United States
Environmentaly'Protection Agency
Office of Wat/r
"Offtee-ef Research and Development
National Exposure Research Laboratory
Cincinnati, OH 45268
Official Business
Penalty for Pri^te
$300
EP/V822/B-00/025
December 2000
www.epa.gov
United States
Environmental Protection
Agency
Office of Water
Washington DC 204
Office of Research/and
Development
Washington DC 20460
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA PERMIT NO. G-35
oEPA
Stressor Identification
Guidance Document
EPA/822/B-00/025
-------
STRESSOR IDENTIFICATION
GUIDANCE DOCUMENT
U.S. Environmental Protection Agency
Office of Water
Washington, DC 20460
Office of Research and Development
Washington, DC 20460
EPA-822-B-00-025
December 2000
-------
Disclaimer
This Stressor Identification Guidance Document provides guidance to assist EPA Regions, States, and
Tribes in their efforts to protect the biological integrity of the Nation's waters, one of the primary
objectives of the Clean Water Act (CWA). It also provides guidance to the public and the regulated
community on identifying stressors that cause biological impairment. While this document constitutes
the U.S. Environmental Protection Agency's (EPA's) scientific recommendations regarding stressor
identification, this document does not substitute for the CWA or EPA's regulations, nor is it a regulation
itself. Thus, it cannot impose legally binding requirements on EPA, States, Tribes, or the regulated
community, and may not apply to a particular situation based upon the circumstances. When appropriate,
State and Tribal decisionmakers retain the discretion to adopt approaches on a case-by-case basis that
differ from this guidance. EPA may change this guidance in the future.
-------
Stressor Identification Guidance Document
Acknowledgments
Primary Authors:
EPA, Office of Research and Development:
Susan Cormier, Ph.D.
Susan Braen Norton, Ph.D.
Glenn Suter II, Ph.D.
EPA, Office of Science and Technology:
Donna Reed-Judkins, Ph.D.
Contributing Authors:
EPA, Office of Science and Technology:
Jennifer Mitchell
William Swietlik
Marjorie Coombs Wellman
EPA, Office of Wetlands, Oceans and Watersheds:
Thomas Danielson
Chris Faulkner
Laura Gabanski, Ph.D.
Molly Whitworth, Ph.D.
EPA, Office of Research and Development:
Edith Lin, Ph. D.
Bhagya Subramanian
EPA, Office of Enforcement and Compliance Assurance
Brad Mahanes
Other Affiliations:
David Altfater, Ohio Environmental Protection Agency
William Clements, Ph.D., Colorado State University, Fort Collins, Colorado
Susan P. Davies, Ph.D., Maine Department of Environmental Protection, Augusta, Maine
Jeroen Gerritsen, Ph.D., Tetra Tech, Owings Mills, Maryland
Martina Keefe, Tetra Tech, Owings Mills, Maryland
Sandy Page, Tetra Tech, Owings Mills, Maryland
Jeffrey Stinson, Ph.D., U.S. Air Force
Technical Editors:
EPA, Office of Research and Development, National Risk Management and Restoration Lab:
Jean Dye, Ph.D.
Scott Minamyer
in
-------
Stressor Identification Guidance Document
Tetra Tech:
Abby Markowitz
Sandra Page
Colin Hill
Brenda Fowler
Stressor Identification and Evaluation Workgroup Members:
Co-leads:
Office of Water: Donna Reed-Judkins, Ph.D., Office of Science and Technology
Office of Research and Development: Susan Cormier, Ph.D., National Exposure Research Lab
Members:
Office of Water:
Office of Science and Technology:
Tom Gardner, Susan Jackson, Jennifer Mitchell, Keith Sappington, Treda Smith,
William Swietlik, Brian Thompson, Marjorie Wellman
Office of Wetlands, Oceans, and Watersheds:
Thomas Danielson, Laura Gabanski, Chris Faulkner, Molly Whitworth, Ph.D.
Office of Research and Development:
National Center for Environmental Assessment:
Susan Norton, Ph.D., Glenn Suter II, Ph.D.
National Health and Environmental Effects Laboratory:
Naomi Detenbeck, Ph.D., Wayne Munns, Ph.D.
National Risk Management and Restoration Laboratory:
Alan Everson, Scott Minamyer
Office of Enforcement and Compliance Assurance
Brad Mahanes
EPA Regions
Toney Ott, Region 8
Other Federal Agencies:
Jeffrey Stinson, Ph.D., U.S. Air Force
States:
Susan Davies, Maine Department of Environmental Protection, Augusta, Maine
Chris O. Yoder, Ohio EPA, Columbus, Ohio
Other Supporting EPA Members:
Don Brady, Alan Hais, Margarete Heber, Mary Sullivan
Contract Support, Tetra Tech, Owings Mills, Maryland:
Michael Barbour, Ph.D., Jeroen Gerritsen, Ph.D., Martina Keefe, Sandy Page
IV
-------
Stressor Identification Guidance Document
Peer Reviewers:
A. Fred Holland, Ph.D., Director, Marine Resources Research Institute of South Carolina.
Kent Thornton, Ph.D., FTN Associates
Wayne Landis, Ph.D., Director, Institute of Environment Toxicology and Chemistry, Western
Washington University
The authors wish to gratefully acknowledge all others, not named above, who helped to prepare this
document. The sum of these efforts contributed to the success of this guidance. Special thanks also Jo all
the EPA and State scientists who participated in the video conference in October, 1999; the Region III
Mid-Atlantic Water Pollution Biology Workshop at Cacapon, West Virginia, in March 2000; and the
Biological Advisory Committee meeting in Cincinnati, in May, 2000. Comments were also provided by
a number of EPA scientists and regulators and by other stakeholders, including the Kansas Department
of Health and Environment, Arizona Department of Environmental Quality, Denver Metro Wastewater
Reclamation District, Pennsylvania Department of Environmental Protection, Proctor and Gamble, The
Nature Conservancy, and the U.S. Geological Survey. Comments from the workshops and other
commenters helped shape the guidance.
The cover illustration was provided by a fifth grade student at Ursula Villa Elementary School, Mount
Lookout, OH. According to the illustrator, the front cover is the river when you first pick up this book,
and the back cover is the river after you've followed the instructions.
-------
Stressor Identification Guidance Document
Table of Contents
Acknowledgments iii
Acronym List xiii
Executive Summary
ES. 1 The Clean Water Act, Biological Integrity, and Stressor
Identification ES-1
ES.2 Intended Audience ES-2
ES.3 Applications of the SI Process ES-2
ES.4 Document Overview ES-3
Chapter 1: Introduction to the Stressor Identification (SI) Process
1.1 Introduction
1.2 Scope of this Guidance . . .
1.3 Data Quality Issues
1.4 Overview of the SI Process
1.4.1 The SI Process
1.4.2 SI Process Iterations
1.4.3 Using the Results of Stressor Identification
1.5 Use of the SI Process in Water Quality Management Programs
-1
-2
-2
-3
-3
-5
-5
-6
Chapter 2: Listing Candidate Causes
2.1 Introduction 2-1
2.2 Describe the Impairment 2-1
2.3 Define the Scope of the Investigation 2-3
2.4 Make the List 2-4
2.5 Develop Conceptual Models 2-5
Chapter 3: Analyzing the Evidence
3.1 Introduction 3-1
3.2 Associations Between Measurements of Candidate Causes and
Effects 3-2
3.3 Using Effects Data from Elsewhere 3-6
3.4 Measurements Associated with the Causal Mechanism 3-9
3.5 Associations of Effects with Mitigation or Manipulation of Causes . . 3-10
Chapter 4: Characterizing Causes
4.1 Introduction 4-1
4.2 Methods for Causal Characterization 4-1
4.2.1 Eliminating Alternatives 4-3
4.2.2 Diagnostic Protocols or Keys 4-7
4.2.3 Strength of Evidence Analysis 4-8
4.2.3.1 Causal Considerations for Strength of Evidence
Analysis 4-9
4.2.3.2 Matching Evidence with Causal Considerations 4-14
4.2.3.3 Weighing Causal Considerations 4-14
4.3 Identify Probable Cause and Evaluate Confidence 4-17
vn
-------
Stressor Identification Guidance Document
Table of Contents (continued)
Chapter 5: Iteration Options
5.1 Reconsider the Impairment 5-1
5.2 Collect More Information on Previous and Additional Scenarios 5-2
Chapter 6: Presumpscot River, Maine
6.1 Executive Summary 6-1
6.2 Background 6-3
6.3 List Candidate Causes 6-5
6.4 Analyze Evidence and Characterize Causes: Eliminate 6-8
6.5 Analyze Evidence and Characterize Causes: Strength of Evidence ..6-11
6.6 Characterize Causes: Identify Probable Cause 6-17
6.7 Significance and Use of Results 6-18
6.8 References 6-18
Chapter 7: Little Scioto River, Ohio
7.1 Executive Summary 7-1
7.2 Introduction 7-4
7.3 Evidence of Impairment 7-5
7.4 List Candidate Causes 7-10
7.5 Analyze Evidence to Eliminate Alternatives 7-13
7.5.1 DataAnalyzed 7-13
7.5.2 Associations between Candidate Causes and Effects 7-14
7.5.3 Measurements Associated with the Causal Mechanism:
Exposure Pathways 7-24
7.5.4 Summary of Analyses for Elimination 7-26
7.6 Characterize Causes: Eliminate 7-26
7.7 Analyze Evidence for Diagnosis 7-28
7.8 Analyze Evidence to Compare Strength of Evidence 7-28
7.9 Characterize Causes: Strength of Evidence 7-31
7.10 Characterize Causes: Identify Probable Cause 7-47
7.11 Discussion 7-48
7.12 References 7-50
7.13 Additional Tables 7-54
APPENDICES
A Overview of Water Management Programs Supported by the SI
B Worksheet Model
C Glossary of Terms
D Literature Cited
INDEX
vin
-------
Stressor Identification Guidance Document
List of Figures
Figure Page
1-1 The management context of the SI process 1-4
2-1 A conceptual model for ecological risk assessment illustrating the effect of
logging in salmon production in a forest stream 2-7
3-1 The flow of information from data acquisition to the analysis phase of the
SI process 3-3
3-2 Plot of toxicity data from a 7-day subchronic test of ambient waters and a
community metric obtained on a common stream gradient 3-4
4-1 A logic for characterizing the causes of ecological injuries at specific sites . . 4-2
6-1 Map of the Presumpscot River showing biomonitoring stations, potential
sources of impairment, and their location relative to the Androscogginn
River 6-4
6-2 Species richness and number of EPT taxa in the Presumpscot River upstream
and downstream of a pulp and paper mill effluent discharge 6-6
6-3 Conceptual model showing the potential impact of stressors on the benthic
community of the Presumpscot River 6-7
6-4 Bottom dissolved oxygen concentration in the Presumpscot River 6-10
7-1 Map of the Little Scioto River, Ohio, showing sites where fish were
sampled 7-6
7-2 Spatial changes in fish IBI (A) and benthic macroinvertebrate ICI (B)
values in the Little Scioto River in 1987 (OEPA 1988) and 1992
(OEPA 1994) 7-7
7-3 Changes in IBI and ICI scores over distance in the Little Scioto River,
1992 7-9
7-4 A conceptual model of the six candidate causes for the Little Scioto
stressor identification 7-12
7-5 Selected QHEI metrics for 1987 and 1992 7-13
7-6 Mean PAH concentrations from the sediment (mg/kg) in the Little Scioto
River 1987-1998 7-15
7-7 Mean metal concentrations from the sediment(mg/kg) in the Little Scioto
River 1987-1998 7-17
IX
-------
Stressor Identification Guidance Document
List of Figures (continued)
Figure Page
7-8 Mean water chemistry values from the Little Scioto River from 1987-1998 . 7-18
7-9 Bile metabolites ((ig/mg protein) measured in white suckers from the
Little Scioto River in 1992 7-25
-------
Stressor Identification Guidance Document
List of Tables
Table Page
ES-1 Summary of the use of Stressor Identification (SI) in water quality
management programs ES-3
1-1 The role of SI in various water management programs 1-6
3-1 Types of associations between measurements of causes and effects among
site data and the evidence that may be derived from each 3-4
3-2 Example associations between site-derived measures of exposure and
measures of effects from controlled studies for different types of stressors . . 3-8
3-3 Example associations between site data and the processes by which
stressors induce effects 3-9
3-4 Types of field experiments and the evidence that may be derived from
each 3-10
4-1 Application of common types of evidence in eliminating alternatives 4-5
4-2 Types of evidence (columns) that contribute to each causal
consideration (rows) 4-15
4-3 Format for a table used to summarize results of an inference concerning
causation in case-specific ecoepidemiology 4-16
6-1 Evidence of biological impairment in the Presumpscot River upstream and
downstream of a pulp and paper mill effluent discharge 6-5
6-2 Physical and chemical parameters measured in the Presumpscot River
upstream and downstream of a pulp and paper mill effluent discharge 6-9
6-3 Considerations for eliminating candidate causes 6-12
6-4 Comparison of TSS loadings in the Presumpscot and Androscoggin
Rivers 6-12
6-5 1996 - 1999 metal concentrations in the pulp and paper mill effluent 6-13
6-6 Strength of evidence of non-attainment in the Presumpscot River 6-14
7-1 Summary of the three impairments that were considered in the Little
Scioto River 7-10
7-2 Spearman rank correlations with selected metrics and the IBI and ICI
from 1992 and selected PAHs 7-19
XI
-------
Stressor Identification Guidance Document
List of Tables (continued)
Table Page
7-3 Spearman rank correlations with selected metrics and the IBI and ICI
from 1992 and selected metals 7-19
7-4 Spearman rank correlations with selected metrics and the IBI and ICI
from 1992 and selected water quality and habitat quality measurements . . . 7-20
7-5 Evidence for eliminating candidates causes at Impairments A, B, and C ... 7-21
7-6 Candidate causes remaining after elimination 7-28
7-7 Cumulative toxic units for PAHs and metals based on the PEL values 7-30
7-8 Comparison of the reported concentration of water quality parameters
(mg/L) with exceedances 7-31
7-9 Strength of evidence analysis for the three candidate causes of
Impairment A, RM 7.9 7-32
7-10 Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5 7-36
7-11 Strength of evidence analysis for the three candidate causes of
Impairment C, RM 5.7 7-43
7-12 Causal characterization 7-48
7-13 Fish metrics for the Little Scioto River 1987 and 1992 7-54
7-14 Macroinvertebrate metrics for the Little Scioto River 1987 and 1992 7-55
7-15 QHEI metrics for the Little Scioto River 1987 and 1992 7-56
7-16 Average concentrations of selected sediment organic compounds (mg/kg)
in the Little Scioto River, Ohio, by river mile in 1987, 1991, 1992 and
1998 7-57
7-17 Average concentrations (mg/kg) of selected metals in sediment from the
Little Scioto River, Ohio, by river mile in 1987, 1991, 1992 and 1998 7-61
7-18 Average concentration of selected water chemistry parameters (mg/L) in
the Little Scioto River, Ohio, by river mile in 1987, 1992 and 1998 7-63
7-19 PAH concentrations at nearest upstream location and locations of
impairments (mg/kg) 7-64
7-20 Metals concentrations at nearest upstream location and locations of
impairments (mg/kg) 7-65
XII
-------
Stressor Identification Guidance Document
Acronym List
303(d) The section of the Clean Water Act that requires a listing by states,
territories, and authorized tribes of impaired waters, which do not meet
the water quality standards that states, territories, and authorized tribes
have set for them, even after point sources of pollution have installed the
minimum required levels of pollution control technology.
305(b) The section of the Clean Water Act that requires EPA to assemble and
submit a report to Congress on the condition of all water bodies across
the Country as determined by a biennial collection of data and other
information by States and Tribes.
7Q10 Lowest average 7 consecutive days flow with average recurrence
frequency of once every 10 years
BAP Benzo[a]pyrene
BOD Biological Oxygen Demand
CERCLA Comprehensive Environmental Response, Compensation, and Liability
Act
COD Chemical Oxygen Demand
CSOs Combined Sewer Outfalls
CWA Clean Water Act
DELTA Deformities, Erosions, Lesions, Tumors, and Anomalies
DDT Dichlorodiphenyltrichloroethane
DNR Department of Natural Resources
DO Dissolved Oxygen
DQA Data Quality Assessment
DQO Data Quality Objectives
ECBP Eastern Cornbelt Plains
EMAP Environmental Monitoring and Assessment Program
EPA U.S. Environmental Protection Agency
EPT Ephemeroptera-Plecoptera-Tricoptera
EROD Ethoxy Resorufin[o]deethylase
xin
-------
Stressor Identification Guidance Document
FACA Federal Advisory Committee Act
GIS Geographic Information System
IBI Index of Biotic Integrity
ICI Invertebrate Community Index
IFIM Instream Flow and Incremental Methodology
KBI Kansas Biotic Index
KDHE Kansas Department of Environmental Protection
MBI Macroinvertebrate Biotic Index
MIWB Modified Index of Weil-Being
MWH Modified Warmwater Habitat
NA Not Applicable/Available
NAPH Naphthalene
NE No Evidence
ND Not Detected
NEP National Estuaries Program
NIH National Institute of Health
NOX Nitrites
NPDES National Pollution Discharge Elimination Act
NFS Non-point Source
NRC National Research Council
OEPA Ohio Environmental Protection Agency
PAHs Polycyclic Aromatic Hydrocarbons
PEL Probable Effect Level
PO4 Ortho-phosphate
POTWs Publicly Owned Treatment Works
QHEI Qualitative Habitat Evaluation Index
xiv
-------
Stressor Identification Guidance Document
RM River Mile
SECs Sediment Effect Concentrations
SEP Supplemental Environmental Protection
SI Stressor Identification
TEL Threshold Effect Level
TKN Total Kjeldahl Nitrogen
TMDL Total Maximum Daily Load
TP Total Phosphorus
TIE Toxicity Identification Evaluation
TRE Toxicity Reduction Evaluation
TSS Total Suspended Solids
USEPA U.S. Environmental Protection Agency
WET Whole Effluent Toxicity
WWH Warm Water Habitat
WWTP Waste Water Treatment Plant
xv
-------
Stressor Identification Guidance Document
Executive Summary
In this Summary:
ES.1 The Clean Water Act, Biological
Integrity, and Stressor Identification
ES.2 Intended Audience
ES.3 Application of the SI Process
ES.4 Document Overview
ES.1 The Clean Water Act, Biological
Integrity, and Stressor
Identification
Since the inception of the Clean Water Act (CWA) in 1972, the rivers, lakes, estuaries,
and wetlands of the United States have indeed become cleaner. The standard for
measuring these improvements are both chemical and biological. Yet, we know that
many waterbodies still fail to meet the goal of the Clean Water Act - to maintain the
chemical, physical and biological integrity of the nation's waters.
The ability to accurately
identify stressors and
defend the evidence
supporting those
findings is a critical step
in developing strategies
that will improve the
quality of aquatic
resources.
Biological assessments have become increasingly important
tools for managing water quality to meet the goals of the
CWA. These methods, which use measurements of aquatic
biological communities, are particularly important for
evaluating the impacts of chemicals for which there are no
water quality standards, and for non-chemical stressors such
as flow alteration, siltation, and invasive species. However,
although biological assessments are critical tools for detecting
impairment, they do not identify the cause or causes of the
impairment.
The Office of Water and Office of Research and
Development of the US EPA have developed a process for
identifying any type of Stressor or combination of stressors
that cause biological impairment. The Stressor Identification
(SI) Guidance is intended to lead water resource managers
through a formal and rigorous process that
* identifies stressors causing biological impairment in aquatic ecosystems, and
+ provides a structure for organizing the scientific evidence supporting the
conclusions.
The ability to accurately identify stressors and defend the evidence supporting those
findings is a critical step in developing strategies that will improve the quality of aquatic
resources.
The Stressor Identification process (SI) is prompted by biological assessment data
indicating that a biological impairment has occurred. The general SI process entails
critically reviewing available information, forming possible Stressor scenarios that might
explain the impairment, analyzing those scenarios, and producing conclusions about
which stressor or stressors are causing the impairment. The SI process is iterative,
usually beginning with a retrospective analysis of available data. The accuracy of the
identification depends on the quality of data and other information used in the SI
process. In some cases, additional data collection may be necessary to accurately
identify the stressor(s). The conclusions can be translated into management actions and
the effectiveness of those management actions can be monitored.
Executive Summary
ES-1
-------
Stressor Identification Guidance Document
ES.2 Intended Audience
This guidance should prove useful to anyone involved in managing impaired aquatic
ecosystems. The results of Stressor Identification investigations are valuable to many
types of environmental managers including land-use
planners, industrial and municipal dischargers, reclamation
Although the Stressor companies, and any individuals or organizations involved in
,,,.,.,. . activities that directly or indirectly affect water quality or
Identification process ,s aquatic habitatg
scientifically rigorous, it
. The process of stressor identification draws upon abroad
is flexible enough to yariety of disciplines and is most effective when the SI
support various water investigator has input from professionals in a number of
environmental areas such as aquatic ecology, biology,
" geology, geomorphology, statistics, chemistry, environmental
requirements. risk assessment, and toxicology. Sophisticated knowledge in
certain fields may increase the tools available to investigators
(e.g., physiological responses to certain stressors), but the SI
process also can be used by investigators with very general tools (e.g., fish population
estimates). Results of general measures, however, may not be as precise as when more
specialized measures are used (e.g., stomach-lining histological evaluations).
ES.3 Applications of the SI Process
Although the Stressor Identification process is scientifically rigorous, it is flexible
enough to support various water management requirements. Some potential applications
of the SI process include the following:
* Characterizing the Quality of the Nation's Waters: Stressor Identification
procedures can assist states in more accurately identifying the causes of
biological impairment in 305(b) reporting.
> Identifying Waterbodies and Wetlands that Exceed Water Quality
Standards: Accurate, reliable stressor identification procedures are necessary
for EPA and the states/tribes to accurately identify the cause(s) of water quality
standards violations for 303(d) listing and Total Maximum Daily Load (TMDL)
calculations. The SI process can help achieve higher degrees of accuracy and
reliability in identifying pollutants causing impacts. The SI process is not
designed, however, to allocate the amount of responsibility for an impact to a
particular source, especially when multiple sources of a stressor are present.
* Regulatory and Non-Regulatory Pollution Management Programs: Stressor
identification procedures can help identify different types of stressors within a
watershed that are contributing to biological impairment. Stressors can then be
prioritized and controlled through a combination of voluntary and mandatory
programs.
Other types of programs in which the SI process is useful include: State/Local Watershed
Management Programs, National Pollutant Discharge Elimination System (NPDES)
Permitting Programs, Dredge and Fill Permitting, Compliance and Enforcement Actions,
Risk Assessments, Preservation and Restoration Programs, and Control Effectiveness
Assessments.
ES-2 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
If a legal challenge to the conclusions drawn is possible, or if costly remediation efforts
are indicated as the means to control a stressor, it is essential to have a high level of
confidence in the accuracy of the identification. However, because requirements for
confidence levels and stressor precision can vary with the intended use of the findings,
managers also require flexibility in evaluation systems. Table ES.l summarizes various
levels of rigor required in eight water quality management programs where the SI
process can be applied.
Table ES.L Summary of the use of Stressor Identification (SI) in water quality
management programs.
Water Program
305(b) Water
Quality Reports
303(d) Impaired
Waterbody Lists
319 Non-point
Source Control
402 Point Source
Permitting
316(b) Cooling
Water Intake
Permitting
401 Water Quality
Certifications
404 Wetlands
Permitting
Water
Enforcement
Type of Program
Advisory
Regulatory
Enforcement
Level of Rigor Needed for SI
Low
Medium
High
ID Source
ES.4 Document Overview
The SI guidance document describes the organization and analysis of available evidence
to determine the cause of biological impairment. The document does not directly
address biological assessment, impairment detection, source allocation, management
actions, or data collection, although these activities interact with SI in significant ways.
This document is intended to guide water resource managers through the Stressor
Identification process.
Section One: The Stressor Identification Process
Introduces SI process and provides detailed guidance on implementing a stressor
identification program. The guidance applies principles of ecoepidemiology to
evaluating causes of biological impairment at specific locations.
Chapter 1: Introduction to the SI Process
Provides the background and justification for the SI process.
Executive Summary
ES-3
-------
Stressor Identification Guidance Document
Chapter 2: Listing Candidate Causes
Provides an overview of and guidance on the first step of the SI process, listing
candidate causes for the impairment.
Chapter 3: Analyzing Evidence
Provides an overview of and guidance on the second step of the SI process,
analyzing new and previously existing data to generate evidence.
Chapter 4: Characterizing Causes
Provides an overview of and guidance on the third step of the SI process, using the
evidence from Step 2 to draw conclusions about the stressors that are most likely to
have caused the impairment.
Chapter 5: Iteration Options
Provides options for stressor identification if no clear cause is found in the first
iteration.
Section Two: Case Studies
Provides two case studies illustrating the SI process.
Chapter 6: Presumpscot River, Maine
Chapter 7: Little Scioto River, Ohio
Appendix A: Overview of Water Management Programs Supported by the SI
Process
Appendix B: Worksheet Model
Appendix C: Glossary of Terms
Appendix D: Literature Cited
ES-4 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Chapter 1:
Introduction to the Stressor
Identification (SI) Process
In this Chapter:
1.1 Introduction
1.2 Scope of this Guidance
1.3 Data Quality Issues
1.4 Overview of the SI Process
1.5 Use of the SI Process in Water
Quality Management Programs
1.1 Introduction
The use of biological assessments and biocriteria in state and tribal water quality
standards programs is atop priority of the U.S. Environmental Protection Agency (EPA).
As such, one of the agency's objectives is to ensure that all States and Tribes develop
water quality standards and programs that
SI is an invaluable
component of any
bioassessment/biocriteria
program concerned with
protecting the biological
integrity of aquatic
ecosystems.
use bioassessment information to evaluate the
condition of aquatic life in all waterbodies,
establish biologically-based aquatic life use
designations,
protect aquatic life use standards with narrative or
numeric biocriteria (see box below),
regulate pollution sources,
assess the effectiveness of water quality
management efforts, and
communicate the condition of their waters.
Although bioassessments are useful for identifying biological impairments, they do not
identify the causes of impairments. Linking biological effects with their causes is
particularly complex when multiple stressors impact a waterbody. Investigation
procedures are needed that can successfully identify the stressor(s) and lead to
appropriate corrective measures through habitat restoration, point and non-point source
controls, or invasive species control. Water management programs have historically
shown that aquatic life protection is best accomplished using integrated information from
various sources. For example, the whole effluent toxicity program has utilized methods
for more than a decade that help resource managers understand and control the toxicity
Defining Terms- Aquatic Life Use and Biocriteria
Aquatic Life Use is a beneficial use designation, identified by a state, in which a waterbody
provides suitable habitat for the survival and reproduction of desirable fish, shellfish, and other
aquatic organisms. Beneficial Use Designation is a management objective defining desirable
uses that water quality should support. Examples include drinking water supply, primary
contact recreation (swimming), and aquatic life use.
Biocriteria are narrative expressions (qualitative) or numeric values (quantitative) describing
the biological characteristics of aquatic communities based on appropriate reference
conditions.
Chapter 1: Introduction to the Stressor Identification (SI) Process
1-1
-------
Stressor Identification Guidance Document
of complex effluents. Similarly, the Stressor Identification process will enable water
resource managers to better understand and control stressors affecting aquatic biota. SI
is an invaluable component of any bioassessment/biocriteria program concerned with
protecting the biological integrity of aquatic ecosystems.
1.2 Scope of this Guidance
The SI guidance covers the organization and analysis of available evidence to determine
the cause of biological impairment. It does not directly address biological assessment,
reference condition, impairment detection, source allocation,
,., . . management actions, data collection, or stakeholder
The SI process may be , . .., ' , .. ...' . . . ... CT.
involvement- although these activities interact with SI in
applied to any level of significant ways. After stressors are identified, the
... . . ... appropriate management actions depend on the nature of
biological organization JT + A ^ f + i A-
those stressors, and on other factors- including economics.
(e.g., individuals, Identifying appropriate management actions is beyond the
. ,. scope of this document, but examples of management actions
populations, F , , ,. ., ' ,. , ., ,. «° . , f
are included in the case studies described in Chapters 6-7 or
communities) and to any this document.
type of waterbody (e.g., ,, +u j +* + TJ
Many methods exist tor measuring impacts, exposure, land-
freshwater streams, use, habitat changes and other parameters that are important
, . ... pieces of evidence in an SI investigation. Descriptions of
estuaries, wetlands, r! .,, , ,., & -.,. .,F ~ CT
those methods are beyond the scope or this guidance. I he SI
etc.). guidance, however, relies on the proper use of many tools to
collect evidence. EPA recognizes the need for a tools
compendium as well as software to help organize evidence, to
make use of available databases and technical publications and to prompt proper
collection of additional data when needed. The SI process should be viewed as a "logic
backbone" in determining the cause of impacts to aquatic biota.
1.3 Data Quality Issues
The SI process is a procedure for analyzing available evidence and determining if the
available evidence is adequate to draw a conclusion about the causes of impairment.
Since evidence may be collected from a variety of sources using a variety of tools, proper
documentation of the data is critical. Each technique for collecting data has associated
quality control measures. The higher the quality of data analyzed, the better the chances
will be of correctly identifying stressors. Guidance on assessing data quality and making
use of various types of data may be found in the Comprehensive State Water Quality
Assessment (305b) guidelines (USEPA 1997) and Ecological Risk Assessment
guidelines (USEPA 1998a, also Chapter 3). Data of unknown or poor quality can
sometimes be used for very rough estimates if the goals of the study allow, but, in
general, the quality of all data should be acceptable and well documented. If the
available data are not adequate, the SI process can show where data are missing or
deficient, but it does not address designing new data collection efforts. Chapter 2,
however, does provide advice on quality control when new data are collected.
After stressors are identified, the appropriate management actions depend on the nature
of those stressors and on other factors, including economics. Evaluating whether Stressor
controls have allowed biological recovery is critically important in verifying that the
stressors were accurately identified.
1 -2 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
1.4 Overview of the SI Process
The SI process may be applied to any level of biological organization (e.g., individuals,
populations, communities) and to any type of waterbody (e.g., freshwater streams,
estuaries, wetlands, etc.). Some of the criteria presented for evaluating evidence may be
specific, however, to a waterbody type (e.g., references to upstream/downstream
associations). Similarly, the logic of the SI process may be applied in straightforward,
single stressor situations or in complex situations with multiple stressors and cumulative
impacts. Complex situations may require investigators to refine the definition of the
study area, gather new data, or do multiple iterations of SI to identify all the important
stressors. The Little Scioto Case Study (Chapter 7) is given as an example of a complex
stressor situation where river segments were analyzed separately because impacts and
stressors differed at each location.
1.4.1 The SI Process
Figure 1-1 provides an overview of the Stressor Identification process within the context
of water quality management and data collection. The SI process is initiated by the
observation of a biological impairment (shown in the topmost box). Decision-maker and
stakeholder involvement is shown along the left-hand side; their involvement is
particularly important in defining the scope of the investigation and listing candidate
causes. At any point in the process of identifying stressors, a need for additional data
may be identified; the acquisition of this data is shown by the box on the right-hand side
of the diagram. The accurate characterization of the probable cause allows managers to
identify appropriate management action to restore or protect biological condition. Once
stressors are identified and management actions are in place to control them, the
effectiveness of the SI process (as demonstrated by improved conditions) can be
monitored using appropriate monitoring tools and designs.
The core of the SI process is shown within the bold line of Figure 1-1 and consists of
three main steps:
1. listing candidate causes of impairment (Chapter 2),
2. analyzing new and previously existing data to generate evidence for each
candidate cause (Chapter 3), and
3. producing a causal characterization using the evidence generated in Step 2 to
draw conclusions about the stressors that are most likely to have caused the
impairment (Chapter 4).
The first step in the SI process is to develop a list of candidate causes, or stressors, that
will be evaluated. This is accomplished by carefully describing the effect that is
prompting the analysis (e.g., unexplained absence of brook trout) and gathering available
information on the situation and potential causes. Evidence may come from the case at
hand, other similar situations, or knowledge of biological processes or mechanisms. The
outputs of this initial step are a list of candidate causes and a conceptual model that
shows cause and effect relationships.
Chapter 1: Introduction to the Stressor Identification (SI) Process 1-3
-------
Stressor Identification Guidance Document
Detect
or Suspect Biological
Impairment
D
(D
O
w'
o
3
a>
?r
(D
^
fi)
a
&>
?r
(D
3"
O.
a
(D
O.
(D
3
(D
s
tressor Identification
LIST CANDIDATE CAUSES
^L
ANALYZE EVIDENCE
n
CHARACTERIZE CAUSES
Eliminate Diagnose Strength of Evidence
1 1 1
Identify Probable Cause
Identify/
Apportion
Sources
(D
O
(D
fi)
O
.a
(D
D
P
rf
(D
3
-------
Stressor Identification Guidance Document
The second step, analyzing evidence, involves analyzing the information related to each
of the potential causes. Virtually everything that is known about an impaired aquatic
ecosystem is potentially useful in this step. For example, useful data may come from
chemical analysis of effluents, organisms, ambient waters, and sediments; toxicity tests
of effluents, waters, and sediments; necropsies; biotic surveys; habitat analyses;
hydrologic records; and biomarker analyses. These data do not in themselves, however,
constitute evidence of causation. The investigator performing the analysis must organize
the data in terms of associations that could support or refute proposed causal scenarios.
Chapter 3 discusses several levels of associations between:
> measurements of the candidate causes and responses,
> measures of exposure at the site and measures of effects from laboratory studies
> site measurements and intermediate steps in a chain of causal processes, and
> cause and effect in deliberate manipulations of field situations or media.
These associations comprise the body of evidence used to characterize the cause.
In the third step, characterize causes, the investigator uses the evidence to eliminate, to
diagnose, and to compare the strength of evidence in order to identify a probable cause.
The input information includes a description of the effects to be explained, the set of
potential causes, and the evidence relevant to the characterization. Evidence is brought
in and analyzed as needed until sufficient confidence in the causal characterization is
reached. In straightforward cases, the process may be completed in linear fashion. In
more complex cases, the causal characterization may require additional data or analyses,
and the investigator may iterate the process.
1.4.2 SI Process Iterations
The SI process may be iterative, beginning with retrospective
analysis of available data. If the stressor is not adequately Although the SI process
identified in the first attempt, the SI process continues using cannot accurately
better data or testing other suspected stressors. The process
repeats until the stressor is successfully identified. The identify stressors
certainty of the identification depends on the quality of without adequate data,
information used in the SI process. In some cases, additional
data collection may be necessary to confidently identify the completing tne o/
stressor(s). Although the SI process cannot accurately identify process is helpful even
stressors without adequate data, completing the SI process is
helpful even without adequate data because the exercise can witnout adequate aata
help target future data collection efforts. because the exercise
1.4.3 Using the Results of Stressor Identification can helP tar9et future
data collection efforts.
Stressor Identification is only one of several activities required
to improve and protect biological condition (Figure 1-1). In
some cases, the most effective management action will be obvious after the probable
cause has been identified. In many cases, however, the investigation must identify
sources and apportion responsibility among them. This can be even more difficult than
identifying the stress in the first place (e.g., quantifying the sources of sediment in a
Chapter 1: Introduction to the Stressor Identification (SI) Process 1-5
-------
Stressor Identification Guidance Document
large watershed), and may require environmental process models. The identification and
implementation of management alternatives can also be a complex process that requires
additional analyses (e.g., economic comparisons, engineering feasibility) and stakeholder
involvement. Once a management alternative is selected and implemented, monitoring
its effectiveness can ensure that biological goals are attained, and provides valuable
feedback to the SI process. All of these important activities are outside the scope of the
current document. However, accurate and defensible identification of the cause through
the SI process is the key component that directs management efforts towards solutions
that have the best chance of improving biological condition.
1.5 Use of the SI Process in Water Quality Management Programs
Identifying the cause of biological impairments is an essential element of many water
quality management programs. Table 1-1 summarizes the stressor identification needs of
several water management programs. An extended discussion of some major regulatory
programs and their requirements is presented in Appendix A.
Table 1 -1. The role of SI in various water management programs.
Program
Type/Name
Purpose
Role of SI
305(b)
Characterizing
the Quality of
the Nation's
Waters
Under section 305(b) of the Clean
Water Act (CWA), states and tribes
are required to assess the general
status of their waterbodies and
identify, in general terms, known or
suspected causes of water quality
impairments, including biological
impairments.
Stressor identification procedures will
assist states and tribes to accurately
identify the causes of biological
impairment. This is a non-regulatory,
information reporting effort. A high
degree of certainty in identifying the
causes of impairment is not always
needed for 305(b) reports.
303(d) Listings
and TMDLs
Identifying
Waterbodies
and Wetlands
that Exceed
Water Quality
Standards
Under section 303(d) of the CWA,
states and tribes are required to
prepare and submit to EPA lists of
specific waterbodies that currently
violate, or have the potential to violate
water quality standards, including
designated uses and numeric or
narrative criteria such as biocriteria.
Wetlands assessment programs are
also being developed and wetlands
may be listed on 303(d) lists.
Accurate, reliable stressor
identification procedures are
necessary for EPA and the
states/tribes to accurately identify the
cause(s) of water quality standards
violations. A high degree of accuracy
and reliability in the stressor
identification process is necessary
and sources will need to be identified.
State/Local
Watershed
Management
Programs
Managing water resources on a
watershed basis involves examining
the quality of a waterbody relative to
all the stressors within its watershed.
Stressors, once identified, are
prioritized and controlled through a
combination of voluntary and
mandatory programs, possibly
employing the CWA 402, 319, 404,
401, and other programs.
Stressor identification procedures will
help to identify the different types of
stressors within a watershed that may
be contributing to biological
impairment. A high degree of
certainty in identifying the causes of
impairment is needed.
1-6
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 1-1 (continued). The role of SI in various water management programs.
Program
Type/Name
Purpose
Role of SI
319 Non-point
Source Control
Program
The 319 Program is a voluntary,
advisory program under which the
states develop plans for controlling
the impacts of non-point source runoff
using guidance and information about
different types of non-point source
pollution.
Stressor identification procedures will
help to identify the different types of
non-point sources within a watershed
that may be contributing to biological
impairment. A high degree of
certainty in identifying the causes of
impairment is not always needed.
NPDES Permit
Program
Under Section 402 of the CWA, it is
illegal to discharge pollutants to
waters of the United States from any
"point source" (a discrete
conveyance) unless authorized by a
National Pollutant Discharge
Elimination System permit issued by
either the states or EPA. NPDES
permits are required whenever a
discharge is found to be causing a
violation of water quality, including
biological impairment.
Accurate stressor identification can
be very critical in NPDES permitting
cases, both for fairness and success
in stressor control. The SI process
can help to determine if the discharge
is the cause of biological impairment.
This is especially important when site-
specific modifications of state
standards or national criteria are
used. A high degree of accuracy and
reliability in the stressor identification
process is necessary and sources will
need to be identified. The SI process
is not designed to allocate the amount
of responsibility for an impact when
multiple sources for a stressor are
present.
316(b) Cooling
Water Intake
Program
Under Section 316(b) of the CWA,
any NPDES permitted discharger
which also intakes cooling water must
not cause an adverse environmental
impact to the waterbody.
To determine if a cooling water intake
structure is causing adverse
environmental impacts to the
waterbody, the overall health of the
waterbody should be known. Where
biological impairments are found,
stressor identification procedures
should be used to identify the different
stressors causing the waterbody to be
impaired, including the intake
structure. A high degree of certainty
is needed.
401 Water
Quality
Certifications
Under Section 401 of the CWA,
different types of federal permitting
activities (such as wetlands dredge
and fill permitting) require a
certification that there will be no
adverse impact on water quality as a
result of the activity. This certification
process is the 401 Water Quality
Certification.
Stressor identification procedures will
help to identify the different types of
stress an activity may place on water
quality that can then be addressed
through conditions in the 401
Certification.
Chapter 1: Introduction to the Stressor Identification (SI) Process
1-7
-------
Stressor Identification Guidance Document
Table 1-1 (continued). The role of SI in various water management programs.
Program
Type/Name
Purpose
Role of SI
Wetlands
Permitting
Under Section 404 of the CWA, the
discharge of dredge and fill materials
into a wetland is illegal unless
authorized by a 404 Permit. The 404
Permit must receive a 401 Water
Quality Certification.
Stressor identification procedures
may help to identify unanticipated
stress from a dredge and fill activity
on water quality or the biological
community after the activity is
underway. Stressor identification
procedures will also help in pre-
permitting evaluations of the potential
impacts of 404 permitting by
assessing different potential stressors
on the wetland in advance.
Compliance
and
Enforcement
Whenever an enforcement action is
taken by a regulatory authority, the
type of pollution, the source, and
other stressors that play a role in
causing the violation need to be
clearly identified and related to the
violating source.
Stressor identification procedures
must be able to clearly identify the
different types of pollution causing the
violation with a high degree of
confidence. Legal defensibility is
required. Identifying the source with a
high degree of confidence is also
needed, though the current SI
process does not provide that
guidance.
Risk
Assessments
Results of bioassessment studies can
be used in watershed ecological risk
assessments to predict risk from
specific stressors and anticipate the
success of management actions.
Accurate stressor identification is an
integral part of this process and can
help ensure that management actions
are properly targeted and efficient in
producing the desired results.
Wetlands
Assessments
States are beginning to develop
wetlands assessment procedures. In
the future, wetlands protection is
expected to be increasingly
incorporated into state water quality
standards.
Stressor identification procedures, as
well as future tools specific to wetland
investigations, are very much needed
by wetlands managers. The biological
assessment methods will allow
resource managers to evaluate the
condition of wetlands and may
provide some indication of the type of
stressor damaging a wetland. Once
bioassessment methods are
completed and incorporated into
monitoring programs, wetlands may
be listed on 305(b) lists as impaired
due to biological impairment. The SI
process should help identify stressors
causing biological impairment so
resource managers can better
remedy the problems.
1-8
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 1-1 (continued). The role of SI in various water management programs.
Program
Type/Name
Purpose
Role of SI
Preservation
Programs
The National Estuary Program (NEP)
was established in 1987 by
amendments to the Clean Water Act
to identify, restore, and protect
nationally significant estuaries of the
United States. The program focuses
on improving water quality in an
estuary, and on maintaining the
integrity of the whole system -its
chemical, physical, and biological
properties-as well as its economic,
recreational, and aesthetic values.
Stressor identification procedures
should be useful to the NEP, and
other preservation programs, by
helping stakeholders identify causes
of impairments. This information
would feed into the development of a
management plan.
Restoration
Programs
The Comprehensive Environmental
Response, Compensation, and
Liability Act (CERCLA), commonly
known as Superfund, was enacted in
1980 (and amended in 1986) for
hazardous waste cleanup.
As in enforcement and compliance
programs, stressor identification
procedures must be able to clearly
identify the different types of pollution
causing the impairment with a high
degree of confidence. Legal
defensibility is required. Identifying
the source with a high degree of
confidence is also needed, though the
current SI process does not provide
that guidance.
Pollution
Control
Effectiveness
A key component of any pollution
control program or watershed
management effort is the ability to
ascertain (or predict) the likely
effectiveness of pollution control
measures or management strategies.
Stressor identification procedures will
help to identify the different types of
pollution a control measure needs to
reduce and the different types of
stressors a management strategy
needs to address.
Chapter 1: Introduction to the Stressor Identification (SI) Process
1-9
-------
Stressor Identification Guidance Document
Chapter 2
Listing Candidate Causes
2.1 Introduction
The first step in the stressor identification
process is to develop the list of candidate
causes, or stressors. This is accomplished by
carefully describing the effect that is
prompting the analysis, and gathering
available information on the situation and
potential causes (see the box below for
definitions of some key terms). Potential
causes are evaluated and those that are
sufficiently credible are retained as candidate
causes used in the analysis stage. The outputs
of this initial step are a list of candidate causes
and a conceptual model that shows the
relationship between the causes and the effect.
In this Chapter:
2.1 Introduction
2.2 Describe the Impairment
2.3 Define the Scope of the Investigation
2.4 Make the List
2.5 Develop Conceptual Models
s
tressor IdentificatiofK /
LIST CANDIDATE CAUSES
4 L
ANALYZE EVIDENCE
3 C
CHARACTERIZE CAUSES
Eliminate Diagnose Strength of Evidence
1 1 1
Identify Probable Cause
Defining Terms - Exposures, Effects, Causes, Sources
An effect is a biological change traceable to a cause.
Exposure is the co-occurrence or contact of a stressor with the biological resource.
A cause is defined as a stressor that occurs at an intensity, duration, and frequency of exposure
that results in a change in the ecological condition.
A source is the origin of a stressor. It is an entity or action that releases or imposes a stressor
into the waterbody.
note: the processes of detecting impairment and identifying sources are beyond the scope of
this document
2.2 Describe the Impairment
The first important piece of information to be documented is a careful description of the
effect that prompted the evaluation. Whenever possible, the impairment should be
described in terms of its nature, magnitude, and spatial and temporal extent (see
worksheet in Appendix B, Unit I, page B-4). Making inferences about causes is easier
when the impairment is defined in terms of a specific effect, or response. The response
should be quantified as a count (abundance of darter species) or continuous variable
(mean length of darters). If multiple effects with different causes are described as a
single impairment, it may be mistakenly assumed that there is only a single cause.
The importance of biological entities as resources and as sentinels of the overall integrity
of ecosystems is recognized in the Clean Water Act as well as in subsequent legislation
and regulations (See Chapter 1). Observations made in streams and rivers can alert
Chapter 2: Listing Candidate Causes
2-1
-------
Stressor Identification Guidance Document
environmental managers or the public to a potential problem. If the biological or
ecological impairment is of sufficient magnitude, it may necessitate identifying the cause
and the potential management controls needed to prevent further damage or to restore the
ecosystem. Observations that might prompt the initiation of a stressor identification
investigation include:
> kills offish, invertebrates, plants, domestic animals, or wildlife,
> anomalies in any life form, such as tumors, lesions, parasites, disease,
> altered community structure such as the absence, reduction, or dominance of a
particular taxon-this can include increased algal blooms, loss of mussels,
increase of tolerant species, etc.,
> loss of species or shifts in abundance,
> response of indicators designed to monitor or detect biological, community, or
ecological condition, such as the Index of Biotic Integrity (IBI) or the
Invertebrate Community Index (ICI),
> changes in the reproductive cycle, population structure, or genetic similarity,
> alteration of ecosystem function, such as nutrient cycles, respiration, and
photosynthetic rates, and
> alteration of the aerial extent and pattern of different ecosystems: for example,
shrinking wetlands, change in the mosaic of open water, wet meadows, sandbars
and riparian shrubs and trees.
It can be important to describe how the observed condition makes the waterbody unfit for
its intended use. This makes the purpose and relative importance of the assessment
clear. For instance, if the fish are covered with lesions, no one wants to fish for them.
In addition to describing the impairment, it is useful to prepare a background statement
articulating the steps taken that revealed the biological impairment. For example, it
might be appropriate to refer to a numerical or narrative biocriterion, or a reference
condition that has been created for this type of waterbody, including the documentation
for its derivation.
If conditions are below expectations, it is important to discuss how the quality or
condition of the stream compares to other streams, or to the same stream in other places
or times. Photographs of the water body provide visual evidence of a lost resource and
can later be used in describing potential pathways that may have lead to the impairment.
Equally important are photographs of what the resource could be like (e.g., taken from
other locations), what it used to be like, or what valued attributes are still retained.
Maps or other geographical representations that show the location and severity of
impairments are essential for orienting the investigators, examining spatial relationships,
and eliciting information from stakeholders (see worksheet in Appendix B, Unit I, page
2-2 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
The scope of the
investigation determines
the extent of the data
sets that will be
analyzed. It defines the
geographic area and
time frame under
consideration and the
types of data that will be
examined.
B-5). Maps can range from simple hand-drawn to computer-
generated versions. Useful geographic information includes
location of the impairment and known point sources, cities,
roads, dams, tributaries, and land use. Examples of maps are
included in the case studies presented in Section 2 of this
document (Chapters 6 and 7). The depiction of this
geographic information is also used to determine the scope of
the evaluation; that is, the overall spatial and temporal extent
of the study.
2.3 Define the Scope of the Investigation
The scope of the investigation influences the selection of
candidate causes, and has ramifications for the final outcome
and the practical use of the entire stressor identification
effort. In a sense, the scope reflects perceptions about the
ecosystem and beliefs about the level of restoration, or
change, that is possible.
The scope of the investigation determines the extent of the
data sets that will be analyzed (see worksheet in Appendix
B, Unit I, page B-4). It defines the geographic area and time
frame under consideration as well as the types of data that
will be examined. The scope of the investigation may be
limited or broad. An example of a limited scope is an
evaluation of whether a particular stressor is responsible for
an impairment. A broader objective would be to evaluate
which, among several candidate causes, could be
responsible for the observed effect. This broader approach
might be appropriate for waters that are not attaining their
designated use, and for which TMDLs (Total Maximum
Daily Loads; see glossary) must be developed.
Early communication
with the stakeholders
will help ensure that
relevant information has
been identified and that
potential causes are
considered.
Several factors influence the overall scope of the investigation, including:
the regulatory context,
the purpose of the investigation,
the relative importance of stressors emanating from outside the watershed,
stakeholder expectations and interests,
logistical constraints,
cost,
personnel, and
available data.
Other factors to consider are the geographic extent of the impairment, and the extent of
knowledge about the impairment. Early communication with the stakeholders will help
ensure that relevant information has been identified, and that potential causes are
considered. After these factors are carefully reviewed, a definition of the geographical
area should be clearly stated. The regulatory context sometimes limits the scope of the
study. For acid rain regulation, the geographical area is very large, whereas an NPDES
violation may involve less than a kilometer of stream reach. The investigators should
Chapter 2: Listing Candidate Causes
2-3
-------
Stressor Identification Guidance Document
document any regulatory authorities involved and discuss the regulatory requirement for
making a causal determination of the impairment.
The depth of the study may be limited by a paucity of data. In this case, it may still be
appropriate to attempt a causal determination with the available data, and then indicate
what additional information is needed to more confidently ascribe the cause.
