal 7-hr/day ozone concentration) as the lowest experimental concentration. This treatment
was assumed to represent an overall natural (not related to human activity) background ozone
concentration for the country. The second treatment was a nonfiltered (NF) air chamber that had a
seasonal average ozone concentration slightly below the ambient level. Two to four higher ozone
concentration treatment chambers also were used in the experiments for a total of four to six
experimental chambers for each replication. Crops were planted following recommended farming
practices. After the plants were several inches tall, field plots of uniform plant material were
identified, and the plots were covered with open-top chambers to control the gaseous environment
around the plants (figure 6-2). Experiments were designed to expose plants to a series of ozone
concentrations so that ozone exposure concentration-crop yield response relationships could be
determined. Growth, yield, and physiological parameters also were measured on the agronomic
crops of primary economic concern. The primary impacts of interest were crop yield on a regional
and national basis and a translation of the yield changes to economic values (Heck et al., 1982).
Additional experimental detail can be found in Heck et al. (1982, 1983, 1984a,b, 1991).
Exposure Concentration-Plant Yield Response Functions. The NCLAN program was
specifically designed to permit the development of ozone exposure concentration-plant yield
response functions. A number of linear and nonlinear functions were tested in this program (Heck
et al., 1982, 1983). The Weibull nonlinear function was finally chosen for several reasons: (1) it
has a flexible form that covers the range of responses observed; (2) its form is biologically
realistic; (3) its parameters have straightforward interpretations; (4) it provides direct estimates of
proportional yields; (5) tests of homogeneity of proportional yield responses over data sets are
readily accomplished; and (6) where homogeneity is found, the common proportional response
models can be used to represent the response of the crop as a species.
The Weibull model is given as:
Y = a exp [-(x/w)X] +£
where Y is the yield and x is the ozone dose. The three parameters to be estimated are a, the
estimated yield at zero ozone concentration; w, the ozone concentration when yield is 0.37a; and
X, a dimensionless shape parameter (X=l gives the exponential loss function, whereas a larger X
[e.g., 4.5] gives a region of almost no loss [a threshold] before the curve starts to drop). The e is
the random error associated with each observation.
Individual and combined response functions were developed in three categories (Somerville
et al., 1989; Lesser et al., 1990):
• Category I—Response functions from the individual experimental designs;
• Category II—A combination of individual experimental designs within a species.
Here the functions showed homogeneity and a homogeneous response function
could be developed; and
6-13
-------
Figure 6-2. Field exposure design (Miller et al., 1989)
6-14
-------
• Category III—A combination of all experimental designs for a given species where
all functions did not show homogeneity; thus, a heterogeneous response function
was generated. The response functions (across sites and years) for all cultivars
within a species were tested for homogeneity and the homogeneous functions
developed were used to predict relative yield loss for specific cultivar groupings
within species (Somerville et al., 1989; Lesser et al., 1990).
Table 6-2 presents estimated relative yield losses in percentage for selected Category I and
II models for most of the crops tested in the NCLAN program (Heck et al., 1991). Relative yield
values for several Category I models (peanut, tobacco, sorghum, a clover/fescue forage) and
several Category II models (cotton and soybean) are shown in figures 6-3 to 6-8 (data are from
Somerville et al., 1989).
Category III response equations are summarized in Somerville et al. (1989) and Lesser et
al. (1990). A Category III response curve for each species is the best-fitting common Weibull
response curve where the model includes the a-terms to account for all sites, year, block, cultivar,
linear SO2, and linear moisture effects. Thus, each equation should be regarded as an
approximation of the average response of the crop. Category III models were used for the crops
included in the economic assessment. Figure 6-9 contains a heterogeneous Weibull response
function for corn, cotton, and soybean using a 12-hr/day seasonal mean value for ozone (Heck et
al., 1991).
Wald confidence interval estimates, at the 95 percent level of predicted relative yield losses
(RYL), are based on the Category I and II models (Somerville et al., 1989, 1990; Lesser et al.,
1990). Selected examples are shown in table 6-3. Somerville et al. (1990) compared the classical
confidence intervals, obtained from first-order linear approximation theory (Wald estimates) for the
NCLAN designs, with the more theoretically correct, but difficult to compute, interval estimates
based on the likelihood ratio test (LRT). Nine Weibull models from nine NCLAN studies were
used to compare the Wald and LRT confidence interval estimates of relative yield loss due to
ozone. Results for sorghum and peanut are shown in figures 6-10 and 6-11.
From earlier discussions and NCLAN results, it is clear that crop species, eultivar, crop
growth stage, environmental conditions, and the presence or absence of insect pests and disease
organisms could affect the response of the crop species to ozone (U.S. EPA, 1986; Heck et al.,
1986, 1988). The interaction with soil moisture was shown in several experimental designs and is
considered the most important environmental variable that might affect the response of the crop
species to ozone. Results from the ozone by soil moisture designs were used to develop a model
(King, 1988) that was used in the economic assessment to adjust for the soil moisture stress.
6-15
-------
Table 6-2. Estimated Relative Yield Losses (Percent) at Four Seasonal (7-hr or 12-hr/day)
Mean Ozone Concentrations Using Homogeneous Weibull Models3
Crop
(Model)b
Coefficient
of
Variation0
Ozone Concentration (nom)
0.04
0.05
0.06
0.07
7-hr/day seasonal means
Bean, kidney (2)
Lettuce (1)
Peanut (1)
Sorghum (1)
Tomato (1)
Turnip (1)
Wheat (2)
Alfalfa (1)
Corn (1)
Cotton (3)
Forage (2)
Soybean (1)
Soybean (3)
Tobacco (1)
15.5
28.2
7.3
5.1
11.8-12.3
33.6
10.9
7.6-8.3
9.9
6.7-17.8
5.6-12.1
6.6-19.8
4.1-18.0
5.3
4.3
0.0
• 6.5
0.9
3.5
7.2
2.8
12-hr/day
3.8
1.2
6.0
3.8
8.0
12.3
6.2
8.9
0.0
12.5
1.7
6.3
14.9
5.8
seasonal means
7.1
3.3
14.0
7.7
15.3
21.5
11.1
14.9
0.1
19.8
2.7
9.5
19.5
9.5
10.7
7.3
26.0
12.5
23.8
31.0
16.4
22.3
0.5
27.9
3.9
12.9
35.7
14.3
14.6
13.9
41.1
18.2
33.0
40.2
21.8
aThe predicted relative yield losses come from table 13 in Somerville et al. (1989) using the
homogeneous models from Lesser et al. (1990). Values shown are mean values from the 95
percent confidence limits table. Yield losses are calculated relative to yield at a seasonal O3
mean of 0.025 ppm. Data points were adjusted to remove all fixed effects except the effects of
03.
"Homogeneous model numbers refer to models from table 1 in Lesser et al. (1990).
The coefficient of variation came from Heagle et al. (1988) and is shown for the study or studies
from which the modeled data were obtained.
6-16
-------
1.2-r
1.0 -
>
e 0.6
I
d 0.4
0.2+
0
0.02 0.04 0.06 0.08 0.1
Ozone (ppm)
0.12 0.14
Figure 6-3. Relative yield response of peanut to O3 using a 7-hr/day seasonal mean O3
0.025 ppm (Somerville et al., 1989)
Y
i
e
I
d
1.2-]
1.0 •
0.8-
0.6-
0.4-
0.2-
*>-*^%
*^""^V
* -1*J>«8s^»
* ^^-•.£
0.
0 0.03 0.06 0.09 0.12
Ozone (ppm)
Figure 6-4. Relative yield response of tobacco to O3 value. The relative yield value is scaled
to show yield relative to yield at O3 = 0.025 ppm (Somerville et al., 1989)
6-17
-------
Y
e
i
i
d
1.2-,
1 n .
I . W
0.8-
0.6-
0.4-
0.2-
n •
•
0,02 0.04 0.06 0.08 0.1
Ozone (ppm)
0.12 0.14
Figure 6-5. Relative yield response of sorghum to O3 using a 7-hr/day seasonal mean O3
value. The relative yield value is scaled to show relative yield at O3 = 0.025 ppm
(Somerville et al., 1989).
Y
i
e
1
d
1.8-j
1.6-
1.4-
1.2-
1.0 •
0.8-
0.6,-
0.4-
0.2-
n .
•
•-••—— -^Sft.,
~ O 09 *>** "**^». ^ 0
O"~~«-''^-«*!« «•
•• **^3*
0.02 0.04 0.06
Ozone (ppm)
0.08
0.1
Figure 6-6. Relative yield from two clover-fescue studies (Forage Model 223) to O3 using a
10-hr/day seasonal mean O3 value. The relative yield value is scaled to show
yield in relation to yield at O3 = 0.025 ppm. Data points were adjusted to
remove all fixed effects except the effects of O3 (Somerville et al., 1989).
6-18
-------
0.02
0.04 0.06 0.08
Ozone (ppm)
0.12 0.14
Figure 6-7. Relative yield response from three cotton studies (Model 323) to O3 using a 12-
hr/day seasonal mean Q3 value. The relative yield value is scaled to show yield
relative to yield at O3 = 0.025 ppm. Data points were adjusted to remove all
fixed effects except the effects of O3 (Somerville et al., 1989).
Y
e
d
1.6 •
1.4-
1.2-
1.0 •
0.8-
0.6'
0.4-
0.2-
m — • *
g * 9f •ijS'Sjj^fr*? •
<• * r» *
0.02 0.04 0.06
Ozone (ppm)
0.08
0.1
Figure 6-8. Relative yield response from six soybean studies (Model 123) to O3 using a 12-
hr/day seasonal mean O3 value. The relative yield value is scaled to show yield
relative to yield at O3 = 0.025 ppm. Data points were adjusted to remove all
fixed effects except the effects of O3 (Somerville et al., 1989).
6-19
-------
S
I
1.0
0.8
£0.6
0.4
1
I
0.2
II IT
0.4 0.6 0.8 l.(
Ozone Concentration (ppra)
Figure 6-9. Category III models (heterogeneous) for all data from each of three species [corn
(A), cotton (•) and soybean (D)]. Relative yield response to O3 using a 12-hr/day
seasonal mean O3 value. The relative yield value is scaled to show yield relative
to yield at O3 = 0.025 ppm. Data points were adjusted to remove all fixed
effects except the effects of O3 (Somerville et al., 1990).
6-20
-------
Table 6-3. Wald Confidence Interval Estimates (95 Percent) of Percentage Yield Loss From
Ozone Relative to Yield at O3 = 0.025 ppm Based on Category I or II Weibull
Response Equations
Estimated Relative Yield Losses (Percent)13
Crop
(Model)a
Alfalfa (1)
Corn (1)
Cotton (3)
Forage (2)
Peanut (l)c
Sorghum (1)°
Soybean (1)
Tobacco (1)
Wheat (2)c
Ozone Concentration fooml
0.03
(0.4, 1.9)
(0, 0.4)
(0.6, 2.0)
(-0.2, 2.3)
(0.9, 2.6)
(-0.5, 1.0)
(1.3, 3.2)
(0, 3.8)
(0.1, 1.4)
0.05
(4.3, 9.8)
(1.0, 5.6)
(10.2, 17.8)
(2.3, 13.0)
(8.7, 16.3)
(-1.7,5.1)
(12.0, 18.6)
(4.2, 18.0)
(2.3, 9.2)
0.07
(11.2, 17.9)
(3.5, 11.1)
(37.3, 44.8)
(12.2, 24.1)
(23.0, 32.8)
(-1.5,9.3)
(29.7, 36.3)
(13.6, 30.0)
(8.9, 19.6)
aModel numbers refer to models from table 1 in Lesser et al. (1990).
bThe estimated relative yield losses come from table 13 in Somerville et al. (1989).
^-hr/day seasonal O3 means were used in these three species; all other species used a 12-hr/day
seasonal O3 mean concentration.
6-21
-------
to
CO
33
CD
0)
73
CD
OC
16
14
12
10
8
6
4
2
0
-2
0.
Relative yield loss estimate
Waid 95% confidence limit
LRT 95% confidence limit
,..—••"
03 0.04 0.05 0.06 0.07 0.08 0.09
Ozone (H.L L'1)
0.1
Figure 6-10. Waid and LRT 95 percent confidence interval estimates of RYL for sorghum
using the Category I concentration-response equation. Point estimates of RYL
are shown with lighter solid line (Somerville et al., 1990).
60i
^
C* 50-
to
to An
.o 40
•o
75 30
—• Relative yield loss estimate
..... Waid 95% confidence limit
— LRT 95% confidence limit
0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Ozone (M-L L"1)
Figure 6-11. Waid and LRT 95 percent confidence interval estimates of RYL for peanut
using the Category I concentration-response equation. Point estimates of RYL
are shown with lighter solid line (Somerville et al., 1990).
6-22
-------
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
*The uncertainties for the ozone concentration-crop yield response data are clearly
represented by the 95 percent confidence limits developed for each homogeneous
response junction for each crop species.
6.3.3. Analysis: Characterization of Exposure
The experimental design ensured that ozone was the stressor of concern. Monitoring data
from around the United States over a number of years showed that ozone was present throughout
the growing season at concentrations capable of causing injury to sensitive vegetation (Heck et al.,
1983). EPA maintains a data bank of ozone-monitoring data from across the country in its Storage
and Retrieval of Aerometric Data (SAROAD) data base. These data, from over 300 sampling
sites, were used in an interpolation process (kriging) to calculate the seasonal 7-hr/day ozone
concentration on a county basis across the country (Heck et al., 1984a,b; Lefohn et al., 1987).
Although the uncertainties 'associated with the ozone data could ,be calculated, this was not done.
However, many data sets were dropped because of incomplete data. Data kriged to several
NCLAN sites compared well with the site data, providing some verification to the kriging model.
Variations in the kriged values would affect the final results, but there was no way to verify the
model or to establish confidence levels. Figures 6-12 and 6-13 show kriged values for two
different years across the United States; the figures are from Knudsen and Lefohn (1988).
Ozone concentrations in the experimental chambers were monitored on a continuous basis
during the experimental period (Heck et al., 1982). The data were initially summarized as hourly
averages. The experimental ozone data sets were then summarized as the 7-hr or 12-hr seasonal
averages (Somerville et al., 1989). These averages were subsequently used with the chamber yield
data to generate ozone exposure concentration-crop yield response functions (Heck et al., 1984a,b;
Lesser et al., 1990; Somerville et al., 1989).
Over the 7 years of the program, the ambient seasonal 7-hr/day ozone concentrations were
between 0.035 and 0.070 ppm both at the experimental sites and from the kriged county-level data.
Monitored levels of ambient ozone suggested that all crops might show some indication of response
in most areas of the country (Heck et al., 1991). Monitored ambient ozone data had low
uncertainties, but these were not calculated.
6-23
-------
Figure 6-12. Location of monitoring sites used to estimate monthly maximum 7-hr O3 for July 1984 (Knudsen and Lefohn,
1988)
-------
ON
Figure 6-13. Map of kriged estimates (concentrations in ppb) of maximum 7-hr O3 for July 1984 (Knudsen and Lefohn, 1988)
-------
Comments on Analysis: Characterization of Exposure
Strengths of the case study include:
• The use of a large data base on ozone measurements as a basis for interpolation
demonstrates how large monitoring networks can be employed in a risk assessment.
Limitations include:
• Uncertainties associated with the ozone data were not calculated although this could
be done.
6.3.4. Risk Characterization
The ozone exposure concentration-crop yield response functions were used in a predictive
fashion to determine the loss of yield in the test species across the country (Somerville et al., 1989;
Lesser et al., 1990). The 95 percent confidence limits for homogeneous response functions for
each species were calculated so that variation in the predicted losses could be determined
(Somerville et al., 1989, 1990; Lesser et al., 1990). These response functions were the most
important component in the overall analysis, because they had the greatest impact on the yield
assessment endpoint.
The county seasonal ozone values and crop inventories were then used in the response
functions to calculate yield losses from ozone on crop species across the country. These yield
losses were then used for the national economic assessment. The summary economic assessment is
shown in table 6-4 for both the 1984 and 1988 assessments (Adams et al., 1984, 1988, 1989).
A possible technique to map crop losses for cotton on a national basis is shown in
figure 6-14. The county seasonal ozone values were used in a composite ozone exposure
concentration-crop yield response function for cotton-growing areas in the United States. The
mean cotton yield loss was estimated in a number of cotton-growing areas. The loss gradient
surface was then developed to represent estimated loss on a geographical basis. Maps such as this
could be developed for any of the crops studied in the NCLAN program.
6-26
-------
Table 6-4. Comparison of the 1984 and 1988 Economic Surplus Estimates in 1982 Dollars3
Total Surplus
Ozone Assumptions ($ millions)
1984 Model
25% Increase -2,165
10% Reduction 699
25% Reduction 1,828
40% Reduction 2,637
1988 Model
25% Increase -2,053
10% Reduction 808
25% Reduction 1,890
40% Reduction 2,780
aFrom Adams et al. (1988). The ozone assumptions are based on the current ambient seasonal ozone
concentrations as determined by the kriging interpolation of the SAROAD data base (Lefohn et al.,
1987).
6-27
-------
PREDICTED MEAN
YIELD LOSS
Cotton Region
PERCENT CROP LOSS
D 0-1 • 15-20
D 1-5 • >20
D 5-10 D Ho Predicted Value
• 10-15
U.S. timiOmtKm >IC1[C1IOI Miner
!»»IIO««l«lll 1ESUICH t*l - COimilS
Pt.iir.l >t IB
Figure 6-14. Estimated mean percent loss of cotton yield based on measured O3 (SUM06) and
a composite O^ exposure-response function. The upper circle map illustrates the
mean cotton yield loss at each site; the sites were used to create the gradient
surface in the lower map. The maps were prepared by S.H. Azevedo and
permission to use them was obtained from D.T. Tingey.
6-28
-------
Comments on Risk Characterization
Strengths of the case study include:
• The NCLAN data are well documented and are contained in a central data base. All
data were verified using extensive cross-checking techniques. All experimental sites
developed and maintained a strong quality assurance/quality control (QA/QC)
program, and the sites were audited on an annual basis.
• The program used yield as both a measurement and an assessment endpoint of
primary concern. However, many other measurements were obtained during the
program that were used for a number of different purposes. Eventually, the other
measurements could be used in model development that could make the predictive
capabilities of the functional relationships more accurate. The economic assessment
was the assessment endpoint of primary interest in the NCLAN program.
• The economic analysis used the experimental crop response data, the kriged county-
level ozone values, and the USDA-generated county-level crop inventory data to
calculate national yield losses from ozone. The data were then used to calculate
economic benefits from different percentage reductions in the seasonal ozone values.
The data that went into the economic analyses were carefully developed and confidence
levels are known.
•Regular and in-depth communications were a primary factor in the success of the
NCLAN program. The ability to interpolate national ozone values and to develop
homogeneous yield responses across sites, years, and cultivarsfor a given species was
another critical factor. Cooperators met annually for a week-long workshop to
develop protocol and adopt consensus approaches to ensure uniformity and
comparability of results.
• This was a valuable case study that addressed a national problem and, because it
was a well-coordinated study using a standardized approach, it will be a useful model
for future studies.
Limitations include:
• The NCLAN program outlined a number of issues that had not received sufficient
attention. These issues are clearly developed in a final NCLAN paper published in an
Air and Waste Management Association Transaction in the spring of 1991; areas for
new research initiatives are recommended in that same publication (Heck et al., 1991).
Several issues are briefly highlighted. The study was limited in size, and only one or
two sites were used (total of six) in four regions of the country. The results on the
effects of soil moisture on plant yield response were not conclusive and very little was
done to determine the effects of multiple abiotic and biotic stressors on crop yield
response. Major scientific issues remaining to be addressed, in addition to those
6-29
-------
Comments on Risk Characterization (continued)
mentioned above, include: (1) effects of the field methodology on plant response to
ozone," (2) the most applicable concentration statistic to use; (3) the accuracy of the
kriging approach; and (4) understanding plant processes in relation to the impact of
ozone so that more process-oriented models can be developed.
*T7ie agricultural production systems under study were much simplified compared with
natural ecosystems. Similarly, the final economic analysis represented an additional
step that may be inappropriate in many ecological risk assessment case studies.
General comment:
+Based on experience with this program, it is recommended that all large-scale
programs have regular researcher meetings and interactions. These activities were
critical to the success ofNCLAN.
6-30
-------
6.4. REFERENCES
Adams, R.M.; Hamilton, S.A.; McCarl, B.A. (1984) The economic effects of ozone on
agriculture. Corvallis, OR: U.S. Environmental Protection Agency, Environ. Res. Lab.
EPA 600/3-84/090.
Adams, R.M.; Glyer, J.D.; McCarl, B.A. (1988) The NCLAN economic assessment: approach,
findings and implications. In: Heck, W.W.; Taylor, O.C.; Tingey, D.T., eds. Assessment
of crop loss from air pollutants. London: Elsevief Science Publishers, pp. 473-504.
Adams, R.M.; Glyer, J.D.; McCarl, B.A.; Johnson, S.L. (1989) A reassessment of the economic
effects of ozone on U.S. agriculture. /. Air Pollut. Control Assoc. 39:960.
Heagle, A.S.; Heck, W.W. (1980) Field methods to assess crop losses due to oxidant air
pollutants. In: Assessment of losses which constrain production and crop improvement in
agriculture and forestry. Proc. of the B.C. Stakman Commemorative Symposium, Misc.,
Pub. #7, Agricultural Exp. Sta., Univ. of Minn., pp. 296-305.
Heagle, A.S.; Kress, L.W.; Temple, P.J.; Kohut, R.J.; Heggestad, H.E. (1988) Factors
influencing ozone concentration-yield response relationships in open-top chamber studies.
In: Heck, W.W.; Taylor, O.C.; Tingey, D.T., eds. Assessment of crop loss from air
pollutants. London: Elsevier Science Publishers, pp. 141-179.
Heck, W.W.; Mudd, J.B.; Miller, P.R. (1977) Plants and microorganism. In: Ozone and other
photochemical oxidants. Washington, DC: National Academy of Sciences, pp. 437-585.
Heck, W.W.; Taylor, O.C.; Adams, R.; Bingham, G.; Miller, J.; Preston, E.; Weinstein, L.
(1982) Assessment of crop loss from ozone. J. Air Pollut. Control Assoc. 32:353.
Heck, W.W.; Adams, R.M.; Cure, W.W.; Heagle, A.S.; Heggestad, H.E.; Kohut, R.J.; Kress,
L.W.; Rawlings, J.O.; Taylor, O.C. (1983) A reassessment of crop loss from ozone.
Environ. Sci. Technol. 17:573A.
Heck, W.W.; Cure, W.W.; Rawlings, J.O.; Zaragoza, L.J.; Heagle, A.S.; Heggestad, H.E.;
Kohut, R.J.; Kress, L.W.; Temple, P.J. (1984a) Assessing impacts of ozone on
agricultural crops: I. Overview. J. Air Pollut. Control Assoc. 34:729.
Heck, W.W.; Cure, W.W.; Rawlings, J.O.; Zaragoza, L.J.; Heagle, A.S.; Heggestad, H.E.;
Kohut, R.J.; Kress, L.W.; Temple, P.J. (1984b) Assessing impacts of ozone on
agricultural crops: II. Crop yield functions and alternative exposure statistics. J. Air Pollut.
Control Assoc. 34:810.
Heck, W.W.; Heagle, A.S.; Shriner, D.S. (1986) Effects on vegetation: native, crops, forests. In:
A.C. Stern, ed. Air pollution, Vol. 6. New York: Academic Press, pp. 247-350.
6-31
-------
Heck, W.W.; Taylor, O.C.; Tingey, D.T., eds. (1988) Assessment of crop loss from air
pollutants. London: Elsevier Science Publishers.
Heck, W.W.; Heagle, A.S.; Miller, I.E.; Rawlings, J.O. (1991) A national program (NCLAN) to
assess the impact of ozone on agricultural resources. In: Berglund, R.L.; Lawson, S.R.;
McKee, D.J., eds. Tropospheric ozone and the environment. Transaction No. 19, Air and
Waste Management Assoc., Pittsburgh, PA, pp. 225-254.
King, D.A. (1988) Modeling the impact of ozone x drought interactions on regional crop yields.
Environ. Pollut. 53:351.
Knudsen, H.P.; Lefohn, A.S. (1988) The use of geostatistics to characterize regional ozone
exposures. In: Heck, W.W.; Taylor, O.C.; Tingey, D.T., eds. Assessment of crop loss
from air pollutants. London: Elsevier Science Publishers, pp. 91-105.
Lefohn, A.S.; Knudsen, H.P.; Logan, J.; Simpson, J.; Bhumralkar, C. (1987) An evaluation of
the kriging method to predict 7-h seasonal mean ozone concentrations for estimating crop
losses. J. Air Pollut. Control Assoc. 37:595.
Lesser, V.M.; Rawlings, J.O.; Spruill, S.E.; Somerville, M.C. (1990) Effects of ozone on
agricultural crops: statistical methodologies and estimated concentration-response
relationships. Crop Sd. 30:148.
Miller, I.E.; Heagle, A.S.; Vozzo, S.F.; Philbeck, R.B.; Heck, W.W. (1989) Effects of ozone
and water stress, separately and hi combination, on soybean yield. J. Environ. Qual.
18:330.
Somerville, M.C.; Spruill, S.E.; Rawlings, J.O.; Lesser, V.M. (1989) Impact of ozone and sulfur
dioxide on the yield of agricultural crops. Tech. Bull. 292. Raleigh, NC: N.C. State Res.
Service.
Somerville, M.C.; Dassel, K.A.; Rawlings, J.O. (1990) Adequacy of interval estimates of yield
responses to ozone estimated for NCLAN data. Crop Sd. 30:836-844.
U.S. Department of Agriculture. (1981) Agricultural statistics, 1980. Washington, DC.
U.S. Environmental Protection Agency. (1978) Air quality criteria for ozone and other
photochemical oxidants. Chapter 10/11. Washington, DC. EPA 600/8-78/004.
U.S. Environmental Protection Agency. (1986) Air quality criteria for ozone and other
photochemical oxidants. Vol. III. Chapter 6. Environmental Criteria and Assessment
Office, Research Triangle Park, NC. EPA 600/8-84/020CF.
6-32
-------
SECTION SEVEN
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
COMPARATIVE ANALYSIS OF MINING TAILING DISPOSAL FOR THE
QUARTZ HILL MOLYBDENUM MINING PROJECT
-------
AUTHORS AND REVIEWERS
AUTHORS
Patricia Cirone
Region 10
U.S. Environmental Protection Agency
Seattle, WA
REVIEWERS
James J. Anderson (Lead Reviewer)
School of Fisheries Research
University of Washington
Seattle, WA
Jay S. Jacobson
Boyce Thompson Institute
Cornell University
Ithaca, NY
Larry Kapustka
Ecological Planning and Toxicology, Inc.
Corvallis, OR
Wayne G. Landis
Institute of Environmental
Toxicology and Chemistry
Western Washington University
Bellingham, WA
John R. Yearsley *'•* - v.'•;',;:.;-y.V-.T". v.''-
Region 10
:U;S. Environmental Protection Agency
Seattle, WA
.
-v ;TJu>ma& Sibley : > ,; ; ; ' ;,
•...^School of Fisheries Research ;t ;, '•:
University,of Washington/ ; t
Seattle, WA
Frieda B. Taub
School of Fisheries Research
University of Washington
Seattle, WA
Donald P. Weston
Horn Point Environmental Laboratory
University of Maryland
Cambridge, MD
7-2
-------
CONTENTS
ABSTRACT .... .................... 7-6
7.1. RISK ASSESSMENT APPROACH . ... . . ..•'. ........ 7-7
7.2. STATUTORY AND REGULATORY BACKGROUND . . . ... .7-7
7.3. CASE STUDY DESCRIPTION ......... . . . ........ ..... 7-9
7.3.1. Problem Formulation ...., . . ,, ....... ... . ..... . . . . . . 7-9
7.3.2. Analysis: Characterization of Ecological Effects . . . . .7-22
7.3.3. Analysis: Characterization of Exposure ... ..... ... . . ............ 7-23
7.3.4. Risk Characterization . . .... ..... . . .> . ; . . . .,. .;.... . 7-26
7.4. REFERENCES ..'. . . \. . . 7-37
7-3
-------
LIST OF FIGURES
Figure 7-1. Structure of analysis for comparative analysis of mining
tailing disposal 7-8
Figure 7-2. Location of Quartz Hill project 7-10
Figure 7-3a. Map of Smeaton Bay alternative for tailing disposal from the
proposed Quartz Hill mining project, Alaska 7-11
Figure 7-3b. Map of Boca de Quadra alternative for tailing disposal 7-11
Figure 7-4a. Smeaton Bay longitudinal bathymetric section 7-12
Figure 7-4b. Boca de Quadra longitudinal bathymetric section 7-12
Figure 7-5a. Probability of exceeding water quality criterion—Year 5 7<-27
Figure 7-5b. Probability of exceeding water quality criterion—Year 20 7-28
Figure 7-5c. Probability of exceeding water quality criterion—Year 55 7-29
LIST OF TABLES
Table 7-1. Ore and Tailings Inorganic Content and Their Associated Toxicity 7-14
Table 7-2. Reagents Used in Quartz Hill Mining Project 7-15
Table 7-3. General Characteristics of Benthic Habitats in Boca de Quadra
and Smeaton Bay 7-20
Table 7-4. Maximum Suspended Solid Concentration Predicted for Upper
Water Column of Smeaton Bay/Wilson Arm and Boca de Quadra 7-25
Table 7-5. Estimation of Extractable Copper Concentration 7-25
Table 7-6. Worst-Case Estimates of Biomass of Demersal Organisms Lost
Over Life of Proposed Quartz Hill Mining Project 7-31
Table 7-7. Comparison of Projects in North America That Discharge
Mining Tailings to Marine Waters 7-33
7-4
-------
LIST OF ACRONYMS
BPJ best professional judgment
EC50 effective concentration for 50 percent of organisms tested
EIS Environmental Impact Statement
EPA U.S. Environmental Protection Agency
LC50 lethal concentration to 50 percent of organisms tested
MIBC methyl isobutyl carbinol
NEPA National Environmental Policy Act
PAH polycyclic aromatic hydrocarbon
USDA U.S. Department of Agriculture
USFS U.S. Forest Service
7-5
-------
ABSTRACT
This case study examines the relative risks of two disposal alternatives for mining tailings.
Quartz Hill is a proposed molybdenum mining site located within Misty Fjords National Monument
in southeast Alaska in a designated non-wilderness area. The project includes an open pit mine
with ore process facilities located nearby. The process wastes (tailings) will be transported from
the Tunnel Creek area by pipeline to a submarine, gravity-flow outfall hi either of two basins.
Two possible waste ore (tailings) disposal sites in nearby fjords (Smeaton Bay and Boca de Quadra)
were evaluated for potential environmental impacts. Steady-state diffusion models were used to
predict the distribution of tailings composed of metals and solids in the surface waters of two
marine fjords. Total recoverable copper, suspended solids, and settled solids were the stressors.
Ecological components included a variety of fish and invertebrate populations. Since the impacts
were presented in probabilistic terms, a relative comparison was completed of each fjord's spatial
and temporal exceedance of water quality degradation criteria and the number of benthic organisms
lost per hectare of viable habitat covered by tailings. Exceedance of the water quality criterion for
copper and loss of benthic habitat was predicted to be greater for the smaller of the two fjords.
However, it is clear that further chemical analytical work or additional site-specific bioassays must
be completed to verify the assessment. The finding from this evaluation is that disposal of tailings
into Boca de Quadra would be the least environmentally damaging alternative of those considered.
The permit to dispose of tailings into Smeaton Bay was denied. The purpose of this ecological risk
analysis was to compare risks; therefore, no conclusion was reached regarding whether either risk
was acceptable.
7-6
-------
7.1. RISK ASSESSMENT APPROACH
A comparative risk assessment for the Quartz Hill molybdenum mining project was
completed to resolve concerns about the choice of alternative locations for disposal of tailings
produced during the mining process (figure 7-1). Before this assessment was prepared, it was
determined that mining was an appropriate activity in this area of Alaska and that open water
disposal was an acceptable method of removing waste tailings. The justification for these
determinations is presented hi the Environmental Impact Statement (EIS) prepared by the U.S.
Forest Service (USFS).
This case! study incorporates data or information from three levels: (1) experimental results
from the revised draft EIS (USDA, 1987); (2) experimental results from other projects or
environments with similar characteristics; and (3) professional judgment based upon experience and
knowledge of physical, chemical, and biological phenomena.
All stages of the project upon which agreement was reached at earlier phases of the
National Environmental Policy Act (NEPA) process are not addressed in this risk assessment. The
following factors were excluded because they were not directly related to estimating the risk of
adverse effects to aquatic organisms from in-water disposal of tailings:
• terrestrial impacts from construction, transportation, or accessories needed for the
operation of the mine, including such things as road construction, energy use, and
transport of tailings;
• discharge of marine terminal wastewater (treated sanitary wastes, runoff, and wash
water) used in the mine operation exclusive of tailings process water; and
• personnel on site (e.g., impacts associated with sewers, water lines).
In this assessment, an attempt was made to quantify risks from exposure to stressors in
either of two basins. However, much of the information that was collected over many years was
not considered acceptable for a variety of reasons. The lack of adequate data limited the
assessment of effects on the aquatic populations. Because this is a comparative risk assessment,
the lack of adequate data did not limit the comparison of risk; it only limited the ability to estimate
all the effects that may result from the disposal of tailings.
7.2. STATUTORY AND REGULATORY BACKGROUND
The USFS, U.S. Environmental Protection Agency (EPA), and other federal and state
agencies prepared an EIS in accordance with NEPA, evaluating the potential environmental
consequences of alternative mine development scenarios. Three alternatives were considered for
tailings disposal: one upland and two in open water. During the initial phase of the NEPA
process, the upland alternative was eliminated. The remaining two alternatives allow disposal of
tailings into either of two fjords: middle-basin Boca de Quadra or Smeaton Bay/Wilson Arm.
These two fjords are technically considered inland waters and therefore are not subject to the
Ocean Discharge Criteria regulations (403c) of the Clean Water Act. However, EPA determined
7-7
-------
Figure 7-1. Structure of Analysis for
Comparative Analysis of Mining Tailing Disposal
PROBLEM FORMULATION
Stressors; physical impact of burial; copper
in water column.
Ecological Components: benthic invertebrates and
fishery resources.
Endpoints; assessment endpoint was habitat loss and
potential effects on benthic invertebrate and fish
populations. Measurement endpoints included estimated
accumuJation of tailings on seafloor and water column
copper concentration.
t
t
ANALYSIS
Characterization of
Exposure
Models were used to
predict burial and to
estimate the concentration
of copper in the water
column.
Characterization of
Ecological Effects
Effects were evaluated with:
- water quality criteria
- toxicity testing
- tissue analysis
- recolonization studies.
RISK CHARACTERIZATION
A comparison of the relative risks of disposal was made
for two fjords. The assessment was based on comparison
of predicted water column copper concentrations with
the water quality criterion for copper, the frequency
of exceedences and estimates of suspended solids
levels, and alteration of benthic habitat.
7-8
-------
that for the Quartz Hill project, the Ocean Discharge Criteria provide a useful framework for
evaluating the impacts of each of the alternatives and for determining National Pollutant Discharge
Elimination System permit conditions. EPA's evaluation entitled A Best Professional Judgment
Evaluation Using the Ocean Discharge Criteria for Mill Tailings Disposal from the Proposed
Quartz Hill Molybdenum Mine is presented as appendix S (U.S. EPA, 1988a) of the revised draft
EIS (USDA, 1987). EPA prepared an ecological risk assessment as a supplement to its Best
Professional Judgment (BPJ) evaluation (U.S. EPA, 1988b).
7.3. CASE STUDY DESCRIPTION
7^3.1. Problem Formulation
Site Description. Quartz Hill is a proposed molybdenum mining site located within Misty
Fjords National Monument in southeast Alaska in a designated non-wilderness area (figure 7-2).
The project includes an open-pit mine with ore process facilities located in the nearby Tunnel
Creek basin. The molybdenum mine tailings would be transported from the ore processing facility
by pipeline to a submarine gravity flow outfall in either of two marine fjords (figure 7-3a, Smeaton
Bay/Wilson Arm, and figure 7-3b, Boca de Quadra).
Wilson Arm is a small embayment at the head of Smeaton Bay and is considered a
subregion of Smeaton Bay for the purposes of this evaluation. The Smeaton Bay/Wilson Arm fjord
extends 20 km from the Wilson River/Blossom River estuary to Behm Canal. An underwater sill
separates the 285-m-deep Smeaton Bay basin from deeper waters in Behm Canal (figure 7-4a).
Wilson Arm is approximately 160 m deep. Discharge of mine tailings to Smeaton Bay/Wilson
Arm is assumed to be located 1.1 km down-fjord from the Wilson River/Blossom River mudflat at
a depth of 45 m. The zone of active deposition is assumed to be the bottom of the fjord, which is
initially approximately 150 m deep. As the fjord fills, the zone of deposition will decrease until at
55 years it is approximately 75 m deep. The discharge of tailings to Wilson Arm eventually would
affect all of Smeaton Bay; thus, tailings discharge to Wilson Arm is considered synonymous with
discharge to Smeaton Bay.
Boca de Quadra is a fjord that extends from the Keta River estuary westward approximately
57 km to the Revillagigedo Channel. Underwater sills divide the fjord into inner, middle, and
outer basins approximately 8, 27, and 33 km long and 170, 400; and 375 m deep, respectively
(figure 7-4b). The tailings outfall for middle-basin Boca de Quadra will be at 45 m depth at a
distance of 6.7 km down-fjord from the mudflat at the mouth of the Keta River. For purposes of
this analysis, it is assumed that the zone of active deposition will be the bottom of the fjord, which
is approximately 150 m deep. Detailed site descriptions are included in USDA (1987) and in U.S.
EPA (1988a).
Stressors. Mine tailings discharge will average 36,290 metric tons/day (mt/d) for the first
4 to 6 years, and approximately 72,570 mt/d for the remaining 49 to 51 years of project life. This
presents approximately 99 percent of the mine materials or approximately 0.84 billion m3. To
obtain the molybdenum ore, the host rock must be crushed, the ore particles physically separated
using a flotation process, and the ore concentrated and sent to a processing plant outside the project
7-9
-------
Hydaburg
.«*. />o
Oondas Island
Legend:
o
L
20
4O
_J
Scale in Miles
Wilderness area
Non-wilderness
area
Figure 7-2. Location of Quartz Hill project (USDA, 1987)
7-10
-------
SHEATON BAY DISCHARGE LOCATIONS
Figure 7-3a. Map of Smeaton Bay alternative for tailing disposal from the proposed Quartz
Hill mining project, Alaska (USDA, 1987)
ALTERNATIVE DISCHARGE LOCATIONS
IN BOCA DE QUADRA
Figure 7-3b. Map of Boca de Quadra alternative for tailing disposal (USDA, 1987)
7-11
-------
SMEATON BAY-WILSON ARM
10
KILOMETERS
Figure 7-4a. Smeaton Bay longitudinal bathymetric section (USDA, 1987)
o •
50
100
iso •
200
250 .
*
i
i
300
350 ^
400 •
ISO
OUTER
SI1L
KfTE ISLRND
SILL
S.O 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 60.0 65.
KILOMETERS
Figure 7-4b. Boca de Quadra longitudinal bathymetric section (USDA, 1987)
7-12
-------
area. The mill tailings effluent will be composed of waste rock particles (median grain size 63
microns), freshwater, seawater, and residual milling chemicals.
The 72,570 mt/d of solids will be mixed with 98,000 tons of seawater (1:1 by weight) prior
to discharge to the fjord. The total mill tailings discharge rate will be 1.35 nrVsec prior to
predilution. For comparison, the mean annual discharge of the Keta River to Boca de Quadra is
23 m3/sec; the combined discharge of the Blossom and Wilson Rivers to Wilson Arm is 53 nrVsec.
A number of types of stressors are present, including metals and organic compounds. A discussion
of each type of stressor is provided below.
Metals. Approximately 94 percent of the waste rock particles will be quartz and feldspar
minerals. Minerals include antimony, arsenic, cadmium, chromium, cobalt, copper, iron, lead,
manganese, mercury, molybdenum, nickel, selenium, silver, vanadium, and zinc (table 7-1).
Two chemical fractions that remain after the ore is processed were selected for this
evaluation: (1) the dissolved metal fraction and (2) the paniculate extractable fraction. The
dissolved metal fraction is that concentration of metal, that is measured in the water after mixing
ore particles with process water and seawater. Concentrations of dissolved metals in the tailings
are expected to be at least one to two orders of magnitude higher than concentrations now observed
in either Boca de Quadra or Smeaton Bay. The particulate extractable fraction is the concentration
of metal that is leached during an acidification procedure described as a total recoverable method in
Ambient Water Quality Criteria for Copper (U.S. EPA, 1985). The extractable fraction is an
estimate of what additional leaching of dissolved metal may occur after the tailings are released
into the fjord. The dissolved and extractable fractions are important in assessing the risks from
metal toxicity.
Of the metals identified in the ore and tailings, silver, copper, and mercury are the most
toxic metals. Silver, copper, and lead were found in the highest concentrations. The toxicological
characteristics of the metals are described in Ambient Water Quality Criteria for Copper (U.S.
EPA, 1985). Based on the water quality criteria (table 7-1) and concentration in the tailings,
copper was selected as the metal that would pose the highest risk to aquatic organisms. In
addition, high copper concentrations were observed in the water at another mining operation in
British Columbia (Island Copper in U.S. EPA, 1988a).
Reagents. A number of reagents must be added to the milling process to control the
separation of the molybdenite from the waste minerals and diesel fuel. Ranked in order of quantity
used, the reagents are diesel fuel, M-502, methyl isobutyl carbinol (MIBC), lime, sodium silicate,
Nokes reagent, and (tied for seventh place) CMC-7 and ALFOL-6. Each reagent (table 7-2) has
been arrayed from two standpoints: (1) predicted mass loading and (2) probable aquatic toxicity.
Because of many confounding factors (often no information on persistence or degradation;
mixtures of multiple chemicals in the final tailing product; occasional presence of more than one
chemical in a given reagent, especially in diesel oil; little or no information on aquatic toxicity,
etc.), all assumptions are intentionally biased toward conservatism. The concentrations and
toxicities assumed to be present in the tailings are thus more likely to represent the "worst case,"
7-13
-------
Table 7-1. Ore and Tailings; Inorganic Content and Associated Toxkitj (U.S. EPA, 1988b)
Antimony
Arsenic
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Selenium
Silver
Vanadium
Zinc
Solid
Ore Tailings
(mg/kg) (mig/kg)
0.002
10.9
2.4
10.0
3.3
90 69.0
16,900
60 47.0
462.0
0.05
2,170 120
17.7
0.1
0.13
17,6
40 46.0
Liquid"
Tailings
6.8
15.0
34.0
35.0
1,790
120.0
' -'. '; . • . _•
1.2
1,080
290
6.6
7.0
77
EPA Acute Water
Quality Criteria
fog/L)
36.0
9.3
50.0
•
'2.9 . ' ,,>..-:-. ,. .-;
: 5.6
0.025
..'.--.• -. , ••
2.9
. • . •• V.,54. ..•..;• ', . '..
2.3
86.0
"Effluent concentration (USDA, 1987; appendix F, table F-2),
7-14
-------
Table 7-2. Reagents Used in Quartz Hill Mining Project (U.S. EPA, 1988a)
., ;^ *V,=« • \' . •>, -,;
'•;"';•>.'•» '• • •':
Reagent
Diesel #2 fuel oil
M-502
MfflC
Lime '
Sodium silicate
Nokes reagent
CMC-7
ALFOL-6
-.• 4-.-t Use per \ •-'•; v-..-i
80,000 Tons/Day
Ob/day) ;AK
50,720
15,920 ^
12,800 ':
10,720 f -
5,040
4,320-,
3,600 W
3,600 ;
{^;4
Application per <
^^n Ore W>-*** >
in , 0.634
: 0.199
'•'• 0.160
?; 0.134
0.063
0.054 *•'-•
># 0.045
y^ 0.045 "V
Approximate
LC50 (mg/L)
0.1-5^0a
1.0b -
i.ob ,,*>-;-
1.0b-
5.ob ;?.-?
0.002° / H
5.0b
5.0b '" '
aLC50 from literature values. v-.'.!.v,i ;? ; aiv^v «V
bNo information 6ii aquatic toxicity in literature; LD50 is assumed based on conservative assumptions.
cAssumes Nokes reagent disassociates tb H2S. ! •'•"
-------
rather than the actual concentrations. Reagents used in the Quartz Hill mining project are
described below.
Diesel fuel is the most commonly utilized reagent at the site. The diesel is used as a
collector of molybdenum, which serves to make the molybdenite particles more hydrophobic and
thus more floatable. Multiple ingredients in the diesel oil, such as polycyclic aromatic
hydrocarbons (PAHs) and other organics, and the differing solubilities and acute and chronic
toxicities of these ingredients make it difficult to estimate the effects that may occur when aquatic
organisms are exposed during discharge of the tailings. Also, most of the relatively sparse aquatic
toxicity data on diesel fuel deal with the water-soluble fractions having lower molecular weights.
One is therefore forced to assume a duality of sorts with diesel, in that some of the material will be
associated with the water column, while other portions will seek out a less hydrophilic environment
(e.g., the marine microlayer, lipids hi the biota). The literature indicates LC50 ranges from 0.01
to 5 mg/L for aquatic toxicity.
M-502 is the trade name for a cationic, quaternary ammonium salt polymer that is used as
a flocculent, to facilitate the settling out of solids in the separation process. Its predicted daily use
is 7,221 kg. Only 10 percent of the M-502 flocculent is predicted to be lost to the environment
daily, due to a great affinity for the clay portion of the tailings. There is no specific information
on aquatic toxicity of this material. However, quaternary ammonium compounds are usually
highly reactive with tissue and can usually be irritants in a general sense. They also could have
other effects on water chemistry, such as buffering, alterations of pH, and so forth.
MIBC may also be used as a flocculent in the ore separation process. Most of the material
is predicted to remain with the liquid phase. Moreover, because of its high vapor pressure, MIBC
is predicted to have a high loss rate via volatilization to the air as ketone (assuming this is in the
form of methyl isobutyl ketone or MIBC). Because of the many uncertainties (e.g., whether cold
water temperatures negate some of the volatile tendencies of the material), and because of its
relatively high predicted use, the loading will be greater than the 10 percent that is predicted.
Nothing could be found in the available literature regarding aquatic toxicity of this material.
Lime is calcium oxide and is to be used at a daily predicted rate of 4,863 kg/day. It is
used as a pH modifier in the separation process. According to the literature, increasing the pH of
the solution to 8.7 makes iron sulfides more floatable, aiding in the separation and purification of
the molybdenum. Calcium oxide is predicted to remain with the liquid portion of the process. The
portions released to the environment would be expected to be readily buffered by the receiving
marine waters. The aquatic toxicity of calcium oxide is probably more indirectly related to its
effect on pH than to other factors.
Sodium silicate is also known as "water glass" and is composed of a complex of SiO2 with
Na2O. It is used as a flotation regulator and as a gangue (slime) depressant, acting to depress
slime via electrostatic charge. At least 50 percent of the material can be expected to remain with
the tailings. Information on the aquatic toxicity of sodium silicate is not available.
Nokes reagent consists of 43.4 percent phosphorus pentasulfide and 56.6 percent sodium
hydroxide. It is extremely toxic and irritating. Nokes reagent is used in the milling process as a
7-16
-------
flotation regulator by generating sulfhydryl ion (SH~), which acts to depress unwanted metals. The
ability of this material to generate SH" is central to its toxicity. The toxicity of the material, as
SH", is assumed to be very significant. In the absence of other data on Nokes reagent per se, the
EPA water quality criterion for sulfide/hydrogen sulfide is used for comparison purposes, assuming
that the SH" generated by the process will follow equilibria in water similar to the known
dissociation reactions for H2S-HS" set forth in the development of the EPA criterion. The criterion
is 2 /ig/L (or 0.002 mg/L) as undissociated H2S. If the concentration exceeds this criterion, there
may be a resultant toxic impact.
CMC-7 is used as a gangue depressant and flotation regulator (as with sodium silicate
mentioned previously). It is composed of sodium carboxyl methyl cellulose. At least 50 percent
of CMC-7 will be associated with the tailings. Information on the degradation and toxicity of the
material is not available.
ALFOL-6 is 1-hexanol, or 1-tetradecanol. It is expected to have a long life in the liquid
phase. ALFOL-6 is intended for use as a substitute flocculent (possibly in lieu of MIBC discussed
earlier), and its alcohol structure suggests that it would tend to have a reasonable affinity for water.
The aquatic toxicity of ALFOL-6 is unknown.
Chemical analyses of tailings prior to dilution with seawater detected no priority pollutants
above 12 jtg/L (U.S. EPA, 1988a). This detection limit is higher than the toxicity levels for most
chemicals of concern. Also, no reagent standards were analyzed. Thus, the evidence suggesting
that reagents will not be present in the effluent is not substantiated by the preliminary analysis. A
worst-case analysis would suggest that the chemicals may be present at toxic concentrations.
Further analyses with lower detection limits and chemicals outside the priority pollutant category
are needed to reduce the uncertainty. A detailed discussion of the potential effects of reagents is
given on pages 17 to 24 of EPA's BPJ report (1988a). While it is clear that exposure to reagents
may be stressful to the aquatic populations, these chemicals were not included in the quantitative
risk assessment. Since this is a comparative risk assessment, an estimation of absolute risks due to
exposure from all stressors was not deemed necessary. However, the limited information on
toxicity of the reagents is a factor that should be included as part of the biological tests that must
be completed in order to fully understand the effects of tailings disposal on these aquatic
ecosystems.
Settled solids. Settled solids are assumed to be that fraction of tailings that is greater than
10 microns hi diameter and settles according to theoretical predictions (U.S. EPA, 1988a). Settled
solids are potentially harmful to the benthic biota of the fjords because they may smother or bury
resident populations and their habitats and prevent community development. Chemical changes in
the sediments may also result in long-term leaching of contaminants from the tailings.
Suspended solids. The suspended solids portion of tailings is assumed to be the 10 percent
fraction with the smallest diameter (median = 5 /zm) that is injected into the water column along
the axis of each fjord (U.S. EPA, 1988a). The suspended solids present a potential harm to
pelagic organisms due to their interference with normal metabolic processes (respiration,
photosynthesis) as well as toxicity of contaminants adsorbed to the particles. Several studies have
demonstrated the tolerance of juvenile salmonids to suspended solids. These studies (Noggle,
7-17
-------
1978; Smith, 1978; Ross, 1982) measured 96-hour LC50s ranging from 1,500 to 54,000 mg/L of
suspended solids. One additional concern raised with respect to juvenile salmonids is the
possibility of eating zooplankton that are covered with participates associated with toxic chemicals
(as opposed to those zooplankton that have ingested contaminated particles). Because of limited
information on the effects of suspended solids, adverse effects are assumed to occur when the
concentrations exceed background levels.
Ecological Components. A detailed discussion of the aquatic populations and their habitats
is presented in EPA's BPJ evaluation (1988a). The populations that were considered as targets for
environmental impacts were salmon, herring, benthic invertebrates, and plankton. Salmon and
herring were selected for their high commercial and recreational value. Benthic invertebrates were
selected for their importance as commercial species (crabs, shrimps) as well as their position as
primary prey for the valued predators (salmon, herring). The plankton were chosen because of
their importance as a primary food source for the invertebrates as well as being the early life stages
of adult species (herring larvae). Marine mammals and birds and aquatic plants were excluded
because of lack of data on the populations or species that may be at risk.
Phytoplankton. Phytoplankton inhabit the upper water column (0 to 25 m) where there is
adequate light for photosynthesis; phytoplankton blooms generally occur from March to August.
Primary production is limited to depths above 8 to 25 m in both bays. Phytoplankton abundance
and primary productivity are the same for each bay.
Zooplankton. Zooplankton inhabit a depth of 0 to 150 m. Copepods are the dominant
zooplankton group. There appears to be some difference in the abundance of zooplankton between
bays. Herbivorous zooplankton, predatory medusae, ctenophores, and chaetognaths dominate the
shallower (0 to 25 m) epipelagic water. The euphausiids and amphipods inhabit the deep (50 to
150 m) mesopelagic zones.
