Jfc United States
Environmental Protection
% W^mU.ilk Agency
Summary Report for the
Workshop on the Causal
Analysis/Diagnosis Decision
Information System
(CADDIS)
Causal Analysis/Diagnosis Decision Information System
-------
EPA/600/R-02/078
October 2002
Summary Report for the Workshop on the
Causal Analysis/Diagnosis Decision Information System
(CADDIS)
U.S. Environmental Protection Agency
Mt. Sterling, OH
August 26-28, 2002
National Center for Environmental Assessment-Washington Office
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency (EPA) policy and has been approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
The drafts of this report were prepared by T.N. and Associates under an EPA contract
(Contract No. 68-C-98-187, Work Assignment No 2002-04) as a general record of discussions
during the Causal Analysis/Diagnosis Decision Information System Workshop. The drafts were
reviewed by EPA personnel, and revised and finalized by the EPA Work Assignment Manager.
Background materials for the workshop were prepared by T.N. and Associates under an EPA
contract (Contract No. 68-C-98-187, Work Assignment Numbers 2001-09 and 2002-02) and
were reviewed and approved for distribution by the EPA Work Assignment Manager. This
workshop was co-sponsored by EPA's National Center for Environmental Assessment and
National Exposure Research Laboratory. As requested by EPA, this report captures the main
points and highlights of discussions held during plenary and break-out sessions. The report is
not a complete record of all details discussed nor does it embellish, interpret, or enlarge upon
matters that were incomplete or unclear. Statements represent the individual views of each
workshop participant; none of the statements represent analyses by or positions of the EPA.
ACKNOWLEDGMENTS
The CADDIS Workshop Steering Committee thanks Seemeen Hashem from Aqueous
Studios for the CADDIS logo. The design is inspired by caddisflies, whose occurrence in
streams often indicates high quality, and the intricate cases they build from pebbles and sticks.
We also thank David Bussard, Kate Smith, Tony Maciorowski, Don Brady, and Jonathon
Herrmann for their support, and Donna Reid-Judkins, who co-chaired the original Stressor
Identification workgroup.
li
-------
CONTENTS
LIST 01 TABLES v
LIST OF FIGURES v
AUTHORS AND REVIEWERS vi
LIST OF ACRONYMS vii
1. EXECUTIVE SUMMARY 1-1
1.1. System Functions Should Focus on Guidance, a Knowledge Base and
User Interaction 1-2
1.2 System Development Should Be User-Focused, Modular and Iterative 1-3
1.3 Conclusions 1-3
2. INTRODUCTION 2-1
2.1 Workshop Objectives 2-1
3. PLENARY SESSION 3-1
3.1 Stressor Identification History 3-1
3.2 Stressor Identification Process Overview 3-2
3.3 CADDIS 3-5
3.4 Design Options and Prototype 3-7
3.5 An Approach for Developing Innovative Software 3-8
4. BREAK-OUT GROUP REPORTS 4-1
4.1 Break-out Groups A and B 4-1
4.1.1 Session 1: Defining the Impairment and Listing Candidate Causes 4-1
4.1.2 Session 2: Analyzing the Evidence 4-7
4.1.3 Session 3: Characterizing the Causes 4-10
4.2 Break-out Group C 4-13
4.2.1 Session 1: Issues Surrounding System Development 4-13
4.2.2 Session 2: QA Requirements and Issues 4-22
4.2.3 Session 3: Interaction With and Among Users 4-26
5. FINAL PLENARY DISCUSSION 5-1
6. CONCLUSIONS 6-1
APPENDIX A: AGENDA, PARTICIPANTS AND BREAK-OUT GROUP
ASSIGNMENTS A-l
APPENDIX B: BACKGROUND MATERIALS
Neptune Paper (CADDIS) B-l
Menzie Paper No. 1 - U.S. EPA Sources B-26
Menzie Paper No. 2 - Additional U.S. EPA Sources (Mid-Atlantic focus) B-48
in
-------
CONTENTS (continued)
APPENDIX C: CASE STUDY MATERIALS AND REFERENCES
Cabin John Case Study C-l
Little Scioto Case Study C-8
iv
-------
LIST OF TABLES
1. Issues to consider when choosing a web-based system and a local application 4-20
LIST OF FIGURES
1. The management context of the SI process 3-3
2. Conceptual diagram of the CADDIS project, research and user communities 3-6
3. Where CADDIS may aid the process of impaired waters 303(d) listing and TMDL
development 4-16
v
-------
AUTHORS AND REVIEWERS
The National Center for Environmental Assessment-Washington Office (NCEA-W) of
EPA's Office of Research and Development was responsible for the preparation of this
document. The first and second drafts were prepared by T.N. and Associates under an EPA
contract (Contract No. 68-C-98-187, Work Assignment No 2002-04). The drafts were reviewed
by the workshop participants, and revised and finalized by the EPA Work Assignment Manager.
AUTHORS
Barbara Brown, U.S. EPA, ORD
Susan Cormier, U.S. EPA, ORD
Doug Norton, U.S. EPA, OW
Susan Norton, U.S. EPA, ORD
Leela Rao, U.S. EPA, ORD
Bhagya Subramanian, U.S. EPA, ORD
Glenn Suter, U.S. EPA, ORD
William Swietlik, U.S. EPA, OW
Paul Black, Neptune and Company, subcontractor to T.N. and Associates
Cheri Butler, Menzie-Cura and Associates, subcontractor to T.N. and Associates
Jackie Little, T.N. and Associates
Charlie Menzie, Menzie-Cura and Associates, subcontractor to T.N. and Associates
Tom Stockton, Neptune and Company, subcontractor to T.N. and Associates
REVIEWERS
The workshop participants (Appendix A) were given the opportunity to review the draft
workshop report for accuracy in the content captured from the plenary and break-out sessions.
The comments received from the participants have been addressed and incorporated into the
report. A second round of review was conducted on the revised draft by two U.S. EPA ORD
workshop participants-Steve Fine, Scott Minamyer-to ensure that the accuracy was maintained.
vi
-------
LIST OF ACRONYMS
BASINS
Better Assessment Science Integrating Point and Nonpoint Source
CADDIS
Causal Analysis/Diagnosis Decision Information System
CD
Compact Disc
CM
Conceptual Model
CWA
Clean Water Act
ECOTOX
Ecological Toxicity Database
EMAP
Environmental Monitoring and Assessment Program
U.S. EPA
U.S. Environmental Protection Agency
MIMS
Multimedia Integrated Modeling System
NCEA
National Center for Environmental Assessment
NGO
Non-government Organization
ORD
Office of Research and Development
OS
Operating System
OW
Office of Water
QA
Quality Assurance
QAPP
Quality Assurance Project Plan
SI
Stressor Identification
SPRC
Strategic Planning and Research Coordination
STAR
Science to Achieve Results
STORET
STOrage and RETrieval database
TMDL
Total Maximum Daily Load
WATERS
Watershed Assessment Tracking and Environmental Results System
vii
-------
1. EXECUTIVE SUMMARY
Increasingly, the regulatory, remedial and restoration actions taken to manage impaired
environments are based on measurement and analysis of the state of the biotic community.