2.4 Make the List
In developing a list of candidate causes, investigators should consider available evidence
from the case at hand, other similar situations, and knowledge of biological processes or
mechanisms (see text box entitled "Using Existing Programs to List Candidate Causes"
and worksheet in Appendix B, Unit I, page B-6). The causes of ecological condition
usually involve multiple spatial and temporal scales; both of which must be considered in
defining the scope of the study and in listing candidate causes. Recent environmental
events are overlaid on historical events, even those spanning geological time. Global and
regional influences form the backdrop for local factors.
In some cases, two or more
stressors must be present
for the effect to occur.
Where multiple stressors contribute to cause an effect, the
stressor that makes the largest contribution is the
principal cause. Usually a principal cause is so dominant
that removing other causes has no effect on the condition
of the resource. For example, if benthic habitat is both
physically altered and chemically contaminated, restoring
the physical habitat may have no effect until the chemical
contamination is removed. In this situation the chemical contamination is the principal
cause. The habitat alteration is still a cause of impairment, but it is ancillary and masked
by the toxic chemical impact. Nevertheless, pervasive ancillary causes like habitat
alteration, nutrient enrichment, and sediment loading can lower the potential
improvement to the waterbody even after the controlling or principal cause is removed.
Finding and Using Existing Lists of Stressors
Monitoring programs conducted by government agencies and non-governmental organizations
may identify types and levels of stressors. For example, EPA's Environmental Monitoring and
Assessment Program (EMAP) has monitored common stressors found in estuarine systems.1
Among those listed are elevated nutrient concentrations, prolonged phytoplankton blooms, low
dissolved oxygen, and sediment contamination.
State agencies and volunteer monitoring programs may also be good sources of information
on stressors. Maryland's Department of Natural Resources (DNR), for example, maintains a
website on which are links to maps indicating long term trends in total nitrogen, total
phosphorus, and total suspended solids for 3rd order and larger streams in the state of
Maryland.2
1 See EPA "Condition of the Mid-Atlantic Estuaries." Office of Research and Development,
Washington, D.C. #600-R-98-147. November, 1998.
2See Maryland DNR website, http://www.dnr.state.md.us/streams/status trend/index.html
2-4
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
In some cases, two or more stressors must be present for the effect to occur. For
example, a moderate level of nutrients poses no toxicological threat, but if sparse
riparian cover permits sufficient sunlight to allow algal growth, then eutrophication can
occur, with a subsequent cascade of effects. Another example is when a combination of
reduced stream flow and lack of shading cause an elevation of temperature beyond the
limit that native species can tolerate. Stressors acting together to cause an effect should
be listed as a single scenario.
There are some ways to simplify the process of identifying and listing candidate causes.
In the beginning, it helps to make a relatively long list and then pare the list down to the
most likely causes. For the initial long list, it is a good idea to include all stressors
known to occur in a waterbody. Even if these stressors have not previously been shown
to cause this type of impairment, someone is likely to want proof that they were not
causal agents. Include stressors that stakeholders have good reason to believe may be
important. Consult other ecologists for potential causes of the impairment.
Knowledge about pollution sources near the waterbody can also suggest potential
stressors. Point sources, such as drainage pipes, outfalls, and ditches are easily identified
as sources. Constituents of the effluent can be listed as candidate causes. Other sources
may be located some distance from the resource, such as motor vehicles and smoke
stacks that generate candidate causes such as acid rain or nitrogen enrichment. Particular
land uses often generate a consistent suite of stressors. For example, siltation and
pesticides are commonly associated with agriculture. Locations of sources and stressors
should be added to the impairment maps developed in Section 2.2.
Once an exhaustive list of candidate causes is developed, the next step is to pare the list
down. Including very unlikely causes can make the identification process unwieldy and
will distract stakeholders and managers from the more likely candidates. Unlikely
stressors are those that are believed to be mechanistically implausible or absent from the
watershed. Although they need not be evaluated, we recommend that you document the
rationale for not including the less likely causes.
2.5 Develop Conceptual Models
The final part of this initial step is to develop conceptual models for the candidate
causes, linking the cause with the effect (see worksheet in Appendix B, Unit I, page B-6).
This part of the process documents a likely explanation of how the stressor could have
caused the impairment. Conceptual models provide a good way to communicate
hypotheses and assumptions about how and why effects are occurring. Models can also
show where different causes may interact and where additional data collection may
provide useful information.
Conceptual models will vary in complexity, depending on the
The conceptual model mechanisms and ecological processes involved. A
..... ,. , generalized conceptual model might show land uses in the
can help the investigator & F 5 .
watershed that generate m-stream stressors impacting valued
see the pathway resources. For instance, if fish communities are impacted by
, , ,. ... , moderate levels of nutrients in a sunlit stream, it is important
between the candidate ' ,
to show that the effect could have occurred via several
cause and the eventual possible pathways, or a combination of pathways, such as:
impact.
Chapter 2: Listing Candidate Causes 2-5
-------
Stressor Identification Guidance Document
> decaying algal blooms that result in low dissolved oxygen,
> the dominance of prey, causing a change in abundance of species,
> conditions favorable for opportunistic pathogens,
> diatom-rich water that is so turbid that sight-feeding fish cannot find prey and
starve, and
> embedded substrates smothered with decaying and overgrown algal mats that
reduce habitat for foraging, refugia, and reproduction.
The primary causes in this example are nutrients and incident sunlight. The secondary
cause in the pathway could be any of the stressors that are formed from the initial cause.
It is usually a good idea to consult with ecologists experienced with similar streams
when developing conceptual models, especially when complex pathways and ecological
process are involved.
Using a pictorial, poster-style conceptual model is useful to introduce the ecological
relationships. Then a box and arrow diagram can be used to show details of the
relationships among stressors, receptors, and intermediate processes. Some models get
too complicated to be helpful. The diagram should show only the pathways and causes
considered in the study. Separate diagrams for each stressor or pathway can keep the
focus on the analysis steps that will follow. Figure 2-1 is an example of a box and arrow
conceptual model illustrating the impacts of logging on salmon production in a forest
stream. Additional examples and advice on conceptual model development can be found
in Jorgensen (1994), Suter (1999), Cormier et al. (2000c), USEPA (1998a) (especially
Appendix C), and in the case studies shown in Chapters 6 and 7.
In addition to helping the investigators to elucidate the relationships among multiple
cause and multiple effects, conceptual models are also powerful tools for communicating
among the investigative team and obtaining additional insights from stakeholders and
managers.
2-6 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Adult Salmon
I
^ Spawning ^
1
Eggs
1
^ Hatching y
1
Alevins
y
{ Survival y
y
Fry
1
/ \
^ Survival y
\ /*
1
Fingerlings
, + V
/ \
^ Survival y
\ y*1
1
Juveniles
^~\Mig
;^
^
C
^
^
^
ln \ A A
ration / 1 Logging I
/ Soil \ / Canopy \
VDisturbancey \DisturbanceJ
V
f Buffer Zone 1 *
i * x^ ^x x Solar \
^JL*^ ^ \ Radiation/
/Sediment\^_ Sediment
^ "" Water Primary
Temperature Production
^
Terrestrial
Detritus
^m ^M
/ \
^ Dissolved ^ / nprnmnncitinnN
^ Oxygen ^ \ / \
\
; |
^
K\
Out \
Migration /
Figure 2-1. A conceptual model for ecological risk assessment illustrating the effect of
logging in salmon production in a forest stream. (The assessment includes a series of
exposures and responses. In the diagram, the circles are stressors, the rectangles are
states of receptors, and the hexagons are processes of receptors. The rectangle with
rounded corners is an intervention, establishment of buffer zones, that is being
considered (Suteretal. 1994).)
Chapter 2: Listing Candidate Causes
2-7
-------
Stressor Identification Guidance Document
Chapter 3
Analyzing the Evidence
In this Chapter:
3.1 Introduction
3.2 Associations Between Measurements of
Candidate Causes and Effects
3.3 Using Effects Data from Elsewhere
3.4 Measurements Associated with the
Causal Mechanism
3.5 Associations of Effects with Mitigation or
Manipulation of Causes
L
Stressoi\ldentification
LIST CANDIDATE CAUSES/
ANALYZE EVIDENCE
3.1 Introduction
The second step in the SI process is to analyze
the information that is related to each of the
candidate causes identified in Chapter 2.
Virtually everything that is known about an
impaired aquatic ecosystem and about the
candidate causes of the impairment may be
useful for inferring causality. Potentially
useful data that may come from studies of the
site include chemical analysis of effluents,
organisms, ambient waters, and sediments;
toxicity tests of effluents, waters, and
sediments; necropsies; biotic surveys; habitat
analyses; hydrologic records; and biomarker
analyses. A similar array of data may be
obtained from other sites and from laboratory
studies (performed ad hoc or reported in the
literature). However, these data do not in themselves constitute evidence of causation.
The investigators performing the causal analysis must
organize and analyze the data in terms of associations that
Existing data are often might support or refute proposed causal scenarios.
CHARACTERIZE CAUSES
Strength of Evidence
Identify Probable Cause
sufficient to determine
the cause of impairment.
The SI process does not require a minimum data set, and
existing data are often sufficient to determine the cause of
impairment. However, the investigator has the responsibility
of evaluating whether the data used are sufficient to support
the SI process. If the investigator decides to generate additional data, its quality must be
assured (see text box entitled "Data Quality Objectives").
The primary inputs to the analysis step are the list of candidate causes and the associated
conceptual models that link the causes with the observed effects (developed in Chapter
2). Other inputs include data and information that come from the case at hand, other
similar cases, the laboratory, and the literature that synthesizes biological and ecological
knowledge (Figure 3-1). In the analysis step, this information is converted into causal
evidence that falls into four general categories of relationships:
1. associations between measurements of the candidate causes and effects
(Section 3.2),
2. associations between measures of exposure at the site and measures of
effects from laboratory studies (Section 3.3),
3. associations of site measurements with intermediate steps in a chain of
causal processes (Section 3.4), and
Chapter 3: Analyzing the Evidence
3-1
-------
Stressor Identification Guidance Document
4. associations of cause and effect in deliberate manipulations of field
situations or media (Section 3.5).
The evidence produced in the analysis step is used to characterize the cause or causes of
the observed effect (see Chapter 4). The analysis and characterization of causes is
usually done iteratively and interactively, as illustrated by the two-way arrows between
the analysis and characterization boxes in Figures 1-1 and 3-1. Evidence is brought in
and analyzed as needed until there is sufficient confidence in the causal characterization.
In straightforward cases, the process may be completed in linear fashion. In more
complex cases, the causal characterization may require additional data or analyses, and
the investigator may repeat the process.
Data Quality Objectives
If new data will be generated for an SI investigation, consider following U.S. EPA's Data Quality
Objectives (DQO) process. The DQO process combines a problem formulation exercise with
conventional sampling statistics to determine the type, quantity, and quality of data needed to
make an environmental decision with a desired probability of error (Quality Management Staff
1994). The DQO process is not directly applicable to SI since it is designed to determine the
probability of exceeding a threshold. However, using a formal process to define the problem,
examine information needs, and determine study boundaries is important in planning any
sampling and analysis program. The criteria for defining an optimum design for an SI study will
vary depending on the circumstances. Following sampling and analysis, a Data Quality
Assessment (DQA) should be performed to determine whether the goals of the DQO process
have been achieved (Quality Assurance Division 1998). The EPA's Quality System, including
requirements for non-EPA organizations, can be found at www.epa.qov/quality/index.html.
Quality Assurance Division. 1998. Guidance for Data Quality Assessment. EPA QA/G-9, QA97
Version, or EPA/600/R-96/084. U.S. EPA, Washington, D.C.
Quality Management Staff. 1994. Guidance for the Data Quality Objectives Process. EPA
QA/G-4, or EPA/600/R-96/055. U.S. EPA, Washington D.C.
3.2 Associations Between Measurements of Candidate Causes and
Effects
The first type of evidence of causation is associations among measurements of candidate
causes and effects (Table 3-1). The objective of this analysis is to provide evidence that:
> the candidate cause and the effect are observed at the same time or place,
> when the candidate cause is not observed, the effect is also not observed, or
> the intensity of the causal factor is related to the magnitude of the effect.
Causal evaluations often begin by examining associations from the case at hand. For
example, effects are observed downstream, but not upstream of a candidate cause. These
associations provide the core of information used for characterizing causes (see
worksheet in Appendix B, Unit II, page B-7). Associations may be revealed by plotting
data on common axes, as shown in Figure 3-2. In this figure, the spatial pattern of a
toxicity bioassay results are clearly associated with the spatial pattern of a community
metric. Causal inference is easier when the stressors and effects are located together (co-
3-2
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
located) in time and space. Inference becomes more difficult
as stressors are dispersed over larger scales, occur
intermittently, or cannot be measured. Inference is also more
difficult when there is a time lag between exposures. For
example, if a stressor, such as a diversion of water flow
prevents salmon from reaching the sea on their out-migration,
the effect (i.e., destruction of the salmon run) may not be
observed until three years later. In some cases, models may
be useful for extrapolating inferences from available measurements.
Causal evaluations often
begin by examining
associations
from the case at hand.
List Candidate Causes
\ \
^L^
Analyze Evidence
Associate measurements of candidate causes wit
observed effects
Associate effects with mitigation or manipulation o
causes
Combine measurements of site exposures with
effects data from elsewhere
Associate measurements with the causal
mechanism
n
Characterize Causes
h
f
\-
r
"""
1.
h
Acquire Data
From Other
Similar Cases
, ,
1 Laboratory/ '
Literature .
1 '
*
1
*
1
1
Figure 3-1. The flow of information from data acquisition to the analysis phase of the SI
process.
Chapter 3: Analyzing the Evidence
3-3
-------
Stressor Identification Guidance Document
Table 3-1. Types of associations between measurements of causes and effects
among site data and the evidence that may be derived from each.
Type of Association
Spatial co-location
Spatial gradient
Temporal relationship
Temporal gradient
Example Evidence
Effects are occurring at same place as exposure
Effects do not occur where there is no exposure
For candidates with discrete sources on streams and
rivers:
Effects occur downstream of a source
Effects do not occur upstream of a source
For candidates with dispersed sources:
Effects occur where there is exposure, but not at
carefully matched reference sites where exposure
does not occur
Effects decline as exposure declines over space
Exposure precedes effects in time
Effects are occurring simultaneously with exposure
(allowing for response and recovery rates)
Intermittent sources are associated with intermittent
exposure and effects
Effects increase or decline as exposure increases or
declines overtime
100 i
80
Z
O
O 60
Q
LJJ
TOXICITYDATA
O COMMUNITY DATA
40
LJJ
O
OH 20
LJJ
Q_
3456
STREAM STATIONS
8
9
Figure 3-2. Plot of toxicity data from a 7-day subchronic test of ambient waters and a
community metric obtained on a common stream gradient (Norberg-King and Mount
1986).
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
The evaluation of associations must consider whether potentially affected organisms may
have moved since exposure. It is helpful to consider the mobility of organisms relative
to the extent of the observed exposed and unexposed reaches or areas. Clearly, fish are
capable of swimming long distances and invertebrates may drift downstream or fly
upstream. However, extensive experience with bioassessment offish and invertebrate
communities has demonstrated that the movements of these organisms are usually not so
great as to prevent the observation of spatial associations. The movement of a few
individual organisms from contaminated reaches to upstream reaches will diminish, but
generally not eliminate, the contrast or gradient among reaches. However, salmon and
other species that regularly move long distances require special consideration when
analyzing spatial associations. In such cases, consider the logic of the situation and
possibly use a GIS as a platform for modeling spatial relationships.
Obtaining measurements of the stressor that can be associated with the effect can be
challenging. In the most straightforward cases, the
measurements of the stressor itself are available; for example,
nutrient concentrations, degree of siltation, dissolved oxygen In some cases, the
concentrations, or chemical concentrations. In some cases, the .
candidate cause is the lack of a required resource, such as canaiaate cause is tne
nesting habitat. In these cases, measurements can establish that lack of a required
the resource is indeed missing at the place and time it would be
required by an organism. When measurements of the stressor resource, sucn as
are not available, surrogates can be used, although the nesting habitat
uncertainty in the analysis will increase. Information on the
location and attributes of sources can be useful surrogates.
This information can be particularly important for stressors that are intermittent in nature
(e.g., high flow events), or degrade quickly (e.g., some pesticides). In these cases, source
information may be used as a surrogate for the stressors. As sources become larger in
scale and more diffuse, information on the sources becomes more difficult to use in site-
specific causal evaluation.
Similarly, measuring the immediate or direct response to a stressor increases the
confidence in a causal evaluation. For example, a fish kill may be associated with
nutrient enrichment, acting through algal growth, decomposition, and oxygen depletion.
Measurements of the initial algal growth and oxygen depletion would increase an
investigator's confidence that nutrient enrichment was the cause of the fish kill.
Conceptual models are very useful for illustrating linkages between complex pathways
of cause and effects, and for illustrating where measurements are (and are not) available.
Whenever possible, associations should be quantified. For categorical data, calculate the
frequencies of associations. For count or continuous data use, linear or nonlinear
models. For example, the abundance of Ephemeroptera at a site may be regressed
against concentration of total sediment PAHs. Similarly, the community data plotted in
Figure 3-2 might be regressed against the toxicity data. If effects data are categorical or
heterogeneous and exposure data are continuous, categorical regression may be used
(Dourson et al. 1997). Select the analysis technique that best illuminates the association,
based on the amounts and types of data available. Some statistical descriptions of the
associations include correlation coefficients, confidence intervals, and p-values.
However, avoid statistical hypothesis testing of the associations (see text box entitled
"Using Statistics and Statistical Hypothesis Testing for Analyzing Observational Data in
Stressor Identification"). Because groups are not randomly assigned in a way that
minimizes the influence of confounding variables, a significant outcome in a hypothesis
test may be falsely attributed to a candidate cause, when in fact it is due to another
Chapter 3: Analyzing the Evidence 3-5
-------
Stressor Identification Guidance Document
factor. On the other hand, the small sample sizes that are usually seen in these studies
decrease the ability to statistically discriminate groups, and may lead to mistakenly
eliminating a true cause.
Often associations between candidate causes and effects can be improved by identifying
and isolating confounding factors in either the receptors or the environment. For
example, the frequency of hepatic neoplasms in fish is associated both with the age
structure of the fish population and the concentration of PAHs in sediment (Baumann et
al.1996). Correction for age offish would increase the consistency and, potentially, the
biological gradient in the relationship between hepatic neoplasm frequency and industrial
contaminants. Similarly, a decline in fish species richness is a common measure of
impairment, but the number of species present generally increases with increasing stream
size (e.g., OEPA 1988a). Therefore, including a correction for stream size could
strengthen the association between the degradation and species loss.
Associations observed from other studies can provide useful supporting information,
particularly when the specific type or constellation of effects is consistently observed in
association with a candidate stressor. Keep in mind that, as evidence, associations
observed from other sites is not as strong as those observed from the study site.
Therefore, if associations of effects and potential causes are analyzed at other sites, they
should be evaluated separately from those at the site of concern.
3.3 Using Effects Data from Elsewhere
Measures of exposure from the case at hand can also be matched with measures of effect
from other situations. The objective of this analysis is to provide evidence showing that
the stressor is present at the study site in sufficient quantity or frequency that the
investigator would expect to see a particular effect based on effect information from
laboratory tests, field tests, or exposure-response relationships developed at other sites
(see worksheet in Appendix B, Unit II, page B-12). This type of evidence is familiar to
ecotoxicologists who combine measures of exposure from the study site with measures
of effect from laboratory tests. For example, concentrations of chemicals measured in
water may be compared to concentrations that are thresholds for effects in toxicity tests,
or they may be used in concentration-response models to estimate the frequency or
magnitude of effects. When doing these comparisons, the investigator should keep in
mind that laboratory conditions or organisms may not accurately represent field
conditions or organisms.
Equivalent measures of exposure and effects are available for non-chemical stressors
(Table 3-2). As in toxicological assessments, it is important to choose the most
applicable high-quality effect measurements. It is also important to ensure that the
measures of exposure and effects are consistent. For example, long-term field exposures
are most appropriately compared with chronic test data. In some cases, exposure-
response information will not be available for a candidate cause, but will be available for
an analogous agent, such as an effluent with a structurally similar chemical or an
introduced species with similar feeding behavior.
3 -6 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Using Statistics and Statistical Hypothesis Testing
for Analyzing Observational Data in Stressor Identification
Statistical techniques are essential tools for summarizing and analyzing environmental measurements for SI.
Good SI uses a variety of techniques, including descriptive statistics (e.g., means, ranges, variances),
exploratory statistics (e.g., multivariate correlations), statistical modeling (e.g, exposure-response relationships),
quality assurance statistics (e.g., accuracy and precision of analyses of duplicates and standard reference
materials) and comparison of alternative models of candidate causes (e.g., goodness-of-fit or maximum
likelihood). However, the use of statistical hypothesis tests is problematic. Statistical hypothesis testing was
designed for analyzing data from experiments, where treatments are replicated and randomly assigned to
experimental units that are isolated from one another. The application of these tests to data from observational
studies can result in erroneous conclusions. In observational studies, treatments are very seldom replicated and
are never randomly assigned to experimental units.
If experimental units are replicated at all, they are replicated within the same water body and hence are likely to
influence one another. As a result, samples are replicated rather than treatments. This is known as
pseudoreplication (Hurlbert 1984). Finally, the location of a candidate cause is a given, rather than being
randomly placed, so it is likely that candidate causes will co-vary with each other and with important natural
attributes of the system (e.g., salinity, depth). The following table summarizes several common analytical
techniques and discusses their use in SI.
Activity
Application
to
observational
data in SI
Comments
Using summary statistics
(e.g., mean water
concentrations, 7Q10 flow
rates) to summarize
measurements
Encouraged
Pay attention to the biological or physical relevance of the
summary statistic used. For example, the mean of chemical
concentrations overtime is often the most relevant (USEPA
1998a). As another example, the bankfull flow event is
considered to be an important determinant of stream
morphology (Rosgen 1996).
Using statistics to determine
the probability that two sets
or samples are drawn from
the same distribution, or
that they differ by a
prescribed amount
Use Caution
Note that this use is not hypothesis testing in that it does not
test a null hypothesis about a treatment (cause). It simply
tells you the likelihood that differences are due to sampling
variance. Also, the conventional criteria for statistically
significant differences are not relevant; the differences must
be shown to be biologically significant and the probabilities
must be shown to affect the overall strength of evidence.
Because the sample sizes are often small relative to
variance, the power to detect real differences may be small.
Using the results of
statistical hypothesis tests
to conclude that a candidate
is (or is not) the cause
Wrong
The assumptions of statistical hypothesis testing are
violated. In observational studies, replicate treatments
cannot be randomly assigned in a way that minimizes the
influence of confounding variables. For this reason, a
significant outcome in a hypothesis test may be falsely
attributed to a candidate cause when in fact it is due to
another factor.
Using correlations or
regression techniques to
quantify relationships
between variables.
Encouraged
The type of data (continuous, ordinal, or categorical) and the
type of relationship (e.g., linear, non linear) will determine
the best technique to use.
Using statistics to determine
the probability that a
relationship is nonrandom,
or that the slope of a
regression differs from zero.
Use Caution
Note that this analysis indicates only the probability that an
apparent relationship is due to sampling variance. It does
not test the hypothesis that the relationship is causal. Also,
the number of samples is likely to be low, so even
correlations or models that are not statistically significant
can be biologically significant and contribute to the strength
of evidence.
Concluding that statistically
significantly correlated
variables have a causal
relationship
Wrong
Correlation does not indicate causation, and a highly
improbable regression model does not indicate that the
independent variable caused the relationship. Because
stressors often covary with each other and with natural
environmental attributes, a strong relationship between a
candidate cause and a biological variable may be due to a
factor other than the candidate cause.
Chapter 3: Analyzing the Evidence
3-7
-------
Stressor Identification Guidance Document
Table 3-2. Example associations between site-derived measures of exposure and
measures of effects from controlled studies for different types of stressors.
Stressor
Chemical
Effluent
Contaminated
Ambient
Media
Habitat
Water
withdrawal/
drought
Thermal
energy
Siltation
(suspended)
Dissolved
oxygen and
oxygen-
demanding
contaminants
(e.g. BOD,
COD)
Siltation
(bed load)
Excess
mineral
nutrients
Pathogen
Non-
indigenous
invasive
species
Characterization of
Exposure:
Intensity, Time, and Space
External concentration in
medium
Internal concentration in
organism
Biomarker
Dilution of effluent
Location and time of
collection
Analysis of medium
Structural attributes
Hydrograph and associated
summary statistics (e.g.,
7Q10)
Temperature
Suspended concentration
(e.g., TSS)
Dissolved Oxygen
Degree of embeddedness,
texture
Dissolved concentration
Presence or abundance of
pathogen
Presence or abundance of
the species
Characterization of
Exposure-Response
Concentration-response or time-
response relationships from
laboratory or other field studies
Effluent dilution - response in the
laboratory (WET)
Lab or in situ tests using the
medium:
Medium dilution - response
Medium gradient - response
Empirical models (e.g., Habitat
suitability models)
In-stream flow models (e.g., IFIM)
Thermal tolerances
Concentration-Response
relationships from laboratory or other
field studies
Oxygen concentration-response
relationships from laboratory or other
field studies.
Empirical siltation-response
relationships from laboratory or other
field studies.
Empirical concentration-response
relationships from laboratory or other
field studies.
Eutrophication models
Disease, Symptoms
Ecological models (food web,
energetics, predator-prey, etc.)
3-8
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
In developing
mechanistic conceptual
models depicting the
induction of effects, it is
often apparent that there
are intermediate steps in
the causal process that
may be observed or
measured.
Laboratory toxicity tests and other controlled studies provide
the bases for models depicting the induction of effects by
particular causes. For example, an acute lethality test of a
chemical provides a concentration-response model which may
be used to determine whether fish kills might be attributable
to observed or estimated ambient concentrations. More
complex causal mechanisms, particularly those involving
indirect causation, require more complex mechanistic models.
As models of causal processes become more complex, it
becomes more difficult to judge whether an individual model
provides an acceptable representation of the causes of
ecological degradation at a site. In such cases, the best
strategy is to generate mechanistic models of each proposed
causal scenario and determine which model best explains the
site data (Hilborn and Mangel 1997).
3.4 Measurements Associated with the Causal Mechanism
In developing mechanistic conceptual models depicting the induction of effects, it is
often apparent that there are intermediate steps in the causal process that may be
observed or measured. Documenting those intermediate steps increases confidence in
the proposed causal mechanism (see worksheet in Appendix B, Unit II, page B-8). This
type of evidence is particularly useful when the ultimate effects of multiple candidate
causes are similar, but act through different mechanistic pathways. Types and examples
of intermediate steps are presented in Table 3-3. In some cases it is sufficient to
document the occurrence of the intermediate step, but in many cases, the level of the
metric must be shown to be adequate. For example, if competition for prey by an
introduced species is the proposed mechanism by which an endpoint species has been
lost, then the investigator should show that the number of prey are reduced sufficiently.
Table 3-3. Example associations between site data and the processes by which
stressors induce effects.
Type of Measurement
Symptoms (i.e., responses
specific to, or characteristic
of, a type of stressor and
causing the overt
impairment)
Biomarkers
Intermediate product of an
ecological process
Changes in abundance of
predators, prey, or
competitors
Example Mechanistic Association
Fish have lesions characteristic of a bacterium
Metalothionine induction is an intermediate step in the
glomerular toxicity of cadmium
Algal abundance and DO are measures of intermediate
steps in the induction offish kill by nutrient additions
Abundance of prey decreases upon introduction of a
new predator
Chapter 3: Analyzing the Evidence
3-9
-------
Stressor Identification Guidance Document
Table 3-3 (continued). Example associations between site data and processes by
which stressors induce effects.
Type of Measurement
Effects on other receptors
Distributions of stressors
and receptors coincide
Example Mechanistic Association
If impairment is defined in terms of effects on fish, then
the responses of invertebrates or plants may suggest
what causes are operating
For a stressor to cause an effect, it must contact or co-
occur with the receptor organisms. For causes that act
through the deprivation of a resource, the deprivation
must actually occur
3.5 Associations of Effects with Mitigation or Manipulation of Causes
Strong causal evidence can be provided by deliberately eliminating or reducing a
candidate cause and noting whether the effects disappear or remain (see worksheet in
Appendix B, Unit II, page B-10). Causes can be eliminated as a part of a field
experiment or by bringing site media into the laboratory (Table 3-4). Field experiments
may also be performed by manipulating the source (see text box entitled "Associating
Effects with Mitigation or Manipulation of a Cause"). For example, cattle may be
fenced away from some locations where they usually have access to a stream channel, or
an effluent may be eliminated for a time due to plant shut-down. These experiments may
be conducted at the site being assessed, or may be conducted at other sites where the
same type of source operates. Occasionally, a regulatory or remedial action may be
treated as an experimental manipulation. Alternatively, experiments may be conducted
that control the exposure of organisms or communities to potential causes. Examples
include caging previously unexposed organisms at contaminated locations, placing
containers of uncontaminated sediments in locations with contaminated water. These
field experiments typically cannot be replicated, so their results are potentially subject to
confounding (see text box "Using Statistics and Statistical Hypothesis Testing for
Analyzing Observational Data in Stressor Identification"). Finally, site media can be
brought into the laboratory and manipulated to eliminate different candidate causes.
Then the results of the manipulation can be tested using laboratory organisms. These
methods have been most extensively developed for the purpose of attributing causality
among different chemicals in effluents.
Table 3-4. Types of field experiments and the evidence that may be derived from each.
Example Experiment
Manipulation of a source
in the field
Manipulation of exposure
in the field
Laboratory manipulation
and testing of media from
the case
Example Evidence Derived from the Experiment
Elimination of a source reduces or eliminates the effect.
Introduction of previously unexposed organisms results in
effects.
Isolation of organisms from one cause reveals the effects
of others.
Extracting site media into fractions containing different
chemical classes results in toxicity being associated with
only one fraction.
3-10
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Associating Effects with Mitigation or Manipulation of a Cause
Biological data collected by the Kansas Department of Health and Environment (KDHE) have
played an increasingly important role in the state's efforts to document water quality
impairments. KDHE historically has applied a modification of Davenport and Kelly's (1983)
macroinvertebrate biotic index (MBI) to identify impairments resulting from nutrient loading and
organic enrichment. Recently, a genus- and species-level indicator known as the Kansas
Biotic Index (KBI) was developed to specifically respond to different stressor categories,
including nutrients and oxygen demanding substances (KBIorg). Data collected by KDHE have
shown that declines in the MBI and KBIorg have been consistently associated with increased
organic enrichment, nutrient loading, and ammonia contamination.
The MBI and KBIorg were used to document the association between effects and the
mitigation or manipulation of causes. After a nitrification process was installed at the city of
Wichita's municipal wastewater treatment facility, median concentrations of total ammonia-
nitrogen in the Arkansas River decreased from 1.1 mg/L (1982-91) to 0.06 mg/L (1992-99).
Concomitant decreases in the upper quartile MBI and KBIorg values were sufficiently large to
justify a formal change in the Arkansas River's 305(b) impairment status. Moreover, city
officials documented the recolonization of this river by several rare or previously extirpated fish
species. Comparable improvements in MBI and KBIorg scores were documented in the Smoky
Hill River below the city of Salina sewage treatment plant after ammonia levels were reduced
by implementing wastewater nitrification and an industrial pretreatment initiative.
Outcome
In the 2000 KDHE 305(b) assessment, the Smoky Hill River was upgraded from non-supporting
to fully supporting of aquatic life.
References
Davenport, E. and H. Kelly. (1983); Huggins, G. and F. Moffett. (1988); KDHE. (1993, 1998,
2000).
Chapter 3: Analyzing the Evidence
3-11
-------
Stressor Identification Guidance Document
Chapter 4
Characterizing Causes
4.1 Introduction
Characterizing causes involves using the
evidence analyzed in Chapter 3 to reach a
conclusion and to state the levels of confidence
in that conclusion. The input information in
this process includes a description of the
effects to be explained, the set of candidate
causes developed in Chapter 2, and the causal
evidence analyzed in Chapter 3.
4.2 Methods for Causal
Characterization
In this Chapter:
4.1 Introduction
4.2 Methods for Causal Characterization
4.3 Identify Probable Cause and Evaluate
Confidence
Stressor^dentification
T
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
\ /
CHARACTERIZE CAUSES
Eliminate Diagnose Strength of Evidence
Identify Probable Cause
After available evidence has been compiled and analyzed, the cause(s) may be obvious.
In other cases, a more systematic method for reaching a conclusion may be needed. The
use of clearly documented inferential logic increases the defensibility of causal
attribution. This chapter describes three methods for using the evidence developed in
Chapter 3 to characterize the cause: (1) eliminating alternatives, (2) using diagnostic
protocols, and (3) weighing the strength of evidence supporting each candidate cause.
Figure 4-1 depicts a procedure that combines these multiple methods to reach a
conclusion of causality. Although this approach uses a combination of methods for
characterizing causes, each method may also be used independently.
This integrated approach does not include all possible methods of causal analysis,
particularly the use of expert judgment. When evidence is ambiguous, the process of
developing consensus among a panel of experts may be more acceptable to stakeholders
than any systematic evaluation of evidence. Utilizing expert judgment is certainly a
more flexible approach in that it does not require any particular data set or type of model.
In addition, experts can reach conclusions on the basis of experience and pattern
recognition. For example, an experienced extension agent
may visit a farm pond that is not producing bass and, without
taking any measurements, know that the pond is too small or
receives too much manure runoff from surrounding pastures
to support bass reproduction. However, when the issue of
causation is contentious, the attempt to develop consensus
may be complicated by experts who represent the interests of
the contending parties. Even when the experts are neutral,
expert consensus may not be acceptable to some parties due
to subjectivity. Finally, the process of developing expert
consensus may not be practical. An NIH consensus
development conference or an NRC panel may be practical
for large-scale issues, such as the carcinogenicity of
electromagnetic fields. It may not be practical to convene an expert panel for each
outfall causing ecological injuries.
Although this approach
uses a combination of
methods for
characterizing causes,
each method may also
be used independently.
Chapter 4: Characterizing Causes
4-1
-------
Stressor Identification Guidance Document
Causal
Evi
List of Candidate Causes
and Conceptual Models
7
1 *
l
Yes
r
Diagnostic
Analysis
Figure 4-1. A logic for characterizing the causes of ecological injuries at specific sites.
(Processes are rectangles, and the three inferential methods have heavy borders.
Decisions are diamonds, and inputs are parallelograms.)
4-2
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
CHARACTERIZE CAUSES
E
liminate |
Diagnose
Strength of Ev
Identify Probable Cause
dence
Inputs to the characterization process (the parallelogram at the top of Figure 4-1) include
a description of the effects to be explained, the list of candidate causes, and the
associated conceptual models (Chapter 2). The set of candidate causes should include
stressors that consist of multiple factors that act together and are not individually
sufficient to cause the effect (i.e., causal scenarios). Other inputs to characterization
include the causal evidence produced in the analysis step (the three parallelograms on the
left side of Figure 4-1). As discussed above, analyses are usually conducted in
combination, as needed, throughout the characterization process. For example, the
evidence necessary for eliminating candidate causes is analyzed first, then evidence for
diagnosis, and, finally if necessary, the strength of evidence for each candidate cause is
analyzed.
4.2.1 Eliminating Alternatives
The causal characterization methods shown in
Figure 4-1 are presented in order, from the
most conclusive to the least conclusive. The
first method, eliminating alternatives, is a
powerful approach to evaluating information.
The ability to eliminate all but one alternative
is a strong standard of proof for causality, and
it is easily understood and widely practiced.
It is the basic technique of literature's most
famous master of inference, Sherlock
Holmes:
"When you have eliminated the impossible, whatever remains, however
improbable, must be the truth."
(Sir Arthur Conan Doyle, Sign of Four, 1890).
Elimination is also an effective way of reducing the numbers of alternatives to be
considered before using another method (e.g., strength of evidence, Section 4.1.3, and
see worksheet in Appendix B, Unit III, page B-15). Eliminating evidence is a
particularly good option for SI when the set of alternatives is limited, and when disproof
does not rely on statistics (see text box in Chapter 3 entitled "Using Statistics and
Statistical Hypothesis Testing for Analyzing Observational Data in Stressor
Identification"). Specifically, if the SI is conducted to support a permitting action,
logical elimination of the permitted source as a potential cause of the observed injury is a
sufficient causal analysis. Because of the complexity associated with ecological systems
and multiple stressors, many SI investigations will not have the evidence necessary to
confidently eliminate causes. These evaluations will rely on a strength of evidence
analysis (Section 4.1.3).
Elimination as a method for establishing causality has strong roots in the philosophy of
science. Popper, Platt, and other conventional philosophers of science have argued that
it is logically impossible to prove a hypothesized relationship, but it is possible to
disprove hypotheses (Platt 1964, Popper 1968). If a set of possible causes has been
identified, once all but one alternative has been eliminated, the remaining hypothesis
must be true. For example, if a body of water is found to be acidic, it is possible to
establish the cause as acid deposition by eliminating acid mine drainage, geologic
sulphate, and biogenic acids as causes (Thornton et al. 1994).
Chapter 4: Characterizing Causes
4-3
-------
Stressor Identification Guidance Document
The elimination of alternatives has three major limitations:
> Due to limited knowledge, it may not be possible to identify a complete set of
candidate. Also, the array of possible causes is potentially infinite, as there is no
clear boundary between plausible and absurd hypothetical causes (Susser 1986b,
Susser 1988).
> The process of elimination is limited by the ability to perform reliable tests and
obtain unambiguous results. Such tests are often difficult in ecology. One may
fail to reject a hypothesis but be uncertain of that result due to sampling
variance, biases, and temporal variance. If all but one cause is rejected on
uncertain grounds, it is difficult to accept the remaining candidate cause with
confidence.
> Elimination of causes should be done with particular care when multiple
sufficient causes may be operating. The evidence for one cause may be so strong
that it masks the effects of another sufficient cause and appears to be the sole
cause. In addition, beware that the temporal sequence of cause and effect may
appear to be wrong when one sufficient cause precedes another. For example, an
industrial effluent may impair a biological community. If the stream is
subsequently channelized, the effects would be obscured by the industrial
effluent. The channelization would have been sufficient to degrade biological
communities within a pristine stream and therefore should be retained as a
candidate cause. As shown in Table 4-1, similar issues are also relevant to
spatial sequences such as those occurring in streams or rivers.
Most often the objective of SI is to identify all sufficient causes (for example, when the
goal is to remediate or restore a water body). In these cases, the elimination step should
be performed iteratively. That is, each cause eliminated during the first round should be
reevaluated to determine if its effects may have been masked by another cause. If so, the
candidate cause should be retained. In extreme cases, the masked secondary causes will
remain unidentified, because the primary causes are so conspicuous. For example, if
channelization has eliminated nearly all fish, it may not be apparent that episodic
pesticide runoff would affect sensitive species. Such occult secondary causes will
become apparent only after the primary causes have been remediated.
Some types of evidence can be used to eliminate candidate causes, and when those
causes might be retained because of masking. Only associations derived from
measurements taken from the case under evaluation are strong
enough to eliminate an alternative. Associations derived from " "
similar cases cannot be used to eliminate alternatives, but are Only associations
useful in strength of evidence analyses which allow for derived from
uncertain or indecisive evidence (Section 4.1.3).
measurements taken
A stressor can be confidently eliminated if case-specific from the case under
measurements clearly show that a necessary step in the causal
chain of events has not occurred. For example, if a chemical evaluation are strong
must be taken up by an organism in order to cause an effect, enough to eliminate an
and it can be demonstrated that uptake has not occurred (e.g.,
though biomarkers or body burdens), the chemical can be alternative.
eliminated as a cause. Similarly, if sedimentation causes
4-4 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
effects by silting-in riffles, and riffles can be demonstrated to be free of silt,
sedimentation can be eliminated as a cause.
Although another potential way to eliminate a candidate cause is through experimental
manipulations, the results of field experiments are seldom sufficiently conclusive to
eliminate a cause. Uncertainties exist in field experiments due to a lack of thorough
knowledge of recovery and recolonization rates following exposure. As a result,
reduction or elimination of exposure may not appear to eliminate the effects. Field
experiment data can, however, be used in the strength of evidence analysis discussed in
Section 4.1.3. In addition, removal of one sufficient cause may unmask the effects of
another. The protocols associated with the Toxicity Identification and Evaluation (TIE)
program can be applied here, but not all effects of concern occur in these tests (e.g.,
tumors). Further, there may be questions concerning the sensitivity of the 7-day tests and
test species relative to field durations and species (USEPA 1993a,b). TIE, therefore, is
considered as part of the strength of evidence analysis.
Table 4-1. Application of common types of evidence in eliminating alternatives.
Type of Evidence
(See Chapter 3)
Associations
between
measurements of
candidate causes
and effects: Did
the stressor
precede the effect
in time?
Associations
between
measurements of
candidate causes
and effects: Is
there an
upstream/
downstream
conjunction of
candidate cause
and effect?
Associations
between
measurements of
candidate causes
and effects: Is
there a reference
site/test site
conjunction of
candidate cause
and effect?
Reason for
Rejection
If the effects
preceded a
candidate cause in
time, it cannot be
the primary cause.
If the effect occurs
upstream of the
candidate cause's
source or does not
occur regularly
downstream (e.g., is
distributed spatially
independently of a
plume, sediment
deposition areas,
etc.), it cannot be
the primary cause.
If a candidate cause
occurs at reference
sites and occurs at
equal or greater
levels, it can be
eliminated.
Masking
Considerations
If the candidate cause
is preceded by both
the effect and another
sufficient cause, its
effects may be
masked, and it should
be retained.
If the candidate cause
is downstream of
another sufficient
cause, its effects may
be masked and it
should be retained.
Causal
Consideration 1
(See Section
4.1.3)
Temporality
Co-occurrence
Co-occurrence
Chapter 4: Characterizing Causes
4-5
-------
Stressor Identification Guidance Document
Table 4-1 (continued). Application of common types of evidence in eliminating
Alternatives.
Type of Evidence
(See Chapter 3)
Associations
between
measurements of
candidate causes
and effects: Is a
decrease in the
magnitude or
proportion of an
effect seen along
a decreasing
gradient of the
stressor?
Measurements
associated with
the causal
mechanism: Has
the stressor co-
occurred with,
contacted, or
entered the
receptor(s)
showing the
effect?
Association of
effects with
mitigation or
manipulation of
causes:
Did effects
continue when a
source or stressor
was removed?
Reason for
Rejection
A constant or
increasing level of
effect with
significantly
decreasing
exposure would
eliminate a cause.
If the candidate
cause never
contacted or co-
occurred with the
receptor organisms,
the cause may be
eliminated. For
appropriate
stressors, if tissue
burdens or other
measures of
exposure are found
not to occur in
affected organisms,
the cause may be
eliminated. For
stressors that act
through a known
chain of events, if a
link in the chain can
be shown to be
missing, the
candidate cause can
be eliminated.
If the effect
continues even after
the stressor is
removed, then the
candidate cause can
be eliminated. This
assumes that there
is no impediment to
recolonization.
Masking
Considerations
If a decreasing
gradient of one
sufficient cause
coincides with an
increasing gradient of
second, recovery from
the first cause may be
obscured.
The effect may also
continue if another
sufficient cause is
present.
Causal
Consideration 1
(See Section
4.1.3)
Biological
Gradient
Complete
Exposure
Pathway
Experiment,
Temporality
Many of the same types of evidence can also be used in the strength of evidence analysis (see
Section 4.1.3). This column denotes the corresponding causal consideration used there.
4-6
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
In some cases all causes but one will be eliminated, and the part of the process is to
describe the level of confidence in the characterization. It is often desirable to perform a
strength of evidence analysis of that cause to demonstrate that it is probable, given all
available evidence. If the true cause was not identified as a candidate, it may be possible
to eliminate all candidate causes. In that case, one must repeat the process of identifying
candidate causes (Chapter 2). In most cases, the elimination of causes will simply
narrow the set of candidates, which is always helpful. Then the process continues to the
next step, which is the use of diagnostic protocols or keys.
4.2.2 Diagnostic Protocols or Keys
CHARACTERIZE CAUSES
Eliminate | Diagnose |
Strength of Evi
Identify Probable Cause
dence
If more than one cause remains after the
elimination step, the next step is to consider
whether any of the causes are subject to a
diagnostic analysis. Whereas the elimination
step relies on negative evidence (e.g., an
exposure pathway is not present), diagnostic
protocols rely on positive evidence (e.g., a
particular symptom is present). Diagnostic
symptoms are also used in the strength of
evidence analysis (under consistency of
association and specificity; see Section 4.1.3). The diagnostic protocols referred to here
have been used and tested sufficiently to be considered authoritative and some have been
formalized into a set of rules or a key (e.g., Meyer and Barclay 1990).
In medicine, diagnostic protocols identify a disease by examining its signs and
symptoms. The diagnostic process requires an understanding of mechanism, so most of
the evidence comes from measurements associated with the causal mechanism (see
Section 3.4 and worksheet in Appendix B, Unit III, page B-
19). As in medical practice, diagnostic information in the SI
process comes from the exposed organisms and includes
symptomatology (i.e., signs of the action of the causal agent
on the organisms), measures of internal exposure (e.g.,
isolation of pathogens or analysis of chemicals in
organisms), or measurements of intermediate processes (e.g.,
a depressed pre-dawn dissolved oxygen level).
The diagnostic approach
is a good alternative for
SI when organisms are
available for
examination, when the
The diagnostic approach is a good alternative for SI when
organisms are available for examination, when the candidate
causes are familiar enough that they have made it into the
protocols, and when there is a high degree of specificity in
the cause, the effect, or both. As an example, protocols for
the investigation offish kills are particularly well established
(e.g., Meyer and Barclay 1990) and consist of collection of
site data concerning candidate causes (e.g., oxygen, pH,
temperature, contaminant levels, and presence of toxic
algae), site data concerning effects (e.g., taxa killed, duration
of event, behavior of live fish), and necropsy results (e.g.,
lesions, pathogens, tissue contamination, or clinical signs
such as blue stomach which indicates molybdenum toxicity).
Meyer and Barclay (1990) even provide a dichotomous key for determining the causes of
fish kills. Since an SI investigation is more likely to examine current biological
candidate causes are
familiar enough that they
have made it into the
protocols, and when
there is a high degree of
specificity in the cause,
the effect, or both.
Chapter 4: Characterizing Causes
4-7
-------
Stressor Identification Guidance Document
community compositions that might reflect past chronic exposures rather than the effects
of acute lethality, the methods for fish kill investigations often are not directly
applicable. However, a diagnostic approach can potentially be employed.
Diagnostic tools are well developed for pathogens and to a slightly lesser extent for
chemicals (e.g., certain bill deformities are diagnostic of exposure to dioxin-like
compounds) (Gilbertson et al. 1991). Diagnostics are also well developed for a few
other agents such as low dissolved oxygen (low blood oxygen, gasping at the surface,
etc.). For many other stressors and for most non-vertebrate aquatic organisms, reliable
diagnostics are seldom available. Expert judgment has been used to assign tolerance
values to taxonomic groups for nutrients and this concept has been extended to other
stressor types (Hilsenhoff 1987, Huggins and Moffet 1988). The utility of using these
tolerance values in multimetric indices along with some recent statistical analyses
indicate that the structure offish and invertebrate communities may prove valuable for
diagnosis (Yoder and Rankin 1995b, Norton et al. 2000). Although the use of
multimetric information for diagnosing cause and effect is not yet widely accepted or
validated, this information can be brought into the strength of evidence analysis
discussed in the next section.
4.2.3 Strength of Evidence Analysis
In many SI cases, the candidate causes are not identified by elimination or diagnosis, and
an analysis of the strength of evidence for each of the candidate causes is required (see
worksheet in Appendix B, Unit III, page B-20). This analysis organizes information so
that the evidence that supports, or doesn't
support, each candidate cause can be easily
compared and communicated. When there are
many candidate causes or when evidence is
ambiguous, strength of evidence analysis is
more useful than elimination of alternatives
because it identifies the alternative that is best
supported by the evidence. Even when a cause
has been identified by a process of elimination
or diagnosis, it is often desirable to complete
the strength of evidence analysis in order to
CHARACTERIZE CAUSES
Eliminal
e Diagnose
| Strength of Ev
Identify Probable Cause
dence
organize all of the evidence for the decision makers and stakeholders.
The strength of evidence analysis discussed in the remainder of this section defines a
group of causal considerations used to organize the information concerning each
alternative. Causal considerations are logical categories of evidence that are consistently
applied to support or refute a hypothesized cause. They are defined in Section 4.2.3.1.
Section 4.2.3.2 discusses how the types of evidence described in Chapter 3 provide
information relevant to each consideration. Finally, Section 4.2.3.3 shows how to
evaluate the strength of each piece of evidence in supporting or refuting a candidate
cause.
For the purposes of this approach, we treat Koch's postulates (see text box entitled
"Koch's Postulates") as a special case of analysis of the strength of evidence. That is,
for pathogens or chemical contaminants, if Koch's postulates are satisfied, the strength
of evidence is particularly high.
4-8
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
4.2.3.1 Causal Considerations for Strength of Evidence Analysis
This section describes various causal considerations used for strength of evidence
analyses. These considerations draw on the work of epidemiologists and ecologists over
the last 30 years (Fox 1991, Hill 1965, Susser 1986a).
Koch's Postulates
Koch's postulates combine different lines of evidence in a formal way to provide compelling
evidence for causation. The approach was originally developed for pathogen-induced
diseases. It has been adapted for demonstrating that particular toxicants cause human
diseases (Yerushalmy and Palmer 1959, Hackney and Kinn 1979) or ecological effects
(Adams 1963, Woodman and Cowling 1987, Suter 1990, Suter 1993), and has been
recommended for ecological risk assessment (EPA 1998a). The following is an adaptation of
Koch's postulates for causal inference in ecological epidemiology for effects of pathogens or
chemicals.
1. The injury, dysfunction, or other potential effect of the pathogen or toxicant must be
regularly associated with exposure to the pathogen or toxicant in association with any
contributory causal factors.
2. The pathogen, toxicant, or a specific indicator of exposure must be found in the
affected organisms.
3. The effects must be seen when healthy organisms are exposed to the pathogen or
toxicant under controlled conditions, and any contributory factors should contribute in
the same way during the controlled exposures.