Fish. Seventy-five species of fish have been identified from near shore, pelagic, and
benthic habitats of both fjords. Dominant species of the near-shore habitat are juvenile salmon,
juvenile herring, and starry flounder. Pelagic habits are utilized by herring, salmon, and an
abundance of larval fishes. Over 40 species of demersal fish have been identified from the benthic
habitat in areas less than 150 m deep. The biomass (kg/km) of demersal fish was greater for Boca
de Quadra than for Smeaton Bay.
Herring. Both bays are important rearing habitats for young-of-the-year and age-1 herring.
During the winter, herring descend to depths of 125 to 150 m (U.S. EPA, 1988a). This appears to
be their preferred depth even when the water is deeper. Approximately 10 to 15 percent of the
herring in southeast Alaska spawn at the Kah Shakes spawning ground at the mouth of Boca de
Quadra. Due to the proximity of Boca de Quadra to the Kah Shakes spawning grounds, there is
speculation that large populations of herring may inhabit the fjord (U.S. EPA, 1988a).
Salmon. Adult salmon are abundant in both fjords. The tributaries to Smeaton Bay
support a much larger salmon run (1.4 million salmon) than the tributaries to Boca de Quadra
(0.4 million salmon). Juvenile salmon probably stay in the top 10 to 20 m of the water column in
saltwater (U.S. EPA, 1988a). Other studies (Straty, 1974) have shown that juvenile sockeye
7-18
-------
salmon are captured within the top 5 m of the water column during daylight and evening hours. In
the sampling from Boca de Quadra and Smeaton Bay/Wilson Ann, it appears that the salmon were
located in the 0 to 20 m depths as expected. The numbers also appear to be greatest during the
summer migration period (March to August). The winter populations are unknown.
Benthic invertebrates. Benthic invertebrates inhabit a wide variety of niches in both fjords
from 0 to 300 m (table 7-3). Rocky intertidal, rocky subtidal, soft-bottom intertidal, and soft-
bottom subtidal benthic habitats occur in both bays. The dominant benthic habitat is the subtidal
soft-bottom habitat, which accounts for the entire bottom below 35 m. The subtidal biological
communities are characterized by distinct shallow (20 to 100 m), mid-depth (100 to 200 m), and
deep (greater than 200 m) benthic assemblages. Shallow and mid-depth communities are more
productive (a higher number of individuals) than deep communities. Mid-depth and deep
communities can be more diverse (Shannon-Weiner species diversity index) than shallow
communities because of a more even distribution of species' abundances. Total infaunal biomass is
greatest in deep communities because of the presence of large deposit-feeding heart urchins.
Subtidal epifaunal assemblages were distributed according to depth. In general, shallow and mid-
depth epifaunal assemblages were richer in number of taxa than deep assemblages. Dungeness
crab and pandilid shrimp were most abundant in trawl catches from inner-basin Boca de Quadra
than the shallower depths of Smeaton Bay. Shrimp caught in pots were most abundant along the
sides of fjords at shallow and mid-depths (0 to 150 m). Total epifaunal biomass was greatest in the
middle basin of Boca de Quadra and the deep areas of Smeaton Bay because of heart urchin and
mud star populations.
The infaunal and epifaunal benthic assemblages and species composition were similar for
both basins. Limited sampling (two consecutive sampling periods per year) suggests that shallow
and mid-depth infaunal communities in Smeaton Bay may be more productive (number of
individuals) than comparable communities in Boca de Quadra's inner basin.
Food Webs. The major sources of energy to the epipelagic, near-shore, and estuarine
habitats are solar radiation for photosynthesis and detritus from terrestrial and aquatic sources.
Most energy flow (detritus and prey organisms) is downward through the water column. Deep
benthic habitats contribute relatively little to habitats in the upper 100 m. Benthic infauna,
epibenthic species of commercial value, and demersal fishes are significantly more abundant in
shallow (20 to 100 m) and mid-depth (100 to 200 m) soft-bottom habitats of both basins.
Critical Habitats. Abiotic factors affecting population distributions are important
measurements for predicting the likelihood of exposure. Abiotic factors that are important in the
development of biological communities include substrate types, depths, riverine loading of
sediments, nutrients, organic carbon, freshwater intrusion, tidal fluctuations, depth of solar
radiation, temperature, and development of a pycnocline during spring and summer.
All pelagic and benthic habitats less than 100 m deep are important to the functioning of the
marine community above this depth. The most densely populated and stable part of the
mesopelagic habitat is between 50 and 150 m depths. Rocky intertidal and rocky subtidal habitats
down to 10 m are an important habitat and source of food for many benthic species. Salmon occur
in the upper 20 m of the water column and herring penetrate down to 150 m. The shallow (to
7-19
-------
Table 7-3. General Characteristics of Benthic Habitats in Boca de Quadra and Smeaton Bay
(VTN Environmental Consulting, 1983b, table 4.2.1., as reported in EPA, 1988a)
Habitat
Rocky intertidal
Soft-bottom
intertidal
Rocky subtidal
Soft-bottom
subtidal
Zone
High intertidal and gradual
slopes
Low intertidal and all
slopes
High intertidal
Middle intertidal
Low intertidal
Vertical walls 0-3 m
3-7 m
7-10 m
Gradual slopes 0-2 m
2-10 m
20-100 m
100-200 m
200-330 m
Location
Throughout fjords
Throughout fjords
Keta and Wilson River mud
flats
Keta and Wilson River mud
flats
Keta and Wilson River mud
flats
Throughout fjords
Throughout fjords
BQ inner basin and
Smeaton Bay
BQ inner basin and
Smeaton Bay
BQ middle and outer basins
Characteristic Organisms8
Rockweed
Barnacles, mussels
Sedge, insects
Rockweed, amphipods
Polychaetes, bivalves,
harpacticoids, eelgrass (Wilson
mud flats only)
Red algae, barnacles, sea urchins,
sea stars
Kelps, red and brown crustose
algae, gastropods
Brachiopods, tunicates
Eelgrass
Sea stars, bivalves
Polychaetes, bivalves, Dungeness
crabs, Tanner crabs, pandalid
shrimps, pinch bug crabs
Polychaetes, bivalves, Tanner
crabs, pandalid shrimps
Polychaetes, bivalves, sidestripe
shrimp, Tanner crabs, heart
urchins, mud stars
"Larger organisms are listed here.
7-20
-------
100 m) subtidal soft-bottom benthic assemblage is the most productive of the subtidal areas. This
habitat type is used more by commercially valuable crabs than are deeper subtidal soft-bottom
areas.
Endpoints. The assessment endpoint was the loss of critical habitat for benthic
invertebrates and fish and the potential for population effects due to discharge of toxicants.
Measurement endpoints included model estimates of benthic habitat loss due to tailings deposition,
predicted exceedance of the water quality criterion for copper, and predicted exceedance of
background suspended solid concentrations.
Comments on Problem Formulation
Strengths of the case study include:
• Hie case study identified as stressors the discharge of mine tailings and associated
reagents used in the ore separation process.
*The case study lists a wide range of important species including salmon, herring,
benthic invertebrates, and plankton. The habitats of the ecological components are
generally sufficiently characterized. Special emphasis is given to quantifying the
potential loss of benthic organisms in the two fiords.
+ The measurement endpoints are clearly defined and are sufficient for a relative
comparison of risk between the two floras.
Limitations include:
•The case study does not adequately explain the risk assessment setting, where it is
located, and whether or not the fjords are unique or contain special populations.
The assessment only considers potential copper toxicity and the covering of benthic
habitat. Effects due to other metals and reagents used in the extraction process are
not considered because of lack of data. The exclusion of factors because of lack of
data or poor data quality has the potential effect of underestimating risk. Poor data
are in effect "rewarded" by being dropped from the study. Since the case study
focuses only on identifying which basin is at least risk, stressors are not described
completely or evaluated throughout the case study. Catastrophic release of reagents
and long-term effects of the project are not addressed.
• While -water column ecological components are identified, they are not carried
through to the later analysis.
• The measurement endpoints are insufficient to determine absolute risk, which would
have to be addressed if tailing discharges are to be allowed at all.
7-21
-------
7.3.2. Analysis: Characterization of Ecological Effects
The relationship between a given water concentration and a predicted biological response is
determined from a series of laboratory tests with specific chemicals and selected organisms. The
EPA water quality criteria are based on such a series of tests. The criteria were developed as a
means of protecting aquatic communities in any given environment without requiring detailed
knowledge of the responses of individual species inhabiting the area. Thus, while concentration-
response curves, effective concentration (EC50), lethal concentration (LC^o), and acute or chronic
studies are not available for all species and chemical forms, the criteria are reasonable guides for a
water quality analysis.
Bioassays. Bioassays on a variety of organisms to evaluate acute, chronic, and sublethal
effects were completed at this site and with mine tailings from similar projects. A detailed
discussion of all bioassays performed on tailings materials is provided in EPA's BPJ report
(1988a). Initial tests with juvenile coho salmon, mussel larvae, amphipods, and euphausiids
indicated that the acute toxicity at exposure periods from 96 hours to 10 days for these species was
low. UCSO and EC50 concentrations ranged from 109,000 mg/L for the euphausiids to 208,000
mg/L for coho salmon (U.S. EPA, 1988a). Subsequent tests with Dungeness crab zoea, mussel
larvae, and amphipods also indicated relatively low toxicity. The LC50 or EC50 concentrations
observed in the studies were 170,000 mg/L for crab zoea, 142,500 mg/L for mussel larvae, and
86,000 mg/L for amphipods (U.S. EPA, 1988a). These tests represented a reasonable preliminary
effort to describe some of the possible impacts of the proposed tailings disposal plan. No
consideration was given in this first series of tests to the possible physical effects of suspended
solids, grain size, or chemical interactions (total organic carbon). Therefore, the results of these
preliminary studies are not considered a definite statement of toxicity. In addition, the studies were
done in static water. Tailings concentrations were only measured at the beginning of the test;
therefore, there was no measure of exposure after day 1.
Acute toxicity studies (U.S. EPA, 1988a) of zooplankton and euphausiids from other
mining sites with similar processing procedures indicate that suspended solid concentrations of 560
mg/L over a 40-day period are necessary before ecologically important effects are noted. It should
be noted (U.S. EPA, 1988a) that the effects were most likely due to physical stress rather than
toxicity.
Chronic and sublethal tests were completed on clam burrowing behavior and phytoplankton
growth. Estimates of absolute numbers of organisms observed colonizing an area (table 4-13,
USDA, 1987) are subject to considerable uncertainty due to sampling procedures as well as
assumptions regarding colonization. VTN Environmental Consulting (1983), as reported by EPA
(1988a), observed colonization rates of up to 25 months with test plots in Boca de Quadra.
Phytoplankton growth did not appear to be affected by exposure to tailings.
Bioaccumulation. Bioaccumulation of metals (cadmium, copper, manganese, molybdenum,
lead, and iron) from tailings was investigated over a 4-month exposure period in the laboratory
with crabs, clams, mussels, and sanddabs. Elevated tissue metal concentrations were not observed
in test organisms. No behavioral or morphological aberrations were noted. These tests did not
address the natural process of chemical uptake that would take place during feeding. Substantial
7-22
-------
uptake of copper, zinc, cadmium, and lead from tailings by the bivalve Yoldia thraciaeformis was
noted at another mine site (U.S. EPA, 1988a).
Recolonization. Studies of tailings deposition (VTN Environmental Consulting, 1983, as
reported in U.S. EPA, 1988a) for the Quartz Hill project indicate that most species would not
survive burial. However, colonization from either larval settlement or lateral migration may
replace the lost benthic communities. Few studies have been completed on the colonization of
tailings material by aquatic organisms.
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
•For a comparison of the two fjords, the characterization of ecological effects is not
highly significant. "The peer reviewers accepted the assumption that all habitats
buried would be lost. The use of the water quality criteria for copper in the water
column precludes the need for developing an exposure-effects relationship for this
endpoint.
Limitations include:
•Although the characterization of effects was attempted for other endpoints, data
were judged inadequate for use. The most serious inadequacy is the lack of tailings
toxicity data.
7.3.3. Analysis: Characterization of Exposure
Oceanographic Processes. A steady-state model, using the distribution of natural and
humanmade conditions, provided the basic framework for estimating the distribution of stressors.
Exposure of ecological components was extrapolated from observations of population abundances
and habitat preferences.
The site descriptions included oceanographic data for existing conditions, collected and
analyzed by the University of Alaska (USDA, 1987). This information has been used to
characterize the flow and density structure in each fjord as a function of time. For each fjord, two
hydrodynamic seasons were identified (USDA, 1987): (1) summer renewal and (2) winter
nonrenewal. For the density structure, six 2-month seasonal periods were identified, based on the
analysis of the hydrography in each fjord (U.S. EPA, 1988b). The fjords were compartmentalized
spatially as well as seasonally. Based on their hydrographic and hydrodynamic characteristics as
well as their biological habitats, each fjord was divided into 12 subregions. The hydrodynamic and
hydrographic information was then used to construct a model of the oceanographic processes that
7-23
-------
could lead to changes in the physical, chemical, and biological states of the two fjords and the
likelihood of exposure of organisms to stressors.
Magnitude. Frequency, and Duration of Exposure. Environmental factors that may result
in variability in the distribution of settled solids include average and maximum short slope stability,
episodic slumping, and in situ compacted density of tailings.
Suspended solid distributions are affected by vertical mixing and upwelling, rate of
exchange of water, breaking of internal waves, formation of an upper-level plume, and down-fjord
turbidity currents. Statistical uncertainty analysis (Monte Carlo method, U.S. EPA, 1988b) was
performed on the interaction of each of these factors, resulting in probabilities, rather than discrete
estimates of concentrations. Simulations that included system variability were performed to
represent the seasonal conditions throughout 1 year, for each of 3 years in the life cycle of the
project. These three periods—year 5, year 20, and year 55—were chosen to represent the initial,
intermediate, and final stages of the project. Two distinct seasons (whiter and summer) were
chosen to keep the problem as simple as possible. The maximum suspended solid concentrations
(table 7-4) indicate that the higher concentrations reaching the upper water column (above 100 m)
are greater for Smeaton Bay/Wilson Arm than for middle-basin Boca de Quadra.
The dissolved and total copper concentrations were measured in several laboratory tests
(USDA, 1987). Most studies of environmental toxicity are based on dissolved metal
concentrations. However, in developing the documentation on water quality criteria, EPA
attempted to account for the possibility that some of the metal content of paniculate matter may
leach when solids are exposed to ambient water conditions. It is an assumption, therefore, that
when applying the water quality criteria in risk analysis, the appropriate form of metal content
should be the acid-soluble form. For purposes of evaluating the potential of copper toxicity, the
total recoverable fraction as well as the dissolved fraction should be analyzed. The total
recoverable metal fraction was not measured according to EPA protocols. However, an alternate
method specified as an "extractable" tailings characterization was presented in the USDA revised
draft EIS (USDA, 1987, appendix F, table F-6). The "extractable" portion of tailings may be
converted to a water column concentration by the following equation:
Tailings Metal Extractable x Water Column _ Water Column Metal
Portion (mg/mg) Suspended Solid Extractable Concentration
Concentration (mg/L) (mg/L)
This "extractable" fraction and the dissolved fraction were taken in sum as the "total
recoverable concentration" (USDA, 1987, appendix E).
In order to provide a basis for comparing the risk due to exposure to rnetals in either fjord,
100 percent of the copper content (90 ng/g) of the ore and 44 percent (40 /ig/g) of the copper
content of the ore were selected (table 7-5) as representative of the range of maximum copper
concentrations that may be experienced during the life of the mine. While these copper
concentrations may not be accurate, they represent in the first case the actual measured maximum
copper content of the ore (total copper) and in the second case an estimate of the average
extractable portion of total ore for all metal fractions (reported in USDA, 1987, appendix F, from
7-24
-------
Table 7-4. Maximum Suspended Solid Concentration Predicted for Upper Water Column
(above 100 Meters) of Smeaton Bay/Wilson Arm and Boca de Quadra (U.S. EPA,
1988b)
Maximum Suspended Solid Concentrations (mg/L)
Year
5
20
55
Smeaton Bay
74
160
170
Boca de Quadra
56
65
65
Table 7-5. Estimation of Extractable Copper Concentration (U.S. EPA, 1988b)a
Smeaton Bay/Wilson Arm
Total ore
Total tailings
44% of ore
44% of tailings
Tailings extraction
Extractable
Portion
(NT3 mg/mg)
0.09
0.069
0.04
0.03
0.022
Extractable
Concentration
Otg/D
13.5
10.4
6.0
4.5
3.3
Boca de Quadra
Extractable
Portion
(10-3 mg/mg)
0.09
0.069
0.04
0.03
0.022
Extractable
Concentration
Otg/L)
5.2
4.0
2.3
1.7
1.3
a From a maximum suspended solid concentration (150 mg/L) in the upper 100 meters of Smeaton
Bay/Wilson Arm at year 55 and the maximum suspended solid concentration (58 mg/L) in the upper
100 meters of middle-basin Boca de Quadra at year 55.
7-25
-------
Burrell, 1983). These estimates of copper content were used only for comparison with the water
quality criterion and are not necessarily the true concentration of extractable copper that may only
be obtained by chemical analysis of tailings using the EPA-prescribed total recoverable method.
The predicted concentration of total recoverable copper in the upper water column (above 100 m)
is higher for Smeaton Bay/Wilson Arm than for Boca de Quadra.
Comments on Analysis: Characterization of Exposure
Strengths of the case study include:
• The dispersion of tailings on the bottom and in the water column is studied with
numerical models that contain most of the relevant physical transport mechanisms.
Concentrations of copper are computed at two levels representing best-guess and
worst-case scenarios.
Limitations include:
*Model verification or comparison is not presented in the case study although some
comparisons are discussed in the supporting documents. The exposure assessment
does not consider low-density hydrocarbons that might disperse in the surface layer.
The model does not include chemical processes.
7.3.4. Risk Characterization
A detailed discussion of the risk characterization is presented in U.S. EPA, 1988b. The
approach used in this study is to define the boundary of that space in each fjord for which the
estimated total recoverable copper exceeds the criterion of 2.9 /t/L. The estimated probability of
water quality exceedances at any location during a given season is the total number of exceedances
at that location divided by the total number of simulations performed. Results indicated that the
water quality criterion for copper would be exceeded in both fjords in the deeper water (greater
than 100 m) during all years of the proposed project. The likelihood of surpassing the criterion
increases with higher levels of extractable copper (from 44 percent to 100 percent) and as the
project proceeds to year 55 (figures 7-5a, 7-5b, and 7-5c). The probability of exceeding the
criterion increases in the upper water column of Smeaton Bay/Wilson Arm as the project proceeds
to year 55. This is not true for Boca de Quadra (figures 7-5a, 7-5b, and 7-5c). The prediction of
lower levels of suspended solids in the upper water column of Boca de Quadra results in a decrease
in the concomitant copper concentration.
The response of aquatic organisms when exposed to copper, suspended solids, and settled
solids was predicted by calculating the maximum loss of organisms due to avoidance or burial
(settled solids), or mortality due to exposure to copper. Since water quality criteria are derived as
a single reference dose rather than a range of doses from a dose/response curve, no data are
available to estimate probabilities of responses other than no effect or a 100-percent response (i.e.,
7-26
-------
Probability of Exceeding Copper Criterion
Smeaton Bay - Year 5
100
200
300
0 Kilometers 5
Risk Level
<0.01
0.01-0.20
0.20-0.40
0.40-0.60
0.60-0.80
0. 80-1.00
Deposition Zone
Probability of Exceeding Copper Criterion
Boca de Quadra - Year 5
100
200
30 D
400
-I
Kiloneters 10
R i sk LeveI
<0. 01
0.01-0.20
0.20-0,40
0.40-0.60
0.60-0.80
0.80-1.00
Deposition Zone
Figure 7-5a. Probability of exceeding water quality criterion—Year 5 (data from U.S. EPA, 1988b)
7-27
-------
Probability of Exceeding Copper Criterion
Smeaton Bay - Year 20
RisX Level
<0.01
o.oi-o.20
o.20-0.40
0. 40-0. 60
0.60-0.80
0.80-1.00
Deposition. Zone
Probability of Exceeding Copper Criterion
Boca de Quadra - Year 20
100
200
300
•400
I
I
I
J
Kilometers 10
Risk Level
<0,D1
0.01-0.20 •
0.20-0.40
0.40-0.60
0.60-0.80
0.80-1.00
Deposition Zone
Figure 7-5b. ProbabiUty of exceeding water quality criterion—Year 20 (data from U.S. EPA, 1988b)
7-28
-------
Probability of Exceeding Copper Criterion
Smeaton Bay - Year 55
100
•200
300
Kiloneters
Risk Level
<0.01
0.01-0.20
0.20-0.-SO
0.40-0.60
0.60-0,80
0.80-1.00
Deposition Zone
Probability of Exceeding Copper Criterion
Boca de Quadra - Year 55
100
200
3QO
Kiloaeters 10
Risk Level
<0.01 .
0.01-0.20
0.20-0.40
0.40-0.60
0.60-0.80
0.80-1.00
Deposition Zone
Figure 7-5c. Probability of exceeding water quality criterion—Year 55 (data from U.S. EPA, 1988b)
7-29
-------
either above or below the criterion for copper). In order to estimate risk, an exposure
concentration in excess of an ambient water quality criterion was assumed to cause 100-percent
mortality of the exposed animals.
The critical habitat for pelagic organisms is above 100 m in both fjords. The impact of
tailings disposal was evaluated with respect to the effect on this important habitat. As the
project proceeds to the 55-year point, the concentration of suspended solids and concomitant
copper will increase in the upper water column, presenting a higher probability of harm to pelagic
organisms such as herring and salmon.
The resulting probability estimates for surpassing the water quality criterion for copper
show that extractable copper concentrations will exceed water quality standards over a large part of
whichever fjord is chosen for the disposal site. In Boca de Quadra, the average concentration of
suspended solids and copper is higher than in Smeaton Bay/Wilson Arm, but the potential impact
upon biota may be greater in Smeaton Bay/Wilson Arm due to the fact that the high concentrations
occur hi the upper water column where there is likely to be more biological activity. The numbers
of zooplankton and euphausiids are also greater in Smeaton Bay, reinforcing the likelihood that the
risks are greater for biological effects in Smeaton Bay.
The risk due to exposure to settled solids is presented in terms of number of benthic
organisms lost per hectare covered with tailings. The adverse effect may be death due to
smothering or burial or dislocation due to loss of suitable habitat. Loss of habitat is considered to
be equal to death by burial. The loss of benthic organisms due to deposition of tailings is
approximately four times greater (table 7-6) for the Smeaton Bay/Wilson Arm disposal option than
for the disposal in Boca de Quadra. Smeaton Bay/Wilson Arm tailings disposal will result in
adverse impacts to 1,660 ha of habitat compared with 1,600 ha adversely affected with the Boca de
Quadra disposal option (U.S. EPA, 1988a).
Herring habitat affected in Smeaton Bay/Wilson Arm will be 320 ha compared with only 20
ha affected in Boca de Quadra (U.S. EPA, 1988a). If recolonization is slow, the loss will be
substantial. It is likely that actual colonization will take much longer than predicted from
laboratory studies, given the distance from a source of organisms, sediment slumping, sediment
toxicity from metals and reagents, the constantly shifting course of undersea channels, and the
continuing deposition of tailings after the organisms have adjusted to a certain distribution of settled
tailings.
The two fjords also will respond differently to changes in geometry as the project evolves.
Little change occurs hi Boca de Quadra with time. In Smeaton Bay/Wilson Arm, there is a
noticeable change in bottom geometry, specifically in the location of the discharge point as the
project progresses. In Smeaton Bay/Wilson Arm, the discharge point is within 75 to 100 m of the
surface during the project's final stage. It is, therefore, not surprising that the model predicts
increasing impacts hi the upper water column as time progresses.
It may be assumed that impacts (reagent toxicity, other metal toxicity) that are not
addressed in this quantitative statement of risk due to oceanographic variation in the two fjords will
also increase the risk estimate for both fjords. Since measurements of other agents were not
7-30
-------
Table 7-6. Worst-Case Estimates of Biomass (kg) of Demersal Organisms Lost Over Life (55
Years) of Proposed Quartz Hill Mining Project (U.S. EPA, 1988b)a
Dungeness Tanner Pot Trawl Walleye
Crab Crab Shrimp Shrimp Pollock Rockfish Flatfish
Boca de
Quadra 22,500 13,750 40,150 45,100 45,100 30,800 17,050
Smeaton
Bay 52,800 129,800 26,400 313,500 33,000 11,002 17,250
Total Demersal Organisms Lost
Boca de
Quadra 175,400
Smeaton
Bay 773,850
aThe organisms are lost due to smothering, avoidance, or toxicity of settleable solids. Assumes
annual colonization.
7-31
-------
completed and expert knowledge regarding impacts is limited, the expected increase in risk
estimates is not included in this analysis.
Uncertainty. Statistical uncertainty analysis (U.S. EPA, 1988b, appendix A) was used to
characterize the natural variability in the oceanographic characteristics of each fjord. The result of
this characterization is a statement of certain differences in the hydrodynamics and hydrography of
each fjord. This oceanographic description of each fjord was used to predict the concentration and
distribution of copper, settled solids, and suspended solids. Verification of the steady-state
diffusion model used to predict copper and suspended solid concentrations hi the basins was not
completed. However, comparison with other mining operations (table 7-7) suggests that the model
predictions are in fact close to realistic estimates of contaminant dispersion.
Comparison with Natural Background. Concentrations of dissolved metals in the tailings
are expected to be at least one to two orders of magnitude higher than concentrations now observed
hi either Boca de Quadra or Smeaton Bay.
Predicted levels of suspended solids have been compared with the ambient concentrations
measured in earlier studies (U.S. EPA, 1988b). These studies found a maximum of 5 mg/L, with
average concentrations less than 1 mg/L for Boca de Quadra. Due to limited data on the "natural
condition," the risk estimate for suspended solids is presented only as a comparison to available
ambient concentrations rather than as a probability statement of potential harm to aquatic
organisms. The average suspended solid concentrations that are predicted for tailings discharges to
middle-basin Boca de Quadra and Smeaton Bay/Wilson Arm exceed natural ambient conditions
throughout both fjords (U.S. EPA, 1988b, figures 9 and 10 in appendix A).
7-32
-------
Table 7-7. Comparison of Projects in North America That Discharge Mining Tailings to
Marine Waters (U.S. EPA, 1988b)
Project
Location
Duration
(years)
Maximum
Discharge
(tons/day)
Island Copper
Kitsault Molybdenum
Quartz Hill
Rupert Inlet, B.C.
Alice Arm, B.C.
Alaska
16
1.5
55
60,000
15,000
80,000
7-33
-------
Comments on Risk Characterization
Strengths of the case study include:
comparison of relative risks in two basins, the risk is adequately
characterized, and Boca de Quadra appears to be less at risk than Smeaton Bay.
Risk of exposure to solids is presented in terms of number ofbenthic organisms
lost per unit area covered. Risk to the water column is presented in terms of the
percentage of times the model predicts that water quality criterion for copper
would be exceeded in the upper and lower layers of the fjords.
9 The uncertainties for oceanographic characteristics are clearly identified. The
modeling approach used to define the zone of greatest impact and to compare
impacts on the two sites is very useful. The areal extent of impacts is an important
addition to the estimate of scale for risk as well as exposure.
Limitations include:
• The discussion of aquatic populations is only addressed qualitatively.
Uncertainty analysis could be applied to the distribution of aquatic populations as
is done for the stressors. Inadequate chemical analysis of tailings and sampling of
biota increases the overall uncertainty in the analysis.
• The assessment is clearly not written as a stand-alone document and should be
considered in light of the discussions contained in the pertinent references, which
give a broader explanation of the environmental impacts.
*The risk characterization does not assess damage to plankton or fish apart from
use of water quality criteria. Risks from normal and catastrophic release of
reagents are not characterized. This information would be required to assess an
absolute level of risk.
• Uncertainty is not used consistently in the case study. This problem is common
because uncertainty can have a variety of meanings. This study uses uncertainty
in three different ways that might be categorized as absolute uncertainty, statistical
uncertainty, and relative uncertainty.
— Absolute uncertainty is the comparison of the analysis with the real
future state of the system. This uncertainty cannot be evaluated because the
future is unknown.
— Statistical uncertainty is assessed only in a rudimentary way by
identifying the percentage of times water quality standards were exceeded.
7-34
-------
Comments on Risk Characterization (continued)
— Relative uncertainty can be assessed by comparing a set of data against a
specific or abstract reference. A high relative uncertainty is assessed if
observations have known problems compared with a standard of a "good
analysis. " Relative uncertainty is implied in the case study. The relative
uncertainty is important to the study, since it is through a comparison of one
set of results with another that we gain an intuitive estimate of uncertainty.
In the case study, the relative uncertainty of the modeling study could have
been more clearly defined by pointing out that the results of the model are in
agreement with observed distributions of tailings in a deep-water discharge
mining operation on Vancouver Island.
• The reviewers felt that this case study had problems in the way the goals are
identified and in the exclusion of some elements of risk analysis on the grounds that
data are lacking or of poor quality.
+While none of the reviewers disagreed with the conclusions that discharge into Boca
de Quadra involved the least risk, they did take strong exception to the contention of
the case study that only relative risk and not absolute risks are necessary to consider
for the completion of the risk analysis.
General comments:
• The regulatory decision rested on which alternative was the least environmentally
damaging; therefore, absolute risks were not necessary for completion of this risk
analysis. Sampling and analysis were completed as part of the Environmental Impact
Statement prior to the initiation of this risk assessment. Since the data were collected
without having first developed an hypothesis or completing a planning assessment,
much of the information that would be needed to do a thorough analysis of risk was
not available. If one accepts risk assessment as an iterative process, this assessment
should be viewed as a first-order iteration. However, decision-making under
conditions of uncertainty is a fact of life in the management of environmental
resources. The project evaluation was considered complete at this stage and a
decision was made not to allow discharge of tailings into Smeaton Bay.
^Hypothetical considerations of food chain and ecosystem effects should provide
some perspective on the importance of impacts at lower organizational levels.
• Using decision analysis prior to the initiation of the investigation would have been
useful in identifying the parameters for which extensive sampling or laboratory
analysis would have reduced the risk characterization uncertainty.
7-35
-------
Comments on Risk Characterization (continued)
^Although the regulatory decision was dependent on relative comparisons, a more
thorough discussion of other metals and reagents would have strengthened the
comparisons as well as achieved a better understanding of the ecosystem level of risk.
7-36
-------
7,4. REFERENCES
Burrell, D.C. (1983) The biogeochemistry of Boca de Quadra and Smeaton Bay, southeast Alaska,
a summary report on investigations, 1980-1983. Prepared for U.S. Borax and Chemical
Corp. and Pacific Coast Molybdenum Co., December 1983.
Noggle, C.C. (1978) Behavioral, physiological, and lethal effects of suspended sediments on
juvenile salmonids. M.S. thesis, Univ. of Washington. 87 pp.
Ross, B.D. (1982) Effects of suspended volcanic sediments on coho and fall chinook salmon
smolts. M.S. thesis, Univ. of Washington. 128 pp.
Smith, D.W. (1978) Tolerance of juvenile chum salmon to suspended sediments. M.S. thesis,
Univ. of Washington. 124 pp.
Straty, R.R. (1974) Ecology and behavior of juvenile sockeye salmon in Bristol Bay and the
eastern Bering Sea. In: Hood, D.W.; Kelley, E.J., eds. Oceanography of the Bering Sea.
Institute of Marine Science, Univ. of Alaska, p. 320.
U.S. Department of Agriculture. (1987) Quartz Hill molybdenum project mine development,
revised draft Environmental Impact Statement, U.S. Forest Service Tongass National Forest
R10-MB-7a.
U.S. Environmental Protection Agency. (1985) Ambient water quality criteria for copper—1984.
EPA 440/5-84/031. 142pp.
U.S. Environmental Protection Agency. (1988a) A BPJ evaluation using the ocean discharge
criteria for mill tailings disposal from the proposed Quartz Hill molybdenum mine. U.S.
EPA, Region 10, Seattle, WA (addendum to USDA, 1987).
U.S. Environmental Protection Agency. (1988b) A risk analysis for Quartz Hill, Alaska. Final
draft ecological risk assessment Quartz Hill molybdenum mining project. U.S. EPA,
Region 10, Seattle, WA (addendum to USDA, 1987).
7-37
-------
-------
I
SECTION EIGHT
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
ASSESSING ECOLOGICAL RISK AT ROCKY MOUNTAIN ARSENAL
-------
AUTHORS AND REVIEWERS
AUTHORS
Carolyn L. Fordham
Terra Technologies
Golden, CO
REVIEWERS
Nancy A. Bryant (Lead Reviewer)
ENSR Consulting & Engineering
Acton, MA
Kenneth L. Dickson
Institute of Applied Sciences
University of North Texas
Denton, TX
Thomas M. Frost
Center for Limnology
University of Wisconsin
Madison, WI
Judy L. Meyer
Institute of Ecology
University of Georgia
Athens, GA
Douglas P. Reagan
Woodward-Clyde Consultants
Denver, CO
Douglas P. Ormrod
Office of Graduate Studies
University of Guelph
Guelph, Ontario, Canada
Randall S. Wentsel
Toxicology Division
U.S. Army Chemical Research,
Development, and Engineering
Center
Aberdeen Proving Grounds, MD
8-2
-------
CONTENTS
ABSTRACT 8-6
8.1. RISK ASSESSMENT APPROACH 8-8
8.2. STATUTORY AND REGULATORY BACKGROUND 8-8
8.3. CASE STUDY DESCRIPTION 8-10
8.3.1. Problem Formulation 8-10
8.3.2. Analysis: Characterization of Ecological Effects 8-18
8.3.3. Analysis: Characterization of Exposure 8-27
8.3.4. Risk Characterization 8-30
8.4. REFERENCES 8-36
APPENDIX A—SUMMARY OF EQUATIONS AND UNCERTAINTY FACTORS ... 8-A1
APPENDIX B—OBSERVED TISSUE CONCENTRATIONS OF CONTAMINANTS
IN TERRESTRIAL ORGANISMS 8-B1
APPENDIX C—OBSERVED TISSUE CONCENTRATIONS IN RAPTOR SAMPLES
OF CHANCE AND BELIEVED CAUSE OF MORTALITY 8-C1
8-3
-------
LIST OF FIGURES
Figure 8-1. Structure of analysis for Rocky Mountain Arsenal 8-9
Figure 8-2. Map of Rocky Mountain Arsenal 8-11
Figure 8-3. Bald eagle food web for contaminants in soil,
sediment, and surface water source 8-22
Figure 8-4. Criteria diagram 8-23
LIST OF TABLES
Table 8-1. Maximum Concentrations of Major Contaminants
of Concern Based on 1985-87 Sampling . 8-13
Table 8-2. Summary of Assessment and Measurement Endpoints 8-16
Table 8-3. Acceptable Concentrations of Major Contaminants v
of Concern in Abiotic Media !......; 8-20
Table 8-4. Summary of Species Where Contaminants of Concern
Differed Significantly (p^O.05) or Approached Significance
Between Controls and Contaminated Areas Onsite 8-33
Table 8-A1. Uncertainty Factors Used in Establishing Acceptable
Water Concentrations 8-A7
Table 8-B1. Contaminant Levels in Terrestrial Ecosystems--
Terrestrial Program Samples 8-B2
Table 8-B2. Miscellaneous Samples: Samples of Chance and
USFWS Supplemental Samples 8-B4
Table 8-C1. Observed Tissue Concentrations in Raptor Samples
of Chance and Believed Cause of Mortality 8-C2
8-4
-------
LIST OF ACRONYMS
AChE acetylcholinesterase
ARAR applicable or relevant and appropriate requirement
AWQC Ambient Water Quality Criteria
BAF bioaccumulation factor
BCF bioconcentration factor
BMP biomagnification factor
BW body weight
t ',''"'"'
CERCLA Comprehensive Environmental Response, Compensation, and Liability Act of
1980 (Superfund)
DBCP dibromochloropropane
EPA U.S. Environmental Protection Agency
FWQC Federal Water Quality Criteria
LCso lethal concentration to 50 percent of organisms tested
LD50 lethal dose to SO percent of organisms tested
LOAEL lowest observed adverse effects level
MATC maximum acceptable tissue concentration
NCP National Contingency Plan
NOEL no observed effects level
NPL National Priorities List
RMA Rocky Mountain Arsenal
SARA Superfund Amendments and Reauthorization Act of 1986
TBC to be considered
USFWS U.S. Fish and Wildlife Service
WQS water quality standards
8-5
-------
ABSTRACT
The Rocky Mountain Arsenal (RMA) is a 27-square-mile U.S. Army installation north of
Denver, Colorado. The site was used from 1942 to 1982 by the U.S. Army and its lessees for
production of chemical and incendiary munitions, pesticides, and other chemicals. In 1985, the
Army initiated studies to determine the nature and extent of contamination in soil, ground water,
surface water, and biota. The Biota Remedial Investigation Report (ESE, 1989) summarized the
nature and extent of contamination in biota at RMA and investigated contaminant effects. By
establishing a quantified relationship between adverse effects on biota and chemical concentrations
in soil and water, studies also determined contaminant levels in the environment that would be
likely to pose a threat to wildlife on RMA. Ecological risk was not quantified in this study
because abiotic data were not fully available; thus, exposure assessment could not be completed.
Studies at RMA are ongoing.
Stressors/Contaminants of Concern. Because of RMA's varied uses in the past both as a
facility for the production of Army chemical and incendiary munitions and for commercial,
chemical manufacturing, there are many different chemicals in the environment. Seven chemicals
were selected as being of major concern to biota based on consideration of parameters such as
toxicity, areal extent, and persistence in the environment. These chemicals were aldrin, dieldrin,
endrin, isodrin, dibromochloropropane, arsenic, and mercury. An additional 32 contaminants were
also evaluated, but not to the extent of the major contaminants of concern.
Ecological Components. RMA provides important habitat for many species, including the
endangered bald eagle, which winters at RMA. Other raptors frequently observed at RMA include
golden eagles, ferruginous hawks, northern harriers, rough-legged hawks, great-horned owls, and
burrowing owls. Resident mammals include mule deer, white-tailed deer, badger, coyote,
cottontail rabbit, jackrabbit, and prairie dog. Waterfowl and wading birds occur in the lake areas.
Criteria for Evaluating Risk. Criteria based on toxicity to terrestrial and aquatic organisms
were determined for surface water and sediment by a method termed "Pathways Analysis" (ESE,
1989), which provided a quantitative means of relating contaminant concentrations in sediment and
water to concentrations and effects in biota. EPA's Ambient Water Quality Criteria for the
Protection of Freshwater Aquatic Organisms were considered in relation to toxicity to aquatic life.
Soil criteria were estimated based on toxicity only to terrestrial organisms (ESE, 1989).
Bioaccumulation was also a factor in determining levels in the environment not expected to pose a
threat to wildlife (ESE, 1989; Fordham and Reagan, 1991). The modeling approach included
quantitative estimation of uncertainty.
Endpoints. A wide array of toxicological and ecological assessment endpoints were
examined. Adverse chemical impacts and chemical contamination of ecological components at
levels that might be expected to impair structure or function were considered as toxicological
assessment endpoints. Decreased population success was considered as an ecological assessment
endpoint.
Contaminant concentrations in plant and animal tissue were one of the measurement
endpoints because of historical correlations of mortality with pesticides in tissue and because many
8-6
-------
of the contaminants of concern were bioaccumulative. Depressed activity of acetylcholinesterase
was examined in birds and mammals as another measurement endpoint for determining chemical
impacts. Population density, occurrence, and age classes were evaluated for certain species at
RMA as a measurement endpoint for determining population success. Reproductive success was
considered a measurement endpoint for avian species population success because some of the
contaminants at RMA were linked with avian reproductive effects.
8-7
-------
8.1. RISK ASSESSMENT APPROACH
The case study described herein does not represent a complete risk assessment (figure 8-1).
However, useful information is provided on problem formulation as well as on characterization of
ecological effects. Ecological risk assessment methodologies are discussed. A complete
characterization of exposure for chemical stressors is lacking but will be completed for the site as a
result of ongoing studies. Historical information and data on the distribution and concentration of
contaminants from Phase I studies of abiotic media (e.g., soil, sediment, water) were used to
determine sites of contamination on Rocky Mountain Arsenal (RMA). These initial data were used
to select sites for field investigation and to determine major contaminants of concern. Detailed
quantitative data were available for biological media only, and exposure was inferred from
measured concentrations of site-related contaminants in tissues, as opposed to measuring
concentrations in tissue compared with colocated, measured concentrations in abiotic media. This
additional step in evaluating exposure will be performed in ongoing risk assessment activities.
A major feature of this risk assessment was the development and use of an exposure
pathway model as a "new tool" to establish a quantitative relationship between concentrations of
bioaccumulative contaminants in abiotic media and concentrations at all levels in the selected food
webs. The model is used to determine if contaminant levels in abiotic media present a risk to
biota, and it can be used to establish ecologically based remediation criteria. The model was used
for both aquatic and terrestrial ecosystems at RMA, and can be adapted for use in any ecosystem.
The overall approach also involves the development of criteria for the selection of key species and
for the identification of contaminants of concern to biota.
Another major aspect of this study was the use of food webs to evaluate exposure pathways.
Species at RMA were organized into both a terrestrial and an aquatic food web. There was some
overlap between the two food webs for species, such as the bald eagle, which preys on terrestrial
mammals as well as waterfowl. The food webs aid in focusing a study design toward species that
may be most affected by exposure as a result of diet. This paper focuses primarily on the
terrestrial food web; for a discussion of the aquatic food web, see Fordham and Reagan (1991).
8.2. STATUTORY AND REGULATORY BACKGROUND
The Comprehensive Environmental Response, Compensation, and Liability Act of 1980
(CERCLA or Superfund) and CERCLA as amended by the Superfund Amendments and
Reauthorization Act of 1986 (SARA) stipulate that environmental health should be protected as well
as human health and welfare. The RMA investigation was conducted according to requirements in
the National Contingency Plan (NCP) and Guidance for Remedial Investigations (U.S. EPA, 1985).
Under CERCLA, Ambient Water Quality Criteria (AWQC) can be considered an applicable
or relevant and appropriate requirement (ARAR) or to be considered (TBC) for ecological risk
assessments. The U.S. Environmental Protection Agency (EPA) develops Federal Water Quality
Criteria (FWQC) under the authority of the Clean Water Act from the Office of Water Regulations
and Standards. These are nonenforceable guidelines that can be used by the states to determine
Water Quality Standards (WQS). The AWQC for the Protection of Freshwater Life and Their
Uses are the only criteria currently available that are specific to the protection of nonhuman life,
8-8
-------
Figure 8-1. Structure of Analysis for
Rocky Mountain Arsenal
PROBLEM FORMULATION
Stressors: seven major and 32 other chemicals.
Ecoloaical Components: veaetation. invertebrates.
/
birds, and mammals.
Endooints: assessment endpoint is reproduction and
survival of species; measurement endpoints are
chemical residues, population data, acetylcholinesterase
activity, behavior and reproductive success.
II
I
I
Characterization of Characterization of
Exposure Ecological Effects
Limited data were availab.e ^"J^S™^
on chemicals in abiotic exposure pathways and
media; data were available develop critical effect levels
for tissue residues; study for abiotic media: toxicol-
is ongoing ogical criteria or benchmarks
were used: field observations
of mortality were made.
V V
RISK CHARACTERIZATION
Risk characterization would involve comparison of
exposure and intakes to benchmarks or criteria;
however, this has not yet been performed because the
study is not complete. Comparisons were made between
control and impacted areas; Monte Carlo analysis
was used.
8-9
-------
although for some chemicals, AWQC are based on protection of humans exposed through
consumption of aquatic life. AWQC were evaluated for each of the major contaminants of concern
to determine that human consumption was not the basis of the criterion.
8.3. CASE STUDY DESCRIPTION
8.3.1. Problem Formulation
Site Description. RMA is a U.S. Army installation consisting of approximately 27 square
miles located north of Denver, Colorado (figure 8-2). In 1942, RMA was used to produce
chemical agents such as blister agent, incendiary munitions, and irritants. From 1945 to 1950,
obsolete World War II ordnance was destroyed at RMA by detonation and burning. Nerve agent
was produced in the 1950s. The Army leased portions of the arsenal to various manufacturers
from 1947 to 1982. Organochlorine pesticides such as DDT, DDE, dieldrin, aldrin, and endrin;
herbicides; adhesives; cutting oils; and other chemicals were produced during this period.
From 1970 until 1984, the primary U.S. Army activity has been demilitarization of
chemical warfare agents. Currently, the sole mission at RMA is to remediate contamination;
present activity at RMA to carry out this mission is limited to land management, wildlife
management, security, technical investigations of contaminant distribution and effects, development
of environmental remediation strategies, and interim response actions.
Disposal practices, spills, and other releases with potential adverse ecological effects have
been historically documented. Some of the most important sources of contamination for biota at
RMA were the wastewater basins in Sections 26, 35, and 36, as well as the four reservoirs known
as the Lower Lakes (ESE, 1989). RMA is a National Priorities List (NPL) site, and certain areas
have been addressed in interim response actions (EBASCO, 1988). RMA is also the site of a U.S.
Fish and Wildlife Bald Eagle Management Area. Visitors are allowed onsite in uncontaminated
areas under controlled conditions to view the eagles in winter at RMA. The Lower Lakes (Upper
and Lower Lake Derby, Lake Ladora, and Lake Mary) have trophy-sized bass and pike and are
fished on a catch-and-release basis only.
Stressors. Contaminants recorded in the RMA environment or inferred from a knowledge
of historical practices at RMA included volatiles, pesticides, herbicides, inorganic salts, Army
chemical agents and their degradation products, and heavy metals. The U.S. Army and U.S. Fish
and Wildlife Service conducted biological monitoring studies during much of the time that RMA
was an active facility. Historical studies (1949-1982) of biota in the basins and Lower Lakes have
indicated death and abnormal behavior for several waterfowl species, other birds, mammals, and
fish (Finley, 1959; USFWS, 1965; USA DPG, 1973; USA EHA, 1976; Linkie and Stiles, 1976;
McEwen, 1981; DeWeese et al., 1982a, b; USFWS, 1983; McEwen, 1983; Thome, 1984).
Organochlorine pesticide levels in tissues were often reported to be elevated in these studies.
Waterfowl losses associated with contaminated sites on RMA ranged from 2,000 to 3,000
individuals annually throughout the 1950s and early 1960s. Waterfowl mortality declined after
1965, at which time the U.S. Army drained Upper and Lower Derby and Ladora Lakes and
removed contaminated sediments.
8-10
-------
BASI,
23
26
lE
•BASIN
NORTH BOQ
NORTH
25 \ PLANTS
BASIN A ,
0\
TOXIC STORAGE '
YARD POND
23
MOTOR
4 POOL (
IRAILROAO X'\
YARD /
1 -—\> HOD AND GUN
\i CLUB POND
STAPL6TON INTERNATIONAL
AIRPORT
N
1/2
1 MILE
Figure 8-2. Map of Rocky Mountain Arsenal (ESE, 1989)
8-11
-------
Data from earlier investigations indicated that organochlorine pesticides might be affecting
avian reproductive success at RMA. The available literature indicates a cause-and-effect
relationship between certain organochlorine pesticides and different aspects of avian reproduction
(St. Omer, 1970; Haegele and Hudson, 1974; Davison et al., 1976; Blus, 1982; Spann et al.,
1986).
Seven chemicals were selected as being of major concern to biota based on consideration of
parameters such as toxicity, areal extent of distribution, and persistence in the environment. These
chemicals were aldrin, dieldrin, endrin, isodrin, dibromochloropropane (DBCP), arsenic, and
mercury. An additional 32 contaminants were also evaluated, but not to the extent of the major
contaminants of concern. Isodrin and DBCP were not analyzed in tissue; isodrin is metabolized to
endrin and DBCP is rapidly metabolized and is not biologically persistent. Arsenic, mercury,
endrin, aldrin, and dieldrin were target analytes in tissue. DDT and DDE also were target analytes
in tissue, although data available at the time did not suggest that they should be included as major
contaminants of concern.
Phase I data on contaminants in abiotic media indicated that the most contaminated areas
were in the areas of chemical manufacturing, storage, and disposal. Sampling of sediments and
water in the lower lakes (Upper and Lower Derby Lakes and Lake Ladora) and soils in the South
Plants, Basins, and Toxic Storage Yard (figure 8-2) indicated elevated levels of chemicals sufficient
to provide a basis for biota field sampling. Maximum concentrations of the major contaminants of
concern in the biosphere (surface water, sediments, and upper 20 ft of soil), based on 1985-1987
sampling, are presented in table 8-1.
Ecological Components. While any of the species that occur at RMA potentially could be
exposed to chemical stressors, this investigation focused on those species that were the most
important or sensitive in the onsite ecosystem, with the assumption that protection of the most
important or sensitive species would ultimately afford protection to all species. By protecting the
most important species in the ecosystem, ecological effects are minimized. Protection of the most
sensitive species indicates that less sensitive species are also protected, thereby minimizing damage
to ecological health. These concepts are fundamental to derivation of the EPA AWQC and are
consistent with human health risk assessment procedures.
Species were classified as important or sensitive if they met the following criteria of
ecological, regulatory, or economic importance: (1) classified as federally threatened or
endangered under the Endangered Species Act of 1973, as amended; (2) listed as endangered or
considered as a Species of Special Concern by the State of Colorado; (3) used as a game species
consumed by humans; or (4) serve as a species of special ecologic value (major prey species,
predator, highly sensitive, bioindicator, etc.). Species selected for tissue monitoring were chosen
after deriving the list of important species with the above criteria. The following additional
selection criteria were used to identify appropriate species for residue monitoring: (1) considered
important components of regional ecosystems, (2) representative of the range of trophic levels in
food chains/webs in regional ecosystems, (3) economically important (e.g., game species), or (4)
representative of higher trophic levels in food chains/webs in regional ecosystems.
8-12
-------
1985-87 Sampling (ESE, 1988)
Analyte
Aldrin
Arsenic
DBCP
Dieldrin
Endrin
Mercury
Isodrin
Soil (to 20 ft)
G*g/g)
40,000
112,00
31,700
7,240
4,650
35,000
3,200
Medium
Sediment
Og/g)
1.38
8.92
1.30 (LT)
1.01
0.74 (LT)
2.30
0.30
Surface Water
Otg/L)
750 (LT)
200,000
1,900 (LT)
470 (LT)
800 (LT)
11
370 (LT)
LT = Less than the Certified Reporting Limit
8-13
-------
The ecological components investigated in this study for contaminant concentrations,
population effects, or reproductive effects included: vegetation (morning glory, sunflowers, and
others), invertebrates (earthworms, grasshoppers, aquatic snails), and vertebrates (mallard duck,
pheasant, mourning dove, raptors [i.e., birds of prey], prairie dog, cottontail rabbit, mule deer,
coyote, badger).
Aquatic species other than snails were investigated by Rosenlund et al. (1986) and other
studies; these data will not be documented in detail here. The aquatic data will be referred to in
this document in a manner consistent with its collection and use hi the Biota Remedial
Investigation. For example, as part of this investigation pathways analysis was performed for the
aquatic food web leading to bald eagles and results of this analysis were used in conjunction with
data collected on contaminant concentrations in the tissue of various components of the aquatic
food web (e.g., aquatic plants, largemouth bass, and northern pike) to establish exposure for bald
eagles.
Eagles and other raptors were studied because they fit the categories of state or federal
endangered or threatened species, because they are predators and thus highly exposed to
bioaccumulative pesticides, and because, as birds, they are sensitive to the effects of
organochlorine pesticides. Raptor populations were surveyed throughout the year. Raptors
observed at RMA include American kestrels; rough-legged, red-tailed, Swainson's, and ferruginous
hawks; golden and bald eagles; northern harriers; and great-horned and burrowing owls. Bald
eagle populations were monitored during the winter months because this species utilizes RMA
primarily for winter roosting and feeding.
The only endangered mammal suspected of occurring at RMA was the black-footed ferret.