When an aquatic community has been identified as impaired, the cause of the impairment must
be determined so that appropriate actions can be taken. The Stressor Identification (SI)
Guidance, published in December 2000, describes a methodology for identifying the causes of
observed impairments in aquatic systems. The ability to identify stressors accurately and defend
those findings with supporting evidence is a critical step in developing strategies that will
improve the quality of aquatic resources.
SI requires extensive knowledge on a variety of stressors as well as depth of knowledge
on the mechanism, symptoms, and stressor-response relationships for specific stressors. The
goal of the Causal Analysis/Diagnosis Decision Information System (CADDIS) project is to
support investigators in Regions, States and Tribes in the determination of causes of aquatic
impairments. We envision a decision support system that will help investigators find, access,
organize and share information useful for causal evaluations in aquatic systems.
At a workshop in Mt. Sterling, OH, August 26-28, 2002, representatives from the Office
of Research and Development (ORD), Office of Water (OW), U.S. Environmental Protection
Agency (U.S. EPA) Regions and States came together to conceptualize CADDIS and to identify
critical research needs for system implementation and population. This report summarizes the
main points and highlights of the workshop. These results may be of interest to scientists
involved with diagnosing impairments, as well as developers of decision support systems. The
report will also be used as the basis of a strategic plan for system development and as a source of
specific recommendations to be incorporated directly into the system's design, platform and
architecture.
This report follows the general structure of the workshop. The workshop began with
plenary presentations summarizing the history and vision for the project, after which the
participants were divided into three break-out groups. Two groups used case studies to provide a
context for identifying high-priority functions and research needs. The third group discussed
issues of system development, quality assurance and the interaction with and among users in the
context of the potential user groups. A final plenary session was used to summarize and
synthesize participant views.
Overall, participants felt that:
1. Professional judgment could be aided, but not replaced by, a computer
1-1
-------
2. CADDIS should focus on providing information, a guide through the SI process and
support for the users, rather than actually making judgments (e.g., through statistical
analysis).
In other words, CADDIS should support the SI process rather than try to perform the process.
Useful feedback on high-priority system functions and considerations for system development
that emerged during the workshop is described in greater detail next.
1.1. System Functions Should Focus on Guidance, a Knowledge Base and User Interaction
The functions that were identified as the greatest help to potential users included
guidance and documentation, a knowledge base, and conceptual model (CM) development.
Specifically, workshop participants stated that the most fundamental feature of CADDIS should
be the capability to lead the user through the SI process and prompt him or her to provide
information, make judgements, or determine the relevance of a step or consideration. This
guidance could include providing the user with blank forms and tables to fill out at appropriate
stages, general advice on avoiding common pitfalls, or suggestions on how best to perform the
analyses and inferences. High-priority areas for more detailed guidance included developing the
initial list of candidate causes, the appropriate use of statistics and the scores used for comparing
strength of evidence. Once the user has completed the SI, CADDIS should then provide some
sort of report that could be reviewed, revised and printed.
Workshop participants also identified the creation of a knowledge base as a high priority.
The knowledge base should include stressor-response relationships, organism tolerances,
probabilities of association for stressors and impairments and a searchable database of case
studies. The need for States to have easy access to stressor-response information, particularly for
the highest priority stressors, was frequently voiced. The highest priority stressors might be
gleaned from the current knowledge of stressors influencing total maximum daily loads
(TMDLs) and known information gaps (e.g., sediment, flow extremes, habitat modifications),
supplemented by user input. In addition to making stressor-response data more accessible, the
need for literature syntheses and shorter synopses was highlighted. Since the development of a
comprehensive set of stressor-response relationships might require additional research, CADDIS
should have a feedback mechanism to inform ORD research planners of user-identified
information gaps.
Providing a database of case studies was seen as a way to enable the community as a
whole to improve expertise in causal evaluations. Making the database searchable would allow
users to find situations similar to the ones they are evaluating. All of the break-out groups
recognized that CADDIS will need to have a mechanism for informing the users of the quality of
the data and information in the knowledge base, as well as prompt the users to keep track of
1-2
-------
quality assurance (QA) information for the data they themselves are using for the causal
evaluation.
A third functionality that the workshop participants identified as a high priority was
aiding the users in CM development. The initial help provided in this area should consist of a
database of standard CMs for typical situations. The users could modify these examples to allow
for customization to a given site. Future developments might include a graphics tool that allows
the user to pick a source or stressor and connect it to effects. Such a tool would not only be
useful for demonstrating the link from SI to TMDL efforts, but may also be attractive to other
Agency and State programs.
1.2. System Development Should Be User-Focused, Modular and Iterative
CADDIS workshop participants felt strongly that early and ongoing input from user
groups was essential to the success of the project. In addition to seeking a representative from
the user community to serve on the development team, feedback could be solicited through the
SI web page, regionally sponsored sessions and visits to scientists in the States and Tribes.
However, it was also noted that the need to engage the user community does not preclude
making initial progress on system development. In fact, it would be useful to have a prototype or
a system with limited functionality available to potential users to help solicit feedback.
The workshop participants supported the concept of a phased, iterative, and modular
development process. A phased and iterative development process would help solicit user and
management advice, allow the development team to showcase successes quickly and often, and
would help keep the system focused on, and relevant to, changing user needs. A modular
CADDIS system would allow a broader user base to take advantage of individual components
(e.g., CM builder, stressor-response databases) of the system. Modularity would also make it
easier to match the strengths of different system platforms with functional needs.
The other major element of system design that participants discussed was
communication/feedback mechanisms for users of CADDIS. A communication mechanism is
important for facilitating both technical support and the sharing of user data and experiences.
Participants expressed interest in dovetailing on existing lines of communication for initial user
contact and marketing (e.g., existing mailing lists, newsletters, websites and conference calls),
but felt that CADDIS should also have its own listserv and online request and submission forms
to conduct technical support and user communication once the system is in place.
1.3. Conclusions
There was a general sense that the project needs to be on a fast track. States are moving
forward on how to address impaired waters listing and TMDL-related issues. Emphasis on early
1-3
-------
design considerations and outreach to user groups by workshop participants means that a
CADDIS development team will need to be in place as soon as possible. This development team
should include not only U.S. EPA technical staff, but also representatives from the user
community and U.S. EPA management. Although early progress is essential, workshop
participants also emphasized that CADDIS would be a long-term commitment for both
developers and users. With the right input and support from all groups, CADDIS will become a
system that can substantially help investigators identify causes of impairments in aquatic
systems.
1-4
-------
2. INTRODUCTION
2.1 WORKSHOP OBJECTIVES
Increasingly, the regulatory, remedial and restoration actions that are taken to manage
impaired environments are based on measurement and analysis of the state of the biotic
community. When an aquatic community has been declared impaired, it becomes necessary to
determine the cause of the impairment so that appropriate actions can be taken. In response to
that need, ORD and OW developed the Stressor Identification (SI) Guidance, which was
published in December 2000.1 The Guidance describes a methodology for identifying the causes
of observed impairments in aquatic systems. The ability to identify stressors accurately and
defend those findings with supporting evidence is a critical step in developing strategies that will
improve the quality of aquatic resources.
Use of the SI process is prompted by a biological impairment. The general SI process
entails critically reviewing available information, forming possible causal 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 flexible enough to support
varying water management requirements and stringencies in identifying stressors.