4. The pathogen, toxicant, or a specific indicator of exposure must be found in the
experimentally affected organisms.
The power of Koch's postulates arises from the way the four types of evidence are combined.
The requirement of regular association in the field ensures that the association is relevant to
the field, but, because field observations are uncontrolled, one cannot determine whether the
association is, in fact, caused by an another agent that happens to be correlated with the
proposed cause. In addition, associations in field data fail to demonstrate the temporal
sequence between the candidate cause and effect. The requirement that the candidate causal
agent induce the effect under controlled conditions eliminated the potential for confounding
and demonstrates that the cause precedes the effect. However, the artificial conditions of
toxicity tests and other experimental studies means that the demonstrated causal association
may not be relevant to the field. The second and fourth postulates provide the ties that bind
the two lines of evidence together. That is, evidence of exposure must be obtained in the field
and must correspond to the experimental exposure. This correspondence of the exposure
metrics makes it highly unlikely that the correspondence of effects in the field and the
experiment are coincidental.
Koch's four postulates were derived for addressing the general issue of whether a stressor
could be a cause at all (i.e., could DDT cause reproductive failure in birds). SI investigations
typically choose among causal scenarios that have already been established as having the
ability to produce impairment. For this reason, the emphasis is placed on postulate 2,
identifying the pathogen, toxicant, or specific indicator of exposure in the affected organisms.
This case-specific information is then combined with previously established information
discussed in postulates 1, 3, and 4. This approach works best for simple causal agents that
have a known indicator of exposure. When causal scenarios have multiple insufficient causes,
the requirements of regular association and experimental evidence can rarely be met for the
specific mixture that is encountered in the field situation. In cases where multiple sufficient
causes can be assumed to be acting independently, the evidence for each cause can be
evaluated separately.
Chapter 4: Characterizing Causes
4-9
-------
Stressor Identification Guidance Document
The first four considerations, co-occurrence, temporality, biological gradient, and
complete exposure pathway draw primarily on associations that are derived from the case
itself. These considerations form the strongest basis for causal inference. The next two
considerations, consistency of association and experiment can be based either on data
from the case at hand or may draw from similar situations. The next four considerations,
plausibility, specificity, analogy, and predictive performance, combine information from
the case at hand with experiences from other cases or test situations, or from knowledge
of biological, physical, and chemical mechanisms. These considerations provide
corroborative information that can be used to supplement the basic observations of
association of observed effects and potential causes from the case. The last two
considerations, consistency and coherency of evidence, evaluate the relationships among
all of the available lines of evidence.
Each of these causal considerations is discussed below:
Co-occurrence - The spatial co-location of the candidate cause and effect. In SI,
this consideration is case-specific; for example, effects may be occurring downstream
but not upstream of an identified source (see text box entitled "Arkansas River Case
Study"). This consideration should be interpreted with caution when several sufficient
causes may be present and when the objective of the analysis is to identify all potential
and contributing causes. In this situation, the causes occurring the furthest upstream may
mask the effects of causes occurring later in the downstream sequence.
Temporality - A cause must always precede its effects. For example, a baseline
monitoring study showing a productive trout population before a dam was built provides
some evidence that the dam caused the subsequent population decline. As with co-
occurrence, this criterion should be applied with caution when several sufficient causes
may be present and when the objective of the analysis is to identify all potential and
contributing causes. In this situation, the causes occurring early in the time sequence
may mask the effects of causes occurring later.
Biological Gradient - The effect should increase with increasing exposure. This
is the classic toxicological requirements that effects must be shown to increase with
dose. Biological gradient is also applicable to other types of causes (see text box entitled
"Arkansas River Case Study"). For example, if fine substrate texture is believed to cause
reduced diversity of benthic invertebrates, then diversity should decline along a gradient
of texture. In SI, evidence for biological gradient is case-specific. Examples include
demonstrating recovery of a community downstream of an outfall, or evidence that an
effect decreases with decreasing concentration of an effluent or with increasing mean
flow. Investigators should be aware that some stressors elicit non-linear response. For
example, community diversity can increase at low levels of nutrient enrichment, then
decline again as enrichment increases. Regression and correlation analyses are common
tools used to quantify biological gradient; both high slopes and large correlation
coefficients increase the strength of evidence.
Complete Exposure Pathway - The physical course a stressor takes from the
source to the receptors (e.g., organisms or community) of interest. If the exposure
pathway is incomplete, the stressor does not reach the receptor, and cannot cause an
effect. Evidence for a complete exposure pathway is case-specific and may include
measurements such as body burdens of chemicals, presence of parasites or pathogens, or
biomarkers of exposure (see text box entitled "Arkansas River Case Study"). For
stressors that do not leave internal evidence (e.g., siltation), measurements that show the
4-10 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
stressor co-occurring in space and time with the receptor may be useful. For causes that
induce effects indirectly, observations or measurements of the intermediate products or
conditions are evidence of a complete exposure pathway (see Chapter 7, Little Scioto
case study).
Consistency of Association - Refers to the repeated observation of the effect and
candidate cause in different places or times (see text box entitled "Lake Washington
Case Study"). A consistent association of an effect with a candidate cause is likely to
indicate true causation. The case for causation is stronger if the number of instances of
consistency is greater, if the systems in which consistency is observed are diverse, and if
the methods of measurement are diverse. Consistency can be demonstrated using
evidence from the case at hand, or may draw on evidence from many cases. For
example, if fish kills repeatedly occur below a particular outfall, there is a consistent
association over time of those incidents with a candidate cause. Less commonly, a
particular case may have multiple instances of exposure to an agent spread over space
rather than time. Consistent association can also be demonstrated across multiple sites
or cases. For example, a decrease in benthic arthropod diversity may be consistently
observed at many different sites having low dissolved oxygen levels. Consistency of
association across many sites is seldom demonstrated because the same causal agent
seldom occurs at multiple sites that are sufficiently similar to demonstrate a consistent
response. However, when it is demonstrated, consistency across sites is stronger
evidence for causation than the simple co-occurrence or temporal association of the
agent with the response in a single case.
Arkansas River Case Study:
Using Strength of Evidence Analysis
This example highlights strength of evidence evaluations used in the SI process. Specifically, the
example presents several lines of evidence used to support the hypothesis that heavy metal exposure
impairs benthic macroinvertebrate communities.
Several sites in the Arkansas River (CO) were monitored over a 10-year span to examine the effects
of cadmium (Cd), zinc (Zn), and copper (Cu) on benthic macroinvertebrates. More specifically, metal
contamination was related to the abundance of heptageniid mayflies. It was found that heptageniid
mayflies were abundant upstream of known metal inputs, and sparse downstream of these inputs, an
example of spatial co-occurrence. In addition, a complete exposure pathway was evident:
concentrations of Cd, Cu and Zn were elevated in benthic invertebrates collected at stations
downstream of the source. Evidence of a biological gradient was observed using multiple regression
analysis; the abundance of heptageniid mayflies decreased with increasing zinc concentrations.
Evidence from other studies was also available and demonstrated that effects from metals would be
plausible based on stressor-response relationships observed in the laboratory. Chronic toxicity
tests of water collected from the Arkansas using Ceriodaphnia dubia and microcosm tests using
mayflies established that effects would be expected at the concentrations of Zn, Cu, and Cd measured
in the Arkansas.
Evidence from other studies also supported the hypothesis that heavy metal exposures reduce
abundance of mayflies. Regional Environmental Monitoring and Assessment Program (R-EMAP) data
from other locations in the Rocky Mountains showed a consistent association between metal
exposures and reduced abundance of heptageniid mayflies.
Finally, efforts were undertaken by several agencies to reduce ambient metal concentrations, an
example of a remedial experiment. Increases in the abundance of heptageniid mayflies were
observed at the sites with greatest metal reduction. Further, little biological improvement was
observed where metal levels have remained elevated.
References: Clements and Kiffney 1994, Kiffney and Clements 1994a, Kiffney and Clements 1994b,
Clements 1994, Clements et al. 2000, Nelson and Roline 1996.
Chapter 4: Characterizing Causes
4-11
-------
Stressor Identification Guidance Document
Experiment - Refers to the manipulation of a cause by eliminating a source or
altering exposure (Hill 1965) (see text boxes entitled "Lake Washington Case Study" and
"Arkansas River Case Study"). Experiments of greatest relevance to SI (see Section 3.3)
include manipulating and testing site media in the laboratory (e.g., using TIE), and
conducting field experiments by controlling a source (e.g., fencing cattle) (USEPA
1991b, 1993a, 1993b). The strongest evidence is case-specific. If evidence from
experiments conducted on a similar situation is used, the relevance to the case at hand
should be described.
Plausibility - Refers to the degree to which a cause and effect relationship would
be expected given known facts. Two types of plausibility are discussed below:
Mechanism: Given what is known about the biology, physics, and chemistry of the
candidate cause, the receiving environment, and the affected organisms, is it
plausible that the effect resulted from the cause? It is important to distinguish a
lack of information concerning a mechanism (e.g., the ability of chemical x to
induce tumors is unknown) from evidence that a mechanism is implausible (e.g.,
chemical x is not tumorogenic). It is also important to carefully consider whether
some indirect mechanism may be responsible. For example, increased nutrient
levels cause algal blooms that decompose and reduce epibenthic oxygen
concentrations, which in turn decrease invertebrate diversity. If a mechanism is
known and there is evidence that the mechanism is operating in a specific case, the
positive evidence is particularly strong.
Stressor-Response: Given a known relationship between the candidate cause and
the effect, would effects be expected at the level of stressor seen in the
environment? The comparison of environmental concentrations to laboratory-
derived concentration-response relationships is a common approach used in
chemical risk assessments. It provides strong evidence of causality if
concentrations are higher than a level that causes a relevant effect (see Table 3-2)
(see text box entitled "Arkansas River Case Study"). Note that exceedence of
water quality criteria or standards does not necessarily imply causation because
regulatory values are intended to be set at safe levels. Whole effluent toxicity tests
may be used with dilution models. Although used mostly for chemical stressors, a
similar approach could also be used for other types of stressors, such as siltation.
Analogy - Examines whether the hypothesized relationship between cause and
effect is similar to any well-established cases. Hill (1965) used the criterion of analogy
to refer specifically to similar causes. For example, a new pesticide with a similar
structure to another one may induce similar effects. The idea can be extended to other
types of stressors. For example, an introduced species that has similar natural history
characteristics to one that had been previously introduced may have similar impacts on
the ecological system.
4-12 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Lake Washington Case Study1
Lake Washington, located in Seattle and draining into Puget Sound, first began receiving street
runoff and raw sewage input from Seattle at the turn of the 20th century. Although the sewer
outlets were eventually replaced by wastewater treatment plant effluents, the growing human
population in the surrounding area put increasing demands on the lake. By 1953, 10
wastewater treatment plants discharged into Lake Washington. Shortly thereafter, the first
report describing nutrient loadings in the lake was issued by researchers at the University of
Washington.
While the problems associated with eutrophication were not widely recognized by the public at
the time, a University of Washington professor, W.T. Edmondson, used the concept of
consistency of association^ make an important observation: the recent discovery of a blue-
green alga (Oscillatoria rubescens) in Lake Washington coincided with other documented
cases where water quality had declined in response to nutrient input. The lakes described in
these reports ranged geographically from Wisconsin to western Europe, yet the highly specific
occurrence of Oscillatoria was identified in each case as an early response to water
enrichment. Thus, Edmondson asserted that the water quality in Lake Washington was
declining in response to nutrient input, and would continue to decline in predictable ways.
Edmondson developed a model based on principles of mass balance and stoichiometry to
define the quantitative relationships between nutrient levels and algal biomass. He used the
model to forecast that water quality in Lake Washington would continue to decline in
predictable ways. This is an example of predictive performance, since continued monitoring
confirmed his assertions.
Outcome
Edmondson's letters and popular science articles describing the problems of the lake
successfully brought about public and political support for the eventual clean-up of Lake
Washington. Between 1963 and 1968, all 10 wastewater treatment plant discharges were
diverted out of Lake Washington and sent to a common collection system that ultimately
discharged deep within Puget Sound.
Until the diversions were constructed, water quality had continued to decline as predicted by
Edmondson, with water transparency at less than 1 m in 1962. However, in the years following
the improvements, nutrient levels decreased substantially. By the 1970s, visibility had reached
12 m, and the presence of the blue-green alga O. rubescens was undetectable. The swift
recovery of Lake Washington following the removal of nutrient inputs in this field experiment
left little uncertainty about the true cause of its water quality decline.
1 Summarized from J. T. Lehman (1986).
Specificity of Cause - Applicable only if the proposed cause is plausible or if it
has been consistently associated with the effect. Specific cause-effect relationships are
more likely to be demonstrated to be causal (see text box "Lake Washington Case
Study"). If an effect (e.g., hepatic tumors in fish) observed at the site has only one or a
few known causes (e.g., PAHs), then the occurrence of one of those causes in association
with the effect is strong evidence of causation. In the extreme, causation is clear when
both effects and causes are specific (x causes specific effect y, and y is caused only by x).
One implication of this consideration is that both effects and causes should be defined as
specifically as possible in order to increase the specificity of the association. For
example, a specific cause such as highly embedded substrate can be more clearly
associated with identified effects than a general cause like overall poor habitat quality.
Predictive Performance - Refers to whether the candidate cause has any initially
unobserved properties that were predicted to occur. Was that prediction confirmed at the
site? The ability to make and confirm predictions is one of the hallmarks of a good
Chapter 4: Characterizing Causes
4-13
-------
Stressor Identification Guidance Document
scientific process. For example, if the proposed cause of a fish kill is drift of an
organophosphate insecticide into a stream, one could make the specific prediction that
cholinesterase levels would be reduced, or the more general prediction that insects and
crustaceans would also be killed. If these predicted conditions are then observed at the
site, it increases confidence in the causal relationship (see text box entitled "Lake
Washington Case Study"). Multiple predictions in both the positive and negative
direction would strengthen this criterion (e.g., plants and protozoa would not be harmed,
but arthropods would be).
Consistency of Evidence - Refers to whether the hypothesized relationship
between cause and effect is consistent with all available evidences. The strength of this
consideration increases with the number of lines of evidence (Yerushalmy and Palmer
1959).
Coherence of Evidence - Examines whether a conceptual or mathematical model
can explain any apparent inconsistencies among the lines of evidence. For example
metal concentrations at the site may be sufficient to impair reproduction in fish, and yet
both juvenile and adult fish occur at the site. This evidence may be coherent if
reproduction is not occurring at the site, but juvenile fish re-colonize the site from
unexposed locations. Another explanation may be that the measured total metal
concentration is not 100% bioavailable. The strength of these explanations depend on
the expertise and judgment of the assessors. It is a weak line of evidence, because of the
possibility that post hoc explanations are wrong. However, the hypotheses may lead to
experiments or predictions in future iterations of the causal assessment (e.g., testing the
bioavailability of the metals), which could support stronger inferences.
4.2.3.2 Matching Evidence with Causal Considerations
Table 4-2 illustrates the different types of evidence discussed in Chapter 3 with the
causal considerations they support. The relationship between types of evidence and
causal considerations is not one-to-one. Each type of evidence may be relevant to
several causal considerations, and a causal consideration may be evaluated using several
different types of evidence. In any specific application of SI, evidence will probably
exist for only some of the causal considerations, and the evidence will be uneven across
the candidate causes. After the evidence relevant to each consideration is identified, it is
evaluated as discussed in the next section.
4.2.3.3 Weighing Causal Considerations
Epidemiologists and ecoepidemiologists have attempted to develop guidance for
weighing the causal considerations described below (Fox 1991, Hill 1965, Susser
1986a). Table 4-3 presents the possible outcomes for each consideration and provides
symbols to represent the influence of each outcome on the inference.
Table 4-3 illustrates a format that can be applied to specific SI cases. In this table, the
causal considerations are listed in the left-hand column. Each of the other columns
presents results for a candidate cause. The rows show the appropriate number of+, -, or
0 symbols associated with the strength of evidence for each consideration evaluated for
each candidate cause. Supporting narratives should describe how the scores were
obtained from the evidence. We do not recommend adding up the scores for each
candidate cause. Adding the scores erroneously implies that each consideration is of
equal importance and is equitable only if the same types of evidence are available across
4-14 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
all candidates. In difficult cases, it may be valuable to compare the evidence for each
individual consideration across the candidate causes. Particular attention should be paid
to negative results, which are more likely to be decisive.
Table 4-2. Types of evidence (columns) that contribute to each causal consideration (rows).
Causal
Considerations
Co-occurrence
Temporality
Biological
Gradient
Consistency of
Association
Complete
Exposure
Pathway
Specificity of
Cause
Plausibility:
Mechanism
Plausibility:
Stressor-
Response
Experiment
Analogy
Predictive
Performance
Types of Evidence
Associations of
Measurements
of Cause and
Effect
o
ro
0
o
o
O
ro
1o
Q.
CO
X
X
X
c
T3
CD
ro
1o
Q.
CO
X
X
o
(D
O
O
O
i
O
O
ro
o
Q.
E
H
X
X
X
c
0
O
-------
Stressor Identification Guidance Document
Table 4-3. Format for a table used to summarize results of an inference concerning
causation in case-specific ecoepidemiology. (Table adapted from Susser (1986a), Fox
(1991), Suter (1998), Beyers (1998).)
Consideration
Results
Score1
Case-Specific Considerations
Co-occurrence
Temporality
Consistency of
Association
Biological Gradient
Complete Exposure
Pathway
Experiment
Compatible, Uncertain, Incompatible
Compatible, Uncertain, Incompatible
Invariant, In many places and times,
At background frequencies or many
exceptions to the association
Strong and monotonic, Weak or other
than monotonic, None, Clear
association but wrong sign
Evidence for all steps, Incomplete
evidence, Ambiguous, Some steps
missing or implausible
Experimental studies: Concordant,
Ambiguous, Inconcordant
+,0,
+,0,
++, +, -
+++, +, -,
++, +, o,
+++, 0,
Considerations Based on Other Situations or Biological Knowledge
Plausibility
Mechanism
Stressor-Response2
Consistency of
Association
Specificity of cause3
Analogy
Positive
Negative
Experiment
Predictive Performance
Actual Evidence, Plausible, Not
known, Implausible
Quantitatively consistent, Concordant,
Ambiguous, Inconcordant
Invariant, In most places, In some
places, At background frequency or
many exceptions to the association
Only possible cause, One of a few,
One of many
Analogous cases: Many or few but
clear, Few or unclear
Experimental studies: Concordant,
Ambiguous, Inconcordant
Prediction: Confirmed specific or
multiple, Confirmed general,
Ambiguous, Failed
++, +, o,
+++, +, 0, -
+++, ++, +1
+++, ++, 0
++, +
+++, 0,
+++, ++, 0,
Considerations Based on Multiple Lines of Evidence
Consistency of Evidence
Coherence of Evidence
Evidence: All consistent, Most
consistent, Multiple inconsistencies
Evidence: Inconsistency explained by
a credible mechanism, No known
explanation
+++, +,
+,o
In addition to the scores noted, there ay be No Evidence (NE) available relevant to the
consideration, or the consideration may be Not Applicable (NA) for the particular case (see
especially stressor-response and specificity).
Stressor-response is not applicable (NA) if the mechanism is clearly implausible.
Specificity of cause is not applicable (NA) if either the mechanism is clearly implausible, or if
there are many exceptions to the association.
4-16
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Other methods for combining different lines of evidence include expert systems based on
the logic of abduction and Bayesian statistical approaches (Josephson and Josephson
1996, Clemens 1986). As of this writing, these more quantitative approaches have not
yet been developed for combining evidence for SI.
4.3 Identify Probable Cause and Evaluate Confidence
CHARACTERIZE CAUSES
Diagnose Strength of Evidence
| Identify Probable Cause |
Whichever method is used to infer causation,
the results of the characterization must be
summarized. That is, the cause must be
described, the logical basis for its Eliminate
determination summarized, and the
uncertainties concerning that determination
presented. As discussed above, there may be
multiple sufficient causes, all of which should
be characterized. In extreme cases, the
effects of the primary causes are so severe
that other potential causes will remain unidentified.
The level of confidence in causal identification may be assessed in quantitative or
qualitative terms. Confidence is determined in part by uncertainty concerning the data,
the models, and the observations that contribute to the inference. The uncertainty
associated with the data may be partially estimated by conventional statistical analysis
(see text box in Chapter 3 entitled "Data Quality Objectives," and "Using Statistics and
Statistical Hypothesis Testing"), but also includes uncertainty concerning the
applicability of the data. If data must be extrapolated between species or life stages, if
old data are used to estimate current conditions, or if, for some other reason, data are not
directly applicable, the associated uncertainty should be estimated. The uncertainty in
statistical models, such as regressions of biological properties against levels of potential
causes, may be estimated using goodness-of-fit statistics or confidence bounds. The
uncertainty due to the parameters in mathematical models, such as models of dissolved
oxygen depression due to nutrient input, may be estimated analytically or by Monte
Carlo simulation (USEPA 1996a, 1999). If a causal inference is logically clear and is
based predominantly on the results of a statistical or mathematical model, the
uncertainties concerning the results may serve to estimate the uncertainties concerning
the inference.
In most cases, unquantified uncertainties will dominate. These include lack of data
concerning the presence or levels of particular stressors, incomplete biological data,
uncertainty concerning the time when the impairment began, and many more. In
addition, most causal inferences are based on the strength of evidence, so that no single
source of uncertainty characterizes the uncertainty concerning the conclusion.
Therefore, the uncertainty concerning most identifications of causes must be
characterized qualitatively. That qualitative judgement should be accompanied by a list
of major sources of uncertainty and their possible influence on the results.
In some cases, investigators will be able to clearly demonstrate that a particular cause is
responsible for the ecological injuries of concern. However, in many if not most cases,
there will be significant uncertainty concerning the relative contributions of alternative
causal factors. In such cases, it is necessary to determine whether the evidence is
sufficient to justify a management action. Standards and criteria for establishing
epidemiological causation are not generally agreed upon. In particular, there is no
Chapter 4: Characterizing Causes 4-17
-------
Stressor Identification Guidance Document
consistent standard for adequacy of proof. While conventional science sets a high
standard to prove causation, the precautionary principle begins by assuming that an agent
is harmful and requires disproof of causation (Botti et al. 1996). Such decisions are
made by risk managers, rather than risk assessors, and may be based on considerations
such as the cost of remediation and the nature and magnitude of the ecological injury.
Ideally, that judgment would be made on the basis of a priori criteria. That is, each
program that uses SI should specify a standard basis for deciding whether the
characterization of the cause is sufficient for the management purpose. For example, for
the permitting of POTW effluents, a particular state might develop standards for proof
that those effluents cause particular types of injuries. However, standards and criteria
for establishing causation are not generally agreed upon, and many decisions are made
ad hoc. That is, the evidence concerning causation may be presented to the risk manager
as a best estimate of causation along with an accompanying analysis of uncertainties.
The risk manager may use that result to help reach a decision.
As discussed in Chapter 1, the SI process may be conducted iteratively until sufficient
confidence in the causal characterization is reached. In the most uncertain and complex
cases, the SI process may best serve to guide further data collection, modeling, or
analysis efforts. Options for iterating the process are discussed further in Chapter 5,
below. If the cause is confidently identified, then the investigation may proceed to
identifying sources, developing and implementing management options, and monitoring
their effectiveness (Figure 1-1).
4-18 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Chapter 5
Iteration Options
In this Chapter:
5.1 Reconsider the Impairment
5.2 Collect More Information
This chapter describes iterations if no clear cause can be identified. If the SI process has
yielded no clear cause for the biological impairment, it may be because (1) there is
actually no effect (Section 5.1), or (2) there is insufficient information concerning the
identified causes or the true cause was not among the list of candidate causes (Section
5.2). These alternatives, all leading to a reiteration of the investigation (Figure 4-1), are
discussed in this section.
5.1 Reconsider the Impairment
When no cause was identified, it may be that there is actually no effect, or the actual
effect may be different from the identified impairment (see worksheet in Appendix B,
Unit V, page B-35). This situation is known as a false positive, or in statistical terms, a
Type I error. It should be noted that both false positive and false negative errors (failure
to detect an effect that exists) are inherent to any detection system, whether it is medical
diagnostics, aircraft radar, or environmental monitoring.
A false positive might result from errors in a biological survey or in the analysis of data.
The samples may have been collected improperly; therefore, the biotic community
appears to be less abundant or species rich than it truly is. The individuals performing
the identifications may have misidentified organisms. There may have been errors in
data recording or analysis. Any of these errors may artificially obscure the responses.
A quality assurance program can minimize, but not entirely eliminate these errors. If the
causal analysis reveals weaknesses in the evidence for the occurrence of a real effect, a
careful audit of the biological survey may be appropriate.
Other reasons for a false positive result include sampling error and the natural variability
of the biological indicators. In any monitoring program, sampling is stratified among
perceived natural classes and subdivisions of systems (e.g., habitat type, salinity,
sediment, elevation, biogeographic region), and often by season (sampling index period
in defined season). A sample may have been taken outside of an index period. A site
may belong to a poorly characterized system type or may have been incorrectly classified
(e.g., cold water system evaluated using warm water criteria). Any unrecognized
misclassification can result in either a false positive or false negative. Intensive
monitoring and characterization of natural systems, combined with quality assurance and
peer review of results, can reduce both types of errors.
In other cases, the impairment may have been defined too broadly or investigators may
have made wrong assumptions about mechanisms when developing their conceptual
model. For example, the first investigations into bird population declines and DDT
focused on mortality rather than egg-shell thinning, and failed to find a connection with
DDT (see text box entitled "Revisiting the Impairment in the Case of DDT"). Careful
reconsideration of the nature of the impairment can put the investigation back on the
right track.
Finally, natural variability of the indicators, not due to any measurement or analytical
errors, can result in both false positives and false negatives. Environmental criteria may
Chapter 5: Iteration Options 5-1
-------
Stressor Identification Guidance Document
be defined by exceedence of a percentile or extreme value of some statistical
distribution. This means that natural, or unimpaired conditions, may also exceed the
criteria at some frequency. Ideally, acceptable error rates should be specified for
decisions resulting from the biological assessment system. If confidence in a finding of
biological impairment is low (that is, if the indicator just exceeds the threshold value),
then increased sampling may reduce uncertainty and increase confidence (see next
section).
5.2 Collect More Information on Previous and Additional Scenarios
If a causal scenario has not been established with sufficient confidence and the effect
appears to be real, management should be consulted to discover if knowing the cause is
still required for decision-making. If so, then more information must be collected (see
worksheet in Appendix B, Unit VI, page B-36). Because the cost of field data collection
and data analysis increases with each iteration, it is important to carefully plan what
additional information is needed to determine the cause of impairment. This information
may include previously considered scenarios for which information was inadequate, or
candidate causal scenarios that were not previously considered.
Revisiting the Impairment in the Case of DDT1
The fact that DDT played a role in the decline of bald eagle and other bird-of-prey populations
(e.g., osprey, brown pelicans) is now commonly appreciated among most biologists. However,
the link between DDT and the eggshell thinning that caused reproductive failure in these birds
was not initially recognized. Ultimately, the connection was made by re-examining the
description of the impairment.
The first link between DDT and diminishing bald eagle and other bird-of-prey populations was
the consistent observation of high body burdens of DDT metabolites. In other words, there was
co-occurrence of the declining bird populations and the candidate cause, DDT. There was
also evidence of a complete exposure pathway to birds based on body burden of DDT.
However, extensive toxicity testing of DDT on adult bird mortality revealed no relationship. This
suggested that the proposed mechanism, toxicity, was implausible. However, lethality was not
the impairment; decline of birds-of prey was the impairment. A new conceptual model was
required that considered other mechanisms that could result in declines in bird populations. In
re-examination of the overall analysis, it became apparent that the species chosen for testing
had been relatively tolerant of DDT exposure compared to those that were affected in the wild,
and that the endpoint observed in these tests (lethality) would not reflect reproductive success
or failure resulting from DDT exposure.
Field observations eventually revealed a potential plausible mechanism of reproductive failure
due to eggshell thinning among bald eagles and other birds-of-prey. Laboratory experiments
showed that DDE could cause eggshell thinning. Field studies showed that field exposures to
DDE, a metabolite of DDT, were sufficient to cause effects in many species of birds based on
the stressor-response relationship. Together these findings provided lines of evidence by
which DDT might cause eggshell thinning and reduce reproductive success, a more specific
impairment than declines in bird population.
Outcome
In 1972, DDT was banned from most uses in the United States. In the years following the ban,
bald eagle and other bird-of-prey populations slowly recovered. The recovery of bird
populations after banning the use of DDT, is an example of mitigation of the effect following
manipulation of the cause, and is very strong evidence that the use of DDT was, in fact, the true
cause of bald eagle and other bird-of-prey population declines.
References
Grier, J.W. 1982; Blus, L.J., and C.J. Henny, 1997
5-2
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Even when the characterization of causes has not determined the cause with sufficient
confidence, the set of candidate causes should have been reduced, and the critical
evidence should be apparent. In particular, it should be possible to design experiments
or observations that will potentially eliminate certain causes (Chapter 4.1.1). However,
such experiments are not always feasible. Alternatively, one may identify critical pieces
of positive evidence that would strongly support one scenario and none of the others. In
most cases, it will be appropriate and prudent to plan a sampling and testing program that
will generate a set of potentially decisive positive and negative evidence.
If all of the most common causes have been eliminated or have been determined to be
unlikely, then additional causal scenarios need to be identified. The process is similar to
that described in Chapter 2. New data may have come to light during the first iteration
of the SI process. These data should be carefully reviewed to determine if there are any
clues to suggest additional causal scenarios. Details of the available data should be
considered, such as weather patterns, new construction, or land use information. If the
descriptions of the effect or the scope were too broad, they may need to be refined or
more clearly defined. Additional potential causal scenarios may include new stressors or
combinations of stressors that occur simultaneously or in a specific sequence. After the
additional candidate causal scenarios are developed, key evidence should be identified
that is likely to allow identification of the cause.
The most important tools to bring to the SI process are experience in multiple disciplines
(especially ecology), careful, deliberate critical thinking, and a strong desire to find the
true cause of biological impairment.
Chapter 5: Iteration Options 5-3
-------
Stressor Identification Guidance Document
Chapter 6
Presumpscot River, Maine
6.1 Executive Summary
The Presumpscot River is located in southern Maine and forms the outlet of one of
Maine's largest lakes, Sebago Lake. From 1984 to 1996, biological monitoring
downstream of a pulp and paper mill discharge consistently revealed non-attainment of
Maine's Class C aquatic life standards. The river is impounded above and below the
discharge. The discharge releases high concentrations of TSS and total phosphorus, and
on occasion releases metals above the chronic criteria but below acute criteria. Upstream
samples consistently indicated attainment of Class C or better standards.
Description of the Impairment
Biological impairment was characterized by a shift in the benthic macroinvertebrate
community from 90% insects upstream of a pulp and paper mill discharge to about 50%
insects downstream. This shift included a 15-35% loss of taxonomic richness, and 40-
60% loss of Ephemeroptera-Plecoptera-Trichoptera (EPT) taxa. Moreover, many insect
taxa found upstream of the discharge were pollution-sensitive, while those found
downstream were primarily pollution-tolerant species, such as snails and worms.
List Candidate Causes
Eight candidate causes for non-attainment were considered in the Stressor Identification
process:
1. Excess toxic chemicals from the discharge;
2. High TSS combined with floe causes high BOD and reduced DO;
3. High TSS combined with floe causes smothering;
4. Excess nutrients (from POTWs, nonpoint sources, and the mill) cause excess
algal growth;
5. Impoundment increases sedimentation that smothers biota;
6. Impoundment decreases flow velocity and causes algal growth, leading to
reduced DO;
7. Impoundment causes low DO; and
8. Impoundment causes loss of suitable habitat.
Characterizing Causes: Eliminate
Four of the eight candidate causes were logically eliminated from examination of the
evidence. Reduced DO sufficient to cause the impairment was not observed in the
Presumpscot River, and bottom-water DO concentrations were stable throughout the
Chapter 6: Presumpscot River, Maine 6-1
-------
Stressor Identification Guidance Document
river, above and below the discharge. Therefore, causal scenarios #2, #6 and #7 could be
eliminated. Although elevated concentrations of total phosphorus (TP) were observed
below the discharge, the increase in chlorophyll a concentration was negligible. Water
column chlorophyll a is a surrogate measure for algal biomass. Because excess algal
growth was necessary for causal pathway #4, and there was none, it was also eliminated
from further consideration.
Characterizing Causes: Diagnose
No evidence strong enough to support diagnosis was available for any of the candidate
causes.
Characterizing Causes: Strength of Evidence
A strength of evidence approach was then used to examine the remaining four candidate
causes. The four remaining causes were toxic chemicals, flocculent TSS causing
smothering, impoundment increasing sedimentation, and impoundment causing loss of
suitable habitat. There was no strong evidence for or against the toxic chemical
hypothesis (#1). Several lines of strong evidence favored the TSS hypothesis (#3):
> The exposure pathway from discharge to biological impairment was complete
and plausible.
> Other rivers with similar elevated flocculent TSS also had impaired biological
assemblage.
> Removal of flocculent TSS from a nearby river resulted in recovery of the
biological assemblage.
Two lines of evidence disfavored the two impoundment hypotheses (#5 and #8):
* Other impoundments with similar potential sediment loadings (not from mill
discharge) and similar habitat support diverse invertebrate assemblages that meet
aquatic life use criteria; and
> a site upstream of the mill effluent, and within the same impoundment, met
aquatic life use criteria.
Characterizing Causes: Identify Probable Cause
The evidence supporting scenario #3, that non-attainment was due to high loads of
flocculent TSS from the discharge, was consistent throughout the lines of evidence.
Strength of association, spatial co-occurrence, and experimental lines of evidence
strongly supported this scenario. Evidence for the toxicity scenario (#1) was extremely
weak. Evidence for the two impoundment scenarios (#5 and #8) was negative. The State
of Maine concluded that high TSS was sufficient for causing the biological impairment.
Quality of the data were adequate, and confidence in the conclusion was high.
Subsequently, the State took management action to reduce loadings of TSS through a
TMDL that was approved by EPA. This was the first time in New England that
bioassessment findings had served as the quantitative response variable for development
of a TMDL and resulting pollutant discharge limits, including the pulp and paper mill.
6-2 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Moreover, it provided a means for Maine to control a pollutant (TSS) for which it had no
specific criterion in its water quality standards.
6.2 Background
This case study is presented as an example of how the stressor identification (SI) process
could have been used to determine the cause of non-attainment of aquatic life use in a
small river in Maine. The case study begins with the presentation of background
information on the regulations in the State of Maine and the geographical location of the
case study. This is followed by a brief discussion of the evidence found at the site and in
other situations. Several causal scenarios are then presented and analyzed separately to
illustrate how the SI process could be used to eliminate four of the eight candidate
causes. A strength of evidence analysis is then used to identify the most likely cause.
The case study concludes with a brief discussion of the management actions taken to
remedy the situation. One of the most significant results of this effort was that the State
of Maine, Department of Environmental Protection, used bioassessment findings to
control a stressor for which the State has no standards.
Impairment Trigger: Biological monitoring in the Presumpscot River in Westbrook,
below a pulp and paper mill discharge, has consistently revealed non-attainment of Class
C aquatic life standards (1984, 1994, 1995, 1996) using standard Maine Department of
Environmental Protection methods (invertebrate) (Davies and Tsomides 1997).
Regulatory Authority
The Maine Department of Environmental Protection (MDEP) issues wastewater
discharge licenses that set the allowable amounts of pollutants that industries may
discharge to waters of the State. These limits are scientifically determined in order to
preserve water quality sufficient to maintain all designated uses and criteria established,
by law, for the river. In recent years USEPA has required that a Total Maximum Daily
Load (TMDL) be established for impaired river systems, such as the Presumpscot, for
which existing, required pollution controls are inadequate to attain applicable water
quality standards.
The State of Maine established minimum standards for three water quality
classifications, Class A, Class B, and Class C. These classes specify designated aquatic
life uses from Class C, the minimum state standard, to the most protected waters with the
Class A/AA designation. Class C requires that the structure and function of the
biological community be maintained and provides for the support of all indigenous fish
species.
Under this system, attainment of the aquatic life classification standards for a given
water body is evaluated using numeric biological criteria. The MDEP numeric aquatic
life criteria are based on statewide data collections over a 14-year period with analysis of
over 400 sampling events. Artificial substrates (rock baskets) are incubated on the
bottom at stream sites, retrieved, and benthic macroinvertebrates that have colonized the
substrates are identified and enumerated (Davies and Tsomides 1997). Aquatic life
classification standards for a given water body are evaluated using numeric biological
criteria that were statistically derived from the statewide database. The criteria are in the
form of a statistical model (linear discriminant model) which yields the probability that a
test sample belongs to one of the 3 water quality classes, or non-attainment of the lowest
Chapter 6: Presumpscot River, Maine 6-3
-------
Stressor Identification Guidance Document
class (Davies et al. 1995). The model uses a set of metrics derived from the species
composition and abundance enumerated from the substrates.
Geography
The Presumpscot River is the outlet of Sebago Lake. The river flows through the most
densely-populated county in the State of Maine, crossing the towns of Gorham,
Windham, Westbrook, Portland, and Falmouth. The Presumpscot then empties into
Casco Bay at the Martins Point Bridge (Figure 6-1).
Pulp Mill Outfall
Westbrook
POTW
295 attainment
Cumberland Dam
238
Presumpscot River
Biomonitoring Stations f\/ Rivers
A Sources H_| Major water areas
Dams I I State
POTW = Publicly Owned
Treatment Works 1 n 1
2 Miles
A
N
Figure 6-1. Map of the Presumpscot River showing biomonitoring stations, potential
sources of impairment, and their location relative to the Androscoggin River (inset).
6-4
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Compared to industrial receiving waters in the State of Maine, the Presumpscot is a
relatively small river, having a drainage area of only 647 square miles. These
circumstances contribute to a low dilution ratio in the lower Presumpscot River.
The river has six impoundments and four industrial and municipal waste discharges.
This study comprises an area immediately downstream of a pulp and paper mill effluent
discharge. Approximately 3.2 km, downstream of the discharge is an impoundment;
upstream is a municipal discharge and (further upstream) two impoundments.
Evidence of Impairment
Biological Evidence: Biological monitoring in the Presumpscot River in Westbrook,
below a pulp and paper mill discharge, consistently revealed non-attainment of Class C
aquatic life standards (1984, 1994, 1995, 1996) using the standard Maine Department of
Environmental Protection methods (Davies et al. 1995, Davies and Tsomides 1997).
Biological evidence indicating impairment on the lower Presumpscot River is
summarized in Table 6-1 and Figure 6-2. Upstream samples consistently indicated
attainment of Class C or better aquatic life standards (Davies et al. 1999). Three
kilometers downstream the Presumpscot within the impounded area did not attain Class
C aquatic life standards.
Table 6-1. Evidence of biological impairment in the Presumpscot River upstream and
downstream of a pulp and paper mill effluent discharge.
Evidence
Aquatic Life Standard
Benthic
Macroinvertebrate
Community
Taxonomic Richness
Sensitive Species
(EPT)
Snails and Worms
Upstream of
Effluent
Class C
90% insects
~
~
Low
Downstream of Effluent
Non-Attainment
50% insects
15%-35% decrease relative to upstream
46%-60% decrease relative to upstream
High
The Presumpscot River biological monitoring samples reveal a shift in the benthic
macroinvertebrate community from 90% insects above the mill to about 50% insects
below the mill, with 15%-35% loss of taxonomic richness and 46%-60% loss of the
sensitive Ephemeroptera-Plecoptera-Trichoptera (EPT) groups (Mitnik 1998). Pollution-
sensitive insect taxa found in the upstream samples were replaced by a predominance of
snails and worms, which are more tolerant of pollution, in the downstream samples.
6.3 List Candidate Causes
Eight candidate causes for the non-attainment of biological standards were considered.
The candidate causes for the biological impairment of the Presumpscot River are shown
in terms of a conceptual model (Figure 6-3), wherein the candidate causes are ordered
from left to right. Each scenario is explained below:
1. Excess Toxic Chemicals - Potentially toxic compounds may be discharged from the
paper mill and these compounds adversely affect aquatic life.
Chapter 6: Presumpscot River, Maine
6-5
-------
Stressor Identification Guidance Document
Biological Indicators of Non-Attainment
] Species Richness
I No. of EPTTaxa
upstream
immediately
dow nstream
approx. 1 mile
dow nstream
Figure 6-2. Species richness and number of EPT taxa in the Presumpscot River
upstream and downstream of a pulp and paper mill effluent discharge.
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
CHARACTERIZE CAUSES
Eliminate
Diagnose
| Strength of Evidence
Identify Probable Cause
2. BOD (produced by high TSS with floe) reduces
DO - Excess total suspended solids (TSS with floe)
may be released by the paper mill effluent, and
these solids create biological oxygen demand
(BOD), reducing dissolved oxygen (DO) levels in
the river. Consequently, the river has insufficient
oxygen to support sensitive species of benthic
invertebrates.
3. TSS with floe - The increased levels of TSS
discharged to the river could impact the benthic
communities by accumulating as (non-
biodegradable) sediment, resulting in fewer
interstitial spaces in which animals can live, and
possibly smothering benthic biota.
4. Excess Nutrients - Excess nutrients, deriving from either upstream, non-point sources
or from the paper mill effluent, may affect water quality by promoting algal blooms. In
this scenario, an overabundance of plant nutrients such as phosphorus is delivered to the
stream, and over-stimulates algal growth (a process known as eutrophication). An
increase of algae in the river may affect benthic macroinvertebrates in two ways. If the
algal growth is severe, the resulting detritus becomes a source of BOD, reducing
dissolved oxygen levels in the river. If the growth is modest, the algae may still affect
the benthic macroinvertebrate community by providing an increased food supply for
opportunistic invertebrates that use algae as a food source. Consequently, the
community would shift in such a way that the opportunistic species would thrive and
outcompete other, less opportunistic species.
6-6
U.S. Environmental Protection Agency
-------
Sources:
Stressors:
Interactions:
Effect:
Altered
Hydrology
(Impoundment)
Toxic Chemicals
Excess Nutrients
Pool Conditions:
reduced flow velocity
stream widening
deepened water
Increased
Sedimentation
Increased TSS
Shift in BenthicMacroinvertebrateCommunity
Figure 6-3. Conceptual model showing the potential impact of stressors on the benthic community of the Presumpscot
River. (Arrow with minus sign (-) indicates inhibition.)
-------
Stressor Identification Guidance Document
The fifth through eighth candidate causes are based on impoundment of the river just
downstream of the paper mill effluent. Each cause begins with the idea that the
impoundment is causing adverse changes in the physical nature of the Presumpscot.
Impoundments generally widen and deepen a stream corridor, reducing flow velocity and
creating pool-like conditions. Such alterations can have several effects:
5. Impoundment Increases Sedimentation - One effect of impoundment is increased
sedimentation due to reduced flow velocity, which leads to fewer interstitial spaces in
which animals can live, and potentially smothers benthic ones.
6. Impoundment Promotes Algal Growth - The pool-like conditions created by the
impoundment become a better habitat for algal growth, and algal blooms occur.
Subsequently, benthic communities shift as a result of oxygen depletion or the
dominance of algae-consuming invertebrates, as described previously.
7. DO Reduction in Impoundment - An impounded river is deeper and slower, which
results in less potential for mixing and more potential for stratification, particularly in
warmer months. As a result, underlying water may not be sufficiently aerated, and
benthic diversity decreases in response to low dissolved oxygen levels.
8. Habitat Degradation caused by Impoundment - Changes in physical conditions of the
river caused by impoundment reduce optimal habitat for benthic organisms. The effect is
a direct one: native benthic macroinvertebrates are unable to thrive under the altered
conditions. Dissolved oxygen levels and other water quality parameters are not a factor.
6.4 Analyze Evidence and Characterize Causes: Eliminate
Physical and Chemical Evidence: Physical and
chemical evidence indicating impairment on the
lower Presumpscot River is summarized in Table 6-
2. Upstream of the pulp and paper mill outfall, it
was possible to see samplers on the river bottom at
2.5 meters of depth, whereas in the effluent plume,
just 600 m downstream, visibility was less than 0.5
meter. Visibility at a sampling station 3.2 km
downstream of the outfall remained significantly
impaired. This evidence was used to eliminate
candidate causes.
1. Toxic Chemicals - No in-stream or sediment
chemistry data were available. Therefore, toxic
chemicals cannot be eliminated as a candidate
cause.
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
CHARACTERIZE CAUSES
Eliminate Dagnose Strength of Evidence
6-8
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 6-2. Physical and chemical parameters measured in the Presumpscot River
upstream and downstream of a pulp and paper mill effluent discharge.
Observation
Visibility
Observations on
Sampling Equipment
(e.g., ropes, nets)
Mean TSS (ppm)1
Mean BOD (ppm)2
DO range (ppm) 3
Mean nitrate - nitrite
(ppm)2
Mean ammonia
(ppm)2
Mean Total
phosphorus (ppb)2
Mean Orthophosphate
(ppb)2
Mean Chlorophyll a
(ppb)2
Source
Mitnik1998
Mitnik1998
Courtemanch
etal. 1997
Mitnik1994
Mitnik1994
Mitnik1994
Mitnik1994
Mitnik1994
Mitnik1994
Mitnik1994
Upstream of
Mill
2.5m
Free of brown
floe
3 ppm
3.96
5.9-8.4
0.03
0.03
12.8
3.5
2.1
Downstream of Mill
<0.5 m (600 m below outfall)
and visibility remained
"significantly impaired" 3.2
km downstream
Coated with brown floe
5.9 ppm
6.19
5.8-8.0
0.05
0.12
61.2
44.3
2.3
Notes
1 Observations from 1995-96; number unknown
2 4 sites above mill and 5 sites below on 3 consecutive days
3 Bottom water; 9 sites above mill and 8 sites below on 6 non-consecutive days
2. BOD (produced by high TSS with floe) reduces DO - Elevated BOD was associated
with the biological impairment in the Presumpscot River. In this candidate scenario,
reduced DO is the actual stressor that acts on the organisms to cause impairment.
Monitoring in the Presumpscot River above and below the mill discharge indicated that
DO concentrations did not decrease upstream to downstream (Table 6-2 and Figure 6-4).
The reported DO measurements were taken at stations indicated on the map (Figure 1).
Most of the sites shown in Figure 1 were impounded water; only 7.7, and 6.3 were free-
flowing. The results reported in Tables 6-2 and 6-3 were all sampled between 0640 and
0850 hours, within 1m of the bottom, in July and August, 1993 (Mitnik 1994). This is
the time, depth, and season at which minimum DO is found in lakes and impoundments,
because of the diurnal cycle of photosynthesis and respiration, and because
photosynthesis (but not respiration) is inhibited in deeper and darker waters. This
analysis demonstrated that low DO does not occur in the Presumpscot River under any of
the candidate causes involving reduced dissolved oxygen. Therefore, candidate causal
scenarios # 2 (High TSS with floe causes high BOD and reduced DO) # 6
(Impoundment promotes algal growth that in turn reduces dissolved oxygen), and # 7
(Impoundment causes low DO through decreased water flow rate) could be eliminated
without further analysis. The elimination of scenario #6 is reinforced by the evidence
described in scenario #4, below.
Chapter 6: Presumpscot River, Maine
6-9
-------
Stressor Identification Guidance Document
j"
E
"*""'
c
>
X
O
o
o
I/)
I/)
b
g
8 -
7 -
6 -
5 -
4 -
3 -
f*
1 -
0
1 t f
i 1 T 1 ^ 1 1 1 i
"" t t ^ t t
T
i
i i i i i i i i i i i i
Max.
Min.
Avg.
c£>cQLnc£>r-~r~r~r~co ^p r~ co ^ co
-------
Stressor Identification Guidance Document
8. Habitat Degradation caused by Impoundment - Again, the biological impairment
found downstream of the paper mill discharge coincided with the presence of an
impoundment (Table 6-3). However, no measurements of habitat quality were available
to determine if the biological impairment was the result of habitat loss caused by the
impoundment. Therefore, scenario #8 could not be eliminated.
Following the process of elimination, 4 candidate causes remained:
1. Excess toxic chemicals.
3. High TSS with floe causes smothering.
5. Impoundment increases sedimentation that smothers biota (with or without
discharge of TSS and floe).
8. Impoundment causes loss of
suitable habitat.
6.5 Analyze Evidence and Characterize Causes: Strength of Evidence
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
CHARACTERIZE CAUSES
Direct observations in the Presumpscot River
during macroinvertebrate and fish tissue sampling
revealed a heavy suspended and settled solids load.
Samplers and gill nets were coated with flocculent
fibers and water clarity was dramatically reduced.
In comparison to other paper mills in the State, the
pulp and paper mill effluent released to the
Presumpscot was considered high strength for
solids. However, the conditions faced on the
Presumpscot were similar to those found below the
discharge from another paper mill on the
Androscoggin River in Jay, Maine. Because of this,
observations in the vicinity of the paper mill on the
Androscoggin River were used to support the
evidence found for this case study.
A comparison of the two rivers and discharge loadings to each is given in Table 6-4.
Paper mill discharges on both rivers were subject to impoundments with similar
hydraulic properties (e.g., velocity and depth) and background TSS concentrations
(about 3 ppm). Two or more dams impounded both rivers upstream of the discharges.
Eliminate Dagnose Strength of Evidence
J
Identify Probable Cause
Chapter 6: Presumpscot River, Maine
6-11
-------
Stressor Identification Guidance Document
Table 6-3. Considerations for eliminating candidate causes.
Candidate
Cause
Toxic
Chemicals
BOD
(produced by
TSS) reduces
DO
TSS with floe
Nutrients and
algal growth
Impoundment
increases
sediment
Impoundment
promotes
algal growth
DO reduction
in
Impoundment
Habitat
degradation
caused by
impoundment
Impairments
occur same
place as
exposure?
NE
BOD Yes;
TSS Yes;
DO No
Yes
Nutrients
Yes;
Algal Yes
Yes
Algal Yes;
DO No
Imp. Yes;
DO No
Yes
Exposure
increased
over
closest
upstream
location?