Black-footed ferret population surveys were conducted in 5,000 acres of prairie dog towns on
RMA, but no evidence of this endangered species was observed. Therefore, black-footed ferrets
were not considered further.
Ring-necked pheasants and mallard populations at RMA are historically linked with
contaminant-related effects. Both species are important prey species in onsite food webs and also
are game species that are consumed by humans. Mourning doves are also prey species, and as
ground feeders could be expected to be highly exposed to contaminants in soils.
Other mammal species were considered in this study. Mule deer fit the criteria of being a
game species consumed by humans and also provide food for predators or scavengers. In addition,
because they are relatively long-lived, they may be expected to accumulate certain contaminants.
Population surveys were conducted for prey species that might be expected to show adverse
responses to chemical contaminants. Many of these species also form important components of
aquatic and terrestrial food webs on RMA. Prairie dogs and rabbits are important prey items,
abundant in the RMA ecosystem, and live in close contact with soil. Grasshoppers, earthworms,
and aquatic snails are also important prey species, are abundant, and have been indicated to
bioaccumulate certain metals and pesticides.
8-14
-------
Vegetation transects were conducted to determine the dominant species of vegetation in both
disturbed and undisturbed areas. Sunflower, cheatgrass, and kochia were some of the species
observed in these surveys. In support of the Biota Remedial Investigation, quantitative vegetation
studies conducted by Morrison Knudsen Corporation, contractors for Shell, indicated that the level
of onsite physical disturbance due to management activities precluded a quantitative analysis of
potential adverse effects on vegetation due to chemical contamination. However, residue analyses
were conducted on two types of vegetation (morning glory and sunflowers) that often occurred in
potentially contaminated areas.
The number, type, and habitat use by mammalian predators were noted, and limited
quantitative studies of abundance and distribution were conducted by Morrison Knudsen, who
provided these data to the Army for use in the ecological evaluation. Mammalian predators
observed at RMA include red fox, gray fox, long-tailed weasel, coyote, and badger. Coyote and
badger were collected as fortuitous samples when animals were found dead or dying.
Endpoint Selection. Several toxicological and ecological assessment endpoints were
examined (table 8-2). Chemical contamination of ecological components at levels that might be
expected to impair structure or function were considered important because many of the
contaminants of concern to biota were both toxic and bioaccumulative. Adverse effects due to
chronic or acute exposure were also considered as toxicological assessment endpoints. Decreased
population success was considered as an assessment endpoint of ecological importance.
Several measurement endpoints were used in this study to evaluate population success.
Species occurrence, population density, and age classes were evaluated as a measurement endpoint
for determining population success for prairie dogs. Population density and occurrence were
measured for large raptors, American kestrels, grasshoppers, and aquatic snails. Total cover,
species richness, and phenology were endpoints for vegetation. Data from contaminated sites on
RMA were compared with data from onsite and offsite control sites.
Reproductive success was a measurement endpoint used to evaluate population success for
avian species. Reproductive success was measured for three avian species: American kestrel,
mallard duck, and ring-necked pheasant. Chemical analysis of eggs and fledglings, egg
measurement (volume, weight, dimensions, and shell thickness), hatching success, and observation
of brood size and fledgling success were measured in samples for RMA and offsite controls.
These variables were considered important in evaluating reproductive success considering the
typical effects of organochlorine exposure for birds.
Bioaccumulative chemicals have the potential to contaminate tissues at concentrations that
may impair the health of the organism or of associated species. Sunflower leaves and flowers were
analyzed because these plants are heavily utilized by small birds and invertebrates as a food source.
If chemicals are taken up by sunflowers, they may enter the food web and be translocated to
different animals and birds. Invertebrate species were analyzed because they are important prey
items and may provide exposure pathways from soil to mammals and birds. Invertebrates sampled
for chemical residues were grasshoppers and earthworms; the sample mass for aquatic snails
precluded chemical analysis.
8-15
-------
Table 8-2. Summary of Assessment and Measurement Endpoints (ESE, 1988)
Assessment Endpoint
Measurement Endpoint
Ecological Component
Population success
oo
Adverse effects due to
exposure
Adverse effects due to
residue concentrations
Occurrence/Distribution
Density/Relative abundance
Age class
Reproductive effects
Residues
Egg measurements
Hatching success
Brood size
Fledging success
Male-female ratio
Phenology
Total cover, height,
density
Species richness
AChE activity
Behavior
Morphology
Physical condition
Chemical analysis
Carnivores, small mammals (mice, moles) mule deer, cottontail
rabbit, prairie dog, raptors, waterfowl, shorebirds, wading birds,
grasshoppers, aquatic snails, earthworms
Cottontail rabbit, prairie dog, raptors, waterfowl, shorebirds,
wading birds, grasshoppers, aquatic snails, earthworms
Prairie dog
American kestrel, mallard, ring-necked pheasant
American kestrel, mallard, ring-necked pheasant
American kestrel, mallard, ring-necked pheasant
American kestrel, mallard, ring-necked pheasant
American kestrel, mallard, ring-necked pheasant
Ring-necked pheasant
Vegetation
Vegetation
Vegetation
Mallard, ring-necked pheasant, prairie dog, cottontail rabbit,
fortuitous samples
Incidental observations during field activities
Incidental observations during field activities
Incidental observations during field activities
Aquatic macrophytes, sunflower, morning glory, earthworms,
grasshoppers, American kestrel, ring-necked pheasant, mallard,
mule deer, prairie dog, cottontail rabbit, fortuitous samples
(i.e., large raptors, badger, coyote)
-------
Contaminant concentration in tissue was chosen as one of the measurement endpoints
because of historical correlations of mortality with pesticides in tissue and because many of the
contaminants of concern were bioaccumulative. Concentrations of the major contaminants of
concern in tissue were monitored hi many of the ecological components. For each species tissues
for analysis were selected based on the probable fate of the organism in the onsite food web or
because of the organism's status as a game species consumed by humans. Thus, for prairie dogs
and other prey species, the carcass as consumed by predators was analyzed (minus head, fur, feet,
stomach, and intestines). For animals consumed by humans, muscle or muscle and liver were
analyzed. Eggs, fledglings, and some adult birds were analyzed to determine contaminants in
avian species. Opportunistic samples were collected from raptors and mammalian predators when
dead or dying individuals were observed. Liver and brain samples were collected and analyzed
from these species and necropsies were performed to determine the possible role of RMA
contaminants in the death of these individuals.
Acetylcholinesterase (AChE) activity was measured because some of the chemicals
historically produced at RMA were cholinesterase inhibitors. Contaminant analysis was performed
on tissues from these animals as an attempt to correlate contaminant concentrations with observed
effects.
Qualitative measurement endpoints to evaluate adverse effects due to exposure were
determined from the organisms collected for contaminant analysis. These endpoints included
behavior (e.g., impaired movement), gross morphology (internal and external), and physical
condition (e.g., presence of normal body fat, emaciated appearance).
Comments on Problem Formulation
Strengths of the case study include:
•The history of the site is documented. Chemicals of major concern are identified to
provide a focus for the assessment. An effort is made to identify a causal relationship
between contaminants and potential adverse responses in the environment. Criteria
are presented for the selection of the major chemicals of concern.
•Components of the entire ecosystem are considered including individuals,
populations, communities, and food webs. Criteria are presented for the selection of
particular ecological components upon which to focus the assessment.
*A number of measurement endpoints are utilized that consider the connection between
the interactions of the ecological components and the chemical properties of the
stressors. A diversity of endpoints is utilized at a number of ecological levels,
including tissue concentrations, biomarkers, and population surveys. This wide
diversity of endpoints provides a holistic examination of the ecosystem, lending greater
confidence in risk estimates.
8-17
-------
Comments on Problem Formulation (continued)
Limitations include:
•Data on chemical concentrations in environmental media were only available for
Phase I sampling. The potential for synergistic effects to occur among compounds is
not addressed. Because the list of contaminants studied is narrow, the question is
raised that some other contaminants may have been missed.
•Details on the life history, habitat, feeding behavior, etc., of the selected ecological
components are not presented. Vegetation could not be better assessed due to physical
disturbances on the site. '
8.3.2. Analysis: Characterization of Ecological Effects
No single reference (control) area was located that was comparable to RMA with respect to
vegetation types, areal extent, and land use (e.g., absence of hunting and grazing pressure). As a
result, different, smaller reference areas were selected to meet specific sampling needs. For
example, the Wellington Wildlife Refuge, more than 50 miles north of" RMA, was selected for
collecting control samples of mallard, mallard eggs, cottontail rabbits, mule deer, and prairie dogs
for chemical analyses, but land adjacent to Barr Lake, only 5 miles north of RMA, was used for
collecting control samples of earthworms and conducting population studies of earthworms in order
to sample within the same soil type. Other reference areas were selected for similar reasons of
appropriateness or species availability. Site characteristics used in selecting reference areas
included: (1) distance from RMA (distances varied, depending on the mobility of the taxa
involved, to ensure that sample organisms from reference areas would not have been potentially
exposed to RMA contamination); (2) similarity in general vegetation type(s); (3) similar land uses
(e.g., protection from hunting and/or grazing); (4) similar topography (e.g., plains, not mountains
or foothill terrain); and (5) similar soil type(s). Different combinations of these features were used
to select reference areas, depending on the objectives of the sampling to be conducted.
Ecological effects were characterized by the following methods: (1) pathways analysis for
criteria development; (2) potentially lethal tissue concentrations; (3) depression in brain AChfe
activity; (4) reproductive success for American kestrel, mallard, and ring-necked pheasant; and (5)
population studies. Population studies (i.e., species occurrence, population density, and age-class
structure) were considered qualitative or semiquantitative indicators of stress response. Many of
these characteristics were not considered quantitative indicators of chemical stress because they can
be influenced by so many factors other than contaminant concentrations (e.g., habitat, prey
availability, noncontaminant-related human disturbance).
Pathways Analysis for Criteria Development. Characterization of ecological effects for the
major contaminants of concern was based on comparison of observed concentrations in abiotic
media and biota with estimated site-specific soil, water, and sediment criteria and the chemical-
8-18
-------
specific AWQC. Site-specific soil, water, and sediment criteria were estimated by examining
potential exposure pathways, a method termed "Pathways Analysis" (ESE, 1989; Fordham and
Reagan, 1991). The criteria can be applied in risk characterization by comparing acceptable
concentrations developed from the model with observed concentrations in the tissues of species
within selected food webs.
The pathways model was developed as a new tool to establish quantitative relationships
between contaminant concentrations in abiotic source media (e.g., soil, sediment) and in biota.
The model was modified from a single food chain model proposed by Thomann (1981) to include
multiple food chains in the food web of key species. The model incorporates estimates of exposure
of various organisms to specific chemicals in the environment through the mechanisms of
bioconcentration (concentration from direct exposure to water in an aquatic environment),
bioaccumulation (concentration from diet plus bioconcentration), and biomagnification (the increase
in concentration as chemicals move along food chains to higher trophic levels). Direct exposure
pathways are also considered. Contaminant concentrations in soil or sediment are linked with
water by the soil-water partition coefficient normalized for organic carbon.
The pathways approach looked at potential lethal and nonlethal forms of biological injury
associated with residue concentrations of the contaminants of concern, as well as direct exposure.
Some of these nonlethal effects were known at the level of the individual organism (e.g., abnormal
behavior, reduced fledgling success in waterfowl) and evaluated at the population level as described
in subsequent sections. Thus, the pathways approach was able to establish and quantify the
exposure pathway, and field sampling for adverse effects was conducted to verify the effect. The
model approach for quantifying exposure pathways was applied to the terrestrial and aquatic
ecosystems on RMA and is adaptable to any ecosystem where basic information is available both
on the species present and the trophic relationships among organisms.
Each potential exposure pathway was evaluated (i.e., ingestion of surface water, ingestion
of prey items) and a criterion for contaminant concentration was developed protective of that
exposure pathway. For a given chemical, all exposure criteria were compared, and the lowest
criterion was selected as protective of sensitive species and sensitive pathways.
Criteria development served several purposes. It served to establish a relationship between
observed tissue concentrations in biota and concentrations in abiotic media that were below
detection and therefore unmeasurable. The criteria also served to indicate a level of contamination
hi the environment that would probably not cause adverse effects. This information was used to
infer where adverse effects might be occurring without directly measuring them. The usefulness of
this approach was confirmed by those instances where observed adverse effects (i.e., avian
mortality) were correlated with tissue concentrations in affected species. Initial acceptable site-
specific abiotic media concentrations developed by this method are presented in table 8-3.
Several exposure pathways were evaluated to obtain criteria; these were as follows:
1. Exposure as a result of surface water ingestion was considered an important exposure
pathway. Toxiciry data obtained from the available literature for terrestrial organisms
were examined for LOAELs and NOELs for an oral exposure route. The most
8-19
-------
Table 8-3. Acceptable Concentrations of Major Contaminants of Concern in Abiotic
Media (ESE, 1988)
Water Sediment Soil
(ppb) (ppm) (ppm)
Aldrin/dieldrin 0.034 0.0055 0.10
Arsenic 100 15 52
DBCP 60 0.086 6.10
Endrin/isodrin 0.032 0.0019 9.2
Mercury 0.004 0.004 1.1
8-20
-------
sensitive (lowest) toxicity value (in mg/kg bw/day) was converted, if necessary, to a
water concentration (mg/L) by dividing by daily water intake (L/kg bw/day) for the
species for which the toxicity value was available. The water concentration was then
divided by uncertainty factors based on data quality (appendix A). Sediment criteria
were developed from the water criteria by multiplying the chemical-specific soil-water
partition coefficient (Koc) and the fraction of organic carbon (foc) by estimated water
criteria (appendix A). This resulted in abiotic criteria specific to toxicity as a result of
ingestion of contaminated media.
2. Exposure to contaminants as a result of ingestion of diet was considered. A
bioaccumulation model (Thomann, 1981) was adapted to reflect an entire aquatic "sink"
food web, as opposed to single food chains. A sink food web is a subset of a
community food web that includes all organisms consumed by the "sink" or top-level
species (Cohen, 1978) (figure 8-3). Total bioaccumulation in the food web (total BAF)
was estimated by modifying Thomann's model by applying a percent diet factor to each
compartment of the food web (Fordham and Reagan, 1991). A maximum acceptable
tissue concentration (MATC) for each contaminant was derived from the available
literature and divided by total BAF to obtain a water concentration (the sink food web
initiated with water; therefore, the BAF for the sink species reflects the amount of
contaminant in water transferred through a food web) (Fordham and Reagan, 1991).
Sediment criteria were obtained from water criteria by applying the Koc and foc.
Uncertainty factors were not applied to the resulting criteria because very conservative
assumptions were made during the modeling process.
3. Toxicity to aquatic life by direct contact was determined by comparison with AWQC.
When AWQC were unavailable or not applicable because they were based on human
consumption of aquatic life as an endpoint, LOAELs and NOELs for aquatic life were
obtained from the available literature. Uncertainty factors were applied based on the
data quality (appendix A).
4. Soil criteria were developed by calculating bioaccumulation factors in a terrestrial food
web. For example, a sink food web consists of multiple food chains. Bioaccumulation
factors for species in each single food chain (i.e., plants, mammals, bald eagle) were
identified from the literature. The bioaccumulation factors were multiplied to obtain a
BAF for each individual food chain. A total BAF was calculated for the sink species
by adjusting each pathway-specific BAF by the relative proportion in the sink species
diet and summing the adjusted factors. Total BMP from all food chains was divided
into an MATC for the sink species to obtain soil criteria. In addition, LOAELs and
NOELs were considered for terrestrial soil fauna based on exposure by direct contact
and soil ingestion.
Criteria for each of the major contaminants of concern developed for each exposure
pathway above were compared to each other (figure 8-4). The lowest criterion for water,
sediment, and soil was selected to represent the acceptable level in the environment that was
unlikely to pose a threat to wildlife populations.
8-21
-------
s
to
SMALL
MAMMALS
PLANTS
SOIL
(Source: ESE, 1989)
1
AQUATIC
PLANTS
BALD EAGLE
_^ m rp_
WATERFOWL
(Mallard)
I
BASS & PIKE
SMALL FISH
A
CHIRONOMIDS
PLANKTON
I
SEDIMENTS
Figure 8-3. Bald eagle food web for contaminants in soil, sediment, and surface water source
-------
MAJOR CONTAMINANTS
OF CONCERN
TERRESTRIAL
PATHWAYS ANALYSIS
SOIL CRITERIA
s
TOXICITY TO
AQUATIC ORGANISMS
TOXICITY TO
TERRESTRIAL BIOTA
(water ingestion)
AQUATIC PATHWAYS
ANALYSIS
LOWEST
SITE-SPECIFIC
WATER CRITERIA
IF APPROPRIATE
IF BIOACCUMULAJES
TOXICITY ASSESSMENT
OF OTHER CONTAMINANTS
OF CONCERN
(Source: ESE, 1989)
CALCULATE FINAL
RESIDUE VALUE
EVALUATE EPA AMBIENT
WATER QUALITY CRITERIA
AND AQUATIC TOXICITY
TOXICITY TO
TERRESTRIAL BIOTA
(water ingeslion)
LOWEST
WATER CRITERIA
TOXICITY TO BIOTA
EXPOSED TO SOIL
-I SOIL CRITERIA I
SEDIMENT
CRITERIA
Figure 8-4. Criteria diagram
-------
Lethal Toxicity. While field sampling focused on selected species from predetermined
areas, animals found dying or recently dead were collected as "samples of chance." Particular
attention was paid to raptors because of their position at the top of site food webs and the known
avian toxicity of some RMA contaminants (e.g., organochlorine pesticides). Necropsies were
performed on carcasses to determine the possible cause of death and to document the condition
(e.g., emaciated) of each animal. For specimens found dying, the condition and behavior of each
were recorded as additional clues to the animal's death. Samples of liver and brain tissue were
collected and analyzed for contaminants.
Lethal dieldrin levels in brain tissue of birds have been reported in the range of 3 to 4
mg/kg (Robinson et al., 1967; Belisle et al., 1972). Three raptors species (ferruginous hawk, red-
tailed hawk, and great-horned owl) were found dead on RMA with levels of dieldrin in the brain
within this range. One great-horned owl, found in an emaciated condition, had dieldrin
concentrations of 9.32 mg/kg in liver tissue and 27.7 mg/kg in brain tissue.
Brain Acetylcholinesterase. Brain acetylcholinesterase (AChE) levels were determined from
various species collected for contaminant analysis. Organophosphates and some metal ions are
known cholinesterase inhibitors. AChE activity can be difficult to interpret as a bioindicator
because this enzyme is also influenced by population variability, diurnal cycles, metabolic function,
and postmortem changes. Levels of AChE activity were measured in brain tissue from
opportunistically collected birds and mammals (if the specimen was considered fresh), rabbits,
prairie dogs, mallards, ring-necked pheasants, and American kestrels. AChE activities in animals
from contaminated areas were considered to be significantly depressed if there was 20 percent or
greater depression hi activity as compared to control animals (Robinson et al., 1988).
Avian species' from RMA (mallard, ring-necked pheasant) had brain AChE levels that were
similar to controls (ESE, 1989). One mourning dove, two golden eagles, and three red-tailed
hawks found dead on RMA had activities slightly elevated with respect to normal values reported
in Hill (1988). It is unknown whether postmortem changes may have influenced enzyme activity
because control specimens dead at the time of collection were not available.
Prairie dogs from two areas on RMA had AChE activity significantly lower than those on
onsite and offsite controls (p<0.01), and even onsite controls were slightly depressed with respect
to levels in offsite controls (p<0.05). Prairie dogs from only one site, the Toxic Storage Yard
(figure 8-2), had significantly depressed levels of AChE. The pattern of AChE depression suggests
contaminant-related effects; however, the effect was not correlated with soil concentrations of
known organic AChE inhibitors. Several inorganics were elevated in soils from the areas where
depressed AChE activity was observed. Arsenic compounds (arsenite ion and to a lesser extent
arsenate) are linked with cholinesterase depression in fish (Olson and Christensen, 1980). Other
metal ions have also been linked to cholinesterase inhibition (Tomlinson et al., 1981). The metal
ions involved appeared to be naturally occurring soil constituents and not RMA contaminants. No
differences between contaminated and control areas were observed in brain AChE in cottontail
rabbits.
8-24
-------
Cholinesterase inhibition is an appropriate measurement endpoint for contaminants that are
known or suspected AChE inhibitors. However, current data on chemical contaminants in abiotic
media at RMA indicated that the occurrence of potential AChE inhibitors on RMA was limited.
Avian Reproductive Success. American kestrel reproductive success has been studied at
RMA for several years (ESE, 1989). The 1986 survey indicated a much higher nesting success
than in USFWS studies in 1982 and 1983 (38 percent nests fledged in 1982, 50 percent in 1983,
and 71 percent in 1986); 1986 data for RMA were not significantly different from controls (ESE,
1989). Reproductive parameters considered were percentage nests hatched and fledged, and mean
number of young hatched per nest.
Collected eggs of American kestrels, ring-necked pheasants, and mallards were measured
for weight, volume, dimensions, and shell thickness. There were few differences between eggs
from RMA and control sites. Kestrel eggs from RMA were slightly larger than controls, and
pheasant eggs from RMA averaged smaller and lighter than controls. Kestrel mean shell thickness
did not differ from controls (ESE, 1989).
Average pheasant brood counts were lower for RMA than for offsite controls (average
young seen per transect: 1.1 on RMA, 3.0 offsite) (ESE, 1989). Total hens and clutch sizes were
also smaller on RMA.
Population Studies. There were no detectable differences for vegetation between RMA and
the offsite controls that could be attributable to contaminant-related effects. Extensive physical
disturbance (e.g., mowing, herbicidal weed control, disking, reseeding, and burning) continued as
part of normal maintenance activities. These activities and soil compaction occurred in both
uncontaminated and contaminated areas of RMA, making it difficult to isolate effects that could be
attributed to RMA contaminants.
The aquatic snail data indicated a high degree of variability among sites and between years.
The covariates of vegetation (substrate) weight, temperature, and pH indicated that these factors
influenced the results. The grasshopper data also did not indicate any differences that were
attributable to contaminant effects. The contaminated sites had decreased species richness, but also
had decreased plant diversity. Population comparisons for earthworms indicated significant
differences between onsite and offsite controls.
The numbers of diving ducks and coots were higher at RMA than at control sites, but the
number of dabbling ducks and geese was lower. The most striking difference was the complete
lack of mallard broods at RMA. Only two mallard nests were located on RMA.
Significantly higher prairie dog adult-young ratios were observed offsite than on RMA, and
density appeared lower onsite than offsite. However, colonies from contaminated areas did not
differ significantly from colonies from uncontaminated areas onsite. Cyclic population fluctuations,
past management practices, predation, and habitat suitability are other factors that could influence
prairie dog populations. At RMA, many of the prairie dog colonies have a high percentage of
cheatgrass as opposed to a diverse mix of vegetative species that provide better habitat and food
sources.
8-25
-------
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
^Pathways Analysis is a useful tool for quantifying the exposure and potential
ecological effects.
^Figure 8-4 provides a helpful flowchart on the relationship between stressors,
ecological components, endpoints, and acceptable criteria for chemical concentrations
in the environment.
limitations include:
^Results are often expressed in general terms when they should be expressed
quantitatively and in tables wherever possible.
General comment/cautionary note:
*The criteria developed in the course of this study are not recommended for use at
other sites because they were developed to be site specific, and more consideration was
given to studies with data regarding species similar to those on KMA. The chemical-
and media-specific criteria were developed for consideration in the absence of
available EPA criteria so that overall risk to aquatic or terrestrial wildlife populations
could be evaluated. At this time, the criteria have not been validated to determine if
these levels are acceptable for protection of ecological health. However, the pathways
modeling approach has proved to be a valuable tool in the assessment of biological
risk and for the development of environmentally based remediation criteria.
•/« addition to considering trophic level when selecting species for tissue analysis, the
exposure variable of species movement in the environment should be considered. In
order to relate abiotic to biotic data, the species collected for chemical analysis should
preferably have a limited home range or limited feeding area. Otherwise, the exposure
assessment becomes very uncertain. Migrant species can be included for contaminant
analysis (i.e., American kestrel), but consideration should be given either to using life
stages that are immobile or to collecting adults after they have been at the site for a
sufficient length of time, where a condition approaching equilibrium may be achieved.
Aquatic species may not be affected by spatial variables to the extent that terrestrial
species are, due to the greater amount of spatial variation observed in a terrestrial
environment as opposed to an aquatic one.
8-26
-------
8.3.3. Analysis: Characterization of Exposure
A full characterization of chemical exposure was not completed during the Biota Remedial
Investigation because all abiotic data collected during the remedial investigations were not
available. However, toxicity assessments were compiled, ecological component populations were
characterized, and exposure pathways were evaluated.
Tissue Concentrations. Biota were observed to contain mercury, arsenic, dieldrin, endrin,
and DDT/DDE (ESE, 1989). Dieldrin was the contaminant most frequently detected in biological
tissue. Tissue concentrations were obtained for various tissues for different organisms on RMA
(appendix B). Samples were collected from several known contaminated areas on RMA, as well as
from areas believed to be relatively undisturbed and from offsite controls. Rosenlund et al. (1986)
and data supplied by Morrison Knudsen were relied upon to determine contaminant effects in fish
and other aquatic life.
Dieldrin was the primary contaminant in eggs from mallards (4.89 mg/kg), ring-necked
pheasants, and American kestrels (mean of 0.504 mg/kg) (ESE, 1989) (appendix B). Dieldrin was
also the primary contaminant in avian carcasses found on RMA (appendix C). Most hawks and
owls found dead on RMA and analyzed for contaminant concentrations had residues of dieldrin in
the brain and liver. Brain concentrations of dieldrin from raptors collected dead in an emaciated
condition ranged from 0.678 to 9.98 mg/kg (appendix C) (ESE, 1989). Lowest reported lethal
brain concentrations of dieldrin in raptors are approximately 3 to 4 mg/kg (Stickel et al., 1969;
Belisle et al., 1972; Stickel, 1973); some raptors retrieved dead from RMA with no other known
cause of death had brain concentrations within this range.
Exposure Pathways. Biota can be exposed to contaminants in many different ways: (1)
direct contact with contaminated surface water or sediment by aquatic life; (2) direct contact with
contaminated soil or surface ponding by soil fauna; (3) ingestion of contaminated surface water,
soil, or food items by terrestrial organisms; (4) inhalation of contaminated air or fugitive dusts; and
(5) dermal contact. For the major contaminants of concern, toxicological literature (Moriarity,
1985) and known distribution and concentration of contaminants in abiotic media indicated that
direct contact and ingestion were the exposure routes expected to be the most important for wildlife
populations. Observation of the types of species apparently most affected by contamination also
indicated the ingestion and direct contact pathways as being the most significant exposure
pathways. High-trophic-level birds and mammals were found dead or with high contaminant levels
in tissues. Burrowing mammals and soil fauna also appeared to have high tissue levels.
Waterfowl and aquatic life, both components associated with ingestion or contact with water or
sediments, appeared affected by contaminants.
Inhalation was determined to be an insignificant exposure pathway based on data provided
hi the Air Remedial Investigation Report (ESE, 1988). In general, air concentrations were low.
Dermal contact by higher level species such as birds or mammals is difficult to evaluate
because the toxicological data and exposure factors for nonhuman biota are lacking. Contaminant
transfer during burrowing, grooming, and other activities across an intact dermal membrane is
highly uncertain. The presence of hair or feathers may impede dermal uptake by keeping
8-27
-------
contaminants adsorbed or absorbed to soil away from skin. Chemical properties also influence
dermal uptake.
Soil ingestion rate was estimated for small mammals from the available literature. The
ingestion rate was compared to the estimated criteria to determine if the criteria would be
protective of exposure of small mammals by soil ingestion.
Magnitude. Frequency, and Duration of Exposure. Some areas of RMA, primarily the
basin areas (Section 35, 36, and 28) where wastes collected and the Lower Lakes, have been
linked to extensive wildlife mortality hi the past. The magnitude and frequency were probably
more extensive during the time when RMA was an active industrial facility. After 1965, the U.S.
Army drained Upper and Lower Derby and Ladora Lakes and removed contaminated sediments.
The lakes were then refilled. Recent analyses by Rosenlund et al. (1986) indicated that sediments
in the Lower Lakes were still contaminated, but that water concentrations of pesticides were below
detection.
Frequency and magnitude of exposure are expected to be dependent on the species in
question. For example, American kestrels do not appear to be affected as severely as in the past,
but mallards and possibly pheasants continue to exhibit contaminant-related effects. This may be
related to the ingestion rate of soil or sediment during pheasant or mallard feeding.
American kestrels did contain pesticides, as did many other predators collected on RMA.
Predators are at risk of contaminant effects when the contaminants of concern are bioaccumulative,
such as the organochlorine pesticides and mercury. For contaminants that are not bioaccumulative
(e.g., arsenic and DBCP), organisms that come into direct contact with the contaminant (e.g.,
terrestrial plants, invertebrates, or aquatic life) are most likely to be at risk.
Some herbivores were relatively free of residues. Deer and rabbits had lower residues than
prairie dogs. This may be related to the frequency and magnitude of exposure. Prairie dogs live
in close contact with soil, spending a large portion of their daily time below ground or involved
with burrowing activities. Grooming would be a likely source of frequent exposure due to
ingestion of soil particles. Prairie dogs have vertical as well as horizontal movement and may be
exposed to chemicals deep in the soil.
Consideration of a species' feeding habits and activity patterns is important in defining
exposure. Species that are sedentary in their habits will have a longer duration of exposure than
free-ranging species. Animals with a small home range are thus more likely to be affected by
contaminant exposure than ones that may move in and out of contaminated areas for feeding.
Prairie dogs, a sedentary species, thus may provide a better index of site-specific exposure than
mule deer or rabbits, which may range over more territory. This is illustrated by the contaminant
concentrations in prairie dogs compared with other mammals with approximately the same food
habits. However, other critical variables may interact besides population movements. Prairie dogs
are likely to have a higher soil ingestion intake due to the extensive burrowing activities performed
by these animals. Avian species are mobile in the environment, and thus their exposure is a result
of a wide range of prey and contaminant levels in prey. For this reason, it is important to consider
population variables such as home range and preferred feeding areas in addition to feeding habits
8-28
-------
and activity patterns when determining risk to wildlife populations from hazardous chemicals in the
environment.
Fate and Transport. Most of the major contaminants of concern to biota were
environmentally persistent chemicals with bioaccumulative properties. Aldrin, dieldrin, endrin,
isodrin, and mercury all accumulate in tissues. Isodrin is metabolized to endrin in tissue, and
aldrin is metabolized to dieldrin. Some aldrin was detected in grasshoppers collected from Section
26 (appendix B), but this might have been a result of dermal contamination from surface soils, and
not actual tissue contamination. Samples were not washed prior to chemical analysis because
analysis was intended to represent a consumer organism's exposure level.
Dieldrin leaches slowly from soils; the bulk of depletion from soil is a function of
volatilization rather than uptake, degradation, or runoff (Beyer and Gish, 1980). Like other
organochlorine pesticides, it is relatively insoluble in water. In terrestrial species, invertebrates
tended to have higher dieldrin concentrations than plants. Avian species had dieldrin levels slightly
lower than those in invertebrates.
A major biological fate of DDT is metabolism to DDE; DDE but not DDT occurred in
waterfowl, pheasants, and raptors (appendix B).
Endrin is persistent in the environment, but is less stable than either aldrin or dieldrin. It is
subject to microbial degradation, particularly under anaerobic conditions such as occur in saturated
soils. Bioaccumulation of endrin is less than for dieldrin. Endrin was not detected as often as
dieldrin (appendix B).
Inorganic arsenic forms relatively insoluble complexes in soil, binding to hydrous oxides on
clays or cations in the soil solution (Woolson, 1983). It occurs in several chemical forms that
affect its bioavailability and transport. In general, the more water-soluble forms occur in areas that
receive little rainfall (Woolson et al., 1971) and are likely to enter surface waters as runoff. In
sandy soils, such as occur at some locations on RMA, arsenic can leach into ground water,
although this generally occurs in shallow surface layers. In aerobic soils, arsenic occurs
predominantly in the arsenate form (Woolson, 1983). Arsenic is not highly bioaccumulative and
tends to be metabolized quickly and excreted. Arsenic was detected in plants and invertebrates and
may have been a result of surface contamination, as samples were not washed prior to analysis.
Arsenic was detected infrequently in vertebrates. Concentrations and frequency of detection in
pheasants from RMA did not differ significantly from controls.
Mercury in sediments tends to be in the inorganic form (Snarski and Olson, 1982).
Because microorganisms are capable of converting inorganic and organic mercury compounds into
highly toxic methylmercury and dimethylmercury, any form of mercury in the environment is
hazardous. The synthesis of methylmercury by bacteria in sediments and water is the major source
of methylmercury in aquatic environments (Boudou and Ribeyre, 1983). Mercury, although
mobile and bioaccumulative in aquatic ecosystems, occurred infrequently in terrestrial species. The
occurrences in terrestrial species were in higher level avian predators, and in invertebrates and
mammals in direct contact with soils. Mercury did occur in species such as mallard and other
waterfowl that feed within the aquatic food web.
8-29
-------
DBCP was not a target for analysis because it is metabolized rapidly and is not expected to
occur in tissue except within hours or several days following exposure. It is slightly soluble in
water and is mobile in the environment, migrating from soil to ground water. It is highly
persistent and decomposes slowly by hydrolysis or microbial action (U.S. EPA, 1987). Movement
of DBCP in soils is greatest in soils with a coarse texture and low organic content (Bigger et al.,
1984). DBCP is metabolized to metallic bromides in plants.
Comments on Analysis: Characterization of Exposure
Strengths of the case study include:
+The case study links the ecology of ecological components to their potential for
exposure to chemicals, and considers the magnitude, frequency, and duration of
exposure.
Limitations include:
+The observations and conclusions drawn would be more strongly substantiated by
quantification of exposure.
8.3.4. Risk Characterization
The risk characterization evaluated the data from the ecological components and compared
estimated exposure concentrations with critical values from the toxicity assessments. The risk
characterization could not be completed because the abiotic data base was incomplete. However,
certain aspects of risk characterization that could be evaluated were the following:
• observed populational effects onsite compared to controls;
• tissue concentrations that exceeded critical values from the literature; and
• development of site-specific criteria to be evaluated and/or applied in later phases of
the ongoing investigation.
These aspects were utilized in deriving a relationship between exposure and adverse effects.
What could not be completed was a location-specific evaluation of ecological risk or definitive
remediation goals. In order to have completed the risk assessment, abiotic media data would have
needed to have been finalized.
Relationship Between Exposure and Adverse Effects. As part of the ecological effects
characterization, assessments were made to determine the relationships between exposure
concentrations and adverse effects. Toxicological effects as a result of exposure, lethal and
8-30
-------
sublethal concentrations from laboratory studies, and data from field studies were compiled.
Populational effects were initially described in the ecological characterization. Populational effects
resulting from exposure were determined by comparison with controls that were unaffected by
chemical hazards found at RMA.
There did not appear to be any exposure-related impacts on plant communities. There were
statistically fewer aquatic snails in the RMA lakes than in offsite lakes (p> 0.001), although
conclusions regarding a relationship with exposure could not be drawn because statistical analysis
also indicated that vegetation, pH, and temperature were covariates that may have confounded
results. There were no significant differences for grasshopper populations, but earthworm
populations were significantly different. However, when residues were analyzed statistically, there
were significant differences only for arsenic, which was higher in controls. Thus, for all lower
level (plant and invertebrate) ecological data, weight of evidence fails to support contaminant-
related impacts. However, mallard ducks had apparently reduced reproductive success. The
reproductive success of kestrels was improved over data from previous studies. Reproductive
success was compared to analytical data for mallards and kestrels collected at RMA.
Organochlorine concentrations in a mallard egg exceeded critical levels from the toxicological
literature of > 1 mg/kg (Blus, 1982) that would be indicative of poisoning. Conversely, mean
concentrations of organochlorines in American kestrel eggs (0.504 mg/kg) were less than the
critical level in eggs (ESE, 1989).
Lethal and sublethal concentrations of the contaminants of concern in tissue were of
particular importance because tissue concentrations were a major endpoint for this study. The
relationship between exposure and adverse effects was documented for the basin areas and the
Lower Lakes by the widespread occurrence of contaminant levels in the tissues of specimens from
RMA. Lethal levels in brain tissue were approached or attained in raptors collected from RMA.
Often, birds were in an emaciated condition, indicative of organochlorine pesticide poisoning.
Arsenic levels in leaves of sunflowers were in the range of phytotoxic concentrations observed in
other plants. Obvious signs of phytotoxicity were not observed in these plants, however. It is
possible that surface dust may have caused the observed concentrations. Levels of the
organochlorine pesticides in grasshoppers were high enough to be a potential hazard to avian
consumers based on comparison with critical levels from the toxicity assessment. Arsenic in
grasshoppers was also high enough to present a potential hazard to insectivores preying on
grasshoppers.
Criteria were estimated for each of the potential exposure pathways (table 8-3). These
pathways included dietary exposure (bioaccumulation), direct contact by soil and aquatic
organisms, and surface water ingestion. Observation of the types of species apparently most
affected by contamination indicated the ingestion and direct-contact pathways as being the most
significant exposure pathways. The Pathways Analysis method of criteria development is under
further investigation at RMA to determine levels in the RMA environment that would not pose a
threat to ecological health. A calibration/validation process is being utilized to reduce model
uncertainty. The criteria provide a means by which to estimate risk; they are not risk estimates.
Comparison of the criteria with exposure data would result in a hazard quotient.
8-31
-------
Depending on how the pathways model is applied, the criteria are protective of populations,
or, in the case of endangered species, of individuals. For instance, if average data for populations
(i.e., average BCFs for different species, average loss rates, average LOAELs) are applied, then
the model will be protective of populations. The model can deliver more conservative results by
altering the parameters to reflect an individual response or by applying uncertainty factors.
Evidence Linking Exposure and Adverse Effects. Comparison to controls indicated
significantly higher tissue concentrations for some species collected from RMA (table 8-4). The
analytes sampled were all site-related contaminants, and the significance levels were quite low for
some analytes and species. Statistics were not possible for samples of chance, or for species where
too few animals were collected. Nonparametric statistics were used due to heteroskedasticity in the
data and small sample sizes (ESE, 1989).
Other evidence linking exposure with adverse effects is presented in appendix C. These
raptors were collected dead, and chemical analysis and autopsies were performed. In some cases,
the emaciated condition of the bird and the levels of organochlorine pesticides are indicative of
organochlorine pesticide poisoning.
Sources of Uncertainty. There are many sources of uncertainty inherent in the analysis of
the nature and extent of contamination. There is analytical uncertainty, although Program Manager
for Rocky Mountain Arsenal protocols specify that the Certified Reporting Limit is the 95 percent
confidence interval for the method. In the implementation of a field sampling design for a large
area, additional uncertainties occur. Due to the size of the study site, there is spatial uncertainty.
The extent to which the samples collected actually represent the population of potential contaminant
detections is unknown. For example, prairie dogs collected from onsite control areas that were
supposedly undisturbed contained dieldrin in the liver. Subsequent sampling confirmed the
presence of dieldrin in surficial soils in the area. Uncertainty is also inherent in estimates of
biological fate of contaminants. Bioavailability of contaminants in soil or surface water may be
highly variable, as may bioconcentrations, bioaccumulations, or depuration.
Animals that are mobile in the environment can alter the actual levels of contaminants to
which they are exposed, such that some populations may be highly exposed while others have
minimal exposure. Therefore, determining representative and worst-case exposure concentrations
over the area of the site is important to determining overall risk to wildlife populations.
Uncertainty can be reduced by emphasizing in contaminant analysis those animals that are in close
contact with their environment and tend to have a limited home range. Sampling avian juveniles
and eggs seemed to provide a better distinction between control and contaminated sites than did the
sampling of adults, although with sampling juveniles of any species, growth dilution of tissue
contaminant load becomes more questionable than when sampling adults.
There is also uncertainty in the measures of ecological effects. The AChE depression
observed in prairie dogs could not be explained from data available during the Biota Remedial
Investigation, although the effect was statistically significant compared with control locations.
Other stressor-response relationships were more clear, such as the presence of dieldrin in
organisms from areas of historical dieldrin contamination. Species-to-species extrapolation (such as
utilizing toxicity data from one species to address risk to others) is as uncertain as a factor of 5 or
8-32
-------
Table 8-4. Summary of Species Where Contaminants of Concern Differed
Significantly (p^O.OS) or Approached Significance Between Controls
and Contaminated Areas Onsite (ESE, 1989)
Species
Analyte
Level of Significance
Earthworms
Grasshoppers
Kestrels
egg
juvenile
Mallards
egg
juvenile
adult
Pheasant
egg
juvenile
Cottontail
Prairie dog
arsenic
arsenic
dieldrin
aldrin
dieldrin
dieldrin
dieldrin
dieldrin
mercury
dieldrin
dieldrin
dieldrin
dieldrin
arsenic
dieldrin
0.05 ;>p > 0.01
0.10 ^ p > 0.05
0.01 ^ p > 0.001
0.10 ^ p > 0.05
0.01 ^ p > 0.001
0.05 ^p > 0.01
p < 0.001
0.05 > p > 0.01
0.05 > p > 0.01
0.10 ;> p > 0.05
p < 0.001
0.10 ^ p > 0.05
0.05 > p > 0.01
0.10 > p > 0.05
p < 0.001
8-33
-------
more. Not only does toxicity differ between species (i.e., LCso ^^ *-D50 data), but in field
situations, the animals' behavior and life history can cause large differences in exposure rates.
Uncertainty in the bioaccumulation exposure pathway, which is one pathway in the
Pathways Analysis method used to estimate site-specific criteria, was quantified with Monte Carlo
analysis. The parameters in the bioaccumulation model were either considered to be fixed, or they
were assigned a distribution, as the available data indicated. Monte Carlo analysis then performs
numerous iterations by randomly selecting points from the parameter distributions. The ultimate
result of the analysis is based on the variability in the parameter distributions and provides a
quantitative measure of uncertainty surrounding the modeling results. The amount of uncertainty in
the model to estimate criteria protective of contaminant effects resulting from bioaccumulation
varied with the contaminant and the availability and quality of the data for each contaminant. For
example, data were available for parameters of depuration bioconcentration (BCF) for the aquatic
food web model, whereas assimilation efficiency data were frequently lacking for wildlife species.
Bioaccumulation (BAF) and dietary proportion were also available from the literature.
Uncertainty in evaluation of the direct exposure pathways was quantified in a manner
similar to assessment of uncertainty in human health risk assessment (appendix A). Uncertainty
factors were applied to literature data in order to derive criteria. If data for acute lethality only
were available, the uncertainty to derive an acceptable concentration was higher. These
uncertainty factors may ultimately produce criteria that are overly conservative; criteria resulting
from this process should not be considered absolute.
Additional sources of uncertainty include the possibility that sensitive species have been
replaced by tolerant organisms during the decades of contamination at RMA. It is also possible
that additional effects resulting from the combined actions of other contaminants have occurred,
although there is currently no basis for evaluating this possibility. The evaluation of a suite of
measurement endpoints at the individual, population, and, ecosystem levels was an important feature
of this assessment that helped reduce its associated uncertainties.
Comments on Risk Characterization
Strengths of the case study include:
• The case study established protective criteria for the ecosystem as represented by
selected ecological components. The study addressed uncertainties throughout the
assessment, including discussion of potential confounding factors and use of Monte
Carlo analysis.
*Toxicological and ecological techniques were used in an integrated manner so
that a large and complex ecosystem was evaluated for chemical impacts. The
methods used to evaluate those impacts can be readily adapted to other sites and
8-34
-------
Comments on Risk Characterization (continued)
other types of chemical impacts. The pathways methodology for relating chemical
concentrations in abiotic media to concentrations and effects in biota and the logic
used to select indicator species, or species for tissue analysis, can also be applied
to other sites. '
Limitations include:
• The study did not complete the integration of effects and exposure data so that
ecological risks could be estimated. The study is ongoing, and these activities will
be completed.
General comments:
•Although the Biota Remedial Investigation contains many of the elements
required to perform an ecological risk assessment, the focus of the study was the
nature and extent of contamination in biological media and quantifying the
interrelationships between contaminants in abiotic media and in biota by use of a
pathways model. Abiotic data were available at the time of the Biota Remedial
Investigation study report for the purpose of selecting study sites. Conclusions on ,
the exposure assessment and assessment of overall risk to biota are continuing with
calibration and validation of the model combined with additional biota sampling.
A complete risk assessment for any site cannot be completed until the abiotic data
are available to the risk assessor, and this is true for ecological or human health
risk assessments.
8-35
-------
8.4. REFERENCES
Belisle, A.; Reichel, W.L.; Locke, L.N.; Lament, T.G.; Mulhern, B.M.; Prouty, R.M.; DeWolf,
R.B.; Cromartie, E. (1972) Residues of organochlorine pesticides, polychlorinated
biphenyls, and mercury and autopsy data for bald eagles, 1969 and 1970. Pestic. Monti. J.
6:133-138.
Beyer, W.N.; Gish, C.D. (1980) Persistence in earthworms and potential hazards to birds of
soil applied DDT, dieldrin, and heptachlor. /. Appl Ecol 17:295-307.
Bigger, J.W.; Nielsen, D.R.; Tillotson, W.R. (1984) Movement of DBCP in laboratory soil
columns and field soils to groundwater. Environ, Geol, 5:127-131.
Blus, L.L. (1982) Further interpretation of the relation of organochlorine residues in brown pelican
eggs to reproductive success. Environ. Pollut. (Series A) 28:15-33.
Boudou, A.; Ribeyre, F. (1983) Contamination of aquatic biocenoses by mercury compounds:
an experimental lexicological approach. In: Nriagu, J.O., ed. Aquatic toxicology.
New York, NY: John Wiley and Sons. pp. 73-116.
Cohen, J.E. (1978) Food webs and niche space. Princeton, NJ: Princeton University Press.
Davison, K.L.; Engebretson, K.A.; Cox, J.H. (1976) p,p'DDT and p,p'DDE effects on egg
production, eggshell thickness, and reproduction of Japanese quail. Bull. Environ. Contain.
Toxicol 15:265-270.
DeWeese, L.R.; McEwen, L.C.; Petersen, B. (1982a) Chemical contaminant residues and, effects
on kestrels and other wildlife of Rocky Mountain Arsenal, part I. U.S. Fish and Wildlife
Service. Microfilm RMA134, Frames 0649-0652. Rocky Mountain Arsenal Information
Center, Rocky Mountain Arsenal, Commerce City, CO.
DeWeese, L.R.; McEwen, L.C., Petersen, B. (1982b) Chemical contaminant residues and effects
on kestrels and other wildlife of Rocky Mountain Arsenal. Progress report parts I and II.
U.S. Fish and Wildlife Service. Rocky Mountain Arsenal Information Center, Rocky
Mountain Arsenal, Commerce City, CO.
EBASCO Services, Incorporated. (1988) Work plans and submittals for the interim action of basin
F hazardous waste cleanup. Prepared for the U.S. Army Corps of Engineers, Omaha
District. February 1988. Rocky Mountain Arsenal Information Center, Commerce City,
Colorado. RIC# 8189R01-05.
Environmental Science and Engineering (ESE). (1988) Final air remedial investigation report
(volume I, version 3.1). Prepared for Office of the Program Manager, Rocky Mountain
Arsenal, Commerce City, CO.
8-36
-------
Environmental Science and Engineering (ESE). (1989) Final biota remedial investigation report
(volumes I, II, III; version 3.2). Prepared for Office of the Program Manager, Rocky
Mountain Arsenal, Commerce City, CO. ' '•
Finley, R.B. (1959) Investigations of waterfowl mortality at the Rocky Mountain Arsenal. Internal
U.S. Fish and Wildlife Report, Denver Wildlife Research Center. Denver Federal Center,
CO. Rocky Mountain Arsenal Information Center, Commerce City, CO. RIC# 87091R-06.
Fordham, C.L.; Reagan, D.P. (1991) Pathways analysis method for estimating water and sediment
criteria at hazardous waste sites. J. Environ. Toxicol. Chem. 10(7):949-960.
Haegele, M.A.; Hudson, R.H. (1974) Eggshell thinning and residues in mallards one year after
DDE exposure. Arch. Environ. Contam. Toxicol, 2:356-363.
Hill, E.F. (1988) Brain cholinesterase activity of apparently normal wild birds. J. Wildl. Dis.
24:51-61.
Linkie, J.G.; Stiles, H.E. (1976) Residue levels cause death to birds, Rocky Mountain Arsenal.
Microfilm RFA002, Frame 0262F-0272. Rocky Mountain Arsenal Information Center,
Commerce City, CO.
McEwen, L.C. (1981) Collection of articles and letters from Fish and Wildlife Service. Microfilm
RSH855, Frame 1544F-1547. Rocky Mountain Arsenal Information Center, Commerce
City, CO.
McEwen, L.C. (1983) Memorandum, 4 March 1983. Microfilm RFA003, Frame 0301-0303.
Rocky Mountain Arsenal Information Center, Commerce City, CO.
Moriarity, F. (1985) Bioaccumulation in terrestrial food chains.'In: Sheshau, P.; Korte, F.; Klein,
W.; Bourdeau, P., eds. Appraisal of tests to predict the environmental behavior of
chemicals. New York, NY: John Wiley and Sons.
Olson, D.M.; Christensen, G.M. (1980) Effects of water pollutants and Other chemicals on fish
acetylcholinesterase. Environ. Res. 21:327-335.
Robinson, J.; Hunter, C.G.; Hunter, K.W. (1967) Aldrin and dieldrin—the safety of present
• exposures of the general populations of the United Kingdom and the United States. Food
Cosmetic Toxicol. 5:781-787.
* . -. •- •
Robinson, S.C.; Kendall, R.J.; Robinson, R.; Driver, C.J.; Lacher, T.E., Jr. (1988) Effects of
agricultural spraying of methyl parathion on cholinesterase activity and reproductive success
in wild starlings (Stumus vulgaris). Environ. Toxicol. Chem. 7:343-349.
Rosenlund, B.; Jenning, S.D.; Karey, B.; Jackson, T.; Bergersen, E. (1986) Contaminants in
aquatic systems at the Rocky Mountain Arsenal—1984. U.S. Fish and Wildlife Service final
report.
8-37
-------
Snarski, V.M.; Olson, G.F. (1982) Chronic toxicity and bioaccumulation of mercuric chloride
in the fathead minnow (Pimephales promelas). Aquatic Toxicol. 2:143-156.
Spacie, A.; Hamelink, J.L. (1985) Bioaccumulation. In: Rand, G.M.; Petrocelli, S.R., eds.
Fundamentals of aquatic toxicology. New York, NY: Hemisphere Publishing Company, pp.
495-525.
Spann, J.W.; Heinz, G.H.; Hulse, C.S. (1986) Reproduction and health of mallards fed endrin. J.
Environ. Toxicol. Chem. 5:755-759.
Stickel, W.H.; Stickel, L.F.; Spann, J.W. (1969) Tissue residues of dieldrin in relation to
mortality in birds and mammals. In: Miller, M.W.; Berg, G.C., eds. Chemical fallout:
current research on persistent pesticides. Springfield, IL: Chas. C. Thomas.
Stickel, L.F. (1973) Pesticide residues in birds and mammals. In: Edwards, C.A., ed.
Environmental pollution by pesticides. New York, NY: Plenum Press, pp. 254-312.
St. Omer, V. V. (1970) Chronic and acute toxicity of the chlorinated hydrocarbon insecticides in
mammals and birds. Can. Vet. J. 11:215-226.
Thomann, R.V. (1981) Equilibrium model of fate of microcontaminants in diverse aquatic food
chains. Can. J. Fish. Aquatic Sci. 38:280-296.
Thorne, D. (1984) Response to congressional inquiry, dated 30 March 1984, by Representative
Patricia Schroeder. Microfilm RMA153, Frame 2150F-2152. Rocky Mountain Information
Center, Commerce City, CO.
Tomlinson, G.; Mutus, B.; McLennan, I. (1981) Activation and inactivation of acetylcholinesterase
by metal ions. Can. J. Biochem. 59:728-735.