Implementing the SI process requires extensive knowledge about a variety of stressors as
well as depth of knowledge on the mechanism, symptoms and stressor-response relationships for
specific stressors. The SI Workgroup acknowledged that making the guidance useful for users in
the Regions and States will require (1) developing easier access to supporting information,
(2) providing investigators with additional support in organizing and analyzing information,
(3) assisting the users in performing logical inferences, and (4) facilitating the sharing of
experience and knowledge gained in conducting causal evaluations. The need for an expert
system to support causal evaluation was also highlighted by the OW/ORD Strategic Planning
and Research Coordination (SPRC) Diagnostics Group. The Group advised that the developed
system should help States and Tribes diagnose causes of biological impairment by linking them
with information they would not otherwise have the resources to access. As such, the overall
objective of the Causal Analysis/Diagnosis Decision Information System (CADDIS) project is to
develop a decision support system that will help investigators find, access, organize, apply and
share information useful for causal evaluations in aquatic systems.
The objectives of the CADDIS workshop were to conceptualize CADDIS and to identify
critical needs necessary to develop and implement the system. This report summarizes the main
points and highlights of the workshop. These results may be of interest to scientists involved
*U.S. Environmental Protection Agency 2000. Stressor Identification Guidance Document. EPA 822-B-00-025.
2-1
-------
with diagnosing impairments, as well as developers of decision support systems. The report will
also be used as the basis of a strategic plan for system development and as a source of specific
recommendations to be incorporated directly into the system's design, platform and architecture.
The workshop began with plenary sessions that were designed to provide participants
with background on the SI process, project history, and system design options. Following
plenary presentations, in-depth discussions were conducted in three break-out groups. Break-out
Groups A and B addressed system functionality and research needs from the perspective of two
case studies (the Little Scioto, OH; and the Cabin John Watershed, MD). Both of these groups
walked through the SI process in three sessions: session one focused on identifying the
impairment and listing candidate causes, session two presented information on analyzing
evidence, and session three concluded the SI process with characterizing and identifying cause.
Appendix C of this report presents background materials and references that were used in
development and presentation of these two case studies.
During each of the workshop sessions, participants discussed issues of system
functionality and research needs in the context of the case study and the information provided for
a particular step in the SI process. When discussing system functions, participants were asked to
think of functionality that would be useful in CADDIS and to prioritize that list of functions.
The participants were also requested to assign priority to information database needs and general
research needs.
Break-out Group C addressed overall system development issues from the perspective of
potential users. The first session specifically considered system development, and the
participants were requested to give feedback on who they envisioned as CADDIS users and how
those users should be engaged in the development process. Once a user base was identified,
Group C discussed possible customer use of the system and implications of various design
options. The second session was devoted to discussing the issue of quality assurance (QA) and
considerations for the development of a Quality Assurance Project Plan (QAPP) for CADDIS.
The final session focused back on the users and how the system might function to facilitate
communication between users, developers and researchers. The participants were asked to
develop some specific design options for communication, as well as to take into consideration
the risks of facilitating communication among these groups.
At the conclusion of the workshop, a final plenary session was held to summarize the
workshop participants' views on approaches to developing the system, high-priority functions,
and critical dates for system development and use. The participants were each given the
opportunity to make one final comment expressing a key point for CADDIS development.
This workshop report generally follows the organization of the workshop: a summary of
the opening plenary sessions, synthesis of the break-out group sessions, and conclusions from
2-2
-------
the final plenary session. Since the conclusions from the two case study groups were similar,
they have been combined and are reported together by session. The outcome from Group C is
presented separately.
2-3
-------
3. PLENARY SESSION
3.1 SI HISTORY
Presented by William Swietlik, U.S. EPA, OW, Office of Science and Technology
Use of biological assessments and criteria as a tool to measure the health of waterbodies
began in the early 1970s. Throughout the 1980s and 1990s, the U.S. EPA published numerous
technical and policy guidances to further the nationwide use of biological assessments and
criteria in surface water management. Currently, active biological assessment programs are in
place in all the States and at least two Tribal Nations. Biological criteria have been formally
adopted into the water-quality standards programs of five States and one Tribal Nation, and
another 28 States are working to do so. Most of these States are now using biological
assessments and criteria, among other tools, to help determine the condition of their waterbodies
under Clean Water Act (CWA) sections 305(b) and 303(d).
With the increased use of biological assessments and criteria to list waterbodies as
impaired under 303(d), States urgently need new tools to help them identify the causes of those
impairments and to isolate and control the sources. States find this task especially difficult when
the cause of the biological impairment is unknown, or when complex multiple causes are
involved. States have even claimed that the lack of scientific tools to properly diagnose causes
of biological impairment was becoming a major impediment to the further use of biological
assessments and criteria in their water-quality programs. This is especially critical given the
large number of legal actions involving biological impairments under the total maximum daily
load (TMDL) program.
As a result, in May 1999, OW and ORD initiated a series of research planning meetings
to identify critical short- and long-term research needs to ensure that States and Tribes would
continue to integrate biological assessments and criteria into their water-quality programs. As a
first effort, the SI workgroup outlined, drafted, and published the Stressor Identification
Guidance Document. This document is designed to lead water-resource managers through a
rigorous scientific process that identifies stressors causing biological impairment in aquatic
ecosystems and provides a structure for organizing the scientific evidence supporting the
conclusions. In addition to this, other important new tools were identified for ORD
development. These included a database system for States to help investigators access, organize
and share information useful for causal evaluations (i.e., CADDIS); models for identifying
sources; stressor-response relationships for pollutants; models to predict response of biota to
stressors; and expanded guidance for complex situations. At this time, OW's highest priority
next step for SI tools is the development of CADDIS.
3-1
-------
3.2 SI PROCESS OVERVIEW
Presented by Glenn Suter, (Modified from the Stressor Identification Guidance Document, EPA
822-B-00-025) U.S. EPA, ORD, National Center for Environmental Assessment (NCEA)
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 SI Process
Figure 1 provides an overview of the SI 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 top 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 and consists of three
main steps:
1. Listing candidate causes of impairment
2. Analyzing new and previously existing data to generate evidence for each candidate
cause
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.
3-2
-------
Detect or Suspect Biological Impairment
Stressor Identification
LIST CANDIDATE CAUSES
ANALYZE EVIDENCE
CHARACTERIZE CAUSES
Eliminate
Diagnose
Strength of Evidence
Identify Probable Cause
Identify/
Apportion
Sources
MANAGEMENT ACTION:
Eliminate or Control Causes;
Monitor Results
Biological Condition Restored or Protected
Figure 1. The management context of the SI process. (The SI process is shown in the center box with bold line. SI
is initiated with the detection of a biological impairment. Decision-maker and stakeholder involvement is
particularly important in defining the scope of the investigation and listing candidate causes. Data can be acquired at
any time during the process. The accurate characterization of the probable cause allows managers to identify
appropriate management action to restore or protect biological condition)
Source: Stressor Identification Guidance Document, EPA 822-B-00-025
3-3
-------
Step 1 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 (CM) that shows cause-and-effect relationships. Step 2,
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. This includes considering 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
Cause and effect in deliberate manipulations of field situations or media
These associations comprise the body of evidence used to characterize the cause.
In Step 3, characterize causes, the investigator uses the evidence to eliminate, to
diagnose, and to compare the strength of evidence 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.
SI Process Iterations
The SI process may be iterative, beginning with retrospective analysis of available data.