NE
BOD Yes;
TSS Yes;
DO No
Yes
Nutrients
Yes;
Algal No
NE
Algal No;
DO No
Imp. Yes;
DO No
NE
Gradient of
recovery at
reduced
exposure?
NE
NE
NE
NE
NA
NA
NA
NA
Exposure
pathway
complete?
NE
No
Yes
No
NE
No
No
NE
Candidate
Causes
Remaining
X
X
X
X
Table 6-4. Comparison of TSS loadings in the Presumpscot and
(Sample points were located below a pulp and paper mill effluent
Androscoggin Rivers.
discharge.)
Mill & Year
Sampled
Aquatic Life Status
TSS treatment
Sampling Months
Flow, cubic
feet/second (cfs)
TSS Discharged,
pounds/day
TSS discharged/flow
Presumpscot
1995
Non-
Attainment
none
June-Aug
418
7454
3.31
1996
Non-
Attainment
none
Aug-Sept
463
8795
3.52
Androscoggin
1995
Non-
Attainment
none
June-Aug
2114
19804
1.74
1996
Attainment
TSS
removal
Aug-Sept
2982
5750
0.36
1997
Attainment
TSS
removal
June-Aug
4116
13495
0.61
6-12
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Moreover, the upstream impoundments on both rivers attained at least Class C aquatic
life standards. However, both rivers were found to be in non-attainment of aquatic life
standards downstream of the paper mill discharges in 1995. Calculated mean ambient
concentrations of TSS in the Presumpscot downstream of the mill were 32% to 39%
greater than ambient levels downstream of the mill on the Androscoggin River. For the
most part, the incremental TSS increase on the Androscoggin River, due to paper mill
discharges, was within 1 ppm of background, while on the Presumpscot, the mill
discharge was about 3 ppm greater than background.
In 1996, efforts were made by another paper mill on the Androscoggin River to reduce
TSS discharge into the Androscoggin River. Following these efforts, the site's
biological score improved and the river met Class C aquatic life standards. This
recovery of biological conditions following TSS reduction provided experimental
evidence that TSS could also be the cause of ecological stress in the Presumpscot River.
Table 6-6 summarizes the types of evidence weighed in the analysis of potential stressors
in the Presumpscot River.
Other evidence used in the strength of evidence comparison is shown in Table 6-5.
Some metals exceeded chronic criteria when the maximum concentration in the effluent
was evaluated with a low flow scenario (Table 6-5). Although low DO was eliminated in
the previous step of this case study. Maine DEP performed an extensive modeling effort
to investigate the potential for low DO below the mill outfall. The modeling results
supported the conclusion that the DO concentrations did not fall below minimum levels
for Class C aquatic life uses (Mitnik 1998). Furthermore, during the same time period as
the biological monitoring, there were not violations of criteria for DO.
Table 6-5.1996 -1999 metal concentrations in the pulp and paper mill effluent.
Metals
Aluminum
Lead
Mercury
Silver
Range |jg/L
in Effluent Grab
Samples
1996-1999
108-1920
3-14
0.0001 - 0.9
10
Maximum
Receiving Water
Concentration
(|jg/L) at Low
Flow1 (7Q102)
207.9
1.52
0.097
1.083
Chronic
Criteria
(M9/L)
87
0.41
0.012
0.12
Acute
Criteria
(M9/L)
750
10.52
2.4
0.92
Notes
1 The receiving water concentration is calculated from the maximum effluent concentration divided
by a dilution factor of 9.
2 7Q10 + 7-day low flow over a ten year period.
Chapter 6: Presumpscot River, Maine
6-13
-------
Table 6-6: Strength of evidence of non-attainment in the Presumpscot River.
Case-Specific
Evidence:
Consideration
Spatial
Co-occurrence
Temporality
Consistency of
Association
Biological
Gradient
Complete
Exposure
Pathway
Experiment
TSS with Floe
Results
Compatible: Non-
attainment observed
in area of high TSS
and floe loading.
Attainment observed
in upstream areas
without TSS loading.
No observations prior
to paper mill
discharge.
No evidence
No evidence
Evidence for all steps:
High TSS and floe
discharge into river
well-documented.
No evidence
Score
+
NE
NE
NE
++
NE
Toxic Compound
Results
Evidence unavailable.
No observations prior
to paper mill
discharge.
No evidence
No evidence
No evidence
No evidence
Score
NE
NE
NE
NE
NE
NE
Impoundment increases
Sedimentation
Uncertain: Non-
attainment observed in
area of impoundment,
but no measurements
of sedimentation were
available.
No observations prior to
impounding
A site within the same
impoundment,
upstream of the mill
met aquatic life uses.
Not Applicable
No evidence
No evidence
Score
0
NE
NA
NE
NE
Impoundment causes Loss
of Habitat
Uncertain: Non-
attainment observed
in area of
impoundment, but no
observations of
habitat quality were
available.
No observations prior
to impounding
A site within the
same impoundment,
upstream of the mill
met aquatic life uses.
Not Applicable
No evidence
No evidence
Score
0
NE
NA
NE
NE
-------
Table 6-6 (continued): Strength of evidence for causes of non-attainment in the Presumpscot River.
Information
from Other
Situations or
Biological
Knowledge:
Consideration
Plausibility -
Mechanism
Plausibility -
Stressor-
Response
Consistency of
Association
Specificity of
Cause
Analogy
Experiment
TSS with Floe
Results
Plausible: Snails and
worms are adapted to
utilization of settled
solids.
TSS response from
Androscoggin study
and modeling
sufficient to cause
impairment.
Invariant: Other sites
on other rivers with
TSS have impaired
biological
communities.
Low: Other causes
elicit similar
responses.
No evidence
Concordant: Removal
of TSS in the
Androscoggin river
improved invertebrate
assembleges.
Score
+
++
+++
0
NE
+++
Toxic Compound
Results
Plausible: Toxic
compounds could
alter community
composition.
Ambiguous:
Assuming low flow
conditions and at the
highest
concentrations
reported for effluent
from the mill, chronic
aquatic life criteria
might be exceeded
for aluminum, lead,
mercury and silver.
However, if we
assume high flows at
the time of sampling
then neither acute nor
chronic aquatic life
criteria are likely to be
exceeded.
In some places:
Possibly could cause
effects if at maximum
values most of the
time, but unlikely
Low: Other causes
elicit similar
responses.
No evidence
No evidence
Score
+
0
0
0
NE
NE
Impoundment increases
Sedimentation
Plausible: Sediment
could alter habitat and
community
composition.
Other impoundments
with similar potential
sediment loadings
support diverse
invertebrate
communities.
Other impoundments
on other rivers are not
impaired.
Low: Other causes
elicit similar responses.
No evidence
No evidence
Score
+
0
NE
NE
Impoundment causes Loss
of Habitat
Plausible: Altered
habitat could change
community
composition.
Other impoundments
with similar habitat
support diverse
invertebrate
communities.
Other impoundments
on other rivers are
not impaired.
Low: Other causes
elicit similar
responses.
No evidence
No evidence
Score
+
0
NE
NE
-------
Table 6-6 (continued): Strength of evidence for causes of non-attainment in the Presumpscot River.
Considerations
Based on
Multiple Lines
of Evidence:
Consideration
Predictive
Performance
Consistency of
Evidence
Coherence of
Evidence
TSS
Results
No evidence
All Consistent.
Score
NE
+++
Toxic Compound
Results
No evidence
Not consistent: data
collected during the
same time period as
the biological
monitoring indicated
that there were no
violations of criteria for
toxic materials (Mitnik
1998).
Could be due to
unmeasured chemical
or episodic exposure.
Score
NE
0
0
Impoundment increases
Sedimentation
No evidence
Not consistent: Other
sites with
impoundments
maintained diverse
communities.
No known explanation.
Score
NE
0
0
Impoundment causes Loss
of Habitat
No evidence
Not consistent:
Other sites with
impoundments
maintained diverse
communities.
No known
explanation.
Score
NE
0
0
-------
Stressor Identification Guidance Document
6.6 Characterize Causes: Identify Probable Cause
Following the process of elimination, four
causal scenarios remained to compare for
strength of analysis (Table 6-6). These
scenarios were: #1 (excess toxic
chemicals), #3 (high TSS with floe causing
smothering), #5 (impoundment increasing
sedimentation that smothers biota, with or
without discharge of TSS and floe), and #8
(impoundment causing loss of suitable
habitat).
CHARACTERIZE CAUSES
Eliminate Diagnose
Strength of Ev
| Identify Probable Cause
dence
The evidence supporting scenario #3, that non-attainment was due to high TSS loads
combined with floe, was consistent throughout the lines of evidence. Moreover, the
strength of association, spatial co-occurrence, plausible stressor-response and experiment
lines of evidence strongly supported this scenario. Therefore, high TSS with floe was
sufficient for causing the biological impairment. The quality of the data are adequate for
this conclusion, and our confidence is high.
In contrast, evidence for the toxicity scenario was weak, because the stressor-response
association was unlikely based on levels of chemicals in the effluent and the likely
dilution provided by the river at the time of discharge. If greater certainty was required,
ambient receiving water toxicity tests could be used.
Likewise, evidence for the candidate causes involving impoundments lacked field
measurements of sedimentation and habitat quality. However, our confidence in
rejecting these scenarios as the primary cause of impairment is strengthened by the fact
that several upstream sites along the Presumpscot River were impounded with no
associated biological impairment (Mitnik 1998, Davies et al. 1999), and within the same
impoundment upstream from the mill, the Presumpscot met aquatic life uses.
Furthermore, several other impounded rivers of the state are able to meet Class B and C
biological criteria (Davies et al. 1999).
Nutrient levels were elevated; however, the algal concentration was not different from
the nearest upstream sampling location. As a result, candidate cause # 4, excess
nutrients, was eliminated; however, it is possible that the growth of algae was inhibited
by other factors, such as shading from floe. If floe were removed, then effects due to
eutrophication might become evident.
Low dissolved oxygen was also eliminated based on spatial patterns of DO along the
river. Other data is also available that increases the confidence that could have been
presented in a strength of evidence analysis. At the site, DO was not below 6 ppm. The
minimum DO level for Class C waters is 5 ppm. Maine DEP also performed an
extensive modeling effort to investigate the potential for low DO below the mill outfall.
The modeling results supported the conclusion that the DO concentrations did not fall
below minimum levels for Class C aquatic life uses (Mitnik 1998).
Chapter 6: Presumpscot River, Maine
6-17
-------
Stressor Identification Guidance Document
6.7 Significance and Use of Results
In December 1998, the U.S. Environmental Protection Agency approved a Total
Maximum Daily Load (TMDL) finding, prepared by Maine Department of
Environmental Protection, for the Presumpscot River. This approval was significant for
several reasons:
1. It was the first TMDL that addressed a listed 303(d) water to be approved in
Region 1 USEPA (the New England States);
2. It was the first time in New England that bioassessment findings had served as
the quantitative response variable from which a pollutant discharge limit was
developed.
The wastewater discharge license that has resulted from this effort requires an initial
30% reduction in the TSS discharge from a pulp and paper mill in Westbrook.
Provisions are included in the license for further reductions (up to 61%) if the initial
levels still fail to provide for attainment of aquatic life standards.
Main Factors Influencing Success
The Department was able to apply this innovative approach to improving water quality
and aquatic life conditions in the Presumpscot River because of the convergence of
several factors:
* The State has a sound legal basis for use of biological monitoring findings to
force action. Clearly defined aquatic life standards exist in the Water Quality
Classification law and technically-defensible numeric criteria have been
established by the Department;
* Data essential to the modeling of the recommended total suspended solids load
reductions on the Presumpscot River had been collected to assess aquatic life
issues on the Androscoggin River (under State requirements for a 401 Water
Quality certification for a hydropower license renewal);
* Teamwork and collaboration between DEP, water quality modelers, and aquatic
biologists resulted in an approach that integrated technical information and
expertise from both disciplines. It also provided a means for the Department to
control a stressor (TSS) for which the State has no standards.
6.8 References
Courtemanch, D.L., P. Mitnik, and L. Tsomides. 1997. Dec. 8, Memorandum to Greg
Wood, Maine Department of Environmental Protection Licensing Section, Augusta,
Maine.
Davies, S.P., L. Tsomides, D.L. Courtemanch, and F. Drummond. 1995. Maine
biological monitoring and biocriteria development program. Maine Department of
Environmental Protection, Augusta, Maine.
6-18 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Davies, S.P. and L. Tsomides. 1997. Methods for biological sampling and analysis of
Maine's inland waters. DEP-LW/07-A97. Maine Department of Environmental
Protection, Augusta, Maine.
Davies, S.P., L. Tsomides, J.L. DiFranco, and D.L. Courtemanch. 1999. Biomonitoring
retrospective: Fifteen year summary for Maine rivers and streams. DEPLW1999-
26. Maine Department of Environmental Protection, Augusta, Maine.
Hilsenhoff, W.L. 1987. An improved biotic index of organic stream pollution. Great
Lakes Entomol. 20:31-39.
Mitnik, P. 1994. Presumpscot River waste load allocation. Maine Department of
Environmental Protection, Augusta, Maine.
. 1998. Presumpscot River supplemental report to waste load allocation. Maine
Department of Environmental Protection, Augusta, Maine.
Chapter 6: Presumpscot River, Maine 6-19
-------
Stressor Identification Guidance Document
Chapter 7
Little Scioto River, Ohio
7.1 Executive Summary
This case study of the Little Scioto River represents an application of the SI process to a
complicated system. Impairment of the Little Scioto River reflected several impacts
caused by different stressors. Originally, the data on the Little Scioto were collected and
analyzed as part of the Ohio Environmental Protection Agency (OEPA) state monitoring
program during 1987, 1991, 1992 and 1998 (OEPA 1988b, 1992, 1994, unpublished data
from 1998) and as research for a USEPA methods development program. The
monitoring data were subsequently analyzed for this SI case study to demonstrate how
data collected from monitoring programs could be used to identify probable causes of
biological impairment.
The SI investigation was initiated because criteria in the state of Ohio's water quality
standards were violated in parts of the Little Scioto, a small river in north-central Ohio
(Yoder and Rankin 1995b). The SI investigation involved a 9-mile stretch of the Little
Scioto River near Marion, Ohio, where there was evidence of biological impairment.
The State of Ohio has a "tiered" set of aquatic life use designations based on narrative
definitions of specific aquatic uses that are protected by a set of numeric biocriteria,
chemical criteria, and habitat criteria. Ohio EPA determines biological impairment of
stream segments by comparing study sites to the numeric biocriteria in their water
quality standards. OEPA uses standard multimetric indices, including the Index of
Biotic Integrity (IBI), the Invertebrate Community Index (ICI) (OEPA 1989a), and the
Qualitative Habitat Evaluation Index (QHEI) (OEPA 1989c). Little Scioto River data
collected in 1987 and 1992 showed a condition of "fair" to "severe impairment" in the
stretch from river mile (RM) 9.2 to where the Little Scioto joins the Scioto River, just
downstream of RM 0.4.
Describe the Impairment
Three distinctive impairments (A, B, and C) were identified for the causal evaluation (at
RM 7.9, 6.5, and 5.7, respectively). Impairment A was characterized by a loss offish
and benthic invertebrate species, a decrease in the number of individual fish, and an
increase in the relative weight offish. Impairment B was characterized by a decrease in
the relative weight offish and a large increase in deformities, fin erosion, lesions, tumors
and anomalies (DELTA). Impairment C was characterized as having a further increase
in DELTA and extirpation of a Tribe of midges, the Tanytarsini.
List Candidate Causes
Stressors impacting the upper portion of the river were identified as mostly non-point
nutrient and sediment loadings associated with agriculture. Beginning at river mile 9.0
and continuing to the mouth, the river is channelized. The Little Scioto River at and
below Marion, Ohio, however, has been notably contaminated with elevated levels of
polycyclic aromatic hydrocarbons (PAH). Creosote and metals in sediment samples and
Chapter 7: Little Scioto River, Ohio 7-1
-------
Stressor Identification Guidance Document
ammonia, phosphorous (P), total nitrogen (N) were detected in water samples (OEPA
1994).
Based on the knowledge about the site and effects, six candidate causes were
hypothesized to account for the three major biological impairments observed in the Little
Scioto study area:
1. Habitat alteration: embedded stream and deepened channel
2. Exposure to PAHs
3. Metal contamination
4. Ammonia Toxicity
5. Low Dissolved Oxygen/High Biological Oxygen Demand
6. Nutrient Enrichment
Characterize Causes: Eliminate
Candidate causes were eliminated because the level of exposure to the candidate cause
did not increase compared to the nearest upstream location. Candidate causes that
remained after the elimination step are listed below:
> Impairment A (RM 7.9) habitat alteration, metal contamination, and nutrient
enrichment remained as probable causes.
* Impairment B (RM 6.5) PAH contamination, metal contamination, ammonia
toxicity, low dissolved oxygen/high biological oxygen demand, and nutrient
enrichment remained as probable causes.
* Impairment C (RM 5.7) metal contamination, ammonia, and nutrient
enrichment remained as probable causes.
Characterize Causes: Diagnose
No evidence strong enough to support diagnosis was available for any of the candidate
stressors.
Characterize Causes: Strength of Evidence
A strength of evidence approach was used to examine the remaining causes with regard
to each impairment. Evidence based on other situations and biological knowledge were
especially important including consistency of association and plausibility of mechanism
and stressor-response.
Characterize Causes: Identify Probable Causes
Impairment A
At Impairment A, the increased relative weight is probably caused by the artificial
deepening of the channel that allows larger fish to live there. The mechanisms were
7-2 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
probable, and consistency of association and experiments from other sites in Ohio and
elsewhere supported this finding for the specific impairments. The extirpation offish
and benthic invertebrates seems to be most likely due to embedded substrates. Although
low DO could also be a cause, upstream locations had even lower DO levels and yet had
a greater variety offish and invertebrate species.
Although metals were present, the likelihood of response at these concentrations is low.
Furthermore, the types of changes in the community, especially an increase in the
relative weight offish, is very unlikely with the candidate cause of metals. Although P
levels are slightly higher, effects are not associated with these phosphorous
concentrations elsewhere, and they do not exceed Ohio proposed criteria values for
effects. PAH and ammonia had already been eliminated because levels were the same or
lower than upstream. Low DO /BOD was also eliminated as an overall pathway;
however, low DO associated with channelization may still play a role, especially with
respect to the slight increase in the percentage of DELTA.
Impairment B
A single probable cause, toxic levels of PAH-contaminated sediments, is likely for the
three manifestations of Impairment B: decreased relative weight, increased DELTA, and
decreased species. All of the evidence support PAH contamination as the cause. There
is a complete exposure pathway at the location, and a clear mechanism of action for each
of the effects. The single most convincing piece of evidence is that the cumulative toxic
units of PAH were more than 300 times the probable effects level.
Metals are at sufficient concentrations to cause effects; however, they are at levels close
to upstream concentrations, and are less than 2% as toxic as the lowest cumulative toxic
units of PAH. Metal concentrations are high enough that they should be considered a
potentially masked cause. Reduced DO resulting from increased BOD is unlikely
because, downstream, even greater levels of BOD did not cause reduction of dissolved
oxygen. Ammonia and nutrient enrichment are unlikely given that state criteria levels
were met and given the much stronger evidence for PAH. Habitat alteration continues to
impair the site, but it is not the cause of the increased DELTA, decreased relative weight,
or the additional decline in the number of species, because the level of embeddedness
was similar to upstream.
Impairment C
At Impairment C, increased % DELTA and % Tanytarsini may have different causes.
Increased DELTA in fish is probably caused by increased P and N. Nutrients, especially
P, have been associated with increased fin erosion and lesions, but some uncertainty
exists since P acts indirectly. Another candidate cause is also probable, namely,
ammonia. Ammonia is slightly higher at Impairment C than at Impairment B, and
exceeded ammonia criterion values. Biological gradients were absent for ammonia;
however, this may have been a statistical artifact given the number of sites available to
perform the analysis, and the potential interference from other stressors downstream.
Metals are considered unlikely, because very specific surface lesions are only
occasionally noted as effects from long-term exposure, and only some metal
concentrations were slightly greater than at Impairment B. Metal concentrations are high
enough that they should be considered a potentially masked cause.
The probable cause of extirpation of Tanytarsini at Impairment C is more uncertain
because less is known about the natural history and stressor-response relationships of
Chapter 7: Little Scioto River, Ohio 7-3
-------
Stressor Identification Guidance Document
these benthic invertebrates. Nutrient enrichment still seems to be the most likely cause
since all of the strength of evidence considerations were consistent.
PAH contamination and habitat alteration continue to impair the site, but they are not the
cause of the increased percent DELTA or extirpation of Tanytarsini.
Identify Probable Cause
The most probable causes were:
> Impairment A (RM 7.9) Siltation and deepened channel are consistent with
impairment A. The magnitude of the alteration and clear difference from
upstream locations strongly support this cause.
> Impairment B (RM 6.5) PAH-contaminated sediments are likely causes for
the three manifestations of Impairment B.
> Impairment C (RM 5.7) The causal characterization at Impairment C is less
certain, but the strength of evidence favors increased nutrient enrichment as the
cause.
The Little Scioto case study is a good example of a complex system requiring a detailed
analysis. Although it was possible to identify the dominant causes of specific
impairments, other causes were present that had the potential to cause impairments if the
dominant cause was removed. For instance, habitat alteration associated with
channelization would still impair the entire river below RM 9.0.
7.2 Introduction
The Little Scioto case study involves a nine-mile stretch of a river suffering from several
impairments with different causes. Typical of similar stressor investigations, the data
examined for this case study were not collected or originally analyzed specifically for the
Stressor Identification Technical Guidance Document. Rather, they were collected as a
part of the Ohio EPA state monitoring program during 1987, 1991, 1992 and 1998
(OEPA 1988b, 1992, 1994, unpublished data from 1998), and as research for a USEPA
methods development program. These monitoring data were subsequently analyzed in
this study to demonstrate how data collected from existing monitoring programs could be
used to identify probable causes of biological impairment.
Various types of data were used in this case study, including chemical analyses
(sediment, water, and fish tissue) and biological assessment (biological community and
physical habitat). Methods for the collection and analysis of chemical data are described
in Ohio EPA (1989c). In 1992, one grab sample was taken, whereas in 1987, multiple
grab samples were taken. Other Ohio EPA data sets included biological assessment data
on fish and invertebrate assemblages and physical habitat measurements. In Ohio,
impairment of stream aquatic life uses are defined by standard multimetric indices
including the Index of Biotic Integrity (IBI) and the Invertebrate Community Index (ICI)
(OEPA 1989a). These indices have been promulgated as numeric biocriteria in the
State's water quality standards. The quality of the habitat is characterized using the
Qualitative Habitat Evaluation Index (QHEI) (OEPA 1989c). These methods are
described in detail by Ohio EPA (1989c). Biochemical measurements of impairment
included bile metabolites measured according to Lin et al. (1996) and ethoxy
7-4 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
resorufin[O]deethylase (EROD) activity measured according to Cormier et al. (2000b).
Although the attempt was made to use biological and chemical data from the same
locations, in some cases, chemical measurements were recorded at a location that did not
exactly coincide with the location of biological assessment (e.g., RM 5.8 and RM 5.7,
respectively). However, the distance between the chemical and biological sample sites
was negligible or overlapped, and the data were able to be used to analyze associations
between candidate causes and the biological impairment.
The Little Scioto River is a small river in north-central Ohio that empties into the Scioto
River (Figure 7-1). It drains primarily farmland in the northeastern quadrant of the
Eastern Corn Belt Plains ecoregion. The soils in this area are glacial till overlying
limestone, dolomite, and shale bedrock. The water table has been lowered in much of
the watershed by extensive use of tile drainage in crop fields. Near Marion, Ohio, the
Little Scioto is biologically impaired.
This causal investigation was initiated because the State of Ohio water quality standards
related to biological criteria were violated (Yoder and Rankin 1995a). The State of Ohio
has a "tiered" set of aquatic life use designations based on narrative definitions of
specific aquatic uses, which are protected by numeric criteria.
The majority of Ohio rivers and streams are designated as Warmwater Habitat (WWH)
(Yoder and Rankin 1995a). This designation is narratively defined as supporting a
balanced, reproducing aquatic community. Quantitatively, the minimum criteria
required to be in attainment of WWH standards are defined as the 25th percentile values
of reference condition scores for a given index, site type, and ecoregion. The choice of
the 25th percentile is considered to be conservative and will likely be influenced by the
inclusion of marginal sites as well as reference quality sites.
The Little Scioto River is considered Warmwater Habitat above RM 7.9 and a Modified
Warmwater Habitat at and below RM 7.9 (see Figure 7-1). The Modified Warmwater
Habitat (MWH) criteria are based on comparisons to a different reference condition than
are used for the WWH criteria (Yoder and Rankin 1995a). The MWH designation is a
non-fishable aquatic life use, and is designed to protect streams that have been too
impacted, or modified, to meet WWH standards. MWH streams are unlikely to recover
sufficiently to meet WWH designation. Consequently, MWH criteria are typically lower
than WWH criteria. In spite of poorer water quality conditions (such as low dissolved
oxygen, high ammonia concentration, and increased nutrient input), MWH streams are
nonetheless able to support permanent assemblages of tolerant species.
7.3 Evidence of Impairment
In 1987 and 1992, sampling and measurements for community and habitat indices (IBI,
ICI, QHEI) were conducted by OEPA along the Little Scioto River. Standardized field,
laboratory and data processing methods followed OEPA procedural guidelines (OEPA
1988a, OEPA 1989a,b,c, Rankin 1989). Fish and macroinvertebrates were sampled at
seven sites along the river, from river mile (RM) 9.5 to 0.4 (Figure 7-1). Index and
metric scores for IBI, ICI, and QHEI used in this study were obtained from data sets that
were generated and made available by OEPA as well as various OEPA reports (1988b,
1992, and 1994).
Chapter 7: Little Scioto River, Ohio 7-5
-------
Stressor Identification Guidance Document
SOURCES OF
RIVER MILE STRESSORS
9.2
Impairment A 7.9
Impairment B 6.5
Impairment C 5.7
Defunct creosote plant
RaU Facility
1 Miles
Figure 7-1. Map of the Little Scioto River, Ohio, showing sites where fish were
sampled. (Approximate locations of significant physical features, tributaries and point
source inputs are noted. The small inset shows the location of the study area in the
state of Ohio. Locations of Impairments A, B and C are also shown.)
Of the seven sites sampled in 1987, the highest IBI score was 34 (out of a possible score
of 60), which occurred at RM 9.2. This score translates to a. fair ranking according to
WWH standards. The remainder of sites were described as severely impaired, with IBI
scores between 25 and 12 (the lowest possible IBI score) (OEPA 1994, Yoder and
Rankin 1995a). In 1992, the IBI score at RM 9.2 decreased by one point to 33.
However, in 1992, the IBI score dropped 9 points to a score of 24 between RM 9.2 and
7-6
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
RM 7.9. Another 5 point drop occurred at KM 6.5 and scores stayed between 19 and 20
through RM 2.7. At RM 0.4, the IBI score climbed back to 25, greater than the adjacent
upstream site's score, but still indicating impairment. Figure 7-2A illustrates the
fluctuation of the IBI at the seven sites during the two sampling years (1987 and 1992).
Figure 7-2B traces a similar pattern of impairment for the invertebrate index during the
1987 and 1992 sampling years. The ICI met WWH aquatic life use standards in 1987
and 1992 at RM 9.2, with scores of 40 and 38, respectively (Figure 7-2). In 1992, the ICI
score declined 22 points at RM 7.9 with a score of 16, considered fair, but below MWH
aquatic life use standards. Scores further declined 12 or more points at RM 6.5, 5.7 and
4.4, with scores ranging between 6 and 10. These scores are indicative of highly
impaired conditions (OEPA 1994, Yoder and Rankin 1995a). ICI scores increased to a
value of 18 downstream at RM 2.7 and RM 0.4. In 1987, both IBI and ICI scores were
greater at RM 6.5 and then declined at RM 5.7, and remained very low to the mouth of
the Little Scioto.
A
B
45 i
4(1
35'
1 30'
> 25.
I20'
« 15 .
u. 10
10'
5'
n
AZ ,
40'
35
30'
0)
I25'
> 20'
()
- 15
10 '
5'
0.
1
\\
"--.AX- A-.. . **'*'
9.2 7.9 6.5 5.7 4.4 3.2 2.7 0.4
River Mile
K
, \
\\
\ \ ,-A- A
4 \ -'
*'* \ -*"*"
V \ ,.-4" ^
9.2 7.9 6.5 5.7 4.4 3.2 2.7 0.4
River Mile
IBI 87
---A---. IBI 92
WWH Criterion
MWH Pritprinn
ICI 87
...^.... Id 92
\finAN I fVi+nrirtn
MWH Criterion
Figure 7-2. Spatial changes in fish IBI (A) and benthic macroinvertebrate ICI (B)
values in the Little Scioto River in 1987 (OEPA 1988) and 1992 (OEPA 1994).
Chapter 7: Little Scioto River, Ohio
7-7
-------
Stressor Identification Guidance Document
The impairments seen below RM 9.2 were more specifically described by examining the
metrics that make up the IBI and the ICI. This information was combined with the
changes seen in the overall IBI and ICI scores to determine whether distinctive patterns
of impairment could be identified. Each distinctive impairment required a separate
causal evaluation.
A subset of the fish and macroinvertebrate metrics, selected to highlight differences in
community patterns, is shown in Figures 7-3A (fish) and 7-3B (macroinvertebrates). The
complete list of values for the metrics is shown in Tables 7-13 and Table 7-14 (Please
note that Tables 7-13 through 7-20 are located in Section 7.13, "Additional Data
Tables"). One of the metrics, relative weight offish, is not a component of the IBI but a
component of another index, the Modified Index of Well-being (MIWB).
Examination of the spatial distribution of the IBI, ICI, and metric patterns in 1992
indicates that at least three distinct impairments occurred:
* Impairment A was seen at RM 7.9 where a marked drop in both the IBI and ICI
occurred relative to the upstream location at RM 9.2. Specific fish metrics that
appeared to correspond to this drop included decreases in the number of
individuals minus tolerant fish, decreased total number of species, and increased
relative weight. In addition, the percentage of mayfly species decreased.
> Impairment B occurred at RM 6.5 and corresponded with an additional decrease
in both the IBI and the ICI. Relative the upstream location at RM 7.9, fish
relative weight decreased, the number of deformities, erosions, lesions, tumors
and anomalies (DELTA) increased, and the percentages of mayflies and
Tanytarsini midges also decreased while the percentage of tolerant organisms
increased.
* Impairment C occurred at RM 5.7. There was no change in the IBI relative to
RM 6.5, although relative weight offish decreased and DELTA increased. The
invertebrates had variable changes depending on the sampling year. In 1987 and
1992, the % Tanytarsini midges decreased or disappeared entirely. Changes in
the metrics at these three locations are summarized in Table 7-1.
The biological assessment data for the remaining locations showed a pattern similar to
Impairment C, with the possibility of intensification at RM 4.4 and some improvement in
metric scores occurring at RM 2.7 and 0.4. A fourth impairment was not hypothesized
for RM 4.4 because the pattern offish and invertebrate metrics were fairly similar to
those seen at RM 5.7.
7-8 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
A L2
xi.
1.0
SJ
a
o .
.1 s
§ 13 0.6
§ ^
M fi
£ 0.4
1
0.2
A rj
u.u
B 12
_1-F
1.0
09
0 0.8
VI ft
i £
U is
M S 0.4
L.
0
0.2
n A
\ ' \
V : V y # fish minus
Channelization :CSO
Begins ;
-
-
r
II
J
WWTP
Landfill
[J
1
1
tolerants
CD # of species
relative weight
CD % DELTA
J
1
-,
I
1
% mayflies
CD % Tanytarsini
-
_
~
IL
In II 1
9.2 7.9
A
6.5
B
5.7 4.4
C
~
1
|
1
1
1
% tolerant
organisms
CD % Cricotopus
,
2.7
0.4
River Mile
Figure 7-3. Changes in the IBI and ICI scores over distance in the Little Scioto
River, 1992. ((A) Changes in the relative scores for the total number of individual
fish minus tolerant fish (# fish minus tolerant), the number of species (# species),
the relative weight of fish (relative weight) and the percentage of DELTA. (B)
Changes in the relative abundances of percent Ephemeroptera, Tanytarsini,
tolerant organisms, and Cricotopus, in the Little Scioto River. Normalized values
were calculated by dividing the value at the individual site by the highest value for
all sites.)
Chapter 7: Little Scioto River, Ohio
7-9
-------
Stressor Identification Guidance Document
Table 7-1. Summary of the three impairments that were considered in the Little
Scioto River. (Each location is scored relative to the location immediately upstream,
based on 1992 data.)
Response
Impairment A
RM7.9
Impairment B
RM6.5
Impairment C
RM5.7
Fish
# of individuals minus
tolerant individuals
# Species
Relative Weight
DELTA
-
-
7.9
0
+
-
-
0
-
0
-
5.7
Invertebrates
% Mayflies
% Tanytarsini midges
% Tolerant taxa
% Cricotopussp.
-
0
0
-
-
-
0
0
-
-
-
0
(+) indicates an increase in the metric relative to the next upstream location
(-) indicates a decrease
(0) indicates no change.
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
7.4 List Candidate Causes
Evidence Used to Develop Candidate Causes
Many point and non-point sources of pollutants are
associated with the Little Scioto River. Stressors
impacting the upper portion of the river are mostly
non-point nutrient and sediment loadings associated
with agriculture. However, the Little Scioto River,
at and below Marion, Ohio, has been notably
contaminated with elevated levels of polycyclic
aromatic hydrocarbons (PAH). Creosote and metals
were found in sediment samples, and ammonia was
detected in water samples (OEPA 1994). The
OEPA has, in fact, recently requested Superfund
support in the clean-up of an abandoned wood
creosote plant suspected of polluting the river since
the 1860's (Edwards and Riepenhoff 1998). An oily
sheen was noted on the river between river miles 6.5 and 5.8 during a site visit in 1992
(Cormier, pers. observ.). In-stream habitat quality was also degraded by channelization
that took place in the early 1900's (OEPA 1994). Locations of the potential sources and
stressors, including a landfill and wastewater treatment plant (WWTP), are shown in
Figure 7-1.
CHARACTERIZE CAUSES
Eliminate
Dagnose
7-10
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
List of Candidate Causes and Scenarios
As noted previously, three distinctive impairments were identified for the causal
evaluation. Based on the knowledge of the sources and effects, six candidate causes
were formulated to account for the impairment observed at each site. A conceptual
model of these candidates is provided in Figure 7-4.
1. Habitat Alteration - Habitat alteration, resulting from channelization, combines a
complex interaction of several stressors. These stressors are evident at RM 7.9 and
continue to the mouth of the river. Channelization can alter biological communities by
changing the physical structure of the stream and the flow characteristics of the water,
ultimately lowering dissolved oxygen, increasing siltation, and reducing substrate
complexity. This complex suite of stressors also includes: decreased woody debris,
which reduces available substrate and changes the energy source; decreased sinuosity,
which changes flow characteristics; erosional patterns and substrates; increased channel
depth that favors larger species offish; loss of pools that act as refugia; and loss of riffles
that oxygenate water and transport sediment (Tarplee et al. 1971, Karr and Schlosser
1977, Yount and Niemi 1990, Allan 1995).
2. PAH and 3. Metals - Biological impairment could also have been caused by toxic
stress. Historically, the river has provided a means of waste disposal for various
industries, whose effluents have contained metals, PAH, and creosote. Waste materials
may have also been buried in the landfill below RM 6.5 (OEPA 1994). All are
potentially toxic to aquatic life, and some have the ability to bioaccumulate through the
food web (Eisler 2000a,b). Thus, two candidate causes emerge: candidate cause #2 is
that biological impairment has occurred due to PAH exposure (with PAH emanating
from creosote deposits), and candidate cause #3 attributes impairment to metal
contamination.
4. Ammonia Toxicity - Ammonia is directly discharged into streams by point sources
(Russo 1985, Miltner and Rankin 1998). Ammonia can also be formed as the result of
nutrient enrichment. When dissolved oxygen levels are low, nitrates are reduced to
ammonium ion. If pH is high, some of the ammonium ion is converted to un-ionized
ammonia, which is toxic to aquatic organisms (Russo 1985). Moreover, pH may rise
during periods of high photo synthetic rates from bicarbonate depletion. High amounts of
nutrients often lead to increased algal growth rates, and the conversion of ammonium to
un-ionized ammonia is expedited (Dodds and Welsh 2000).
5. Low Dissolved Oxygen/ High Biological Oxygen Demand - Depletion of DO
commonly occurs from organic enrichment (Smith et al. 1999). Organic enrichment is
the most common cause of increased biological oxygen demand (BOD) (Allan 1995).
Potential sources of excess organic matter within the study area include a waste water
treatment plant (WWTP) and several combined sewer outfalls (CSOs), as well as
upstream, non-point sources. Organic matter is also produced by excess algal growth
from nutrient enrichment (Dodds and Welsh 2000). Algal blooms themselves result in
increased organic matter regardless of DO depletion. The algal bloom may suffice to
raise BOD so that DO is depleted. Because no chlorophyll a or algal biomass data were
collected in this study, the cause of BOD to the river can only be estimated from BOD,
measured at several points, and COD (chemical oxygen demand) measured at point
sources such as the WWTP above RM 5.4 in 1998.
Chapter 7: Little Scioto River, Ohio 7-11
-------
Sources
Channelization
Stressors
jndfill
Hazardous
Waste Site
i
Historical
Industrial
Discharge
r
^
Con
S
Ove
Altered stream morphology.
decreased wood debris,
decreased sinuosity,
decreased spatial diversity of
flow and depth, decreased
do, increased maximum
stream flow, increased
sediment deposition
Effects
Combined
Sewer
Overflows
>
Municipal
Waste
Treatment
Plant
/
Upstream
Fertilizer
Use
Nutrients.
hosphorus
Organic Carbon
Increased Toxic
Substances. PAHs
and Metals
Increased Total
Ammonia
(NH4+&NH3)
Increased
BOD
Reduced
Dissolved
Oxygen
Increased
Un-ionized
Ammonia
(NH3)
Shift in Fish and Benthic Macromvertebrate Community Structure
Figure 7-4. A conceptual model of the six candidate causes for the Little Scioto stressor identification. (Potential sources are
listed in top most rectangles. Potential stressors and interactions are located in ovals. Candidate causes are
numbered 1 through 6. Note that some causes have more than one stressor or more than one step associated with it.
The impairments are located in the lower rectangle.)
-------
Stressor Identification Guidance Document
6. Nutrient Enrichment - The sixth and final candidate cause is a less extreme form of
nutrient enrichment. Primary production and organic matter loading to the sediments are
increased, but not enough to reduce DO. This can cause changes in fish and benthic
macroinvertebrate assemblages, including changes in dominant species, and greatly
increased abundance and biomass (Carpenter et al. 1988, Rankin et al.1999, Smith et al.
1999, Dodds and Welsh 2000, Edwards et al. 2000). This form of nutrient enrichment is
also associated with fin erosion (Rankin et al. 1999).
7.5 Analyze Evidence to Eliminate Alternatives
7.5.1 Data Analyzed
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
Habitat alteration-related data
Data on the spatial location of habitat alteration was
obtained by using the Qualitative Habitat Evaluation
Index (QHEI). The QHEI incorporates measures of
habitat condition and has been correlated with the
IBI. This index uses eight interrelated metrics,
which assess substrate type and quality; in-stream
cover type and amount; channel morphology;
riparian width and quality and bank erosion; pool /
riffle characteristics including depth, current, pool
morphology, substrate stability and riffle
embeddedness; and finally gradient (Rankin 1989).
Based on these metrics, a total score is assigned to a
stream reach out of a possible 100 points, with
greater scores indicating higher quality. The channel morphology and substrate metrics
are particularly relevant for this case because of the channelization (Figure 7-5). Values
for the QHEI and its component metrics are given in Table 7-15 (see Section 7.13).
Stressor Identification
CHARACTERIZE CAUSES
95 79 65 58
95 79 65 58
27 04
Figure 7-5. Selected QHEI metrics for 1987 and 1992. (Scores are qualitative
ranks.)
Chapter 7: Little Scioto River, Ohio
7-13
-------
Stressor Identification Guidance Document
Chemical Data
Data on sediment and in-stream chemistry were used to evaluate the spatial location of
the remaining candidate causes (#2-6). Nutrient concentrations measured in water
included ammonia, nitrates and nitrites (NOX), phosphorus (P), and BOD. Ambient
levels of potential toxic chemicals were determined for sediment and water. Results of
chemical analyses are presented in Tables 7-16, 7-17 and 7-18 (See Section 7.13), and
Figures 7-6, 7-7, and 7-8.
While PAHs were not detectable at the upstream sites (RM 9.5 and 7.9), many PAHs
were detected between RM 6.5 and 0.4 (Table 7-16) (Figure 7-6). Spearman Rank
Correlations between chemical and biological data from 1992 at RM 5.7 to 0.4 are
shown in Table 7-2 through 7-5.
Metals were found in sediments at relatively high concentrations at RM 6.5 and
downstream (Table 7-17; see Section 7.13) (Figure 7-7). These included lead, cadmium,
copper, chromium, zinc, and mercury. Arsenic was relatively high at upstream reference
and study sites. Spearman rank correlations between metals and biological data from
1992 at RM 5.8 to 0.4 are shown in Table 7-3. Strong correlations having the sign that is
consistent with the hypothesis were noted for copper and mercury.
The water quality parameters ammonia, nitrates and nitrites (NOX), and BOD increased
substantially at RM 5.8, and remained elevated. Dissolved oxygen declined at 7.9 and
remained low to RM 0.4 (Table 7-18; see Section 7.13) (Figure 7-8). Spearman rank
correlations of water chemistry and biological endpoints are presented in Table 7-4.
Percent Tanytarsini are significantly correlated with DO, BOD, NOX and P, and the
negative direction of the slope was consistent with ecological theory. Percent DELTA
was correlated with the same parameters (DO, BOD, NOX and P) but at the 0.8 level,
whereas percent Cricotopus was associated with ammonia and the QHEI.
7.5.2 Associations between Candidate Causes and Effects
The associations between candidate causes and effects were analyzed by combining data
on the location of the three impairments with data on habitat quality and chemical
concentrations in water and sediments. The analyses evaluated whether the candidate
causes and each of the three impairments were spatially co-located, and whether a
gradient in recovery corresponded with a decrease in the candidate cause. These
associations are organized in table format (Table 7-5).
The first objective of the analysis was to determine if there was evidence that the
candidate cause occurred at the same place as the impairment but not where that
particular impairment was absent. Plots of the channel quality and substrate metrics
from the QHEI are shown in Figure 7-5. The chemistry values relevant to each of the
causal scenarios are shown in Figures 7-6, 7-7 and 7-8. Each graph shows the level or
concentration of the parameter. The presence or absence of candidate causes at the
locations of Impairments A, B, and C are summarized in Table 7-5.
7-14 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
500
800
c
* 600 -
c
dj 200
a! i*n .
i 15°
Vmn
CO
2 50 -
£ 300
n 200 -
* 100 -
°- o -
i 400 -
^ ^uu
200 -
S_^
11.1 9.5 79 7.1 65
O
11.1 9.5 7.9 7.1 6.5
In
0 0 II INI
11.1 9.5 7.9 7.1 6.5 i
0* 0 o| |
11.1 9.5 7.9 7.1 6.5
.
rn« I
11.1 9.5 79 7.1 6.5
11.1 9.5 7.9 7.1 6.5
II
o o -^o| |
11.1 9.5 7.9 7.1 6.5 !
_,_ o
5.8 4.4 2.7 0.4
~l» o _ OH n
5.8 4.4 2.7 0.4
~
n _ OB n
>.8 4.4 2.7 0.4
0 0
58 4.4 2.7 0.4
0 _|-l Or-, r-,
5.8 4.4 2.7 0.4
- _ o
5.8 44 2.7 0.4
O
5.8 4.4 2.7 0.4
1987
D1991
1998
Figure 7-6. Mean PAH concentrations from the sediment (mg/kg) in the Little Scioto
River 1987-1998. ((o) indicates below detection limit. Absence of bar indicates no data
available.)
Chapter 7: Little Scioto River, Ohio
7-15
-------
Stressor Identification Guidance Document
a 150 -
(0
o>
300
c 250
200
5 50
c 700 -
100
rf Eft
0
350
2 300
m o -J
o 150 -
§ 50
0 0 ^
m
N 250
o>
o>
n
o 50
S o
01
o 160 -
IE 80
v n _
CO °
T
I_
11.1 9.5 7.9 7.1 6.5
11.1 9.5 7.9 7.1 6.5
11.1 9.5 7.9 7.1 6.5
I I
ol I
11.1 9.5 7.9 7.1 6.5
n
o o ,_, J I
11.1 9.5 7.9 7.1 6.5
I I
o on B-r-B
11.1 9.5 7.9 7.1 6.5
_
dn
11.1 9.5 7.9 7.1 6.5
n
0 0 I
11.1 9.5 7.9 7.1 6.5
n
1
h-, n _ o_ ,_,
5.8 4.4 2.7 0.4
_ 0 0 0
5.8 4.4 2.7 0.4
0 0
5.8 4.4 2.7 0.4
, ,
| | o
5.8 4.4 2.7 0.4
1 |
n 11
5.8 4.4 2.7 0.4
. .
_ ,-,
5.8 4.4 2.7 0.4
o_
5.8 4.4 2.7 0.4
- on ^
5.8 4.4 2.7 0.4
1987
D 1992
1998
Figure 7-6 (continued). Mean PAH concentrations from the sediment (mg/kg) in the
Little Scioto River 1987-1998. ((o) indicates below detection limit. Absence of bar
indicates no data available.)
7-16
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
16
I 0
Q) p
R .
^ d
12
m
E 10
E 6
0
0 2
500
300
O 100
140
120 -
g- fin
o 4n
20 -
0 -
250
200
"O 1RD -
_3 100
50
0 -
]
>. 0-8
g U.b
A n 4 -
S 02
0 -
800
700
u 500 -
300
200
0 -
n
L
n
III
11.1 9.5 7.9 7.1 6.5 5.8
n
m* n n 1 ~i* n~
11.1 9.5 7.9 7.1 6.5 5.8
-i i-n
11.1 9.5 7.9 7.1 6.5 5.8
I
I I
11.1 9.5 7.9 7.1 6.5 5.8
~Z^f
r
11.1 9.5 7.9 7.1 6.5 5.8
m ^
r-ri n r-i in
11.1 9.5 7.9 7.1 6.5 5.8
In r
i r
11.1 9.5 7.9 7.1 6.5 5.8
|
4
1,
4
4
|
1
1,
4
-|
1,
4
1
1
4
4
-^
4
4
4
4
4
4
4
2
2
2
,1
2
2
2
2
I 1
| |
7 0
-
7 0
1
7 0
J
1 1
7 0
,
7 0
M
rl
7 0
~L
n
7 0
4
:
4
-|
4
4
1
4
-
4
4
1987
n 1991
D 1992
Figure 7-7. Mean metal concentrations from the sediment (mg/kg) in the Little Scioto
River from 1987-1998. (Absence of bar indicates no data available.)
Chapter 7: Little Scioto River, Ohio
7-17
-------
Stressor Identification Guidance Document
10
g
m 4
10
1 1
1 1
I I
25
2
n 1.5
Z 1
0.5
500
400
g1 300
« 200
100
3
2.5
2
0. 1.5
1
0.5
10
O
0 4
25
20
1 =
10
5
0
11.1 9.5 7.9 7.1 6.5 5.8 44 27 0.4
tm
11.1 95 7.9 7.1 6.5 58 44 27 04
11.1 9.5 7.9 7.1 6.5 5.8 4.4 2.7 0.4
MSB
1
11.1 9.5 7.9 7.1
6.5 5.8
4.4
i
11.1 9.5 7.9 7.1 6.5 5.8 4.4 2.7 0.4
m
11.1 9.5 7.9 7.1 6.5 5.8 4.4 2.7 0.4
11.1 9.5 7.9 7.1 6.5 5.8 4.4 2.7 0.4
11.1 9.5 7.9 7.1 6.5 5.8 4.4 2.7 0.4
2.7 0.4
1987
D 1992
1998
Figure 7-8. Mean water chemistry values from the Little Scioto River from 1987-1998.
(BOD, NOX, Ammonia, CaCO3, PO4, are all mg/L, Temperature (°C). DO is also mg/L
and is the minimum value obtained from grab samples for each year. Absence of a bar
indicates no data for that year.)
7-18
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-2. Spearman rank correlations with selected metrics and the IBI and ICI from
1992 and selected PAHs. (Reflects only values from RM 5.8 to 0.4. Correlations N=4).
Parameter
Anthracene (#2)
Benzo[a]anthracene (#2)
Benzo[ghi]perylene(#2)
Benzo[a]pyrene (#2)
Chrysene (#2)
Dibenzo[a,h]anthracene
(#2)
Fluoranthene (#2)
Fluorene (#2)
Naphthalene (#2)
Phenanthrene (#2)
Pyrene (#2)
DELTA
0.60
0.00
0.00
0.00
0.80*
-0.21
0.00
0.74
0.26
0.60
0.00
% Tanytarsini
Midges
-0.74
-0.21
-0.21
-0.21
-0.95*
-0.06
-0.21
-0.89*
-0.54
-0.74
-0.21
% Cricotopus
-0.20
-0.40
-0.40
-0.40
0.40
-0.21
-0.40
0.11
0.26
-0.20
-0.40
Correlations above 0.8
Table 7-3. Spearman rank correlations with selected metrics and the IBI and ICI from
1992 and selected metals. (Reflects only values from RM 5.8 to 0.4. Correlations N=4).
Parameter (Candidate
Cause)
Arsenic (#3)
Cadmium (#3)
Chromium (#3)
Copper (#3)
Lead (#3)
Mercury (#3)
Zinc (#3)
DELTA
0.74
0.20
0.80*
1.00*
0.40
1.00*
0.40
% Tanytarsini
Midges
-0.89*
0.11
-0.63
-0.95*
-0.32
-0.95*
-0.32
% Cricotopus
0.11
-0.60
0.40
0.20
0.80*
0.20
0.80*
* Correlations above 0.8
Chapter 7: Little Scioto River, Ohio
7-19
-------
Stressor Identification Guidance Document
Table 7-4. Spearman rank correlations with selected metrics and the IBI and ICI from
1992 and selected water quality and habitat quality measurements. (Reflects only
values from RM 5.8 to 0.4. Correlations N=4).