U.S. Army Dugway Proving Ground (USA DPG). (1973) Incident report on the wildlife mortalities
at Rocky Mountain Arsenal during the period April 4 through June 14, 1973. RIC#
84284R-01. Rocky Mountain Information Center, Commerce City, CO.
U.S. Army Environmental Hygiene Agency (USA EHA). (1976) Rocky Mountain Arsenal bird kill.
Entomological Special Study No. 44-123-76. Aberdeen Proving Grounds, Aberdeen, MD.
RIC# 83020R-03. Rocky Mountain Information Center, Commerce City, CO.
U.S. Environmental Protection Agency (EPA). (1985) Guidance on remedial investigations under
CERCLA. Office of Solid Waste, Cincinnati, OH. EPA 540/6-85/002.
U.S. Environmental Protection Agency (EPA). (1987) Health advisories for 16pesticides. Office
of Drinking Water, Washington, DC. PB87-200176.
8-38
-------
U.S. Fish and Wildlife Service (USFWS). (1965) Effects of pesticides on fish and wildlife: 1964
research findings of the U.S. Fish and Wildlife Service. USFWS Circular #226. RIC#
87091R-02. Rocky Mountain Information Center, Commerce City, CO.
U.S. Fish and Wildlife Service (USFWS). (1983) Organochlorine chemical residues in brains of
birds and one mammal found dead at Rocky Mountain Arsenal in 1982. RIC# 87091R-05.
Rocky Mountain Information Center, Commerce City, CO.
Woolson, E.A. (1983) Emissions, cycling, and effects of arsenic in soil ecosystems. In: Fowler,
B.A., ed. Biological and environmental effects of arsenic. New York, NY: Elsevier
Science Publishers, pp. 51-120.
Woolson, E.A., Axley, J.H., Kearney, P.C. (1971) The chemistry and phytotoxicity of arsenic
in soils: I. contaminated field soils. Soil Sci. Soc. Am. Proc. 35:938-943.
8-39
-------
-------
APPENDIX A
SUMMARY OF EQUATIONS AND UNCERTAINTY FACTORS
8-A1
-------
APPENDIX A. SUMMARY OF EQUATIONS AND UNCERTAINTY FACTORS
Equations
Values such as reference doses or slope factors are unavailable for wildlife species.
Instead, abiotic concentrations predicted to be safe for the most sensitive or most important species
were derived, and assumed acceptable to the less sensitive or less important members of the
ecosytem. The toxicological literature was surveyed for the lowest NOELs or LOAELs:
LQAEL QR NQEL fmg/kg bw/day) = acceptable concentration in surface water (mg/L) (1)
water intake (L/kg bw/day)
A similar process is applied to derive soil criteria for an ingestion pathway. Soil ingestion
rate (kg/kg bw/day) is substituted into equation (1) above to yield soil criterion in mg/kg.
To determine dietary exposure from bioaccumulative contaminants, a model for single food
chains (Thomann, 1981) was adapted for an aquatic food web (multiple food chains) by adding a
term for the percentage of prey organism in the predator organism's diet, and determining
equations by which the multiple food chains could be summed. The model structure is:
Trophic Levels
(1) BCF = Cb/Cw where: Cb = tissue 1 (2)
Cw = water
(2) BAF2 = BCF2 + f2^CFl 2 (3)
(3) BAF3 = BCF3 + f3BCF2 + f^ECF^ 3 (4)
(4) BAF4 = BCF4 + f4BCF3 + f4f3BCF2 + f^BCFj 4 (5)
Additional discussion regarding the model as developed by Thomann (1981) can be found
hi Spacie and Hamelink (1985).
BMP (sum of all BAFs in the food web) was calculated with variations of the following
general equation:
i + E fi BAFM (6)
BMP is used to imply that results are for consideration of multiple food chains.
When BCF = 0 for organisms whose trophic level exceeds 1 (i.e., terrestrial predators
feeding in the aquatic food web):
i = E fi BCF1 (7)
8-A2
-------
Equation (7) essentially sums the BAFs. In our food web, this was applied to mallards
(trophic level 2) feeding on different components assumed to be trophic level 1. For the purposes
of the model, not only plants but invertebrates were assumed to be trophic level 1; this was
because the surface area to volume ratio was high, and it was assumed that BCF would make
uptake from diet appear negligible for the contaminants of concern for this study. This assumption
might not be applicable to other sites.
When BCF = 0 for organisms whose trophic level exceeds 1, BAFs must not be added
directly because BCF for the organism is counted every time a pathway they appear in is added to
another. Thus, BMP equations vary with trophic level and the structure of the food web:
BMF2 = BCF2 + E f2BCF1 (8)
BMF3 = BCF3 + E f3(BMF2) (9)
BMF4 = f4(BMF3) + f3(BMF2) (10)
Equation (8) was applied to bluegill (trophic level 2 in the food web). Bluegill have a BCF
component, as well as uptake from the various lower trophic levels. Equation (9) was applied to
pike in our food web. Pike were assumed to feed exclusively on smaller fish (represented by
bluegill). This assumption is conservative, since small pike might take invertebrates as well.
However, at RMA large numbers of alternative food for pike (ducklings, invertebrates) were not
observed. Since only one food source was modeled, a "sum of f" term was not used. In other
food webs, where trophic level 3 could be modeled with multiple food chains, the equation below
is the simplest approximation:
BMF3 = BCF3 + E f3BMF2 (11)
Equation (10) was used for the bald eagle. Where two food web pathways converged and
BCF for the top trophic level was negligible, adding the concentration from each pathway was
done by adding the total accumulation factors for the pathways. The bald eagle occurred at
multiple trophic levels, reflected by the different numbers.
The food term contains variables that may often be derived from the available literature,
although some of these variables must be extrapolated between species, or worst-case assumptions
must be made:
(12)
Where: a = assimilation efficiency (fig absorbed)
g ingested
R = dietary intake (g/g bw/day)
8-A3
-------
k2 = depuration rate (day"1)
% = percentage of prey item in diet of high-trophic-level organism
The equations for the terrestrial food web are less complex, in part due to lack of data for
loss and assimilation rates for terrestrial species. Bioaccumulation factors are derived for each
species from toxicological studies; chronic, dietary data are preferred since these are more
applicable to field studies. For a simple food web of the following species, hypothetical data were
used as an example.
Species Source of BAF BAF
Plant [Plant]/[Soil] 2
Earthworm [Earthworm]/[Soil] 100
Songbird [Whole Body]/[Diet] 10
Deer mouse [Whole Bodyj/[Diet] 10
American kestrel [Whole Body]/[Diet] 10
In the above example, higher trophic levels have the same laboratory BAF from diet. In a
food chain/food web, the difference in dietary exposure is derived as follows:
Food
Chain BAF
2 10 10
(1) Soil —> Plant > Deer mouse > Kestrel 200
2 10 10
(2) Soil—> Plant > Songbird > Kestrel 200
100 10 10
(3) Soil > Earthworm —> Songbird > Kestrel 10,000
If the kestrel fed 100 percent in any food chain, the BAFs above would apply. Given the
following dietary proportion, however, a BMP for kestrel from all combined food chains is
derived:
Species Prey Proportion
Kestrel
Songbird
Deer mouse
Plant
Earthworm
0.10
0.90
0.50
0.50
Songbird
Deer mouse Plant 1.0
8-A4
-------
For the model, proportions for plants and earthworms for a "diet" of soil are considered
1.0. The proportions are multiplied together and with the pathway BAF to obtain the adjusted
BAF; these values can then be summed to obtain a food web BMP. Note that the songbird may
provide two food chains to the kestrel.
Food Chain BAF Relative Food Chain Importance Adjusted BAF
Songbird Deer mouse Kestrel
(1) 200 1.0 0.90 180
(2) 200 0.5 0.10 10
(3) 10,000 0.5 0.10 500
Total BMP 690
Several items of interest may be pointed out. Since the kestrel does not subsist on food
chain (3), overall BMP is lower than BAF for food chain (3) only. The diet may have seasonal or
annual fluctuations. Thus, when addressing ecological risk, future as well as current projections
should be made. If food chain(s) became highly utilized, risk estimates would differ.
Also, use of chronic BAF data negates the need for loss or assimilation data. Careful
selection of parameter data is important, however, depending on the food web. If the sink species
is a juvenile and not an adult, BAFs for juveniles should be used. Results for growing animals
must be interpreted cautiously, since growth dilution produces lower BAFs and diets change with
age. Hydroponic solution data may not be applicable to use as BAF for plants, as this may differ
from field data.
Site-specific variables may cause the predicted and observed values to differ. High soil
moisture, pH, cation exchange capacity, or foc may alter soil to biota BAFs from literature data.
Thus, calibration with site data is recommended.
Finally, bioaccumulation may occur from ingested soil as well as diet. This can be
addressed by adding a soil to sink or other species pathway. The proportion of soil is incorporated
as with a dietary item.
In order to relate the estimated concentration of contaminants in the food web back to an
acceptable concentration in water, a health effects endpoint, the Maximum Acceptable Tissue
Concentration (MATC), hi mg/kg bw, must be defined from review of the toxicological literature.
When divided by the Total BMF (the BMF for the target or top-level species) for the food web,
which is unitless or can be considered as L/kg bw, a water concentration is obtained in mg/L:
MATC = C^ (13)
Total BMF
8-A5
-------
Because the original food web equations began with tissue concentration in relation to water
concentration (BCF), the final concentration factor also relates back to water. A sediment
concentration can be estimated by assuming partitioning occurs between water and sediment as
follows:
f00 (14)
Further details of the modeling can be obtained in Fordham and Reagan (1991) and ESE
(1989).
Uncertainty Factors
Uncertainty factors were applied based on the availability and relevance of available data.
Uncertainty factors were not applied to the food web model because many conservative
assumptions were built into the model. Uncertainty factors were applied to the estimation of
surface water criteria (equation 1) from LOAELs and NOELs and intake rates in order to make the
criteria protective. The uncertainty factor approach is outlined in table 8-A1.
8-A6
-------
Table 8-A1. Uncertainty Factors Used in Establishing Acceptable Water
Concentrations (ESE, 1989)
Health Effects
Chronic NOEL
Chronic LOAEL
Subchronic NOEL
Subchronic LOAEL
Acute NOEL
Acute LOAEL, LDcn
Factor Used to
Convert Effect
to a Chronic
NOEL
~
5
10
50
100
1,000
Factor Applied
for Interspecific
Variation
5
5
5
5
5
5
Total
Uncertainty
Factor
5
25
50
250
500
5,000
8-A7
-------
-------
APPENDIX B
OBSERVED TISSUE CONCENTRATIONS OF
CONTAMINANTS IN TERRESTRIAL ORGANISMS
8-B1
-------
Table 8-B1. Contaminant Levels in Terrestrial Ecosystems—Terrestrial Program Samples (ESE, 1989)
Species Tissue
TERRESTRIAL PLANTS
Homing Whole Plant
Glory Whole Plant
Sunflouer Floors
Flowers
Leaves
Leaves
INVERTEBRATES
Earthworms Whole
Whole
Whole
00 Grasshoppers Whole
i
5
i^»
VERTEBRATES
Mallard Juvenile Carcass
Adult Carcass
Juvenile Carcass
Adult Carcass
Effi
Esg
Ring-necked Juvenile Carcass
pheasant Adult Carcass
Juvenile Carcass
Adult Carcass
Egg
ttiscle**
Location
(Section)
ttA, (26, 36)
BMA Control (20)
RMA, Basin A
RMA Control (19)
RMA Basin A
FMA Basin C
RMA (19)
RMA, South Plants
RMA Control (5)
Offpost Control
BMA Section 26
RMA Section 36
KMA Control (7, 8)
Offpost Control
RMA***
RMA
Offpost Control
Offpost Control
RMA (1)
Offpost Control
RMA
RWi
OEfpost Control
OfEpost Control
RHA
RMk
Offpost Control
Concaiinant Level in parts per nillion (me/kz we ueieht basis)(RwKeAesn*)
Arsenic (n/nt)
<0.250-5.35 (1/5)
BDL (1)
BDL (6)
Bad)
<0.250-4.51 (4/5)
1.37
BDL (1)
EEC (1)
BDL (1)
0.618-1.53 (8/8)
1.03
BDL (2)
BDL (4)
0.905-6.60 (4/4)
3.17
BDL (3)
BDL (2)
NRQ
NRQ
NRQ
NRQ
twq
NRQ
<0.250-1.82 (3/11)
BDL (4)
<0.250-1.40 (2/11)
BDL (2)
BDL (10)
<0.250-4.07 (2/20)
BDL (2)
Mercury (n/nt)
BDL (5)
BDL (1)
BDL (6)
BDL(l)
BDL (5)
BDL (1)
BDL (1)
<0.050->2.35 U/2)
<0.050-0.245 (2/8)
BDL (2)
BDL (4)
<0.050-0.108 (2/4)
0.058
BDL (3)
BDL (2)
<0.050-0.066 (2/3)
0.051
BDL (8)
BDL (6)
<0.050-0.061 (1/8)
0.173-0.185 (2/2)
0.179
<0.050-0.186 (5/10)
0.068
BDL (11)
BDL (4)
BDL (11)
BDL (2)
BDL (11)
BDL (20)
BDL (2)
Aldrin
BDL
BCL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
0.046-5
1.
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
(n/nt)
(5)
(1)
(6) .
(1)
(5)
(1)
(1)
(1)
(7)
(1)
.8 (4/4)
59
(4)
(3)
(2)
(3)
(8)
(6)
(8)
(2)
(10)
(12)
(4)
(14)
(3)
(11)
(20)
(2)
BieUrin (n/nc)
<0.046-0.034 (2/5)
BDL (1)
BDL (6)
BEL (1)
BDL (5)
X).300 (1)
BDL (1)
1.93 (i)
<0.062-5.30 (1/7)
. BDL (1)
0.496-7.2 (4/4)
2.53
0.271-0.446 (4/4)
0.381
BDL (3)
BDL (2)
<0.031-0.522 (2/3)
. 0.201
<0.031-4.53 (3/8)
BDL (6)
BDL (8)
3.0-4.89 (2/2)
3.94
BDL (10)
<0.031-1.33 (5/12)
<0.031-2.92 (3/4)
0.767
<0.031-J8.6 (1/14)
BDL (3)
<0.031-5.38 (9/11)
1.12
<0.018-0.063 (2/20)
BDL (2)
Endrin (n/nt)
BDL (5)
BDL(l)
BDL (6)
Bit (1)
BDi (5)
0.188 (1)
BDL(l)
BDL
<0.080-0.
BDL
(1)
914 (1/7)
(1)
<0.064-1.65 (3/4)
p.p-DDE (n/nt) p
NHQ
fflQ
NRQ
NRQ
NRQ
NRQ
NRQ
BDL (1)
BDL (8)
BDL (1)
BDL (4)
,p-BDT (n/nt)
WQ
NRQ
NRQ •
NRQ
NBQ
NRQ
NRQ
BDL
BDL
BDL
H3L
(1)
(8)
(1)
(4)
0.528
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
B1TL
DDL
BDL
(4)
(3)
(2)
(3)
(8)
(6)
(8)
(2)
(10)
(12)
(4)
(14)
(3)
BDL (4)
BDL (3)
BE, (2)
•ffl.094-0.507 (1/3)
BDL-0.360 (4/8)
0.239
BDL (6)
<0.094-1.02 (2/8)
0.606-0.919 (2/2)
0.762
<0.094-1.35 (6/10)
0.302
BDL (11)
BDL (3)
ffl.lW-l,* (1/12)
BIJ, (2)
<0.40-0.143 (1/11) BDL (10)
BDL
BDL
(20)
(2)
BDL (20)
BDL (2)
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BIX.
BDL
BDL
BDL
BDL
sra.
BDL
BDL
SDL
(i)
(3)
(2)
(3)
(8)
(6)
f8)
(2)
(2)
(11)
(3)
(12)
(2)
(in)
(20)
(2)
-------
Table 8-B1. Contaminant Levels in Terrestrial Ecosystems—Terrestrial Program Samples (continued)
00
Cd
Species
Tissue
Location
Contaminant Level in parts per million (mg/kp wet weight basis)(Range/Mean*)
(Section) Arsenic (n/nt)
Ring- necked
pheasant
American
Kestrel
Prairie Dog
Cottontail
Mule Deer
Liver**
Eg*
Juvenile
Juvenile
EgR
Egg
Carcass
Carcass
Carcass
Carcass
Carcass
Carcass
Kidneys
Muscle
Muscle
Muscle
Liver
Liver
Muscle
Muscle
Btt
Offpost Control
Offpost Control
Carcass KMA
Carcass Offpost Control
RMA
Offpost Control
FMA (36) Sunnier <0,
IMA (36) Winter
RHA, TSf <0
RMA Control Simmer (19, 20)
RMA Control Winter (20)
Offpost Control Summer
EMA, (36) Winter
FMA, (36)
RMA Control (19, 20)
Offpost Control .
RMA
Offpost Control
RMA
Offpost Control
NRQ
NPQ
BDL (10)
NK)
NK)
NRQ
NRQ
.250-0.741 (2/9)
BDL (5)
.250-4.22 (1/5)
BDL (9)
BDL (5)
BDL (9)
BDL (5)
BDL (7)
BDL (7)
BDL (7)
BDL (14)
BDL (2)
BDL (14)
BDL (2)
Mercurv (n/nt)
NBQ
NRQ
BDL (11)
BDL (10)
BDL (8)
<0.050-0.405 (8/34)
<0.050-0.057 (1/11)
BDL (9)
BDL (5)
BDL (5)
BDL (9)
BDL (5)
BDL (9)
<0.1(H).356 (3/5)
0.178
BDL (7)
BDL (7)
BDL (7)
BDL (14)
BDL (2)
BDL (14)
BDL (2)
Aldrin (n/nt)
BEL (6)
BEL (2)
BDL (11)
BDL (10)
BEL (8)
BDL (33)
BDL (11)
BDL (9)
BEL (5)
BDL (5)
BDL (9)
BDL (5)
BEL (8)
BDL (5)
BDL (7)
BDL (7)
BDL (7)
BDL (14)
BDL (2)
BDL (14)
BDL (2)
Dieldrin (n/nt)
Endrin (n/nt) p,p-DDE (n/nt) p
<0.018-2.3 (4/6) BDL-0.091 (1/6) BDL-Q.44 (1/6)
0.655
BDL (2)
BDL (11)
<0.031-1.01 (5/10)
0.316
BDL (8)
<0.031-3.63 (17/33)
XJ.512
BDL (11)
0.233-13.4 (9/9)
2.03
0.119-6.18 (5/5)
1.44
O.Ott-0.155 (5/5)
0.114
<0.031-0.346 (2/9)
<0.031-0.096 (1/5)
BDL (8)
<0.248-1.54 (2/5)
-------
Table 8-B2. Miscellaneous Samples: Samples of Chance and USFWS Supplemental Samples (ESE, 1989)
00
Spec!e<
Blue-winged
tnl
Redhud
American Coot
Mourning Dove
Bald Eagle
Golden Eagle
Ferruginous
Hawk
Red-tailed
Hawk
Great-homed
Owl
Northern
Harrier
Coyote
Badger
Tissue
Livw
Muide
liver
Muscle
Liver
Muscle
Carcass
Liver
Egg
Liver
Brain
Uver
Brain
Liver
Brain
Liver
Brain
Egg
Liver
Liver
Kidneys
Location
(Section)
RMA
Upper Derby
RMA
Upper Derby
RMA
Upper Derby
RMA
Upper Derby
RMA
Upper Derby
RMA
Upper Derby
RMA (35)
RMA(1)
Barr Lake
RMA"
RMA
RMA
RMA
RMA
RMA
RMA
RMA
RMA
RMA (25)
RMA (25)
RMA (25)
Cootimlninl Level In ptra per mfflton (aiftl wet weight bun) (Rmjc/Mtin*)
Arsenic (o/ot)
BDL (3)
BDL (3)
BDL (5)
BDL (5)
BDL (9)
BDL (9)
BDL (2)
BDL (1)
BDL
BDL(1)
BDL (2)
BDL (5)
BDL (5)
BDL (3)
BDL (3)
BDL (4)
BDL (4)
BDL (2)
BDL(1)
BDL (1)
NRQ
Mercury (n/nt)
0371-L64 (3/3)
1.07
O259-0559 (3/3)
0391
0080-0368 (5/5)
0.211
•C0050-0.073 (Iff)
0300-1.77 (90)
1.08
<0.050-0339 (8/9)
0.179
BDL (2)
BDL (1)
0099
•eO.050-0.216 (1/2)
O120
<.098-a57 (2)
<0.050-0.293 (1/5)
<0050-0.152 (1/5)
<0050-0345 (1/3)
<0.050-0.0» (1/3)
<0,050-0.086 (2/4)
0.047
BDL (4)
BDL (2)
BDL(1)
BDL (1)
NRQ
AMrin(nAl)
BDL (3)
BDL (3)
<003(W>.OSS (US)
BDL (5)
BDL (9)
BDL (9)
<0.633-133 (2/2)
1.23
BDL(l)
BDL(l)
BDL (2)
BDL (2)
BDL (5)
BDL (5)
BDL (3)
BDL (3)
BDL (4)
BDL (4)
BDL (2)
BDL(l)
BDL(1)
BDL(1)
DfcHrin (nhv)
0.1834281 (3/3)
O239
0.090-0.164 (3/3)
0.127
0307-0.747 (5/5)
0.458
0.117-0320 (5/5)
0203
-------
APPENDIX C
OBSERVED TISSUE CONCENTRATIONS IN
RAPTOR SAMPLES OF CHANCE AND BELIEVED CAUSE OF MORTALITY
8-C1
-------
Table 8-C1. Observed Tissue Concentrations in Raptor Samples of Chance and Believed
Cause of Mortality (ESE, 1989)
Species
Ferruginous Hawk
ferruginous Hawk
Ferruginous Hawk
Ferruginous Hawk
Ferruginous Hawk
Red-tailed Hawk
Red-tailed Hawk
Red-tailed Hawk
Great-horned Owl
Great-horned Owl
Great-horned Owl
Great-horned Owl
Golden Eagle
Golden Eagle
Age*
A
A
1
A
1
1
A
1
A
A
A
A
1
Physical
Condition
Emaciated
Good
Emaciated/
Convulsions
Good
No body fat
Emaciated
Unknown
Unknown
Unknown
Emac i ated
Good
Unknown
Unknown
Good
Contaminant Levels
lercury
BDL/BDL
0.152/0.293
BOL/BDL
BDL/BDL
BDL/BDL
BOL/BDL
0.093/0.345
BDL/BDL
BDL/0.086
BDL/BDL
: BDL/BDL
BDL/ 0.051
0.257/0.216
BDL/BDL
0 1 e i dr i n
0.678/0.527
7.73/4.79
9.98/3.45
BDL/0.263
6.85/4.26
9.44/5. 19
9.2/6.59
BDL/0.52
15.6/10.8
9.32/27.7
BDl/0.143
10.2/8.89
SDL/0.221
BDL/BDL
of Brain/Liver Cause of
DDE
BDL/BDL
BDL/BDL
BDL/BDL
BDL/BDL
.. BDL/BDL
BDL/0.529
BDL70.759
BDL/BDL
10.3/15.5
0.475/2.47
BDL/BDL
2.24/5.49
BDL.'BDL
BDL/BDL
Death
Unknown
Unknown
Unknown
Electrocution***
Unknown?**
Unknown***
Unl-nown
Electrocution
Unknown
Unknown
Unl- nown
Entero to -.em i a***
Untnown*"*
Respiratory failure***
« A » Adult
1 « Immature
•• On uet weight basis
BDL * Below Detection Limit
*«* Determined by Dr. Leroy Eggleston. DVM.
or Dr. Terry Spraker. DVM.
. 8-C2
-------
SECTION NINE
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
SELENIUM EFFECTS AT KESTERSON RESERVOIR
-------
AUTHORS AND REVIEWERS
AUTHORS
Margaret McVey
ICF International, Inc.
Fairfax, VA
REVIEWERS
Wayne G. Landis (Lead Reviewer)
Institute of Environmental
Toxicology and Chemistry
Western Washington University
Bellingham, WA
James J. Anderson
School of Fisheries Research
University of Washington
Seattle, WA
Jay S. Jacobson
Boyce Thompson Institute
Cornell University
Ithaca, NY
Larry Kapustka
Ecological Planning and Toxicology, Inc.
Corvallis, OR
Bruce A. Macler
Region 9
U.S. Environmental Protection Agency
San Francisco, CA
Thomas Sibley
School of Fisheries Research
University of Washington
Seattle, WA
Frieda B. Taub
School of Fisheries Research
University of Washington
Seattle, WA
Donald P. Weston
Horn Point Environmental
Laboratory
University of Maryland
Cambridge, MD
9-2
-------
CONTENTS
ABSTRACT 9-7
9.1. RISK ASSESSMENT APPROACH 9-8
9.2. STATUTORY AND REGULATORY BACKGROUND 9-8
9.3. CASE STUDY DESCRIPTION . 9-8
9.3.1. Problem Formulation 9-8
9.3.2. Analysis: Characterization of Ecological Effects 9-17
9.3.3. Analysis: Characterization of Exposure 9-19
9.3.4. Risk Characterization 9-29
9.4. REFERENCES 9-36
9-3
-------
LIST OF FIGURES
Figure 9-1. Structure of analysis for selenium effects at Kesterson Reservoir 9-9
Figure 9-2. Site location of Kesterson Reservoir . 9-10
Figure 9-3. Simplified selenium transfer diagram with key to transfer and diet factors . . . 9-23
Figure 9-4. Probability distribution of predictions of selenium concentration
in receptor species diet for Flexible Response Plan 9-30
Figure 9-5. Probability distribution of predictions of selenium concentration
in receptor species diet for Onsite Disposal Plan 1 9-31
Figure 9-6. Probability distribution of predictions of selenium concentration
in receptor species diet for Onsite Disposal Plan 2 9-32
9-4
-------
LIST OF TABLES
Table 9-1. Frequency of Dead or Deformed Embryos or Chicks in Nests
of Aquatic Birds at Kesterson Reservoir, 1983-1985 9-12
Table 9-2. Kesterson Bird Populations and Mortalities 9-13
Table 9-3. Summary of Contaminant Levels in KR Media 9-14
Table 9-4. Summary of Dose Response Reported for Avian Species 9-18
Table 9-5. Selenium Concentration (ppm, dry weight) in Composite
Samples of Plants, Invertebrates, and Mosquitofish,
May 1983 9-21
Table 9-6. Transfer and Diet Factors for Simplified Selenium Transfer
Diagram for Mallard, American Coot, Tricolored Blackbird,
and Black-Necked Stilt 9-24
Table 9-7. Transfer and Diet Factors for Simplified Selenium Transfer
Diagram for Eared Grebe and Mosquitofish 9-25
Table 9-8. Transfer and Diet Factors for Simplified Selenium Transfer
Diagram for San Joaquin Kit Fox
Table 9-9. Summary of Data Used to Derive Transfer Factors 9-27
Table 9-10a. Percentage of Diet Selenium Predictions That Are Below the
5 mg/kg Estimated Harmful Level for Each Key Species and
Each Cleanup Alternative 9-33
Table 9-10b. Percentage of Diet Selenium Predictions That Are Below the
10 mg/kg Estimated Harmful Level for Each Key Species and
Each Cleanup Alternative 9-33
9-5
-------
LIST OF ACRONYMS
CVWQCB Central Valley Water Quality Control Board
DFG California Department of Fish and Game
EIS Environmental Impact Statement
EPA U.S. Environmental Protection Agency
FRP Flexible Response Plan
FWS U.S. Fish and Wildlife Service
KR Kesterson Reservoir
LC50 lethal concentration to 50 percent of organisms tested
NRDC Natural Resources Defense Council
SLD San Luis Drain
SWRCB California State Water Resources Control Board
IDS total dissolved solids
USBR U.S. Bureau of Reclamation, Department of the Interior
9-6
-------
ABSTRACT
Subsurface drainage of agricultural water in the San Joaquin Valley, California, to
Kesterson Reservoir over several years resulted in a variety of adverse effects on waterfowl nesting
in the reservoir area by 1983. Between 1983 and 1985, embryonic mortality and deformity rates at
the site were significantly higher than those at control sites for several species of waterfowl.
Review of information related to past and present contamination at Kesterson Reservoir
indicated that selenium was a major contaminant of concern because it had exceeded water quality
guidelines and criteria, had accumulated in reservoir soils, had migrated into the ground water in
some locations, and had been linked experimentally and observationally to wildlife effects.
Measured tissue selenium levels at the site were elevated relative to controls for aquatic plants,
invertebrates, mosquitofish, and several species of waterfowl and small mammals. In response to
public pressure and the provisions of the Clean Water Act and the Federal Migratory Bird Treaty
Act, the U.S. Bureau of Reclamation commissioned a study of possible remediations at the site.
This ecological assessment was conducted to estimate the effectiveness of three cleanup alternatives
in reducing the selenium levels in Kesterson Reservoir to levels protective of waterfowl and other
wildlife.
To estimate the reduction in risk associated with the three cleanup alternatives, seven
representative wildlife species in the area were selected for the risk assessment. For this
assessment, food chain exposure was considered the most important pathway for exposure of fish
and wildlife to selenium. The mallard, American coot, black-necked stilt, tricolored blackbird,
mosquitofish, eared grebe, and San Joaquin kit fox were used to represent species that could be
exposed to selenium via the midwater, benthic, and aquatic rooted plant, fish, and terrestrial
pathways. Detailed food chain exposure diagrams for each of these species were developed into
simplified selenium transfer models. These models were used with calculated transfer factors
derived from experimental data and a Monte Carlo simulation technique to estimate the probability
distribution of predictions of selenium concentration hi the diets of the key species under each of
the three remedial alternatives. The risk assessment indicated that each of the three remediation
plans might present some risks to wildlife.
This case study represents a true risk assessment (i.e., for each remedial alternative, the
probability that residual contamination of food chain organisms would be below harmful levels to
key species at the refuge was calculated). Moreover, a weight-of-evidence approach was used to
attribute the adverse effects of Kesterson Reservoir on waterfowl to one of the drainwater
contaminants (i.e., selenium). The use of food web models was a good attempt to account for
ecosystem processes in estimating exposures. The results of this risk assessment led to
abandonment of the proposed remediations and adoption of site mitigation. The wetland was
closed and converted to a terrestrial habitat. A larger wetland was created to provide for the lost
wetland habitat at Kesterson.
9-7
-------
9.1. RISK ASSESSMENT APPROACH
This case study represents a complete risk assessment in that it incorporates all the elements
that constitute the risk assessment process (figure 9-1). The analysis included modeling of
probabilities of levels of dietary selenium exposure to selected ecological components (species).
Modeled exposures were based on environmental measurements at the site and incorporated data on
transfer of material through food chain elements. Dietary exposures were compared with no
observed adverse effect levels based on several measurement endpoints.
9.2. STATUTORY AND REGULATORY BACKGROUND
This risk assessment for Kesterson Reservoir (KR) was conducted by the U.S. Department
of the Interior, Bureau of Reclamation (USER), in response to a suit filed by the Natural Resources
Defense Council (NRDC) against the California State Water Resources Control Board (SWRCB).
The NRDC suit claimed violation of state water quality standards at KR and resulted in the
SWRCB issuing Order WQ85-1 to the Central Valley Water Quality Control Board (CVWQCB) to
define water quality standards and cleanup actions for KR (SWRCB, 1986). Under provisions of
the Federal Migratory Bird Treaty Act, the USER (Mid-Pacific Region) agreed to assist the
CVWQCB to comply with the order and sponsored the assessment of remedial alternatives for KR
(USER, 1986a). This order also directed USER to control conditions that caused any threat to
wildlife or humans from operation of KR and resulted in its closure and mitigation for loss of the
wetlands.
9.3. CASE STUDY DESCRIPTION
The overall objective of this ecological risk assessment was to provide, to the extent feasible
with existing information, a quantitative analysis of the magnitude and uncertainty of estimates of
potential adverse impacts on fish and wildlife populations that could result from implementation of
KR cleanup alternatives. Three cleanup alternatives to reduce the selenium levels at KR were
developed and qualitatively evaluated in the Kesterson Program Environmental Impact Statement
(EIS) (USER, 1986b). All alternatives involved closing the drains that fed KR with selenium, and
all alternatives were designed to reduce food chain selenium exposure. From these, three scenarios
were selected for analysis as providing high and low ends of the risk spectrum. The Flexible
Response Plan (FRP) involved flooding ponds having the highest selenium concentrations with low-
selenium water to dilute remaining selenium and disrupt food chains. Ponds with lower selenium
content were to be dried out and disked to reduce vegetation. In Onsite Disposal Plan 1 (Plan 1),
soils above 4 mg/kg selenium (342,000 m3) would be excavated and disposed in a landfill. In
Onsite Disposal Plan 2 (Plan 2), all contaminated soils (760,000 m3) would be excavated and
landfilled.
9.3.1. Problem Formulation
Site Description. Kesterson Reservoir is located on the Kesterson National Wildlife Refuge
in Merced County in the San Joaquin Valley of California (figure 9-2). The San Joaquin Valley
has historically been a region where intermittent ponding occurs from storm events and runoff.
Agriculture along the western margin of this region requires irrigation, and the use of irrigation
9-8
-------
Figure 9-1. Structure of Analysis for
Selenium Effects at Kesterson Reservoir
PROBLEM FORMULATION
Stressors: selenium in sediments and surface waters.
Ecoloaical Corrmonents: birds (five soecies), kit fox.
A
and mosquitofish.
Endpoints: assessment endpoint was health and condition
of local populations of selected fish and wildlife
species; measurement endpoints included a variety of
toxicological indices.
II
I
I
Characterization of Characterization of
Exposure Ecological Effects
Selenium concentrations Effects of selenium were
were measured in sediment, evaluated based on
water, and tissues. Food laboratory feeding studies
chain transfer factors were ancj aquatic bioassays with
developed. Food chain fish- Fie|d observations
model was used to predict documented effects.
exposure.
1 A
4 *
RISK CHARACTERIZATION
Monte Carlo simulations were used to estimate probability
distributions of selenium in the diet under various
remedial scenarios. Risk estimates were developed
by combining dietary exposure and effects information.
The habitat impacts associated with the physical effects
of remediation were not considered.
9-9
-------
KESTERSON
NWR
KESTERSON
RESERVOIR
Figure 9-2. Site location of Kesterson Reservoir (Ohlendorf et al., 1989)
9-10
-------
increased substantially during the 1960s and 1970s. Natural levels of salts in the soils of the
region, coupled with water application and near-surface perched aquifers, led to accumulation of
saline ground water in root zones of cropped species, which reduced productivity. It became
apparent that drainage of these lands would be necessary to keep them in production. Subsurface
drains were installed, flowing into the San Luis Drain (SLD), which was to discharge into the
Sacramento/San Joaquin River Delta. This final link was never completed; rather, the drainwater
was delivered via the SLD to KR. KR consisted of 12 shallow ponds (1 m to 1.5 m deep), totaling
about 500 ha, designed to serve as evaporation and holding basins for this drainage water
(Ohlendorf et al., 1988). Besides drainage to KR, more than 500 million m3 of drainage water are
discharged annually into other surface aquatic ecosystems in California, primarily the San Joaquin
River and its western tributaries, the Delta-Mendota Canal, evaporation ponds in the Tulare Basin,
and the Salton Sea and its principal tributaries (Skorupa and Ohlendorf, 1991).
Concurrent with agricultural and other development, more than 90 percent of the Central
Valley's historic wetlands have been lost. Remnant wildlife populations have been concentrated
onto the remaining wetlands, including those wetlands, such as KR, receiving drainage water.
Frequently, drainage water evaporation ponds are the most common type of wetland available to
wildlife during the spring. The shallow, nutrient-enriched waters of these ponds lead to high
primary and secondary productivity and, therefore, are particularly attractive to breeding
waterbirds. By this route, the ponds provide a pathway for wildlife exposure to contaminants in
drainage water (Skorupa and Ohlendorf, 1991). KR is located on a major migratory bird flyway,
and the ponds provide nesting and feeding grounds for several bird species.
Stressors. During 1983, it became clear that aquatic birds nesting at KR were experiencing
poor reproductive success. A high frequency of both embryo mortality and developmental
abnormalities occurred in most species during the 1983 to 1985 breeding seasons at KR (table 9-1)
(Ohlendorf, personal communication in USBR, 1986a; Ohlendorf et al., 1986a, b). In contrast,
researchers found no abnormalities in embryos from nests monitored through late stages of
incubation or hatching at the Volta Wildlife Area, a control site located 10 km away in an area that
did not receive agricultural drainage waters. The expected incidence of major external
malformations in hatchlings of uncontaminated wild populations of birds and in embryos of
laboratory-incubated mallard eggs is less than 1 percent (Pomeroy, 1962; Gilbertson et al., 1976;
Hoffman, 1978; Hill and Hoffman, 1984). Table 9-2 provides estimates of avian population
densities and past KR-related avian mortalities. These estimates are based on data from published
literature, unpublished surveys by the U.S. Fish and Wildlife Service (FWS) and the California
Department of Fish and Game (DFG), and consultation with personnel from these agencies and
other local experts (USBR, 1986a).
Delivery of agricultural drainwater to KR ceased in June 1986. Contamination remaining at
KR was that portion of delivered contaminants that had accumulated in soils and biota. The
SWRCB evaluated several contaminants in SLD drainwater and KR surface water, ground water,
soils, and biota to determine if there was any evidence for residual contamination that could result
in future harmful effects to wildlife. Seven of ten drainwater constituents warranted further
analysis because they exceeded water quality guidelines and criteria in historic SLD drainwater
and KR surface water: boron, chromium, mercury, molybdenum, selenium, total dissolved solids
(TDS), and zinc (table 9-3).
9-11
-------
Table 9-1. Frequency of Dead or Deformed Embryos or Chicks in Nests of Aquatic Birds at
Kesterson Reservoir, 1983-1985 (Ohlendorf et al., 1989)
Frequency of Occurrence*
Species
Year
Coot0
1983
Grebed
1983
Ducks0
1983
1984
1985
Stilt
1983
1984
1985
Avocet
1983
1984
1985
Killdeerf
1984
1985
Total
Dead
Nestsb
59/92
141/163
30/42
13/36
17/27
101/125
63/189
69/96
16/16
19/51
22/35
12/32
16/25
578/929
No.
35
84
5
6
6
17
7
20
0
0
4
0
7
191
%
(59.3)
(59.6)
(16.7)
(46.2)
(35.3)
(16.8)
(H.l)
(29.0)
(0)
(0)
(18.2)
(0)
(43.8)
(33.0)
Deformed
No.
25
22
3
0
1
17
12
23
0
0
4
0
3
110
%
(42.4)
(15.6)
(10.0)
(0)
(5.9)
(16.8)
(19.0)
(33.3)
(0)
(0)
(18.2)
(0)
(18.8)
(19.0)
Total
No.
38
89
7
6
6
24
14
30
0
0
5
0
8
227
%
(64.4)
(63.1)
(23.3)
(46.2)
(35.3)
(23.8)
(22.2)
(43.5)
(0)
(0)
(22.7)
(0)
(50.0)
(39.3)
aDead = number of nests (and percentage) with one or more dead embryos; deformed = nests
with one or more deformed embryos or chicks; total = sum of all nests with at least one dead or
deformed embryo or chick. All percentages calculated by dividing by number of monitored nests.
bMonitored/found: nests monitored to hatching or from which a late-stage embryo was
collected/nests found during study, including those lost to predation, flooding, and desertion.
cNo coot nests found hi 1984 or 1985, although adults were present throughout the nesting
season.
"Adult birds present throughout the nesting season only during 1983.
''Mallard, gadwall, cinnamon teal, and northern pintail.
fSpecies not studied in 1983.
9-12
-------
Table 9-2. Kesterson Bird Populations and Mortalities (see also Ohlendorf et al., 1989)
Kesterson
Reservoir
Mortality
Kesterson
Reservoir
Population
San Joaquin
Valley
Population
California
Population
Pacific
Flyway
Population
Mallard
1 found dead
in 1986s
17-22 ducks
"lost"b
45-100 per
dayd
89,142h
435,421h
l,759,800n
American
Coot
438 "lost"
in 1983b
9,489e
216,623h
427,415h
562,400"
Black-Necked
Stilt
197 "lost"
in!985b
50-60 individuals
nesting in
1986"
No data
~100,000k
No data
Tricolored
Blackbird
82,150 "lost"
eggs and
chicks in
1986°
47,000^
85,850"
133,000*
No data'
Eared
Grebe
411 "lost"
in 1983b
17f
No data
730,250'
No data
San Joaquin
Kit Fox
No data
15-2QS
5,294)
10,000-14,800™
Same as
statewide
population
a Personal communication from Mary Coakley, wildlife hazer, 10-2-86. This value does not reflect nestling mortalities.
b Unpublished memo from Dr. Harry Ohlendorf to Mr. Ken Anderson of the General Accounting Office, 5-16-86.
Includes dead or deformed embryos or chicks and those presumed to have hatched but that failed to survive. Some losses were
due to predation, but these cannot be accurately separated from possible mortality due to selenium toxicosis.
0 Rough calculation assuming about 23,500 nesting pairs, a clutch size of 3.5, and mortalities of all but 100 fledglings:
d Unpublished FWS hazing data from Kesterson Reservoir, March 1986.
c Maximum average daily use total from unpublished FWS data 1984-86. See draft EIS.
Paveglio (personal communication b). Average daily use for 1982, 1984, and 1985. Peak use is 125-175 during
migration.
8 Paveglio (personal communication b). Total adult population in adjacent 25,000-acre range. Up to five individuals (including
pups) have been seen simultaneously at Kesterson Reservoir.
h Unpublished FWS data, 1973-77 average Pacific Flyway mid-winter waterfowl survey, 1986.
1 Population estimates by Rich DeHaven (FWS) for the period 1969-72. Note that this species is largely endemic to California,
so the statewide population approximately equals global population.
J Based on adult population estimates by Morrell (1975) for Kern, Tulare, Kings, Fresno, San Benito, Merced, Stanislaus,
and San Joaquin Counties.
k "Ballpark" estimate by Ron Jurek, DFG, 10-2-86.
1 Maximum extrapolated population estimate from Mono Lake, August 30, 1976 (Winkler et al., 1977); actual statewide
population is probably much higher than this (Gould, personal communication).
m Range of adult population estimates for California by Morrell (1975).
n Unpublished FWS data, 1955-85 average from Pacific Flyway midwinter waterfowl survey.
9-13
-------
Table 9-3. Summary of Contaminant Levels in KR Media (USER, 1986a)
Constituents
Boron
Cadmium
Chromium
Copper
Manganese
Mercury
Molybdenum
Nickel
Selenium
Zinc
TDS
Drainwater,
Surface Water,
or Shallow
Ground Water >
Standards?
Yes
No
Yes
No
No
Yes
Yes
No
Yes
Yes
Yes
KR Soils >
Background?
No
—
No
—
—
No
No
—
Yes
No
NAa
KR Water Supply
Ground Water >
Standards?
Yes
~
No
—
—
No
No
—
No
No
1 Yes
KR Biota >
Background?
Yes
—
.No
—
__
No
No
~
Yes
No
NAa '
aNA = not applicable; TDS not applicable to water.
9-14
-------
In 1983, researchers concluded that selenium was the most likely cause of avian deaths and
deformities at KR because the types of deformities found in avian embryos and young were typical
of those induced by exposure to high selenium levels, but not high levels of the other contaminants
of concern. Boron, molybdenum, and strontium all were characterized as below toxic levels, but
few data supported these claims. Although boron had been shown to cause mortality and
teratogenic development of eggs, the dietary concentrations used in these studies were much higher
than the levels in most dietary elements at KR (Hothem and Ohlendorf, 1989). Studies by Heinz et
al. (1987) indicated that when mallards were fed diets containing selenomethionine, some embryos
had deformities similar to those observed at KR. Studies with poultry and quail had shown several
toxic responses to dietary ingestion of selenium compounds, including reduced growth;
reproductive impairment; embryonic, hatchling, and adult mortality; and gross deformities
(Rosenfeld and Beath, 1964; National Research Council, 1976, 1980; Shamberger, 1981/1983).
Field studies had correlated high levels of dietary selenium with low hatchability, embryonic
deformity, and high mortality in wild birds at KR (Ohlendorf et al., 1986a, b; Saiki, 1986).
Selenium's potential for bioaccumulation caused additional concern. Studies by Lemly
(1985) in an aquatic ecosystem indicated that plankton could concentrate selenium to 750 times the
concentration in water and that fish could concentrate selenium to levels 4,000 times that in water.
Ecological Components. The wildlife species present at KR represented a variety of trophic
levels and, hence, selenium exposure potential. Selection of key fish and wildlife species to
represent potential exposure pathways in the risk assessment was based on several considerations.
A species was selected if it was a terminus species of a major KR food chain exposure pathway, if
impacts of KR on the species had been observed in the past, if it was a rare or endangered species,
if it was a species with particularly sensitive life stages, or if information was available on the
effects of selenium exposure for the species. Not all of the species selected satisfied all of these
criteria. A weakness of this assessment was that it ignored primary producers and organisms at the
lowest levels of the food chains. Descriptions of the key species and rationales for their selection
are provided below.
• Mallard. Adult Anas platyrhynchos are omnivorous; however, during nesting and
egg laying, the diet of adult females changes from primarily vegetation to primarily
aquatic invertebrates.
» American coot. The adult coot (Fulica americand) feeds primarily on terrestrial and
aquatic plants, insects, and other epiphytal fauna.
• Black-necked stilt. Adult Hitnantopus mexicanus feed while wading in the water
primarily on littoral benthic epifauna.
• Tricolored blackbird. Adult Agelaius tricolor feed their young almost exclusively
on adult insects and aquatic insect larvae. The status of the tricolored blackbird as
a candidate for federal threatened and endangered species listing also was a factor
hi its selection.
9-15
-------
* Eared grebe. The eared grebe (Podiceps nigricollis) is an invertebrate- and fish-
eating bird whose population at KR had been shown previously to be adversely
affected.
" Mosquitofish. The mosquitofish (Gambusia affinis) is the only fish that still existed
in KR waters in 1986; it is highly resistent to selenium toxicosis.
• San Joaquin kit fox. The kit fox was included as the terrestrial food chain receptor
because it is a federal- and state-listed endangered species.
For each of these species, a review of the scientific literature was conducted to quantify their
food habits, and dietary preferences were summarized by life stage, sex, and season, as
appropriate. The home range size of each species was estimated from literature values.
Bndpoints. The assessment endpoint was the health and condition of local populations of
selected fish and wildlife species under alternative remedial scenarios. Measurement endpoints
included field, laboratory, and literature investigations of adverse effects on birds (reduced
reproductive success, growth, and survival), kit fox (liver changes and heart, kidney, and spleen
effects), and mosquitofish (survival).
Comments on Problem Formulation
Strengths of the case study include:
^Detailed description of the site and profiles of metal contamination are available,
+The review of criteria for the inclusion of the selected ecological components
(species) is sufficiently detailed. In addition, substantial data were collected from
thefield.
Limitations include:
•Boron is not included as a stressor because of the lack oftoxicity data. Some of
the resources of this study could have been allocated toward providing that
information.
^Habitat alterations are not included as a stressor, although these would have
occurred from the proposed remediations. These remediations would largely destroy
the existing habitat and severely affect existing food chains at all levels. The lack of
consideration of habitat alteration was a result of the Federal Migratory Bird Treaty
Act, which focuses on chemical impacts, such as those from selenium.
9-16
-------
Comments on Problem Formulation (continued)
• Ecological components include key species of social interest. However, the
ecosystem as a whole and species important to ecosystem stability are not
addressed. In particular, lower food chain elements are ignored. Many plants are
sensitive to elevated boron, and damage to these organisms at low trophic levels
could lead to large changes in the ecosystem. By focusing on the removal of
selenium from incorporation into the food chains at KR, consideration of overall
survival of the key species is ignored.
General comment:
9 This study is based on a $2 to $3 million research effort to characterize the
impacts of metal contamination on the waterfowl and endangered species ofKR,
The studies were initiated due to the discovery of malformation and other
teratogenic effects in waterfowl chicks. These teratogenic impacts caused a large
public outcry and the implementation of the Federal Migratory Bird Treaty Act.
The emphasis on avion species within KR was driven by these factors.
9.3.2. Analysis: Characterization of Ecological Effects
Birds. Selenium toxicity has been reviewed extensively and documented in poultry and quail
in studies dating back to the 1930s (Rosenfeld and Beath, 1964; National Research Council, 1976,
1980; Shamberger, 1981, 1983). The early studies of chickens receiving selenium in their diet
from cereal grains grown in seleniferous soils showed effects ranging from both reduced growth
and reproductive impairment to complete failure of hatching (Moxon, 1937; Poley et al., 1937;
Poley and Moxon, 1938; Moxon and Rhian, 1943). The selenium content of the grain was
speculated to be as low as 10 ppm (Heinz et al., 1987). Ort and Latshaw (1978) identified 5 ppm
selenium in feed (added as sodium selenite) as the threshold for reduced hatching success in
chickens. Edema of the head and neck was seen at 7 to 9 ppm. Heinz et al. (1987) induced
reproductive impairment and embryonic deformities in mallards at dietary levels as low as 10 ppm
selenium as selenomethionine. Based on the existing information for birds, the range of harmful
dietary selenium threshold concentrations was assumed to be 5 to 10 ppm. Field studies had
correlated high levels of dietary selenium with low hatchability, embryonic deformity, and high
mortality in wild birds at KR (Ohlendorf et al., 1986a, b, 1988; Saiki, 1986), but extreme
conditions at KR provided little opportunity to assess thresholds for selenium toxicity. A summary
of ecological effects reported for avian species is shown in table 9-4.
Mammals. Selenium concentrations of 8 to 30 ppm (dry weight) in the diet have been
associated with chronic toxicity in mammals (Wilber, 1980). Indications of toxicity such as liver
changes and heart, kidney, and spleen effects have been observed in mammals following chronic
exposures to feed containing 1.4 to 3.0 ppm selenium (Anspaugh and Robinson as cited in USBR,
1986a). On the other hand, Halverson et al. (1966) observed no significant effect on growth in
9-17
-------
Table 9-4. Summary of Dose Response Reported for Avian Species
Dose
(ppm)
100
78
40
25
10
VO
SS 12
10
10
7
7
6
5
5
8
Chemical Form
Sodium selenite
Selenium selenite
Sodium selenite
Sodium selenite
Selenomethionine
Sodium selenite
Selenomethionine
Sodium selenite
Selenium selenite
Selenium selenite
Selenomethionine
Sodium selenite
Sodium selenite
Sodium selenite
Response
Mortality of adults; weight loss
Lowered egg production
Reduced chick survival
Mortality of adults
Depressed body weight - adult
Reduced egg laying
Reduced duckling survival
Lower Radcliffe index
Depressed body weight - duckling
Reduced duckling survival
Low hatching success
18% abnormal embryos
Multiple malformations
Depressed body weight
Lower hatchability
Low hatching success
Multiple malformations
Lowered hatching success
Reduced egg weight
No effect on adult or egg
production
Reduced growth
Impaired hatching success
Lowered chick survival
Test
Organism
Mallard
Chicken
Chicken
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Mallard
Japanese quail
Chicken
Mallard
Chicken
Chicken
Chicken
Chicken
Chicken
Japanese quail
Reference
Heinz et al., 1987
Arnold etal., 1973
Arnold et al., 1973
Heinz etal., 1987
Heinz et al., 1987
Heinz et al., 1987
Heinz etal., 1987
Heinz etal., 1987
Heinz etal., 1987
Heinz et al., 1987
Heinz et al., 1987
Heinz et al., 1987
Heinz et al., 1987
Heinz et al., 1987
El-Begearmi et al., 1977
Poley and Moxon, 1938
Heinz et al., 1987
Ort and Latshaw, 1978
Ort and Latshaw, 1978
Moksnes, 1983
Jensen, 1975
Ort and Latshaw, 1978
El-Begearmi etal., 1977
-------
rats exposed to 1.6 to 4.8 ppm selenium in their feed. A diet of 6.4 ppm selenium in feed in the
form of sodium selenite or seleniferous wheat caused significant growth depression, and death
occurred in the postweanling rats after 4 weeks at levels of 8.0 to 11.2 ppm in the diet. Earlier
studies had shown a toxic response in rats maintained on a diet containing 5 ppm selenium
(Moxon, 1937; Franke and Painter, 1938). Based on these and other studies, the threshold range
of harmful dietary selenium concentrations was estimated to be 2 to 5 ppm.