If the stressor is not adequately identified in the first attempt, the SI process continues using
better data or testing other suspected stressors. The process repeats until the stressor is
successfully identified. The certainty of the identification depends on the quality of information
3-4
-------
used in the SI process. In some cases, additional data collection may be necessary to confidently
identify the stressor(s). Although the SI process cannot accurately identify stressors without
adequate data, completing the SI process is helpful even without adequate data because the
exercise can help target future data collection efforts.
Using the Results of SI
SI is only one of several activities required to improve and protect biological condition
(Figure 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 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.
3.3 CADDIS
Presented by Susan Norton, U.S. EPA, ORD, NCEA
The CADDIS workshop's objective is to conceptualize a decision support system that
will help investigators find, access, organize, apply and share information useful for causal
evaluations in aquatic systems. In addition, critical research needs for system implementation
and population will be identified.
CADDIS is needed because implementing the SI Guidance asks much of investigators. It
requires wide knowledge about a variety of stressors as well as depth of knowledge on
mechanism, symptoms, and stressor-response relationships for specific stressors. High-quality
data from field and laboratory studies may not be readily at hand. Research results may need
additional development to be directly useful for the process. Finally, documenting the process
for presentation to others may be burdensome.
3-5
-------
CADDIS will supplement the experiences of investigators by providing a guide to the SI
process, facilitating access to relevant information and knowledge, providing tools to analyze
and interpret evidence, and helping investigators organize, quantify, and share results (Figure 2).
Close interactions with the research community will facilitate linkages to existing databases,
construction of new databases, and the identification of research needs. Close interaction with
the user community will facilitate the production of case studies, the application to new
investigations, and the prioritization of system functions and research needs. CADDIS will help
make investigations defensible, transparent, and repeatable. It will allow investigators to learn
from each other by creating opportunities for users to submit information and examples and
share experiences.
Process description & guidance
Find relevant information
Help analyze and interpret evidence
Help organize, quantify, share results
Build, maintain
CADDIS PROJECT
Case studies
Real cases
USER COMMUNITY
databases Do research
RESE
COMIV
ARCH
IUNITY
Figure 2: Conceptual Diagram of the CADDIS Project, Research and User Communities. The three
stacked boxes under CADDIS Project and Research Community represent the three phases of the SI
Process: list candidate causes, analyze evidence, and characterize causes.
3-6
-------
3.4 DESIGN OPTIONS AND PROTOTYPE
Presented by Tom Stockton, Neptune and Company, Inc.
Another intent of the workshop is to provide input on the design of CADDIS. To
facilitate workshop participation in CADDIS design, a white paper that describes three possible
design options was distributed to workshop participants (Appendix B). These options provide
examples of the types of material that could be included in CADDIS and were used to initiate
discussion on these and other possibilities. The three options present visions of CADDIS at
increasing levels of complexity.
The three CADDIS system design options, CADDIS1, CADDIS2 and CADDIS3, vary in
complexity and resource levels needed to develop and implement. These options build on each
other sequentially to define a resource-efficient path forward for developing CADDIS. The
move to more complex approaches is in large part a move from streamlining access to
information in a static environment towards an adaptive decision-making framework wherein
uncertainties and decision consequences are taken into account.
CADDIS1 is primarily a static system in which information is made available through
web-based links, and the user is stepped through the SI process as described in the guidance. In
CADDIS1, SI worksheets can be completed externally following guidance provided by the
system. CADDIS2 builds on CADDIS1 and includes options for analysis of SI data and
completion of the SI process interactively. This is a more dynamic system in which analysis
capabilities are included to allow synthesis of site-specific information with relevant database
and reference information. CADDIS3 involves more dynamic functionality in terms of
implementing the SI process. This option allows for building conceptual and numerical models
and designing the numerical models to derive causal effects in a probabilistic setting.
A table of the functionality CADDIS could provide was also presented for discussion at
the workshop (Appendix B-l). For each function or feature, this table indicates which of the
three option levels applies, as well as the anticipated level of effort and resources needed for
implementation.
A prototype web-based version of CADDIS was developed in association with the white
paper so that the workshop audience could share some of the vision of what could be included in
CADDIS. This is presented on a "localhost" web server. The look, feel, and technical approach
for implementing the desired functionality are presented using this prototype. The prototype
CADDIS takes the user through the SI guidance on a path that includes:
3-7
-------
Features and placeholders for all CADDIS1 options (access to information sources)
Examples of how CADDIS2 could be implemented, including passage through the SI
process, access to mapping and analytical support, and completion of the process of
identifying causes for the stressors using the Little Scioto case study
An overview of the possibilities for the CADDIS3 option
The prototype CADDIS website is intended as a tool for exploring how a user might be
guided through the SI process and which technical tools are needed to implement CADDIS.
This site is not intended to be a model for the aesthetics of CADDIS or to provide complete
functionality. In general, the web tools used in this prototype are Open Source and freely
available. The prototype serves as a point of reference and a point of departure for consideration
in the workshop. Along with the web-based prototype, a table of the functionality CADDIS
could provide is also presented as a point of departure for the workshop. For each function or
feature, this table indicates which of the three option levels applies, as well as the anticipated
level of effort and resources needed to implement the feature.
3.5 AN APPROACH FOR DEVELOPING INNOVATIVE SOFTWARE
Presented by Steve Fine, U.S. EPA, ORD, NERL
Typical software engineering approaches do not address the needs of innovative projects
because the traditional techniques are often unsuitable for situations where requirements
continue to evolve during the life of the project. U.S. EPA's Multimedia Integrated Modeling
System (MIMS) is an example of such an innovative project. MIMS will include software to
support the composition, configuration, application and evaluation of complex systems of
multimedia models. How to do this is a complex research problem that requires experimentation
and flexibility as new approaches are identified. To address these demands, MIMS is following
several approaches. Principles of agile software development are being applied. This includes
providing rapid feedback to software developers, using designs that are simple but leave future
options open, making incremental changes, and building quality products. MIMS also applies
several design techniques that simplify the process of incorporating new features, reuse good
designs and software that others have developed and that support our objectives, and maintain a
balanced effort with a focus on MIMS's customers' needs. The MIMS project has been
successful in iteratively developing software currently used by a small group. Broader
distribution of MIMS will occur this fall and provide additional feedback on the value of this
3-8
-------
development approach.
3-9
-------
4. BREAK-OUT GROUP REPORTS
This section discusses the results from the three break-out groups. Each of the three
groups worked through three sessions. For the case study groups, the sessions corresponded
with the steps of the SI process. The third group focused on different system development issues
during each of the sessions. Charges to the break-out groups are also presented, although the
session summaries do not necessarily follow the questions since each group approached the
charges differently.
4.1 BREAK-OUT GROUPS A AND B
Groups A and B focused on case studies for the Little Scioto River, OH; and the Cabin
John Watershed, MD respectively. These case studies were intended to offer food for thought to
help the break-out groups identify how a diagnostic system such as CADDIS could help in the SI
process. The break-out groups identified key functionalities for each of the three main steps in
the SI process. Specifically, the questions Groups A and B were charged to answer in each
session were:
Of the possible functions for CADDIS, which ones are highest priority and why?
• Are there other CADDIS functions?
What information would you like to access that is currently hard to find?
Is there information/data that would be useful but does not exist or is one or two steps
away from being directly usable?