Parameter (Candidate
Cause)
Channel Metric (#1)
QHEI (#1)
Ammonia, N (#4)
Dissolved oxygen maximum
(#5)
Dissolved oxygen minimum
(#5)
BOD (#5)
Nitrate-nitrite, N (#4,5,6)
Phosphorus, total P (#5,6)
DELTA
0.77
0.20
0.40
0.80*
0.60
0.80*
0.80*
0.80*
% Tanytarsini
Midges
-0.82*
-0.32
-0.32
-0.95*
-0.74
-0.95*
-0.95*
-0.95*
% Cricotopus
0.77
1.00*
0.80*
0.40
-0.20
0.40
0.40
0.40
Correlations above 0.8
The second objective was to determine if the cause increased compared to the nearest
upstream location. Statistical analyses were not used to determine an increase because
the power would be very weak due to small sample sizes. Even a small increase was
accepted since it might represent a threshold for the effect (Table 7-5).
The third objective of the analyses was to evaluate whether a gradient in the intensity of
the potential cause corresponded to a gradient of recovery in impairment. The gradient
analysis was conducted only for Impairment C, which was observed at four contiguous
locations (i.e., RM 5.8 to 0.4). The recovery of Impairment B could not be analyzed
since it would be masked by Impairment C. Similarly, any recovery of Impairment A
would be masked by both B and C. The gradients in environmental parameters and the
IBI and ICI were examined visually by comparing Figures 7-2 and 7-3 with Figures 7-5
through 7-8. The IBI and ICI metrics for 1987 and 1992 data are shown in Table 7-13
and Table 7-14, respectively. In addition, Spearman's rank correlations were calculated
using the 1992 data set to relate the biological metrics (shown in Figure 7-3) with each of
the parameters related to the candidate causes. The results of this analysis are shown in
Tables 7-2 through 7-4.
Two metrics are more severe at Impairment C: % DELTA and % Tanytarsini midges
decrease and % Cricotopus increases. Percent DELTA were significantly correlated
with copper and mercury, and moderately correlated with chrysene, chromium, BOD,
nitrate, phosphorous, and maximum DO. The change in tanytarsini midges was
negatively and strongly correlated with chrysene, copper, mercury, BOD, nitrate,
phosphorous, maximum dissolved oxygen, and moderately correlated with fluorene,
arsenic, and the channel metric. The change in % Cricotopus was strongly positively
correlated with QHEI and moderately correlated with lead, zinc, and ammonia.
7-20
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-5. Evidence for eliminating candidates causes at Impairments A, B, and C.
Impairment A
Impairment B
Impairment C
Habitat Alteration (Candidate Cause 1)
Is there exposure at the
same location as the
impairment?
Is exposure increased
over the closest
upstream location?
Is there a gradient of
recovery as exposure
decreases?
Is the exposure pathway
complete?
Yes
Yes
NA*
(Gradient in
impairment is
masked by B
and C)
Yes
Yes
No
NA
(Gradient in
impairment is
masked by C)
Yes
Yes
No
No
(Correlation
coefficients have
the wrong signs,
with % DELTA and
% Tanytarsini)
Yes
PAH Contamination (Candidate Cause 2)
Is there exposure at the
same location as the
impairment?
Is exposure increased
over the closest
upstream location?
Is there a gradient of
recovery as exposure
decreases?
Is the exposure pathway
complete?
No
No
NA
(Gradient in
impairment is
masked by B
and C)
No
Yes
Yes
NA
(Gradient in
impairment is
masked by C)
Yes
Yes
No
(based on
metabolite values
in fish)
Inconclusive
(Mixed results)
Yes
Chapter 7: Little Scioto River, Ohio
7-21
-------
Stressor Identification Guidance Document
Table 7-5 (continued). Evidence for eliminating candidates causes at Impairments
A, B, and C.
Impairment A
Impairment B
Impairment C
Metal Contamination (Candidate Cause 3)
Is there exposure at the
same location as the
impairment?
Is exposure increased
over the closest
upstream location?
Is there a gradient of
recovery as exposure
decreases?
Is the exposure pathway
complete?
Yes
Yes
(all metals
greater in some
years)
NA
(Gradient in
impairment is
masked by B
and C)
Yes
Yes
Yes
(all metals
greater)
NA
(Gradient in
impairment is
masked by C)
Yes
Yes
Yes
(copper and zinc
increased)
Yes
(Tanytarsini
midges and %
DELTA are
strongly correlated
with copper and
mercury)
Yes
Ammonia (Candidate Cause 4)
Is there exposure at the
same location as the
impairment?
Is exposure increased
over the closest
upstream location?
Is there a gradient of
recovery as exposure
decreases?
Is the exposure pathway
complete?
Yes
No
NA
(Gradient in
impairment is
masked by B
and C)
No
Yes
Yes
NA
(Gradient in
impairment is
masked by C)
Yes
Yes
Yes
NA
(ammonia
increases below
RM 5.8)
Yes
7-22
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-5 (continued). Evidence for eliminating candidates causes at Impairments
A, B, and C.
Impairment A
Impairment B
Impairment C
Low Dissolved Oxygen/High BOD (Candidate Cause 5)
Is there exposure at the
same location as the
impairment?
Is exposure increased
over the closest
upstream location?
Is there a gradient of
recovery as exposure
decreases?
Is the exposure pathway
complete?
Yes
No
(DO is
depressed,
BOD
unchanged)
NA
(Gradient in
impairment is
masked by B
and C)
Yes
Yes
Yes
(BOD is two
times greater in
1992, DO slightly
less)
NA
(Gradient in
impairment is
masked by C)
Yes
Yes
No
(BOD is elevated,
but DO is greater
than either RM 7.9
orRM6.5)
NA
(ammonia
increases below
RM 5.8)
No
Nutrient Enrichment (Candidate Cause 6)
Is there exposure at the
same location as the
impairment?
Is exposure increased
over the closest
upstream location?
Is there a gradient of
recovery as exposure
decreases?
Is the exposure pathway
complete?
Yes
Yes
NA
(Gradient in
impairment is
masked by B
and C)
Yes
Yes
Yes
NA
(Gradient in
impairment is
masked by C)
Yes
Yes
Yes
Yes
(% Tanytarsini and
% DELTA are
strongly correlated
with NOX and Total
P)
Yes
NA* = not applicable
Chapter 7: Little Scioto River, Ohio
7-23
-------
Stressor Identification Guidance Document
7.5.3 Measurements Associated with the Causal Mechanism: Exposure Pathways
The exposure pathways are shown in Figure 7-4. Lines of evidence for each exposure
pathway are discussed below and are summarized in Table 7-11. To refute an
hypothesis, a step in the pathway must be absent.
Habitat Alteration (#1) - Channelization results in a constellation of stressors, including
loss of riffles with increased sediment deposition, and decreased DO. The QHEI metrics
can yield insights into specific changes: for example, riffle scores are zero throughout
the channelized portion of the stream (Table 7-15; see Section 7.13), substrate quality
and embeddedness due to fine sediment drops at RM 7.9, and DO also drops at RM 7.9.
The co-occurrence of macroinvertebrates with changes in physical structure may be
somewhat lessened, because Hester-Dendy samplers create an artificial solid substrate
for colonization. The ICI score does include a qualitative kick net sample that is
independent of the artificial substrates. The exposure pathway for habitat alteration is
complete for Impairments A, B, and C.
PAHs (#2) - Exposure to PAHs involves two steps: direct contact with external tissues
and uptake into the organism. Because the PAH information in this case is from the
sediments, we assume that fish and benthic invertebrates between river miles 6.5 and 0.4
will contact this contamination. Concentrations of PAH in the sediment were used only
from samples collected in 1992, as it was the only year in which we were confident that
the samples were collected from the top six inches. It is unlikely that fish or
invertebrates would be exposed to deeper sediments.
The exposure pathway for PAHs could be interrupted if there was no sign of internal
exposure. Aquatic contaminants such as PAHs have been monitored by measuring the
metabolites of xenobiotics in fish bile (Roubal et al. 1977, Gmur and Varanasi 1982,
Varanasi et al. 1983). Samples from white suckers (Catostomus commersoni) taken in
1992 from the Little Scioto River were analyzed for concentrations of benzo[a]pyrene
(BAP) and naphthalene (NAPH)-type metabolites. Results of the analysis of PAH bile
metabolites in white suckers from the Little Scioto River are shown in Figure 7-9.
Biomarkers of NAPH and BAP are elevated from RM 6.5 to the mouth of the river,
providing evidence that the exposure pathway is complete at these locations. Exposure
criteria, concentrations considered to be above background, were exceeded at RMs 6.5
through 0.4. PAHs are also known to cause induction of detoxifying enzymes such as
EROD. EROD activity was elevated at RM 6.5 - 0.4. Based on the absence or presence
of bile metabolites, the exposure pathway for PAHs is incomplete at Impairment A, and
complete at Impairments B and C.
Metals (#3) - Metals must be taken into organisms to cause adverse effects. Data from
fish tissue sampled in 1992 confirm uptake of lead and zinc. For common carp
(Cyprinus carpio) at RM 9.2, zinc concentrations were 79.6 mg/kg, at RM 6.5, zinc
concentrations were 68.3 mg/kg. For white suckers at RM 6.5, zinc concentrations were
17.8 mg/kg, and lead concentrations were 81.4 mg/kg. At RM 2.7, fish tissues levels
were 15.8 mg/kg for zinc and 0.34 mg/kg for lead. For the other metals, we have
conservatively assumed that external exposure will represent internal exposure for fish.
Making this assumption, increased exposure to at least one of the metals occurs at all
sampling locations in the reach RM 7.9 to 0.4. Concentrations of metals in sediment
were from samples taken in 1992 from the top six inches of sediment. For 1987 and
1998 data, the depth of samples is unknown.
7-24 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
7.9
6.5
0.5yMg/mg BAP
and 80/ig/mg NAPH
5.7 4.4
River Mile
I I BAP
NAPH
Figure. 7-9. Bile metabolites (|jg/mg protein) measured in white suckers
from the Little Scioto River in 1992. (Median levels of PAH metabolites
below RM 7.9 were up as much as 4 times the Exposure Criteria, (dashed
horizontal line) which are upper limits of background for the state of Ohio.
The numbers above the bars equal number offish sampled. Vertical lines
are standard errors.)
Ammonia (#4) - There are several interweaving pathways by which ammonia can be
produced in the river and cause effects. We have evidence for two of these steps: total
ammonia, and nitrate and nitrite concentrations that are converted to ammonia when DO
is low. Toxic unionized ammonia is formed at high pH. Hard water streams of the
Eastern Corn Belt Plains typically have pH from 7.5-8; pH may rise even above 9.0 in
the summer during maximum photosynthesis in nutrient-enriched waters. Data on pH
are not available in 1992, however, in 1998 grab samples, pH ranged between 7.4 to 8.0.
The Little Scioto is highly enriched, and it is highly likely that there are periods when pH
is greater than indicated by grab samples. Thus, we assume that the exposure pathway is
complete in the Little Scioto when total ammonia is present. This occurs from RM 11.1
to 0.4. Because ammonia concentrations are measured in the water column, both fish
and macroinvertebrates are exposed.
Low Dissolved Oxygen/High Biological Oxygen Demand (#5) - Dissolved oxygen can be
depleted by high BOD due to the bacterial respiration associated with allochthonous
organic matter or decaying algal mats. We have measurements of several relevant
parameters: NOX, total P, BOD and DO concentrations. This exposure pathway is
considered complete under two scenarios: (1) BOD is elevated and DO is reduced
compared with the most upstream location, or (2) if BOD data is unavailable, NOX and P
are elevated and assumed to cause algal growth, and DO is reduced as compared with the
most upstream location. At RM 7.9, DO is reduced, but BOD is unchanged, so that the
exposure pathway is considered incomplete. RM 6.5 is more difficult to evaluate
because data are scanty and are used from different years. In 1987, DO data were low at
Chapter 7: Little Scioto River, Ohio
7-25
-------
Stressor Identification Guidance Document
RM 6.5, and in 1992, the BOD was slightly elevated; thus, the pathway is complete. At
RM 5.8 to 0.4, because BOD is elevated but DO is similar or greater than at 7.9, the
exposure pathway is considered incomplete.
Nutrient Enrichment (#6) -We have evidence for the presence of elevated levels of both
NOX and total P concentrations. This exposure pathway appears to be complete at RM
5.8 to 0.4, and at RM 7.9.
7.5.4 Summary of Analyses for Elimination
The results of the analysis of spatial associations are summarized in Table 7-5 (pages 7-
21 to 7-23). The table addresses four questions for each combination of impairment and
candidate cause. If any of the answers are no, then the candidate cause can be
eliminated:
* The first question is whether a candidate cause and impairment are spatially co-
located. Regardless of concentration, the answer is yes if the stressor is present.
If the stressor is not present, the answer is no and the impairment could not have
been caused by exposure to that stressor.
* The second question asks whether the exposure is elevated compared to the
closest upstream location where the impairment does not occur. The candidate
cause could have been responsible for the impairment only if exposure increased.
The candidate cause can be eliminated if the answer to the second question is no.
* The third question asks whether there is a decrease in exposure that corresponds
with recovery of the impairment. As discussed above, this question is relevant
only to Impairment C. If the answer is no with results clearly showing a
lessening of impairment with consistent exposure, then the candidate cause can
be eliminated.
* The last question asks if the exposure pathway is complete. If it is interrupted or
clearly incomplete so that exposure could not have taken place, then it can be
eliminated as a potential cause.
7.6 Characterize Causes: Eliminate
Potential causes may be eliminated if the
evidence indicates that they do not co-occur
with effects, if effects decrease with increasing
influence of the cause, or if the exposure
pathway is incomplete. Each of the three
Impairments (A, B, and C) are discussed below
in relation to the elimination of specific causes.
Conclusions about which candidate causes
remain for each impairment are also listed.
CHARACTERIZE CAUSES
| Eliminate
| Diagnose
Strength of Ev
Identify Probable Cause
dence
Impairment A: RM 7.9
Habitat alteration and metal contamination are the only candidate causes known to
co-occur at RM 7.9 and to increase compared to upstream locations. All metals were
slightly greater at RM 7.9 compared to RM 9.2. PAHs and ammonia were not
elevated at RM 7.9 relative to the upstream reference, thus candidate causes #2 and
7-26
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
#4 are eliminated. DO concentrations were about 30% lower than upstream, but
BOD concentrations were not different from the upstream reference location (RM
9.2), thus candidate cause #5 is eliminated. NOX increased from 1.2 mg/L to 1.4
mg/L. The shift is small, but precludes elimination of candidate cause #6.
Conclusion: Habitat Alteration (#1), Metal Contamination (#3), and Nutrient
Enrichment (#6) remain.
Impairment B: RM 6.5.
At this site, only candidate cause #1 can be eliminated because the degree of habitat
alteration is not elevated compared with those at RM 7.9. The decline in QHEI score
is associated with the obvious presence of organic chemical contamination rather
than physical stream characteristics. Organic chemicals, including benzo[a]pyrene
and naphthalene, were present and were elevated above concentrations at RM 7.9.
Exposure to these organic chemicals was demonstrated by internal concentrations of
metabolites. The metals chromium, copper, lead, and mercury were elevated
compared to upstream concentrations in all years for which there is data, including
1988, 1991, 1992 and 1998. Dissolved oxygen levels were among the lowest in the
river in 1987, and BOD levels were slightly greater than upstream locations.
Ammonia concentrations were also slightly greater, and total P concentrations were
0.02 mg/L greater.
Conclusion: PAH Contamination (#2) Metal Contamination (#3), Ammonia
Toxicity (#4), Low Dissolved Oxygen/High Biological Oxygen Demand (#5), and
Nutrient Enrichment (#6) remain.
Impairment C: RM 5.7.
In this reach of the river, the degree of habitat alteration and PAH levels were similar
or lower than at RM 6.5, thus candidates #1 and #2 are eliminated. Candidate cause
#5, low DO/high BOD, can be eliminated, even though BOD, P and NOX are elevated
because the subsequent event in the pathway, decreased DO, did not occur. DO is
unchanged from RM 7.9 in 1992, and RM 6.5 in 1987. The metals (copper and zinc)
increased slightly, and the copper gradient was significantly correlated with %
Tanytarsini midges and % DELTA, thus candidate cause #3 remains. NOX and P
were elevated in the reach compared to upstream locations and were significantly
correlated with % Tanytarsini midges. Candidate cause #6 remains. Candidate
cause #4 could be eliminated since ammonia was not correlated with the specific
impairments. However, the increase in ammonia was 10 times greater than
upstream, and because the data available for correlations were very limited, a
conservative decision could be made to retain this cause for further evaluation by the
strength of evidence approach.
Conclusion: Metal Contamination (#3), Ammonia (#4), and Nutrient
Enrichment (#6) remain.
A summary of the candidate causes that remain after the elimination process are listed in
Table 7-6. Only those causes remaining need to be evaluated by diagnostic or strength of
evidence analyses.
Chapter 7: Little Scioto River, Ohio 7-27
-------
Stressor Identification Guidance Document
Table 7-6. Candidate causes remaining after elimination.
#1 Habitat alteration
#2 PAH Contamination
#3 Metal Contamination
#4 Ammonia
#5 Low DO/BOD
#6 Nutrient Enrichment
Impairment A
X
X
X
Impairment B
X
X
X
X
X
Impairment C
X
X
X
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
7.7 Analyze Evidence for Diagnosis
Diagnosis is the identification of causes based on
characteristic signs or symptoms (see 4.2.2). No
evidence strong enough to support diagnosis was
available for any of the candidate stressors.
However, the pattern of community change is
considered to be suggestive, and is used in the
strength of evidence analysis below.
The deformities, fin erosion, rumors, physical
lesions and anomalies on fish that constitute the
DELTA are pathologies that are also potentially
subject to diagnosis. Some DELTA are strongly
associated with known toxic substances and others
with increased nutrients (Yoder and Rankin 1995b).
However, no pathologist has examined the fish in
question. DELTA cannot be used to distinguish among toxic substances unless specific
anomalies are identified, and even these may be too non-specific to diagnose without
additional information (e.g., histopathology).
7.8 Analyze Evidence to Compare Strength of Evidence
CHARACTERIZE CAUSES
Eliminate
1
Dagnose
| Strength of Evidence |
| Identify Probable Cause
J
All of the remaining candidate causes are subjected
to a strength of evidence analysis to verify the
elimination step and to identify the most likely
cause from the multiple hypothesized causes that
remained after the elimination process. The
strength of evidence analysis examined case
specific evidence as well as evidence from other
situation and biological knowledge.
Case Specific Evidence
The evidence presented earlier for the elimination
step is useful here as well. In addition, some data
on loadings are available from the Waste Water
Treatment Plant (WWTP), which discharges at KM
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
CHARACTERIZE CAUSES
Eliminate
Diagnose
Strength of Evidence
7-28
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
6.2 and also has combined sewer overflows that discharge during wet weather periods.
No clear trends were evident in the loadings of total non-filterable residue or biological
oxygen demand between 1977 and 1992. Ammonia values were generally low, with fifty
percent of loading below 10 kg/day between 1977 and 1991. The highest ammonia
loading occurred during 1992, with a median of 12.6 kg/day and a maximum of 130
kg/day (OEPA 1994).
Evidence from Other Situations or Biological Knowledge
This section presents evidence that uses information from other studies that are related to
either exposures or effects found in segments of the Little Scioto River. In particular,
associations are made between the exposures known at the site and reports of effects
caused by similar exposures. This section also uses levels of effects seen at the site and
effects seen at other sites where the same candidate cause occurred. It also considers
special experimental evidence; that is, reports about places with similar stressors and
effects that improved when the stressor was removed, and laboratory studies of candidate
cause-effect relationships.
Exposure-response data are available for PAHs and metals, although not for the
community parameters of greatest interest for this study. Sediment effect concentrations
(SECs) developed for Hyalella azteca and Chironomus riparius were considered, but
only Hyalella azteca was used since Chironomus riparius values were always less
sensitive. Sediment effect concentrations for Hyalella azteca are expressed as threshold
effect level (TEL) and probable effect level (PEL) (Table 7-19; see Section 7.13)
(USEPA 1996b).
The TEL and PEL are sediment concentrations associated with toxicity in laboratory
tests. The interpretation is that toxicity rarely occurs below the TEL and frequently
occurs above the PEL (USEPA 1996b). Values were derived from a data set consisting
of many similar studies, and they consider both effect and no-effect data for field-
contaminated sediments. The TEL and PEL values used in this study are listed in Tables
7-19 and 7-20 (see Section 7.13). Since many metals and PAHs were present at sites,
partial toxicity contributed by individual chemicals were calculated and summed to
estimate the overall toxicity of metals and PAH at each site. TELs and PELs are used
with caution because they are based on sediments with multiple contaminants.
The TEL and PEL values were compared with the concentrations seen at the locations of
impairment in Table 7-19. As shown in Table 7-19, the most striking result is that no
PAH exceeded any criterion level at Impairment A for 1992. For metals only, the TEL
for arsenic was exceeded at Impairment A in 1992. At Impairment B and C, the Hyalella
azteca PEL and TEL were exceeded for all PAH that were measured and in every year
except 1992, when there were more samples below the detection limit. Hyalella azteca
TEL values were exceeded for most metals, but only a few PEL values were exceeded,
including those for lead, copper, and chromium.
For PAHs, the cumulative toxic units were exceeded at Impairments B and C in every
year (Table 7-7). Exceedances ranged from 339 to 18,820 times the value that would
probably kill Hyalella azteca. For metals, the cumulative toxic units were also exceeded
at Impairments B and C in every year. However, exceedances were never more than six
times the cumulative probable effect level.
Chapter 7: Little Scioto River, Ohio 7-29
-------
Stressor Identification Guidance Document
Table 7-7. Cumulative toxic units for PAHs and metals based on the PEL values.
(Values greater than 1.0 exceed PEL*).
Chemical
PAH
Metals
Cumulative Toxic Units
Nearest
Upstream
Location
/0\
(0)
[1.2]*
/0.4\
(0.6)
[0.9]
Impairment A
/0\
(0)
[2.5]*
/0.7\
(1.1)*
[0.9]
Impairment B
/604.5V
(339.4)*
[18819.9]*
/4.3\*
(5.1)*
[1.6]*
Impairment C
79697. 8\*
(821)*
[1633.4]*
/1 .5\*
(2.8)*
[5.8]*
* Exceeds PEL and TEL./\= 1987-1991, () = 1992, [] =1998. Zero = below
detection.
Criteria are also available for ammonia (USEPA 1998b) (Table 7-8). The toxicity of
total ammonia (which includes NH3 and NH4+) varies with pH. Dehydration of
ammonium ion (NH4+) to un-ionized ammonia is controlled by ambient pH, such that
excess hydroxide ions (high pH) increase the concentration of the more toxic, un-ionized
form. Hard water streams of the Eastern Corn Belt Plains (ECBP) typically have pH
from 7.5-8; in the summer, during maximum photosynthesis in nutrient enriched waters,
pH may rise above 9.0. In 1998, pH values ranged between 7.4 and 8.4, and appeared to
be independent of location. Total ammonia concentrations at RM 5.8 through 2.7 would
have exceeded the ammonia criterion for water having a pH 8.0 to 8.5 in 1992 (Table 7-
8). In 1998, the criterion would have been exceeded at pH 8.5.
Ohio's criteria for dissolved oxygen (causal candidate #5) are 4.0 mg/1 for warm water,
and 3.0 mg/1 for modified warm water. In 1992, no locations had dissolved oxygen
below the modified warm water criterion, and only RM 2.7 had dissolved oxygen
concentrations below the warm water criterion, based on a single measurement.
However, in 1987, continuous data were collected by Datasonde (in-stream Hydrolab)
and violations were detected at Impairments A and B (Table 7-8).
Ohio's proposed state-wide criterion for modified warm-water habitat for nitrate and
nitrite is 1.6 mg/L for wadeable streams in the ECBP having a drainage greater than 20
mi2 and less than 200 mi2. For total phosphorus, the proposed state-wide criterion for
modified warm-water habitat is 0.28 mg/L (Rankin et al. 1999). These are exceeded at
RM 5.8 (Table 7-8).
A state-wide study by Yoder and Rankin (1995b) indirectly examined the plausibility of
specific community changes associated with nine types of sources, including waste water
treatment plants, industrial point sources, conventional municipal sources, combined
sewer overflows, channelization, and agricultural non-point sources. They found that
deformities, erosions, lesions, tumors and anomalies (DELTA) in fish were associated
with industrial discharges (Yoder and Rankin 1995b) and nutrient enrichment (Rankin et
al. 1999). In the Little Scioto, the greatest % DELTA values are associated with the
greatest nutrient concentrations. Among macroinvertebrates, the loss of Tanytarsini
midges and the increase ofCricotopus sp. are both associated with industrial discharges
(Yoder and Rankin 1995b). In the Little Scioto, the disappearance of Tanytarsini midges
and an increase in Cricotopus are associated with Impairment C.
7-30
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-8. Comparison of the reported concentration of water quality parameters
(mg/L) with exceedances.
Sediment
Parameter
Criteria mg/L
Ammonia3
0.57 mg/L at pH 8.5
1. 27 mg/L at pH 8.0
Dissolved Oxygenb
3.0 mg/L for MWH
Nitrate-nitrite0
1 .6 mg/L
Total phosphorusd
0.28 mg/L
RM7.9
[RM 7.1]
(<0.05)
[0.11,<0.05]
{4.6-2.8}*
(7.9, 5.7)
(1.4)
[0.73, <0.1]
(0.07)
[0.36*, 0.13]
RM6.5
(0.1)
{7.2-1 .9}*
(NA)
(0.8)
(0.09)
RM5.8
[RM 6.2]
(1.2)
[0.35, 0.69]
{8.3, 4.2}
(8.23,4.21)
(8.1)*
[0.33, 2.37]*
{1 .65}*
(2.17)*
[1.9, 1.21]*
No Entry = No data for that year. {}=1987, ( )=1992, [] =1998.
a USEPA (1998b) recommended ammonia criterion
bOEPA (1994) dissolved oxygen criterion
c Rankin et al. (1999) proposed nitrate-nitrite criterion
d Rankin et al. (1999) proposed total phosphorus criterion
* Exceedance of criterion
Dissolved oxygen values are maximum and minimum. Ammonia, nitrate-nitrite, total
phosphorus measured in August and October, 1998.
7.9 Characterize Causes: Strength of Evidence
Strength of evidence analysis uses all of
the evidence generated in the analysis
phase to examine the credibility of each
remaining candidate cause. The causal
considerations for the strength of evidence
analyses used three types of evidence:
case-specific evidence, evidence from
other situations or biological knowledge,
and evidence based on multiple lines of
evidence (Section 4.3.3). All the evidence
was evaluated for consistency or
coherence with the hypothesized causes.
The results of the strength of evidence analysis are presented in Tables 7-9 to 7-11.
Following the strength of evidence analysis, the candidate causes are characterized
(Table 7-12). This involves describing the causal evidence and identifying the probable
cause.
CHARACTERIZE CAUSES
Eliminat
e Diagnose
| Strength of Evidence |
Identify Probable Cause
Chapter 7: Little Scioto River, Ohio
7-31
-------
Stressor Identification Guidance Document
Table 7-9. Strength of evidence analysis for the three candidate causes of Impairment A, RM 7.9.
Causal
Considera-
tion
Evidence
Score
Evidence
Score
Evidence
Score
Case-Specific Considerations
Co-
occurrence
Temporality
Consistency
of
Association
Biological
Gradient
Complete
Exposure
Pathway
Experiment
Habitat Alteration
Compatible: At and
below RM 7.9, the
habitat of the Little
Scioto is altered as a
result of channelization.
The degree of habitat
alteration remains about
the same to the mouth
of the river. The
upstream reference is
not channelized and
habitat is good.
No evidence
No evidence
Not applicable: Other
downstream candidate
causes interfere with
this consideration.
Evidence for all steps:
The fish and
invertebrates inhabit the
channelized reach
where the habitat is
altered.
Channel was deepened.
DO was depressed.
Substrate was
embedded.
No evidence.
+
NE
NE
NA
++
NE
Metals Contamination
Compatible: All
sediment metal
concentrations were
slightly higher at RM 7.9
compared to upstream.
No evidence
No evidence
Not applicable: Other
downstream candidate
causes interfere with
this consideration.
Incomplete evidence:
No internal
concentrations of
metals were measured.
Metals were present in
sediment and exposure
could occur from
ingestion or by
respiration of epibenthic
water or sediment
particles or through the
food chain.
No evidence.
+
NE
NE
NA
+
NE
Nutrient Enrichment
Compatible: N was
elevated by 0.2mg/L in
1992 compared to
upstream.
P is the same or
decreases compared to
upstream.
No evidence
No evidence
Not applicable: Other
downstream candidate
causes interfere with
this consideration.
Incomplete evidence:
Fish and invertebrates
inhabit stream where
nutrients are elevated.
Concentrations of algae
or chlorophyll a were not
measured.
No evidence.
+
NE
NE
NA
+
NE
NE = no evidence; NA = not applicable/not available
7-32
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-9 (continued). Strength of evidence analysis for the three candidate causes of Impairment
A, RM7.9.
Causal
Considera-
tion
Evidence
Score
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge
Plausibility:
Mechanism
Plausibility:
Stressor-
Response
Habitat Alteration
Increased Relative
Weight: Plausible:
Artificially deepened
channel allows larger
sized fish to survive.
Increased DELTA: Not
known: No obvious
mechanism other than
stress.
Loss of species:
Plausible: Embedded
sediments remove
forage, reproductive,
and cover habitats for
benthic fish including
darters and benthic
invertebrates including
mayflies. Low DO is not
tolerated by many
species (Karr and
Schlosser 1977, Yount
and Niemi 1990, Rankin
1995).
Increased Relative
Weight: No evidence.
Increased DELTA: No
evidence.
Loss of species: No
evidence.
No quantitative
evidence.
Habitat alteration
associated with
channelization is
generally believed to be
an all or none situation
affected by it's spatial
extent and severity.
+
0
+
NE
NE
NE
Metals Contamination
Increased Relative
Weight: Implausible: No
known mechanism for
metals. Metals usually
cause a decrease in the
relative weight offish
(Eisler 2000b).
Increased DELTA:
Implausible: Metals do
not cause fin erosion
and lesions (Eisler
2000b).
Loss of species:
Plausible: Metals are
known to cause lethal
and sub-lethal effects to
invertebrates and fish
that can extirpate
species from a site
(Eisler 2000b).
Metals usually cause a
decrease in the relative
weight offish (Eisler
2000b).
Increased Relative
Weight: Not applicable:
Implausible mechanism.
Increased DELTA: Not
applicable: implausible
mechanism.
Loss of species:
Inconcordant.
No metals exceeded
Hyalella azteca PEL
values in 1 987, 1 992 or
1998. The TEL value
for arsenic was
exceeded only in 1 992.
Metals cumulative toxic
units exceeded PEL in
1992, but only by 0.1
units (USEPA1996b).
+
NA
NA
Nutrient Enrichment
Increased Relative
Weight: Implausible: N
is a nutrient for algal
growth. Greater
production of algae
could provide additional
food, increasing fish
growth. However, the
mechanism is
implausible because N
is generally not limiting
(Allan 1995).
Increased DELTA:
Plausible: Nutrients are
believed to create
conditions that favor
opportunistic pathogens
and fungi that cause
lesions, fin erosion and
interfere with wound
healing.
Loss of species:
Plausible: Switching to
an autochthonous
energy source could
alter species survival
and community
composition offish and
invertebrates.
Increased Relative
Weight: Not applicable:
Implausible mechanism.
Increased DELTA:
Inconcordant:
magnitude of nutrient
change too small to
cause effect.
Loss of species:
Inconcordant.
The magnitude of
nutrient change was too
small to account for the
dramatic shifts in
invertebrate and fish
metrics. Proposed
nitrogen criterion for
Ohio was not exceeded
(Rankin et al. 1999).
+
+
NA
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-33
-------
Stressor Identification Guidance Document
Table 7-9 (continued). Strength of evidence analysis for the three candidate causes of Impairment
A, RM7.9.
Causal
Considera-
tion
Evidence
Score
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge (cont'd)
Consistency
of
Association
Habitat Alteration
Increased Relative
Weight: In most places.
Increased DELTA: In
most places.
Loss of species: In most
places.
Moderate increase in
DELTA and loss of
species are commonly
associated with habitat
alteration associated
with channelization
(Yoder and Rankin
1995b). Increased
Relative Weight is also
commonly increased
with deepened channels
(Personal Observation).
Agricultural areas with
channelization having
similar stressors
showed decreases in
IBI and ICI component
metrics (Edwards et al.
1984, Sheildsetal.
1998).
++
++
++
Metals Contamination
Increased Relative
Weight: Many
exceptions.
Increased DELTA:
Many exceptions.
Loss of species: Many
exceptions.
At other sites in Ohio
with similar metals
concentrations, Relative
Weight and DELTA
were not increased and
species were abundant.
Personal observation of
Ohio database.
-
.
-
Nutrient Enrichment
Increased Relative
Weight: No evidence.
Increased DELTA:
Many exceptions. At
many sites in Ohio,
DELTA was not
increased by these
levels of N (Rankin et al.
1999).
Loss of species: Many
exceptions. At many
sites in Ohio, IBI and ICI
scores were high at
these levels of N
(Rankin et al. 1999).
High IBI and ICI cannot
be achieved when many
species are lost.
NE
-
-
NE = no evidence; NA = not applicable/not available
7-34
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-9 (continued). Strength of evidence analysis for the three candidate causes of Impairment
A, RM7.9.
Causal
Considera-
tion
Evidence
Score
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge (cont'd)
Specificity of
Cause
Analogy
Experiment
Predictive
Performance
Habitat Alteration
Increased Relative
Weight: One of a few:
Deep channels or pools
required for larger fish.
Relative weight offish is
significantly correlated
with drainage area, a
surrogate for channel
depth (Norton 1999).
Increased DELTA: One
of many.
Loss of species: One of
many.
Not applicable
Increased Relative
Weight: No evidence
Increased DELTA: No
evidence
Loss of species:
Concordant: Artificial
riffle and pools
improved invertebrate
assemblage in the
channelized Olentangy
River (Edwards et al.
1984), and fish in
Mississippi River
(Sheildsetal. 1998).
No evidence
++
0
0
NA
NE
NE
+++
NE
Metals Contamination
Increased Relative
Weight: Not applicable:
Implausible mechanism.
Increased DELTA: Not
applicable: Implausible
mechanism.
Loss of species: One of
many.
Not applicable
Increased Relative
Weight: No evidence
Increased DELTA: No
evidence
No evidence
NA
NA
0
NA
NE
NE
NE
Nutrient Enrichment
Increased Relative
Weight: Not applicable:
Implausible mechanism.
Increased DELTA: One
of many.
Loss of species: One of
many.
Not applicable
Increased Relative
Weight: No evidence
Increased DELTA: No
evidence
No evidence
NA
0
0
NA
NE
NE
NE
Considerations from Multiple Lines of Evidence
Consistency
of Evidence
Coherence
of Evidence
Habitat Alteration
Increased Relative
Weight: All consistent.
Increased DELTA: All
consistent.
Loss of species: All
consistent.
Increased Relative
Weight, Increased
DELTA, Loss of
species: None.
+++
+++
+++
0
Metals Contamination
Increased Relative
Weight: Inconsistent:
Implausible mechanism.
Increased DELTA:
Inconsistent:
Implausible mechanism.
Loss of species:
Inconsistent - Although
metals are present, the
concentrations are
unlikely to cause
species extirpation.
Increased Relative
Weight, Increased
DELTA, Loss of
species: None.
0
Nutrient Enrichment
Increased Relative
Weight: Inconsistent:
Magnitude of change
inconsistent with
magnitude of effect.
Increased DELTA:
Inconsistent: Magnitude
of change inconsistent
with magnitude of effect.
Loss of species:
Inconsistent: Magnitude
of change inconsistent
with magnitude of effect.
Increased Relative
Weight, Increased
DELTA, Loss of
species: None.
0
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-35
-------
Stressor Identification Guidance Document
Table 7-10. Strength of evidence analysis for the five candidate causes of Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Case-Specific Considerations
Co-occurrence
Temporality
Consistency of
Association
Biological Gradient
Complete Exposure
Pathway
Experiment
PAH contamination
Compatible: Sediment PAH
concentrations were several orders
of magnitude greater at RM 6.5 than
upstream (Table 13).
No evidence
No evidence: only one location.
Not Applicable: Other candidate
causes downstream interfere with
this consideration.
Actual evidence for all steps: PAHs
were present in the sediment, and
bottom-feeding fish and benthic
invertebrates are typically exposed
to sediment contaminants. Both BAP
and NAPH metabolites were found in
fish. EROD, a detoxifying enzyme
known to be induced by PAH, was
elevated.
No evidence
+
NE
NE
NA
+ +
NE
Metals Contamination
Compatible: Lead, chromium, copper
and mercury concentrations in
sediment were two to ten times
greater at RM 6.5 than upstream.
Cadmium and zinc were also greater,
but to a lesser degree.
No evidence
No evidence: only one location.
Not Applicable: Other candidate
causes downstream interfere with this
consideration.
Actual evidence for all steps: Metals
were present in sediment and
exposure could occur from ingestion
or by respiration of epibenthic water of
sediment particles or through the food
chain. Zinc and lead were detected
in fish tissues.
No evidence
+
NE
NE
NA
+ +
NE
NE = no evidence; NA = not applicable/not available
7-36
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-10 (continued). Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge
Plausibility:
Mechanism
Plausibility:
Stressor-Response
Consistency of
Association
Specificity of Cause
PAH contamination
Decreased relative weight:
Plausible: PAHs are known to
reduce growth. Toxic compounds
can shorten life span resulting in
smaller fish (Eisler 2000a).
Increased DELTA: Plausible: PAHs
are known to cause eroded barbels,
fin erosion, lesions and internal and
external tumors (Eisler 2000a).
Decreased species: Plausible: PAHs
are known to be toxic and cause
reproductive impairments which
could extirpate species (Eisler
2000a).
Decreased relative weight:
Concordant: Toxic levels are
consistent with decreased fish
growth.
Increased DELTA: Quantitatively
consistent: PAHs are at levels that
cause tumors and other DELTA.
Decreased species: Quantitatively
consistent: The Hyalella azteca
PEL's were exceeded for all PAHs.
The cumulative PAH toxic units
ranged between 339 to 18,820 times
the PEL value (USEPA 1996b).
Decreased relative weight: In most
places: Decreased relative weight is
associated with complex toxic
exposures (Yoder and Rankin
1995b).
Increased DELTA: Invariant: Tumors
and other DELTA are associated
with fish exposed to high
concentrations of PAH in fresh and
marine waters (Albers 1995).
Decreased species: Invariant: At
more than 25 locations associated
with PAH contamination that
exceeded exposure criteria in Ohio,
IBI and ICI scores were below 30
(Cormier et al. 2000a). IBI and ICI
are known to be depressed even
when habitat quality is high (Cormier
et al. 2000b, OEPA 1992a). IBI and
ICI scores of less than 30 only occur
when some species are extirpated.
Decreased relative weight: One of
many.
Increased DELTA: One of many.
PAHs are known to cause external
lesions seen at Impairment B.
Decreased species: One of many.
+
+
+
+
+++
+++
++
+++
+++
0
0
0
Metals Contamination
Decreased relative weight: Plausible:
Metals are known to reduce growth.
Toxic compounds can shorten life
span resulting in smaller fish (Eisler
2000b).
Increased DELTA: Implausible:
Metals do not cause fin erosion and
lesions
Decreased species: Plausible: Metals
are known to cause lethal and sub-
lethal effects to invertebrates and fish
that can extirpate species from a site
(Eisler 2000b).
Decreased relative weight:
Ambiguous. Toxic levels are
consistent with decreased fish growth
(Eisler 2000b).
Increased DELTA: Not applicable:
mechanism is implausible.
Decreased species: Quantitatively
consistent. Lead exceeded Hyalella
azteca PEL values in 1 988-1 991 and
1992 and chromium in 1992. The
cumulative toxic units values for all
metals range from 1.6 to 5.1 (USEPA
1996b).
Decreased relative weight: In most
places: Decreased relative weight is
associated with complex toxic
exposures (Yoder and Rankin 1995).
Increased DELTA: Not applicable.
Decreased species: In most places:
Hickey and Clements (1998) reviewed
changes in invertebrate community
associated with metals in water
column.
Decreased relative weight: One of
many.
Increased DELTA: Not applicable.
Decreased species: One of many.
+
+
0
NA
+++
++
NA
++
0
NA
0
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-37
-------
Stressor Identification Guidance Document
Table 7-10 (continued). Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge (cont'd)
Analogy
Experiment
Predictive
Performance
PAH contamination
Not applicable
Decreased relative weight:
Concordant: Following dredging in
the Black River, Ohio, the age
structure of the brown bullheads
increased (Baumann and
Harshbarger 1995).
Increased DELTA: Concordant: In
the Black River Ohio, removal of
PAHs by dredging resulted in lower
levels of DELTA (Baumann and
Harshbarger 1995) and PAH bile
metabolites (Lin et al. submitted).
Decreased species: Concordant:
Following dredging the composition
of species at this site also changed
(Baumann, pers. comm.).
No evidence
NA
+++
+++
+++
NE
Metals Contamination
Not applicable
No evidence: No references sought.
No evidence: No references sought.
No evidence: No references sought.
No evidence
NA
NE
NE
Considerations from Multiple Lines of Evidence
Consistency of
Evidence
Coherence of
Evidence
PAH contamination
Decreased relative weight: All
consistent.
Increased DELTA: All consistent.
Decreased species: All consistent.
+++
+++
+++
Metals Contamination
Decreased relative weight: All
consistent.
Increased DELTA: Multiple
inconsistencies.
Decreased species: All consistent.
Increased DELTA: No known
explanation.
+++
+++
0
NE = no evidence; NA = not applicable/not available
7-38
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-10 (continued). Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Case-Specific Considerations
Co-occurrence
Temporality
Consistency of
Association
Biological
Gradient
Complete
Exposure
Pathway
Experiment
Ammonia Toxicity
Compatible:
Ammonia
concentration was
doubled relative to
Impairment A.
No evidence
No evidence: Only
one location.
Not applicable:
Other downstream
candidate causes
interfere with this
consideration.
Evidence for all
steps: Fish and
invertebrates
inhabited stream
where ammonia
was present.
No evidence
+
NE
NE
NA
+ +
NE
Low Dissolved oxygen/High
BOD
Compatible: In
1992, BOD was
double the
upstream value
and the lowest DO
levels measured
were 0.9 mg/L less
than upstream.
No evidence
No evidence: Only
one location.
Not applicable:
Other downstream
candidate causes
interfere with this
consideration.
Evidence for all
steps: Fish and
invertebrates
inhabited stream
where conditions
of low DO and high
BOD occurred.
No evidence
+
NE
NE
NA
+ +
NE
Nutrient Enrichment
Compatible:
Compared to RM
7.9, P was
elevated by 0.02
mg/L. N was less.
No evidence
No evidence: Only
one location.
Not applicable:
Other downstream
candidate causes
interfere with this
consideration.
Evidence for all
steps: Fish and
invertebrates
inhabit stream
where P was
elevated.
No evidence
+
NE
NE
NA
+ +
NE
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-39
-------
Stressor Identification Guidance Document
Table 7-10 (continued). Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge
Plausibility:
Mechanism
Plausibility:
Stressor-
Response
Ammonia Toxicity
Decreased relative
weight: Plausible:
Ammonia toxicity
could reduce
growth and
survival. Low
survival could alter
the age structure
resulting in
smaller, younger
fish.
Increased DELTA:
Plausible:
Ammonia has
been associated
with anomalies
(Dyer, pers.
comm.).
Decreased
species: Plausible:
Ammonia is known
to be toxic to fish
and invertebrates
(USEPA1998b).
Decreased relative
weight: No
evidence.
Increased DELTA:
No evidence.
Decreased
species:
Inconcordant: The
ammonia
concentrations
were not great
enough to cause
the dramatic
effects seen at
Impairment B.
Ammonia criteria
were not
exceeded.
(USEPA1998b).
+
+
+
NE
NE
Low Dissolved oxygen/High
BOD
Decreased relative
weight: Plausible:
Stress could
reduce growth and
survival. Low
survival could alter
the age structure
resulting in more
smaller, younger
fish.
Increased DELTA:
Not known: No
known mechanism.
Decreased
species: Plausible:
Low DO can kill
fish and
invertebrates
(Allan 1995).
Decreased relative
weight: No
evidence.
Increased DELTA:
Not applicable.
Decreased
species: DO levels
are below Ohio
criteria for MWH
(OEPA1992b).
+
0
+
NE
NA
+
Nutrient Enrichment
Decreased relative
weight:
Implausible:
Increased nutrients
are usually
associated with
increased algal
growth that
augment the
energy available
for growth.
Increased DELTA:
Plausible: Nutrients
are believed to
create conditions
that favor
opportunistic
pathogens and
fungi that cause
lesions, fin erosion,
and interfere with
wound healing
(Rankin et al.
1999).
Loss of species:
Plausible:
Switching to an
autochthonous
energy source
could alter species
survival and
community
com position for
fish and
invertebrates (Allan
1995).
Decreased relative
weight:
Inconcordant.
Increased DELTA:
Inconcordant.
Decreased
species:
Inconcordant: The
magnitude of P
change was not
great enough to
cause dramatic
effects seen at
Impairment B.
Proposed P
criterion was not
exceeded (Rankin
etal. 1999).
+
+
-
NE = no evidence; NA = not applicable/not available
7-40
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-10 (continued). Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge (cont'd)
Consistency of
Association
Specificity of
Cause
Analogy
Experiment
Predictive
Performance
Ammonia Toxicity
No evidence
Decreased relative
weight: One of
many.
Increased DELTA:
One of many.
Decreased
species: One of
many.
Not applicable
No evidence: No
reference sought.
No evidence
NE
0
0
0
NA
NE
NE
Low Dissolved oxygen/High
BOD
No evidence.
Decreased relative
weight: One of
many
Increased DELTA:
Not applicable.
Decreased
species: One of
many.
Not applicable
No evidence: No
reference sought.
No evidence
NE
0
NA
0
NA
NE
NE
Nutrient Enrichment
Decreased relative
weight: Many
exceptions.
Increased DELTA:
Many exceptions:
DELTA are
associated with
increased P at
many sites in Ohio,
but at a higher
concentration of P
(Rankin et al.
1999).
Decreased
species: Many
exceptions:
Reduced species
are associated with
many sites in Ohio
increased P, but at
a higher
concentration
(Rankin et al.
1999).
Decreased relative
weight: Not
applicable
Increased DELTA:
One of many.
Decreased
species: One of
many.
Not applicable
No evidence: No
references sought.
No evidence
-
-
NA
0
0
NA
NE
NE
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-41
-------
Stressor Identification Guidance Document
Table 7-10 (continued). Strength of evidence analysis for the five candidate causes of
Impairment B, RM 6.5.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Considerations from Multiple Lines of Evidence
Consistency of
Evidence
Coherence of
Evidence
Ammonia Toxicity
Decreased relative
weight: All
consistent.
Increased DELTA:
All consistent.
Decreased
species:
Inconsistent:
Magnitude of
change
inconsistent with
magnitude of
effect.
Decreased
species: No
known explanation.
+++
+++
0
Low Dissolved oxygen/High
BOD
Decreased relative
weight: Most
consistent.
Increased DELTA:
Many
inconsistencies:
No known
mechanism.
Decreased
species: Most
consistent.
Increased DELTA:
No known
explanation.
+
+
0
Nutrient Enrichment
Decreased relative
weight: Many
inconsistencies.
Increased DELTA:
Many
inconsistencies:
Magnitude of
change
inconsistent with
magnitude of
effect.
Decreased
species: Many
inconsistencies:
Magnitude of
change
inconsistent with
magnitude of
effect.
Decreased relative
weight, Increased
DELTA, Decreased
species: No
known explanation.
0
NE = no evidence; NA = not applicable/not available
7-42
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-11. Strength of evidence analysis for the three candidate causes of Impairment C,
RM5.7.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Case-Specific Considerations
Co-occurrence
Temporality
Consistency of
Association
Biological
Gradient
Metals Contamination
Uncertain: There
were only slight
changes in metal
concentrations in
sediment at RM
5.7 compared to
RM6.5. Only
copper and zinc
increased slightly
and possibly
cadmium. All
others declined.
No evidence
Similar patterns of
fish and
invertebrate
communities are
seen at RM 5.7,
4.4 and 2.7
Increased DELTA:
Strong and
monotonic: From
RM5.7toRM0.4,
copper and
mercury are
strongly correlated
with % DELTA.
Decreased
Tanytarsini: Strong
and monotonic:
The decline in %
tanytarsini was
also strongly
correlated with
copper and
mercury.
0
NE
+
++
++
Ammonia Toxicity
Compatible:
Ammonia
concentrations
were 1 0X or
greater than at RM
6.5. from RM 5.7 to
RM2.7
No evidence
Similar patterns of
fish and
invertebrate
communities are
seen at RM 5.7,
4.4 and 2.7.
Increased DELTA:
None: No
correlation of
ammonia with %
DELTA.
Decreased
Tanytarsini: None:
No correlation of
ammoniawith the
decline in %
Tanytarsini.
+
NE
+
-
Nutrient Enrichment
Compatible: Total
phosphorus and
nitrogen
concentrations are
elevated at RM 5.7
through 2.7. P
values are more
than 24X greater
than at RM 6.5 and
more than 1 0X
greater for nitrogen
than upstream.
No evidence
Similar patterns of
fish communities
are seen at RM 5.7,
4.4 and 2.7.
Increased DELTA:
Strong and
monotonic: %
DELTA was
moderately
correlated with
BOD, N and P.
Decreased
Tanytarsini: Strong
and monotonic:
BOD, nitrate-nitrite
and phosphorus
were all strongly
correlated with
decline in %
Tanytarsini midges
and the ICI.