Fish. In the development of ambient water quality criteria for selenium, the U.S.
Environmental Protection Agency (1980) summarized a data base of 23 studies of 8 freshwater
species. The acute toxicity (96-hour LC50) values ranged from 0.62 to 28.5 mg selenium/L for the
bluegill (Lepomis macrochirus); 96-hour LC50 concentrations of 2.1 and 5.2 mg selenium/L were
determined for fathead minnow (Pimephales promeles) fry and juveniles, respectively.
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
• Several species and toxicity evaluations are used in estimating the toxic range of
selenium.
Limitations include:
•Effects of selenium are extrapolated from laboratory studies and are based on various
forms of selenium; these studies may not reflect the conditions at KR. There also may be
differences in bioavailabittty of selenium between the laboratory studies and the actual
field conditions.
*No consideration is given to the physical effects of the remedial actions on fish and
waterfowl habitat. (It had been determined to close KR and open an adjacent wetland
habitat.) The removal or capping of the water and surrounding habitat contaminated by
selenium would cause a major alteration of the habitat in which several endangered
species are able to reproduce. In addition, it is important to consider that the material
accumulates as succession occurs and recolonizfition by plants redistributes any buried
selenium back into the ecosystem.
9.3.3. Analysis: Characterization of Exposure
As described in the Kesterson Program Final EIS (USBR, 1986b), the potential existed for
residual soil/sediment selenium contamination to move into terrestrial and aquatic food chains.
Selenium in the environment may occur in numerous chemical forms due to the processes of
oxidation and reduction and biologically mediated transformations. Selenate is the most mobile
form of selenium and makes up the majority of selenium that had been delivered to KR via the
9-19
-------
SLD. The various forms of selenium were not distinguished in this risk assessment, however,
because the toxicity evaluation included all forms.
For this assessment, food chain exposure was considered the most important pathway for
exposure of fish and wildlife to selenium. Detailed food chain exposure diagrams for each of the
key species previously noted were developed into simplified selenium transfer models. These
models were used with calculated transfer factors derived from experimental data and a Monte
Carlo simulation technique to estimate the probability distribution of predictions of selenium
concentration in the diets of the key species under each of the three remedial alternatives.
Measured Selenium Concentrations. Mosquitofish captured at KR in May 1982 contained
high levels of selenium—about 135 ppm (Saiki, 1986). Because of this, FWS began intensive
studies at KR and at Volta Wildlife Area (a control area 10 km to the southwest of the site [figure
9-2] that is not contaminated by agricultural drainwater) to further define the effects and extent of
contamination resulting from drainwater delivery to KR. Of the six contaminants of concern that
SWRCB identified, FWS found that only selenium exhibited elevated concentrations in KR soils
compared with Volta (table 9-3). Boron, chromium, mercury, molybdenum, selenium, TDS, and
zinc were elevated in Kesterson waters. Examination of the biota at KR showed that only selenium
and boron were significantly higher in tissues versus biota from Volta (Ohlendorf et al., 1986b;
table 9-3); therefore, only boron and selenium were considered of potential concern for KR.
(Because of limited lexicological data, a boron risk assessment was not completed.) Food chain
organisms and fish were sampled in detail at KR and found to have significantly higher selenium
concentrations than at Volta (table 9-5).
Selenium concentrations in eggs of aquatic birds at KR were far higher than those reported
elsewhere in the United States (U.S. Department of the Interior, 1984; Ohlendorf et al., 1986a),
and selenium concentrations in the livers of aquatic birds sampled at KR significantly exceeded
selenium concentrations hi livers of birds from Volta (Ohlendorf et al., 1986b). A study of
potential selenium contamination of mammals at KR was conducted in 1984. Preliminary data for
the four most abundant species in this sample (California vole, harvest mouse, house mouse, and
ornate shrew) indicated that selenium levels at KR were 10 to 1,000 times higher than those at
Volta (USBR, 1986a).
Estimating Selenium Concentrations for the Remedial Alternatives. To estimate the
reduction in selenium risk associated with the three cleanup alternatives, changes in exposure were
modeled based on selenium movement through food chains to the key species. The terrestrial food
chain represented the dry areas around KR. It included selenium flux from drainwater deposition
through uptake into land plants, herbivores, and carnivores, ending with the kit fox. The aquatic
food chains represented the ponds and seasonally wet areas of KR and included benthic, aquatic
(water column), and rooted plant pathways. For the nonpiscivorous aquatic birds, selenium flux
was from sediments through the water column into plants and herbivorous insects. It ended with
mallards (aquatic plants and invertebrates), American coots (plants and epiphytal fauna), tricolored
blackbirds (aquatic insects), and black-necked stilts (benthic epifauna). A fish/piscivorous bird
pathway considered selenium flux from sediments through the water column herbivores into
mosquitofish and the fish-eating eared grebe.
9-20
-------
Table 9-5. Selenium Concentration (ppm, dry weight) in Composite Samples of Plants,
Invertebrates, and Mosquitofish, May 1983 (Ohlendorf et al., 1986b)
Sample
Filamentous algae
Rooted plants
Net plankton
Water boatmen
(Corixidae)
Midge larvae
(Chirononudae)
Dragonfly nymphs
(Anisoptera)
Damselfly nymphs
(Zygoptera)
Mosquitofish
(Gambusia qffinis)
Na
0/4
1/1
4/4
5/5
3/3
2/2
2/2
5/5
Volta
Meanb
NDC
0.43
2.03
1.91
2.09
1.29
1.45
1.29
Kesterson
(Range)
(1.4-2.9)
(1.1-2.5)
(1.5-3.0)
(1.2-1.4)
(1.2-1.7)
(1.2-1.4)
N
6/6
18/18
7/7
2/2
3/3
6/6
3/3
12/12
Mean
35.2
52.1
85.4
22.1
139
122
175
170
(Range)
(12-68)
(18-79)
(58-124)
(20-24)
(71-200)
(66-179)
(118-218)
(115-283)
a Number with measurable concentrations/number analyzed.
b Geometric means computed only when selenium was measurable in at least 50 percent of
samples. When only one sample was analyzed, the concentration is shown in this column.
c ND = not detected.
9-21
-------
Detailed food chain exposure diagrams for each of the seven key species were abstracted into
simplified selenium transfer models, as illustrated in figure 9-3 and table 9-6. The simplified
pathways contain all of the basic selenium transfer pathways present in the complex transfer
diagrams. The transfer of selenium through the food chain and concentrations of selenium in food
groups were estimated for each cleanup alternative using transfer factors derived from the studies
at Kesterson and elsewhere (tables 9-6, 9-7, and 9-8) and diet factors based on the dietary habits of
the receptor species. The empirical transfer and diet factors served as the basis for the
mathematical model used to predict the relationship between selenium in each trophic level and the
exposure of key species to selenium.
Transfer factors were based on empirical observations (table 9-9). For example, the transfer
factor between water and nonrooted plants at KR was calculated using paired observations of
selenium concentrations in water and nonrooted plants. This transfer factor reflects the relationship
between water and plant selenium concentrations at a water selenium concentration that was higher
than is expected to result from any of the cleanup alternatives. Because the uptake and metabolism
of selenium probably does not have a linear relationship with concentrations in water, a transfer
factor appropriate for the predicted range (2 to 15 ng selenium/L) was derived from literature
reviewed by Lemly (1985) and Ohlendorf et al. (1986b). The same procedure was followed for
derivation of other transfer factors in the aquatic food chain. Transfer factors derived from the
literature were generally about two to three times higher than those observed in high water
selenium concentrations at KR (USSR, 1986a).
Diet factors were used to model the fraction of the whole diet of each organism that was
contributed by each compartment contained in the simplified selenium transfer diagram. Diet
factors were based on species food preferences as described in the literature, their home range size
relative to the size of KR (see Ecological Components in section 9.3.1), and the relative abundance
of different types of prey species at KR. For species that consumed aquatic invertebrates, for
example, the relative abundance of herbivorous and carnivorous aquatic invertebrates measured at
KR was used to specify the composition of these organisms in the diet.
Because water applied to KR in each of the three remedial alternatives would have very low
selenium concentrations (i.e., < 1 jxg/L), the major potential source of selenium for biological
uptake was the sediments. The amount of selenium present in sediments in 1986 was quite variable
throughout KR, but it tended to be greater in southern ponds than in northern ponds. In the FRP
alternative, where no removal of sediment was involved, a value of 7 ± 7 mg selenium/kg
(standard deviation), representative of the southern ponds, was used (USBR, 1986a). The variance
component of this estimate reflects the spatial heterogeneity of selenium measured at KR. Plan 1
was estimated to leave an average sediment selenium concentration of 3 ± 2 mg/kg. Plan 2 would
leave a sediment selenium concentration of 1.5 + 1 mg/kg.
A Monte Carlo model simulated the selenium concentration in each compartment along the
exposure pathways from sediment to target organism by multiplying the selenium concentration in
the previous compartment by the appropriate transfer factor. Key species' average dietary
selenium concentration was calculated by weighting each component of the diet by the appropriate
diet factor. Initial conditions for the model were those selenium concentrations estimated in KR
sediments or surface water after implementation of the cleanup alternatives.
9-22
-------
vp
N>
UJ
JLJL
SURFACE
WATER
SUPPLY
DETRITUS/
MICROBES
8
DETRITIVORES
9
CARNIVORES
OFFS1TE
FOOD
SOURCES
Figure 9.3. Simplified selenium transfer diagram with key to transfer and diet factors (see table 9-6) (USBJR, 1986a)
-------
Table 9-6. Transfer and Diet Factors for Simplified Selenium Transfer Diagram for Mallard,
American Coot, Tricolored Blackbird, and Black-Necked Stilt (see figure 9-3) (USER,
1986a)a
Sediment Cone, (rag/kg d.w.)
Surface Water Supply (tng/L)
TRANSFER FACTORS
1 Sediment - Rooted Plants
2 Sediment- Water4
3 Water - Rooted Plants
4 Water - Nonrootcd Plants
5 Nonrooted Plants -
Herbivores
6 Herbivores - Carnivores
7 Sediment- Detritus/
Microbes
8 Detritus/Microbes -
Dctritivores
9 Dctritivorcs - Carnivores
RELATIVE SUPPLY FACTORS8
a Sediment - Rooted Plants
b Water - Rooted Plants
c Surface Water Supply - Water
d Sediment - Water
DIET FACTORS' FOR PAST
CONDITION, FLEXIBLE RESPONSE,
AND ONSI1E DISPOSAL
e Rooted Plants - Receptor
f Water -Receptor
g Nonrooted Plants - Receptor
h Herbivores - Receptor
i Carnivores (1) - Receptor
j Detritus/Microbes - Receptor
k Dctritivores - Receptor
1 Carnivores (2) - Receptor
m Oflsite Food Sources
Past
Condition
7(7)
0.3
2.8(3)
0.0003-0.002
81(21)
187 (22)
2.2 (0.8)
0.7 (0.3)
2.4 (2.9)
2.2 (0.8)
0.7 (0.3)
25
75
100
0
Adult Female
Mallard
Nesting
14(3)
5(2)
33(7)
41(6)
7(2)
0
0
0
0
Flexible
Response
7(7)°
0
2.8(3)
0.0003-0.002
81(21)
500 (50)
4.0 (2.0)
1.5(0.5)
2.4 (2.9)
2.2 (0.8)
0.7 (0.3)
25
75
0
100
Adult
American
Coot
30(5)
5(2)
18(3)
35(5)
7(2)
KD
2(1)
2(1)
0
Onsite
Disposal lb
3(2)
0
2.8 (3)
0.0003-0.002
81 (21)
500 (50)
4.0 (2-0)
1.5 (0.5)
2.4 (2.9)
2.2 (0.8)
0.7 (0.3)
25
75
0
100
Tricolored
Blackbird
Nestling
3(2)
0
2(1) '
79(8)
16 (5)
0
0
0
0
Onsite
Disposal 2
1.5(1)
0
2.8 (3)
0.0003-0.002
81 (21)
500 (50)
4.0 (2.0)
1.5(0.5)
2.4 (2.9)
2.2 (0.8)
0.7 (0.3)
25
75
0
100
Adult
Black-Necked
Stilt
0
5(2)
0
38(5)
7(3)
5(1)
38(5)
7(3)
0
' Standard deviations are in parentheses.
° For the seasonally wet areas. Also applicable to FRP seasonally wet areas in the northern ponds.
°. For ponds that will be wet all year (southern ponds).
" Uniform distribution; therefore, range is given.
• In cases where two routes of selenium supply exist, their ratio of supply is defined.
1 Percent of total diet from each compartment.
9-24
-------
Table 9-7. Transfer and Diet Factors for Simplified Selenium Transfer Diagram for Eared
Grebe and Mosquitofish (USER, 1986a)a
Sediment Cone, (mg/kg d.w.)
Surface Water Supply (mg/L)
Past
Condition
7(7)
0.3
Flexible
Response
7(7)"
0
Onsite
Disposal lb
3(2)
0
Onsite
Disposal 2
1.5 (1)
0
TRANSFER FACTORS ,
1 Sediment - Water4
2 Water - Nonrooted Plants
3 Nonrooted Plants - Herbivores
4 Herbivores - Carnivores
RELATIVE SUPPLY FACTORS6
a Surface Water Supply - Water
b Sediment - Water
DIET FACTORSf FOR PAST
CONDITIONS, FLEXIBLE RESPONSE,
AND ONSITE DISPOSAL
c Water - Receptor
d Carnivores (1) - Receptor
0.0003-0.002
187 (22)
2.2 (0.8)
0.7 (0.3)
100
0
0.0003-0.002
500 (50)
4.0 (2.0)
1.5 (0.5)
0
0
0.0003-0.002
500 (50)
4.0 (2.0)
1.5 (0.5)
0
0
0.0003-0.002
500 (50)
4.0(2.0)
1.5 (0.5)
0
0
Eared Grebe
10(5)
90(5)
Mosquitofish
8 Standard deviations are in parentheses.
b For the seasonally wet areas. Also applicable to FRP seasonally wet areas in the northern ponds.
0 For ponds that will be wet all year (southern ponds).
d Uniform distribution; therefore, range is given.
6 In cases where two routes of selenium supply exist, their ratio is defined.
f Percent of total diet from each compartment.
g NA = not applicable.
9-25
-------
Table 9-8. Transfer and Diet Factors for Simplified Selenium Transfer Diagram for San
Joaquin Kit Fox (USER, 1986a)a
Past Flexible Onsite Onsite
Condition Response Disposal lb Disposal 2
Sediment Cone, (mg/kg d.w.) 3 (2) 3 (2)c 3 (2) 1.5 (1)
Surface Water Supply (mg/L) 0.3 0 0 0
TRANSFER FACTORS
1 Soil - Terrestrial Plants 10(5) 10(5) 10(5) 10(5)
2 Terrestrial Plants -
Herbivores
3 Herbivores - Carnivores
DIET FACTORS'1
a Herbivores - Receptor
b Carnivores - Receptor
c Offsitc Pood Sources
0.3 (0.3)
4(2)
22.5 (20)
2.5(2)
75 (25)
0.3 (0.3)
4(2)
9(5)
1(1)
90 (10)
0.3 (0.3)
4(2)
9(5)
KD
90 (10)
0.3 (0.3)
4(2)
9(5)
1(1)
90 (10)
* Standard deviations are in parentheses.
b For the seasonally wet areas. Also applicable to FRP seasonally wet areas in the northern ponds.
c For ponds that will be wet all year (southern ponds).
d Percent of total diet from each compartment.
9-26
-------
Table 9-9. Summary of Data Used to Derive Transfer Factors8
Nonpiscivorous Bird Pathway
Past
Sediment
.Rooted Plants
Water
Nonrooted Plants
Noribenthic Herbivores
Nonbenthic Carnivores
Detritus
Detritivores
Benthic Carnivores
Value
7±7
26±18
300
56±5.5
127 ±23
92±19
26±18
56 ±5. 5
127±23
Data
Source
1
1
1
2,3
2,3,4
2,3,4
6
2,3,4
2,3,4
Flexible Response
Value
7±7
26±18
2-15
0.3-20°
d
2-270*
26±18
56±5.5
I27±23
Data
Source
1
1
2,3
5
5
5
6
2,3,4
2,3,4
Onsite Disposal
Value
3 ±2.
1.5±lb
26±18
2-15
0.3-20°
d
2-2706
26±18
56±5.5
127±23
Data
Source
1°
1
2,3
5
5
5
6
2,3,4
2,3,4
Fish/Piscivorous Bird Pathway
Sediment
Water
Nonrooted Plants
Nonbenthic Herbivores
Mosquitofish/Carnivores
Soil
Terrestrial Plants
Herbivores
Carnivores
Data Sources:
7±7
300
56±5.5
127 ±23
104±25
3 ±2
30±10
10±24
48±17
1
1
2,3
' 2,3,4
2,3
1
1
7
7
7±7
2-15
0.3-20e
d
2-2706
Terrestrial
3 ±2
30±10
10±24
48 ±17
1
2,3 '
5
5
5
Pathway
1
1
7
7
3 ±2,
1.5±lb
2-15
0.3-20e
d
2-270e
3 ±2
1.5±lb -
30±10
10±24
48 ±17
lc
2,3
5
5
5
1°
1
7
7
1 USBR, 1986a. Standard QA/QC procedures, all data sources given in EIS.
2 Presser and Barnes, 1984.
3 Saiki and Lowe, 1987.
4 Ohlendorfet al., 1986a. QA/QC procedures specified.
5 Lemly, 1985. QA/QC procedures not specified.
6 No specific reference. Assumed majority of detritus composed of rooted macrophytes.
7 Clark, personal communication.
Note: All standard deviations estimated by dividing range by 6 except those from Presser and Barnes, 1984, and Saiki and Lowe, 1987, which
were given in the references. This is based on the assumptions of a normal distribution and that > 99 percent of values are within ±3
standard deviation.
• AM units mg/kg (d.w.) except water, which is pg/L.
b Two values given for Onsite Disposal No. 1 and 2, respectively.
° Mean value from USBR, 1986a.
d Inferred from reference data.
c Range only was given in reference.
9-27
-------
Each estimated transfer factor and diet factor had an associated standard distribution, either
uniform or lognormal. The lognormal distribution was used to represent uncertainty in the transfer
and diet factors because it is a common distribution of selenium concentration data observed in
nature. In addition, the lognormal distribution has several statistically desirable properties, such as
estimating only positive values. It is a skewed distribution that produces rare large values more
often than does the normal distribution. The best estimate of the transfer and diet factors was
taken as the mean of the distribution, and uncertainty was expressed as a standard deviation or
range. Simulations were run for each cleanup alternative for each species. The model also was
run under past conditions with the application of drainwater to KR as a control.
|>
Estimates of Key Species Population Sizes. The estimates of population densities and
estimates of past KR-related mortalities (table 9-2) are based on data from published literature,
unpublished surveys by FWS and California DFG, and consultation with personnel from these
agencies and other local experts (USER, 1986a). These data were provided to put in perspective
the relative risks of selenium exposure of each population. The population data are indices of
density, and in most cases the actual values are unknown. The risks of contamination-induced
mortality vary greatly between these species. The data for mallards, American coot, and black-
necked stilt suggest that these species are at low risk because of their small population at KR
relative to San Joaquin Valley and statewide populations, even though both the American coot and
black-necked stilt suffered significant losses at KR. In contrast, the tricolored blackbird population
at KR represented more than one-half of the San Joaquin Valley population and one-third of the
statewide population. Tricolored blackbirds are endemic to California; thus, the statewide
population approximately represents the global population. Eared grebe and kit fox populations
were not known at the site but were estimated by observation to be fewer than 20 each.
Comments on Analysts: Characterization of Exposure
Strengths of the case study include:
*As detailed above, extensive data exist to demonstrate that the organisms were
exposed to selenium. The Monte Carlo analysis is an important contribution to the
analysis of exposure. The distributions are delineated in the supporting
documentation.
•The food chain model is well documented and maximizes use of site-specific
information on selenium concentrations in various food chain compartments.
Limitations include:
*The assessments of exposure are dependent on the transfer factors in the elaborate
food chain models. Transfer factors determining the bioaccumulation of selenium
are taken as constants. However, the uncertainty in the variability of these
9-28
-------
Comments on Analysis: Characterization of Exposure (continued)
transfer factors is high. An attempt to calibrate the prediction of the models using
measured selenium concentration in birds would have reduced the uncertainty
associated with the simplified model assumptions.
•There is little technical basis for the selection of the statistical distributions used in
the Monte Carlo analysis.
9.3.4. Risk Characterization
This study contained two components of risk characterization: (1) attribution of adverse
effects observed in KR to selenium and (2) prediction of the effectiveness of three remedial
alternatives in reducing food chain exposures to selenium to below harmful levels for key species.
Attribution. Selenium was determined to be the most likely cause of the adverse
reproductive effects in waterfowl at KR for two reasons. First, the effects observed at KR were
similar to those produced in several avian species in the laboratory by administering excess
selenium in the diet, whereas the other five contaminants of concern generally produce other types
of adverse effects or their effects are unknown. Second, measurements of drainwater constituents in
ground water, soils, and biota indicated that only selenium and boron were elevated in KR samples
relative to control areas. Boron was excluded from the remainder of the assessment, however,
because of limited lexicological information.
Risks Associated with Remedial Alternatives. The Monte Carlo simulation of the
selenium transfer models estimated the probability distribution of selenium concentrations in the
diets of the key species under each of the remedial alternatives (figures 9-4, 9-5 and 9-6). As an
example, the results of the simulation for the FRP are presented in figure 9-4. For any combination
of cleanup alternative and key species, the 50-percent probability level represents the dietary
selenium concentration predicted from the mean transfer and diet factors for that particular
condition. The uncertainty of the exposure estimate is shown by the probability distribution about
the mean.
Using the mallard as an example, the 50-percent probability level represents a dietary
selenium concentration of about 5 ppm (5 mg/kg in figure 9-4). Therefore, based on the selenium
transfer model and on the uncertainty of transfer and diet factors, the FRP has about a 50-percent
chance of resulting in a mallard dietary selenium concentration of less man 5 ppm.
Estimated selenium dietary concentrations were compared with threshold selenium toxicities
for the key species. Tables 9-10a and 9-10b summarize the results for each cleanup alternative as
the percentage of diet selenium predictions that are below estimated harmful levels for key species
(USER, 1986a). These results indicated that predicted risks are greatest for the FRP, less for
9-29
-------
0£-6
-------
I £-6
-------
Z£-6
-------
Table 9-10a. Percentage of Diet Selenium Predictions That Are Below the 5 mg/kg Estimated
Harmful Level for Each Key Species and Each Cleanup Alternative (USBR, 1986a)
Key Species
Birds"
Mallards
Coots
Black-Necked Stilts
Tricolored Blackbirds
Eared Grebes
Mammals0
San Joaquin Kit Fox
Fishc'd
Mosquitofish
FRP
5
50
50
25
35
35
2
90
3
20
Cleanup Alternative
Onsite-l*
5
75
80
50
60
60
2
95
3
35
Onsite-2b
5
95
95
80
85
80
2
-100
3
70
Table 9-10b. Percentage of Diet Selenium Predictions That Are Below the 10 mg/kg Estimated
Harmful Level for Each Key Species and Each Cleanup Alternative (USBR, 1986a)
Cleanup Alternative
Birds0
Mallards
Coots
Black-
Necked Stilts
Tricolored
Blackbirds
Eared
Grebes
Mammals0
San Joaquin
Kit Fox
Fish0'"
Mosquitofish
FRP
10
80
75
50
65
60
5
-100
5
35
Onsite-r
10
95
95
80
90
85
5
-100
5
50
Onsite-2"
10
100
100
95
95
90
5
-100
5
90
*450,000 cubic yards.
" 1,000,000 cubic yards.
c Diet selenium concentration (mg/kg).
d Mosquitofish not harmed by these selenium levels, although other species may be.
9-33
-------
Plan 1, and least for the Plan 2 remedial alternative. The results suggest that while none of the
alternatives would clearly fail, none would clearly succeed in removing selenium risk to the key
populations.
Uncertainties. The exposure estimates for each alternative are based on assumptions
regarding the steady-state relationship between selenium concentration in sediments and the
resulting concentration in surface water. Although this relationship is based on existing knowledge
of selenium chemistry and field and laboratory experiments, the model results do not take into
account the length of time necessary to achieve steady-state conditions.
The model does not describe uptake and loss rates of selenium by the components of the
exposure pathways; rather, the model used transfer and diet factors observed in the laboratory and
field. The uncertainty estimates of these factors are based on these observations, but they do not
necessarily simulate exact conditions at KR.
Because insufficient information exists to develop quantitative dose-response relationships for
diet selenium exposures specific for the key species at KR, the model results cannot be used to
make quantitative estimates of the impact of cleanup alternatives to the exposed population. The
toxicity profiles, however, can be used with the model results to determine the uncertainty of the
relative safety of the cleanup alternatives.
Considerations for Other KR Organisms. The impact of each cleanup alternative on each
species can be considered in terms of the fraction of the total population that resides at KR.
Although only seven species were selected for selenium exposure evaluation, predicted impacts also
may apply to other species that were not directly considered in the analysis because the trophic
levels of the seven species are representative of those of a large number of species at KR. The
transfer and diet factors (and associated uncertainties) and, thus, the selenium exposure estimates
have broad applicability. Applicability for organisms at lower trophic levels is limited.
Habitat Changes. The model does not address the potential indirect effect of implementation
of each alternative on wildlife populations as a result of physical changes in habitat. Each
alternative would affect the habitat of KR to a variable degree. Furthermore, diet factors may
change in the case of opportunistic organisms in response to changes in food availability brought
about by implementation of a particular alternative.
Comments on Risk Characterization
Strengths of the case study include:
•Given the constraints of the assumptions, the risk characterization is applicable.
Certainly a strong point of this case study is the derivation of the probabilities as
expressed in the figures.
9-34
-------
Comments on Risk Characterization (continued)
Limitations include:
• Because of the narrow focus on selenium and terrestrial vertebrates in the KR study,
some important factors may have been missed. Boron is not considered as a toxicant
because of lack of data. The influence of habit alteration as a factor also is not
property addressed. This last factor is crucial because each of the cleanup alternatives
involved major changes in habitat quantity and quality. To the extent that some habitat
changes were known (e.g., drainage or covering), it may have been possible to model
this into the risk assessment to provide a quantitative analysis along with uncertainty
estimates.
General comment:
•/« accordance with the Kesterson Program EIS (USBR, 1986), the Bureau
implemented a phased approach to KR cleanup. During the course of the continuing
research program that was conducted concurrent with the action, it became clear that
an alternative remedial plan (which had not been considered in the EIS) would be most
appropriate. Because that plan had not been identified before the risk assessment, it
was not evaluated. In the end, the risk assessment indicated the inadequacy of
alternatives and directed the responsible parties to consider site mitigation as the
alternative of choice. KR was closed and convened to a terrestrial habitat. A much
larger wetland was created to replace that lost at KR.
9-35
-------
9.4. REFERENCES
Arnold, R.L.; Olson, O.G.; Carlson, C.W. (1973) Dietary selenium and arsenic additions and
their effects on tissue and egg selenium. Poult. Sci. 52:847-854.
El-Begeanni, M.M.; Sunde, M.L.; Gauther, H.E. (1977) A mutual protective effect of
mercury and selenium in Japanese quail. Poult. Sci. 56:313-322.
Franke, K.W.; Painter, E.P. (1938) A study of the toxicity and selenium content of seleniferous
diets: with statistical consideration. Cereal Chem. 15:1-24.
Gilbertson, M.; Morris, R.D.; Hunter, R.A. (1976) Abnormal chicks and PCB residue levels in
eggs of colonial birds on the lower Great Lakes (1971-1973). Auk 93:434-442.
Halverson, A.W.; Palmer, I.S.; Guss, P.L. (1966) Toxicity of selenium to post-weanling rats.
Toxicol. Appl. Pharmacol. 9:477-484.
Heinz, G.H.; Hoffman, D.J.; Krynitsky, A.J.; Weller, D.G. (1987) Reproduction of mallards
fed selenium. Environ. Toxicol. Chem. 6:423-433.
Hill, E.F.; Hoffman, D.J. (1984) Avian models for toxicity testing. J. Am. Coll. Toxicol.
3:356-357.
Hoffman, D.J. (1978) Embryotoxic effects of crude oil in mallard ducks and chicks. Toxicol.
Appl Pharmacol. 46:183-190.
Hothem, R.L.; Ohlendorf, H.M. (1989) Contaminants in foods of aquatic birds at Kesterson
Reservoir, California, 1985. Arch. Environ. Co/item. Toxicol. 18:773-786.
Jensen, L.S. (1975) Modification of a selenium toxicity in chicks by dietary silver and
copper. /. Nutr. 105:769-767.
Lemly, A.D. (1985) Ecological basis for regulating aquatic emissions from the power industry:
the case with selenium. Reg. Toxicol. Pharmacol. 5:465-486.
Moksnes, K. (1983) Selenium deposition in tissues and eggs of laying hens given surplus of
selenium as selenomethionine. Acta Vet. Scand. 24:34-44.
Morrell, S.H. (1975) San Joaquin Mtfox distribution and abundance in 1975. California
Department of Fish and Game. Administrative Report #75-3.
Moxon, A.L. (1937) Alkali disease or selenium poisoning. S.D. Agric. Exp. Stn. Bull. 311.
Moxon, A.L.; Rhian, M. (1943) Selenium poisoning. Physiol. Rev. 23:305-337.
National Research Council. (1976) Selenium. Washington, DC: National Academy Press.
9-36
-------
National Research Council. (1980) Mineral tolerance of domestic animals. Subcommittee on
Mineral Toxicity in Animals, Committee on Animal Nutrition. Washington, DC: National
Academy Press.
Ohlendorf, H.M.; Hothem, R.L.; Bunck, C.M.; Aldrich, T.W.; Moore, J.F. (1986a)
Relationships between selenium concentrations and avian reproduction. Trans. N. Am.
Wildl. Natl Resour. Conf. 51:330-342.
Ohlendorf, H.M.; Hoffman, D.J.; Saiki, M.K.; Aldrich, T.W. (1986b) Embryonic mortality and
abnormalities of aquatic birds: apparent impacts of selenium from irrigation drainwater. Sci.
Total Environ. 52:49-63.
Ohlendorf, H.M.; Kilness, A.W.; Simmons, J.L.; Stroud, R.K.; Hoffman, D.J.; Moore, J.F.
(1988) Selenium toxicosis in wild aquatic birds. /. Toxicol. Environ. Health 24:67-92,
Ohlendorf, H.M.; Hothem, R.L.; Welsh, D. (1989) Nest success, cause-specific nest failure,
and hatchability of aquatic birds at selenium-contaminated Kesterson Reservoir and a
reference site. Condor 91:787-796.
Ort, J.F.; Latshaw, J.D. (1978) The toxic level of sodium selenite hi the diet of laying
chickens. /. Nutr. 108:1114-1120.
Poley, W.E.; Moxon, A.L.; Franke, K.W. (1937) Further studies of the effects of selenium
poisoning on hatchability. Poult. Sci. 16:219-225.
Poley, W.E.; Moxon, A.L. (1938) Tolerance levels of seleniferous grains in laying rations.
Poult. Sci. 17:72-76.
Pomeroy, D.E. (1962) Birds with abnormal bills. Breeding Birds 55:48-72.
Presser, T.S.; Barnes, I. (1984) Selenium concentrations in waters tributary to and in the vicinity
of the Kesterson National Wildlife Refuge, Fresno and Merced Counties, California. USGS.
Water Resources Investigation Report 84-4122.
Rosenfeld, J.; Beath, O.A. (1964) Selenium: geo-botany, biochemistry, toxicity, and nutrition.
New York, NY: Academic Press.
Saiki, M.K. (1986) A field example of selenium contamination in an aquatic food chain.
In: First Annual Environmental Symposium: Selenium in the Environment—proceedings.
California State University, Fresno, June 10-12, 1985. CATI-860201. Fresno, CA:
California Agricultural Technology Institute, pp. 67-76.
Saiki, M.K.; Lowe, T.P. (1987) Selenium hi aquatic organisms from subsurface agricultural
drainage water, San Joaquin Valley, California. Arch. Environ. Contam. Toxicol.
16:657-670.
9-37
-------
Shamberger, R.J. (1981) Selenium in the environment. Sci. Total Environ. 17:59-74.
Shamberger, R.J. (1983) Biochemistry of selenium. New York, NY: Plenum Press.
Skorupa, J.P.; Ohlendorf, H.M. (1991) Contaminants in drainage water and avian risk
thresholds. In: Dinar, A.; Iberman, D.Z., eds. The economics and management of water and
drainage in agriculture. Norwell, MA: Kluwer Academic Pub.
State Water Resources Control Board. (1986) Regulation of agricultural drainage to the San
Joaquin River. California SWRCB Order No. WQ85-1, Tech. Comm. Staff, SWRCB
and CVWQCB.
U.S. Bureau of Reclamation. (1986a) Risk Assessment Kesterson Program. Prepared for
USBR, Mid-Pacific Region, by CH2M Hill, Sacramento, CA.
U.S. Bureau of Reclamation. (1986b) Kesterson Program Final Environmental Impact
Statements, volumes I and II: USBR, Mid-Pacific Region, in cooperation with U.S. Fish and
Wildlife Service and U.S. Army Corps of Engineers, Sacramento, CA.
U.S. Department of Interior. (1984) Conference on toxicity problems at Kesterson Reservoir.
December 5-7, 1983. Sacramento, CA: U.S. Department of Interior.
U.S. Environmental Protection Agency. (1980) Ambient water quality criteria for selenium.
Office of Water Regulations and Standards, Washington, DC. EPA 440/5-80/070.
Wilber, C.G. (1980) Toxicology of selenium: a review. Clin. Toxicol. 17:171-230.
Winkler, D.W.; Weigen, C.P.; Engstrom, F.B.; Birch, S.E. (1977) Ornithology. In: Winkler,
D.W., ed. An ecological study of Mono Lake, California. Davis, CA: University of
California, Institute of Ecology, publication #12.
9-38
-------
SECTION TEN
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
RISK CHARACTERIZATION METHODS USED IN DETERMINING THE
EFFECTS OF SYNTHETIC PYRETHROIDS ON TERRESTRIAL
AND AQUATIC ORGANISMS
-------
AUTHOR AND REVIEWERS
AUTHOR
Greg R. Susanke
Office of Pesticide Programs
U.S. Environmental Protection Agency
Washington, DC
REVIEWERS
Arthur L. Buikema (Lead Reviewer) Clifford Hupp
Biology Department Water Resources Division
Virginia Polytechnic Institute and State U.S. Geological Survey
University Reston, VA
Blacksburg, VA
Ronald J. Kendall
Joel Brown Institute of Wildlife and
Department of Biological Sciences Environmental Toxicology
University of Illinois Clemson University
Chicago, IL Clemson, SC
Carolyn L. Fordham
Terra Technologies
Golden, CO
Christian E. Grue
Washington Cooperative Fish
and Wildlife Research Unit
U.S. Fish and Wildlife Service
Seattle, WA
10-2
-------
CONTENTS
ABSTRACT '. 10-6
10.1. RISK ASSESSMENT APPROACH 10-7
10.2. STATUTORY AND REGULATORY BACKGROUND 10-7
10.3. CASE STUDY DESCRIPTION 10-9
10.3.1. Problem Formulation 10-9
10.3.2. Analysis: Characterization of Ecological Effects 10-11
10.3.3. Analysis: Characterization of Exposure 10-20
10.3.4. Risk Characterization 10-25
10.4. REFERENCES 10-30
APPENDIX A—TIER TESTING SYSTEM FOR ECOLOGICAL EFFECTS 10-A1
10-3
-------
LIST OF FIGURES
Figure 10-1. Structure of analysis for evaluating risks associated with pyrethroids 10-8
LIST OF TABLES
Table 10-1. Toxicity of Synthetic Pyrethroids to Freshwater Organisms 10-13
Table 10-2. Toxicity of Synthetic Pyrethroids to Marine and Estuarine Organisms .... 10-14
Table 10-3. Toxicity of Synthetic Pyrethroids to Birds and Mammals 10-15
Table 10-4. Residues Found on Various Substrates After Application 10-21
Table 10-5. EECs or Actual Measured Residues Exceeding Toxicity Concentrations . . . 10-26
10-4
-------
LIST OF ACRONYMS
ASTM American Society for Testing and Materials
BCF bioconcentration factor
EEB Ecological Effects Branch
EEC estimated environmental concentration
EPA U.S. Environmental Protection Agency
EXAMS Exposure Analysis Modeling System
FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
LC50 lethal concentration to 50 percent of organisms tested
LD50 lethal dose to 50 percent of organisms tested
LOEL lowest observed effect level
MATC maximum acceptable toxic concentration
NOEC no observed effect concentration
NOEL no observed effect level
OPP Office of Pesticide Programs
PRZM Pesticide Root Zone Model
SWRRB Simulator for Water Resources in Rural Basins
10-5
-------
ABSTRACT
The process by which the U.S. Environmental Protection Agency's Office of Pesticide
Programs conducts ecological risk assessments for the registration and reregistration of pesticide
products is presented in this case study, which focuses on a group of insecticides known as
synthetic" pyrethroids. The ecological risks associated with their application to cotton and
sunflowers are assessed according to a combination of the quotient method (toxicity and exposure)
and the weight of evidence (field studies). Although the case study considers ecological
components in aquatic and terrestrial habitats, emphasis is placed on aquatic biota. The assessment
endpoint is the health and survival of biological components or key species. Measurement
endpoints include the following: (1) acute and chronic effects on aquatic invertebrates and fish,
(2) system effects that adversely affect fish and aquatic invertebrate populations, and (3) potential
adverse effects on breeding waterfowl and hatchlings due to reduction in the food base.
10-6
-------
10.1. RISK ASSESSMENT APPROACH
The Ecological Effects Branch (EEB) of the Office of Pesticide Programs (OPP) has an
established procedure by which it conducts risk assessments for the registration and reregistration
of pesticides. The various components of these assessments are discussed in this case study. The
order in which they are presented has been changed to conform to the case study format (figure
10-1). Risk assessments performed by EEB are usually pesticide specific, but for the purpose of
this case study, a group of pesticides known as synthetic pyrethroids has been chosen.
To evaluate the risks associated with synthetic pyrethroid use, EEB has characterized the
ecological effects and exposure for these chemicals. Ecological effects were characterized using
acute and chronic laboratory toxicity testing and field studies. Exposure was evaluated based on
estimates of residues in the terrestrial and aquatic environment likely to affect ecological
components. Field studies were used to verify whether effects indicated by toxicity testing and
exposure estimates did, in fact, manifest following applications of the chemical. Laboratory and
field data were used with the exposure estimates to characterize the risk.
A strength 'C'f this case study is the broad and fairly complete toxicity data base that is
obtained when data from the different generations of synthetic pyrethroids are combined. Exposure
has been characterized by using measurements, estimates based on assumptions for fate and
transport, and computer models. By focusing on synthetic pyrethroids as a group, this case study
illustrates various methods that have been used to estimate exposure.
In addition to the laboratory studies of effects, the assessment of ecological risks due to
pyrethroid application currently relies on the results of only two field studies. Other field studies
have been required and are in progress or have been submitted and are under review.
Pyrethroid applications to cotton and sunflowers are used in the case study to illustrate the
method by which risks have been assessed. Applications to cotton fields are discussed in some
detail because they are the major site where synthetic pyrethroids are applied. Sunflowers are
discussed because of the potential adverse effect on waterfowl in the prairie pothole region, the
major sunflower production area.
10.2. STATUTORY AND REGULATORY BACKGROUND
A pesticide product may be sold or distributed in the United States only if it is registered or
exempt from registration under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA),
as amended (1988, 7 U.S.C. 136 et seq.). Before a product can be registered unconditionally, it
must be shown that it can be used without "unreasonable adverse effects on the environment"
(FIFRA section 3[c][5]); that is, without causing "any unreasonable risk to man or the
environment, taking into account the economic, social and environmental costs, and benefits of the
use of the pesticide" (FIFRA section 2[bi>]). The burden of proving that a pesticide meets this
standard for registration is at all times on the proponent of initial or continued registration. If at
any time the U.S. Environmental Protection Agency (EPA) determines that a pesticide no longer
meets this standard for registration, the Administrator may initiate proceedings to cancel the
registration under FIFRA section 6.
10-7
-------
Figure 10-1. Structure of Analysis for
Evaluating Risks Associated with Pyrethroids
PROBLEM FORMULATION
Stressors; several generations of pyrethroid pesticides.
Ecological Components: aquatic invertebrates, fish,
terrestrial insects (bees), and waterfowl.
Endpoints: assessment endpoint is health of aquatic
organisms and terrestrial wildlife. Measurement
endpoints include acute and chronic toxicity tests and
aquatic field studies.
f
ANALYSIS
Characterization of
Exposure
Models and simple
algorithms were used to
estimate a range of
exposure concentrations.
Measurements were made
of residues.
Characterization of
Ecological Effects
Laboratory, mesocosm, and
a few field studies were
used to examine pyrethroid
toxicity.
RISK CHARACTERIZATION
Estimated exposure concentrations were compared
to Ecological Effects Concentrations (EECs) using the
quotient method. Field and mecocosm studies were
employed as part of an overall "weight-of-evidence"
approach.
10-8
-------
The special review process provides a mechanism through which the Agency gathers
information about pesticides that appear to pose risks of adverse effects to human health or the
environment. Risk evidence submitted to and/or gathered by the Agency must be evaluated and
considered in light of benefit information. If the Agency determines that risks appear to outweigh
the benefits, the Agency can initiate action under FIFRA to cancel, suspend, and/or require
modification of the terms and conditions of registration.
10.3. CASE STUDY DESCRIPTION
10.3.1. Problem Formulation
Stressor. Synthetic pyrethroids may be applied to a wide variety of crops: alfalfa, barley,
citrus, corn, cotton, fruit trees, nuts, oats, peanuts, potatoes, small fruits, soybeans, sugar beets,
sugarcane, sunflowers, tobacco, vegetables, and wheat. Pyrethroids are also applied to anthills,
lawns, livestock yards, ornamentals, pine, conifers, rangeland, turf, and for mosquito control.
These uses for pyrethroids are conditionally registered pending the outcome of data requirements.
Pesticides that have been conditionally registered (FIFRA section 3 [c][7]) are allowed to be
applied throughout the duration of the registration process. Basically, the registration decision
weighs both the economic consequences of denying registration and the environmental effects of
granting registration. At this time, only permethrin has been unconditionally registered for
agricultural crop uses (1982). Data requirements needed to support registration are based on a
tiered testing system (appendix A).1
Since their first review by EEB in the 1970s, synthetic pyrethroids have caused a great
concern for potential adverse effects on the aquatic environment, including a reduction in numbers
of micro- and macroinvertebrates and fish. The available data on synthetic pyrethroids show that
these compounds, as a class of insecticides, all appear to be highly toxic to aquatic organisms, both
acutely and chronically. They also have a high potential to persist in water and sediment and to
bioaccumulate in aquatic organisms. EEB believes that all synthetic pyrethroids are likely to
exhibit these properties. Consequently, EEB has proposed a specific set of aquatic organism
studies that will be required to support the registration of all synthetic pyrethroids. Based on
EEB's experience in reviewing synthetic pyrethroids, many of the data that were specified as
reserved (upper tier tests) are now required. These include: (1) a freshwater aquatic invertebrate
life-cycle study, (2) an estuarine/marine invertebrate life-cycle study, (3) a fish full-life-cycle
study, (4) an aquatic organism accumulation study, and (5) simulated (mesocosm) or actual (pond)
aquatic field studies.
lrThe manufacturer (registrant) of a specific pesticide is responsible for submitting the data necessary
to support registration. Data requirements in the first three tiers of the testing protocol are
bioassays. Initially, acute toxicity tests are conducted. Depending on the acute toxicity values,
subchronic and chronic toxicity tests may be required (see Preston and Hitch, 1982, for a
determination of when chronic testing is required). The last tier is the requirement for simulated or
actual field testing. Field testing is required on a chemical-by-chemical basis, as determined by the
available information on potential effects and exposure. The field study is used to verify if effects
indicated in the bioassays manifest in the field applications.
10-9
-------
EEB is also concerned with the adverse effects of synthetic pyrethroids on waterfowl (U.S.
EPA, 1987b). It has concluded that a reduction in the diversity of species and number of
organisms of an aquatic habitat may affect waterfowl recruitment and hatchling survival because of
the reduced prey base.
Ecological Components. This case study focuses on aquatic invertebrates and fish at the
individual, population, and community levels; waterfowl; and a few target species (e.g., bees) of
terrestrial habitats.
Endpoint Selection. The assessment endpoint is the health of aquatic organisms and
terrestrial wildlife. Measurement endpoints include laboratory acute and chronic toxicity tests with
selected species (aquatic fish and invertebrates, nontarget insects, and waterfowl) and aquatic field
tests (mesocosm and pond studies).
Comments on Problem Formulation
Strengths of the case study include:
• The assessment outlines a framework of sequential testing within which field data
are used to supplement initial toxicity data in order to verify potential effects. This
is especially important for identifying ecological effects that could not have been
predicted from laboratory studies.
• The assessment considers potential indirect effects as well as the direct toxic effects
of the chemicals. It notes the importance of considering food chain effects as they
relate to biomagniflcation and potential secondary effects such as the reduction in
the food base for higher trophic levels.
Limitations include:
• The potential effects on waterfowl due to a reduced prey base were considered to
be an important issue, but it was recognized that there are few data available for
evaluating these effects. This observation indicates the lands of information that may
need to be developed in the future. The endpoint should be effects on waterfowl
productivity rather than direct toxicity.
• Original laboratory toxicity test and field data were not provided for peer review.
10-10
-------
10.3.2. Analysis: Characterization of Ecological Effects
Overview of Test Methods. Several aquatic and avian species commonly used in toxicity
testing have been chosen by EEB to represent all aquatic organisms and terrestrial wildlife in the
United States. It is understood that these species are not necessarily the most sensitive to all
pesticides tested, but they are generally recognized as good indicator species. Those species
chosen by EEB have met selection criteria (Urban and Cook, 1986) and are consistent with the
recommendations of various sources such as the American Society for Testing and Materials
(ASTM) standards (1980, E 729-80, E 1022-84), the Committee on Methods for Toxicity Tests
with Aquatic Organisms (1975), and the National Water Quality Laboratory Committee on Aquatic
Bioassays (1971). Different indicator species are used depending on the bioassay performed.
Typical species used as indicators for each test type (tiers 1 through 3) are listed in the tables
presented in the Analysis: Characterization of Exposure section.
In mesocosm testing, the indicator fish species used is the bluegill. No predetermined
aquatic invertebrate species are used as indicators. The species that are used are those found in
similar aquatic habitats near the study site. Actual pond or terrestrial field studies investigate the
effects to naturally occurring species of fish, aquatic invertebrates, birds, and small mammals.
Current policy is that toxicity testing not be performed on amphibians and reptiles. It is
assumed that they are protected when fish, mammals, and birds are protected. Mammalian toxicity
testing is performed for the human risk characterization process. The results are also used by
EEB. The species used to represent mammalian wildlife are the laboratory rat, mouse, rabbit, and
dog. Nontarget insect toxicity data are required for outdoor uses that may result in honey bee
exposure. Cotton and sunflowers are such uses. Honey bees were selected to represent nontarget
insects because of their economic importance.
Laboratory Testing. Numerous bioassays have been performed on synthetic pyrethroids by
various sources under various testing protocols. However, for the registration of pesticides the
bioassays must follow the protocols described in the Pesticide Assessment Guidelines: Subdivision
E Hazard Evaluation—Wildlife and Aquatic Organisms (Preston and Hitch, 1982) and Subdivision
L Hazard Evaluation—Nontarget Insects (Hitch, 1982). These protocols are consistent with
standards recommended by ASTM (1980, E 729-80, E 1022-84), the Committee on Methods for
Toxicity Tests with Aquatic Organisms (1975), and the National Water Quality Laboratory
Committee on Aquatic Bioassays (1971). The aquatic and avian toxicity values presented in the
following tables were obtained from bioassays that were conducted according to these protocols and
thereby satisfied data requirements.
The toxicity values are presented as the LC50, LD50, and MATC. The LC50 is the median
lethal concentration or the concentration at which 50 percent of the test organisms die. The LD50
is the median single lethal dose or the dose at which 50 percent of the test organisms die. The
MATC is the maximum acceptable toxic concentration or the range of concentrations between the
no observed effect concentration (NOEC) and the lowest observed effect concentration (LOEL).
The MATC for chronic studies is a function of survival, reproduction, and growth. Toxicity
values are expressed as milligrams per liter (mg/L), micrograms per liter (/t/L), or nanograms per
10-11
-------
liter (ng/L) for pyrethroid concentrations in water and as milligrams per kilogram (mg/kg) or
micrograms per kilogram (/tg/kg) for concentrations in nonaqueous media.
Synthetic pyrethroids are highly toxic to all freshwater organisms. Tables 10-1, 10-2, and
10-3 incorporate acute and chronic toxicological data from the various synthetic pyrethroids. The
toxicity values presented show the approximate ranges found within each study type. These values
come from EBB and HED (Health Effects Division; mammalian data) reviews of studies submitted
to GPP. Both the studies themselves and the reviews are unpublished and generally unavailable to
the general public except via the Freedom of Information Act. The typical indicator test species
are listed in parentheses.
Synthetic pyrethroids also appear to be very highly toxic to marine and estuarine
organisms. The ranges of acute and chronic toxicity values among the various generations of
synthetic pyrethroids tested to date are presented in table 10-2 for each study type. The typical
indicator test species are found in parentheses.
Synthetic pyrethroids are slightly toxic to practically nontoxic to birds and are moderately
toxic to practically nontoxic to mammals. The ranges of LD50 and LC50 toxicity values among the
various generations of synthetic pyrethroids tested to date are presented in table 10-3 for each study
type. The typical indicator test species are listed.
There is wide variation in avian reproductive effects among synthetic pyrethroids. The
avian NOELs are as high as 900 to 1,000 ppm for fluvalinate (study I.D. 073443, 1985) and
cyfluthrin (study I.D. 254820, 1985), respectively. But fenvalerate (study I.D. 96385, 1978)
increased bobwhite quail eggshell cracking at < 25 ppm, and cyhalothrin (study I.D. 073989,
1988) affects the number of viable embryos, hatchlings, and 14-day survivors of eggs incubated
(set) at 50 ppm (LOEL). i
Synthetic pyrethroids are generally highly toxic to honey bees on an acute contact LD50
basis. Restrictive labeling is required.
Field Testing. EEB has determined that aquatic field studies are required to complete the
characterization of ecological effects for synthetic pyrethroids. These tests are required to verify
results of laboratory bioassays. If no significant effects are observed in acceptable field studies at
the estimated environmental concentration (EEC), then it will be concluded that there are no
aquatic risks at this exposure level.
At the time this case study was prepared, one simulated (mesocosm) and one actual (pond)
field study were found to be scientifically acceptable. These were the studies for lambda
cyhalothrin and bifenthrin, respectively. The remainder of the field studies for other synthetic
pyrethroids are either pending or currently under review. The results discussed in this section
come from EEB reviews of field studies submitted to OPP. Both the studies themselves and the
reviews are unpublished and generally unavailable to the general public except via the Freedom of
Information Act.
10-12
-------
Table 10-1. Toxicity of Synthetic Pyrethroids to Freshwater Organisms
Study
Acute
Bluegill
Rainbow trout
Invertebrate
(Daphnia magna)
K- Chronic
9
u> Fish early life stage
(fathead minnow)
Fish life cycle
(fathead minnow)
Invertebrate life cycle
(Daphnia magna)
Test
LC50
LC50
LCso
LC50
LCso
LC50
NOEL
LOEL
NOEL
LOEL
NOEL
LOEL
Toxicity
0.13 ppb
2.8 ppb
0.06 ppb
9.8 ppb
0.038 ppb
8.3 ppb
>64pptr
<410pptr
>31pptr
<95pptr
> 1.3 pptr
<45pptr
Chemical
Tefluthrin
Tralomethrin
Tefluthrin
Permethrin
Tralomethrin
Flucythrinate
Fluvalinate
Permethrhi
Lambda Cyhalothrin
Bifenthrin
Bifenthrhi
Fluvalinate
EPA I.D.