The recommendations from the two groups had much in common; therefore, their
recommendations have been combined and are synthesized by session next.
4.1.1 Session 1: Defining the Impairment and Listing Candidate Causes
The first break-out session of the workshop focused on the SI phase of defining (refining)
the impairment and listing candidate causes. The four steps to this phase of SI are:
1. Mapping the situation
2. Refining the description of the impairment
4-1
-------
3. Listing the candidate causes developed from data collated for the case
4. Developing a CM of the situation
The assistance envisioned from CADDIS for all steps falls into three categories:
1. Educating the user
2. Providing foundational information for the step
3. Providing tips, or a "wizard," to prompt and guide the user through the process and
highlight considerations that the user should make during that step
Step 1. Map the situation
Functionalities for this step are:
Education
- Identify the kinds of things a user would want to include on a map, e.g., National
Pollutant Discharge Elimination System sites, gaging stations, industrial plants
• Foundational information
- Provide links to resources (if they exist) that can provide those kinds of data
NOTE: The groups felt that there should not be a mapping capability in CADDIS
itself; instead, the system should have the ability to upload user-generated maps
developed off-line or available from external sources (e.g., EnviroMapper).
However, some need was indicated for creating figures that depict the site (e.g.,
cartoon figures, line drawings, thumbnail sketches), even if mapping was not
provided directly in CADDIS. (This is also, in part, a CM issue.)
Tips/"wizard"
- Provide prompts when maps might be useful in the SI process or be helpful for
documenting implementation in CADDIS
4-2
-------
Step 2. Refine the description of the impairment
Functionalities for this step are:
Education
- Provide a description of what an index of biotic integrity is, as well as which
metrics and thresholds have been used (recognize that this might vary by State or
county and the problems this would create with consistency)
- Provide a description of stressor-response relationships
Foundational information (e.g., taxa lists and associated categorizations,
tolerant/intolerant, feeding guilds)
- Provide the ability for the user to modify the databases given by the system to
customize them for their own situation/place
- Identify particular suites of effects that indicate particular categories of change
(e.g., nutrient cycle, energy, diversity)
Tips/"wizard"
- Provide prompts through the thought process on how/what to refine in stating the
biological effect, such as appropriate metrics that might be indicative of particular
stressors, ways to disaggregate from index to indicators, or maybe, how to look at
the raw data to best describe biological effect
- Help the user define the scale of the problem (e.g., pervasive over a large area or
point source issue, stream or watershed). Impairment could be caused by either
local or global issues (e.g., air deposition)
Step 3. List candidate causes developed from available information from the case, from the
region, and from conferring with stakeholders
Functionalities for this step are:
Education
- Provide a description of stressor-response relationships
Foundational information (searchable "library")
- Provide a compendium of known source-stressor relationships
- Provide a compendium of known stressor-effect relationships
4-3
-------
- Provide a compendium of known thresholds (e.g., temperature tolerances for
various fish species)
- Provide links to already existing databases that provide information above
- Provide a compendium/database of case studies that can be searched for similar
situations to those that the user is evaluating
- Provide a "glossary" or two-page synthesis of knowledge about each stressor
Tips/"wizard"
- Provide prompts through what to consider, potential causes for particular stressors
or sources
- Provide a way for users to list their sources, then have the system list the typical
stressors for each source (pick list); allow the users to then modify for their own
situations
- Identify logical ways to shorten the list to what is possible for user's situation
(e.g., is a particular kind of source not there?)
- Provide a reminder (or maybe track through CADDIS) to document rationale for
taking something off the list
- Provide a way to extract new relationships (source-stressor or stressor-effect)
from the case studies at hand
Step 4. Draw CM
Functionalities for this step are:
Education
- Provide a description of what a CM is and how to use it; provide example CMs
• Foundational information
- Provide a database of "standard" CMs for typical situations (modifiable to allow
customization for a user's particular situation)
- Include current land use in the CM
Tips/"wizard"
- Possibly provide a graphics tool to allow the user to pick source or stressor, tool
4-4
-------
draws and connects
- Provide a way to click on an individual box and get underlying data if user has
loaded it in
- Provide the ability to nest, or develop hierarchical models (e.g., start with a
general relationship, click to more detailed model that shows more specific
interactions)
Group A talked about using the term "change" instead of "impairment" to reflect the fact
that what SI really measures is whether the system being investigated is undergoing change over
time. This is what leads to change in stream-measured values (e.g., indices) and causes streams
to become listed.
In general, flexibility should be provided for developing impairment and cause lists. It
would be helpful if common associations were made available within CADDIS, but user-defined
associations need to be accommodated as well. Metrics or measures that the user chooses must
also be accommodated. Some discussants thought that categorization of stressor-response
relationships might help the user sift through possible choices or options (e.g., biodiversity,
energy, nutrients, metals, toxic chemicals, habitat alteration, point sources, non-point sources).
The participants wanted long, fixed (common) lists as a starting point, but the ability to have
users expand and add to lists as well, and wanted to be able to link impairments and causes.
Following this discussion was one about how to pare down the lists in response to
implementation of the rest of CADDIS. The lists and relationships included in the foundations
above could be used to define impairments and causes. They could be used initially to generate
a broad-based list of stressor-response pairs, which could be managed (reduced or expanded) in
response to progress through the SI process; an elimination step will result in change to the lists.
There was recognition that the process is iterative, and so the active lists will have to be revised
in response to implementation within CADDIS. In general, this revision is done by the user. For
example, in the Little Scioto case, a lot of work and options were considered prior to the
presentation in the papers. For another site, working through the process did not locate the cause
initially, but another potential cause was added after the SI process failed to identify a cause at
the outset (Dye Facility). Flexibility within the system is needed to accommodate these changes
while the process is being implemented. Some concern was expressed that if lists were provided,
users could be anchored to those lists and would not define their own site-specific items.
However, it was felt that broad lists would be preferable.
Also, once list items are activated, the foundation is then laid for CM building, in which
4-5
-------
case CADDIS could generate CM diagrams linked directly to stressor-response pairs included in
the application. If CM diagrams are developed, then these diagrams should be updated with the
change, and the record of how and why the changes were made should be tracked. Creation of
CMs, graphically or otherwise, was considered very important to the process, since the CM
describes the process. Tracking can be used to document changes in thinking in response to new
information and evidence.
Some mention was also made of understanding the incremental cost-benefit from any
action that might be taken. For example, it may be more effective to take action on a less
impacted stream because recovery would be greater than to take a similar action on a more
impacted stream that may not recover as well. This point was not considered beyond the initial
discussion because it was felt that this came under the jurisdiction of management decisions,
which is not within the scope of CADDIS at this time.
In general, there was a sense that professional judgment could not be replaced by a
computer, and that CADDIS should focus on providing information, some guidance through the
SI process, and capability to implement the process by capturing the professional judgments of
the user, but not by making the judgments (e.g., through statistical analysis). In other words,
CADDIS should support the process rather than trying to perform it. (The user must answer the
site question, not CADDIS; CADDIS is simply a repository for supporting the user's judgments.)
Following this discussion, group members considered ways to pare down the lists in
order to implement the rest of CADDIS.