+
NE
+
++
++
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-43
-------
Stressor Identification Guidance Document
Table 7-11 (continued). Strength of evidence analysis for the three candidate causes of
Impairment C, RM 5.8
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Case-Specific Considerations (cont'd)
Complete
Exposure
Pathway
Experiment
Metals Contamination
Incomplete
evidence: Lead
and zinc were
detected in water
samples (OEPA
1992a). In
sediment, many
metals were
detected. No
internal
concentrations of
metals were
measured. Water
hardness may
have reduced
metal availability.
No evidence.
+
NE
Ammonia Toxicity
Evidence for all
steps: Ammonia
levels measured in
water column, so
exposure possible
for fish and
invertebrates.
Ammonia is
directly discharged
into streams by
point sources.
Temperature and
pH conditions are
favorable for
forming unionized
ammonia, the toxic
form of ammonia.
Conditions are
favorable for
conversion of
nitrites to ammonia
(low DO).
No evidence.
+ +
NE
Nutrient Enrichment
Incomplete
evidence: Nutrient
and phosphorus
concentrations were
measured in water
column, and would
be available for
algal, fungal and
bacterial growth.
Neither algal nor
chlorophyll a
concentrations, the
direct effect of
nutrient enrichment,
nor bacterial
concentrations were
not measured.
No evidence.
+
NE
Considerations Based on Other Situations or Biological Knowledge
Plausibility:
Mechanism
Increased DELTA:
Implausible: Metals
do not cause fin
erosion and
lesions (Eisler
2000b).
Decreased
Tanytarsini:
Plausible: Metals
are known to
cause lethal and
sub-lethal effects
to invertebrates
that can extirpate
species from a
site. In a literature
review, lead and
copper were
associated with
mortality and other
metals with
mortality,
reproduction,
growth and
behavior changes
(Eisler 2000b).
-
+
Increased DELTA:
Plausible:
Ammonia has
been associated
with DELTA (Dyer,
pers. comm.).
Decreased
Tanytarsini:
Plausible:
Ammonia is toxic
to benthic
macroinvertebrates
(USEPA1998b).
+
+
Increased DELTA:
Plausible: Nutrients
are believed to
create conditions
that favor
opportunistic
pathogens and fungi
that cause lesions,
fin erosion and
interfere with wound
healing (Rankin et
al. 1999).
Decreased
Tanytarsini:
Increased nutrients
are known to
change community
structure primarily
by changing the
food source (Allan
1995).
+
+
NE = no evidence; NA = not applicable/not available
7-44
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-11 (continued). Strength of evidence analysis for the three candidate causes of
Impairment C, RM 5.8.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Considerations Based on Other Situations or Biological Knowledge (cont'd)
Plausibility:
Stressor-
Response
Consistency of
Association
Specificity of
Cause and Effect
Analogy
Experiment
Predictive
Performance
Metals Contamination
Increased DELTA:
Not applicable.
Mechanism not
plausible.
Decreased
Tanytarsini:
Ambiguous: The
cumulative toxic
units exceed PEL
by 1.5 to 2.8 times
in 1988/91 and
1992, respectively.
The cumulative
toxic units for PEL
decreased
compared to
upstream in
1988/91 and 1992.
In 1998,
cumulative PEL
was 3.5 times
greater than at
Impairment B, but
this occurred after
the impairment had
already occurred
(USEPA1996b).
Increased DELTA:
Many exceptions.
Ohio EPA
database.
Decreased
Tanytarsini: No
evidence.
Increased DELTA:
Not applicable.
Decreased
Tanytarsini: One of
many.
Not applicable
No evidence
No evidence
NA
0
NE
NA
0
NA
NE
NE
Ammonia Toxicity
Increased DELTA:
Concordant
Decreased
Tanytarsini:
Quantitatively
consistent:
Ammonia
concentrations are
in a plausible
range to cause
toxic effects
especially on
warm, sunny days.
Conservatively,
ammonia was two
times the USEPA
chronic criteria
(USEPA 1996b).
Increased DELTA:
In most places
(Rankin et al.
1999).
Decreased
Tanytarsini: No
evidence.
Increased DELTA:
One of a few.
Decreased
Tanytarsini: One of
many.
Not applicable
No evidence
No evidence
+
+++
++
NE
++
++
NA
NE
NE
Nutrient Enrichment
Increased DELTA:
Quantitatively
consistent:
%DELTA consistent
with associations of
P concentrations
found in streams
throughout Ohio
(Rankin et al, 1999)
Decreased
Tanytarsini:
Concordant.
Nutrient criteria are
proposed for Ohio
and were exceeded
at RM 5.7 through
RM 0.4 for both
nitrate-nitrite and
phosphorus. At RM
5.7, nitrogen
concentration was
five times the
proposed criterion
value. P
concentration was
more than seven
times the proposed
phosphorus criterion
(Rankin et al. 1999).
Increased DELTA:
In most places
(Rankin et al. 1999).
Decreased
Tanytarsini: No
evidence.
Increased DELTA:
One of a few.
Decreased
Tanytarsini: One of
many.
Not applicable
No evidence
No evidence
+++
+
+ +
NE
++
++
NA
NE
NE
NE = no evidence; NA = not applicable/not available
Chapter 7: Little Scioto River, Ohio
7-45
-------
Stressor Identification Guidance Document
Table 7-11 (continued). Strength of evidence analysis for the three candidate causes of
Impairment C, RM 5.8.
Causal
Consideration
Evidence
Score
Evidence
Score
Evidence
Score
Considerations from Multiple Lines of Evidence
Consistency of
Evidence
Coherence of
Evidence
Metals Contamination
Increased DELTA:
Multiple
inconsistencies.
Decreased
Tanytarsini: Most
consistent.
Although metals
are toxic the
magnitude and
type of effect do
not seem to
indicate that
metals caused
either the increase
% DELTA or shifts
in invertebrate
metrics. However,
mercury and
copper are both
significantly
correlated with %
DELTA and %
tanytarsini.
Increased DELTA:
No known
explanation.
0
0
0
Ammonia Toxicity
Increased DELTA:
Most consistent.
Decreased
Tanytarsini: Most
consistent.
Ammonia may
have toxic effects,
but % DELTA not
likely to be caused
by ammonia. No
biological
correlation.
Increased DELTA:
Biological gradient
based on few
observations and
may be
confounded by
other stressors
downstream.
Decreased
Tanytarsini:
Biological gradient
based on few
observations and
may be
confounded by
other stressors
downstream.
+
+
0
0
Nutrient Enrichment
Increased DELTA:
All consistent.
Decreased
Tanytarsini: All
consistent.
Reasonable
evidence to suspect
that nitrogen and
phosphorus are
creating conditions
that favor
opportunistic
pathogens.
Proposed criteria
values are
exceeded and high
% DELTA
consistent with
effects seen even in
the absence of
toxics. Shifts in
invertebrate metrics
more uncertain.
+++
++
NE = no evidence; NA = not applicable/not available
7-46
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
7.10 Characterize Causes: Identify Probable Causes
CHARACTERIZE CAUSES
Eliminate
Diagnose
Strength of Evi
Identify Probable Cause
dence
Impairment A (RM 7.9). At RM 7.9,
there is a decline in IBI and ICI that is
characterized by an increase in the
relative weight offish and percent
DELTA, a decreased number offish
and species offish, and a decreased
percentage of mayflies. Candidate
Causes #2, PAH, #4, ammonia, and #5,
low DO/BOD were eliminated (Tables
7-5 and 7-6). Candidate causes #1,
habitat alteration, #3, metal
contamination, and #6, nutrient enrichment, were evaluated in a strength of evidence
analysis (Tables 7-9, 7-10 and 7-11). An artificially deepened channel was identified
as the probable cause for an increase in the relative weight offish. An embedded
stream bed was identified as the probable cause for decreased numbers and species
offish and decreased percentage of mayflies. The stream bed may have been
susceptible to becoming embedded due to a lower gradient than upstream. The
probable cause for the low but measurable increase in percent DELTA remained
uncertain. The strength of evidence analysis strongly supports this causal
relationship. The quality of the data is high, and the consistency of the evidence is
good.
Impairment B (RM 6.5). At RM 6.5, there is a further decline in the IBI and ICI.
Specific impairments include an increase in % DELTA, a decrease in the relative
weight and numbers of species offish, and an additional decrease in percent
mayflies. Habitat alteration was eliminated as a candidate cause (Tables 7-5 and 7-
6). In the strength of evidence analysis a single probable cause, PAHs, was found to
be sufficient to cause all of the specific impairments (Tables 7-10 and 7-12). Habitat
alteration continued to impair the site but was not the cause of the increased DELTA,
decreased relative weight, or the additional decline in the number of species. The
strength of evidence analysis strongly supports this causal relationship. The quality
of the data is high, and the consistency of the evidence is very good.
Impairment C (RM 5.7). At RM 5.7, there is a notable further increase in %
DELTA and a decrease in % Tanytarsini. Altered habitat and PAH still cause
impairments, but since the level of alteration remains about the same or decreases,
these candidate causes were eliminated (Tables 7-5 and 7-6). In the strength of
evidence analysis, nutrient enrichment, candidate cause #6, was identified as the
probable cause for both impairments. Nevertheless, ammonia toxicity may still be
important. We have moderate confidence in this characterization.
The causal characterization of the Little Scioto River could be strengthened by evidence
from published literature that reports associations applying to plausible mechanism and
stressor-response, consistency of association, specificity, and others. It was not the
intent of this document to prepare an exhaustive list of appropriate evidence, but such a
resource is certainly needed to make these types of evidence accessible for future
characterizations. This case study does demonstrate the stressor identification process
and the importance of clearly presenting the reasoning and evidence.
Chapter 7: Little Scioto River, Ohio
7-47
-------
Stressor Identification Guidance Document
Table 7-12. Causal characterization.
Impairment A - RM 7.9
Impairment B - RM 6.5
Impairment C - RM 5.7
Probable Cause: Habitat Alteration
Probable Cause: PAH
Contamination
Probable Cause: Nutrient Enrichment
Increased Relative Weight: Is
probably caused by the artificial
deepening of the channel that allows
larger fish to live there.
Increased DELTA: The percentage of
DELTA is commonly associated with
channelized streams, but the specific
aspect of the channelization that
increased DELTA is unknown.
Loss of species: Many factors could
contribute to the loss offish and
benthic invertebrate species; however,
embedded substrates seem to be the
most likely stressor since upstream
locations had even lower DO levels
and yet had a greater variety offish
and invertebrate species.
Although metals are present, the
likelihood of response at the these
concentrations are low. Furthermore,
the types of changes in the
community, especially an increase in
the relative weight offish, is very
unlikely with the candidate cause of
metals.
Although P levels are slightly higher,
effects are not associated with these
phosphorous concentration elsewhere
and they do not exceed Ohio's
proposed criteria values for effects.
Candidate Causes #2, PAH, and #4,
Ammonia, were eliminated because
levels were the same or lower than
upstream. Candidate Cause #5, Low
DO /BOD , was also eliminated as an
overall pathway; however, low DO
associated with channelization may
still play a roll especially in DELTA.
Siltation and deepened channel are
consistent with Impairment A. The
magnitude of the alteration and clear
difference from upstream location
strongly support this cause.
A single cause is likely for the
three manifestations of
Impairment B: decreased
relative weight, increased
DELTA, and decreased
species:
The probable cause of
Impairment B is toxic levels
of PAH-contaminated
sediments. All of the
evidence support PAH
contamination as the cause.
There is a complete exposure
pathway at the location and
clear mechanism of action for
each of the effects. The
single most convincing piece
of evidence is that the
cumulative toxic units of PAH
were more than 300 times
the probable effects level.
Metals are at sufficient
concentrations to cause
effects; however, they were
sometimes at levels close to
upstream levels and were
less than 2% as toxic as the
lowest cumulative toxic units
of PAH. Metal concentrations
are high enough that they
should be considered a
potentially masked cause.
Candidate cause #5 is
unlikely because even
greater levels of BOD did not
cause reduction of dissolved
oxygen downstream.
Candidate Causes #4,
Ammonia, and #6, Nutrient
Enrichment, are unlikely
given that state criteria levels
were met and the much
stronger evidence for PAH.
Habitat alteration continues
to impair the site, but it is not
the cause of the increased
DELTA, decreased relative
weight, or the additional
decline in the number of
species.
At Impairment C increased % DELTA
and % Tanytarsini may have different
causes. Increased DELTA in fish is
probably caused by increased P and
NOX. Nutrients, especially P, have
been associated with increased fin
erosion and lesions but some
uncertainty exists since P acts
indirectly.
Ammonia is slightly higher than at
Impairment B and exceeded ammonia
criteria values. Biological gradients
were absent; however, this may have
been a statistical artifact given the
number of sites available to perform
the analysis and potential interference
from other stressors downstream.
Metals are considered unlikely
because surface lesions are only
occasionally noted as effects from long
term exposure and only some metal
concentrations were slightly greater
than at Impairment B. Metal
concentrations are high enough that
they should be considered a
potentially masked cause.
The probable cause of extirpation of
Tanytarsini at Impairment C is more
uncertain because less is known about
the natural history and stressor
response relationships of these
benthic invertebrates. Candidate
cause #6, nutrient enrichment, still
seems to be the most likely cause
since all of the strength of evidence
considerations were consistent.
PAH contamination and habitat
alteration continue to impair the site,
but they are not the cause of the
increased % DELTA or extirpation of
Tanytarsini.
The causal characterization at
Impairment C is less certain, but the
strength of evidence favors cause #6,
increased nutrients.
7.11 Discussion
An important, practical aspect of this study is that even though the primary cause was
identified in each case, it is obvious that other causes are also present that would
constrain the biological community if the dominant cause was removed. For instance, if
PAHs could be independently removed from the river, metals might be high enough to
7-48
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
impair the biological assemblage. Likewise, if metals were removed, habitat alteration
would still affect the biological community and would lower IBI and ICI scores at
Impairments B and C.
Another issue is the impact of habitat alteration and its influence on modifying the
assimilative capacity of the river. In other words, if the physical habitat were improved,
would the impacts of PAH contamination be lessened? At Impairment B, this is unlikely
based on evidence from at least one river elsewhere that has very good physical habitat
qualities, yet has an impoverished biological community replete with high levels of %
DELTA due to high PAH concentrations (OEPA 1992b, Cormier et al. 2000b). The
strength of evidence analysis can provide these insights for the next step in managing
ecosystems, which is to find ways to identify and apportion the sources for the identified
causes and then take action to restore and protect the resource.
At Impairment C, a physical habitat that included wetlands, riparian wetlands, and
riparian cover might improve the assimilative capacity of the river by providing sinks for
the nutrient and ammonia loadings. However, since PAH and metals contamination are
still high at Impairment C, removal of nutrient loading alone would result in only a very
small improvement in biological condition.
At Impairment B, nutrient enrichment was retained as a candidate cause, even though the
increase in phosphorous was minute. Nutrient enrichment was an unlikely cause, but the
reasons for it being improbable come from ecological knowledge from examples in other
watersheds, not from evidence that permits elimination. The reason nutrient enrichment
was retained was because it failed to meet the criteria for elimination. The strength of
evidence is the proper way to show this evidence.
There are other uncertainties. Wet weather flow data was not available for review.
Events, especially near the combined sewer overflow at RM 6.0, could be undetected
sources of candidate causes. Downstream from Impairment C, persistent impairments
may have other causes. For instance, BOD is elevated at RM 5.8; however, its effects
are usually associated with a certain lag time that results in low DO.
The results from this particular causal analysis could have several practical applications.
If it is determined that the river conditions must be improved due to state regulations,
federal TMDL (total maximum daily load) rules, citizen action, or other reasons, one
option is to remove or decrease all potential stressors identified in the causal analysis;
that is, remove both channel modification as well as water and sediment contamination.
However, there may be intermediate pathways that may be more cost effective. Factors
that should be considered in choosing an option include the desired or expected level of
improvement in river condition, and the usefulness of the river's resources versus the
cost to restore the river. Another factor to consider is the mode of restoration. For
instance, both PAH and metal remediation may require dredging of the contaminated
sediments. Knowing which agents (PAH, metals, or a combination of the two) may
satisfy our curiosity, but it may not change the management action or ecological
outcome. However, it might be determined that knowing the cause is important for
assigning the financial responsibility for clean-up. In the latter case, additional
information may be needed, especially if restoration costs are high.
Chapter 7: Little Scioto River, Ohio 7-49
-------
Stressor Identification Guidance Document
7.12 References
Albers, P. 1995. Petroleum and individual polycyclic aromatic hydrocarbons, Pages
330-355 in Hoffman, David J. et al, ed. Handbook ofEcotoxicology, CRC Press,
Boca Raton, Florida.
Allan, J.D. 1995. Stream ecology: structure andfunction of'running waters. Chapman
and Hall Publishers, London.
Baumann, P.C., and Harshbarger, J.C. 1995. Decline in liver neoplasms in wild brown
bullhead catfish after coking plant closes and environmental PAHs plummet.
Environ. Health Perspect. 103:168-170.
Carpenter, S.R., N.F. Caraco, D.L. Correll, R.W. Howarth, A.N. Sharpley and V.H.
Smith. 1988. Non-point pollution of surface waters with phosphorus and nitrogen.
Ecological Applications 8(3):559-568.
Cormier, S.M., E.L.C. Lin, F.A. Fulk and B. Subramanian. 2000a. Estimation of
exposure criteria values for biliary polycyclic aromatic hydrocarbon metabolite
concentration in white suckers (Catostomus commersoni). Environmental
Toxicology and Chemistry 19:1120-1126.
Cormier, S.M., E.L.C. Lin, M.R Millward, M.K. Schubauer-Berigan, D. Williams, B.
Subramanian, R. Sanders, B. Counts and D. Altfater. 2000b. Using regional
exposure criteria and upstream reference data to characterize spatial and temporal
exposures to chemical contaminants. Environmental Toxicology and Chemistry
19:1127-1135.
Dodds, K. and E.B. Welsh. 2000. Establishing nutrient criteria in streams. The North
American Benthological Society 19:186-196.
Edwards, R., and J. Riepenhoff. 1998. State turns to feds for cleanup. Columbus
Dispatch, April 28, 1998.
Edwards, A.C., H. Twist and G.A. Codd. 2000. Assessing the impact of terrestrially
derived phosphorus on flowing water systems. Journal of Environmental Quality.
29:117-124.
Edwards, C.J., B.L. Griswold, RA. Tubb, E.G. Weber, L.C. Woods. 1984. Mitigating
effects of artificial riffles and pools on the fauna of a channelized warmwater stream.
North American Journal of Fisheries Management 4:194-203
Eisler, R. 2000a. Polycyclic aromatic hydrocarbons. Pages 1343-1411 in Handbook of
Chemical Risk Assessment. Vol. II. Lewis Publishers, Boca Raton, FL.
. 2000b. Handbook of Chemical Risk Assessment. Vol.1. Lewis Publishers,
Boca Raton, FL.
Gmur, D.J., and U. Varanasi. 1982. Characterization of benzo[a]pyrene metabolites
isolated from muscle, liver, and bile of a juvenile flatfish. Carcinogenesis 5:1397-
1403.
7-50 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Hickey, C.W. and W.H. Clements. 1998. Effects of heavy metals on benthic
macroinvertebrate communities in New Zealand streams. Environmental Toxicology
and Chemistry 17:2338-2346
Karr, J.R., and I.J. Schlosser. 1977. Impact of near stream vegetation and stream
morphology on water quality and stream biota. EPA-600-3-77-097. U.S.
Environmental Protection Agency, Environmental Research Laboratory, Athens, GA.
Lin, E.L.C., S.M. Cormier, and J.A. Torsella. 1996. Fish biliary polycyclic aromatic
hydrocarbon metabolites estimated by fixed-wavelength fluorescence: comparison
with HPLC-fluorescent detection. Ecotoxicol. Environ. Safety 35:16-23.
Lin, E.L.C., T.W. Neiheisel, B. Subramanian, D.E. Williams, M.R. Millward, and S.M.
Cormier. Historical monitoring of biomarkers of exposure of brown bullhead in the
remediated Black River, Ohio and two other Lake Erie tributaries. Submitted to
Journal of Great Lakes Research.
Long, E.R., L.J. Field, and D.D. MacDonald. 1998. Predicting toxicity in marine
sediments with numerical sediment quality guidelines. Envir. Toxicol. Chem.
17:714-727.
Meyer, P.P. and L.A. Barclay. 1990. Field manual for the investigation offish kills.
Resource Pub. 177. U.S. Fish and Wildlife Service, Washington, B.C.
Miltner, R.J. and E.T. Rankin. 1998. Primary nutrients and the biotic integrity of rivers
and streams. Freshwater Biology 40:145-158.
Norton, S.B. 1999. Using biological monitoring data to distinguish among types of
stress in streams of the Eastern Cornbelt Plains Ecoregion. Ph.D. Dissertation,
Georgetown University, Fairfax, VA.
Ohio Environmental Protection Agency (OEPA). 1988a. Biological criteria for the
protection of aquatic life: Vol. II. users manual for biological assessment ofohio
surface waters. Division of Water Quality Planning and Assessment, Ecological
Assessment Section, Columbus, OH.
. 1988b. Biological and water quality study of the Little Scioto River
watershed, Marion County, OH. OEPA Technical Report prepared by State of Ohio
Environmental Protection Agency, Division of Surface Water, Columbus, OH.
. 1989a. Addendum to: Biological criteria for the protection of aquatic life:
Volume II. users manual for biological assessment of Ohio surface waters. Division
of Water Quality Planning and Assessment, Ecological Assessment Section,
Columbus, OH.
. 1989b. Biological criteria for the protection of aquatic life: Volume III.
standardized field and laboratory methods for assessing fish and macroinvertebrate
communities. Division of Water Quality Planning and Assessment, Ecological
Assessment Section, Columbus, OH.
. 1989c. Manual of Ohio EPA surveillance methods and quality assurance
practices. Division of Environmental Services, Columbus, OH.
Chapter 7: Little Scioto River, Ohio 7-51
-------
Stressor Identification Guidance Document
1992a. Bottom sediment evaluation, Little Scioto River, Marion, Ohio.
Division of Water Quality Planning and Ecological Assessment Section, Columbus,
OH.
. 1992b. Biological and water quality study of the Ottawa River, Hog
Creek, Little Hog Creek, and Pike Run. OEPA Technical Report EAS/1992-9-7.
Prepared by State of Ohio Environmental Protection Agency, Division of Surface
Water, Columbus, OH.
. 1994. Biological, sediment, and water quality study of the Little Scioto
River, Marion, Ohio. OEPA Technical Report EAS/1994-1-1. Division of Surface
Water, Ecological Assessment Section, Columbus, OH.
Rankin, E.T. 1989. The qualitative habitat evaluation index (QHEI): rationale,
methods, and application. State of Ohio Environmental Protection Agency, Division
of Water Quality Planning and Assessment, Ecological Assessment Section,
Columbus, OH.
. 1995. Habitat indices in water resource quality assessments. Pages 181-
208 in W.S. Davis and T.P. Simon (editors). Biological Assessment and Criteria.
Lewis Publishers, Boca Raton, Florida.
Rankin E., R. Miltner, C. Yoder and D. Mishne. 1999. Association between nutrients,
habitat, and the aquatic biota in Ohio rivers and streams. Ohio EPA Technical
bulletin MAS/1999-1. Ohio EPA, Columbus, OH.
Roubal, W.T., T.K. Lallier, and D.C. Malins. 1977. Accumulation and metabolism of
C-14 labeled benzene, naphthalene, and anthracene by young coho salmon
(Oncorhynchus kisutch) and starry flounder (Platichthys stellatus). Arch. Environ.
Contam. Toxicol. 5: 513-529.
Russo, R.C. 1985. Ammonia, nitrate and nitrite. Pages 455-471 in G.M. Rand and S.A.
Petrocelli (editors). Fundamentals of Aquatic Toxicology. McGraw Hill,
Washington, D.C.
Sheilds, F.D. Jr., S.S. Knight and C.M. Cooper. 1998. Rehabilitation of aquatic habitats
in warmwater streams damaged by channel incision in Mississippi. Hydrobiologica
382:63-86
Smith, V.H., G.D. Tilman and J.C. Nekola. 1999. Eutrophication: impacts of excess
nutrient inputs on freshwater, marine and terrestrial ecosystems. Environmental
Pollution 100:179-196.
Tarplee, W.H. Jr., D.E. Louder, and A.J. Weber. 1971. Evaluation of the effects of
channelization on fish populations in North Carolina's coastal plain streams. North
Carolina Wildlife Resources Commission, Raleigh, NC.
U.S. Environmental Protection Agency (USEPA). 1996b. Calculation and evaluation of
sediment effect concentrations for the amphipod Hyallela azteca and the midge
Chironomus riparius. Assessment and Remediation of Contaminated Sediments
(ARCS) Program. Great Lakes National Program Office, Chicago, IL. EPA 905-
R96-008.
7-52 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
1998b. 1998 update of ambient water quality criteria for ammonia. Office
of Water, Washington, DC. EPA 822-R-98-008.
Varanasi, U., J.E. Stein, M. Nishimoto, and T. Horn. 1983. Benzo[a]pyrene metabolites
in liver, muscle, gonads, and bile of adult English sole (Parophrys vetulus). Pages
1221-1234 in Cooke, M. and A.J. Dennis, eds. Polynuclear Aromatic Hydrocarbons:
Formation, Metabolism, and Measurement. Battelle, Columbus, OH, USA.
Yoder, C.O., and E.T. Rankin. 1995a. Biological criteria program development and
implementation in Ohio. Pages 109-144 in W.S. Davis and T.P. Simon, eds.
Biological Assessment and Criteria: Tools for Water Resource Planning and
Decision Making. Lewis Publishers, Boca Raton, FL.
. 1995b. Biological response signatures and the area of degradation value:
New tools for interpreting multi-metric data. Pages 236-286 in Biological
Assessment and Criteria: Tools for Water Resource Planning and Decisionmaking
Lewis Publishers, Boca Raton, FL..
Yount, J.D., and G.J. Niemi. 1990. Recovery of lotic communities and ecosystems from
disturbance; A narrative review of case studies. Environmental Management
14:547-569.
Chapter 7: Little Scioto River, Ohio 7-53
-------
Table 7-13. Fish metrics for the Little Scioto River 1987 and 1992.*
Response
Total No. of Species
No. of Darter Species
No. of Sunfish species
No. of Sucker Species
No. of Intolerant Species
Percent Tolerant Species
Percent Omnivores
Percent Insectivores
Percent Pioneering Species
No. of Individuals
Percent Simple Lithophilic
Species
Percent DELTA
Relative Weight
IBI
River Mile
{9.2}
(9.2)
{19.3}
(22)
{5}
(5.5)
{3}
(4)
{1.3}
(2.5)
{1}
(D
{35.45}
(60.69)
{33.28}
(56.21)
{53.41}
(35.96)
{35.49}
(69.85)
{808.5}
(1104.6)
{21 .24}
(12.8)
{0.13}
(0.14)
{4.021}
(8.6)
{332
(33)
(7.9)
(13)
(0)
(5)
(3)
0
(69.12)
(44.95)
(53.07)
(28.41)
(206)
(18.19)
(1.64)
(74.9)
(23)
{6.5}
(6.5)
{13}
(8.3)
{0}
(0)
{3}
(2.3)
{1}
(1.3)
{0}
(0)
{82.43}
(85.12)
{57.2}
(56.72)
{40.99}
(39.77)
{22.52}
(26.39)
{416.3}
(335)
{2.7}
(42.69)
{0.0}
(9.98)
{34.2}
(38.7)
{24}
(19)
{6.0}
(5.7)
{10}
(10.7)
{0}
(0)
{1.3}
(3.3)
{1.7}
(1.7)
{0}
(0)
{94.28}
(68.14)
{72.47}
(46.84)
{16.73}
(47.8)
{21.94}
(21.76)
{237.33}
(174)
{24.01}
(26.2)
{16.46}
(14.51)
{29.773}
(17.031)
{14}
(19)
(4.4)
(7.7)
(0)
(3)
(D
(0)
(82.75)
(71.37)
(21.91)
(17.81)
(137)
(31.45)
(22.37)
(7.2)
(18)
(3.1)
{3.3}
{0}
{0.7}
{0.7}
{0}
{98.2}
{94.15}
{3.95}
{4.05}
{84.7}
{5.88}
{32.8}
{10.7}
{12}
{2.7}
(2.7)
{6}
(7.7)
{0}
(0)
{0.3}
(2.7)
{1}
(1.3)
{0}
(0)
{94.95}
(70.85)
{85.4}
(51.77)
{10.18}
(42.51)
{7.58}
(23.31)
{237.33
}
(94)
{9.14}
(24.91)
{14.22}
(10.99)
{24.482
}
(6.3)
{13}
(19)
{0.1}
(0.3)
{8.7}
(9.7)
{0.3}
{1}
(2.7)
{2}
(2.7)
{0}
{63.41}
(38.64)
{62.72}
(31.92)
{32.74}
(55.36)
{5.33}
(22.1)
{78}
(75)
{19.01}
(28.08)
{16.19}
(10.04)
{46.079}
(21.1)
{14}
(25)
Q.
Q.
rt;
o'
3
D)
D)
CT
(D
-------
Table 7-14. Macroinvertebrate metrics for the Little Scioto River 1987 and 1992.*
Response
Total Number of Macroinvertebrates
Total No. of Taxa collected at a Site, both
Qualitative and Quantitative
Total No. of Quantitative Taxa
No. of Mayfly Taxa
No. of Caddisfly Taxa
No. of Dipteran Taxa
No. of Qualitative EPTTaxa
Percent Mayfly Taxa
Percent Caddisfly Taxa
Percent Tanytarsini Midges
Percent Dipterans
Percent Non-insects
Percent Tolerant Organisms
Percent Cricotopus
ICI
River Mile
{9.2}
(9.2)
{773}
(1464)
{51}
(47)
{34}
(36)
{6}
(7)
{2}
(3)
{19}
(20)
{10}
(8)
{56.016}
(58.811)
{4.657}
(6.557)
{1.552}
(3.347)
{26.132}
(32.445)
{7.762}
(1.639)
{4.916}
(8.607)
{0.388}
(0.48)
{40}
(38)
(7.9)
(1952)
(38)
(30)
(2)
(0)
(18)
(1)
(16.393)
(0)
(3.381)
(74.795)
(5.43)
(15.061)
(0)
(16)
{6.5}
(6.5)
{1116}
(2815)
{38}
(29)
{25}
(18)
{3}
(2)
{2}
(0)
{15}
(12)
{2}
(0)
{20.251}
(5.009)
{0.179}
(0)
{4.48}
(2.345)
{55.018}
(57.336)
{20.43}
(37.549)
{37.993}
(77.371)
{6.631}
(0)
{22}
(8)
{5.8}
(5.7)
{207}
(1600)
{26}
(32)
{13}
(18)
{2}
(2)
{0}
(0)
{7}
(13)
{1}
(1)
{3.382}
(2)
{0}
(0)
{0.966}
(0)
{37.198}
(91)
{56.039}
(7)
{61.353}
(67.75)
{0}
(8)
{8}
(6)
(4.4)
(1899)
(27)
(20)
(1)
(1)
(13)
(0)
(0.263)
(0.053)
(0)
(74.829)
(21.959)
(54.766)
(2.53)
(10)
{3.2}
{763}
{28}
{14}
{1}
{0}
{10}
{4}
{1.573}
{0}
{3.67}
{95.937}
{0.524}
{20.315}
{0}
{8}
{2.7}
(2.1)
{1779}
(5242)
{28}
(41)
{13}
(23)
{0}
(2)
{0}
(3)
{11}
(14)
{2}
(6)
{0}
(0.114)
{0}
(0.267)
{0}
(0.343)
{23.834}
(97.138)
{75.998}
(2.461)
{89.545}
(29.569)
{4.947}
(0.301)
{4}
(18)
{0.4}
(0.4)
{645}
(1151)
{24}
(37)
{16}
(26)
{2}
(3)
{0}
(1)
{12}
(16)
{1}
(2)
{1.24}
(3.215)
{0}
(0.087)
{0}
(2.085)
{61.085}
(91.659)
{37.674}
(4.344)
{52.713}
(59.34)
{4.961}
(2.172)
{6}
(18)
*{}=1987; ()= 1992
-------
Stressor Identification Guidance Document
Table 7-15. QHEI metrics for the Little Scioto River 1987 and 1992.*
Metric
Substrate
Cover
Cover
Types
Channel
Riparian
Pool
Riffle
Gradient
QHEI
River Mile
{9.2}
(9.2)
{18}
(16)
{10}
(14)
{3}
(6)
{18}
(17)
{9}
(6)
{8}
(11)
{5}
(6)
{6}
(6)
{74}
(76)
(7.9)
(5)
(10)
(4)
(10)
(5.5)
(8)
(0)
(4)
(42.5)
{6.5}
(6.5)
{1}
(1)
{9}
(11)
{2}
(6)
{6}
(10)
{4}
(4)
{6}
(8)
{0}
(0)
{4}
(4)
{30}
(38.5)
{6.0
}
{1}
{13}
{6}
{10}
{4}
{8}
{0}
{4}
{40}
(5.7
)
0)
(10)
(4)
(10)
(6)
(9)
(0)
(4)
(40)
(4.4)
(5)
(10)
(4)
(10)
(6)
(6)
(0)
(2)
(39)
{3.1}
{1}
{11}
{4}
{11}
{5}
{6}
{0}
{2}
{36}
{2.7}
(2.7)
{1}
(5)
{13}
(11)
{6}
(6)
{10.5}
(10)
{6}
(8)
{8}
(8)
{0}
(0)
{2}
(2)
{40.5}
(42)
{0.1}
(0.3)
{1}
(5)
{12}
(9)
{5}
(6)
{10}
(7)
{8}
(5.5)
{8}
(8)
{0}
(0)
{4}
(4)
{43}
(38.5)
= 1987;() = 1992
7-56
U.S. Environmental Protection Agency
-------
Table 7-16. Average concentrations of selected sediment organic compounds (mg/kg) in the Little Scioto River, Ohio, by river mile
in 1987, 1991, 1992 and 1998.*
Compound
Acenaphthene
Anthracene
Benzo(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
River Mile
[11.1]
[0.59]ND
[0.59]ND
[0.072]J
[0.068]J
[0.058]J
/9.42\
(9.5)
[9.21]
(ND)
[0.7]ND
(ND)
[0.70]ND
(ND)
[0.043]J
(ND)
[0.052]J
(ND)
[0.052]J
(7.9)
(ND)
(ND)
(ND)
(ND)
(ND)
/7.15\
[7.09]
[0.047]J
[0.037]J
/15\J
[0.059]J
/25\
[0.051]J
[0.046]J
{6.5}
/6.6\
(6.5)
[6.6]
{14.8}
(ND)
[760]J
{66.8}
(ND)
[100]J
{44.7}
/15VJ
(8.2)J
[310]J
{23.6}
/20\J
(18.1)
[200]J
{213.2}
(9.9)J
[160]J
/5.8\
(5.8)
[6.2]
/150\
(5)
[5]
/360\
(27.1)
[41]
/185\
(16.5)
[42]J
/215\
(16.8)
[95]
(12.87)
[80]
(4.4)
(4.3)
(7.9)
(6.9)
(6.9)
(4.6)
{2.7}
/2.7\
(2.7)
[2.65]
{1.3}
(ND)
[0.930]J
{2.3}
(ND)
[3.7]
{4.3}
(2)J
[8.2]
{2.0}
(1.6)J
[12]
{21.3}
(ND)
[10]
(0.4)
(ND)
(3.3)
(15.8)
(13.8)
(10.5)
-------
Table 7-16 (continued). Average concentrations of selected sediment organic compounds (mg/kg) in the Little Scioto River, Ohio,
by river mile in 1987, 1991, 1992 and 1998.*
Compound
Benzo(ghi)perylene
Benzo(a)pyrene
Chrysene
Dibenzo(a,h)anthracene
Fluoranthene
River Mile
[11.1]
[0.052]J
[0.067]J
[0.087]J
[0.59]ND
[0.19]J
/9.42\
(9.5)
[9.21]
(ND)
[0.044]J
(ND)
[0.053]J
(ND)
[0.065]J
(ND)
[0.7]ND
(ND)
[0.097]J
(7.9)
(ND)
(ND)
(ND)
(ND)
(ND)
/7.15\
[7.09]
/10VJ
[0.030]J
/10VJ
[0.043]J
/15VJ
[0.081]J
[0.56]N
D
/20VJ
[0.20]J
{6.5}
/6.6\
(6.5)
[6.6]
{144.1}
(49.5)
[150]ND
{141.1}
(14.8)J
[210]J
{119.5}
/15\J
(16.5)
[390]J
{33.3}
(ND)
[150]ND
{78.4}
/50\
(8.2)J
[100]J
/5.8\
(5.8)
[6.2]
/65\
(11.2)
[19]
/125\
(15.8)
[14]
/305\
(20.8)
[13]
(4.6)
[16]ND
/550\
(37.6)
[44]J
(4.4)
(4.9)
(7.2)
(9.9)
(ND)
(13.5)
{2.7}
/2.7\
(2.7)
[2.65]
{16.5}
(ND)
[13]
{11.4}
(ND)
[12]
{9.7}
(1.6)J
[13]
{2.1}
(ND)
[3.7]
{6.3}
(ND)
[14]
(0.4)
(6.9)
(11.5)
(ND)
(3.3)
(22.4)
-------
Table 7-16 (continued). Average concentrations of selected sediment organic compounds (mg/kg) in the Little Scioto River, Ohio,
by river mile in 1987, 1991, 1992 and 1998.*
Compound
Fluorene
lndeno(1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
River Mile
[11.1]
[0.590]ND
[0.045]J
[0.59]ND
[0.14]J
/9.42\
(9.5)
[9.21]
(ND)
[0.70]ND
(ND)
[0.037]J
(ND)
[0.70]ND
(ND)
[0.053]J
(7.9)
(ND)
(ND)
(ND)
(ND)
/7.15\
[7.09]
[0.059]J
/5\J
[0.56]ND
[0.56]ND
[0.1 1]J
{6.5}
/6.6\
(6.5)
[6.6]
{18.3}
(ND)
[830]J
{156.0}
(13.2)J
[150]ND
{22.9}
(ND)
[260]
{88.3}
/40\J
(ND)
[230]
/5.8\
(5.8)
[6.2]
/200\
(7.0)
[20]
/60\
(14.5)
[16]
/70\
(4.6)
[18]J
/470\
(24.1)
[38]J
(4.4)
(4.0)
(6.6)
(ND)
(12.9)
{2.7}
/2.7\
(2.7)
[2.65]
{1.2}
(ND)
[0.98]J
{18.6}
(ND)
[10]
{1.6}
(ND)
[0.28]J
{2.0}
(ND)
[2.8]J
(0.4)
(ND)
(10.5)
(ND)
(2.6)J
{}= 1987;/\= 1991; () = 1992;[] = 1998
-------
Table 7-16 (Continued). Average concentrations of selected sediment organic compounds (mg/kg) in the Little Scioto River, Ohio,
by river mile in 1987, 1991, 1992 and 1998.*
Compound
Pyrene
River Mile
[11.1]
[0.2]J
/9.42\
(9.5)
[9.21]
(ND)
[0.1 ]J
(7.9)
(ND)
/7.15\
[7.09]
/15VJ
[0.2]J
{6.5}
/6.6\
(6.5)
[6.6]
{67.5}
/30\J
(ND)
[810]J
/5.8\
(5.8)
[6.2]
/405\
(23.8)
[32]J
(4.4)
(10.2)
{2.7}
/2.7\
(2.7)
[2.65]
{5.2}
(ND)
[10]
(0.4)
(17.5)
{} = 1987 data from OEPA 1988, sample depth unknown
/ \ = 1991 data from OEPA 1992a, sample depth unknown
() = 1992-93 data from OEPA 1994, sample from 1-6" except RM 7.9 sample from 8-12"
[ ] = 1998 data from OEPA unpublished, sample depth unknown
J is an estimated value that is above zero but below the practical quantitation limit.
-------
Table 7-17. Average concentrations (mg/kg) of selected metals in sediment from the Little Scioto River, Ohio, by river mile in 1987,
1991,1992 and 1998.*
Metal
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
River Mile
[11.1]
[3.6]J
[0.1]ND
[8.1]J
[15.7]
[23.8]
[0.1]ND
/9.4\
(9.5)
[9.2]
/<10\
(<10)
[8.3]J
/<1 .0\
(<1.0)
[0.1]ND
/5.8\
(7.3)
[14.3]J
(7.4)
[24.1]
/<10\
(12.1)
[20.4]
/<0.1\
(<0.1)
[0.2]J
(7.9)
(12.4)
(<1.0)
(13.6)
(17.2)
(19.1)
(<0.1)
/7.2\
[7.1]
/<10\
[6.0]J
/<1 .0\
[0.2]
/13.2\
[8.9]
[22.9]
/25.5\
[24]J
/<0.1\
[0.1]J
{6.5}
/6.6\
(6.5)
[6.6]
{11.2}
/<10\
(<10)
[10.8]J
{1.8}
/3.4\
(<1.0)
[0.1]ND
{47.6}
/415\
(208)
[32.3]J
{68}
(79)
[39.2]
{170}
/175.5\
(172)
[46.4]
/0.3\
(0.33)
[0.3]J
/5.8\
(5.8)
[6.2]
/<10\
(13.8)
[9.8]J
/1.0\
(<1.0)
[2.0]
/39.2\
(60.9)
[50.4]
(56.0)
[133]
/59.5\
(84.6)
[220]J
/0.2\
(0.2)
[0.6]J
(4.4)
(11.3)
(10.5)
(302)
(76.8)
(93.4)
(0.8)
{2.7}
/2.7\
(2.67)
[2.65]
{9.49}
(<10)
[9.0]
{4.39}
(1.0)
[1.4]
{134}
(71.2)
[77.1]
{83}
(42.4)
[79.3]
{160}
(108)
[180]J
(0.12)
[0.4]J
(0.36)
(<10)
(1.6)
(48.6)
(24.5)
(38)
(<0.1)
-------
Table 7-17 (continued). Average concentrations (mg/kg) of selected metals in sediment from the Little Scioto River, Ohio, by river mile in
1987, 1991,1992 and 1998.*
Metal
Zinc
River Mile
[11.1]
[48.2]
/9.4\
(9.5)
[9.2]
(30.6)
[81.4]
(7.9)
(79.0)
/7.2\
[7.1]
[66.6]
{6.5}
/6.6\
(6.5)
[6.6]
{187}
(173)
[89.2]
/5.8\
(5.8)
[6.2]
(141)
[280]J
(4.4)
(226)
{2.7}
/2.7\
(2.67)
[2.65]
{760}
(408)
[316]J
(0.36)
(96.8)
{} = 1987 data from OEPA 1988, sample depth unknown
/ \ = 1991 data from OEPA 1992, sample depth unknown
() = 1992-93 data from OEPA 1994, sample from 1-6" except RM 7.9 sample from 8-12"
[ ] = 1998 data from OEPA unpublished, sample depth unknown
J is an estimated value that is above zero but below the practical quantitation limit.
-------
Table 7-18. Average concentrations of selected water chemistry parameters (mg/L) in the Little Scioto River, Ohio, by river mile in 1987,
1992 and 1998.*
Compound
Ammonia
Dissolved
oxygen**
BOD
Nitrate- nitrite,
NOX
Phosphorus,
total P
Hardness,
CaCO3
River Mile
[11.1]
[0.1,0.3]
[<2.0, 6.6]
[0.7,3.3]
[0.5,0.6]
[222,250]
(9.2)
[9.2]
(<0.05)
[<0.05,<0.0
5]
(12.2,8.8)
(1.0)
[<2.0, <2.0]
(1.2)
[0.4, 0.2]
(0.06)
[1.8,0.1]
(329)
[275, 269]
{7.9}
(7.9)
(<0.05)
{4.6,
2.8}
(7.9,
5.7)
(1.0)
(1.4)
(0.07)
(327)
[7.1]
[0.11,
<0.05]
[<2.0, 2.1]
[0.73, <0.1]
[0.36,0.13]
[281 , 407]
{6.5}
(6.5)
(0.12)
{7.27, 1.9}
(2.3)
(0.8)
(0.09)
(389)
{5.8}
(5.8)
[6.2]
(1.16)
[0.35, 0.69]
{8.3, 4.2}
(8.23,
4.21)
(4.7)
[4.6,13]
(8.1)
[0.33, 2.37]
{1.65}
(2.17)
[1.9, 1.21]
(278)
[224,261]
{4.4}
(4.4)
(1.44)
{8.8, 3.2}
(5.2, 4.3)
(4.2)
(6.6)
(1 .96)
(280)
{2.7}
(2.7)
[2.7]
(2.10)
[0.67, 1.1]
{6.67, 2.0}
(4.1,3.0)
(3.5)
[3.3,4.1]
(4.5)
[3.5, 0.9]
{2.71}
(1 .80)
[1.18, 1.31]
(306)
[228,210]
{0.4}
(0.4)
(0.58)
{6.74, 2.5}
(5.6, 4.4)
(2.2)
(4.47)
(1.34)
(320)
{} = 1987 (OEPA 1988b; () = 1992-1993 (OEPA 1994) [ ] = 1998 (OEPA August and October, unpublished data).
Dissolved Oxygen {maximum, minimum}, data from 1987 (OEPA, 1988b).
(maximum, minimum from box plots), data from 1992 (OEPA, 1994.
-------
Stressor Identification Guidance Document
Table 7-19. PAH concentrations at nearest upstream location and locations of
impairments (mg/kg). (Hyalella azteca sediment effects concentrations, PEL and TEL,
normalized to sediment WET weight.)
Chemical
PEL TEL
Benzo(a)pyrene (BAP)
0.32 0.03
Naphthalene (NAPH)
0.14 0.02
Fluorene
0.15 0.01
Phenanthrene
0.41 0.02
Anthracene
0.17 0.03
Fluoranthene
0.32 0.04
Pyrene
0.49 0.02
Benzo[a]anthracene
0.28 0.03
Chrysene
0.41 0.02
Benzo(g,h,i)perylene
0.25 0.01
PAH sediment concentration
Nearest
Upstream
Location
(0)
[0.053] #
(0)
[0]
(0)
[0]
(0)
[0.053] #
(0)
[0]
(0)
[0.097] #
(0)
[0.076] #
(0)
[0.043] #
(0)
[0.065] #
(0)
[0.044] #
Impairment
A
(0)
[0.043] #
(0)
[0]
(0)
[0.059] #
(0)
[0.11]#
(0)
[0.037] #
(0)
[0.2] #
(0)
[0.16]#
(0)
[0.059] #
(0)
[0.081 ]#
(0)
[0.03] #
Impairment
B
/141.1\*
(14.8)*
[210]*
/22.9\*
(0)
[260]*
(0)
[830] *
(0)
[230] *
(0)
[100]*
(8.2) *
[100]*
(0)
[810]*
(8.2) *
[310]*
(16.5)*
[390] *
(49.5) *
[150]*
Impairment
C
/125\*
(15.8)*
[14]*
/70\*
(4.6) *
[18]*
/200\ *
(7)*
[20]*
/470\ *
(24.1)*
[38]*
/360\ *
(27.1)*
[41]*
/550\ *
(37.6) *
[44]*
/405\ *
(23.8) *
[32]*
/185\*
(16.5)*
[42]*
/305\ *
(20.8) *
[13]*
/65\*
(1 1 .2) *
[19]*
(*) exceeds PEL and TEL; (#) exceeds TEL.
Zero = below detection; No Entry = No data
/\= 1987-1991, ( ) = 1992, [] = 1998.
for that year
7-64
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Table 7-20. Metals concentrations at nearest upstream location and locations of
impairments (mg/kg). (Hyalella azteca sediment effects concentrations, PEL and TEL,
normalized to sediment wet weight.)
Chemical
PEL TEL
As
48.4 10.8
Cd
3.2 0.58
Cr
119.4 32.3
Cu
101.2 28
Pb
81.7 37.2
Zn
544 98.1
Nearest
Upstream
Location
/5\
(5)
[8.3]
/0.5\
(0.5)
[0]
/5.8\
(7.3)
[14.3]
(7.4)
[24.1]
(12.1)
[20.4]
(30.6)
[81 .4]
Impairment
A
/8\
(12.4)#
[6]
/0.5\
(0.5)
[0.2]
/13.2\
(13.6)
[8.9]
(17.2)
[22.9]
(19.1)
[24]
(79)
[66.6]
Impairment
B
/11.2\#
(0)
[10.8] #
/1.8\#
(0.5)
[0.1]
/47.6\#
(208) *
[32.3] #
/68\#
(79) #
[39.2] #
/170\*
(1 72) *
[46.4] #
/187\#
(1 73) #
[89.2]
Impairment
C
/8\
(13.8)#
[9.8]
/1\
(0.5)
[2]#
/39.2\#
(60.9) #
[50.4] #
(56) #
[133]*
/59.5\#
(84.6) *
[220] *
(141)#
[280] #
(*) exceeds PEL and TEL; (#) exceeds TEL. *ND= not detected, NA = not available, / \:
1987-1991,( ) = 1992, [ ] = 1998. Zero = below detection; No Entry = No data for that
year
Chapter 7: Little Scioto River, Ohio
7-65
-------
APPENDIX A
OVERVIEW OF WATER
MANAGEMENT PROGRAMS
SUPPORTED BY THE SI
-------
Stressor Identification Guidance Document
Appendix A
Overview of Water
Management Programs
Supported by the SI
The following sections describe several major water management programs and how the
SI process can support them.
A.1 Water Quality Assessment Reports Under CWA Section 305(b)
In 1987, EPA's Office of Water recommended that regulatory authorities increase the
use of biological monitoring to better characterize aquatic systems. State and Tribal
agencies were directed to protect the fishable and swimmable goals of the Clean Water
Act. Under Section 305(b), States, Territories, the District of Columbia, interstate water
commissions, and participating American Indian Tribes are required to assess and report
on the quality of their waters (USEPA 1997). The results of 305(b) assessments are not
raw data, but rather are statements about the degree to which each waterbody supports
the uses designated in state or tribal water quality standards. Each State and Tribe
aggregates these assessments and extensive programmatic information in a 305(b) report,
which is a detailed document usually including information from multiple agencies.