Number
261402
072123
261402
096699
UCCES
11507-74-03
098288
250141
096699
41519001
40791301
41156501
250141
Year of EEB
Review
1986
1984
1986
1978
1981
1979
1983
1978
1990
1989
1989
1983
-------
Table 10-2. Toxicity of Synthetic Pyrethroids to Marine and Estuarine Organisms
Study
Acute
Fish (sheepshead minnow)
Mollusc (eastern oyster)
Crustacean (mysid shrimp)
Chronic
Invertebrate life cycle
(mysid shrimp)
Fish early life stage
(sheepshead minnow)
Test
LC50
LC50
LC50
NOEL
LOEL
NOEL
LOEL
Toxicity
0.13 ppb
17.5 ppb
3.42 ppb
>1 ppm
3.97 pptr
840 pptr
> 0.17 pptr
< 0.93 pptr
> 24 pptr
< 620 pptr
Chemical
Tefluthrin
Bifenthrin
Cyfluthrin
Tefluthrin
Bifenthrin
Tralomethrin
Cyfluthrin
Tralomethrin
Cyfluthrin
Cyfluthrin
EPA I.D.
Number
40161137
264646
262443
40161135
264647
070692
262443
264510
265895
265895
YearofEEB
Review
1987
1987
1987
1987
1987
1982
1988
1987
1987
1987
-------
Table 10-3. Toxicity of Synthetic Pyrethroids to Birds and Mammals
Study
Rat
Mallard duck
Northern bobwhite quail
Mallard duck
Northern bobwhite quail
Test Toxicity
LD50 56 mg/kg
1,000 mg/kg
LD50 1,089 mg/kg
9,932 mg/kg
LD50 1,800 mg/kg
2,510 mg/kg
LC50 1,280 ppm
12,488 ppm
LC50 2,354 ppm
15,000 ppm
Chemical
Lambda Cyhalothrin
Cyfluthrin
Fenpropathrin
Fenvalerate
Bifenthrin
Fluvalinate
Bifenthrin
Cyhalothrin
Cyhalothrin
Tefluthrin
EPA I.D.
Number
259805
072008
249939
96385
AROTAL06
241388
AROTAL05
073221
073221
261402
YearofEEB
Review
1986
1982
1983
1978
1984
1980
1984
1985
1985
1986
-------
Ideally, a field study should be conducted for each pesticide use. However, this is often
impractical due to cost or time limitations. Therefore, field studies should be designed to
incorporate as many uses as possible without compromising the study or limiting the ability to
make regulatory decisions at the environmentally relevant concentrations. The lambda cyhalothrin
mesocosm that was conducted adjacent to cotton fields was designed in this manner. The tested
concentrations provide a range that includes the estimated concentrations of several different uses.
The loading concentrations of lambda cyhalothrin into adjacent waterways from cotton
fields were determined from modeling (see Analysis: Characterization of Exposure). The
concentrations are approximately equal to the levels used in the middle- and high-dose ponds of the
mesocosm. There were significant effects on invertebrates and fish populations at these dose levels
and also at the lowest dose level, which was equivalent to 1.7 percent of the exposure expected in
ponds. Residues of 58 ^g/kg were found in sediments tested for lambda cyhalothrin 2 months after
the last application.2 The chemical has also been shown to bioaccumulate in fish. Tests were
performed with carp because it is a bottom feeder, thereby having direct contact with residues in
the sediment. Residue concentrations in whole fish were 4,600 to 5,000 times greater than those
found in the sediment.
A mesocosm study attempts to incorporate the structure and function of an ecosystem
(Touart, 1988). The results of the lambda cyhalothrin mesocosm (U.S. EPA, 1989) show that
invertebrate populations in the ponds were reduced in numbers when compared with the control
ponds. Some groups of invertebrates in the high-dose ponds were essentially decimated after one
or two treatments. Reduced numbers at all dose levels were evident into the posttreatment period.
The fish hi each treatment group had statistically significant lower weights and body lengths than
those in the control group. The weights were reduced 20 to 30 percent for all fish and 30 to 40
percent when only the first year (young-of-the-year) fish are considered. This slowing of growth
may affect maturation, reproduction, and ultimately the population structure. These effects
occurred at all concentrations tested; the lowest concentration was 0.6 to 1 ng/L, which is 0.01 of
the label application rate for cotton.
This mesocosm study has confirmed the toxic effects indicated by laboratory testing. The
use of lambda cyhalothrin can reasonably be anticipated to cause significant adverse effects to the
aquatic ecosystem. There were no additional data requirements for the registration of lambda
cyhalothrin.
A pond study with bifenthrin (U.S. EPA, 1990, 1991) investigated effects to aquatic
organisms resulting from application to cotton according to label instructions. The results have
qualitatively shown that bifenthrin has the potential to "change the natural balance and degrade the
ecological integrity of aquatic ecosystems." However, due to the lack of replications in this study,
a quantitative assessment could not be made.
Bifenthrin was found to be more persistent than any other synthetic pyrethroid. For more
than a year following application, residues were measurable in water (4 ng/L), sediment (37 /t/kg),
2Umts for sediments and tissues are on a mass wet weight basis.
10-16
-------
and all 28 fish species analyzed (5 to 9 /tg/kg except gizzard shad, which exhibited 78 fig/kg).
There was also a severe reduction in survival and reproductive potential ofDaphnia and snails.
Zooplankton such as calanoid copepods were eliminated. Macroinvertebrate populations were also
affected. There was a severe reduction in the chironomid insect population, and the entire
population of mayflies Caenis and damselflies Enallagma were eliminated after the first
application. Mayflies remained extremely rare throughout posttreatment, thereby indicating that
recovery would take longer than a year.
Although bifenthrin did not cause high mortality or overt reproductive failure in bluegill
sunfish, it did cause adverse effects on the population. A significant reduction occurred in the
condition factor (physical condition of the fish) and in the growth rate of juvenile bluegill.
,,/i
Gizzard shad were also adversely affected. Almost the entire population of shad (1,600)
died the winter following application. At the time of the kill, residue measurements in shad
reached 440 ng/kg. This was 55,000 to 147,000 times greater than the pond water (0.003 to 0.008
jug/L) and 9 to 11 times greater than the sediment (40 to 50 ng/kg). The field-measured
bioconcentration factors (BCFs) from water approximate the estimated BCF of 50,000 (5 percent of
the octanol-water partition coefficient) and the BCFs calculated from laboratory bioaccumulation
studies. In these studies, bluegill sunfish were exposed to 1.2 ng/L concentration bifenthrin 6,090
times (whole body measurement). It could then be estimated that fish would bioaccumulate
bifenthrin 15,000 to 41,000 times the water concentrations that were present at the time of the kill
(0.003 to 0.008 /tg/L). The various BCFs are similar despite the differences between laboratory
and field conditions and the species measured.
It was concluded that the shad were not able to metabolize and excrete bifenthrin, but
rather accumulated it in fatty tissues or reproductive organs. It was hypothesized that at times of
stress, i.e., cold temperatures and/or diminished food supply, the shad relied on stored energy and,
in the process, the residual bifenthrin was released at toxic concentrations. The data also indicated
that stored residue may adversely affect reproduction (no young-of-year class). Given the
persistence and bioaccumulative potential of bifenthrin, many other aquatic species may experience
the same physiologic response as the gizzard shad.
EEB is very concerned about bifenthrin's apparent potential to cause devastating and lasting
effects to aquatic ecosystems and population structure. Bifenthrin is extremely toxic; i.e., a NOEL
of less than 1 ng/L, so low that the pesticide cannot be detected by traditional analytic means at
concentrations that elicit a biological effect. It is apparent from the pond study that pesticide
residues as well as biological effects persist longer than 1 year after application ceases. It is EEB's
opinion that bifenthrin has the potential, in the long term, to cause extreme population shifts in
aquatic ecosystems and possibly eradication of certain aquatic species.
The results of a residue monitoring study with cypermethrin applied to cotton (U.S. EPA,
1982) showed that repeated aerial applications (16) at 0.125 lb a.i./A at 5-day intervals will result
in cypermethrin residues in a stream ecosystem 8 miles downstream from the treated fields.
During the study there were three runoff events. Maximum cypermethrin residues found in the
stream were 24 ng/L at 165 meters from the point of runoff, 13 ng/L at 2 miles from the point of
runoff, and 3 ng/L 8 miles from the point of runoff. Because of the results of this residue
10-17
-------
monitoring study, an actual pond study was required as the condition for a cotton registration. The
pond study has since been found to be invalid, and a mesocosm study is currently in progress.
Field studies are typically designed to investigate the ecological effects from a single
application season. These effects can be qualitatively and quantitatively measured; however,
multiple-year testing would be required to determine long-term effects on affected populations. At
this time the extent of the population effect is not required by the testing to characterize risk. A
limiting factor in the investigation of long-term effects is the lack of scientific knowledge about
how to interpret the significance of effects manifested at one level of organization (e.g., on
individuals or local populations) on higher levels of organization (e.g., populations, communities,
and ecosystems). Because of the lack of understanding of ecological systems, the reviewer of a
field study must rely on good scientific judgment in interpreting the results.
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
* Acute and chronic laboratory toxicological data for fish and aquatic invertebrates
are scientifically sound, and the amount of data is sufficient to establish general
toxitity levels for synthetic pyrethroids.
Limitations include:
+EEB does not have field study data quantifying population effects of synthetic
pyrethroids to waterfowl. A study of this magnitude would require multiple-year
testing and would need to incorporate natural seasonal predator and prey population
fluctuations. The lack of scientific information in this area would make a study of
this magnitude difficult to evaluate.
*Systemwide effects on aquatic components could be evaluated only on the basis of
two field studies. The other field studies still are being conducted or are in review.
• Because synthetic pyrethroids adsorb to sediment, laboratory bioassays need to be
developed to address the effects of sediment residues on benthic dwelling micro- and
macroinvertebrates. The probability of chemicals becoming released by the sediment
and becoming resuspended into the water column also needs to be investigated.
• The field study does not consider long-term recovery of affected populations, the
effects of multiple-year exposures, or long-term impact on fish reproduction during
the second year. Under the current field test protocols, fish populations are exposed
after the peak reproduction of fishes; no information is available on the effects of
bioaccumulation on egg and sperm production or spawning success.
10-18
-------
Comments on Analysis: Characterization of Ecological Effects (continued)
• Ecological Junction is not well understood, and the ability to examine interactions
among organisms is not well developed. It is not known what level of an effect is
needed to sustain a long-term impact on fish and waterfowl populations. The high
natural variability of biota, both temporally and spatially, within and among ponds
in the field studies also adds to the uncertainty because it may be difficult to
separate out the effects of a chemical.
General comments:
iK
• Several questions were raised by reviewers on this assessment. What uncertainties
exist with the current selection of test species and methods? Is the data base
adequate or should other tests be required on a case-by-case basis? Will addition of
other tests reduce uncertainty in the risk assessment process? For example, should
sediment bioassays be required because synthetic pyrethroids bind to soil sediments
and are persistent? Because field studies have identified the loss of sensitive
cosmopolitan species, such as mayflies, should these animals be suggested as
additional test species? Because synthetic pyrethroids are known to affect
chlorophyll concentrations, should toxidty tests be conducted with phytoplankton?
Reviewers indicated that it was not reasonable to assume that the toxidty data for
fish will protect reptiles or amphibians. Amphibian toxidty test methods exist and
could be used. Biomagnification of pesticides may occur in snakes if they are the
top carnivore in the test area.
* Aquatic invertebrate populations have been adversely affected. The population of
some species has been severely reduced and in others it has been completely
eliminated. Based on available information, population effects to fish are
anticipated because there was a reduction in biomass and a slowing offish growth.
These effects will delay reproduction and thereby cause population effects. The
extent to which the population will be affected is unknown. Multiple-year testing
would be necessary in order to assess the long-term impact on fish and aquatic
invertebrate populations.
«Space limitations prevented a more thorough discussion of the laboratory and field
studies methodologies. For the field studies, a brief overview of a typical mesocosm
protocol would have been useful to emphasize the statistical considerations used in
the experimental design, such as the use of replicated treatment and control ponds
and the temporal and spatial considerations that drive the sampling of biota, water,
and residues.
10-19
-------
10.3.3. Analysis: Characterization of Exposure
Approximately 10 million acres of cotton were harvested in 1987 (U.S. Department of
Commerce), and the largest cotton-producing states were Texas, California, and Mississippi. The
bulk of the United States cotton crop is planted during April but may not be completed until mid-
June, depending on seasonal conditions. Synthetic pyrethroids may be sprayed on cotton as early
as 7 days after planting.
Approximately 2 million acres of sunflowers were harvested nationwide in 1987 (U.S.
Department of Commerce). Sunflowers are grown in much the same way as 'corn. Planting dates
in the Northern Great Plains begin May 1 and extend through late June. Synthetic pyrethroid
treatments are applied as necessary throughout the season up to 28 days preharvest (early
September).
Pesticide residues expected to be found in terrestrial and aquatic ecosystems are calculated
by various methods that will be explained in subsequent sections. These calculations are expressed
as the EEC.
EECs are derived from worst-case scenarios for single or multiple applications over a
single season. A worst-case scenario incorporates practical extremes in determining the EEC, such
as using the maximum application rate, lowest application interval, maximum residue
concentration, etc. This is done to provide a margin of safety for most application situations.
Several synthetic pyrethroids will be used as examples in these calculations.
Terrestrial Exposure. EEB has developed a nomograph (Urban and Cook, 1986) from
Hoerger and Kenaga (1972) and Kenaga (1973). These calculations estimate pesticide residues
found on terrestrial wildlife food items. Residues are determined according to the rate of a single
spray application and the surface area to which it is applied. The application of lambda cyhalothrin
is used as an example. It can be applied at 0.03 Ib active ingredient per acre (a.i./A) (maximum
label rate for cotton). The residues shown in table 10-4 are maximum values expected to be found
on these various substrates immediately after application (U.S. EPA, 1988).
A computer program designed by R. Lee (unpublished) of EEB is used to calculate the
daily accumulated terrestrial or aquatic concentrations from multiple applications and degradation
rates. It is based on the first order kinetic rate equation (Williams and Williams, 1967). Multiple
applications of lambda cyhalothrin, up to a maximum of 0.2 Ib a.i./A per season, can be applied to
cotton by ground or air equipment. To calculate lambda cyhalothrin residues on cotton leaves from
multiple applications, the following information was used: an initial estimated residue of 4 mg/kg,
which is found on leaves and leafy crops (as determined from the previous nomograph example),
an application interval of 3 days (as stipulated on the label), 7 applications per season (maximum
application rate per season 0.2 Ib a.i./A -5- maximum rate per application of 0.03 Ib a.i.), and a
23-day foliar half-life (determined by the equation presented below). The average and maximum
residues estimated from this model are 15 and 22 mg/kg, respectively. The maximum residues on
small insects was 11 mg/kg, and the average residue was 7 mg/kg (U.S. EPA, 1988).
10-20
-------
Table 10-4. Residues Found on Various Substrates After Application (U.S. EPA, 1988)
Maximum
Substrate Residues (mg/kg)
Short range grass 7
Long grass , 3
Leaves and leafy crops (cotton)8 4
Forage (alfalfa, clover) 2
and small insects
Pods containing seeds (sunflower 4
seeds) and large insects
Fruit 0.21
Soil (top 0.1 inch) 0.66
Top 6 inches of water 0.022
(direct application)
10-21
-------
Because a foliar half-life bioassay had not been conducted, the 23-day foliar half-life value
was estimated from information obtained from residue chemistry testing. The following formulas,
which are based on the first-order kinetic rate equation, were used.
,, f, . x log (original concentration) - log (remaining concentration)
i\. ^oecay rate.) — — - ———-—
t (time elapsed in days)
The relationship between the decay rate (K) and the decay half-life (tl/2) is
tl/2 (foliar half-life) = 0-0693 (constant)
Aquatic Exposure. To obtain an approximation of an aquatic EEC, the (unpublished)
algorithms presented below are used. The EEC is determined for a 1-acre pond that receives
pesticide runoff and drift from the surrounding 10 acres that have been treated once by aerial or
mist blower application. Cyfluthrin's maximum label application rate of 0.1 a.i./A on cotton will
be used as an example (U.S. EPA, 1987a). Synthetic pyrethroids adsorb strongly to soil particles;
therefore, their potential to runoff is lessened. However, as in the case of cyfluthrin, the soil half-
life is approximately 60 days. Therefore, the runoff event does not have to occur immediately
after application in order for residue concentrations to reach toxic levels. Nevertheless, aquatic
exposure is most likely to result from spray drift.
Runoff Contribution for EEC
Runoff =A*B*C*D
where:
A = the amount of active ingredient applied (0.1 Ib)
B = the application efficiency (0.6 or 60 percent was chosen)
C = the runoff coefficient (0.001 or 0.1 percent was assumed)3
D = the area of drainage basin (10 acres was selected)
Using the above equation, 0.0006 Ib of active ingredient was estimated in runoff.
3The percentage runoff is usually 1, 2, or 5 percent. These percentages correspond to the
solubility of the pesticide, respectively < 10 mg/L, 10 to 100 mg/L, and > 100 mg/L. Because
the solubility of synthetic pyrethroids is extremely low (cyfluthrin = 0.002 mg/L), 0.1 percent
was used.
10-22
-------
The EEC in a 1-acre pond was estimated for two water depths (6 feet and 6 inches) as
follows:
6 feet deep = 61 /ig/L (conversion factor)4 X 0.0006 Ib a.i. = 0.037 /tg/L
6 inches deep — 734 /ig/L (conversion factor) x 0.0006 Ib a.i. = 0.44 /ig/L
Drift Contribution for EEC
Drift = A * E
where:
A = the amount applied (0.1 Ib)
E = the amount of drift (0.05 or 5 percent was used)5
Using the above equation, a drift of 0.005 Ib of active ingredient was estimated.
The EEC in a 1-acre pond due to drift was estimated as follows:
6 feet deep = 61 /tg/L x 0.005 Ib a.i. = 0.305 /ig/L
6 inches deep = 734 /ig/L x 0.005 Ib a.i. = 3.67 /ig/L
If pertinent environmental fate data are available, the EEC can be calculated using
computer simulation models. These are the Simulator for Water Resources in Rural Basins
(SWRRB) model (Arnold et al., 1990), the Pesticide Root Zone Model (PRZM; Carsel et al.,
1984), and the Exposure Analysis Modeling System (EXAMS II) (Burns, 1990). SWRRB and
PRZM are surface-water runoff (hydrologic) models that estimate loading or the amount of residue
that will enter a body of water either adsorbed to soil particles or having been desorbed from them.
The environmental fate half-lives for photolysis, soil metabolism, and soil adsorption are necessary
to run this model. EXAMS II is a chemical fate (kinetic) model that estimates the chemical fate
and concentration of a pesticide once it has reached a body of water. Calculations are adjusted
according to whether the water is lotic or lentic as in the case of streams or ponds. The
environmental fate half-lives for photolysis, hydrolysis, microbial breakdown, and sediment
adsorption are necessary to run this model. The models can be run to determine the highest
461 /ig/L and 734 /ig/L are standard concentrations calculated from 1 Ib of pesticide per acre
applied to water with a depth of 6 feet and 6 inches, respectively.
5On average, 5 percent (range is 2 to 10 percent) of the application rate per acre can be expected
to drift from a field without a buffer zone to the center of an adjacent 1-acre pond (i.e., 105 feet
from the edge of the field) (Akesson and Yates, 1964; Garner and Harvey, 1984). Nigg et
al. (1984) also reported 5 percent drift.
10-23
-------
maximum EEC found immediately after application or the average EEC over a given time period.
The average EEC could be determined for single or multiple applications throughout the time
period. The average EEC calculated over a given time period of 21 or 32 days, for example,
could then be compared to the effect levels in the Daphnia magna Life Cycle (21 days) Chronic
Toxicity Test or the Fish Early Life Stage Test (32 days).
Cotton. Cyfluthrin is used as an example to calculate the aquatic EEC on cotton (U.S.
EPA, 1987a). Based on the SWRRB/EXAMS II results, the initial concentration may reach 500
ng/L in the water column of a 1-acre pond 6 feet in depth, which receives runoff and some spray
drift from nearby cotton fields. The dissolved concentration reaching streams exiting the ponds
would be much less (20 ng/L). These estimates are based on an application rate of 0.089 Ib
a.i./A, 9 applications per year with a 5-day spray interval, and a half-life of 193 days at pH 7.
Sunflowers. SWRRB/EXAMS II model estimates for another synthetic pyrethroid,
tralomethrin, show water column concentrations of up to 8,800 ng/L in peaks that dissipate rapidly,
and concentrations of up to 50 ng/L that persist more than 7 days. A drift of only 1 percent would
result in concentrations in temporary potholes of more than 150 ng/L (based on 0.01 x application
rate applied directly to 1 acre with 6-inch average depth). Water residues measured in an actual
pond study were as high as 37 ng/L (U.S. EPA, 1987b).
Comments on Analysis: Characterization of Exposure
Strengths of the case study include:
•The EECs -were determined from various methodologies. The predicted adverse
effects based on toxicity data and EECs were confirmed by field testing.
limitations include:
•The characterization of exposure would have benefited by the inclusion of information
on numbers of acres per crop that could be sprayed with synthetic pyrethroids, the
range of concentrations used per application, the maximum seasonal concentration that
can be applied, the number of applications per growing season, the interval between
applications, the method of application, and the relationship between application and
timing of various ecological parameters of interest (e.g., fish reproduction).
•It was difficult to judge the reasonableness of the exposure estimates.
10-24
-------
10.3.4. Risk Characterization
Risk to Ecological Components in Aquatic Environment. The risk characterization methods
used in Ms case study are a combination of the quotient method and the overall weight of
evidence. OPP has presented data showing that aquatic organisms could be exposed during
applications to synthetic pyrethroids at levels that equal or exceed acute and chronic toxicily values.
In the case of cotton use, the highest instantaneous EEC is 500 ng/L. Actual measured residue
concentrations are included for bifenthrin. The lowest LCSO (acute) values ranged from 2.4 to 38
ng/L, and the lowest NOELs (chronic) ranged from 0.17 to 46 ng/L. (Note that the same species
were not necessarily used in both acute and chronic testing). The number of times the EEC or
actual measured residues exceed toxicity concentrations is presented in table 10-5.
v ?)t''
The EECs for these synthetic pyrethroids exceed the acute toxicity levels by factors ranging
from 1 to 208 times and the chronic toxicity levels from 1 to 3,000 times. Synthetic pyrethroid
residues found in the water column can be expected to be adsorbed to the sediment Aquatic
sediment testing has revealed residue concentrations of lambda cyhalothrin at 58 ug/kg 2 months
after application, and 37 ug/kg of bifenthrin a year after application. Even if synthetic pyrethroid
residue concentrations in water or sediment are below acute toxicity levels, it has been shown that
fish can bioaccumulate lambda cyhalothrin residues 4,600 to 5,000 times and bifenthrin residues
56,000 times (bioconcentration factor). Because of bioaccumulative capabilities of these chemicals,
persistence becomes more of a concern man the ambient water or sediment concentrations.
Based on the available toxicity and exposure data, EEB has determined that the use of
synthetic pyrethroids on cotton and sunflowers poses a significant risk to aquatic organisms. The
Special Review criteria in 40 CFR Part 154.7(a)(3) of the regulations have been met because the
EEC >l/2 LC50 and >1/20 LC50 (Urban and Cook, 1986). Consequently, there is a high risk of
adverse effects to nonendangered and endangered aquatic species, respectively. One-half and l/20th
of the LCSO values are used to incorporate margins of safety.
The lambda cyhalothrin mesocosm (U.S. EPA, 1989) demonstrated that fish were more
sensitive under field conditions than in the laboratory. The NOEL in the fish full-life-cycle study
was 31 ng/L. This concentration is 52 times greater than the mesocosm concentrations where
adverse effects to fish occurred (0.6-1 ng/L). It should be noted that bluegill were used in me
mesocosm and fathead minnows were tested in this bioassay. Both species are used as indicator
species; therefore, the toxicity levels can be compared.
Risk to Ecological Components of the Terrestrial Habitats. The terrestrial EECs for both
cotton and sunflower uses are not expected to approach acute toxicity values for avian or
mammalian species; therefore, the risk concern for direct acute effects to these species is low.
Most of the synthetic pyrethroids have adverse effects on eggshell integrity and other reproductive
impairments, but toxicity to avian species is not expected because the terrestrial EECs are lower
than chronic effect levels. To date, field studies have not addressed adverse effects of synthetic
pyrethroids on waterfowl. Nevertheless, EEB has concluded that the use of synthetic pyrethroids on
sunflowers may pose a significant risk to breeding waterfowl and hatchlings because of their
reliance on aquatic prey organisms.
10-25
-------
Table 10-5. EECs or Actual Measured Residues Exceeding Toxicity Concentrations
(Cotton Use Pattern)
Chemical
Cyfluthrin
Tralomethrin
Bifcnthnn
Lowest Acute
LCM (jig/L)
2.4
(40069501,
1987)'
38
(11507-74-03,
1981)
35
(264647,
1987)
Lowest
Chronic
NOEL (ng/L)
0.17
(262443,
1988)
46
(41860701,
1991)
13
(41156501,
1989)
Estimated EEC
Estimated Exceeds
EEC from Measured
SWRRB/ Residues
Exams (ng/L) (ng/L) LCM
500 (water) - 208x
(189625, 1987)
50 (water) - Ix
(194619, 1987)
4 (water) 2 (water) Ix
(water)
17,000 10,000 - 60,000
ng/kg (sediment) ng/kg (sediment)
(279-EUP-RNR, (40981801, 1991)
1984)
NOEL
3,000x
Ix
3x
(water)
* EPA identification number, year of EBB review.
10-26
-------
Risk to the Waterfowl Prey Base. Ninety percent of the sunflower growing region is in
North Dakota, South Dakota, and Minnesota. These states are in the prairie pothole region, which
is responsible for more than 50 percent of the annual duck production in North America (Smith et
al., 1964). During the reproduction period, adults and ducklings are highly dependent on aquatic
macroinvertebrates (Krapu and Swanson, 1975; Sheehan et al., 1987). Sheehan et al. (1987) also
state that "inadequate nutrition adversely affects reproduction" and that calcium and protein
requirements come from aquatic invertebrates that supply 2 to 4 times the protein levels of plants.
It has been shown that laying hens consumed more than 70 percent animal food and that some
species are totally dependent on invertebrates (Sheehan et al., 1987).
Essential amino acids must be in the proper ratios for protein formation. Krapu and
Swanson (1975) state that when pintail hens feed] on several invertebrate prey items, they are
balancing their amino acid needs. Joyner (1980) investigated seven morphological and biological
variables that might affect pond selection by ducks. It was concluded that only invertebrate
numbers and taxa present were correlated with duck usage. This information shows the importance
of prairie potholes containing large invertebrate populations with adequate species diversity.
Laying hens are opportunistic feeders and will switch prey when there is a seasonal change
in species abundance. Laying efforts generally terminate, however, when the abundance of
selected prey items starts to decline. Females also terminate renesting attempts when prey
availability is reduced (Sheehan et al., 1987). Renesting attempts are very important, especially in
the pothole regions because of the high destruction rate of nests from agricultural operations
(Krapu and Swanson, 1975).
Evidence indicates that ducklings have a high dependence on animal food, primarily
invertebrates, for periods of 1 to 7 weeks post-hatch. A diet of greater than 50 percent animal
protein is required to sustain a growth rate in mallard ducklings. Growth rate is correlated with fat
deposit, plumage development (both important in temperature control), strength, and activity.
Therefore, duckling growth is roughly equivalent to survival, and fast growth increases their
likelihood of survival (Sheehan et al., 1987).
The National Research Council of Canada (1986) has stated that the survival of young birds
that feed in ponds adjacent to pyrethroid-treated agricultural lands may be affected by the timing of
application. Fenvalerate is conditionally registered for use on sunflowers and can be expected to
be applied in the prairie pothole region during the stages of waterfowl reproduction that rely most
heavily on aquatic invertebrates and emerging insects. These stages of reproduction are the last
half of the duck egg-laying phase and the entire hatching and early duckling phase.
Risk to Bees and Plant Pollination. Sunflower production used to be almost entirely
dependent on bees and other insects for pollination. Most new sunflower hybrids have been
selected to self-pollinate without pollinator activity. But some hybrid plants such as the Fl type
still require over 20 bees per 100 heads in bloom (McMullen, 1985). Synthetic pyrethroids are
generally highly toxic to honey bees under laboratory conditions. But some of them have repellent
properties that make exposure less likely in the field.
10-27
-------
Sources of Uncertainty. There are uncertainties in characterizing risks because variations
and assumptions have been made in the exposure and toxicity assessments. Environmental fate
data were not available to be used in determining the terrestrial EEC from multiple applications,
but a method was presented hi which foliar half-life was estimated. The aquatic EEC equation
contains assumed values for application efficiency and percentage drift. Both the SWRRB and
EXAMS II models contain variations that are inherent hi chemical fate data because such data are
usually expressed as a range and not a fixed number. These models also contain hydrologic
variations due to regional climatic, topographical, soil type, and soil texture differences. It is
assumed that the limited number of test species used in toxicity testing represent all wildlife
species. Adverse effects to waterfowl were based not on direct field testing but on indirect
evidence obtained from the literature and a measured decrease in their prey base (supported by
field studies). Even though there are uncertainties in«the risk characterization process, EEB has
been able to accurately predict when certain chemicals are expected to cause adverse ecological
effects. As in the case with synthetic pyrethroids, both the mesocosm study and pond study have
confirmed EEB's presumption of risk.
Comments on Risk Characterization
Strengths of the case study include:
+The case study'presented for synthetic pyrethroids is a good example of the process
used by EPA for the risk assessment of pesticides. Its several strengths are given
below.
—The case study illustrates how various types of information are used and integrated;
these include laboratory acute and chronic toxicity data, estimated environmental
concentration of pesticides from models, and field studies on actual or simulated
aquatic ecosystems. The field studies were used to confirm the presumption of risk
indicated by the toxicity data and exposure assessment.
—The selection of synthetic pyrethroids was excellent because these chemicals are
extremely toxic to aquatic invertebrates and fish, persistent in the environment, and
bioaccumulated.
—The case study illustrates the sensitivity of biological systems to chemical
concentrations that are below the level of detection.
—Because the case study concentrated on conditionally registered pesticides, the study
illustrated how the Agency deals with data gaps in the risk assessment process.
—This case study illustrated a range of potential ecological effects for different use
patterns, cotton and sunflowers, that would likely affect two different types of aquatic
habitats, ponds and prairie potholes.
—Given the state of the art in ecological risk assessment, this study illustrated that both
quantitative and qualitative data are useful in making professional judgments.
10-28
-------
Comments on Risk Characterization (continued)
• It has been shown that aquatic invertebrates are an essential food source for breeding
waterfowl hens and hatchings. Therefore, an indirect qualitative relationship between
synthetic pyrethroids and waterfowl can be drawn. If synthetic pyrethroid use severely
reduces or eliminates aquatic invertebrate populations, then adverse effects to waterfowl
may occur.
Limitations include:
• The assessment could have benefited from a better integration of field and laboratory
data.
• No rationale was provided for using the LC50 in quotient method. The technical
basis for using 1/2 and 1/20 the LC50 to trigger other laboratory and field studies
was not described, although a literature citation is given.
• The relationship between pesticide-induced aquatic invertebrate population reductions
and adverse effects to waterfowl requires additional documentation.
General comment:
• The process for risk assessment creates myriad problems for the registrant because of
the length of time it takes to provide the data for the risk assessment and the rapid
changes in methods for measuring ecological effects and predicting exposure. The
registrant must use state-of-the-art methods to generate data to be used in the risk
assessment several years later. During this time, the process must be open to receive
new information, especially information not required currently to characterize potential
ecological effects.
10-29
-------
10.4. REFERENCES
Akesson, N.B.; Yates, W.E. (1964) Problems related to application of agricultural chemicals and
resulting drift residues. Ann. Rev. Entomol. 9:285-318.
Arnold, J.G.; Williams, J.R.; Nicks, A.D.; Sammons, N.B. (1990) SWRRB: a basin scale
simulation model for soil and water resources management. College Station, TX: Texas
A&M University Press.
ASTM Standard E 729-80. (1980) Practice for conducting acute toxicity tests with fishes,
macroinvertebrates, and amphibians. Philadelphia, PA: American Society for Testing and
Materials. \
Burns, L.A. (1990) Exposure analysis modeling system: user's guide for EXAMS II version 2.94.
Athens, GA: U.S. EPA, Environmental Research Laboratory. EPA 600/3-89/084.
Carsel, R.F.; Smith, C.N.; Mulkey, L.A.; Dean, J.D.; Jowise, P. (1984) User's manual for the
Pesticide Root Zone Model (PRZM). Athens, GA: U.S. EPA, Environmental Research
Laboratory. EPA 600/3-84/109.
Committee on Methods for Toxicity Tests with Aquatic Organisms. (1975) Methods for acute
toxicity tests with fish, macroinvertebrates, and amphibians. U.S. EPA, Ecol. Res. Series.
EPA 660/3-75/009.
Federal Insecticide, Fungicide, and Rodenticide Act as Amended, 1988. U.S. EPA, Office of
Pesticide Programs. EPA 540/09-89/012.
Garner W.Y.; Harvey, J., Jr. (1984) Chemical and biological controls in forestry. ACS
Symposium Series 238. Washington, DC: American Chemical Society.
Hitch, R.K. (1982) Pesticide assessment guidelines subdivision L hazard evaluation: nontarget
insects. EPA 540/9-82-019.
Hoerger, P.; Kenaga, E.E. (1972) Pesticide residues on plants: correlation of reproductive data as
a basis for estimation of their magnitude in the environment. Environmental quality and
safety; global aspects of chemistry, toxicology and technology as applied to the
environment. Vol. 1. New York, NY: George Thieme Publishers, Stuttgart Academic
Press, Inc.
Joyner, D.E. (1980) Influence of invertebrates on pond selection by ducks in Ontario. J. Wildl.
Manage. 44(3):700-705.
Kenaga, E.E. (1973) Factors to be considered in the evaluation of the toxicity of pesticides to birds
in their environment. Environmental quality and safety; global aspects of chemistry,
toxicology, and technology as applied to the environment. Vol. 2. New York, NY: George
Thieme Publishers, Stuttgart Academic Press, Inc.
10-30
-------
Krapu, G.L.; Swanson, G.A. (1975) Some nutritional aspects of reproduction in prairie nesting
pintails. /. Wildl. Manage. 39(1): 156-162.
McMullen, M.P. (1985) Sunflower production and pest management. North Dakota Cooperative
Extension Service. Extension Bulletin No. 25.
National Research Council of Canada. (1986) Pyrethroids: their effects on aquatic and terrestrial
ecosystems. Publication No. NRCC 23476 of the Environmental Secretariat, Ottawa,
Canada.
National Water Quality Laboratory Committee on Aquatic Bioassays. (1971) Recommended
bioassay procedure for fathead minnow- Pimephales promelas (Rafmesque) chronic tests.
In: Biological field and laboratory methods. U.S. EPA. EPA 670/4-73/001. pp. 15-24.
Nigg, H.N.; Stamper, J.H.; Queen, R.M.; Knapp, J.L. (1984) Fish mortality following application
of phenthoate to Florida citrus. Bull. Environ. Contam. Toxicol. 32(5):587-596.
Preston, W.; Hitch, R.K. (1982) Pesticide assessment guidelines subdivision E hazard evaluation:
wildlife and aquatic organisms. EPA 540/9-82/024.
Sheehan, P.J.; Baril, A.; Mineau, P.; Smith, O.K.; Harfenist, A.; Marshall, W.K. (1987) The
impact of pesticides on the ecology of prairie nesting ducks. Technical Report Series No.
19, Canadian Wildlife Service, Headquarters.
Smith, A.G.; Stoudt, J.H.; Gollop, J.B. (1964) Prairie potholes and marshes. In: Waterfowl
tomorrow. Washington, DC: U.S. Government Printing Office, pp. 39-50.
Touart, L.W. (1988) Hazard Evaluation Division technical guidance document: aquatic mesocosm
tests to support pesticide registration. EPA 540/09-88/035.
Urban, D.; Cook, N. (1986) Hazard Evaluation Division standard evaluation procedure, ecological
risk assessment. EPA 540/9-86/167.
U.S. Department of Commerce. (1987) Census of agriculture. Geographic Area Series, Vol. 1.
Washington, DC: U.S. Government Printing Office.
U.S. Environmental Protection Agency. (1982) Ecological Effects Branch review—cypermethrin
residue samples in a runoff study. File/Registration No. 10182AL.
U.S. Environmental Protection Agency. (1987a) Ecological Effects Branch review—cyfluthrin use
on cotton. File/Registration No. 3125-GLR. Record No. 189625.
U.S. Environmental Protection Agency. (1987b) Ecological Effects Branch review—tralomethrin
use on sunflowers. File/Registration No. 34147-2. Record No. 194619.
10-31
-------
U.S. Environmental Protection Agency. (1988) Ecological Effects Branch review—lambda
cyhalothrin use on cotton. File/Registration No. 10182-OA. Record No. 21292.
U.S. Environmental Protection Agency. (1989) Ecological Effects Branch review—lambda
cyhalothrin mesocosm review. MRID 40515901.
U.S. Environmental Protection Agency. (1990) Ecological Effects Branch review—bifenthrin pond
study review. MRID 40981801.
U.S. Environmental Protection Agency. (1991) Ecological Effects Branch review—amended data
evaluation record for the bifenthrin pond study. MRID 40981801.
' f *..
Williams, V.R.; Williams, H.B. (1967) Basic physical chemistry for the life sciences. San
Francisco, CA: W.H. Freeman and Company.
10-32
-------
APPENDIX A
TIER TESTING SYSTEM FOR ECOLOGICAL EFFECTS
10-A1
-------
TIER TESTING SYSTEM FOR ECOLOGICAL EFFECTS
TIER I
Terrestrial:
Mammalian toxicity data
Acute oral LD50 test—bird
Dietary LC50 test—bird
Seed germination/seedling emergence and vegetative vigor
Acute contact LD50 test—honey bee
quatic:
96-hour LC50 test—coldwater fish
96-hour LC50 test—warmwater fish -^
48-hour (or 96-hour) LC50/EC50 test—freshwater aquatic invertebrate
96-hour EC50 test—algae
TIER II
Terrestrial:
• Wild mammal toxicity data
• Avian reproductive studies
• Special studies with avian or mammalian species (e.g., avian cholinesterase test,
secondary toxicity)
• Honey bee residue on foliage
• Seed germination/seedling emergence and vegetative vigor
Aquatic:
• 96-hour LC50 test—estuarine/marine fish
• 96-hour LC5o test—estuarine/marine crustacean
• 48-hour £€50 test—bivalve embryo-larvae or
96-hour £€50 test—bivalve shell deposition
• Fish early-life-stage MATC or Effect/No Effect Level
• Aquatic invertebrate life cycle MATC or Effect/No Effect Level
• Fish bioaccumulation factor, e.g, 1000X
• Special aquatic organism test data (e.g., fish acetylcholinesterase levels)
• Aquatic plant growth testing
TIERS III and IV
Terrestrial:
• Simulated and actual field testing with avian and/or mammalian species
• Field test for pollinators
Aquatic:
• Fish full-life-cycle MATC or Effect/No Effect Level (Tier III)
• Field testing for aquatic organisms (Tier IV) (e.g., mesocosm or pond study)
10-A2
-------
SECTION ELEVEN
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
TOXIC DISCHARGES TO SURFACE WATERS: ASSESSING THE RISK TO
AQUATIC LIFE USING NATIONAL AND SITE-SPECIFIC
WATER QUALITY CRITERIA
-------
AUTHORS AND REVIEWERS
AUTHORS
Robert L. Spehar
Environmental Research Laboratory-Duluth
U.S. Environmental Protection Agency
Duluth, MN
REVIEWERS
Anthony F. Maciorowski
(Lead Reviewer)
Fish Culture and Ecology Laboratory
U.S. Fish and Wildlife Service
Kearneysville, WV
Robert Brooks
School of Forest Resources
Pennsylvania State University
University Park, PA
G. Allen Burton
Biological Sciences Department
Wright State University
Dayton, OH
Rick D. Cardwell
Parametrix, Inc.
Bellevue, WA
Michael J. Dover
The Cadmus Group, Inc.
Peterborough, NH
Anthony R. Carlson
Environmental Research Laboratory-Duluth
U.S. Environmental Protection Agency
Duluth, MN
Joseph Makarewicz
Department of Biological Sciences
State University of New York
at Brockport
Brockport, NY
Thomas P. O'Connor
National Status and Trends Program
National Oceanic and Atmospheric
Administration
Rockville, MD
Jerry R. Schubel
Marine Sciences Research Center
State University of New York
at Stony Brook
Stony Brook, NY
11-2
-------
CONTENTS
ABSTRACT 11-6
11.1. RISK ASSESSMENT APPROACH 11-7
11.2. STATUTORY AND REGULATORY BACKGROUND 11-7
11.3. CASE STUDY DESCRIPTION 11-9
11.3.1. Problem Formulation 11-9
11.3.2. Analysis: Characterization of Ecological Effects 11-11
11.3.3. Analysis: Characterization of Exposure 11-22
11.3.4. Risk Characterization 11-24
11.4. REFERENCES 11-27
APPENDIX A—WATER QUALITY CRITERIA FOR AQUATIC LIFE 11-Al
11-3
-------
LIST OF FIGURES
Figure 11-1. Structure of analysis for assessing risk using national
and site-specific criteria 11-8
LIST OF TABLES
Table 11-1. Recalculation Procedure: Acute Toxicity Data
for Cadmium From National Criteria Document
for Resident Species of the St. Louis River 11-15
Table 11-2. Indicator Species Procedure: Acute Values (LC50) for
Indicator Species Exposed to Cadmium in St. Louis River
and Reconstituted Water 11-17
Table 11-3. Acute (LC50) and Chronic Values for Aquatic Organisms
Exposed to Cadmium in St. Louis River Water 11-17
Table 11-4. Chronic Toxicity Values of Two Species Exposed to Cadmium
in St. Louis River and Lake Superior Water 11-19
Table 11-5. Resident Species Procedure: Minimum Data Set of
Resident Aquatic Species Exposed to Cadmium in
St. Louis River Water 11-19
Table 11-6. Cadmium Water Quality Criteria Derived From National
and Site-Specific Procedures 11-20
11-4
-------
LIST OF ACRONYMS
CCC criterion continuous concentration
CMC criterion maximum concentration
EC50 effective concentration for 50 percent of organisms tested
EPA U.S. Environmental Protection Agency
C^i
FAV final acute value
FCV final chronic value
FPV final plant value
FRY final residue value
LA load allocations
LC50 lethal concentration for 50 percent of organisms tested
NOEL no observed effects level
NPDES National Pollutant Discharge Elimination System
WLA wasteload allocations
WQC water quality criteria
11-5
-------
ABSTRACT
Guidelines recommended by the U.S. Environmental Protection Agency (EPA) for deriving
national and site-specific water quality criteria for the protection of aquatic life were evaluated in a
case=study of the St. Louis River basin in Duluth, Minnesota. The recalculation, indicator species,
and resident species procedures of the site-specific approach were used to modify the national
cadmium criteria to site-specific criteria. The procedures accounted for differences in species
sensitivity and the biological availability and/or toxicity of cadmium due to physical and/or
chemical characteristics of site water. The national and site-specific criteria approaches were used
as a tool to assess the risk of ambient metal exposure to resident aquatic life.
These guidelines provide an approach to applying national water quality criteria on a site-
specific basis to reflect local environmental conditions. Because the national criteria are designed
to protect the biological integrity of all water bodies, these criteria serve as benchmarks and may
require adjustments for site-specific applications (e.g., risk assessments). Consideration of local
conditions assures that criteria for a particular body of water are tailored specifically to its aquatic
life and uses.
A major strength of the site-specific guidelines is that they are based on national guidelines
that, have undergone extensive scientific review to assure their general applicability. However, the
water quality criteria guidelines constitute only one approach to assessing the risk of pollutants on
aquatic systems. For comprehensive, ecologically based risk assessments, this approach should be
used in conjunction with other EPA procedures, such as the whole-effluent approach used for
dealing with mixtures of chemicals, as well as procedures for developing sediment, wildlife, and
biological criteria.
11-6
-------
11.1. RISK ASSESSMENT APPROACH
This case study does not represent a complete risk assessment as defined in the Framework
for Ecological Risk Assessment (figure 11-1). Although it provides useful information on the '
stressor and its anticipated ecological effects, this case study does not characterize the exposure
where all stressors to the St. Louis River Estuary are considered. Instead, cadmium was measured
from a limited monitoring program at this site so that this metal could be used as an example of a
stressor. To conduct a complete risk assessment; a significant amount of information,' as defined
by the objectives of any case study, would need to be gathered to identify and quantify all stressors
for a site (Stephan et al., 1985).
The following case study is based on an%arlier study by Spehar and Carlson (1984a, b) that
was designed to demonstrate how to use the U.S. Environmental Protection Agency's (EPA's) site-
specific approach for modifying national water quality criteria. The current case study was
developed as an example by reorganizing information from the earlier study into the present risk
assessment format. The results show how the principles of deriving site-specific criteria can be
used as a tool for making ecological risk assessments.
The procedures of the national and site-specific guidelines (Carlson et al., 1984; Stephan et
al., 1985), along with their stated assumptions and limitations, need to be understood before the
current case study will have meaning in such assessments. These procedures constitute only one
approach—the chemical-specific approach—that EPA uses in its water quality-based program for
limiting toxins hi surface waters. The chemical-specific approach focuses on the protection of
aquatic life and does not completely address other issues involving wildlife or terrestrial
communities. To conduct a comprehensive risk assessment of a particular site, the chemical- '
specific approach should be used hi combination with the whole-effluent approach (for
characterizing mixtures) and EPA's procedures for deriving sediment, wildlife, and biological
criteria.
11.2. STATUTORY AND REGULATORY BACKGROUND
The Clean Water Act of 1977 (Section 304[al[l]) requires EPA to review and publish water
quality criteria necessary to protect public water supplies and to safeguard the propagation of
shellfish, fish, and wildlife. The criteria provide scientific data and guidance on the environmental
effects of pollutants and can be used to derive water quality-based regulatory requirements such as
effluent limitations, water quality standards, or toxic pollutant effluent standards (U.S. EPA,
1980a).
National water quality criteria have been developed by applying a set of guidelines (Stephan
et al., 1985) to data for certain pollutants designated as toxic under Section 307(a)(l) of the Clean
Water Act of 1977 pursuant to an agreement in the case of Natural Resources Defense Council et
al. v. Train, 1976. According to these guidelines, water quality criteria should be based on an
array of data for the types of plant and animal species that occupy various trophic levels. Based on
these data, criteria can be derived that should adequately protect the types of species necessary to
support an aquatic community.
11-7
-------
Figure 11-1. Structure of Analysis for Assessing
Risk Using National and Site-Specific Criteria
PROBLEM FORMULATION
Stressors: cadmium; other stressors interact with this
metal but were not considered.
Ecological Components: invertebrates and fish species.
'ix,
Endpoints: assessment endpoint is protection of aquatic
community structure functions and integrity. Measurement
endpoint includes single species toxicity tests.
A
ANALYSIS |
Characterization of
Exposure
A limited number of
measurements were made
of cadmium concentrations
in river water.
1
Characterization of
Ecological Effects
Data supporting the
national criterion for
cadmium and new toxicity
tests using St. Louis River
water were used to
establish site-specific
criteria for cadmium.
RISK CHARACTERIZATION
The limited data on cadmium concentrations in the St.
Louis River were compared to the national and site-
specific criteria. The limitations of the approach
are described.
11-8
-------
Although the criteria represent a reasonable estimate of pollutant concentrations consistent
with the maintenance of designated uses, each state may modify these values to reflect local
conditions. Because national criteria may be either underprotective or overprotective, EPA has
developed guidelines (U.S. EPA, 1983; Carlson et al., 1984) to adapt national water quality criteria
to local conditions (i.e., site-specific criteria). Most national criteria (see appendix A) are based on
information obtained from toxicity and bioconcentration tests conducted in the laboratory. Some
criteria, however, as stated in the national guidelines, can be based on assessments of adequate field
data from actual sites. These assessments include information on aquatic life (residue content) that
may be useful for protecting wildlife populations.
In other cases, toxicological information obtained on laboratory-tested aquatic species may
not be applicable to species in specific water bodies. The species at a particular site may be more
or less sensitive than those included in the national criteria data base, or the physical and/or
chemical characteristics of the water at the site may alter the biological availability and/or toxicity
of the material.
This case study uses laboratory data and EPA's site-specific approach as a tool for assessing
the risk posed by cadmium to the St. Louis River Estuary.
11.3. CASE STUDY DESCRIPTION
11.3.1. Problem Formulation
Site Description. The St. Louis River system, located primarily in southern St. Louis
County, Minnesota, encompasses approximately 1,400 miles of streams. The St. Louis River and its
tributaries account for 815 of these 1,400 miles (Spehar and Carlson, 1984a). The mouth of the St.
Louis River is a freshwater estuary containing approximately 11,500 acres of water. It has been
developed into a major industrial port that serves as the economic base for the cities of Duluth,
Minnesota, and Superior, Wisconsin. The site chosen for a source of dilution water in this study
was located approximately 34 miles upstream from the mouth of the Duluth-Superior estuary, at the
State Highway 33 crossing in the city of Cloquet, Minnesota.
The St. Louis River Estuary was chosen as the specific site for the case study because
enough information on the biological and chemical characteristics of this site was known, hi
addition, sufficient information was available to meet the requirements for conducting a risk
assessment by using the components of the risk assessment framework. For example, a limited
water quality monitoring data set (Hammermeister et al., 1983) was available and resident aquatic
species for the site were previously characterized and documented (Spehar and Carlson, 1984a) for
use in defining ecological components.
Stressors. Cadmium was used as an example to demonstrate EPA's site-specific approaches
to risk assessment (Spehar and Carlson, 1984a, b). Because this was a sample exercise, other
stressors that were known to exist at the site, such as different metals, organic chemicals, and
conventional pollutants (as single chemicals or as mixtures in the water column and/or sediment),
were not considered for use in this case study. Cadmium was chosen in the earlier publication
because it is known to be highly toxic to aquatic organisms.
11-9
-------
Cadmium is commonly found in the environment in treated municipal wastes (U.S. EPA,
1980b). In addition, its chemistry in water may be influenced by changes in water quality (Geisy
et al., 1977; Calamari et al., 1980; Reid and McDuffie, 1981), which would be a major
consideration in modifying the present national criteria for use in this case study. Detailed
lexicological characterizations were conducted according to EPA's guidelines for deriving national
and site-specific criteria (Carlson et al., 1984; Stephan et al., 1985).
Ecological Components. Although the components in this case study are the resident
species of the St. Louis River Estuary (Spehar and Carlson, 1984a), they are represented in this
case study by the array of species defined by the minimum data sets in the national and site-specific
guidelines (Stephan et al., 1985; Carlson et al., 1984, respectively). The specific components used
in this representation are delineated in the procedures described in the characterization of ecological
effects section of this case study.
An assessment of the biota in the St. Louis River was conducted according to the
procedures delineated below for modifying the national criteria for cadmium at this site. The
approach assumes that protection of 95 percent of the tested species according to EPA's guidelines
(Carlson et al., 1984; Stephan et al., 1985) will adequately protect community structure, function,
and integrity.
Endpoints. The assessment endpoint is protection of aquatic community structure, function,
and integrity. Measurement endpoints included the acute and chronic toxicity test results used in
this exercise to represent the array of resident species at this site, as well as those delineated by the
guidelines for deriving both the national and the site-specific criteria.
Comments on Problem Formulation
Strengths of the case study include:
•The causal relationship for cadmium exposure and aquatic life toxicity is well
documented.
• Site-specific water quality criteria may be better suited than the national criteria for
cadmium.