1. The break-out groups did not envision that CADDIS would provide analytical tools
and be a place to do on-line analysis of data, although Group A did believe that it
might be useful to provide the ability for the user to present data graphically (spatially
or temporally), perform exploratory data analysis, and calculate some basic
descriptive statistics within the system (e.g., bar charts that were presented in the
Little Scioto example). There was recognition that statistical analysis is used to
short-list impairments based on site data, but it was felt that this capability did not
have to be provided within CADDIS. Instead, CADDIS could simply be a repository
for the results of those analyses (e.g., short-list of impairments). The break-out
groups also agreed that CADDIS should not be envisioned as a place to go to get raw
data for a particular case; the users would bring their own data to the process.
Analyses of those data would best be conducted off-line, with the capability to upload
the results of those analyses into the system for documentation of the process.
2. The break-out groups felt one of the strongest benefits CADDIS could bring to the SI
process would be to serve as a knowledge base to add value to the user's evaluation
by (1) synthesizing stressor information and providing it in one place (as opposed to
4-6
-------
having to search the Internet), and (2) providing a repository of case studies that
provide knowledge of what everyone else has had problems with and which solutions
have been effective.
3. However, Groups A and B strongly felt that the final needs and requirements for the
system should be defined by the intended users. This leads to the question of who is
the intended audience. Once decided, those users should be contacted to see what
they feel would be useful. There was some discussion in Group A of whether
CADDIS should be a computer-based system or if the SI process is sufficiently well
defined and should be used more as a paper system before proceeding with
development of a computer system.
4.1.2 Session 2: Analyzing the Evidence
This phase of the SI process analyzes the available data from the site and elsewhere to
generate evidence concerning potential causation. The two case study groups considered the
support that they would like from CADDIS for this phase. The consensus of the two groups, and
a few areas of differences, are described in this section. The discussions are best summarized in
terms of the priorities for desired features.
Walk-through with Prompts
A fundamental feature CADDIS must have is the capability to lead the user through the
process and prompt the investigator to provide information, make judgements, or determine the
relevance of a step or consideration. In this phase of SI, this feature could be best performed by
providing a table of types of associations that the user would fill in with relevant evidence. In
general, it was felt that CADDIS also should provide an indication of where the user is in the
process and that the context of the work be understood.
Information
There was strong agreement that CADDIS users would want access to outside
information concerning potential causal agents. This would include: (1) stressor-response
information, (2) diagnostic responses including biomarkers, (3) characteristics of the agents such
as degradation rates and other physical/chemical properties, (4) reference or background levels
or characteristics of the agents, and (5) results of remediation or restoration projects. In addition,
some biological information is needed concerning the impaired organisms, populations, or
communities, such as spawning habits and seasons or trophic characteristics. The information
may be provided outside the system, and CADDIS may simply provide web links or references
to published reviews.
4-7
-------
CADDIS may provide a useful context-based search engine. It could also provide a
bibliographic database that would save the exposure-response and other information that the user
had accumulated and the related bibliographic sources. However, there was a consensus of both
groups that it would be very useful for CADDIS to generate and provide summaries of relevant
exposure-response and diagnostic information for common stressors and classes of stressors.
Finally, once CADDIS is in use, it will be important to provide a means for information sharing
among the users. This might be a separate listserv, the capability might be built into CADDIS,
or other means may be identified.
Expert Guidance
The concepts and terminology of CADDIS may be unclear to beginning users.
Therefore, CADDIS should, at minimum, provide links to definitions and explanations of
individual terms and concepts and even entire steps in the process. A later addition to the system
would be suggestions for how best to perform the analyses or inferences. These might include,
rules of thumb (e.g., how to relate temporal patterns of exposure in the field to laboratory test
durations) or built-in warnings about using the same evidence twice. They might also include
examples of analyses and inferences taken from case studies or generated for this purpose. User
feedback can guide the development of this feature.
Report Generation
At a minimum, CADDIS should save the results of work performed by the user in a
format that can be reviewed, revised, and printed. For this step in the process, it would be
valuable to generate tables of the evidence generated for each type of association. More
advanced versions of this feature would generate reports that would fully document and explain
the process so that it could be understood by managers and individuals other than the users.
Data Management and Analysis
This feature inspired considerable differences of opinion. Some participants believed it
would be useful to incorporate simple data management and analysis capabilities, particularly for
the types of analyses recommended by the expert guidance. Others felt that users already have
data management and analysis capabilities that they are familiar with and that are suited to their
data sources. Those users might benefit little from data management and analysis capabilities in
CADDIS. Based on user feedback, more advance analytical tools might be included. These
might be built into the system, or CADDIS might provide macros or formulas for use in common
4-8
-------
statistical platforms.
Quality Assurance
This feature also caused differences of opinion. Because CADDIS users will depend
primarily on existing data, they have little control over data quality and cannot take advantage of
conventional data QA approaches. Hence, the data QA is performed by the data generators, and
the CADDIS user should document that by importing the relevant meta data along with the
primary data. However, it may be important to evaluate and track the quality of the analyses and
inferences used in the causal analysis itself. For example, one may identify a spatial trend based
on 3 points or 30. Hence, several participants felt that it would be important to include a feature
that would prompt the user to evaluate the uncertainty or quality of an analysis or inference at
each step in the process and carry that information into the report.
Case Studies
These might include rules of thumb (e.g., how to relate temporal patterns such as
exposure in the field to laboratory test durations), or built-in warnings about using the same
evidence twice.
Link Forward
It may be useful to link forward from the analysis of evidence to the characterization
phase. In particular, some people felt that they are likely to see opportunities for elimination of a
candidate cause during this phase. They would like to jump forward to the elimination step,
eliminate that cause, and then return to the analysis phase.
Link Back to the Conceptual Model
Participants thought that it might be useful to be able to easily go back and modify the
CM during this phase. As the data are analyzed, the user is likely to gain an understanding of the
impaired ecosystem that could result in revision of the CM. Also, the CM could be updated as
candidate causes are eliminated or added, based on the analysis and subsequent characterization,
so as to reflect progress in the causal analysis.
The minimal CADDIS capabilities for the analysis of evidence must include a walk-
though of the process, with some ability to define or explain terms and concepts, and saving and
tracking results. It was felt that users will want some information very early in the development
of CADDIS, particularly related to exposure-response. Other features may be added in later
versions, based on user feedback.
4-9
-------
4.1.3 Session 3: Characterizing the Causes
Eliminating Impossible Causes
During the analysis phase of the investigation, associations are made using evidence
collected at the study site. These associations can now be used to eliminate impossible causes.
CADDIS can help by creating a form having all candidate causes depicted across one axis and
the specific impairments along the other axis. Icons could appear that allow the user to seek a
description of elimination, clarification of associations using texts, examples, a diagram or
cartoon, and pitfalls to avoid.
The questions and answers developed in the analysis phase should be available for
review by both the user and the investigator. The system could guide the user through each
elimination step or allow a quick mechanism to highlight those causes that should be eliminated.
Another option could allow the investigator to chose an automated process that eliminates all
causes that have been identified as impossible causes of the impairment.
Space should be provided for caveats, such as why a candidate cause is not eliminated
even though there is a "no" entered into the analysis. CADDIS could provide tips and options
for deciding if and when a concentration is "really" elevated or not from one location or time to
another, as well as tips on interpreting correlations to facilitate evaluation of biological gradients.