EPA then uses individual 305(b) reports to prepare a biennial National Water Quality
Inventory Report to Congress. This report is the primary vehicle for informing Congress
and the public about water quality conditions in the United States.
Most of the information contained in 305(b) assessments is based on data collected and
evaluated by states, tribes, and other jurisdictions over the two-year period immediately
preceding issuance of the report. The Report to Congress contains national summary
information about water quality conditions in rivers, lakes, estuaries, wetlands, coastal
waters, the Great Lakes, and groundwater. The report also contains information about
public health and aquatic ecosystem concerns, water quality monitoring, and state and
federal water pollution management programs.
States and Tribes base their 305(b) water quality determinations on whether waterbodies
are clean enough to support basic uses, such as aquatic life, swimming, fishing, and
drinking supply. These uses, along with appropriate national criteria and anti-
degradation statements, are part of the water quality standards set by each state or tribe
to protect its waters. These standards must be approved by EPA.
Water quality for each individual use is rated as either:
* Good/Fully Supporting
* Good/Threatened
> Fair/Partially Supporting
> Poor/Not Supporting
> Poor/Not Attainable
Appendix A: Overview of Water Management Programs A-1
-------
Stressor Identification Guidance Document
For waterbodies with more than one use, information is consolidated into a summary use
support designation of general water quality conditions. These uses are characterized as
either:
> Good/Fully Supporting All Uses
* Good/Threatened for One or More Uses
* Impaired for One or More Uses
Once a state or tribe has determined, under section 305(b), that a waterbody is impaired
for one or more uses, the state or tribe is required to identify the source and cause of
impairment. Some causes are much easier to identify than others. For example, a case
where impairment is caused by a specific chemical from a point source discharge might
be straightforward and easily analyzed. Monitoring programs, however, must deal with
impacts caused not only by chemical toxicity, but also conventional pollutants (e.g.,
temperature, pH and dissolved oxygen) and anthropogenic pollutants from non-point
sources. Monitoring agencies need the ability to evaluate the relative impact that a
particular pollutant or other stressor has on the biological integrity of a receiving water.
A.2 303(d) Lists and TMDLs
Section 303 of the 1972 Clean Water Act requires States, Territories and authorized
Tribes to establish water quality standards and Total Maximum Daily Loads TMDLs) for
EPA review and approval. Water quality standards identify the uses for each waterbody
(e.g., drinking water supply, contact recreation, aquatic life support) and the water
quality criteria to support that use. Water quality criteria can be either numeric (e.g., no
more than 10 |-ig/L of copper) or narrative (e.g., nutrients are not to exceed levels which
cause an imbalance of aquatic flora and fauna). Water quality standards also include
antidegradation policies to prevent deterioration of existing high quality waters.
Under Section 303(d), States, Territories and authorized Tribes must identify impaired
waters and establish TMDLs for these waters. Impaired waters are those that do not
meet applicable water quality standards, even after point sources of pollution have
installed the minimum required levels of pollution control technology. States, Territories
and authorized Tribes are required to submit their list of impaired every two years.
States, Territories and authorized Tribes are required to establish priority rankings for
impaired waters on the 303(d) lists and develop TMDLs for these waters. A TMDL
specifies the maximum amount of a pollutant that a waterbody can receive and still meet
water quality standards, and allocates pollutant loadings among point and nonpoint
pollutant sources. EPA must approve or disapprove lists and TMDLs established by
States, Territories and authorized Tribes. If a State, Territory or authorized Tribe
submission is inadequate, EPA must identify the impaired waters and establish the
TMDL.
TMDLs are a critical component of the water quality program. They provide the analytic
underpinning for watershed decisions and promote integrated program planning,
implementation, and funding. For example, controlling sediment and/or nutrient
loadings can protect aquatic habitat, wetlands, endangered species, and drinking water
sources. As requirements are strengthened and public communication emphasized,
sound procedures for identifying stressors and management solutions will become more
important.
A-2 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Development of a TMDL varies based on numerous factors including environmental
setting, waterbody type, source type/behavior, and pollutant type/behavior. However,
TMDL development generally includes the following activities:
1. Problem Identification: characterization of the impairment and
identification of the pollutant causing the impairment;
2. Identification of Water Quality Targets: establishment of the TMDL
endpoint or target value, which is typically the applicable numeric water
quality criterion or a numeric interpretation of the narrative water quality
standard;
3. Source Assessment: estimation of the point, nonpoint and background
sources of pollutants of concern, including magnitude and location of
sources;
4. Allocations: identification of appropriate wasteload allocations for point
sources and load allocations for nonpoint sources;
5. Link Between Numeric Target(s) and Pollutant(s) of Concern: Analysis
of the relationship between numeric target(s) and identified pollutant
sources. For each pollutant, describes the analytical basis for conclusion
that sum of wasteload allocations, load allocations, and margin of safety
does not exceed the loading capacity of the receiving water(s).
6. Calculation of the explicit or implicit margin of safety for each pollutant
and description of accounting for seasonal variations and critical
conditions in the TMDL.
A.2.1 Causes for Impairment: Pollutants and Pollution
Waterbodies are impaired by a variety of stressors. Recent data indicate that the top
causes for impairment include sedimentation/siltation/turbidity and suspended solids
(16%), nutrients (13%), pathogens (13%), and dissolved oxygen ( 10%). These stressors
are often associated with sources or activities that fall under the Clean Water Act
definition of pollutant, or pollution. Pollution is defined in Section 502(19) as the "man-
made or man-induced alteration of the chemical, physical, biological, and radiological
integrity of water."
Section 303(d) requires the identification and listing of all impaired waterbodies
regardless of the origin or source of the pollution or pollutant. Current regulations
require that TMDLs be calculated only for pollutants. Pollutants are defined in Section
502(6) as "dredged spoil, solid waste, incinerator residue, sewage, garbage, heat, and
industrial, municipal, and agricultural waste discharged into water."
Both pollution and pollutants are "stressors" that can be identified and evaluated using
the SI process. Under current regulations, those calculating TMDLs will benefit directly
from guidance on identifying stressors considered pollutants under the Clean Water Act.
The SI guidance can also assist in establishing the causal linkage between a pollutant and
the biological impairment, and thus provide a basis for the development of a TMDL. For
example, if a pollutant causes ecosystem changes that alter the fish community, the
Appendix A: Overview of Water Management Programs A-3
-------
Stressor Identification Guidance Document
altered biological community is an impairment that can be traced to a pollutant for which
a TMDL can be calculated.
A.2.2 EPA Actions to Implement the TMDL Program
In an effort to speed the Nation's progress toward achieving water quality standards and
improving the TMDL program, EPA began, in 1996, a comprehensive evaluation of
EPA's and the states' implementation of their Clean Water Act section 303(d)
responsibilities. EPA convened a committee under the Federal Advisory Committee Act,
composed of 20 individuals with diverse backgrounds, including agriculture, forestry,
environmental advocacy, industry, and state, local, and tribal governments. The
committee issued its recommendations in 1998. These recommendations were used to
guide the development of proposed changes to the TMDL regulations, which EPA issued
in draft in August, 1999. After a long comment period, hundreds of meetings and
conference calls, much debate, and the Agency's review and serious consideration of
over 34,000 comments, the final rule was published on July 13, 2000. However,
Congress added a "rider" to one of their appropriations bills that prohibits EPA from
spending FY2000 and FY2001 money to implement this new rule. The current rule
remains in effect until 30 days after Congress permits EPA to implement the new rule.
TMDLs continue to be developed and completed under the current rule, as required by
the 1972 law and many court orders. The regulations that currently apply are those that
were issued in 1985 and amended in 1992 (40 CFR Part 130, section 130.7). These
regulations mandate that states, territories, and authorized tribes list impaired and
threatened waters and develop TMDLs.
A.2.3 Stressor Identification and the TMDL Program
EPA developed the SI process to assist water resource managers in identifying and
delineating stressors causing biological impairments to waterbodies. While not all water
quality impairments listed under 303(d) are linked directly to biological components of
waterbodies, a sample of submittals from 19 states indicate that approximately one-half
of waterbodies listed as impaired under 303(d) are not meeting biological designated
uses (e.g., aquatic life, cold water fishery). The SI process will have direct utility to
States, Tribes, and EPA by providing sound approaches to evaluating the causes of
biological impairments under the TMDL Program.
As used in the SI process, the term Stressor is synonymous with the terms pollutant and
pollution which, under Section 303(d), are considered causes of impairment. The
identification of pollutant stressors resulting in biological impairment to waterbodies,
and the diagnostic evaluation of the sources of these stressors, is an essential first step in
calculating Total Maximum Daily Loads under Section 303(d) of the Clean Water Act.
For pollution stressors (e.g., habitat degradation, water control structures), for which
TMDLs are not calculated, SI results can be used to identify the sources of the pollution
for use in alternative watershed management activities.
A.3 State/Local Watershed Management
Since 1991, EPA has promoted a watershed protection approach to help address the
nation's remaining water resource challenges (USEPA 1991a). The watershed approach
is an integrated, holistic strategy for protecting and managing surface water and
groundwater resources by watershed, a naturally defined hydrologic unit. For any given
watershed, the approach considers not only the water resource; such as a stream, river,
A-4 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
lake, estuary, or aquifer; but all of the land from which water drains into that resource.
The watershed approach uses all aspects of water resource qualityphysical (e.g.,
temperature, flow, mixing, habitat); chemical (e.g., conventional and toxic pollutants,
such as nutrients and pesticides); and biological (e.g., health and integrity of biotic
communities, biodiversity). EPA's Office of Water has worked to orient and coordinate
point source, non-point source, surface water, wetlands, coastal, groundwater, and
drinking water programs within a watershed context.
The watershed approach is not a program but a way to organize programs, so that the use
of SI will vary with the program conducting the investigation. The watershed approach,
however, can facilitate an SI investigation since information is already integrated from
various sources, such as point source discharges and non-point source runoff. This
integrated information can help investigators make sense of disturbances through
knowledge of potential sources of stressors that might feed into that location or might
affect the food source or some other essential ecosystem component by affecting the
natural continuum (Vannote et al. 1980).
The challenge for identifying stressors for watershed-based programs is proper scaling.
Even though the SI may be initiated by a program using the watershed approach, the
impairment may not be watershed wide. Impairment to the biological system may be
difficult to determine on a watershed scale. Similarities among biota tend to follow
ecoregions, rather than watersheds. Several ecoregions may exist within a watershed,
especially where elevation differences are great. The biota within any given ecoregion
may respond differently to a given stressor than the biota within a neighboring ecoregion.
Accurate scaling of the problem is important any time a biological impairment is found,
but especially with the watershed approach, to ensure that the information is used to full
advantage in identifying and characterizing stressors.
A.4 Non-point Source 319 Management
The 1987 Water Quality Act Amendments to the Clean Water Act added section 319,
which established a national program to assess and control non-point source (NFS)
pollution. Under this program, states and tribes are asked to assess their NFS pollution
problems and submit their assessments to EPA. The assessments included a list of
navigable waters within the State or Tribal Territories, which without additional action
to control NFS pollution, cannot reasonably be expected to attain or maintain applicable
water quality standards or the goals and requirements of the Clean Water Act. Section
319 also requires identification of categories and subcategories of NFS pollution that
contribute to impairment of waters, descriptions of procedures for identifying and
implementing best management practices, control measures for reducing NFS pollution,
and descriptions of State, Tribal, and local programs used to abate NFS pollution.
NFS programs need to identify and control NFS pollutants. Since NFS pollutants can be
difficult to trace, identifying the source of these pollutants is probably the greatest
challenge for NFS programs. The SI process can help investigators obtain greater
confidence that stressors have been accurately identified. Attributing responsibility to a
particular source can be very straightforward and obvious or very difficult. Mechanisms
used to attribute responsibility need to be assessed for each situation, and common sense
should be used. For example, runoff may be obviously coming from one farm. In
another situation, runoff may encounter multiple potential sources of pollution, including
a poultry farm, a cattle feedlot, and an abandoned mine. In the latter situation, if nutrient
Appendix A: Overview of Water Management Programs A-5
-------
Stressor Identification Guidance Document
loading is the identified stressor, attributing responsibility between the poultry farm and
cattle feedlot may be difficult, but ruling out the abandoned mine would be simple.
A.5 Permitting Programs
A.5.1 NPDES Permits
All discrete sources of wastewater are required to obtain a National Pollutant Discharge
Elimination System (NPDES) permit (or State equivalent) that regulates the facility's
discharge of pollutants. This approach to controlling and eliminating water pollution is
focused on pollutants determined to be harmful to receiving waters and sources of such
pollutants. Authority for issuing NPDES permits is established under Section 402 of the
CWA. A summary of the Water Quality-based "Standards to Permits" Process for
Toxics Control (adapted from the Technical Support Document for WQ-based Toxics
Control, TSD, USEPA 1991a) lists nine steps:
1. Define water quality objectives, criteria, and standards;
2. Establish priority waterbodies;
3. Characterize effluent - chemical-specific or Whole Effluent Toxicity
(WET);
a) evaluate for excursions above standards,
b) determine reasonable potential, and
c) generate effluent data;
4. Evaluate exposure (critical flow, fate modeling, and mixing) and
calculate wasteload allocation;
5. Define required discharge characteristics by the waste load allocation;
6. Derive permit requirements;
7. Evaluate toxicity reduction and/or investigate indicator parameters (as
needed, for permits containing WET monitoring or limits);
8. Issue final permit with monitoring requirements - average monthly and
maximum daily average weekly for publicly operated treatment works)
limits; and
9. Track compliance.
Sometimes the monitoring requirements include biological assessment of the receiving
water. The permit can contain a reopener clause to allow the limits and monitoring
requirements to be adjusted if biological impairment is found in the receiving water.
The SI guidance is somewhat analogous in function to the Toxicity Reduction Evaluation
(TRE) and Toxicity Identification Evaluation (TIE) guidance used in Step 7 above
(USEPA 1988a,b,c, 1991b, 1993a,b). In the permitting process, toxicity is controlled
through limits for specific chemicals and limits for whole effluent toxicity. When permit
monitoring shows that an effluent has toxicity above the amount allowed by the permit,
the discharger is often required to conduct a TRE to determine if a simple solution exists
A-6 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
for reducing the toxicity, e.g., housekeeping procedures for cleaning fluids, or pH
buffering of the effluent. If the solution is not apparent from the TRE, additional TIE
procedures may be required. TIE procedures guide investigators through additional data
collection to determine the toxic component(s) of the waste stream. These procedures
include both aquatic toxicity methods and chemistry methods.
When WET or chemical testing show that the effluent is toxic, this does not mean that an
impairment will necessarily be found in the aquatic biota within the zone of influence of
the discharge. Effluent limits include safety factors in their calculations. The waste load
allocation (Step 4, above) is calculated based on worst-case estimations. For example,
effluent limits for toxicity or for a toxic chemical are based on low-flow conditions in
streams and rivers (often the lowest seven-day flow in a ten-year period). Effluent limits
may be exceeded, a TRE/TIE conducted, and the problem solved without incurring
measurable impairment in the receiving water biota. The current trend is to lessen this
safety buffer by customizing water quality-based permit limits to local conditions
through such mechanisms as dynamic modeling of waste load allocation (USEPA 1991a)
and recalculation of water quality standards or use of the water-effects ratio (USEPA
1994).
Conversely, ambient biological assessments may show impairment in the aquatic biota
below a permitted discharge without a measured permit limit exceedence. The role of
the effluent in causing the impairment is not readily apparent in this case. The effluent
stream could have been toxic during periods when toxic parameters were not being
measured; effluent toxicity tests could have been insufficiently sensitive through
inappropriate selection of test organisms or operator error; or impairment could have
been caused by stressors other than effluent discharge. Accurate attribution of
responsibility can be very critical in NPDES permitting cases, both for fairness and
success in stressor control. A SI should be conducted to distinguish effects caused by the
effluent discharge and effects from other stressors.
A.5.2 Cooling Tower Intake 316(b) Permitting
Under section 316(b) of the CWA, any NPDES permitted discharger which intakes
cooling water must not cause an adverse environmental impact to the waterbody. To
determine if a cooling water intake structure is causing adverse environmental impacts to
the waterbody, the overall health of the waterbody should be known. Where biological
impairments are found, stressor identification procedures should help investigators
identify the different stressors causing the waterbody to be impaired, including the intake
structure. A high degree of certainty is needed.
A. 5.3 Dredge and Fill Permitting
Under Section 401 of the CWA, different types of federal permitting activities (such as
wetlands dredge and fill permitting) require a certification that there will be no adverse
impact on water quality as a result of the activity. This certification process is the 401
Water Quality Certification. Under Section 404 of the CWA, the discharge of dredge
and fill materials into a wetland is illegal unless authorized by a 404 Permit. The 404
Permit must receive a 401 Water Quality Certification.
Stressor identification procedures will help investigators identify the different types of
stress an activity may place on water quality that can then be addressed through
conditions in the 401 Certification. Stressor identification procedures may help to
Appendix A: Overview of Water Management Programs A-7
-------
Stressor Identification Guidance Document
identify unanticipated stress from a dredge and fill activity on water quality or the
biological community after the activity is underway. Stressor identification procedures
may also help in pre-permitting evaluations of the potential impacts of 404 permitting by
assessing different potential stressors on the wetland in advance.
A.6 Compliance and Enforcement
Since 1972, Section 309 of the Clean Water Act has provided statutory authority for a
range of enforcement responses for entities or individuals who fail to comply with the
Act. At the extreme end of this range, actions can result in criminal penalties. EPA has
national and regional programs in place to investigate and prosecute cases. States and
Tribes may have their own compliance and enforcement investigation programs.
A.6.1 Investigations
When a violation occurs, an investigator must first ascertain what must be done to
achieve compliance with the Clean Water Act. Under a Section 309 order, the violator
must come in full compliance with the Clean Water Act; which, under Article 101,
directs the restoration and maintenance of the biological integrity of the nation's waters.
When non-compliance is due to biological impairment or non-attainment of biological
integrity, the investigator must determine the cause of the impairment before
implementing a program to restore biological integrity and achieve compliance. This is a
direct use of the SI process.
The degree of environmental harm is a very important factor that investigators and
judges evaluate when assessing criminal penalties. The SI process should be helpful in
determining whether the causes of impairment are consistent with the causes that would
likely have resulted from the source under investigation. The SI process can also help to
determine the likelihood that one Stressor versus another caused the impairment. In
cases where separation of Stressor mechanisms is fairly clear cut, the SI process can help
investigators determine the significance of the available evidence in determining whether
the alleged Stressor caused the noted environmental harm. However, the SI process is
limited to evaluating causes. If more than one Stressor or source are involved, allocating
the relative contribution of each Stressor or source to the environmental harm may
require additional tools, such as allocation methodologies, that are beyond the scope of
this document.
A.6.2 Enforcement Proceedings
In an enforcement action, the enforcement official seeks for a court to order the
defendant to cease the harmful action, or give injunctive relief. Identifying the causes of
impairment is a crucial step in identifying the actions that would constitute injunctive
relief. The SI process should benefit enforcement officials and expert witnesses by
helping them identify responsible stressors and organize cogent evidence supporting the
identified causal scenario. The SI process adds uniformity to the organization and
analysis of data.
A special program that is often used to grant injunctive relief is the Supplemental
Environmental Project (SEP). Under this program, a judge may allow a defendant to
improve the environment in lieu of paying a portion of a federal fine to the National
Treasury. The environmental benefit gained through an SEP may not directly alter the
harm that the defendant caused originally, but is seen as alternate compensation. For
A-8 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
example, rather than paying a fine of $ 1 million, a defendant might pay a $600,000 fine
and build a bike path with a 30-foot riparian buffer zone (for runoff reduction) along the
impacted creek, or even a neighboring stream.
When the SI process identifies multiple stressors as the cause of impairment, the
information can still be valuable to the SEP program because the alternate stressors may
help direct compensatory action. If, for example, the SI process identifies a stressor
scenario with two stressors working in conjunction and the defendant is responsible for
only one of the two stressors, a judge might approve a plan for the defendant to use
resources to conduct an SEP project that reduces the second stressor, in lieu of a portion
of the fine.
Targeting resources is very important to investigation and enforcement efforts. EPA
often uses 303d lists of impaired waterbodies to target these efforts. The SI process can
supplement the information in the 303d lists so that stressors may be targeted within
targeted waterbodies. Targeting may also be important in assessing future legislative
needs when mechanisms for stressor control are inadequate in national rules and policies,
and in current state and tribal statutes. Targeting stressors for increased control may
identify changes to instigate.
A.7 Risk Assessment
Risk assessment is a scientific process that includes stressor identification, receptor
characterization and endpoint selection, exposure assessment, stress-response
assessment, and risk characterization (USEPA 1998a, Suter 1993). Risk management is
a decision-making process that combines human-health and ecological assessment results
with political, legal, economic, and ethical values to develop and enforce environmental
standards, criteria, and regulations. Risk assessment can be performed on a site-specific
basis, or can be geographically-based (e.g., watershed scale). It can be used to assess
human health or ecological risks.
Results of bioassessment studies can be used in watershed ecological risk assessments to
develop broad-scale empirical models of biological responses to stressors. Such models
can be combined with exposure information to predict risk from specific stressors and
anticipate the success of management actions. Accurate stressor identification is an
integral part of this process and can help ensure that management actions are properly
targeted and efficient in producing the desired results.
A.8 Wetlands Assessments
Although few states have fully incorporated wetlands into water quality standards or
biological assessment programs, a growing number have started to develop biological
assessment methods for wetlands. During the past five years, several state and federal
agencies have independently started to develop bioassessment methods for wetlands.
Minnesota, Montana, North Dakota, and Ohio have been pioneers among the states. The
Biological Resource Division of the U.S. Geological Service, Wetlands Science Institute
of the Natural Resources Conservation Service, and EPA have been the leading federal
agencies.
The SI process and tools specific to wetlands investigations are very much needed by
wetlands managers. In recent 305(b) Reports, states identified sedimentation, nutrient
enrichment, fill and drainage, pesticides, and flow alterations as the major causes of
Appendix A: Overview of Water Management Programs A-9
-------
Stressor Identification Guidance Document
wetlands degradation. Biological assessment methods will allow resource managers to
evaluate the condition of wetlands and may provide some indication of the types of
stressors involved. Once bioassessment methods are completed and incorporated into
monitoring programs, wetlands may be listed as impaired due to biological impairment.
SI methods will be needed to identify stressors causing biological impairment so that
resource managers can better remedy the problems. More information about wetland
bioassessments is available at the EPA Wetlands Division web page
(www .epa. go v/o wow/wetlands).
A.9 Preservation and Restoration Programs
Preservation and restoration programs like the National Estuary Program and the
Superfund Program can also benefit from the SI process.
A. 9.1 National Estuary Program
The National Estuary Program (NEP) was established in 1987 by amendments to the
Clean Water Act to identify, restore, and protect nationally significant estuaries of the
United States. Unlike traditional regulatory approaches to environmental protection, the
NEP targets a broad range of issues and engages local communities in the process. The
program focuses not only on improving water quality in an estuary, but also on
maintaining the integrity of the whole system, its chemical, physical, and biological
properties, and its economic, recreational, and aesthetic values.
The NEP is designed to encourage local communities to take responsibility for managing
their own estuaries. Each NEP is made up of representatives from federal, state and
local government agencies responsible for managing the estuary's resources, as well as
members of the community citizens, business leaders, educators, and researchers.
These stakeholders work together to identify problems in the estuary, develop specific
actions to address those problems, and create and implement a formal management plan
to restore and protect the estuary. Twenty-eight estuary programs are currently working
to safeguard the health of some of our nation's most important coastal waters.
The SI process should be useful to the NEP, and other preservation programs, by helping
stakeholders identify sources and causes of impairments. This information would feed
into the development of a management plan.
A. 9.2 Superfund
The Comprehensive Environmental Response, Compensation, and Liability Act
(CERCLA), commonly known as Superfund, was enacted in 1980 (and amended in
1986) for hazardous waste cleanup. This law created a tax on the chemical and
petroleum industries and provided federal authority to respond to releases or threatened
releases of hazardous substances that may endanger public health or the environment.
The money collected from the taxation went to a trust fund for cleaning up abandoned or
uncontrolled hazardous waste sites. CERCLA also established prohibitions and
requirements for closed and abandoned hazardous waste sites; defined liability of
persons responsible for releases of hazardous waste at these sites; and established
funding for cleanup when no responsible party could be identified.
Since the basis for actions is whether the hazardous substance may endanger public
health or the environment, identifying the stressor(s) causing environmental harm is
A-10 U. S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
important. For cleanup sites where other stressors (e.g., habitat alteration) are also likely
causes of impairment, any cleanup and ecosystem recovery plans would need to take into
account the effects of these stressors. Allocating the amount of responsibility that may
be attributed to each stressor is beyond the scope of the SI process, but knowledge of any
additional stressors that may be causing effects can be valuable in determining expected
outcomes of recovery activities.
Appendix A: Overview of Water Management Programs A-11
-------
APPENDIX B
WORKSHEET MODEL
-------
Stressor Identification Guidance Document
Appendix B
Worksheet Model
The following pages contain a worksheet model that may be used with the SI process.
This is only an example and may not fit every case without alterations.
B.1 Instructions for Using the Worksheet Model
This worksheet follows the SI process outlined in this document. The wortajheet was
designed to be flexible. At certain points, the user will be asked to stop (^g^g) and
consider the evidence gathered thus far, in order to determine whether the process is
complete or requires further analysis. For detailed guidance, the user will need to refer
to the sections of the document that are cited at each step.
1. To begin, write the name of the investigator and date for reference.
2. Fill in the appropriate information in Unit I: List Candidate Causes. To
determine the types of information to include throughout the worksheet,
please refer to the cited sections of the document.
3. Summarize and document the data and analyses in Unit II, Part A. Then,
you may use either of the following options:
* Option 1: Analyze the strongest evidence. If you feel that you have
enough case specific data to eliminate some causes, analyze this data
using Unit II, Part B and proceed to Unit III, Step 1: Eliminate
Alternatives. Note: You may also look at other types of evidence that
can be used for elimination in Unit II, Parts C and D. To do this, fill in
only the blanks in Parts C and D that are designated by the letter E (for
elimination) under the heading Associated Causal Characterization
Method in Unit III. Review this additional evidence to see if it allows
you to eliminate any alternatives.
* If you still have more than one likely causal scenario that could not be
readily eliminated, or if you want to thoroughly review all evidence,
proceed to Unit II, Parts C and D. Complete relevant sections of Parts
C and D for each candidate cause that you listed in Unit I. Then
proceed to Unit IIIand characterize the cause using diagnosis or strength
of evidence, as appropriate (described under #4 below).
* Option 2: List all available evidence in Unit II before going on to Unit
III: Characterize Causes. Using either option, you may still choose to
do additional iterations if the available evidence is insufficient.
* Go to Unit III, Characterize Causes. For those candidate causes listed
in Unit I that were not eliminated while analyzing the evidence listed in
Unit II (i.e., those causes not designated asE in Parts C and D under
the heading Associated Causal Characterization Method in Unit III),
complete Step 1: Eliminate Alternatives and try to further eliminate
Appendix B: Worksheet Model B-l
-------
Stressor Identification Guidance Document
causes. Analyze this evidence carefully; if the evidence is not strong
enough to eliminate a candidate, it still may be useful for the strength of
evidence analysis. Using the worksheet in Unit III, Step 1, determine:
* If the primary cause is so dominant that it masks the effects of others,
then re-evaluate whether the other stressors should be retained. A cause
should not be eliminated if it is potentially masked. Instead, strength of
analysis should be used.
> If only one candidate cause remains, go to Unit IV: Sufficiency of
Evidence. Note: You still may want to look at the diagnostic and
strength of evidence information to strengthen your case. If so, go to
Unit III, Step 2.
> If more than one candidate cause remains, go to Unit III, Step 2 to look
for diagnostic evidence.
> If no candidate causes remain, go to Unit V. You will need to do another
iteration with more information.
* Next, try diagnosis. Look for evidence designated as D under the
column labeled Associated Causal Characterization Method in Unit III
in Unit II, Part C tables. Using the worksheet in Unit III, Step 2,
determine:
> If only one candidate cause remains, go to Unit IV: Sufficiency of
Evidence. Note: You may still want to do a strength of evidence
analysis to strengthen your case. If so, go to Unit III, Step 3.
> If more than one candidate cause remains, go to Unit III, Step 3
(Strength of Evidence Analysis).
> If no candidate causes remain go to Unit Fand do another iteration with
more information.
> Many investigators will want to complete the strength of evidence
analysis even if elimination or diagnosis have identified the stressor.
This part of the SI process helps determine how strong a case an
investigator can make for a particular stressor. Look for evidence
designated as S under the column labeled Associated Causal
Characterization Method in Unit III in Unit II Part C, and also consider
the evidence gathered in Part D. Analyze this evidence carefully using
the worksheet in Unit III, Steps 3, 4, and 5.
* Unit III Steps 3, 4, and 5 allow the investigator to compare evidence,
side-by-side, for candidate causes. The step used depends on the type of
evidence. Scores are assigned to each candidate cause to reflect that
cause's relevance to each causal consideration. (For more detailed
information on comparing stressors, refer to the sections cited in the
worksheets). Compare scores among the candidate causes, and then go
to Unit IV, Sufficiency of Evidence.
B-2 U.S. Environmental Protection Agency
-------
* List the most likely cause in Unit IV, and determine if the evidence is
sufficient for the intended use.
+ If yes, your SI is complete, report results.
+ If no, go to Unit V, Reconsider Impairment.
* Reconsider whether the impairment was real and describe the results.
* If no, your SI is complete, report results.
+ If yes, go to Unit VI, Collect More Data.
* Determine whether all reasonable causes were analyzed.
* If no, complete Unit VI, Follow-on 1 to determine whether additional
scenarios should be analyzed (back to Unit I), or whether the process
should be ended and the results reported as inconclusive.
* If yes, go to Unit VI, Follow-on 2 to determine whether additional data
should be collected and another iteration begun (back to Unit I), or
whether the process should be ended and the results reported as
inconclusive.
-------
Stressor Identification Guidance Document
Stressor Identification Worksheet
Investigator
Date Completed^
UNIT I. LIST CANDIDATE CAUSES
Describe the impairment.
(see Chapter 2. 2)
Make a map. (Unit I part A)
(see Chapter 2. 2)
Define the Scope of the Investigation.
(see Chapter 2. 3)
List the candidate causes
(see Chapter 2. 4)
Develop a conceptual model for the case. (Unit
I, part B)
(see Chapter 2. 5)
Results / Notes
Candidate Causes
# i.
#2.
#3.
Go to Unit II, Analyze Evidence.
B-4
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
UNIT I. LIST CANDIDATE CAUSES
Part A. Make a map to document geographic features relevant to the analysis.
Draw a map or insert map of study area.
Include natural and man-made features such as dams, sources, tributaries, landfills, dredge areas, jetties, sand
bars, waterfalls, wetlands, salt water intrusion, etc. See Chapter 2.2.
Show location of impairment.
Appendix B: Worksheet Model
B-5
-------
Stressor Identification Guidance Document
UNIT I. LIST CANDIDATE CAUSES
Part B. Make a conceptual model of the case.
Draw a conceptual model of the case. See Chapter 2.5.
Include hypothesized sources, stressors and important environmental processes that lead to the impairment.
Label candidate causes.
B-6 U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
UNIT II
Part A. Summarize and document associations between the candidate cause and the effect from the case.
Insert tables, graphs and/or figures of relevant data. See Chapter 3.1.
Insert statistical analyses including correlations, geographic associations, etc. See Chapter 3, textbox 3-2.
You may want to look at other types of evidence that can be used for elimination in Unit II, Part B and C.
If you feel that you have enough case specific data to eliminate some causes, proceed to Unit III
Step 1 (Eliminate Alternatives). If not, proceed to Unit II Part B.
Appendix B: Worksheet Model
B-7
-------
Stressor Identification Guidance Document
UNIT II
Part B. Measurements associated with the causal mechanism (Chapter 3.3).
Evidence can be used for Elimination (E) Diagnosis (D) or Strength of Evidence (S), as noted below.
Prepare a separate table for each candidate cause.
Use this as a reminder of types of data that could be used in the analysis. Not all questions may be appropriate.
Candidate Cause:
Example Questions:
Yes/No/
Question Not
Relevant
Associated Causal
Characterization
Method in Unit
III*
Supporting Analysis
Are symptoms or other responses specific to
or characteristic of a type of stressor found in
organisms from the impaired community?
D, S
Are there internal measures of exposure (e.g.,
body burdens, biomarkers) found in
organisms from the impaired community?
E,D, S
Is an intermediate product of an ecological
process present?
E, S
Do distributions of stressors and receptors
coincide?
E, S
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Example Questions:
Yes/No/
Question Not
Relevant
Associated Causal
Characterization
Method in Unit
III*
Supporting Analysis
Have there been expected changes in the
abundance of predators, prey, or competitors?
Are there expected effects on other receptors?
Other
'E = Elimination; D = Diagnosis; S = Strength of Evidence
If you feel that your evidence can be used to identify the cause through diagnosis, go to Unit III,
Step 2. If not, continue with the analysis of evidence in Unit II Parts C and D.
Appendix B: Worksheet Model
B-9
-------
Stressor Identification Guidance Document
UNIT II
Part C. Associations of effects mitigation with manipulation of causes (Chapter 3.4).
Evidence can be used for elimination ONLY if it is from the site.
Prepare a separate table for each candidate cause.
Use this as a reminder of the types of data that could be used in the analysis. Not all questions may be appropriate.
Candidate Cause:
Questions:
Yes/No/
Information not
available/
Question not
Applicable
Asso-
ciated
Causal
Charac-
teriza-
tion
Method
in Unit
III*
Supporting Analysis
Does elimination of the source reduce or
eliminate the effect?
S,E
Does the introduction of previously
unexposed organisms result in an effect?
Does the isolation of organisms from one
cause reveal the effects of others?
B-10
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Questions:
Does the testing of chemical fractions of site
media result in toxicity being associated with
a particular fraction (i.e., TIE)?
Other
Yes/No/
Information not
available/
Question not
Applicable
Asso-
ciated
Causal
Charac-
teriza-
tion
Method
in Unit
III*
S
Supporting Analysis
'E = Elimination; D = Diagnosis; S = Strength of Evidence
If you have enough data to determine the cause, proceed to Unit III Step 1 (Elimination) or Step
2 (Diagnosis) or Step 3 (Strength of Evidence), as appropriate. If not or uncertain, proceed to Unit
II Part D.
Appendix B: Worksheet Model
B-ll
-------
Stressor Identification Guidance Document
UNIT II
Part D. Using effects data from elsewhere (Chapter 3.2).
Use this table to incorporate data from other situations that support the analysis. Not all questions may be appropriate for a
given candidate cause.
This evidence is applicable to Strength of Evidence (S) characterization method.
Prepare a separate table for each candidate cause.
Candidate Cause:
Type of
Candidate
Cause
Characterization
of Exposure
(Intensity, Time,
and Space)
Data
Available?
Yes (note
location of
data)/No
Exposure-Response
(E-R) Relationship
E-R Available?
Yes (note
location of data)
/No/Not
Relevant
Would effects
be expected at
the
environmental
conditions
seen in the
case?
(Yes/No)
Location of supporting
analysis
Chemical
What is the
concentration in
the medium at the
site?
What is the
concentration-response
relationship (seen in the
lab or the field)?
What is the
internal
concentration in
organisms at the
site?
What is the internal
external concentration-
response relationship
(seen in the lab or the
field)?
What is the
concentration in
the biomarker at
the site?
What is the biomarker-
response relationship?
B-12
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Type of
Candidate
Cause
Effluent
Contaminated ambient
media
Habitat
Water Withdrawal or
Drought
Thermal Energy
Characterization
of Exposure
(Intensity, Time,
and Space)
What is the
dilution of the
effluent at the
location of the
impairment?
What were the
location and time
of collection and
the results of
analyses?
What are the
structural
attributes of the
habitat?
Are hydrograph
readings and
summary statistics
(e.g., 7Q10)
available?
Are temperature
records available?
Data
Available?
Yes (note
location of
data)/No
Exposure-Response
(E-R) Relationship
What are the laboratory
test (i.e., WET) results
from 100% effluent or
diluted effluent?
What are the results of
laboratory tests of
ambient media?
Are empirical models
available that relate
habitat characteristics
to biological responses
What are the results of
instream flow models
(e.g., IFIM)?
What are the thermal
tolerances of the
impacted organisms?
E-R Available?
Yes (note
location of data)
/No/Not
Relevant
Would effects
be expected at
the
environmental
conditions
seen in the
case?
(Yes/No)
Location of supporting
analysis
Appendix B: Worksheet Model
B-13
-------
Stressor Identification Guidance Document
Type of
Candidate
Cause
Siltation
(Suspended)
Siltation
(Bed-load)
Dissolved Oxygen and
Oxygen-Demanding
Contaminants
(e.g., BOD, COD)
Excess Mineral
Nutrients
Characterization
of Exposure
(Intensity, Time,
and Space)
What is the total
suspended solids
(TSS)
concentration?
What is the degree
of embeddedness
and texture of the
silt?
Review the
dissolved oxygen
data (esp.
predawn).
Review the BOD,
COD data from
the source.
What were the
dissolved mineral
nutrient
concentrations?
Data
Available?
Yes (note
location of
data)/No
Exposure-Response
(E-R) Relationship
What is the
concentration-response
relationship (seen in the
lab or field)?
Are empirical models
available to
characterize the
effects?
What is the
concentration-response
relationship (from lab
or other field studies)?
Are there oxygen
demand models that
can be used to predict
effects?
What is the
concentration-response
relationship (from lab
or other field studies)?
Are there
nutrient/eutrophication
models that can be used
to predict effects?
E-R Available?
Yes (note
location of data)
/No/Not
Relevant
Would effects
be expected at
the
environmental
conditions
seen in the
case?
(Yes/No)
Location of supporting
analysis
B-14
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Type of
Candidate
Cause
Nonindigenous
Species
Pathogen
Other
Characterization
of Exposure
(Intensity, Time,
and Space)
Is a nonindigenous
species present or
abundant?
Is a pathogen
present? If so, is it
abundant?
Data
Available?
Yes (note
location of
data)/No
Exposure-Response
(E-R) Relationship
Are ecological models
available to
characterize the
effects?
Are any symptoms or
diseases observed?
E-R Available?
Yes (note
location of data)
/No/Not
Relevant
Would effects
be expected at
the
environmental
conditions
seen in the
case?
(Yes/No)
Location of supporting
analysis
Go to Unit III, Characterize Causes.
Appendix B: Worksheet Model
B-15
-------
Stressor Identification Guidance Document
UNIT III. CHARACTERIZE CAUSES
Step 1. Eliminate Alternatives (Section 4.1.1) and compare supporting evidence where causes were
eliminated.
For each candidate cause indicate Yes, No, No Evidence (NE), or Not Applicable (NA).
If more than one stressor is necessary for a cause to be sufficient (i.e., temperature and dissolved oxygen), indicate
response for each stressor.
Use extra pages for more than 3 candidate causes.
Provide comments as necessary.
Case-Specific
Consideration
Candidate Cause # 1
(Yes / No / NE / NA)
Candidate Cause # 2
(Yes / No / NE / NA)
Candidate Cause # 3
(Yes / No / NE / NA)
Temporal Co-occurrence
Did the effect precede the stressor in time?
(If the effects preceded a proposed cause
and effects are not obscured by another
sufficient cause, then it cannot be the
primary cause.)
Temporal Gradient
Did the effect increase or decrease over
time in association with an increase or
decrease in the stressor?
(If the effect increases or decreases over
time without a corresponding increase or
decrease in the stressor, then the stressor
cannot be the primary cause.)
B-16
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Case-Specific
Consideration
Candidate Cause # 1
(Yes / No / NE / NA)
Candidate Cause # 2
(Yes / No / NE / NA)
Candidate Cause # 3
(Yes / No / NE / NA)
Spatial Co-occurrence
Is there an upstream/downstream
conjunction of candidate cause and effect?
(If the effect occurs upstream of the source
or does not occur regularly downstream,
e.g., is distributed spatially independently
of a plume, sediment deposition areas, etc.,
and effects are not obscured by another
sufficient cause, then the candidate cannot
be the primary cause).
Co-occurrence with Reference Site(s)
Is there a reference site/impaired site
conjunction of candidate cause and effect?
(If the cause occurs at reference sites as
well as the impaired sit, it can be
eliminated.)
Spatial Gradient
Does the effect increase or decrease across
a given region in association with an
increase or decrease in the stressor?
(If the effect increases or decreases over a
given region without a corresponding
increase or decrease in the stressor, then
the stressor cannot be the primary cause.)
Appendix B: Worksheet Model
B-17
-------
Stressor Identification Guidance Document
Case-Specific
Consideration
Candidate Cause # 1
(Yes / No / NE / NA)
Candidate Cause # 2
(Yes / No / NE / NA)
Candidate Cause # 3
(Yes / No / NE / NA)
Biological Gradient Is a decrease in the
magnitude or proportion of an effect seen
along a decreasing gradient of the stressor?
(A constant or increasing level of effect
with decreasing exposure would eliminate
a cause.)
Complete Exposure Pathway,
Question 1: Is there evidence that the
stressor did not co-occur with, contact, or
enter the receptor(s) showing the effect?
(If there is no route of exposure, or, for
appropriate stressors, if tissue burdens or
other measures of exposure were not found
to occur in affected organisms, the cause
may be eliminated.)
Complete Exposure Pathway,
Question 2: Is there evidence that a
necessary intermediate step in the causal
chain of events did not occur?
(If a link in a known chain of events can be
shown to be missing, the cause may be
eliminated.)
B-18
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Case-Specific
Consideration
Experiment, Temporality
Did the effects continue when the candidate
cause was removed (allowing for rates of
recovery)?
(If effects continue despite elimination of
the candidate cause, that cause can be
eliminated.)
Other
Candidate Cause # 1
(Yes / No / NE / NA)
Candidate Cause # 2
(Yes / No / NE / NA)
Candidate Cause # 3
(Yes / No / NE / NA)
After completing Step 1 (above) for each candidate cause listed in Unit I:
If only one candidate cause remains, elimination is definitive. Go to Unit IV.
If more than one candidate cause remains, go back to Unit II, Part B. If Unit II Part B is
complete, go to Unit III Step 2.
If no candidate causes remains, go to Unit V.
Appendix B: Worksheet Model
B-19
-------
Stressor Identification Guidance Document
UNIT III
Step 2. Characterize cause using diagnostic evidence (Section 4.1.2).
If diagnostic evidence was found in Unit n Part D, determine if the evidence is sufficient to define the cause using this
table.
If evidence is not sufficient to diagnose the cause, it may still be used in the strength of evidence in Unit HI Step 3 .
Use extra pages for more than 3 candidate causes.
Candidate Cause
# 1
#2
#3
Type of Diagnostic
Evidence
Description of Evidence
After completing Step 2 for all causes remaining after the elimination step (Step 1):
If diagnosis is definitive. Go to Unit IV.
If diagnosis is uncertain, go back to Unit II Parts B, C and D. If Unit II Parts B, C, and D
are complete, proceed to Unit III Step 3.
B-20
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
UNIT III
Step 3. Analyze strength of evidence (Section 4.1.3) for Case-Specific Considerations.
Use extra pages for more than 3 candidate causes.
Causal
Considerations
and possible
scores
Candidate Cause # 1
Evidence and Literature
Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and Literature
Citation
Score
Co-occurrence
Compatible (+),
Uncertain (0),
Incompatible (),
No evidence (NE)
(The stressor has
either contacted
the affected
organisms, their
food source, or
some parameter
that can affect the
organisms.)
Appendix B: Worksheet Model
B-21
-------
Stressor Identification Guidance Document
Causal
Considerations
and possible
scores
Temporality
Compatible (+),
Uncertain (0),
Incompatible ( ),
No evidence (NE)
(A cause must
always precede its
effects.)
Consistency of
Association
Invariant (++),
In many places and
times (+),
At background
frequencies (-),
No Evidence (NE)
(The repeated
observation of a
similar
relationship of the
effect and
candidate cause in
different places
and times.)
Candidate Cause # 1
Evidence and Literature
Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and Literature
Citation
Score
B-22
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Causal
Considerations
and possible
scores
Biological
Gradient
Strong and
monotonic (+++),
Weak or other than
monotonic (+),
None (-),
Clear association
but wrong sign
Not applicable
(NA)
(The effect
increases in a
regular manner
with increasing
exposure.)
Candidate Cause # 1
Evidence and Literature
Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and Literature
Citation
Score
Appendix B: Worksheet Model
B-23
-------
Stressor Identification Guidance Document
Causal
Considerations
and possible
scores
Complete
Exposure
Pathway
Evidence for all
steps (++),
Incomplete
evidence (+),
Ambiguous (0),
Some steps missing
or implausible (-),
No evidence (NE)
(The stressor co-
occurs with or
contacts the
receptor(s).)
Candidate Cause # 1
Evidence and Literature
Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and Literature
Citation
Score
B-24
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Causal
Considerations
and possible
scores
Experiment
Experimental
studies Concordant
Ambiguous (0),
Inconcordant ( )
No evidence (NE)
(Toxicity tests or
other controlled
experimental
studies
demonstrated that
the candidate
cause can induce
the observed
effect.)
Candidate Cause # 1
Evidence and Literature
Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and Literature
Citation
Score
Appendix B: Worksheet Model
B-25
-------
Stressor Identification Guidance Document
UNIT III
Step 4. Analyze strength of evidence (Section 4.1.3) using Evidence from Other Situations or from
Biological Knowledge.
Use extra pages for more than 3 candidate causes.
Causal
Consideration and
possible scores
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
Plausibility:
Mechanism
Evidence of
Mechanism (++),
Plausible (+),
Not Known (0),
Implausible (-)
(It is plausible that
the effect resulted
from the cause given
what is known about
the biology, physics,
and chemistry of the
candidate cause, the
receiving
environment, and the
affected organisms.)
B-26
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Plausibility:
Stressor- Response
Quantitatively
consistent (+++),
Concordant (+),
Ambiguous (0),
Inconcordant (-), No
evidence (NE)
(Given a known
relationship between
the candidate cause
and the effect, effects
would be expected at
the level of Stressor
seen in the
environment.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
Appendix B: Worksheet Model
B-27
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Consistency of
Association
Invariant (+++), In
most places (++),
In some places (+),
At background
frequency (-),
Not applicable (NA)
(The repeated
observation of the
effect and candidate
cause is similar in
different places and
times.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
B-28
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Analogy: Positive
Analogous cases:
Many or few but
clear (++),
Few or unclear (+),
None (0)
(The hypothesized
relationship between
cause and effect
similar to other well-
established cases.)
Analogy: Negative
Analogous cases:
Many or few but
clear (- -),
Few or unclear
None (0)
(The hypothesized
relationship between
cause and effect is
dissimilar to other
well-established
cases.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
Appendix B: Worksheet Model
B-29
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Specificity of
Cause*
Note: only
applicable if the
cause is plausible or
is consistently
associated with the
effect.
Only possible cause
One of a few (+),
One of many (0), Not
applicable (NA)
(The effect observed
at the site is known
to have only one or a
few known causes.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
B-30
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Experiment
Experimental
studies: Concordant
Ambiguous (0),
Inconcordant
No evidence (NE)
(Toxicity tests or
other controlled
experimental studies
demonstrated that
the candidate cause
can induce the
observed effect.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
Appendix B: Worksheet Model
B-31
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Predictive
Performance
Prediction:
Confirmed specific
or multiple (+++),
Confirmed general
(++), Ambiguous (0),
Failed (- - -),
No evidence (NE)
(The candidate cause
has any initially
unobserved
properties that were
predicted to occur
and the prediction
was subsequently
confirmed at the
site.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
B-32
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
UNIT III
Step 5. Analyze strength of evidence (Section 4.1.3) based on multiple lines of evidence.
Use extra pages for more than 3 candidate causes.
Causal
Consideration and
possible scores
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
Consistency of
Evidence
All consistent (+++),
Most consistent (+),
Multiple
inconsistencies
(...)
(The hypothesized
relationship between
the cause and effect
is consistent across
all available
evidence.)
Appendix B: Worksheet Model
B-33
-------
Stressor Identification Guidance Document
Causal
Consideration and
possible scores
Coherence of
Evidence
Evidence:
Inconsistency
explained by a
credible mechanism
No known
explanation (0)
No entry if all
consistent
(A mechanistic
conceptual model
explains any
apparent
inconsistencies
among the lines of
evidence.)
Candidate Cause # 1
Evidence and
Literature Citation
Score
Candidate Cause # 2
Evidence and Literature
Citation
Score
Candidate Cause # 3
Evidence and
Literature Citation
Score
Compare evidence among the candidate causes, then go to Unit IV to summarize your findings.
B-34
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
IV. SUFFICIENCY OF EVIDENCE (Chapter 4.2)
Most Likely Candidate Cause:
Is Evidence Sufficient for the Management Purpose?
Q YES SI COMPLETE, REPORT RESULTS
D NO GO TO UNIT V, RECONSIDER
IMPAIRMENT
Summary of Characterization
Candidate Cause
# 1.
#2.
#3.
Cause
Reasoning & Confidence
Appendix B: Worksheet Model
B-35
-------
Stressor Identification Guidance Document
V. RECONSIDER IMPAIRMENT
Does Biological Impairment Really Exist?
(Section 5.1)
Reconsider the impairment by auditing the quality of the methods used to generate and manage the data, by using better
analysis tools, and by eliminating any suspicious data or analyses.
Describe Reconsideration:
Were effects real?
D NO SI COMPLETE, REPORT RESULTS.
D YES GO TO UNIT VI, COLLECT MORE INFORMATION.
B-36
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
VI. COLLECT MORE INFORMATION (Section 52)
Were all reasonable causes analyzed?
D NO Go to Follow-on 1.
D YES Go to Follow-on 2.
Follow-on 1: Make sure that all reasonable causes were analyzed.
If additional scenarios are indicated, repeat process, beginning at Unit 1.
If a good faith effort was implemented with reasonable time and resource expenditures, consult management goals
and determine if the process should be ended with inconclusive results.
SI COMPLETED, REPORT RESULTS AS INCONCLUSIVE.
Follow-on 2: Look at the supporting evidence in Unit II, Analyze Evidence.
Prioritize information needs for likely candidate causes, collect new information and repeat the process, beginning
at Unit 1.