+The procedures are designed to derive site-specific water quality criteria by
allowing substantial flexibility in the method used. This flexibility should permit
regulatory agencies to choose the most appropriate and efficient means of obtaining
the information needed to modify national criteria for each particular site for use in
risk assessments. Site-specific water quality criteria for cadmium and the St. Louis
River Estuary obtained from the site-specific guidelines appear to be logical because
they take into account the national cadmium criteria and the physical, chemical, and
biological characteristics of water at this site.
11-10
-------
Comments on Problem Formulation (continued)
Limitations include:
*77ie endpoints, although derived in a rigorous and standardized manner, do not
relate directly to meaningful ecological endpoints that can be readily measured in the
field. Water quality criteria guidelines encourage the use of field studies for deriving
criteria; however, field studies are usually not available for most chemicals:
Consequently, in most cases, laboratory data are used. Although criteria derived
from single species responses from laboratory tests have been validated with
ecologically based studies (U.S. ERA, 1989), it is difficult to demonstrate that
criteria derived in this manner do in fact produce the desired results at the ecosystem
level.
+The water quality-based approach assumes that the aquatic ecosystem is at risk
when water quality criteria are exceeded, but does not delineate the actual species,
populations, communities, or other system components at risk.
+The case history is limited to the toxicological risk of cadmium without considering
other contaminants or ecosystem stressors.
11.3.2. Analysis: Characterization of Ecological Effects
Background. Two types of stress-response assessments were used for, this case study. The
first derived ambient freshwater aquatic life water quality criteria for cadmium (U.S. EPA, 1985)
and was completed before the present case study according to the national guidelines (Stephan et
al., 1985). The second assessment derived freshwater site-specific water quality criteria for
cadmium in the St. Louis River (Spehar and Carlson, 1984a, b) as outlined in the following
procedures.
EPA criteria consist of three parameters (magnitude, frequency, and duration) that are
developed for two levels of effect (acute and chronic). Uncertainties associated with the derivation
of EPA national and site-specific criteria are not detailed here; however, validation studies are
available (U.S. EPA, 1983) that support the use of these approaches until more meaningful
ecological endpoints can be measured readily on an ecosystem basis.
The following is an abbreviated list of endpoints and definitions that will be helpful in
understanding the present procedures for deriving site-specific criteria for cadmium. A more
complete list can be found in the national and site-specific guidelines for deriving water quality
criteria (U.S. EPA, 1983; Stephan et al., 1985).
• Acute value: A 48- to 96-h LC50 or ECs0, depending on the species.
11-11
-------
« Chronic value: The geometric mean of the lower chronic limit (highest tested
concentration in an acceptable chronic test that did not cause significant decreases
from the control) and the upper chronic limit (lowest tested concentration in an
acceptable chronic test that caused significant decreases from the control).
• Acute/chronic ratio: The ratio of an acute value for a species to a comparable
chronic value for that species tested in the same water.
• Final acute value (FAV): An estimate of the concentration of a material
corresponding to a cumulative probability of 0.05 in the acute toxicity values for the
genera with which acute tests have been conducted for that material. (For an
exception and the effects of water quality characteristics on this value, see Stephan
etal., 1985.)
» Final chronic value (FCV): An estimate of the concentration of a material
corresponding to a cumulative probability of 0.05 in the chronic toxicity values for
genera with which chronic tests have been conducted for that material, usually
obtained by dividing the FAV by the final acute/chronic ratio. (For an exception
and the effects of water quality characteristics on this value, see Stephan
etal., 1985.)
• Final acute/chronic ratio: The geometric mean of all the species' mean
acute/chronic ratios available for both freshwater and saltwater species. (For
variations in the calculation of this value, see Stephan et al., 1985.)
• Final residue value (FRY): The lowest of the residue values obtained by dividing
the maximum permissible tissue concentrations by the appropriate bioconcentration
or bioaccumulation factors.
• Criterion maximum concentration (CMC): A criterion value that is equal to one-
half of the FAV.
• Final plant value (FPV): The lowest result from a test with an important aquatic
plant species in which the concentrations of the test material were measured and the
endpoint was biologically important.
• Criterion continuous concentration (CCC): The criterion value that is the lowest of
the FCV, FRV, and FPV, unless other data from field studies or laboratory tests
show that a lower value should be used.
National Criteria for Cadmium. With the possible exception of a locally important but
highly sensitive species, freshwater aquatic organisms and their uses should not be affected
unacceptably if the 4-day average concentration (CCC in jig/L) of cadmium does not exceed the
numerical value given by e(0-7852[ln(hardness)]-3.490) more tjjan once every 3 years on average and if
the 1-hour average concentration (CMC in /tg/L) does not exceed the numerical value given by
e(l.l28Dn(hardness)]-3.828) more ^an once every 3 years on average. For example, at hardness values
11-12
-------
of 50, 100, and 200 mg/L as CaCO3, the 4-day average concentrations of cadmium are 0.66, 1.1,
and 2.0 ug/L, respectively, and the 1-hour average concentrations are 1.8, 3.9, and 8.6 /Kg/L,
respectively (U.S. EPA, 1985).
Site-Specific Criteria. Numerous assumptions are associated with the site-specific
guidelines, most of which also apply to and are included in the national guidelines. A few
assumptions need to be emphasized. The principal assumption is that the species sensitivity
ranking and toxicological effect endpoints (e.g., survival, growth, and reproduction) derived from
laboratory tests will be similar to those derived from site situations. Another assumption is that the
protection of all of the site species all of the time is not necessary because aquatic life can tolerate
some stress and occasional adverse effects.
It is also assumed that the site-specific guidelines are an attempt to improve the protection
of the various uses of aquatic life by accounting for toxicological differences in species sensitivity
or the biological availability and/or toxicity of a material at specific sites. Modification of the data
set must always be scientifically justifiable and consistent with the assumptions, rationale, and spirit
of the national guidelines.
In these types of assessments, EPA criteria are developed (using toxicological data) as
national recommendations to assist states in developing water quality standards. Standards are then
adopted by the states to designate uses and to define ambient characteristics of receiving waters.
These standards must be maintained to allow those uses and must be met before wastewaters can be
discharged legally.
Procedures for calculating site-specific criteria. Three procedures from the site-specific
guidelines (U.S. EPA, 1983; Carlson et al., 1984) were used in this study to modify the national
CMC and CCC for cadmium (see the rationale for the site-specific guidelines listed in these
publications). All procedures were conducted by using the pertinent species, exposures, and
calculations that would be needed to modify the national criteria. This characterization was made
even though all three procedures would not necessarily be used in an actual site modification.
The three procedures used to calculate the site-specific criteria for cadmium in the St. Louis
River Estuary were as follows: (1) the recalculation procedure, to account for differences in
cadmium sensitivity between species resident in the St. Louis River Estuary and those species
contained in the national cadmium criteria document (U.S. EPA, 1985); (2) the indicator species
procedure, to account for differences in the biological availability and/or toxicity of cadmium due
to physical and/or chemical characteristics of the St. Louis River water and laboratory water by
deriving a water-effect ratio (toxicity in site water divided by the toxicity in laboratory water); and
(3) the resident species procedure, to account simultaneously for differences in both resident
species sensitivity and differences that may be attributed to water quality. Acute lethality was the
only endpoint considered in the recalculation procedure; the indicator species and resident species
procedures utilized test data on lethality, growth, and reproductive potential.
Procedures to determine a final residue value or a final plant value (which are required in
the national guidelines for deriving water quality criteria) are not used in this case study because
they are not sensitive endpoints for cadmium. Cadmium is not lipid-soluble and will not
11-13
-------
biomagnify in aquatic systems and affect higher food chain organisms such as wildlife. Plants are
not as sensitive as aquatic animals to cadmium and would not be affected at criteria levels based on
animal tests. Thus, the criteria concentrations calculated in this case study are based on the most
sensitive FAVs and FCVs, which were determined from toxicity tests with aquatic animals.
A detailed description of how to define a site; the rationale, assumptions, and limitations of
the site-specific procedures; and the relationship of site-specific procedures to those used for
deriving national water quality criteria are included in the site-specific guidelines (U.S. EPA, 1983;
Carlson etal., 1984).
Recalculation procedure. The recalculation procedure modifies the national CMC for
cadmium (U.S. EPA, 1985) by eliminating data for nonresident species from the national
data base.
The data set for resident species in the St. Louis River was sufficient to meet the minimum
data set requirements of the national guidelines (Stephan et al., 1985) (see table 11-1). Thus, no
additional acute tests in laboratory water were needed. The site-specific FAV for this procedure
and cadmium was 1.4 fig/L for the data set when calculated by using the procedure described in
the national guidelines (Stephan et al., 1985). The CMC was derived by the following equation:
site-specific CMC = site-specific FAV/2.
The value obtained from this equation is 0.7 /tg/L. However, because the toxicity of
cadmium has been related to the hardness of the water, this relationship was taken into account by
using the method described in the national guidelines to adjust for hardness before the site-specific
CMC was calculated (table 11-1). The site-specific CMC for cadmium from the preceding
equation, adjusted for the hardness of the St. Louis River, was 0.8 /ig/L when derived by using the
recalculation procedure. (A hardness value of 55 mg/L as CaCO3 was used as an example for this
exercise and was determined from the lowest hardness measured monthly in this water over a year;
a more comprehensive monitoring data program would probably be needed to determine the most
appropriate value for this site on a seasonal basis.)
The recalculation procedure does not require testing to determine a site-specific CCC. A
site-specific FCV can be derived by dividing the site-specific FAV by the national final
acute/chronic ratio; if the national final acute/chronic ratio was not used to calculate the national
FCV, the national FCV then becomes the site-specific FCV. A national final acute/chronic ratio
for cadmium was not derived because enough chronic values were available to calculate an FCV
directly (U.S. EPA, 1985). Therefore, the site-specific FCV determined by using the recalculation
procedure is the same as the national FCV adjusted for the hardness of the St. Louis River, or 0.7
Indicator species procedure. The indicator species procedure is based on the determination
of a water effect ratio to account for the differences in the toxicity of cadmium in the St. Louis
River and in laboratory water due to physical and/or chemical characteristics of these waters.
Tests with two sensitive species, one fish and one invertebrate species, were required for this
procedure. A cladoceran (Simocephalus serrulatus) and rainbow trout (Oncorhynchus mykiss) were
selected as indicator species. Tests for each species were conducted in both types of water under
11-14
-------
Table 11-1. Recalculation Procedure: Acute Toxicity Data for Cadmium From National
Criteria Document for Resident Species of the St. Louis River (modified from
Spehar and Carlson, 1984a)a
Genus
Mean Acute Value
Rank fcig/L)
19 8,325
18 5,708
17 3,800
16 3,641
15 3,514
14 2,310
13 1,200
12 322.8
11 215.5
10 204.9
9 156.9
8 104.0
7 83.02
6 62.55
5 55.72
4 43.74
3 30.5
2 4.481
1 1.638
Species
Goldfish (Carassius auratus)
Channel catfish (Ictalurus punctatus)
Snail (Amnicola sp.)
Green sunfish (Lepomis cyanellus)
Pumpkinseed (Lepomis gibbosus)
Bluegill (Lepomis macrochirus)
White sucker (Catostamus commersonf)
Mayfly (Ephemerella grandis)
Midge (Chironomus sp.)
Mayfly (Paraleptophlebia praepedita)
Common carp (Cyprinus carpio)
Amphipod (Hyalella azteca)
Snail (Physa gyrina)
Snail (Aplexa hypnorum)
Cladoceran (Ceriodaphnia reticulatd)
Amphipod (Gammarus pseudolimnaeits)
Amphipod (Gammarus sp.)
Cladoceran (Daphnia pulex)
Cladoceran (Simacephalus serrulatus)
Cladoceran (Simocephdliis vetulus)
Fathead minnow (Pimephales promelas)
Coho salmon (Oncorhynchus kisutch)
Chinook salmon (Oncorhynchus tshawytscha)
Rainbow trout (Oncorhynchus mykiss)
Brown trout (Salmo trutta)
Species
Mean Acute Value
fog/L)
8,325
5,708
3,800
5,147
1,347
6,961
3,514
2,310
1,200
322.8
215.5
204.9
156.9
104.0
83.02
55.90
70.00
55.72
45.93
41.65
30.50
5.894
4.254
3.589
1.638
"Site-specific final acute value (calculated for a hardness of 50 mg/L from genus mean acute values) = 1.416 /tg/L.
Site-specific criterion maximum concentration = (1.416 /tg/L)/2 = 0.708 /tg/L (for a hardness of 50 mg/L).
In (site-specific criterion maximum intercept) = In (0.708) - [slope x In (50)] = - 4.758.
Site-specific criterion maximum concentration = e [1.128 (In hardness) - 4.758 ] = 0.788 /tg/L adjusted for a hardness of 55
mg/L.
11-15
-------
similar test conditions. A water-effect ratio was calculated by using the following equation:
water-effect ratio = site water LC50/lab water LC50.
Measured LC50 values for a toxicant must be significantly different (U.S. EPA, 1983;
Carlson et al., 1984) in the two waters for this procedure to be valid. If the values are not
different, then the national CMC becomes the site-specific CMC.
The 96-h LC50 values for cladocerans and rainbow trout were statistically different in site
water and laboratory water (table 11-2), and their water-effect ratios were similar. Consequently,
they could be used to calculate a site-specific CMC in the following equation: site-specific CMC
(7.4 jig/L) = geometric mean water-effect ratio (4.6) X national CMC (1.6 /tg/L). The national
CMC was adjusted for the hardness of the laboratory water (45 mg/L as CaCO3) before the site-
specific CMC was calculated.
The site-specific CCC for the indicator species procedure can be derived from three
optional methods (U.S. EPA, 1983; Carlson et al., 1984):
• by calculating (no testing required) the national acute/chronic ratio (if one is
present) and applying it to the site-specific FAV; if the national acute/chronic ratio
was not used to establish a national FCV, the national FCV may be used as the site-
specific FCV;
• by performing two acute and two chronic tests with both a fish and an invertebrate
species in site water and applying the resulting acute/chronic ratio to the site-
specific FAV;
• by conducting chronic tests with both a fish and an invertebrate species in both site
water and laboratory water and by applying the chronic water-effect ratio to the
national FCV.
When derived by using the first method, the site-specific CCC from the indicator species
procedure was 0.7 /ig/L, or the same as the national CCC. A final acute/chronic ratio was not
used to establish the national FCV, so the site-specific FCV is the same as the national value.
(Note that the FCVs and CCCs for both the national and site-specific values are the same in this
example because the FRVs and FPVs were not calculated in this exercise [see explanation above];
therefore, the FCVs become the CCCs for both procedures.)
The site-specific CCC calculated by using the second method was 0.3 /ig/L, based on a
geometric mean acute/chronic ratio of 50 (from tests performed in the St. Louis River water [table
11-3]) and the following equation: site-specific FCV = site-specific FAV/(site-specific final
acute/chronic ratio, or 14.8/50 = 0.3 /tg/L). The site-specific chronic value was obtained
by using a site-specific FAV of 14.8 (twice the site-specific CMC obtained from the indicator
species procedure).
11-16
-------
Table 11-2. Indicator Species Procedure: Acute Values (LC50) for Indicator Species Exposed
to Cadmium in St. Louis River and Reconstituted Water (modified from Spehar
and Carlson, 1984a)
Organism
Cladoceran (Simocephalus serrulatus)
Rainbow trout (Oncorhynchus mykiss)
St. Louis
River Water
fog/L)
123
10.2
Reconstituted
Water
fag/L)
24.5
2.3
Water-
Effect
Ratio8
5.0
4.4
"Geometric mean water-effect ratio = 4.6.
Table 11-3. Acute (LC50) and Chronic Values for Aquatic Organisms Exposed to Cadmium in
St. Louis River Water (Spehar and Carlson, 1984a)
Organism
Fathead minnow (Pimephales promelas)
Cladoceran (Ceriodaphnia reticulata)
Acute
Value
Otg/L)
1,830
129
Chronic
Value
(Atg/L)
18.9
5.0
Acute/
Chronic
Ratio8
97
26
aGeometric mean acute/chronic ratio = 50.
11-17
-------
The site-specific CCC calculated by using the third method (the chronic water-effect ratio),
based on studies for two species (fathead minnow and cladoceran), was determined to be 1.0
because the chronic values obtained from tests in site and laboratory water were not significantly
different (the chronic limits overlapped) (U.S. EPA, 1983) (table 11-4). Since the mean chronic
water-effect ratio was not different from 1.0, the site-specific FCV is the same as the national FCV
adjusted for the hardness of the St. Louis River, or 0.7 fig/L, using this method. Although tests
were not conducted specifically to obtain a site-specific chronic value by using this third method,
comparisons were made between present chronic tests in St. Louis River water and tests conducted
at different times with the same species hi Lake Superior water for use as an example of a chronic
water-effect ratio. According to the site-specific guidelines (U.S. EPA, 1983; Carlson et al.,
1984), tests in both waters should be run at the same time with organisms from the same
population and under the same test conditions.
Resident species procedure. The resident species procedure allows for modification of the
national criteria for cadmium on the basis of tests conducted in site water with a set of resident
species of the St. Louis River (table 11-5). Because the minimum data set requirements for
resident species were met at the St. Louis River site, substitute families were not needed. (Note:
A family in a phylum other than Arthropoda or Chordata [e.g., Rotifera, Annelida, Molluska, etc.]
was not included in this data set because it was not a requirement of the national guidelines at the
time these tests were conducted.) The site-specific FAV calculated by using the prescribed method
for resident species (Stephan et al., 1985) was 3.8 /ig/L. The resident species site-specific CMC
was calculated as follows: site-specific CMC = site-specific FAV/2, or 3.8/2 =1.9 jig/L.
The site-specific CCC of the resident species procedure was obtained by using the first two
methods described under the indicator species procedure, based on a site-specific FAV of 3.8
ftg/L. The site-specific FCVs for these methods were 0.7 and 0.1 /ig/L, respectively. The third
method should not be used to calculate a site-specific FCV using the resident species procedure
(U.S. EPA, 1983).
Summary of criteria calculation. Comparison of cadmium water quality criteria derived
from the national and site-specific procedures showed that criteria values varied according to the
procedure used (table 11-6). Site-specific criteria derived from the recalculation procedure were
similar or slightly lower than those of the national criteria. The lower acute site-specific criterion
was derived using a smaller number of genera to calculate the site-specific FAVs than was used to
calculate the national criterion (19 versus 44).
Site-specific criteria for cadmium derived from the indicator species procedure were higher
on an acute basis (CMC) than those derived from the national and recalculation procedures. This
result was expected because criteria derived from the indicator species procedure were based on a
water-effect ratio attributed to site water characteristics that decreased the toxicity of cadmium.
Cadmium was found to be less toxic to several species in St. Louis River water than in laboratory
water. On a chronic basis (CCC), site-specific criteria using this method were the same as or
slightly lower than the national criteria. The lower value was due to a large acute/chronic ratio
(50, see table 11-3) measured for two species in tests conducted in site water.
11-18
-------
Table 11-4. Chronic Toxicity Values of Two Species Exposed to Cadmium in St. Louis River
and Lake Superior Water (Spehar and Carlson, 1984a)
Organism
Fathead minnow
(Pimephales promelas)
Cladoceran
(Ceriodaphnia reticulatd)
St. Louis River
Water
18.9
(13-26)b
5.0
(3.4-7.2)
Chronic Value Otg/L)
Lake Superior
Water
13°
(9-18)
5.2d
(3.6-7.5)
Chronic Water-
Effect Ratio*
1.0
1.0
aChronic water-effect ratios are 1.0 because values in site and laboratory are not different (chronic
limits overlap).
bChronic limits.
°Data from Carlson et al. (unpublished manuscript).
dData from D.I. Mount (unpublished manuscript).
Table 11-5. Resident Species Procedure: Minimum Data Set of Resident Aquatic Species
Exposed to Cadmium in St. Louis River Water (Spehar and Carlson, 1984a)a
Rank
8
7
6
5
4
3
2
1
Organism
Bluegill (Lepomis macrochirus)
Channel catfish (Ictalurus punctatus)
Fathead minnow (Pimephales promelas)
Mayfly (Paraleptophlebia praepeditd)
Amphipod (Hyalella azteca)
Cladoceran (Simocephalus serrulatus)
Amphipod (Gammarus pseudolimnaeus)
Rainbow trout (Oncorhynchus mykiss)
LC50 (^g/L)
8,800
7,900
3,390
449
285
123
54
10
aSite-specific final acute value for this resident species data set = 3.8.
11-19
-------
Table 11-6. Cadmium Water Quality Criteria Derived From National and Site-Specific
Procedures (modified from Spehar and Carlson, 1984a)
Criterion
Derivation
Procedure
National
Criterion Maximum
Concentration (ug/L)
2.0a
Criterion Continuous
Concentration (pg/L)
0.7a
Site-specific
recalculation 0.8a 0.7
indicator 7.4b-° 0.7(a)d
0.3(b)
0.7(c)
Resident 1.9 0.7(a)
"Adjusted for a water hardness of the St. Louis River of 55 mg/L as CaCO3.
"The national data base containing 44 genera was used in this calculation.
°A national criterion of 1.6 ug Cd/L (adjusted for a hardness of 45 mg/L hardness [as CaCO3])
was used in this procedure.
dJLetters in parentheses indicate optional methods used for calculating the site-specific criterion
continuous concentration.
11-20
-------
The large acute/chronic ratio indicates that chronic toxicity was not greatly affected by
water quality at this site. Chronic values were similar to those from tests with similar species that
were conducted in laboratory water (U.S. EPA, 1985). However, acute values were higher than
previously reported, indicating that the water quality characteristics of this site decreased toxicity
on a short-term basis, probably through mechanisms affecting the bioavailability of this chemical.
The site-specific CMC derived from the resident species procedure was approximately two
times higher than the criterion derived from the recalculation procedure, but was lower than that
obtained from the indicator species procedure. This result also was expected and was attributed to
a combination of the use of a limited data base of eight species (which resulted in a lower
criterion) and the use of site water tests (which raised the criterion due to the mitigating effects of
the site water). In addition, the site-specific CCC derived from the chronic water-effect ratio
method (the third method under the indicator species procedure) was the same as the national FCV
(adjusted for the hardness of the laboratory water) because the mean chronic water-effect ratio was
not significantly different from 1.0. The low CCC obtained from the second method for both the
indicator and resident species procedures was attributed to the large site-specific acute/chronic ratio
(50) used in the calculation of the site-specific FCV.
Although all of the procedures from the site-specific guidelines (U.S. EPA, 1983) were
tested, only one would probably be used in an actual site criteria modification. If species
sensitivity was the important factor (for example, where species at a particular site are more or less
sensitive than those used to derive the national criteria), then the recalculation procedure would be
the recommended approach because it would require no testing. The indicator species procedure
would be appropriate when water quality at a site may mitigate the toxicity of a chemical and the
resident species are similar to those used to calculate the national criteria. This is especially true
for metals like cadmium, for which biological availability and/or toxicity are significantly affected
by variations in water quality characteristics of the site water. When both species sensitivity and
water quality are important considerations for a particular site, the resident species procedure
would be the best approach because it is designed to account for differences due to both factors.
Recommended Site-Specific Procedure. For the St. Louis River and Estuary, the indicator
species procedure would be the recommended approach for deriving site-specific criteria for
cadmium. The present study showed that cadmium was less toxic in site water than in laboratory
water, which resulted in a water-effect ratio. This ratio was designed to account for water quality
effects, and its use with the national CMC should provide a site-specific CMC that would
adequately protect the resident aquatic species. The recalculation procedure would not be
appropriate in this case because there was no real difference between the sensitivity range of
species represented in the national data set and species found at this site. This rationale would also
apply to the resident species procedure, which in part accounts for species sensitivity. The resident
species procedure could be used for this site, but additional testing in site water would be required.
11-21
-------
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
• This case study illustrates how a site-specific tool (water quality criteria) can be
developed for evaluating potential effects,
limitations include:
•Acute toxicity of cadmium to 25 aquatic species indigenous to the St. Louis River
is ranked for the recalculation procedure. Rankings are based on sensitivity in
laboratory studies rather than on relative importance in the river system. For the
recalculation procedure, lethality is the only test endpoint.
^ The hardness values used for the test site require a larger data base.
• The stress-response assessment is limited to toxicological effects of site water on
selected surrogate species, with no consideration of higher dimensional ecological
phenomena.
11.3.3. Analysis: Characterization of Exposure
A study by Hammermeister et al. (1983) provided monitoring data for cadmium from a
relatively clean water upstream sampling station and a downstream sampling station below major
municipal and industrial wastewater discharges into the St. Louis River Estuary. This study was
used as an example to provide an exposure concentration for use in this exercise. A total of six
grab samples per station was taken over a 6-month period beginning on January 29, 1983, and
ending on June 23, 1983. Cadmium was found to be present at 1.0 /ig/L, a concentration that
slightly exceeded both the national and site-specific CCCs (table 11-6). Cadmium concentrations
above these criteria were measured in only one of the downstream samples.
The averages of monthly acute cadmium toxicity values for larval fathead minnows exposed
in St. Louis River water (from the present procedures) were approximately five and two times
higher than values obtained from concurrent cadmium exposures conducted in reconstituted and
Lake Superior water, respectively, according to studies by Spehar and Carlson (1984a, b). These
studies reported similar toxicity differences in four of five juvenile species exposed to St. Louis
River water. The findings indicate that physical and/or chemical characteristics of the St. Louis
River water reduced the toxicity of cadmium on an acute basis from what was observed in
laboratory water.
The acute cadmium toxicity values calculated from the above larval fish exposures in river
water varied by a factor of three and increased with increases in the concentration of suspended
11-22
-------
solids, total organic carbon, turbidity, and dissolved solids. Linear regression correlation
coefficients for acute toxicity were calculated, and these parameters were 0.58, 0.60, 0.68, and
0.77, respectively. The UC^0 values for tests conducted hi reconstituted and Lake Superior water
varied by a factor of less than 2. The larger variation hi values obtained from tests conducted in
site water was attributed to high and low stream flows, which influenced water quality factors
throughout the year. The large degree of binding or complexing of cadmium that occurred during
times when concentrations of particulates in this water were highest was the apparent cause of
reduced cadmium toxicity. Although this effect on acute toxicity was not large in the present tests,
larger variations in toxicity may occur in streams where paniculate loads change significantly
during different times of the year. The frequency of testing required to perform risk assessments
by using the site-specific approach will depend on this seasonal variability.
The effect of seasonality on the physical and chemical characteristics of water and
subsequent effects on biological availability and/or toxicity of cadmium justify the use of seasonally
dependent site-specific criteria for the St. Louis River Estuary. A major implication of seasonally
dependent criteria is whether the most sensitive time of the year coincides with the time when the
flow is the basis for waste treatment facilities design or National Pollutant Discharge Elimination
System (NPDES) permits. That is, if the physical and chemical characteristics of the water during
low-flow seasons increase the biological availability and/or toxicity of cadmium, the permit
limitations may be more restrictive than if the converse relationship were to apply.
The national and site-specific water quality criteria concentrations stated above contain
duration (averaging period) and frequency (or average recurrence interval) periods (to account for
excesses) that are based on biological, ecological, and toxicological data and are designed to
protect aquatic organisms and their uses from unacceptable effects (Stephan et. al., 1985).
Numeric criteria are used as the basis for determining wasteload allocations (WLAs) for point
sources of contaminants or load allocations (LAs) for nonpoint sources (U.S. EPA, 1991). Limits
on wastewater loads are set and nonpoint source allocations are established so that receiving water
concentrations (i.e., for cadmium in the St. Louis River Estuary) do not exceed water quality
criteria. Information on conducting exposure and wasteload allocation procedures are detailed in
EPA's Technical Support Document for Water Quality-Based Toxics Control (U.S. EPA, 1991).
Comments on Analysis: Characterization of Exposure
Limitations of the case study include:
+The case study does not focus on exposure.
• The exposure assessment includes the magnitude, duration, and frequency of
exposure limits, but includes no implicit discussion of probable exposure to
populations in the St. Louis River.
11-23
-------
11.3.4. Risk Characterization
The indicator species procedure was selected as the appropriate approach to use in
conducting a risk assessment for this site. This procedure takes into account factors that would
affect the bioavailability and/or the toxicity of the identified stressor (cadmium). However,
because the indicator species procedure is used as an example, the actual data have limited
practical use. To fully characterize risks to ecological components at this particular site, more
information would be needed on exposure to all stressors and the specific components (i.e.,
aquatic, terrestrial, and other species such as wildlife) that are of concern within the scope of the
study. The water quality criteria (including site-specific) approach is only intended to protect
aquatic life and their uses (but simultaneously may protect wildlife to some degree) and is only one
approach used by EPA to assess aquatic effects. This approach needs to be used in combination
with the whole-effluent approach used for assessing the impact of mixtures of chemicals and with
other approaches used to develop sediment, wildlife, and biological criteria.
The ambient concentration of 1.0 /ig/L observed for cadmium in the St. Louis River
Estuary was slightly above the site-specific values based on chronic tests (CCC), but was generally
less than the site-specific criteria values based on acute tests (CMC). The site-specific CMC,
based on the indicator species procedure, was higher than the national criteria because the site
water decreased cadmium toxicity by decreasing its bioavailability to the resident species in the
estuary. Because the ambient concentration of cadmium at the site was lower than the site-specific
CMC, it would appear that cadmium would not pose a risk to the resident aquatic species on a
short-term basis. However, because ambient concentrations are above site-specific criteria based
on chronic tests, some sensitive resident species may be affected on a long-term basis. Additional
risk characterization techniques would be needed to determine wasteloads from point sources or
loadings from nonpoint sources to see if cadmium, mixtures of chemicals that include cadmium, or
other stressors to the system do indeed cause in-stream toxicity when criteria are exceeded. WLA
and LA models are available, as noted above, to conduct such characterizations. After these types
of analyses are conducted, specific toxin limits can be defined by using models to predict design
flows that should prevent ambient exposures from exceeding criteria levels for longer durations or
at more frequent intervals than allowed.
Although some laboratory microcosm/mesocosm tests and field studies support the
approaches of this case study, additional research is recommended to revise and improve the
guidelines that form the underpinnings of the criteria derivation process. Recommendations for
making these revisions and for proposed research to improve EPA's guidelines are discussed in
U.S. EPA (1989). A refinement of these suggested revisions is currently being made by EPA as
recommended by a workshop held in December 1990 to revise the National Water Quality Criteria
Guidelines.
11-24
-------
Comments on Risk Characterization
Strengths of the case study include:
• Water quality criteria (WQC) are valuable impact assessment tools that can be
considered for risk assessments when used with appropriate hypotheses and exposure
scenarios. Although the case study is not an example of a risk assessment, the
development of a benchmark (criterion) can be used in a quotient approach to risk
assessment.
•This case study provides a good example of an ecological effects stressor-response
assessment.
Limitations include:
• Sole reliance on water quality criteria compliance might be misleading. One
cannot assume that ecosystem protection is "acceptable" if WQC are met, due to
interactions or unknown exposures. Ecological field data may be necessary to verify
or validate toxicological phenomena. Biosurveys are important to bridge the gap
between the laboratory and field environments.
• A site-specific criterion maximum concentration of 7.4 fig/L is justified by the
study data, but the relative risk and uncertainty are characterized rhetorically rather
than quantitatively.
•A toxicity-based approach to water quality assumes ecological risk to populations
and communities, but relies solely on organismic endpoints that do not address all
ecological concerns. Without some type of community assessment, ecological risk
and damage may be improperly assessed.
General comments:
*The development of ecological (i.e., biological) criteria in parallel with chemical-
specific criteria would help validate predictions and generate biological endpoints
that can be measured in the field.
•Exposure-response curves, rather than a single endpoint such as the no observed
effects level (NOEL), could be used and an acceptable response level on the curve
selected to set criteria.
+The effects of water quality characteristics (e.g., pH, temperature, organic
content) on the bioavailability of pollutants, particularly metals, are only partially
understood. Similarly, predictions oftoxidty and persistence can be inaccurate.
Therefore, laboratory techniques are needed to better simulate ambient conditions
and field methods are needed to evaluate the predictions.
11-25
-------
Comments on Risk Characterization (continued)
•Duration and frequency criteria should be based on effects that occur as a
consequence of exposures similar to actual excess levels rather than worst-case
scenarios. Better information is needed on actual "time-till-effect" relationships for
specific pollutants and recovery times for ecosystems responding to moderate as well
as catastrophic effects.
•Better techniques are needed for integrating pollutant-specific assessments.
Mixtures of pollutants may interact in complex ways, or they may simply exhibit the
effects of the predominant constituents). The tools available (i.e., whole-effluent
approach) can be used to evaluate these interactions better.
11-26
-------
11.4. REFERENCES
Calamari, D.; Marchetti, R.; Vailati, G. (1980) Influence of water hardness on cadmium toxicity
to Salmo gairdneri (Rich.). Water Res. 14:1421-1426.
Carlson, A.R.; Brungs, W.A.; Chapman, G.A.; Hansen, DJ. (1984) Guidelines for deriving
numerical aquatic site-specific water quality by modifying national criteria. EPA 600/
3-84/099. Available from National Technical Information Service, Springfield,
VA.
Geisy, J.P., Jr.; Leversee, G.J.; Williams, D.R. (1977) Effects of naturally occurring aquatic
organic fractions on cadmium toxicity to Simocephalus serrulatus (Daphnidae) and
Gambusia affiniis (Poeciliidae). Water Res. 11:1013-1020.
Hammermeister, D.; Northcott, C.; Brook, L.; Call, D. (1983) Results of analysis of St. Louis
River water sampled from an upstream and downstream site, January-June 1983. Center
for Lake Superior Environmental Studies, University of Wisconsin-Superior, Superior, WI.
U.S. EPA Cooperative Agreement No. CR809234040.
Reid, J.D.; McDuffie, B. (1981) Sorption of trace cadmium on clay minerals and river sediments:
effects of pH and Cd(II) concentrations in a synthetic river water. Water Air Soil Pollut.
15:375-386.
Spehar, R.L.; Carlson, A.R. (1984a) Derivation of site-specific water quality criteria for cadmium
and the St. Louis River Basin, Duluth, Minnesota. Available from National Technical
Information Service, Springfield, VA. Order No. PB84-153196.
Spehar, R.L.; Carlson, A.R. (1984b) Derivation of site-specific water quality criteria for cadmium
and the St. Louis River Basin, Duluth, Minnesota. Environ. Toxicol Chem. 3:651-665.
Stephan, C.E.; Mount, D.I.; Hansen, D.J.; Gentile, G.A.; Chapman, G.A.; Brungs, W.A. (1985)
Guidelines for deriving numerical national water quality criteria for the protection of
aquatic organisms and their uses. U.S. Environmental Research Laboratories, Duluth, MN;
Narragansett, RI; and Corvallis, OR. Available from National Technical Information
Service, Springfield, VA. PB85-227049.
U.S. Environmental Protection Agency. (1980a) Water quality criteria documents availability.
Federal Register 45:79318-79379.
U.S. Environmental Protection Agency. (1980b) Treatability manual. Vol. 11. Industrial
description. PB80-223068.
U.S. Environmental Protection Agency. (1983) Water quality standards handbook. Guidelines for
deriving site-specific water quality criteria. Office of Water Regulations and Standards,
Washington, DC.
11-27
-------
U.S. Environmental Protection Agency. (1985) Ambient water quality criteria for cadmium—1984.
EPA 440/5-84/032. Available from National Technical Information Service, Springfield,
VA. PB85-227031.
U.S. Environmental Protection Agency. (1989) Status of the development of water quality
criteria and advisories. In: Water quality standards for the 21st century, proceedings of a
national conference. EPA Office of Water. Dallas, TX. March 1-3, 1989. pp. 163-169.
U.S. Environmental Protection Agency. (1991) Technical support document for water quality-based
toxics control. Office of Water. EPA 505/2-90/001. Available from National Technical
Information Service, Springfield, VA. PB91-127415.
11-28
-------
APPENDIX A
WATER QUALITY CRITERIA FOR AQUATIC LIFE
11-A1
-------
APPENDIX A
Water Quality Criteria for Aquatic Life
available from the
National Technical Information Service (NTIS)
5285 Port Royal Road
Springfield, VA 22161
(703-487-4650)
Note: Multiple entries are given for those pollutants for which corrections
and/or revised criteria have been published.
Pollutant
Acenaphchene
Acrolein
Acrylonitrile
Aldrin/Dieldrin
Ammonia
Antimony
Arsenic
Asbestos
Benzene
Benzidine
Beryllium
Cadmium
Federal
Register
Notice*
1
1
1
1
4
1
1
4
1
1
1
1
1
4
EPA Number
EPA 440/5-80-015
EPA 440/5-80-016
EPA 440/5-80-017
EPA 440/5-80-019
EPA 440/5-85-001
EPA 440/5-80-020
EPA 440/5-80-021
EPA 440/5-84-033
EPA 440/5-80-022
EPA 440/5-80-018
EPA 440/5-80-023
EPA 440/5-80-024
EPA 440/5-80-025
EPA 440/5-84-032
NTIS Number
PB81-117269
PB81-117277
PB81-117285
PB81-117301
PB85-227114
PB81-117319
PB81-117327
PB85-227445
PB81-117335
PB81-117293
PB81-117343
PB81-117350
PB81-117368
PB85-227031
11-A2
-------
Pollutant
Acenaphthene
Acrolein
Carbon tetrachloride
Chlordane
Chlorinated benzenes
Chlorinated ethanes
Chloroalkyl ethers
Chlo'rinated naphthalene
Chlorinated phenols
Chlorine
Chloroform
2-Chlorophenol
Chlorpyrifos
Chromium
Copper
Cyanide
DDT
Dichlorobenzenes
Dichlorobenzidine
Federal
Register
Notice*
1
1
1
1
1
1
1
1
1
4
1
1 .
6
1
4
1
4
1
1
1
1
EPA Number
EPA 440/5-80-015
EPA 440/5-80-016
EPA 440/5-80-026
EPA 440/5-80-027
EPA 440/5-80-028
EPA 440/5-80-029
EPA 440/5-80-030
EPA 440/5-80-031
EPA 440/5-80-032
• EPA 440/5-84-030
EPA 440/5-80-033
EPA 440/5-80-034
EPA 440/5-86-005
EPA 440/5-80-035
EPA 440/5-84-029
EPA 440/5-80-036
EPA 440/5-84-031
EPA 440/5-80-037
EPA 440/5-84-028
EPA 440/5-80-038
EPA 440/5-80-039
EPA 440/5-80-040
NT IS Number
P381-117269
P381-117277
PB81-117376
PB81-117384
PB81-117392
P381-117400
PB81-117418
PB81-117426
PB81-117434
PB85-227429
PB81-117442
PB81-117459
PB87-105267
PB81-117467
PB85-227478
PB81-117475
PB85-227023
PB81-117483
PB85-227460
PB81-117491
PB81-117509
PB81-117517
11-A3
-------
Pollutant
Acenaphthene
Acrolein
Dichloroethylenes
2 , 4-Dichlorophenol
Dichloropropanes/Dichloropropenes
2 , 4-Dime Chylphenol
Dinitrotoluene
Diphenylhydrazine
Dissolved oxygen
Endosulfan
Endrin
Ethylbenzene
Fluor an thene
Haloethers
Halomethanes
Heptachlor
Hexachloroobutadiene
Hexachlorocyclohexane
Hexachlorocyclopentadiene
Isophorone
Federal
Register
Notice*
1
1
1
1
1
1
1
1
5
1
1
1
1
1
1
1
1
1
1
1
EPA Number
EPA 440/5-80-015
EPA 440/5-80-016
EPA 440/5-80-041
EPA 440/5-80-042
EPA 440/5-80-043
EPA 440/5-80-044
EPA 440/5-80-045
EPA 440/5-80-062
EPA 440/5-86-003
EPA 440/5-80-046
EPA 440/5-80-047
EPA 440/5-80-048
EPA 440/5-80-049
EPA 440/5-80-050
EPA 440/5-80-051
EPA 440/5-80-052
EPA 440/5 -80 -05 3
EPA 440/5-80-054
EPA 400/5-80-055
EPA 440/5-80-056
NTIS Number
PB81-117269
PB81-117277
PB81-117525
PB81-117533
.PB81-117541
PB81-117588
PB81-117566
PB81-117731
PB86-208253
PB81-117574
PB81-117582
PB81-117590
PB81-117608
PB81-117616
PB81-117624
PB81-117632
PB81-117640
PB81-117657
PB81-117655
PB81-117673
11-A4
-------
Pollutant
Acenaphthene
Acrolein
Lead
Mercury
Naphthalene
Nickel
Nitrobenzene
Nitrophenols
Nitrosamines
Parathion
Pentachlorophenol
Phenol
Phthalate esters
Polychlorinated biphenyls (PCBs)
Polynuclear aromatic hydrocarbons
Selenium
Silver
Federal
Register
Notice'
1
1
1
4
1
2
4
1
1
6
1
1
1
6
1
6
1
1
1
1
1
1
EPA Number
EPA 440/5-80-015
EPA 440/5-80-016
EPA 440/5-80-057
EPA 440/5-84-027
EPA 440/5-80-058
Correction
EPA 440/5-84-026
EPA 440/5-80-059
EPA 440/5-80-060
EPA 440/5-86-004
EPA 440/5-80-061
EPA 440/5-80-063
EPA 440/5-80-064
EPA 440/5-86-007
EPA 440/5-80-065
EPA 440/5-86-009
EPA 440/5-80-066
EPA 440/5-80-067
EPA 440/5-80-068
EPA 440/5-80-069
EPA 440/5-80-070
EPA 440/5-80-071
NTIS Number
PB8 1-11 7269
PB81-117277
PB81-117681
PB85-227437
PB81-117699
--
PB85-227452
PB81-117707
PB81-117715
PB87-105359
PB81-117723
PB81-117749
PB81-117756
PB87-105383
PB81-117764
PB87-105391
PB81-117772
PB81-117780 •
PB81-117798
PB81-117806
PB81-117814
PB81-117822
11-A5
-------
Pollutant
Acenaphthene
Acrolein
Tetrachloroethylene
2,3,7 , 8-Tetrachlorodibenzo-p-dioxin
Thallium
Toluene
Toxaphene
Trichloroethylene
Vinyl chloride
Zinc
Federal
Register
Notice*
1
1
1
3
1
1
1
6
1
1
1
7
EPA Number
EPA 440/5-80-015
EPA 440/5-80-016
EPA 440/5-80-073
EPA 440/5-84-007
EPA 440/5-80-074
EPA 440/5-80-075
EPA 440/5-80-076
EPA 440/5-86-006
EPA 440/5-80-077
EPA 440/5-80-078
EPA 440/5-80-079
EPA 440/5-87-003
NTIS Number
PB81-117269
PB81-117277
PB81-117830
PB81-117848
PB81-17855
PB81-117863
PB87-105375
PB81-117871
PB81-117889
PB81-117897
PB87-153581
Guidelines for Deriving Numerical
National Water Quality Criteria for
the Protection of Aquatic Organisms
and Their Uses.
PB85-227049
^Federal Register Notices:
1. Federal Register, Vol.
2. Federal Register, Vol.
3. Federal Register, Vol.
4. Federal Register, Vol.
5. Federal Register, Vol.
6. Federal Register, Vol.
7. Federal Register, Vol.
45, pp. 79318-79379, November 28, 1980
46, pp. 40919, August 13, 1981
49, pp. 5831-5832, February 15, ;1984
50, pp.. 30784-30796, July 29, 1985
51, pp. 22978, June 24, 1986
51, pp. 43665-43667, December 3, 1986
52, pp. 6213, March 2, 1987
11-A6
-------
SECTION TWELVE
ECOLOGICAL RISK ASSESSMENT CASE STUDY:
MODELING FUTURE LOSSES OF BOTTOMLAND FOREST WETLANDS AND
CHANGES IN WILDLIFE HABITAT WITHIN A LOUISIANA BASIN
-------
AUTHORS AND REVIEWERS
AUTHORS
Michael S. Brody
Office of Policy, Planning, and Evaluation
U.S. Environmental Protection Agency
Washington, DC
Yvonne Valette
Region 6
U.S. Environmental Protection Agency
Dallas, TX
REVIEWERS
Clifford Hupp (Lead Reviewer)
Water Resources Division
U.S. Geological Survey
Reston, VA
Joel Brown
Department of Biological Sciences
University of Illinois
Chicago, IL
Arthur L. Buikema
Biology Department
Virginia Polytechnic Institute
and State University
Blacksburg, VA
Michael E. Troyer
Office of Technology Transfer and
Regulatory Support
U.S. Environmental Protection Agency
Washington, DC
Carolyn L. Fordham
Terra Technologies
Golden, CO
Christian E. Grue
Washington Cooperative Fish and
Wildlife Research Unit
U.S. Fish and Wildlife Service
Seattle, WA
Ronald J. Kendall
Institute of Wildlife and
Environmental Toxicology
Clemson University
Clemson, SC
12-2
-------
^ CONTENTS
ABSTRACT 12-6
12.1. RISK ASSESSMENT APPROACH 12-7
12.2. STATUTORY AND REGULATORY BACKGROUND 12-7
12.3. CASE STUDY DESCRIPTION 12-9
12.3.1. Problem Formulation 12-9
12.3.2. Analysis: Characterization of Ecological Effects 12-17
12.3.3. Analysis: Characterization of Exposure 12-22
12.3.4. Risk Characterization . 12-23
12.4. RECOMMENDATIONS AND FUTURE RESEARCH NEEDS 12-34
12.5. REFERENCES 12-36
12-3
-------
LIST OF FIGURES
Figure 12-1. Structure of analysis for modeling losses of bottomland forest wetlands
and changes in wildlife habitat within a Louisiana basin 12-8
Figure 12-2. Location of Lake Verret Basin 12-10
Figure 12-3. Diagram of dynamics contained in FORFLO 12-21
Figure 12*4. Succession in wet bottomland hardwood site when a subsidence rate
of 0.5 cm/yr is assumed 12-25
Succession in the canopy and subcanopy of the dry bottomland hardwood
site when a subsidence rate of 0.5 cm/yr is assumed 12-27
Succession and production in the cypress-tupelo swamp when a
subsidence rate of 0.5 cm/yr is assumed 12-28
Figure 12-5.
Figure 12-6.
Figure 12-7. Comparison of FORFLO simulation results with field observations 12-32
LIST OF TABLES
Table 12-1. Sample Selection Matrix for Wildlife Species 12-13
Table 12-2. Tree Species Observed in Three Simulated Sites . . . 12-14
Table 12-3. List of All Variables Required hi Five HSI Models 12-16
Table 12-4. Waterlogging Tolerance Ranking of Major Tree Species
in Lake Verret Basin 12-19
Table 12-5. Initial Tree Composition and Density for Three Simulated Sites . 12-24
Table 12-6. Current and Future Habitat Suitability (HSI) Indices for Five
Species Evaluated 12-29
Table 12-7. Current and Future Suitability Indices (SI) for Variables
of Five HSI Models 12-30
12-4
-------
LIST OF ACRONYMS
ATP adenosine triphosphate
COE U.S. Army Corps of Engineers
DBH diameter at breast height
EIS Environmental Impact Statement
EPA U.S. Environmental Protection Agency
HEP Habitat Evaluation Procedure
HSI Habitat Suitability Index
HU Habitat Units
MSL mean sea level
NEPA National Environmental Policy Act of 1969
NOAA National Oceanic and Atmospheric Administration
> ;
RCRA Resource Conservation and Recovery Act
SAB Science Advisory Board
i
SI Suitability Indices
USFWS U.S. Fish and Wildlife Service
USGS U.S. Geological Survey
12-5
-------
ABSTRACT
The Lake Verret Basin, a part of the Atchafalaya River floodplain in southern Louisiana, is
composed largely of bottomland hardwoods and cypress-tupelo swamps. These forested wetlands
tolerate variable flood durations and provide high-quality habitat (e.g., food and shelter) for many
wildlife species. However, if hydrologic changes such as long-term excess flooding become too
great, these habitats can become nonforested marshes or open-water areas less suitable for wildlife
species. Given the Administration's policy of "no net loss of wetlands" and the difficulty of
replacing forested wetlands, any future destruction or degradation of these habitats should be
avoided. Artificial levees first erected in the Atchafalaya River floodplain in the 1930s have
deprived the Lake Verret Basin of sediment deposition, contributing to a net subsidence rate of
25 cm every 50 years. Subsequently, increased backwater flood heights and durations have caused
the bottomland forest to succeed toward cypress-tupelo communities and have reduced opportunities
for lower-elevation cypress-tupelo to regenerate. In this case study, an analysis is conducted on the
ongoing processes hi the Lake Verret Basin. Although construction of more levees has been
proposed, this analysis predicts future conditions given current subsidence problems from past
levee construction. Thus, this analysis serves as a set of baseline conditions with which to
compare the changes in drainage and water flows that future levee projects would cause. This case
study is based on work by Brody et al. (1989) and Conner and Brody (1989) to add the effects of
subsidence into FORFLO, a bottomland forest succession model, and to forecast temporal and
spatial forest community impacts in swamp areas and wet and dry bottomlands within the Lake
Verret Basin. Model outputs of FORFLO are coupled with Habitat Suitability Index models to
determine present and future wildlife habitat values for two species of birds and three species of
mammals.
12-6
-------
12.1. RISK ASSESSMENT APPROACH
This case study includes all major components of an ecological risk assessment (figure
12-1). Information originally provided by Brody et al. (1989) and Conner and Brody (1989) on the
incorporation of FORFLO and Habitat Suitability Index (HSI) models allowed synthesis of
previously collected impact assessment data into a format useful for assessing the probability of
ecological risk as it pertains to modifications in hydrology and subsequent changes in habitat.
Effects of other anthropogenic stressors (e.g., chemical pollutants) are not considered.
Because of the interaction of environmental factors in influencing forested wetland biology
and succession, the FORFLO bottomland forest succession model (Pearlstine et al., 1985; Brody
and Pendleton, 1987; Brody et al., 1989) was used for predicting the future plant communities in
the Lake Verret Basin. The succession model outputs of FORFLO were used as input values to
HSI models for assessing future value of the habitat to wildlife. FORFLO predicts tree species
presence and abundance, individual tree size, canopy closure, flood Duration, and other habitat
measures that affect wildlife. It provides these outputs on an annual oasis for as many years as
required for the impact analysis. As such, a mechanism is available to quantify a field assessment
of the current habitat values (HSI models), to quantify future habitat conditions (FORFLO), and to
assess future habitat values (HSI models).
12.2. STATUTORY AND REGULATORY BACKGROUND
The National Environmental Policy Act of 1969 (NEPA) establishes a national
environmental policy and goals for the protection, maintenance, and enhancement of the
environment and provides a process for implementing these goals within federal agencies, including
the U.S. Environmental Protection Agency (EPA). The NEPA process consists of an assessment
of the environmental effects of federal projects including all alternatives and impacts. Such
assessments (and management decisions) appear to be increasingly risk based since the risk
assessment approach can help identify the relative efficiency and effectiveness of different risk
reduction options (Stakhiv, 1986; Russell and Gruber, 1987).
Under the Clean Water Act, EPA is responsible for restoring and maintaining the physical,
chemical, and biological integrity of the nation's waters, including wetlands. Although no
comprehensive wetlands management program exists, Section 404 of the Act provides the primary
legislative authority behind federal efforts to protect wetland use; however, the level of protection
is limited to the regulation of "discharges" of dredged or fill materials into waters of the United
States. Wetlands continue to be altered in many other ways, including excavation, draining,
clearing, flooding, and other water diversions. It is estimated that Section 404 does not cover
about 80 percent of the nation's wetland losses (Office of Technology Assessment, 1984).
Currently, it appears that the scope of federal wetlands protection will change as a result of
technical and policy revisions to the 1989 Federal Manual for Identifying and Delineating
Jurisdictional Wetlands and Congress' reauthorization of the Clean Water Act. Whether these
changes will enhance or degrade America's wetland resources is a subject of current debate.
The U.S. Fish and Wildlife Service (USFWS) developed HSI models (USFWS, 1981) to
define a habitat-based approach for assessing the environmental impacts of federal water projects
12-7
-------
Figure 12-1. Structure of Analysis for
Modeling Losses of Bottomland Forest Wetlands
PROBLEM FORMULATION
Stressors: changes in water level elevations due to
subsidence and other factors.