The table created for the elimination of causes should then show those causes that will
continue to the diagnosis and strength-of-evidence characterizations. It would be handy for the
table to provide a reference to the appropriate association developed in the analysis portion of
the documentation. Another way of tracking the process would be by showing a graphic of the
pathway used to eliminate the candidate cause.
The elimination step should occur prior to the collection of evidence using observations
from the study area and from the literature; associations with cleanly eliminated candidate causes
may not be needed, and this information would make it easier for CADDIS to create a table for
the strength-of-evidence arguments.
Using Svmptomology to Diagnose a Probable Cause
Observations and measurements about the specific impairment at the study location may
permit the investigator to use symptomology to diagnose a cause. Symptomology collected from
the published literature assembled in the analysis phase could be made available to the
investigator for easy viewing. This would probably include a table of symptoms and an
indication that each symptom was or was not present and weights, if appropriate.
4-10
-------
CADDIS could provide a location for documenting the characterization and rationale for
diagnosing the cause of an impairment. A graphic of the decision points could help to lead the
investigator through this step; it would document the process that the investigator used to either
diagnose or proceed to the strength-of-evidence analysis.
CADDIS should also provide space for composing text to support the tables. The team
has not fully explored the best way to analyze and present evidence for diagnosis. Options for
providing a realistic and complete example are needed.
Organizing "Strength-of-evidence" Arguments
During the analysis phase, associations were made using evidence gathered from the
literature and by comparing them with measurements and observations taken at the study area.
These associations could then be used to weigh various considerations within the strength-of-
evidence arguments. CADDIS could facilitate this process by providing information to help the
user through the process. Help should be optional so that experienced investigators can proceed
to completing the characterization of causes more rapidly. One option would be to show icons
that allow the user to seek a description of the strength-of-evidence process, including all
considerations, examples, and diagrams or cartoons, as well as pitfalls to avoid.
To characterize each consideration, for all the impairments and candidate causes,
CADDIS could return the investigator to a screen that shows a consideration, the evidence,
impairment, and the candidate cause. There is a particular need to enable the investigator to
become comfortable with scoring. Participants commented that the table in the SI Guidance did
not provide enough assistance.
CADDIS could provide a list of options appropriate for each consideration, and the
investigator could choose from among them and automatically score the consideration. The
score should then be automatically recorded in a summary table and a complete strength-of-
evidence table (including a reference for each piece of evidence); the investigator can then
compare all the evidence for a particular candidate cause or for a consideration among multiple
candidate causes. An option should be provided to use similar previous cases as templates or to
pull in a column of evidence from the strength-of-evidence table for a particular candidate cause.
CADDIS can help by creating either a summary form for each impairment (that shows all
the candidate causes depicted across one axis, and the considerations along the second axis), or a
similar form with the specific impairment information repeated below each consideration.
The interest in creating forms and tables elicited discussions of the challenges for coding
the software. The coding had to be flexible enough to accommodate variation among studies
4-11
-------
with different numbers of candidate causes and impairments.
Generating the Final Report
Some suggestions for basic report elements included:
Chapters for each section of the SI process, including an introduction describing the
regulatory or other impetus for undertaking the causal analysis
Various products such as tables, lists, CMs, bibliographies, and text should appear in
the appropriate sections
The template could be exported to the user's preferred platform for word processing
or presentation software.
Research Needs
Two situations were identified where further consideration is needed concerning the
availability of data, guidance on how to address the situation, and modifications to CADDIS to
accommodate these situations. The first was the scenario wherein multiple stressors could act
synergistically to cause an impairment.
The second was the need for flexibility in characterizing causes at different spatial scales.
For example, the impairments and/or causes could vary for different parts of a river or
components within a watershed. How CADDIS could accommodate analysis of various
impairments for different waterbodies and scales was discussed.
There also was discussion about the potential for a more advanced scoring process that
could include independent scoring developed by individual users. The process could
accommodate user uncertainty through some form of weighting process, as well as scoring
among candidate causes for each consideration.
Other Comments
There were many reminders that causal analysis is an iterative process. One situation
was described to illustrate this point. If an investigator was gathering information from the
literature and discovered that the impairment could be caused by a new candidate cause, then it
might be added to the list of candidate causes and the CM, and require analysis and
characterization.
Because of the complexity of performing both analysis and characterization, it seemed
important to remind investigators to seek informal consultation with many colleagues with
4-12
-------
diverse expertise. This was considered absolutely essential if, after completing the strength-of-
evidence process, no cause had sufficient evidence.
Throughout the causal analysis, a navigator would be useful to show the investigator
where he or she is in the process and which resources and work already done might be useful for
that activity. This would permit the user to move to that resource or to another part of the
process. Appearance could be via a list of completed steps or more complex processes. This is
especially important during the characterization step, because it is necessary to continually refer
to previous work.
Almost all of the functions listed in Appendix B-l of the Design Options Paper were
useful and often essential; however, some uncertainty existed concerning the definition and,
therefore the usefulness of the modeling and weighting wizards for the strength-of-evidence
analysis, as well as having the probabilistic modeling wizard for evaluating uncertainties.
There needs to be a balance between being instructional and having a place to create a
document. Generally, CADDIS should prompt users for information in a step-wise fashion and
record user responses to document decisions in the causal analysis. For cases where no evidence
was generated in the analysis phase, CADDIS could inform the investigator that there is no
reason to evaluate that consideration or could prompt the user to seek additional information.
4.2 BREAK-OUT GROUP C
Break-out Group C was tasked with addressing system development issues from the
perspective of the potential users in three different sessions. The general topics discussed in
each session included: (1) the issues surrounding system development, (2) QA requirements and
issues and (3) interaction with and among users.
4.2.1 Session 1: Issues Surrounding System Development
In Session 1, the group addressed the following specific questions:
Who are the potential users of the CADDIS system?
How will the users use the system?
Who might want to independently modify the code?
How and when should users be involved in the development process. How do we
make this happen?
4-13
-------
What are the pros and cons associated with different options for the system's long-
term home?
What design and implementation strategy would maximize the likelihood of CADDIS
becoming a long-lived system?
The results of these discussions are summarized next.
Potential Users of CADDIS
Group C brainstormed a list of potential users of the system. It was suggested that by
identifying as many potential users as possible, marketing the system would be easier and more
effective. It was also suggested that the better the developers understood the users, the better the
system requirements could be defined. However, one caution was noted that by trying to address
the needs of all potential users, the design of the system might be overwhelmed by a large
number of diverse needs. Group C identified the following potential users:
States: Biologists
Tribes: Biologists
Water Quality Managers
Local Governments: Stormwater Managers; Utilities Personnel; Biologists; Planners
Consultants
Citizen/Watersheds Groups
U.S. EPA Regions
Other Federal Agencies
• U.S. EPA, ORD
• U. S. EPA, OW
External Stakeholders, including potential sources of stressors
• Enforcement Personnel
Educators: Universities (students learn the process and then get hired by States/local
4-14
-------
governments)
From this list, Group C identified that the primary users would be States, at least initially. Local
governments may also be primary users, along with certain stakeholders who might be identified
as the source of the stressors causing an impairment in a waterbody.