If a good faith effort was implemented with reasonable time and resource expenditures, consult management goals
and determine if the process should be ended with inconclusive results.
SI COMPLETED, REPORT RESULTS AS INCONCLUSIVE.
Appendix B: Worksheet Model B-37
-------
APPENDIX C
GLOSSARY OF TERMS
-------
Stressor Identification Guidance Document
Appendix C
Glossary of Terms
Ambient monitoring:
Ambient waters:
Analogy:
Bioassessment
(biological assessment):
All forms of monitoring conducted beyond the
immediate influence of a discharge pipe or injection
well and may include sampling of sediments and living
resources.
water bodies that are in the environment.
a comparison of two things, based on their similarity in
one or more respects. In SI, the criterion of an analogy
refers specifically to similar causes.
evaluation of the condition of an ecosystem that uses
biological surveys and other direct measurements of the
resident biota.
Biocriteria
(biological criteria):
Biogenic:
Biological gradient:
Biomarker:
Body burden:
Candidate cause:
Categorical regression:
numerical values or narrative expressions that describe
the reference biological condition of aquatic
communities inhabiting waters of a given designated
aquatic life use. Biocriteria are benchmarks for
evaluation and management of water resources
produced by biological processes. For example, organic
acids produced by decomposition of plant litter are
biogenic acids.
a regular increase or decrease in a measured biological
attribute with respect to space (e.g., below an outfall),
time (e.g., since a flood), or an environmental property
(e.g., temperature).
contaminant-induced physiological, biochemical, or
histological response of an organism.
the concentration of a contaminant in a whole organism
or a specified organ or tissue.
a hypothesized cause of an environmental impairment
which is sufficiently credible to be analyzed.
regression analysis in which the dependent variable is
defined by a categorical scale rather than as a count or
continuous variable.
Appendix C: Glossary of Terms
C-l
-------
Stressor Identification Guidance Document
Causal analysis:
Causal mechanism:
Causal relationship:
Causal association:
Causal evidence:
Causal inference:
Causal characterization:
Causal considerations:
Cause:
Co-occurren ce:
Coherency of evidence:
a process in which data and other information are
organized and evaluated using quantitative and logical
techniques to determine the likely cause of an observed
condition.
the process by which a cause induces an effect.
the relationship between a cause and its effect.
a correlation or other association between measures or
observations of two entities or processes which occurs
because of an underlying causal relationship.
the results of an analysis of data to reveal an association
between the environmental condition and a candidate
cause.
the component of a causal analysis that is specifically
concerned with the interpretation of the evidence to
determine the most likely cause.
a step in the stressor identification process in which the
proposed cause is described, the evidence for its causal
relationship to the impairment is summarized, and
uncertainties are presented.
logical categories of evidence that are consistently
applied to support or refute a hypothesized cause. A
causal consideration (e.g., biological gradient) is
evaluated using causal evidence (e.g., a regression of
benthic invertebrate diversity against sediment PCB
concentration).
1. that which produces an effect (a general definition).
2. a stressor or set of stressors that occur at an intensity,
duration and frequency of exposure that results in a
change in the ecological condition (a Si-specific
definition).
the spatial co-location of the candidate cause and effect.
the final consideration in a strength of evidence analysis.
If the results of all of the causal considerations in a
strength of evidence analysis are not consistent, they
may still be coherent, if a mechanistic conceptual or
mathematical model explains the apparent
inconsistencies.
C-2
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Complete exposure
pathway:
Concentration-response
model:
Consideration:
Consistency of association:
Consistency of evidence:
Diagnostic analysis:
Diagnostic protocol:
Dilution ratio:
Ecoepidemiology:
Endpoint species:
Eutroph ication:
Experiment:
Expert judgement:
Exposure:
the physical course a stressor takes from the source to
the receptors (e.g., organisms or community) of interest.
(Evidence for a complete exposure pathway is case-
specific and may include measurements such as body
burdens of chemicals, presence of parasites or
pathogens, or biomarkers of exposure.)
a quantitative (usually statistical) model of the
relationship between the concentration of a chemical to
which a population or community of organisms is
exposed and the frequency or magnitude of a biological
response.
see Causal consideration.
the degree to which an effect and candidate cause have
been determined to co-occur in different places or times.
the degree to which the causal considerations in a
strength of evidence analysis are in agreement
concerning a candidate cause.
a type of causal analysis in which effects that are
characteristic of a particular cause are used to determine
whether that candidate cause may be responsible for an
impairment.
a standard procedure for performing a diagnostic
analysis.
the ratio of the stream flow to the wastewater flow
the study of the nature and causes of effects on
ecological systems.
a species that is the object of an assessment or test.
enrichment of a water body with nutrients, resulting in
high levels of primary production, often leading to
depletion of dissolved oxygen.
the manipulation of a candidate cause by eliminating a
source or altering exposure so as to evaluate its
relationship to an effect.
a method of causal inference based on the knowledge
and skill of the assessors rather than a formal method.
the co-occurrence or contact of a stressor and the
resource that becomes impaired.
Appendix C: Glossary of Terms
C-3
-------
Stressor Identification Guidance Document
Exposure-response
relationships:
Impairment:
Indirect causation:
Indirect effects:
Inferential logic:
Initial response:
Intermediate processes:
Internal exposure:
Logic of abduction:
Mechanism:
Necropsy:
Negative evidence:
Opportun istic:
Pathogens:
a qualitative or quantitative (usually statistical) model of
the relationship between an exposure metric (e.g., the
concentration of a chemical or the abundance or an
exotic species) to which a population or community of
organisms is exposed and the frequency or magnitude of
a biological response.
a detrimental effect on the biological integrity of a water
body that prevents attainment of the designated use.
the induction of effects through a series of cause-effect
relationships, so that the impaired resource may not
even be exposed to the initial cause.
changes in a resource that are due to a series of cause-
effect relationships rather than to direct exposure to a
contaminant or other stressor.
a process for reasoning from the evidence to a necessary
and specific conclusion.
the response of an organism, population or community
to direct exposure to a stressor.
processes that occur between the occurrence of a
stressor in an ecosystem and the induction of the effect
of concern. For example, the reduction in algal
abundance is an intermediate process between the
introduction of a non-native filter feeder and the
reduction in abundance of native planktivorous species.
exposure of an organism to bioaccumulated
contaminants.
inference from data to the hypothesis that best accounts
for the data.
the process by which a system is changed.
a post-mortem examination or inspection intended to
determine the cause of death or the nature of
pathological changes.
evidence that tends to refute a candidate cause.
having the ability to exploit newly available habitats or
resources.
organisms that are capable of inducing a disease in a
susceptible host.
C-4
U.S. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Plausibility:
Positive evidence:
Predictive performance:
Principal cause:
Pseudoreplication:
Publicly Owned Treatment
Works (POTW):
Receptors:
Replicate:
Source:
Spatial gradient:
Specificity:
Specificity of cause:
Specificity of effect:
the degree to which a cause and effect relationship
would be expected, given known facts.
evidence that tends to support a candidate cause.
the degree to which a candidate cause has led to
predictions concerning conditions in the receiving
system which have been subsequently confirmed by
observation or measurement.
the cause that makes the largest contribution to the
effect.
the treatment of multiple samples that are subject to the
same treatment as replicates for statistical purposes. For
example, multiple samples of benthic invertebrates taken
in a channelized stream are pseudo- replicates because
they are not independent. True replicates would be
taken from different channelized streams.
a water treatment facility, as defined by Section 212 of
the Clean Water Act, that is used in the storage,
treatment, recycling, and reclamation of municipal
sewage or industrial wastes of a liquid nature and is
owned by a municipality or other governmental entity.
It usually refers to sewage treatment plants.
organisms, populations, or ecosystems that are exposed
to a contaminant or other stressor.
(a) one of a set of independent systems which have been
randomly assigned a treatment; or (b) to generate a set
of such systems.
an origination point, area, or entity that releases or emits
a stressor. A source can alter the normal intensity,
frequency, or duration of a natural attribute, whereby the
attribute then becomes a stressor.
a graded change in the magnitude of some quantity or
dimension measured on a transect
the quality of being specific rather than general.
only one candidate cause or a few similar causes can
induce the observed effect.
one type of effect is characteristically induced by a
candidate cause. The absence of that effect is evidence
for eliminating the candidate cause.
Appendix C: Glossary of Terms
C-5
-------
Stressor Identification Guidance Document
Strength-of-evidence
analysis:
Strength of association:
Stressor:
Supplemental
Environmental Project
(SEP):
Symptomatology:
Temporal relationship:
Temporal gradient:
Total Maximum Daily
Load (TMDL):
Toxicity Reduction
Evaluation (TRE):
Toxicity Identification
and Evaluation (TIE):
an inferential process that uses all relevant evidence in a
systematic process to determine which candidate cause
is most likely to have induced the effect of concern.
the size of the effect produced by an increment in the
candidate cause. A candidate cause that is associated
with a large change in the level of effect is more likely
to be the true cause than one that is weakly associated.
any physical, chemical, or biological entity that can
induce an adverse response.
a special program that is often used to grant injunctive
relief.
a set of signs of the action of a causal agent on
organisms. A set of symptoms with a common cause
constitutes a symptomatology.
the relationship between the time of occurrence of a
candidate cause and of the effect of concern.
a graded change in the magnitude of some quantity or
dimension measured over time.
the total allowable pollutant load to a receiving water
such that any additional loading will produce a violation
of water-quality standards.
a site-specific study conducted in a stepwise process
designed to identify the causative agent(s) of effluent
toxicity, isolate the sources of toxicity, evaluate the
effectiveness of toxicity control options, and then
confirm the reduction in effluent toxicity.
a process that identifies the toxic components of an
effluent or ambient medium by a process of chemically
manipulating the effluent or medium and testing the
resulting material.
C-6
U.S. Environmental Protection Agency
-------
APPENDIX D
LITERATURE CITED
-------
Stressor Identification Guidance Document
Appendix D
Literature Cited
Albers, P. 1995. Petroleum and individual polycyclic aromatic hydrocarbons, Pages
330-355 in Hoffman, David J. et al, ed. Handbook of Ecotoxicology, CRC Press,
Boca Raton, Florida.
Adams, D.F. 1963. Recognition of the effects of fluorides on vegetation. /. Air Pollut.
Control Assoc. 13:360-362.
Allan, J.D. 1995. Stream ecology: structure and function of running waters. Chapman
and Hall Publishers, London.
Baumann, P.C., and J.C. Harshbarger. 1995. Decline in liver neoplasms in wild brown
bullhead catfish after coking plant closes and environmental PAHs plummet.
Environ. Health Perspect. 103:168-170.
Baumann, P.C., I.R. Smith, and C.D. Metcalfe. 1996. Linkage between chemical
contaminants and tumors in benthic Great Lakes fish. /. Great Lakes Res.
22(2):131-152.
Beyers, D.W. 1998. Causal inference in environmental impact studies. /. North Amer.
Benthol. Soc. 17: 367-373.
Botti, C., P. Comba, F. Forastiere, and L. Settimi. 1996. Causal inference in
environmental epidemiology: the role of implicit values. Sci. Total Environ.
184:97-101.
Blus, L.J., and C.F. Henny. 1997. Field studies on pesticides and birds: unexpected and
unique relations. Ecological Applications. 7:1125-1132.
Carpenter, S.R., N.F. Caraco, D.L. Correll, R.W. Howarth, A.N. Sharpley, and V.H.
Smith. 1988. Non-point pollution of surface waters with phosphorus and
nitrogen. Ecological Applications 8(3):559-568.
Clemen, R.T. 1986. Making hard decisions. An introduction to decision analysis. 2nd
ed. Duxbury Press, Belmont, CA.
Clements, W.H. 1994. Benthic invertebrate community responses to heavy metals in the
Upper Arkansas River Basin, Colorado. /. N. Am. Benthol. Soc. 13:30-44.
Clements, W.H., and P.M. Kiffney. 1994. Integrated laboratory and field approach for
assessing impacts of heavy metals at the Arkansas River, Colorado.
Environmental Toxicology and Chemistry 13:397-404.
Clements, W.H., D.M. Carlisle, J.M. Lazorchak, and P.C. Johnson. 2000. Heavy metals
structure benthic communities in Colorado mountain streams. Ecol. Appl.
10:626-638.
Appendix D: Literature Cited D-l
-------
Stressor Identification Guidance Document
Cormier, S.M., E.L.C. Lin, F.A. Fulk, and B. Subramanian. 2000a. Estimation of
exposure criteria values for biliary polycyclic aromatic hydrocarbon metabolite
concentration in white suckers (Catostomus commersoni). Environmental
Toxicology and Chemistry 19:1120-1126.
Cormier, S.M., E.L.C. Lin, M.R. Millward, M.K. Schubauer-Berigan, D. Williams, B.
Subramanian, R. Sanders, B. Counts, and D. Altfater. 2000b. Using regional
exposure criteria and upstream reference data to characterize spatial and
temporal exposures to chemical contaminants. Environmental Toxicology and
Chemistry 19:1127-1135.
Cormier, S.M., M. Smith, S. Norton, and T. Neiheisel. 2000c. Assessing ecological risk
in watersheds: a case study of problem formulation in the Big Darby Creek
Watershed, Ohio, USA. Env. Tox. and Chemistry Vol. 19, No. 4(2) 1082-1096.
Courtemanch, D.L., P. Mitnik, and L. Tsomides. 1997. Dec. 8, Memorandum to Greg
Wood, Maine Department of Environmental Protection Licensing Section,
Augusta, Maine.
Davenport, T.E., and M.H. Kelly. 1983. Water resource data and preliminary trend
analysis for the Highland Silver Lake Monitoring and Evaluation Project,
Madison County, Illinois. Phase II. Report No. IEPA/WPC/83-013. Illinois
Environmental Protection Agency, Springfield.
Davies, S.P., L. Tsomides, D.L. Courtemanch, and F. Drummond. 1995. Maine
biological monitoring and biocriteria development program. Maine Department
of Environmental Protection, Augusta, Maine.
Davies, S.P., and L. Tsomides. 1997. Methods for biological sampling and analysis of
Maine's inland waters. DEP-LW/07-A97. Maine Department of Environmental
Protection, Augusta, Maine.
Davies, S.P., L. Tsomides, J.L. DiFranco, and D.L. Courtemanch. 1999. Biomonitoring
retrospective: Fifteen year summary for Maine rivers and streams.
DEPLW1999-26. Maine Department of Environmental Protection, Augusta,
Maine.
Dodds, K., and E.B. Welsh. 2000. Establishing nutrient criteria in streams. The North
American Benthological Society 19:186-196.
Dourson, M.L., L.K. Teuschler, P.R. Durkin, and W.M. Stiteler. 1997. Categorical
regression of toxicity data: A case study using aldicarb. Reg. Toxicol.
Pharmacol. 25:121-129.
Edwards, R., and J. Riepenhoff. 1998. State turns to feds for cleanup. Columbus
Dispatch, April 28, 1998.
Edwards, A.C., H. Twist, and G.A. Codd. 2000. Assessing the impact of terrestrially
derived phosphorus on flowing water systems. Journal of Environmental
Quality. 29:117-124.
D-2 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Edwards, C.J., B.L. Griswold, R.A. Tubb, E.G. Weber, and L.C. Woods. 1984.
Mitigating effects of artificial riffles and pools on the fauna of a channelized
warmwater stream. North American Journal of Fisheries Management 4:194-
203.
Eisler, R. 2000a. Poly cyclic aromatic hydrocarbons. Pages 1343-1411 in Handbook of
Chemical Risk Assessment. Vol. II. Lewis Publishers, Boca Raton, EL.
. 2000b. Handbook of Chemical Risk Assessment. Vol.1. Lewis Publishers,
Boca Raton, EL.
Fox, G.A. 1991. Practical causal inference for ecoepidemiologists. /. Toxicol. Environ.
Health 33:359-373.
Gilbertson, M., T. Kubiak, J. Ludwig, and G. Fox. 1991. Great Lakes embryo mortality,
edema, and deformities syndrome (GLEMEDS) in colonial fish-eating birds:
similarity to chick-edema disease. /. Toxicology and Environmental Health
33:455-520.
Gmur, D.J., and U. Varanasi. 1982. Characterization of benzo[a]pyrene metabolites
isolated from muscle, liver, and bile of a juvenile flatfish. Car cino genesis
5:1397-1403.
Grier, J.W. 1982. Ban of DDT and subsequent recovery of reproduction in bald eagles.
Science 218:1232-1234.
Hackney, J.D., and W.S. Kinn. 1979. Koch's postulates updated: A potentially useful
application to laboratory research and policy analysis in environmental
toxicology. Amer. Rev. Respir. Dis. 1119:849-852.
Hickey, C.W., and W.H. Clements. 1998. Effects of heavy metals on benthic
macroinvertebrate communities in New Zealand streams. Environmental
Toxicology and Chemistry 17:2338-2346.
Hilborn, R., and M. Mangel. 1997. The ecological detective: confronting models with
data. Princeton U. Press, Princeton, NJ.
Hill, A.B. 1965. The environment and disease: Association or causation. Proceed. Royal
Soc. Medicine 58:295-300.
Hilsenhoff, W.L. 1987. An improved biotic index of organic stream pollution. Great
Lakes Entomol. 20:31-39.
Huggins, D.G., and M.F. Moffett. 1988. Proposed biotic and habitat indices for use in
Kansas streams. Report No. 35, Kansas Biological Survey, Lawrence.
Hurlbert, S.H. 1984. Pseudoreplication and the design of ecological field experiments.
Ecological Mono. 54:187-211.
Josephson, J.R., and S.G. Josephson. 1996. Abductive Inference. Cambridge University
Press, Cambridge.
Appendix D: Literature Cited D-3
-------
Stressor Identification Guidance Document
Jorgensen, S.E. 1994. Fundamentals of ecological modelling. Elsevier, Amsterdam.
Kansas Department of Health and Environment (KDHE). 1993. Kansas list of water
quality limited surface waters (303(d) list). Kansas Department of Health and
Environment, Division of Environment, Office of Science and Support, Topeka.,
KS.
. 1998. Kansas water quality limited segments (303(d) list). Kansas
Department of Health and Environment, Division of Environment, Bureau of
Environmental Field Services, Topeka, KS.
. 2000. Kansas water quality assessment (305(b) report). Kansas
Department of Health and Environment, Division of Environment, Bureau of
Environmental Field Services, Topeka, KS.
Karr, J.R., and I.J. Schlosser. 1977. Impact ofnearstream vegetation and stream
morphology on water quality and stream biota. EPA-600-3-77-097. U.S.
Environmental Protection Agency, Environmental Research Laboratory, Athens,
GA.
Kiffney, P.M., and W.H. Clements. 1994a. Structural responses of benthic
macroinvertebrate communities from different stream orders to zinc.
Environmental Toxicology and Chemistry 13:389-395.
. 1994b. Effects of heavy metals on a macroinvertebrate assemblage from a
Rocky Mountain stream in experimental microcosms. /. N. Am. Benthol. Soc.
13:511-523.
Lehman, J.T. 1986. Control of eutrophication in Lake Washington: case study. Pages
301-316 in Ecological knowledge and environmental problem-solving: Concepts
and case studies. National Academy Press, Washington, D.C.
Lin, E.L.C., S.M. Cormier, and J.A. Torsella. 1996. Fish biliary polycyclic aromatic
hydrocarbon metabolites estimated by fixed-wavelength fluorescence:
comparison with HPLC-fluorescent detection. Ecotoxicol. Environ. Safety
35:16-23.
Lin, E.L.C., T.W. Neiheisel, B. Subramanian, D.E. Williams, M.R. Millward, and S.M.
Cormier. Historical monitoring of biomarkers of exposure of brown bullhead in
the remediated Black River, Ohio and two other Lake Erie tributaries. Submitted
to Journal of Great Lakes Research.
Long, E.R., L.J. Field, and D.D. MacDonald. 1998. Predicting toxicity in marine
sediments with numerical sediment quality guidelines. Envir. Toxicol. Chem.
17:714-727.
Meyer, P.P., and L.A. Barclay. 1990. Field manual for the investigation offish kills.
Resource Pub. 177. U.S. Fish and Wildlife Service, Washington, D.C.
Miltner, R.J., and E.T. Rankin. 1998. Primary nutrients and the biotic integrity of rivers
and streams. Freshwater Biology 40:145-158.
D-4 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Mitnik, P. 1994. Presumpscot River waste load allocation. Maine Department of
Environmental Protection, Augusta, Maine.
. 1998. Presumpscot River supplemental report to waste load allocation.
Maine Department of Environmental Protection, Augusta, Maine.
Nelson S.M., and R.A. Roline. 1996. Recovery of a stream macroinvertebrate
community from mine drainage disturbance. Hydrobiologia 339:73-84.
Norberg-King, T., and D. Mount. 1986. Validity of effluent and ambient toxicity tests
for predicting biological impact. Skeleton Creek, Enid, Oklahoma. EPA/600/30-
85-044. Duluth Environmental Research Laboratory, Minnesota.
Norton, S.B. 1999. Using biological monitoring data to distinguish among types of
stress in streams of the Eastern Cornbelt Plains Ecoregion. Ph.D. Dissertation,
Georgetown University, Fairfax, VA.
Norton S.B., S.M. Cormier, M. Smith, and R.C. Jones. 2000. Can biological assessment
discriminate among types of stress? A case study for the eastern cornbelt plains
ecoregion. Environ. Toxicol. Chem. 19(4): 1113-1119.
Ohio Environmental Protection Agency (OEPA). 1988a. Biological criteria for the
protection of aquatic life: Vol. II. users manual for biological assessment of
Ohio surface waters. Division of Water Quality Planning and Assessment,
Ecological Assessment Section, Columbus, OH.
. 1988b. Biological and water quality study of the Little Scioto River
watershed, Marion County, OH. OEPA Technical Report prepared by State of
Ohio Environmental Protection Agency, Division of Surface Water, Columbus,
OH.
. 1989a. Addendum to: Biological criteria for the protection of aquatic life:
Volume II. users manual for biological assessment of Ohio surface waters.
Division of Water Quality Planning and Assessment, Ecological Assessment
Section, Columbus, OH.
. 1989b. Biological criteria for the protection of aquatic life: Volume III
standardized field and laboratory methods for assessing fish and
macroinvertebrate communities. Division of Water Quality Planning and
Assessment, Ecological Assessment Section, Columbus, OH.
. 1989c. Manual of Ohio EPA surveillance methods and quality assurance
practices. Division of Environmental Services, Columbus, OH.
. 1992a. Bottom sediment evaluation, Little Scioto River, Marion, Ohio.
Division of Water Quality Planning and Ecological Assessment Section,
Columbus, OH.
. 1992b. Biological and water quality study of the Ottawa River, Hog
Creek, Little Hog Creek, and Pike Run. OEPA Technical Report EAS/1992-9-7.
Prepared by State of Ohio Environmental Protection Agency, Division of
Surface Water, Columbus, OH.
Appendix D: Literature Cited D-5
-------
Stressor Identification Guidance Document
_. 1994. Biological, sediment, and water quality study of the Little Scioto
River, Marion, Ohio. OEPA Technical Report EAS/1994-1-1. Division of
Surface Water, Ecological Assessment Section, Columbus, OH.
Platt, J.R. 1964. Strong inference. Science 146:347-353.
Popper, K.R. 1968. The logic of scientific discovery. Harper and Row, New York.
Rankin, E.T. 1989. The qualitative habitat evaluation index (QHEI): rationale,
methods, and application. State of Ohio Environmental Protection Agency,
Division of Water Quality Planning and Assessment, Ecological Assessment
Section, Columbus, OH.
. 1995. Habitat indices in water resource quality assessments. Pages 181-
208 in W.S. Davis and T.P. Simon (editors). Biological Assessment and
Criteria. Lewis Publishers, Boca Raton, Florida.
Rankin E., R. Miltner, C. Yoder and D. Mishne. 1999. Association between nutrients,
habitat, and the aquatic biota in Ohio rivers and streams. Ohio EPA Technical
bulletin MAS/1999-1. Ohio EPA, Columbus, OH.
Rosgen, D. 1996. Applied river morphology. Wildland Hydrology Books, Pagosa
Springs, CO.
Roubal, W.T., T.K. Lallier, and D.C. Malins. 1977. Accumulation and metabolism of
C-14 labeled benzene, naphthalene, and anthracene by young coho salmon
(Oncorhynchus kisutch) and starry flounder (Platichthys stellatus). Arch.
Environ. Contam. Toxicol. 5: 513-529.
Russo, R.C. 1985. Ammonia, nitrate and nitrite. Pages 455-471 in G.M. Rand and S.A.
Petrocelli (editors). Fundamentals of Aquatic Toxicology. McGraw Hill,
Washington, D.C.
Sheilds, F.D. Jr., S.S. Knight, and C.M. Cooper. 1998. Rehabilitation of aquatic habitats
in warmwater streams damaged by channel incision in Mississippi.
Hydrobiologica 382:63-86
Smith, V.H., G.D. Tilman, and J.C. Nekola. 1999. Eutrophication: impacts of excess
nutrient inputs on freshwater, marine and terrestrial ecosystems. Environmental
Pollution 100:179-196.
Susser, M. 1986a. Rules of inference in epidemiology. Regulatory Toxicology and
Pharmacology 6:116-186.
. 1986b. The logic of Sir Karl Popper and the practice of epidemiology. Am.
J.Epidemiol. 124:711-718.
. 1988. Falsification, verification and causal inference in epidemiology:
Reconsideration in light of Sir Karl Popper's philosophy. Pages 33-58 in K.J.
Rothman (ed.). Causal Inference. Epidemiology Resources Inc., Chestnut Hill,
MA.
D-6 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Suter, G.W., II. 1990. Use of biomarkers in ecological risk assessment. Pages 419-426
in J.F. McCarthy and L. L. Shugart (eds.). Biomarkers of Environmental
Contamination. Lewis Publishers, Ann Arbor, Michigan.
. 1993. Ecological risk assessment. Lewis Publishers, Boca Raton, FL.
_. 1998. Retrospective assessment, ecoepidemiology, and ecological
monitoring. Pages 177-217 in P. Calow (ed.). Handbook of Environmental Risk
Assessment and Management. Blackwell Scientific, Oxford, UK.
. 1999. Developing conceptual models for complex ecological risk
assessments. Human & Ecolog. Risk Assess. 5:375-396.
Suter, G.W. II, J.W. Gillett, and S. Norton. 1994. Characterization of exposure.
Chapter 4 in Ecological Risk Assessment Issue Papers. EPA/630/R-94/009.
U.S. Environmental Protection Agency, Washington, D.C.
Tarplee, W.H. Jr., D.E. Louder, and A.J. Weber. 1971. Evaluation of the effects of
channelization on fish populations in North Carolina's coastal plain streams.
North Carolina Wildlife Resources Commission, Raleigh, NC.
Thornton, K.W., G.E. Saul, and D.E. Hyatt. 1994. Environmental monitoring and
assessment program assessment framework. EPA/620/R-94/016. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
U.S. Environmental Protection Agency (USEPA). 1988a. Generalized methodology for
conducting industrial toxicity reduction evaluations. EPA/600-2-88/070.
. 1988b. Toxicity reduction evaluation protocol for municipal waste-water
treatment plants. EPA/600/2-88/062.
. 1988c. Methods for aquatic toxicity identification evaluations: phase 1,
toxicity characterization procedures (EPA/600/3-88/034); Phase 2, toxicity
identification procedures (EPA/600/3-88/035); and Phase 3, toxicity
confirmation procedures. EPA/600/3-88/036.
. 1991a. The watershed protection approach: An overview. Office of
Water. Washington, DC. EPA 503/9-92-001.
. 1991b. Methods for aquatic toxicity identification evaluations, phase I
toxicity characterization procedures, 2nd ed. U.S. Environmental Protection
Agency, Office of Research and Development, Environmental Research
Laboratory, Duluth, MN. EPA/600/6-91/003.
. 1993a. Methods for aquatic toxicity identification evaluations, phase II
toxicity identification procedures for samples exhibiting acute and chronic
toxicity. EPA/600/R-92/080. (NTIS: PB94-114907).
Appendix D: Literature Cited D-7
-------
Stressor Identification Guidance Document
1993b. Methods for aquatic toxidty identification evaluations, phase HI
toxicity confirmation procedures for samples exhibiting acute and chronic
toxidty. U.S. Environmental Protection Agency, Office of Research and
Development, Environmental Research Laboratory, Duluth, MN.
EPA/600/R-92/081.
. 1994. Interim guidance on determination and use of water-effect ratios
for metals. U.S. Environmental Protection Agency, Office of Science and
Technology, Washington, DC. EPA/823/B-94/001.
. 1996a. Summary report of workshop on Monte Carlo analysis. Office of
Research and Development, Risk Assessment Forum, Washington, DC.
EPA/630/R-96/010.
. 1996b. Calculation and evaluation of sediment effect concentrations for
the amphipod Hyallela azteca and the midge Chironomus riparius.
Assessment and Remediation of Contaminated Sediments (ARCS) Program.
Great Lakes National Program Office, Chicago, IL. EPA 905-R96-008.
. 1997. Guidelines for preparation of the comprehensive state water quality
assessments (305(b) Reports) and electronic updates: Report contents. U.S.
Environmental Protection Agency, Office of Water. Washington, DC 20460.
EPA-841-B-97-002A.
. 1998a. Guidelines for ecological risk assessment. Office of Research and
Development. Risk Assessment Forum, Washington, D.C. EPA/630/R-95/002F.
. 1998b. 1998 update of ambient water quality criteria for ammonia. Office
of Water, Washington, DC. EPA 822-R-98-008.
. 1999. Report of the workshop on selecting input distributions for
probabilistic assessments. Office of Research and Development, Risk
Assessment Forum, Washington, DC. EPA/630/R-98/004
Vannote, R.L., G.W. Minshall, K.W. Cummins, J.R. Sedell, and C.E. Gushing. 1980.
The river continuum concept. Canadian Journal of Fisheries and Aquatic
Sciences 37:130-137.
Varanasi, U., I.E. Stein, M. Nishimoto, and T. Horn. 1983. Benzo[a]pyrene metabolites
in liver, muscle, gonads, and bile of adult English sole (Parophrys vetulus).
Pages 1221-1234 in Cooke, M. and A.J. Dennis, eds. Polynuclear Aromatic
Hydrocarbons: Formation, Metabolism, and Measurement. Battelle, Columbus,
OH, USA.
Woodman, J.N., and E.B. Cowling. 1987. Airborne chemicals and forest health.
Environ. Sci. 21:120-126.
Yerushalmy, J., and C.E. Palmer. 1959. On the methodology of investigations of
etiologic factors in chronic disease. /. Chronic Disease 10(1):27-40.
D-8 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Yoder, C.O., and E.T. Rankin. 1995a. Biological criteria program development and
implementation in Ohio. Pages 109-144 in W.S. Davis and T.P. Simon, eds.
Biological Assessment and Criteria: Tools for Water Resource Planning and
Decision Making. Lewis Publishers, Boca Raton, FL.
. 1995b. Biological response signatures and the area of degradation value:
New tools for interpreting multi-metric data. Pages 236-286 in Biological
Assessment and Criteria: Tools for Water Resource Planning and
Decisionmaking Lewis Publishers, Boca Raton, FL..
Yount, J.D., and G.J. Niemi. 1990. Recovery of lotic communities and ecosystems from
disturbance; A narrative review of case studies. Environmental Management
14:547-569.
Appendix D: Literature Cited D-9
-------
Stressor Identification Guidance Document
Index
A
Algal growth, 6-6-6-8, 6-10, 6-12, 7-11
Ammonia concentrations, 7-2-7-3, 7-11, 7-14, 7-18-7-22, 7-25-7-31, 7-39-7-48, 7-63
Analogy
description, 4-12
Little Scioto River case study, 7-35, 7-38, 7-41, 7-45
Presumpscot River case study, 6-15
Androscoggin River case study, 6-11-6-13
Aquatic life standards, 6-1, 6-3, 6-5, 6-13
Aquatic life use
defined, 1-1
Arkansas River case study, 4-11
B
Beneficial use designation
defined, 1-1
Benthic macroinvertebrates
effects of heavy metal exposure, 4-11
Little Scioto River case study, 7-1-7-8, 7-24
Presumpscot River case study, 6-1-6-18
Biocriteria
defined, 1-1
Biological gradient
Arkansas River case study, 4-11
described, 4-10
Little Scioto River case study, 7-32, 7-36, 7-39, 7-43
Presumpscot River case study, 6-14
Biological integrity
describing impairments, 2-1-2-3
overview of Stressor Identification, 1-3-1-5
role of Stressor Identification process in water management programs, 1-6-1-9
Stressor Identification process, ES-1
using results of Stressor Identification, 1-5-1-6
water quality management, ES-2
Biological oxygen demand
Little Scioto River case study, 7-11, 7-14, 7-18, 7-20, 7-23-7-28, 7-39-7-49,
7-63
Presumpscot River case study, 6-6-6-7, 6-9, 6-12
BOD. See Biological oxygen demand
C
Candidate causes
analyzing evidence, 3-1-3-11
categories of relationships, 3-1-3-2
characterizing causes, 4-1^-18
conceptual models, 2-5-2-7
describing the impairment, 2-1-2-3
key terms, 2-1
listing, 2-4-2-5
Index 1-1
-------
Stressor Identification Guidance Document
overview of Stressor Identification, 1-3-1-5
principal causes, 2-4
scope of the investigation, 2-3-2-4
unlikely stressors, 2-5
using existing lists of stressors, 2-4
Case studies
Androscoggin River, 6-11-6-13
Arkansas River, 4-11
DDT, 5-2
Lake Washington, 4-13
Little Scioto River, 7-1-7-65
Presumpscot River, 6-1-6-18
Causal evidence
associations between measurements of candidate causes and effects, 3-2-3-6,
4-5^-6
associations of effects with mitigation or manipulation of causes, 3-10-3-11, 4-6
causal considerations, 4-9-4-14
confidence evaluation, 4-17^-18
diagnostic analysis, 4-7^-8
eliminating alternatives, 4-3^-7
identifying probable cause, 4-17^4-18
matching evidence with causal considerations, 4-14
measurements associated with the causal mechanism, 3-9-3-10, 4-6
methods for characterization, 4-1^-17
strength of evidence analysis, 4-8^4-17
using effects data from elsewhere, 3-6-3-9
weighing causal considerations, 4-14-4-11
Cause
defined, 2-1
CERCLA. See Comprehensive Environmental Response, Compensation, and Liability
Act
Channelization, 7-6, 7-11, 7-13, 7-24, 7-47, 7-48
Chemical contaminants. See also Toxic compounds
Little Scioto River Case Study, 7-11, 7-14-7-22, 7-24-7-25, 7-27, 7-29-7-30, 7-
32-7-48
Koch's postulates, 4-9
Chemical oxygen demand, 7-11
Chironomus riparius, 7-29
Chlorophyll a, 6-2, 6-10
Class C aquatic life standards, 6-1, 6-3, 6-5, 6-13
Clean Water Act, 1-6, 1-9, A-l-A-11, ES-1
319 program, 1-7, A-5-A-6
404 Permits, 1-8, A-7-A-8
section 303(d), 1-6, A-2-A-4
section 305(b), 1-6, A-1
section 309, A-8
section 316(b), 1-7, A-7
section 319, A-5
section 401, 1-7, A-7-A-8
section 402, 1-7, A-6
section 502, A-3
1-2 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Co-occurrence
Arkansas River case study, 4-11
DDT case study, 5-2
description, 4-10
Little Scioto River case study, 7-32, 7-36, 7-39, 7-43
Presumpscot River case study, 6-14
COD. See Chemical oxygen demand
Coherence of evidence
description, 4-14
Little Scioto River case study, 7-35, 7-38, 7-42, 7-46
Presumpscot River case study, 6-16
Colorado
Arkansas River case study, 4-11
Combined sewer outfalls, 7-11
Community data plot, 3-4
Complete exposure pathway
DDT case study, 5-2
description, 4-1 CM-11
Little Scioto River case study, 7-32, 7-36, 7-39, 7-44
Presumpscot River case study, 6-15
Comprehensive Environmental Response, Compensation, and Liability Act, 1-9, 7-10, A-
10-A-ll
Comprehensive State Water Quality Assessment, 1-2
Conceptual models, 2-5-2-7, 5-2
Consistency of association
description, 4-11
Lake Washington case study, 4-13
Little Scioto River case study, 7-32-7-39, 7-41, 7-43, 7-45
Presumpscot River case study, 6-14-6-15
Consistency of evidence
description, 4-14
Little Scioto River case study, 7-35, 7-38, 7-42, 7-46
Presumpscot River case study, 6-16
Cooling tower intake permitting, A-7
Cooling water intake program, 1-7
Cricotopus sp., 7-9-7-10, 7-19-7-20, 7-30
CSOs. See Combined sewer outfalls
CWA. See Clean Water Act
D
Data Quality Assessment, 3-2
Data quality issues, 1-2
Data Quality Objectives process, 3-2
DDT case study, 5-2
Deformities, fin erosion, lesions, tumors and anomalies, 7-1-7-4, 7-8-7-10, 7-20-7-23,
7-27-7-30, 7-33-7-49
DELTA. See Deformities, fin erosion, lesions, tumors and anomalies
Department of Natural Resources (Maryland)
website, 2-4
Dissolved oxygen
Little Scioto River case study, 7-3, 7-11, 7-14, 7-20, 7-23-7-31, 7-39-7-49, 7-63
Presumpscot River case study, 6-6-6-10, 6-13, 6-17
Index 1-3
-------
Stressor Identification Guidance Document
DNR. See Department of Natural Resources
DO. See Dissolved oxygen
DQA. See Data Quality Assessment
DQO. See Data Quality Objectives process
Dredge and fill permitting, A-7-A-8
E
Eastern Corn Belt Plains, 7-30
Ecological Risk Assessment, 1-2
Edmondson, W.T., 4-13
Effect
defined, 2-1
Elimination of alternatives, 4-3-4-7, 6-8-6-11, 7-26-7-27
EMAP. See Environmental Monitoring and Assessment Program
Enforcement actions
EPA responsibilities, A-8-A-9
role of Stressor Identification process, 1-8
Environmental Monitoring and Assessment Program, 2-4
EPA. See U.S. Environmental Protection Agency
Ephemeroptera-Plecoptera-Trichoptera, 6-1, 6-5-6-6, 7-9
EPT. See Ephemeroptera-Plecoptera-Trichoptera
EROD. See Ethoxy resorufin[O]deethylase
Ethoxy resorufin[O]deethylase, 7-5, 7-24
Eutrophication, 6-6
Experiments
Arkansas River case study, 4-11
DDT case study, 5-2
description, 4-12
Lake Washington case study, 4-13
Little Scioto River case study, 7-35-7-36, 7-38-7-39, 7-41, 7-44-7-45
Presumpscot River case study, 6-14-6-15
Expert judgment, 4-1
Exposure
defined, 2-1
F
False positives, 5-1
Federal Advisory Committee Act, A-4
Field experiments
types of, 3-10
Fill permitting, A-7-A-8
Fish kills
diagnostic protocols, 4-7
Floe. See TSS with floe
G
Glossary of terms, C-l-C-6
H
Habitat degradation
Little Scioto River case study, 7-11, 7-13, 7-21, 7-24-7-28, 7-32-7-36
Presumpscot River case study, 6-8, 6-11, 6-12, 6-14-6-17
1-4 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
Heavy metals. See Metals
Hyalella azteca, 7-29, 7-33, 7-64-7-65
I
IBI. See Index of Biotic Integrity
ICI. See Invertebrate Community Index
Impoundment, 6-7-6-8, 6-10-6-12, 6-14-6-17
Index of Biotic Integrity, 2-2, 7-1, 7-4-7-9, 7-13, 7-19-7-20, 7-47
Invertebrate Community Index, 2-2, 7-1, 7-4-7-9, 7-19-7-20, 7-47
K
Kansas Biotic Index, 3-11
Kansas Department of Health and Environment
water quality documentation, 3-11
KBI. See Kansas Biotic Index
KDHE. See Kansas Department of Health and Environment
Koch's postulates, 4-9
L
Lake Washington case study, 4-13
Landfills, 7-6, 7-10
Little Scioto River case study
analyzing evidence for diagnosis, 7-28
characterizing causes, 7-26-7-28
comparing strength of evidence, 7-28-7-31
conceptual model of candidate causes for stressor identification, 7-12
discussion, 7-48-7-49
eliminating alternatives, 7-13-7-26
evidence of impairment, 7-5-7-10
executive summary, 7-1-7-4
fish metrics, 7-54
identifying probable causes, 7-47-7-48
introduction, 7-4-7-5
list of candidate causes, 7-10-7-13
macroinvertebrate metrics, 7-55
map, 7-6
metals concentrations, 7-61-7-62, 7-65
PAH concentrations, 7-64
QHEI metrics, 7-56
sediment organic compounds concentrations, 7-57-7-60
strength of evidence analysis, 7-28-7-46
water chemistry parameters, 7-63
M
Macroinvertebrate biotic index, 3-11
Macroinvertebrates. See Benthic macroinvertebrates
Maine
Presumpscot River case study, 6-1-6-18
Maine Department of Environmental Protection, 6-3, 6-18
Maps
describing impairments, 2-2-2-3
Little Scioto River case study, 7-6
Index 1-5
-------
Stressor Identification Guidance Document
Presumpscot River case study, 6-4
Maryland
Department of Natural Resources website, 2-4
Mayflies, 7-9-7-10
MBI. See Macroinvertebrate biotic index
MDEP. See Maine Department of Environmental Protection
Mechanisms
description, 4-12
Little Scioto River case study, 7-33, 7-37, 7-40, 7-44
Presumpscot River case study, 6-15
Mechanistic conceptual models, 3-9-3-10
Metals
Arkansas River case study, 4-11
Little Scioto River case study, 7-2-7-3, 7-11, 7-14, 7-17, 7-19, 7-22, 7-24-7-38,
7-61-7-62, 7-65
Presumpscot River case study, 6-13
Midges. See Tanytarsini midges
MIWB. See Modified Index of Well-being
Modified Index of Well-being, 7-8
Modified Warmwater Habitat, 7-5, 7-7
Monte Carlo simulation, 4-17
MWH. See Modified Warmwater Habitat
N
National Estuary Program, 1-9, A-10
National Pollutant Discharge Elimination System permit program
monitoring requirements, A-6-A-7
role of Stressor Identification process, 1-7
National Water Quality Inventory Report to Congress, A-l
NEP. See National Estuary Program
Nitrates, 7-14, 7-18, 7-20, 7-27, 7-30-7-31, 7-63
Nitrification, 3-11
Nitrites, 7-14, 7-18, 7-20, 7-27, 7-30-7-31, 7-63
Nitrogen, 7-2
Non-point source pollution
management under section 319 of the CWA, A-5-A-6
role of Stressor Identification process in control program, 1-7
NPDES. See National Pollutant Discharge Elimination System permit program
NFS. See Non-point source pollution
Nutrients
enrichment, 7-13, 7-23, 7-26-7-28, 7-32-7-36, 7-39-7-49
excess, 6-6-6-7, 6-10, 6-12
loading, 3-11
O
OEPA. See Ohio Environmental Protection Agency
Ohio
Little Scioto River case study, 7-1-7-65
Ohio Environmental Protection Agency, 7-1, 7-5, 7-10
Organic enrichment, 3-11,7-11
Ortho-phosphate, 6-10
1-6 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
P
PAH. See Polycyclic aromatic hydrocarbons
Pathogens
Koch's postulates, 4-9
PEL. See Probable effect levels
Permitting programs, A-6-A-8
pH levels, 7-30
Phosphorous, 7-2-7-3, 7-14, 7-18
Phosphorus, total
Little Scioto River case study, 7-20, 7-27, 7-30-7-31, 7-63
Presumpscot River case study, 6-1-6-18
Plausibility
Arkansas River case study, 4-11
DDT case study, 5-2
description, 4-12
Little Scioto River case study, 7-33, 7-37, 7-40, 7-44-7-45
Presumpscot River case study, 6-15
Pollutants
defined, A-3
Pollution
defined, A-3
Pollution control
measuring effectiveness, 1-9
Polycyclic aromatic hydrocarbons, 7-1-7-4, 7-10-7-30, 7-36-7-38, 7-47-7-49, 7-64
Predictive performance
description, 4-13^-14
Little Scioto River case study, 7-35, 7-38, 7-41, 7-45
Presumpscot River case study, 6-16
Preservation programs, 1-9, A-10
Presumpscot River case study
background information, 6-3-6-5
biological indicators of non-attainment, 6-6
comparison with Androscoggin River, 6-11-6-13
conceptual model of stressor impact, 6-7
eliminating candidate causes, 6-8-6-12
executive summary, 6-1-6-3
identifying probable cause, 6-17
list of candidate causes, 6-5-6-8
map, 6-4
significance of results, 6-18
strength of evidence analysis, 6-11-6-16
using results, 6-18
Probable effect levels, 7-29-7-30, 7-33, 7-64-7-65
Pseudoreplication, 3-7
Pulp and paper mill discharge, 6-1-6-18
Q
QHEI. See Qualitative Habitat Evaluation Index
Qualitative Habitat Evaluation Index, 7-1, 7-4-7-5, 7-13-7-14, 7-20, 7-24, 7-27, 7-56
Quality System website, 3-2
Index 1-7
-------
Stressor Identification Guidance Document
R
R-EMAP. See Regional Environmental Monitoring and Assessment Program
Regional Environmental Monitoring and Assessment Program, 4-11
Restoration programs, 1-9, A-10-A-11
Risk assessment, 1-8, A-9
S
SECs. See Sediment effect concentrations
Sediment effect concentrations, 7-29
Sediment organic compounds, 7-57-7-60
Sedimentation, 6-7-6-8, 6-10, 6-12, 6-14-6-17
SEP. See Supplemental Environmental Project
SI. See Stressor Identification
Source
defined, 2-1
Spatial co-location associations, 3-4
Spatial co-occurrence, 4-11, 6-14
Spatial gradient associations, 3-4
Spearman rank correlations, 7-14, 7-19-7-20
Specificity of cause
description, 4-13
Little Scioto River case study, 7-35, 7-37, 7-41, 7-45
Presumpscot River case study, 6-15
Statistical techniques
analyzing observational data in Stressor Identification, 3-7
evaluating confidence in causal identification, 4-17
Stressor Identification
analyzing evidence, 3-1-3-11
applications of the process, ES-2-ES-3
associations between measures of exposure and measures of effects, 3-8
characterizing causes, 4-1^-18
data quality issues, 1-2
document overview, ES-3-ES-4
EPA objectives, 1-1
flow of information from data acquisition to analysis phase, 3-3
function and description, ES-1
intended audience, ES-2
iteration options, 5-1-5-3
listing candidate causes, 2-1-2-7
management context, 1-4
mechanistic association with site data, 3-9-3-10
overview of process, 1-3-1-5
process iterations, 1-5
role in water management programs, 1-6-1-9
scope of guidance, 1-2
TMDL program and, A-4
using results, 1-5-1-6
using statistical techniques for analyzing observational data, 3-7
water management programs, A-l-A-11
worksheet model, B-l-B-37
Stressor-responses
Arkansas River case study, 4-11
1-8 £7.5. Environmental Protection Agency
-------
Stressor Identification Guidance Document
DDT case study, 5-2
description, 4-12
Little Scioto River case study, 7-33, 7-37, 7-40, 7-45
Presumpscot River case study, 6-15
Superfund, 1-9, 7-10, A-10-A-11
Supplemental Environmental Project, A-8-A-9
T
Tanytarsini midges, 7-1, 7-3-7-4, 7-8-7-10, 7-19-7-23, 7-27, 7-43-7-48
TEL. See Threshold effect levels
Temporal gradient associations, 3-4
Temporal relationships, 3-4
Temporality
description, 4-10
Little Scioto River case study, 7-32, 7-36, 7-39, 7-43
Presumpscot River case study, 6-14
Threshold effect levels, 7-29-7-30, 7-33, 7-64-7-65
TIE. See Toxicity Identification and Evaluation program
TMDL. See Total Maximum Daily Load
Total Maximum Daily Load
Clear Water Act requirements, A-2-A-3
EPA actions, A-4
Presumpscot River case study, 6-2-6-3, 6-18
Stressor Identification process, 1-6, ES-2
Total phosphorus
Presumpscot River case study, 6-1-6-18
Toxic compounds. See also Chemical contaminants
Little Scioto River Case Study, 7-11, 7-14-7-22, 7-24-7-25, 7-27, 7-29-7-30, 7-
32-7-48
Presumpscot River case study, 6-5-6-8, 6-12, 6-14-6-17
Toxicity data plot, 3-4
Toxicity Identification and Evaluation program, 4-5, A-6-A-7
Toxicity Reduction Evaluation, A-6-A-7
TP. See Total phosphorus
TRE. See Toxicity Reduction Evaluation
TSS with floe
Presumpscot River case study, 6-1-6-18
Type I error, 5-1
U
U.S. Environmental Protection Agency
compliance and enforcement of CWA, A-8-A-9
Data Quality Objectives process, 3-2
Environmental Monitoring and Assessment Program, 2-4
Quality System website, 3-2
TMDL program implementation, A-4
Wetlands Division website, A-10
W
Warmwater Habitat, 7-5-7-7
Washington
Lake Washington case study, 4-13
Index 1-9
-------
Stressor Identification Guidance Document
Waste water treatment plants, 7-6, 7-10-7-11, 7-28
Water chemistry parameters, 7-63
Water hardness, 7-18, 7-63
Water management programs, A-l-A-11
Water quality
overview of Stressor Identification, 1-3-1-5
ratings, A-l-A-2
Stressor Identification process, 1-6-1-9, ES-2-ES-3
Water Quality Act Amendments, A-5
Water Quality Certification
dredge and fill permitting, A-7
role of Stressor Identification process, 1-7
Water Quality Classification, 6-18
Watershed management programs
role of Stressor Identification process, 1-6
state and local programs, A-4-A-5
Wetlands assessments
methods, A-9-A-10
role of Stressor Identification process, 1-8
Wetlands permitting
role of Stressor Identification process, 1-8
Worksheet model, B-l-B-37
WWH. See Warmwater Habitat
WWTP. See Waste water treatment plants
1-10 £7.5. Environmental Protection Agency
------- |