Ecological Components: tree species in bottomland
habitat and five wildlife species (three mammals; two birds)
Endpoints: assessment endpoint was physical alteration
or change in the forest community and associated habitat
value. Measurement endpoints included the vegetation,
hydrologic, and other input data required for the FORFLO
and HSImodels.
I
f
ANALYSIS
Characterization of
Exposure
A baseline change in water
elevation was estimated
from other studies for
regional subsidence rates.
Characterization of
Ecological Effects
The FORFLO model was
used to estimate changes in
forest vegetation; the HSI
models were used to relate
these changes to effects
on habitat for five species.
RISK CHARACTERIZATION
Changes (risks) to forest vegetation were estimated
under a baseline exposure regime represented by
regional subsidence.
Changes (risks) to wildlife were estimated for
the future conditions forecast for forest vegetation.
The uncertainties associated with the use of the
FORFLO and HSI models were described.
12-8
-------
on fish and wildlife resources (i.e., for Environmental Impact Statements [EISs] required by NEPA
and other reports required by the Fish and Wildlife Coordination Act). To evaluate the effects of
hydrologic changes on forested wetland habitats, the USFWS developed a bottomland forest
succession model, FORFLO (Pearlstine et al., 1985), that simulates the growth, reproduction, and
competition of a mixed-tree-species forest stand.
In 1987, at the request of the EPA Administrator, a National Wetlands Policy Forum
convened to suggest ways to improve wetland regulation and management. In its final report,
Protecting America's Wetlands, the Forum recommended "...no overall net loss of the nation's
remaining wetlands base, as defined by acreage and function." This policy has been endorsed by
EPA. Furthermore, several of the Forum's recommendations have reappeared in EPA testimony to
Congress during Clean Water Act reauthorization hearings in 1991. At present, EPA lacks risk
assessment and management approaches for considering physical habitat alteration and biological
diversity. The Science Advisory Board's (SAB's) report Reducing Risk: Setting Priorities and
Strategies for Environmental Protection (U.S. EPA, 1990a) states that wetlands have
"extraordinary value" and that the alteration and destruction of natural habitats (including wetlands)
pose a high risk to the natural ecology and human welfare of very large areas. The SAB also
considers the related issue of species extinction and overall loss of biological diversity (including
genetic diversity) as a high-risk problem. As a result, the SAB has called on EPA managers to
direct their efforts toward reducing risks posed by these and other "critical environmental
problems" (U.S. EPA, 1990b).
12.3. CASE STUDY DESCRIPTION
12.3.1. Problem Formulation
Site Description. The Lake Verret Basin, a part of the Atchafalaya floodplain in south-
central Louisiana, is bounded by the East Atchafalaya Basin Protection Levee on the west, the
natural levee ridges of the Mississippi River and Bayou LaFourche on the east, Bayou Plaquemine
to the north, and Louisiana Highway 20 and U.S. Highway 90 to the south (see figure 12-2). The
watershed occupies approximately 99,000 ha, of which 48 percent are seasonally flooded
bottomland hardwood areas and cypress-tupelo swamps (Soil Conservation Service, 1978). The
basin lies entirely within the Mississippi River Deltaic Plain, and the land is low and flat with
elevations ranging from 1 m mean sea level (MSL) in the northern portion of the basin to less than
1 m MSL in the southern portion.
Stressors. Nature and human activities are recognized as major contributors to declining
wetland resources. In Louisiana, naturally occurring wetlands loss results from subsidence, rising
sea level, normal wave action, pounding storm surges, and saltwater intrusion into freshwater
areas. Human causes of wetlands loss include levee construction along the Mississippi and
Atchafalaya Rivers, dredging and soil disposal, drainage, mineral extraction, wave action from
vessel traffic, and agricultural, urban, and industrial expansion. An important factor contributing
to the risk of deteriorating wetlands along coastal areas is the alteration of hydrologic conditions
for flood control or navigation.
12-9
-------
Laveas
Water Gauges
WETLAND
Q UPLAND
Figure 12-2. Location of Lake Verret Basin (Conner et al., 1986)
12-10
-------
The subsidence that occurs in Louisiana can be divided into two general categories:
tectonic subsidence and consolidation/compaction. Tectonic subsidence refers to the large-scale
downward geologic displacement caused by sedimentary loading and associated settlement
processes of deltaic formations. The consolidation/compaction aspect of subsidence is attributed to
a variety of causes including overlying weight (levees, spoil mounds); subsurface withdrawal (oil
and gas exploration); and dewatering (drainage and reclamation projects). Because of the difficulty
in separating the effects of subsidence and sea level rise during any analysis of relative changes
between land and water levels, the two factors are frequently identified by researchers as "relative
sea level rise." Tidal gauges along the coast of Louisiana indicate that the rate of relative sea level
rise is 9 to 13 mm per year (3 to 4 ft per century) (Slater, 1986).
Sedimentary processes are responsive to changes in hydrologic and biologic processes, and
the rates of sediment accretion affect the ability of plants to adapt to the direct and indirect effects
of relative and absolute variations in water level (Gaboon and Turner, 1986). Sediment input and
organic accumulation counteract compaction and contribute to land accretion, but the supply and
distribution of sediments are not static in recent times. According to Meade and Parker (1984),
suspended sediments in the Mississippi River apparently have declined by more than 50 percent
since the early 1950s.
Prior to the construction of artificial flood-control levees, the Verret Basin was part of the
Mississippi-Atchafalaya River floodplain, and under natural conditions, floodwaters would bring
sediment-laden waters into these forested areas and the swamp forests would be replaced by
bottomland hardwood forests. However, artificial levees, erected for flood protection along the
east side of the Atchafalaya floodplain, have deprived the basin of seasonal overbank flooding and
sediment deposition, contributing to an approximate net subsidence rate of 25 cm over 50 years
(Slater, 1986).1 Water levels in the basin have since been influenced mainly by backwater
flooding and subsidence and secondarily by rainfall and upland runoff. Additionally, the
Atchafalaya River is lengthening its course due to the active deposition of sediments at its mouth
and the subsequent development of its delta. The river gradient (slope) is steadily decreasing and
has resulted in higher water levels at the mouth. Rising water levels in the lower part of the
Atchafalaya Basin reduce the hydraulic gradient of drainage from the upper part of the Verret
watershed, further increasing water levels (Boesch et al., 1983).
In this case study, an analysis is conducted on the ongoing processes in the Lake Verret
Basin. This analysis predicts future conditions, given current subsidence problems from past levee
construction. The primary hazard of concern in this case study is the rate and magnitude of water
level changes over time (i.e., hydroperiod) and the resultant changes on forest community species
and dynamics (i.e., habitat alteration). The analysis incorporates the three habitat types found
1In reality, the net subsidence rate probably varies within some range across the Atchafalaya Basin.
Slater (1986) provided the best known estimate of net subsidence in the basin at the time of this
study. The reader should note that predictions made by FORFLO and other models depend on the
most accurate environmental data available as inputs. As such, emphasis should be placed on
obtaining actual field data instead of using "default values" that may or may not be representative
of the study site.
12-11
-------
within the basin (drier bottomland hardwoods, wetter hardwoods, and cypress-tupelo swamps) with
their current value to five wildlife species, their probable fates due to basin-wide hydrologic
changes, and their most probable value to these same wildlife species in £0 years.
Ecological Components. As with most (if not all) assessments of the natural environment,
everything cannot be measured. Instead, one must decide on a subset of parameters that, within a
given timeframe and budget, will likely provide useful data for answering a scientific or regulatory
question. If the objectives of the study are to assess the impacts of one or more stressors on an
entire ecological community, then an ecologically based approach is desirable. Selection of
"components" (or species to be evaluated) may be based on the approach of choosing a suite of
representative species to provide an ecological perspective of the study area. Ideally,
representative species should be sensitive to specific land-use actions, serve as indicators for a
large segment of the wildlife community, and represent groups of species that use a common
environmental resource (e.g., representative species for various trophic guilds).
Fauna present at the study site were not surveyed directly. Instead, the wildlife species or
"components" used in this study were chosen from a list of common biota known to occupy the
Atchafalaya Basin by an interagency evaluation team composed of representatives of the USFWS,
U.S. Army Corps of Engineers (COE), National Marine and Fisheries Service, Louisiana
Department of Wildlife and Fisheries, Louisiana Department of Natural Resources, and EPA
(Gerry Bodin, USFWS, Lafayette, LA, personal communication). Wildlife components were
selected to represent (i.e., act as surrogates or status indicators of) the various guilds of wildlife
using ground cover, forest canopy, and water resources in the basin for feeding or reproduction.
Species also were chosen for their commercial, recreational, or social importance, and their
sensitivity to hydrological impacts (i.e., candidate species were screened for those variables found
in nature and the HSI models that would be sensitive to hydrologic changes in the basin). The
following five wildlife indicator species representative of habitat value in forested wetlands were
chosen for the HSI models: gray squirrel, Sciurus carolinensis (Allen, 1982); swamp rabbit,
Sylvilagus aquaticus (Allen, 1985); mink, Mustela vison (Allen, 1983); wintering wood duck, Aix
sponsa (Sousa and Farmer, 1983); and downy woodpecker, Picoldes pubescens (Schroeder, 1983).
The selection matrix used to evaluate and choose these species is shown in table 12-1.
Detailed biotic assessments of the study site focused on surveying dominant canopy tree and
wildlife cover vegetation in the wet bottomland hardwood, dry bottomland hardwood, and swamp
areas. Tree species encountered in all three areas are presented in table 12-2.
Endpoints. The primary assessment endpoint was spatial and temporal change in the forest
community and the associated habitat value of the forested wetlands. Measurement endpoints
included the data input requirements for the FORFLO and HSI models, as described below.
FORFLO Model Data Inputs. Data needs for FORFLO may be divided into the categories
of vegetation, hydrologic data, and other site data. These categories describe "ideal" data for input
for each FORFLO application. Shortcuts can be taken occasionally, and extrapolations can be
made with useful results still possible (Brody and Pendleton, 1987).
12-12
-------
Table 12-1. Sample Selection Matrix for Wildlife Species
Cover Types
Species
Socioeconomic
Value
Ground Canopy Water Indicator for
Gray Recreational X
squirrel
Swamp X
rabbit
Mast production
Utilization of ground surface
Downy Social
woodpecker
Mink
Wood
duck
Commercial
Recreational
and social
X
X Snag density and canopy
cover
X Edge effect between ground
and water surface
X Wintering habitat and
mast production
12-13
-------
Table 12-2. Tree Species Observed in Three Simulated Sites (Brody et aL, 1989)
Species Dry Bottomland Wet Bottomland Swamp
Acer rubrum (red maple) X X
Carya aqmtica (water hickory) X X
Ceuis laevigata (sugarberry) X
Diospyros virginiana (persimmon) X
Fraxinus pennsylvanica (green ash) X X
Gleditsia triacanthos (honey locust) X
Liquidambar styratiflua (sweetgum) X X
Nyssa aquatica (water tupelo) X
Populus deltoides (cottonwood) X
Quercus fyrata (overcup oak) X
Quercus nigra (water oak) X
Quercus nuttalUi (nuttall oak) X X
Quercus virginiana (live oak) X
Salix nigra (black willow) X
Taxodium distichum (bald cypress) X X
Ulmus americana (American elm) X X
12-14
-------
Vegetation. For each site to be modeled, species composition, relative abundance, and
density of canopy trees must be obtained. As currently formulated, FORFLO primarily models the
growth of canopy tree species. For each tree species, estimates of the average diameter at breast
height (DBH) and its standard deviation must be made. These data may be collected from either
line transects or plots and should be at relatively constant elevations. Although not as important as
DBH, average age and its standard deviation are required for each tree species.
Dominant species may be cored to establish age, or logging records may be used to
establish stand ages. Development of age/DBH regressions also can be a reliable way to provide
age data.
Other biological vegetation data required as model inputs are the beginning and ending
dates and the length of the growing season. Finally, maximum potential stand biomass must be
estimated as the total above-ground tree biomass.
Hydrologic data. The model requires hydrologic inputs in the form of an annual water
stage hydrograph. FORFLO breaks the year into 24 half-month (15-day) periods; for each of these
periods, the average water stage (height) and standard deviation of the water stage are required.
These data, typically available from long-term gauge readings of the COE or the U.S. Geological
Survey (USGS), describe current hydrologic conditions. When predicting impacts of projects that
will in some way alter this hydrologic regime, future hydrologic conditions assuming project
completion should be provided by the agency developing the project. The average water table
depth during the growing season must be measured or estimated. For example, wells 4 feet deep
can be easily dug, then measured on a weekly, biweekly, or monthly basis.
Other site data. To relate hydrologic information from a gauge reading to the bottomland
site, the elevation of each sample vegetation plot must be established, typically to the nearest foot.
The basic soil type must be generically described, for example, whether the soil is primarily sand,
clay, or loam. Annual degree-days to a 42-degree base and standard deviation can be determined
from National Oceanic and Atmospheric Administration (NOAA) climatological records.
Field studies based on point-centered quarter transects (Mueller-Dombois and Ellenberg,
1974) across a range of wet to dry sites from the top to the bottom of the basin were used to
establish the initial conditions for FORFLO. Tree species frequency, dominance, density, and
relative importance were estimated. At a more detailed level, tree species numbers, DBH, and
replacement of tree species were measured for 2 years in wet and dry plots at the bottom of the
basin, and in wet, transitional, and dry plots at the top of the basin. Tree bands during this period
documented seasonal growth patterns, and tree corings in the study plots documented long-term
growth increments. Basin water levels were estimated from a series of COE gauges.
HSI Model Data Inputs. The variables required for the gray squirrel, mink, downy
woodpecker, swamp rabbit, and wood duck HSI models are listed in table 12-3. Tree species and
size variables also were estimated from the point-centered quarter transects. Water regime
variables were estimated from COE water gauge readings, maps developed by the National
Wetlands Inventory (USFWS), Soil Conservation Service, and USGS, and aerial photography.
Data also were collected at the field sites to estimate values of the cover variables. Sampling was
12-15
-------
Table 12-3. List of All Variables Required in Five HSI Models (adapted from Brody et al.,
1989)
Species Model Variables
Squirrel Percentage of canopy closure of hard mast trees
Squirrel Number of tree species that produce hard mast
Squirrel, rabbit Percentage of tree canopy closure
Squirrel Average DBH of overstory trees
Squirrel Percentage of shrub crown cover*
Rabbit, mink Annual flood duration
Mink Percentage of tree, shrub, and persistent
emergent vegetation canopy closure3
Woodpecker Basal area
Woodpecker Number of snags
Wood duck Percentage of water surface covered
by winter cover*
aVariable not predicted by FORFLO.
12-16
-------
stratified by the three major habitat types: wet bottomland hardwoods, dry bottomland hardwoods,
and cypress-tupelo swamp. Within habitat types, circular (0.04 ha) plots were established
randomly at each site. Estimates were made by visual observation using the method of Hays et
al. (1981).
Future values for HSI model variables were taken directly from outputs of FORFLO for
seven variables (table 12-3). Because FORFLO did not predict midstory or ground cover strata
(i.e., percentage of shrub crown cover, percentage of shrub and persistent emergent vegetation
canopy closure, and percentage of water surface covered by winter cover for the squirrel, mink,
and wood duck models, respectively); predictions of future values of these variables used the field
data directly. The field samples provided an average value of midstory and ground cover strata for
each of the three habitat types. If a site changed habitat type, according to FORFLO predictions,
these variables were assigned new values based on the average value determined by the field
sampling. Thus, for each habitat type, midstory and ground cover strata were assigned the same
average values hi both the present and the future.
Comments on Problem Formulation
Strengths of the case study include:
stressors, ecological components, and endpoints are clearly identified. A
rationale is presented for selecting specific indicator species.
• This case provides a good example of problem formulation when the stressor
involves physical stressors and habitat alterations.
Limitations include:
• The case study considers effects on the habitat of five animal species. A
rationale is presented for their selection. However, there will always be some
limitation with regard to the species chosen with respect to how well they may
represent wildlife in general.
•Presentation of a species list for each habitat type would have been helpful. This
would have permitted the reader to relate the selected species to the ranges that
were present as well as to evaluate the implications of changes in habitat.
12.3.2. Analysis: Characterization of Ecological Effects
Plant Response to Flood Conditions. In general, plant adaptations to flood stress may be
characterized as physical or metabolic (Wharton et al., 1982). Physical adaptations include
the ability to restore or maintain root structures in flooded conditions or to produce anatomically
different roots that enhance survival in saturated soil conditions (e.g., more porous roots,
12-17
-------
pneumatophores). The primary metabolic plant adaptation to flood stress is a shift from the normal
three-step process of glucose metabolism and energy (adenosine triphosphate [ATP]) production
(glycolysis/Kreb's citric acid cycle/oxidative phosphorylation) to only glycolysis (an anaerobic
process). In most flood-intolerant tree species, only glycolysis occurs in the absence of free
oxygen (i.e., during flooded conditions) and, as a result, ethanol (an end product of glycolysis)
may accumulate to phytotoxic concentrations. Flood-tolerant tree species have the capability of
producing organic acid end products that can be used hi cellular synthesis in stems and leaves
instead of ethanol. This capability allows flood-tolerant tree species to avoid ethanol toxicity.
Despite a number of ecological studies of forested wetlands (see Conner and Day, 1982,
for a review), current knowledge of vegetation dynamics, especially in response to flooding in
wetland forests, is incomplete. Mature cypress (Taxodium cttstichum) and tupelo (Nyssa aquaticd)
do well under flooded conditions (Dickson et al., 1972; Kennedy, 1982). Increased flooding,
however, can sometimes have serious consequences even for the most flood-tolerant trees. In
Florida, Harms et al. (1980) found that in water from 20 to 100 cm deep, 0 to 16 percent of the
cypress trees died hi 7 years. In water over 120 cm deep, 50 percent of the cypress died after 4
years. In Louisiana, a long-term study of cypress survival was conducted in Lake Chicot
(Penfound, 1949; Eggler and Moore, 1961). After 4 years of flooding with water 60 to 300 cm
deep, 97 percent of the cypress were still alive. Eighteen years after flooding, 50 percent of the
cypress were still alive. Most of the living trees in the deep water had dead tops (Eggler and
Moore, 1961), but the numerous cypress trees still alive in the area indicate that cypress can
survive for long periods in a permanently flooded situation. At Catahoula Lake in northwest
Louisiana, Brown (1943) found cypress growing in waters with a seasonal variation in water level
greater than 7.5 m. From the available data on flooding and cypress growth and survival, it
appears that cypress can adapt to shallow (< 120 cm), permanent flooding. Even in deep water
(> 120 cm), death and decline are gradual (Hall et al., 1946; Eggler and Moore, 1961; Harms et
al., 1980; Klimas, 1987).
Other bottomland hardwood tree species are less tolerant of flooding than cypress and
tupelo. Many factors such as age and size of the tree, soil type, depth of flooding, time of
flooding, duration of flooding, and state of the floodwaters exert an influence on tree growth and
survival (Hook and Scholtens, 1978). The relative waterlogging tolerance ranking of the major
tree species in the Lake Verret Basin is summarized in table 12-4.
The references reviewed above provide some idea of what happens when water levels are
raised suddenly and large increases in depth occur, but the rate of change in vegetational
communities caused by a gradual rise in water level as documented in the coastal forests of
Louisiana (Conner and Day, 1988) is harder to determine.
Selection of the FORFLO and HSI Models. Because of the interaction of environmental
factors hi influencing forested wetland biology and succession, FORFLO was chosen for predicting
the future plant communities in the Lake Verret Basin. FORFLO is a simulation model developed
by modifying FORET (Shugart and West, 1977), a well-known upland deciduous forest succession
model. A number of other upland forest models have been derived from FORET, such as FORAR
(Meilke et al., 1977, 1978), FORMIS (Tharp, 1978), BRIND (Shugart and Noble, 1981), and
FORICO (Doyle, 1981). Shugart (1984) and Dale et al. (1985) provide a more comprehensive
12-18
-------
Table 12-4. Waterlogging Tolerance Ranking of Major Tree Species in Lake Verret Basin
(adapted from Hook, 1984)
Ranking* Species
Most tolerant Nyssa aquatica (water tupelo)
Taxodium distichum (bald cypress)
Highly tolerant Carya aquatica (water hickory)
Moderately tolerant Acer rubrum (red maple)
Fraxinus pemsylvanica (green ash)
Liqiddambar styradflua (sweetgum)
Quercus nuttaltii (nuttall oak)
Ulmus americana (American elm)
Weakly tolerant Celtis laevigata (sugarberry)
Quercus nigra (water oak)
aMost tolerant—those species capable of living from seedling to maturity in soils waterlogged
almost continually, except for short durations during droughts.
Highly tolerant—those species capable of living from seedling to maturity in soils waterlogged for
50 to 75 percent of the year.
Moderately tolerant—those species capable of living from seedling to maturity in soils
waterlogged about 50 percent of the year.
Weakly tolerant—those species capable of living from seedling to maturity in soils temporarily
waterlogged for 1 to 4 weeks of the year or about 10 percent of the growing season.
Least tolerant—those species capable of living from seedling to maturity in soils occasionally
waterlogged for only a few days, usually <2 percent of the growing season.
12-19
-------
discussion of these and other forest succession models. SWAMP (Phipps, 1979) was the only
other forested wetland model available, but FORFLO, which includes some modified growth
functions from SWAMP, was developed for use in regulatory applications.
The wildlife assessment processes of HSI models were linked to FORFLO's ability to
forecast future habitat conditions. HSI models are based on published studies of the basic food,
shelter, and reproductive requirements of selected wildlife species. The succession model outputs
of FORFLO provided input values to HSI models for assessing future value of the habitat to
wildlife. FORFLO predicts tree species presence and abundance, individual tree size, canopy
closure, flood duration, and other habitat measures that affect wildlife, on an annual basis for as
many years as required for the impact analysis. Thus, a mechanism is available to quantify a field
assessment of the current habitat values (HSI models), to quantify future habitat conditions
(FORFLO), and to assess future habitat values (HSI models).
The parameters influencing tree species presence and growth in the FORFLO model are
shown in figure 12-3. The model contains a library of tree data for common bottomland hardwood
species that defines how tree growth and reproduction respond to changes in site quality, flooding,
and the presence of other tree species. For example, each tree species has an assumed range of
annual duration in which it can survive, a maximum growth rate at an optimal water level, and
particular needs for wet and dry periods throughout the year for seed germination, seedling
survival, and growth (for reviews of water-tolerance characteristics of bottomland tree species, see
Bedinger, 1971; Teskey and Hinckley, 1977; Hook and Scholtens, 1978; Bedinger, 1979; Clark
and Benforado, 1981). Temperature is represented by the degree-days at a site. The range of a
tree species is determined by its maximum and minimum degree-day requirements, and it is
assumed that the tree species grows best at the center of its range. The effects of shading and
crowding are incorporated and represent competition among trees on a site, which may reduce their
growth. Trees are entered into the model either as seedlings or sprouts. Flood duration,
browsing, and soil variables reduce or enhance successful recruitment. Subsidence or accretion of
ground elevation on a site directly affects flood durations and depth of the water table. The
probability of a tree dying increases as the tree approaches the maximum age for the species.
Trees also have a high probability of death in the model when their growth slows to less than 10
percent of their optimum growth. The model tracks the species type, DBH, and age of each tree
on the simulated plot from the time the tree enters the plot as a seedling or sprout until it "dies."
12-20
-------
Figure 12-3. Diagram of dynamics contained in FORFLO (Pearlstine et al., 1985)
12-21
-------
Comments on Analysis: Characterization of Ecological Effects
Strengths of the case study include:
• The FORFLO model is based on a range of studies that relate tree growth and
survival to changes in water-level elevation.
• The HSI models provide a framework for evaluating potential ecological effects
associated with habitat modification.
•This case study is a good illustration of the application of ecological effects models.
Limitations include:
• The FORFLO and HSI models were developed independently. As such, the output
of FORFLO does not satisfy all the input requirements for the HSI model.
• Reviewers felt it would have been helpful to have more discussion of the sensitivity
of model output to the selected input variables.
• The presence and success of a wildlife species will depend on more than the
availability of suitable habitat. Other factors not considered in the model (such as
completion, predation, disease, etc.) may be important, and this introduces some
uncertainty into the analysis.
12.3.3. Analysis: Characterization of Exposure
The exposure regime was simulated by the FORFLO model, which simulates changes in
water elevation and assesses the impacts of the timing, duration, and magnitude of hydrological
effects. For this case study, this simulation model was used to represent the change in
hydrological conditions associated with natural subsidence, providing a baseline against which other
conditions can be compared. A value of 0.5 cm/yr was estimated to be the lowest estimate of
subsidence in the basin. Selected simulations also were conducted using a higher subsidence rate
of 1.0 cm/yr.
The results of applying the FORFLO simulation model for the two subsidence rates are
presented as part of the risk characterization (section 12.3.4). Exposure and effects information,
however, is incorporated within the simulation model.
12-22
-------
Comments on Analysis: Characterization of Exposure
Strengths of the case study include;
*The study examines a baseline case for subsidence. This can be used to gauge the effects
of more severe exposure conditions.
Limitations include:
• The model does not address physical burial of seedlings by sedimentation and other
similar factors. It should be noted that sedimentation processes independent of water level
are not addressed by the model. This factor introduces uncertainty into the analysis.
• Limited information was available for selecting subsidence rates. It would be helpful to
know the projected range of subsidence rates for the area.
•Levee construction is one of several processes that can influence hydroperiod (and depth
of flooding). Since the flood gauge readings used as input to the FORFLO model integrate
water-level changes from many sources, relative source contributions are not easily
separated.
• The model may not apply in other situations where the detrimental effects of
sedimentation, such as burial of seedlings, are critical.
12.3.4. Risk Characterization
Risks to Vegetation Due to Subsidence. FORFLO was first used to test the effects of
subsidence on the bottomland forest communities. For illustrative purposes, subsidence was set at
a low enough rate to enable the development of a mature community before it was replaced by a
more water-tolerant community. The results showed classic succession as the forest community
responded to increased flooding durations. Upland tree species first responded favorably to the
increased wetting of the soil, but as the flood durations continued to increase, the upland tree
species were replaced by a bottomland hardwood community and finally cypress-tupelo. As the
modeled flooding conditions became too great to support cypress-tupelo, the site became
nonforested to marsh or open water. This trend also was observed in the subcanopy with the
exception of a quicker response.
FORFLO was applied to a Lake Verret Basin wet bottomland hardwood site as
characterized in table 12-5. With a low subsidence rate of 0.5 cm/yr, succession from wet
bottomland hardwood to water tupelo (Nyssa aquaticd) dominance occurred within 50 years
(figure 12-4). As the simulation continued, bottomland hardwoods were completely replaced by
water tupelo and some bald cypress (Taxodium distichum) within 120 years, and water tupelo was
no longer sustained after 240 years.
12-23
-------
Table 12-5. Initial Tree Composition and Density for Three Simulated Sites (Brody et al.,
1989)
Species
Acer rubrum (red maple)
Carya aquatica (water hickory)
Celtis faevigata (sugarbeny)
Dlospyros virginiana (persimmon)
Fraxinus pemsylvanica (green ash)
Gleditsia triacanthos (honey locust)
Hquidambar styraciflua (sweetgum)
Nyssa aquatica (water tupelo)
Populus deltoides (cottonwood)
Quercus fyrata (overcup oak)
Quercus nigra (water oak)
Quercus nuttallii (nuttall oak)
Quercus virginiana (live oak)
Salix nigra (black willow)
Taxodium distichum (bald cypress)
Ulmus americana (American elm)
Wet Bottomland
Hardwood
0.02
0.09
0.39
0.02
0.28
—
0.02
—
0.14
—
—
0.02
—
0.28
0.02
0.02
Dry
Bottomland
Hardwood
0.03
0.03
—
—
0.05
0.05
0.15
—
—
0.05
0.39
0.05
0.26
—
—
0.10
Swamp
—
—
—
—
—
—
0.78
—
—
—
--
--
—
2.96
—
12-24
-------
157 -
§
~ 126-
X
*4
E
o 94-
o s *
M
o 63""
-------
Because 0.5 cm/yr was the lowest estimate of subsidence in the basin, the same plot was
simulated at a 1.0 cm/yr subsidence rate. With this rate, bottomland hardwoods were completely
replaced by the swamp community in 70 years, followed by the complete removal of swamp
species from the plot by year 140. The conservative estimate of 0.5 cm/yr was used for the
remaining simulation of the Lake Verret Basin forest succession. If longevity and continued
conservation of bottomland hardwoods hi the basin are the desired results, then these simulations
represent the best-case scenario.
For a drier bottomland hardwood site (table 12-5), water hickory (Carya aquaticd) was
dominant by year 50, but bottomland hardwoods were no longer regenerating and were replaced by
the recruitment of water tupelo and cypress (figure 12-5). When a swamp site (table 12-5) was
modeled, cypress and tupelo quickly stopped regenerating as flood durations increased, but almost
200 years elapsed before all the mature trees on the site were dead (figure 12-6). Net production
hi the swamp began declining almost immediately with the increased flooding duration, indicating
that while the trees were not immediately killed, they were stressed many years before dying.
Potential Risks to Wildlife. Current and 50-year predictions of future HSI values show a
general trend toward a loss of wildlife values in all habitat types (table 12-6). HSIs for squirrel,
woodpecker, and wood duck decreased hi all but a few cases where they remained virtually the
same. The swamp habitat lost almost all values to wildlife species. Habitat for two modeled
wildlife species, mink and swamp rabbit, increased in value as the flood duration in some areas
increased and retained some cover vegetation. Loss of squirrel habitat value was caused primarily
by the disappearance of hard mast-producing tree species (table 12-7). The only remaining mast-
producing tree was the flood-tolerant water hickory. Although not a preferred mast source of the
gray squirrel, this tree species is used if it is the only type available. Habitat values for the downy
woodpecker declined as the potential for snags for nest sites decreased. During the onset of forest
decline there may be a short-term increase in the number of dead or dying trees, but that also
diminishes as the rate of replacing old trees with new trees declines. The suitability values (table
12-7) for wood duck seem counter-intuitive, with the dry forest having higher value than swamp.
Within the swamp, however, water levels are permanently too deep for winter-persistent
herbaceous cover. This variable for cover is necessary in the model to supply quality wintering
habitat: Drier areas have much more cover and are often flooded in winter and spring months,
with permanently flooded areas nearby. The increasing flood durations cause the eventual decline
hi winter cover for wood duck.
Summary of Risks to Vegetation and Wildlife. Regional subsidence is a major process
altering the Lake Verret Basin. Flood duration and heights are increasing and apparently have
been for many years. Quantitative descriptions of the ongoing changes must concentrate on the
rates of these changes as well as their structure. Even with the conservative estimates of
subsidence that were assumed in this study, FORFLO predicted a rapid decline in bottomland
hardwoods and in the well-being of swamp tree species. Eventual nonforested conditions were
predicted throughout most of the basin as a result of increases in water levels caused by land
subsidence. These sites may succeed to fresh marsh or open water. Higher subsidence rates may
be realistic (Conner et al., 1986; Slater, 1986) and would accelerate these trends.
12-26
-------
135
o
"108
I *1
i 54
3
< 27 -
CANOPY
C - Catya aquatic*
O • C«iu laavigata
L • Uquidarnbar alyncAia
N • NytM aquano
N- 'N
™ a
— 254 -|
60
ro
80
90 100
(Yr»)
.0« .10
je
.30
.45
.so
.70
.89 Annual Flood
Duration
Figure 12-5. Succession in the canopy and subcanopy of the dry bottomland hardwood site
when a subsidence rate of 0.5 cm/yr is assumed (Brody et al., 1989)
12-27
-------
N . Nyssa aquaiica
T * Taxodium disticfium
s
I
i
2
<
1*
VI
m
a
135 -
90-
45 -
T T"\.T
"^T.
T-*^
^ N^^ N >.
^^^^ ^f
' " ^V ^"^^NJ^
,"163
» Total Production
• Leaf Production
a Stern Production
20 40 GO 80 100 120 140 160 180 200 Time(Yrs)
.70 .85 .90
1.00
Annual Flood
Duration
Figure 12-6. Succession and production in the cypress-tupelo swamp when a subsidence rate
of 0.5 cm/yr is assumed (Brody et al., 1989)
12-28
-------
Table 12-6. Current and Future Habitat Suitability Indices (HSI) for Five Species Evaluated
(Brody et al., 1989)
Dry Bottomland
Hardwood
Wet Bottomland
Hardwood
Swamp
Gray squirrel*
Minka
Downy
woodpecker
Swamp rabbit
Wood duck*
Current
HSI
0.66
0
0.5
0.57
0.7
Future
HSI
0.43
0.37
0.28
0.77
0.7
Current
HSI
0.47
0.60
0.50
0.42
0.62
Future
HSI
0.27
0.83
0.28
0.10
0.30
Current
HSI
0
1.00
0.50
0
0.54
Future
HSI
0
0.38
0.09
0
0
aFORFLO did not predict midstory or ground cover strata variables for the gray squirrel, mink, and
wood duck HSI models (i.e., percentage of shrub crown cover, percentage of shrub and persistent
emergent vegetation canopy closure, or percentage of water surface covered by winter cover).
Predictions of future values of these variables were based on collected field data that provided an
average value of midstory and ground cover strata for each of the three habitat types.
12-29
-------
Table 12-7. Current and Future Suitability Indices (SI) for Variables of Five HSI Models
(Brody et al., 1989)
Dry Bottomland
Hardwood
Variable
Mink
Percentage of tree, shrub,
and persistent emergent
vegetation canopy closure
Annual flood duration
Swamp rabbit
Percentage of tree
canopy closure
Flood duration
Gray squirrel
Percentage of canopy
closure of hard-mast trees
Number of tree species
that produce hard mast
Percentage of tree
canopy closure
Average DBHa of
overstory trees
Percentage of shrub
crown cover
Downy woodpecker
Number of snags
Basal area
Wood duck
Percentage of water
surface covered by
winter cover
Current
SI
1
0
0.57
1
0.55
0.8
0.86
1
1
1
0.5
0.7
Future
SI
0.97
0.14
0.96
0.8
0.93
0.2
1
1
0.96
0.28
1
0.7
Wet Bottomland
Hardwood
Current
SI
1
0.36
0.53
0.8
0.45
0.5
0.82
1
0.95
1
0.5
0.62
Future
SI
0.72
0.96
1
0.1
0.38
0.2
1
1
1
0.28
1
0.3
Swamp
Current
SI
1
1
0.76
0
0
0
0.95
1
0.62
1
0.5
0.54
Future
SI
0.15
1
0.44
0
0
0
0.28
1
0.64
0.09
0.88
0
aDBH = diameter at breast height.
12-30
-------
Several processes changing the basin's forested wetland are affecting wildlife and have
probably been doing so for many years. The decline of hard mast-producing tree species means
that less winter food is available. Some oak populations remain, but as the simulations show, they
have no chance of any significant regeneration. There will be fewer tree species, as only the most
flood-tolerant will regenerate. FORFLO provides no direct prediction of ground cover, but it may
be inferred from the flood duration predictions of the model that ground cover will eventually
disappear. The positive side of these changes is that wetter conditions will probably create short-
term benefits to some aquatic wildlife species.
The changes in wildlife habitat are both an indirect consequence of the changing hydrology
altering the forest communities and their habitat structure and a direct consequence of an even
longer period of flooding. Once again, the rate of change is as significant an attribute as the
changes themselves. This factor is particularly true for practical applications of this assessment
methodology to federal projects. In these applications, the project's impact on wildlife habitat
typically will be considered for time periods representing the "life of the project." Thus, the
analysis using HSI models was performed at the 50-year point, but simulated forest changes were
applied for far longer periods to provide insight on longer-term impacts on wildlife habitat in the
basin.
Uncertainties and Limitations. This section presents the uncertainties and limitations
associated with the case study results. The discussion begins first with the FORFLO model and its
validation, then proceeds to the HSI models and their relationship with the Habitat Evaluation
Procedure (HEP).
FORFLO. FORFLO primarily assesses the presence or absence of trees as a function of
hydrological conditions over time. Not all interspecific and intraspecific interactions between
wildlife and plant species are represented in the model (e.g., understory shrub species, exotics).
Nevertheless, since canopy-level vegetational succession is predictable to a certain extent, the
structural and physical features of habitats also are predictable. Thus, future habitat values were
projected by incorporating the outputs of FORFLO as inputs to HSI models. However, the number
of individuals fluctuates naturally over time and is often independent of the structural and physical
features of the available habitat. These fluctuations can be difficult to measure or predict and are
often caused by other stochastic events such as disease, predation, and competition. FORFLO does
not factor in chemical stresses and thus cannot express these types of impacts for interpretation
within the HSI models.
Confidence in the predictions of FORFLO models has been established by independent field
validation studies. For example, results of a study by Pearlstine et al. (1985) showed good
agreement between FORFLO simulations and field observations of forest composition along a
25-km reach of forested floodplains in South Carolina (see figure 12-7). Field observations in this
particular study were conducted by both the USFWS and the University of South Carolina.
Importance values (the sum of the relative density, relative dominance, and relative frequency of
each tree species) are used to describe observed and predicted forest composition. The tree species
listed along the bottom of the graph are all the species available to the FORFLO model for
recruitment (Pearlstine et al., 1985).
12-31
-------
3001
100-
iS 90-
3 80-
> 70-
ill
0 60-
£ 50-
g 40-
| 30-
20-
10-
n
;m
— U.S. FISH AND WILDLIFE SERVICE, CHARLESTON. SC
,-. UNIVERSITY OF SOUTH CAROLINA
FORFLO SIMULATION. 50 RUN MEAN
p
I
1 1 :
ii i i. !l ,
I
i ! 1.
I
1
!
-100
•90
-BO
-70
-60
-50
-40
-30
r20
-10
n
Figure 12-7. Comparison of FORFLO simulation results with field observations (Pearlstine
et al., 1985)
12-32
-------
HEP and HSI models. HSI models are an integral part of the HEP developed by the
USFWS. In HEP analyses, the end result is to derive and compare Habitat Units (HUs), which
represent both the quality and quantity of habitat for chosen wildlife indicator species (in this case
study, gray squirrel, mink, downy woodpecker, swamp rabbit, and wood duck). Results of HSI
models represent habitat quality; habitat quantity is expressed as some unit of surface area
(typically acres). Hus for each wildlife species are derived by the following equation:
HU = (HSI X Acreage)
Present and future Hus were beyond the scope of the original study by Brody et al. (1989).
As a result, the conclusions in this case study do not take into account instances, for example,
where low-quality, high-acreage habitats may be more beneficial to the wildlife indicator species
than high-quality, low-acreage habitats.
Like any other approach used for impact or risk assessment, HSI models and HEP have
limitations that define the limits of application and identify potential problem areas where good
professional judgment is required. A habitat approach basically limits application of the
methodology to those situations hi which measurable and predictable habitat changes are important
variables, but there are no assurances that wildlife populations will exist at the potential levels or
optimal levels predicted by habitat analyses. Another limitation is that the wildlife species/
habitat-based assessment methodology is applicable only for the wildlife species evaluated and does
not necessarily relate to other wildlife species associated with other ecosystem components.
Nevertheless, this should not prevent users of HEP and the FORFLO/HEP linkage from making
scientifically sound, qualitative statements about potential beneficial or adverse impacts to other
important flora and fauna in the study area. Keep in mind, however, that such statements must
include caveats that they are qualitative inferences, not reliable facts.
HEP does not provide the user with any guidance for performing future predictions.
Therefore, projected impacts are only as reliable as the user's ability to predict future conditions.
FORFLO is designed to provide a better methodological approach for predicting future habitat
conditions. ,
Comments on Risk Characterization
Strengths of the case study include:
•The case study is a good example of an ecological risk assessment where physical
alteration and habitat modification are the stressors.
*The case study illustrates a methodology that could be used to assess future
alterations (e.g., accelerated rates of change in water elevation).
12-33
-------
Comments on Risk Characterization (continued)
*The study was considered to be of good scientific quality. FORFLO provides a
valuable tool for predicting changes in habitat quality as input to HSL
•The case study illustrates how ecological effects-exposure models may be used to
predict changes (risks) associated with forecasted changes in exposure. Further, the
case study indicates an effort to use and combine available tools to assess risks.
Limitations include:
•Even if habitat is optimal, wildlife populations may not exist.
*HSIs are applicable for the species modeled, but not necessarily for other wildlife
species.
•Only limited validation of FORFLO/HSI model predictions is available; thus, there
is uncertainty in the analysis.
12A. RECOMMENDATIONS AND FUTURE RESEARCH NEEDS
This case study illustrates that FORFLO can improve decisions on environmental problems
regarding terrestrial or wetland forests. Despite model limitations, there are obvious benefits to be
derived from continued work on HSI models, FORFLO, and the interface between them. HSI
models are already fairly well known, and FORFLO is an appropriate choice to forecast habitat
changes in floodplain forests. FORFLO has potential for providing EPA managers with more
insight on (1) the physical impacts of proposed Resource Conservation and Recovery Act (RCRA)
sites, dredge and fill activities, and other Agency actions affecting forested wetlands; (2) the
probable success of wetland remediation efforts; (3) the likelihood that socially beneficial functions
associated with forested wetlands will be present through time; and (4) possible hydrological
criteria. FORFLO has been linked with geographic information systems and also may play an
important role hi assessing cumulative impacts.
EPA's Science Advisory Board (U.S. EPA, 1990b) has ranked physical habitat alteration
and biological depletion as posing high ecological risks and has called upon EPA to direct its
efforts to "those most critical environmental problems where the greatest risk reduction can be
obtained." As a result, ecological risk assessors within the Agency should expand beyond
chemical-specific assessments and consider the potential regulatory applications of FORFLO,
HSI models, and other ecological models and techniques that assess physical environmental
stressors.
Future research to improve FORFLO could include better linkages with HSI models or
other regulatory applications, better stress-response descriptions of the effects of hydrological
12-34
-------
changes on tree species, stress-response information on shrub species and tree species not already
hi the model, and the incorporation of properties that would enable the model to make predictions
at the landscape or regional level.
12-35
-------
12.5. REFERENCES
Allen, A.W. (1982) Habitat Suitability Index models: gray squirrel. U.S. Fish and Wildlife
Service, Division of Biological Services, Washington, DC. FWS/OBS-82/10.19.
Allen, A.W. (1983) Habitat Suitability Index models: mink. U.S. Fish and Wildlife Service,
Division of Biological Seivices, Washington, DC. FWS/OBS-82/10.61.
Allen, A.W. (1985) Habitat Suitability Index models: swamp rabbit. U.S. Fish and Wildlife
Service, Division of Biological Services, Washington, DC. Biological Report 82(10.107).
Bedinger, M.S. (1971) Forest species and indicators of flooding in the lower White River Valley,
Arkansas. U.S. Geological Survey Professional Paper 750-C.
Bedinger, M.S. (1979) Forest and flooding with special reference to the White River and Ouachita
River Basins, Arkansas. U.S. Geological Survey Open-File Report-79-68.
Boesch, D.F.; Levin, D.; Nummedal, D.; Bowles, K. (1983) Subsidence in coastal Louisiana:
causes, rates, and effects on wetlands. U.S. Fish and Wildlife Service, Division of
Biological Services, Washington, DC. FWS/OBS-83/26.
Brody, M.; Pendleton, E. (1987) FORFLO: a model to predict changes in bottomland hardwood
forests. U.S. Fish and Wildlife Service, Research and Development, Office of Information
Transfer, Fort Collins, CO.,
Brody, M.; Conner, W.; Pearlstine, L.; Kitchens, W. (1989) Modeling bottomland forest, and
wildlife habitat changes in Louisiana's Atchafalaya Basin. In: Sharitz, R.R.; Gibbons,
J.W., eds. Freshwater wetlands and wildlife. CONF-8603101, DOE Symposium Series,
No. 61. USDOE Office of Scientific and Technical Information, Oak Ridge, TN.
Brown, C.A. (1943) Vegetation and lake-level correlations at Catahoula Lake, Louisiana.
Geograph. Rev. 33:435-445.
Gaboon, D.R.; Turner, R.E. (1986) Accretion and canal impacts in a rapidly subsiding wetland II.
Feldspar marker horizon technique. Estuaries 12(4):260-268.
Clark, J.R.; Benforado, J., eds. (1981) Wetlands of bottomland hardwood forests. In:
Developments in agricultural and managed-forest ecology 11. Proceedings of a Workshop
on Bottomland Hardwood Forest Wetlands of the Southeastern United States held at Lake
Lanier, GA, June 1-5, 1980. Elsevier Scientific Publishing Company, NY.
, . : ' * ;
Conner, W.H.; Day, J.W., Jr. (1982) The ecology of forested wetlands in the southeastern United
States. In: Gopal, B.; Turner, R.E.; Wetzel, R.G.; Whigham, D.F., eds. Wetlands:
ecology and management. Jaipur, India: International Scientific Publishers, pp. 69-87.
12-36
-------
Conner, W.H.; Slater, W.R.; McKee, K.; Flynn, K.; Mendelssohn, LA.; Day, J.W. (1986)
Factors controlling the growth and vigor of commercial wetland forests subject to increased
flooding in the Lake Verret, Louisiana, watershed. Baton Rouge, LA: Louisiana State
University Center for Wetland Resources, Final Report, Louisiana Board of Regents'
Research and Development.
Conner, W.H.; Day, J.W., Jr. (1988) Rising water levels hi coastal Louisiana: implications for
two coastal forested wetland areas in Louisiana. /. Coastal Res. 4:589-596.
Conner, W.H.; Brody, M. (1989) Rising water levels and the future of southeastern Louisiana
swamp forests. Estuaries 12(4):318-323.
Dale, V.H.; Doyle, T.W.; Shugart, H.H. (1985) A comparison of tree growth models. Ecol
Modelling 29:145-169.
Dickson, R.E.; Broyer, T.C.; Johnson, C.M. (1972) Nutrient uptake by tupelogum and bald
cypress from saturated or unsaturated soil. Plant Soil 37:297-308.
Doyle, T.W. (1981) The role of disturbance in the gap .dynamics of a Montane rain forest: an
application of a tropical forest succession model. In: West, D.C.; Shugart, H.H.; Botkin,
D.B., eds. Forest succession: concepts and application. New York: Springer-Verlag, pp.
56-73.
Eggler, W.A.; Moore, W.G. (1961) The vegetation of Lake Chicot, Louisiana, after eighteen
years of impoundment. Southwest Nat. 6:175-183.
Hall, T.F.; Penfound, W.T.; Hess, A.D. (1946) Water level relationships of plants in the
Tennessee Valley with particular reference to malaria control. Report ReeTfoot Lake Biol.
Sta, 10:18-59.
Harms, W.R.; Schreuder, H.T.; Hook, D.D.; Brown, C.L. (1980) The effects of flooding on the
swamp forest in Lake Ocklawaha, Florida. Ecology 61:1412-1421.
Hays, R.L.; Summers, C.; Seitz, W. (1981) Estimating wildlife habitat variables. U.S. Fish and
Wildlife Service, Division of Biological Services, Washington, DC. FWS/OBS-81/47.
Hook, D.D. (1984) Waterlogging tolerance of lowland tree species of the south. South. J. Appl.
For. 8:136-149.
Hook, D.D.; Scholtens, J.R. (1978) Adaptations and flood tolerance of tree species. In: Hook,
D.D.; Crawford, R.M.M., eds. Plant life in anaerobic environments. Ann Arbor, MI:
Ann Arbor Science Publishers Inc., pp. 299-331.
Kennedy, H.E., Jr. (1982) Growth and survival of water tupelo coppice regeneration after six
growing seasons. South. J. Appl. For. 6:133-135.
12-37
-------
Klimas, C.V. (1987) Bald cypress response to increased water levels, Caddo Lake, Louisiana-
Texas. Wetlands 7:25-37.
Meade, R.H.; Parker, R.S. (1984) Sediments in rivers of the United States. National water supply
summary. Vicksburg, MS: U.S. Geological Survey Water-Supply Paper 2275.
Meilke, D.L.; Shugart, H.H.; West, D.C. (1977) User's manual for FORAR, a stand model for
upland forests of southern Arkansas. Oak Ridge, TN: Oak Ridge National Laboratory.
ORNL/TM-5767.
Meilke, D.L.; Shugart, H.H.; West, D.C. (1978) A stand model for upland forests of southern
Arkansas. Oak Ridge, TN: Oak Ridge National Laboratory. ORNL/TM-6225.
Mueller-Dombois, D.; Ellenberg, H. (1974) Aims and methods of vegetation ecology. New York,
NY: John Wiley & Sons.
Office of Technology Assessment (OTA) (1984) Wetlands: their use and regulation. Washington,
DC: U.S. Congress, Office of Technology Assessment. OTA-0-206.
Pearlstine, L.; McKellar, H.; Kitchens, W. (1985) Modelling the impacts of a river diversion on
bottomland forest communities in the Santee River Floodplain, South Carolina. Ecol.
Modelling 29:281-302.
Penfound, W.T. (1949) Vegetation of Lake Chicot, Louisiana, in relation to wildlife resources.
Proc. La. Acad. Sci. 12:47-56.
Phipps, R.L. (1979) Simulation of wetlands forest vegetation dynamics. Ecol. Modelling .
7:257-288.
Russell, M.; Gruber, M. (1987) Risk assessment in environmental policy-making. Science
236:286-290.
Schroeder, R.L. (1983) Habitat Suitability Index models: downy woodpecker. U.S. Fish and
Wildlife Service, Division of Biological Services, Washington, DC. FWS/OBS-82/10.38.
Shugart, H.H. (1984) A theory of forest dynamics: the ecological implications of forest succession
models. New York, NY: Springer-Verlag.
Shugart, H.H.; West, D.C. (1977) Development and application of an Appalachian deciduous
forest succession model. /. Environ. Management 5:161-179.
Shugart, H.H.; Noble, I.R. (1981) A computer model of succession and fire response of the high-
altitude eucalyptus forest of the Brindabella Range, Australian Capital Territory. Aust. J.
Ecol. 6:149-164.
12-38
-------
Slater, W.R. (1986) Long-term water-level changes and bottomland hardwood growth in the Lake
Verret Basin, Louisiana. Baton Rouge, LA: Louisiana State University, M.S. thesis.
Soil Conservation Service. (1978) Lake Verret watershed, final revised Environmental Impact
Statement. Alexandria, VA: U.S. Department of Agriculture.
\t
Sousa, P.J.; Farmer, A.H. (1983)s Habitat Suitability Index models: wood duck. U.S. Fish and
Wildlife Service, Division of Biological Services, Washington, DC. FWS/OBS-82/10.43.
Stakhiv, E.Z. (1986) The status of risk analysis in water resources engineering. In: Haimsed,
Y.Y.; Stakhiv, E.Z., eds. Risk-based decision-making in water resources. New York, NY:
American Society of Civil Engineers, pp. 111-128.
Teskey, R.O.; Hinckley, T.M. (1977) Impact of water level changes on woody riparian and
wetland communities. Vol. II: southern forest region. U.S. Fish and Wildlife Service,
Biological Services Program, Washington, DC. FWS/OBS-77/59.
Tharp, M.L. (1978) Modeling major perturbations on a forest ecosystem. Knoxville, TN:
University of Tennessee, M.S. thesis.
U.S. Environmental Protection Agency, Science Advisory Board. (1990a) Reducing risk-setting
priorities and strategies for environmental protection. SAB-EC-90-021, September 1990.
U.S. Environmental Protection Agency, Science Advisory Board. (1990b) Relative risk reduction
project: reducing risk, appendix A. The report of the Ecology and Welfare Subcommittee.
EPA SAB-EC-90-021A, September 1990.
U.S. Fish and Wildlife Service. (1981) Standards for the development of Habitat Suitability Index
models. 103 ESM. U.S. Fish and Wildlife Service, Division of Ecological Services,
Washington, DC. |
Wharton, C.H.; Kitchens, W.M.; Pendleton, E.G.; Sipe, T.W. (1982) The ecology of bottomland
hardwood swamps of the Southeast: a community profile. U.S. Fish and WildlifeUervice,
Biological Services Program, Washington, DC. FWS/OBS-81/37.
12-39
*U.S GOVERNMENT PRINTING OFFICE 1993 -750 -002/ 80272
------- |