Use of the System
Major uses of the system identified by Group C included 303(d) listing of impaired
waterbodies and the development of TMDLs (Figure 3), where causes needed to be confirmed
and sources needed to be identified. Both States and U.S. EPA would use the system in this
way. U.S. EPA Regional Offices may also use CADDIS for checking State lists of impaired
waterbodies in developing TMDLs or in TMDL litigation. States might also use the system to
confirm previously identified causes of biological impairment (using best professional
judgement or other techniques) to make their determinations more legally defensible. CADDIS
may be used to provide added value for current systems [STORET, BASINS] and to improve
cost-effectiveness in SI. It was also suggested by the group that CADDIS could be a valuable
training tool to enhance consistency in SI.
Modification of the System's Code
There was general agreement that the potential users listed in this section would probably
not need to modify the code. However, it will be essential for the users to be able to add or
change the data in the system. Given this, it was suggested by the group that data input and
format be flexible for different users and sources/types/formats of information. The group also
suggested that the system be modular to maintain flexibility and that adequate documentation
should be available to provide confidence in the TMDL decisions when modifications to the
system are made by users.
4-15
-------
Priority ranking
of part 1
Develop
schedule
Separate list
into 4 parts
303(d) Listing
Process
Submit
303(d) list
Identify impaired
and threatened waters
Start of Listing Cycle.
Develop listing
methodology
NO
Initiate TMDL development
according to schedule.
YES
Update next YFC,
listing cycle, Ytt>
NO
Problem
identification
EPA establishes
or approves TMDL?
Target
analysis
TMDL submittal
TMDL
Establishment
Process
Source
assessment
Implementation and
monitoring plan
Linkage of
source and target
Allocation
Figure 3: Where CADDIS may aid the processes of impaired waters
303(d) listing and TMDL development. Boxes around steps indicate where
CADDIS might be used.
4-16
-------
A discussion of the design of the system also ensued. However, a number of members in
the group suggested that it was premature to ask design questions prior to getting enough input
from the potential users. The group suggested that a series of questions could be posed as a user
requirements study to help focus user input on design. The questions that were suggested are:
What is the actual process you use when assessing an impaired stream?
What are the top five things the computer can do for you to make your job easier?
At what point in the process will you need to use the system?
What would help you to get through the TMDL process?
If you were to use the SI process, where would you need help?
What could we provide that would make you want to use this process?
What are the minimum features that CADDIS should have to make you want to use
the process? (e.g., searchable relevant literature)
What are the current weaknesses of SI that you think need to be addressed?
What kind of training/educational component would be needed to use this system?
User Involvement in the Development Process
The group had a fairly long discussion of this important issue. One general agreement
that emerged was that users need to be factored into the development process very early on, or
right up front, and to be involved often. Some members of the group seemed to agree that as
soon as a preliminary prototype was developed, it should be given to users to get their reaction
and feedback. Other members of Group C suggested that users should even be involved in the
development of a preliminary prototype. Some members of the group also suggested it would be
good to have user(s) on the development team and not just programmers and U.S. EPA project
managers. It was expressed that users need to invest the time in development to have the payoff
of acceptance in the end. One member offered that CADDIS should be a system created by users
for users. A few members of the group suggested that CADDIS should come out with something
quickly that has value to users (core functions), and iterate system development from there.
The group also recommended that there needs to be good communication between the
users and the developers - communication goes both ways. For good communication,
4-17
-------
developers of the system should have a prioritized list of requirements collected from the user
community. Developers should also have a schedule that shows that the list of user requirements
will be addressed; functions on the list will not be ignored. The group also suggested that system
developers need to give the users feedback on the time/cost needed for the requirements. This
may change the prioritization for the system development.
Group C also recognized it was important to keep U.S. EPA senior management involved
and informed about system development. It was suggested that CADDIS development team
should think big, but work carefully in small steps and show the progress. Continual funding of
CADDIS will be needed to maintain the system and support any enhancements. The
development team could use demonstrated user involvement, such as lists of features and
enhancements requested by users, to help justify a plan for incremental funding to develop and
improve the system. CADDIS needs to demonstrate careful planning and flexibility for
investment, one that is reasonable and affordable over time.
The group next discussed how to engage users. The group agreed that to best engage
users, contact needs to be face to face, at least initially. Later contact can be informal by e-mail,
web-based feedback, or discussion groups. There could also be monthly or bi-weekly meetings
of an "architecture group" - developers and substantially involved users. It was pointed out that
the users that are involved in development are those that want to get involved and have the time
to commit to the process. Since not all users may have time to be involved, there need to be
different levels of involvement by users; give users a choice of the level of his/her involvement.
The group suggested that there should be initial meetings with the user base: go out to the
Regions (Regions can invite State representatives). The group suggested having a workshop, or
series of workshops with users. One member suggested that U.S. EPA Regional buy-in is very
important for encouraging State use.
The group next discussed the Development Team. The group suggested the team should
consist of:
• Programmers
Project Managers
• Line Management
User Group (tiered)
One person intimately involved
Others less involved, maybe just from time to time
4-18
-------
A Long-term Home for the System
Several options were identified for the long-term home of the CADDIS system. This was
identified as a key question that would impact both the system design and the functionality that
could be built into CADDIS. The first option is that CADDIS could be developed as a local
application that resides on the user's computer. A local implementation of CADDIS could
assume either an Internet connection or no Internet connection. Alternatively, CADDIS could be
developed as a web-based application that resides wholly on a web server (e.g., the U.S. EPA
National Computing Center in RTP, NC). An intermediate option, web-based with applets, was
also discussed. This was considered intermediate since an applet is a Java program that may be
downloaded over a network connection and run on a local machine via a web browser. The
issues involved in deciding between a web-based application versus a local application are
outlined in Table 1. The group generally felt that the system needed to rely on Internet access
and that avoiding applets is desirable. No clear consensus on the appropriate path forward was
identified by the group on whether to proceed with a web-based or local application.
The U.S. EPA and a non-government organization (NGO) were discussed as two
possibilities for maintaining a web-based implementation of CADDIS. Advantages of a U.S.
EPA-based home for CADDIS included U.S. EPA maintaining control over CADDIS and the
credibility lent to CADDIS by the epa.gov domain name. Advantages of a professional
organization hosting CADDIS included increased flexibility and functionality allowed by not
having to meet U.S. EPA web specifications for security, accessibility and presentation, as well
as the possibly of a built-in peer review mechanism for case studies and expert system rules.
Also of concern was identifying a group within U.S. EPA or an organization outside of U.S.
EPA that could ensure long-term funding of the system.
4-19
-------
Table 1. Issues to consider when choosing between a web-based system and a local application.
Issues
Web only
Web with applets
Local application
with wph arrpss
Local application
without wph arrpss
Web Access
Application can be accessed
by anyone, anywhere. Access
could be disrupted by U.S.
EPA policy, funding
decisions, server problems.
Performance will be
significantly impacted
by access speed.
High-speed access
will be needed if large
remote datasets are
accessed.
No access needed but
removes linking to
information
resources, limiting
functionality.
Deployment
Deployment to U.S. EPA
server requires U.S. EPA
National Computing Center
Review that can hold up
release and will require a
security plan. Deployment
will also require an office to
take full responsibility for the
website/ application.
Same as web but
requires testing to
include a variety of
desktops to test
performance of
applets.
Could deploy via the
web and/or compact
disc (CD).
Have to deploy by
CD. Requires
maintaining mailing
lists, and entails
postage and mailing
costs.
Updating
Automatic. Users never see
updating process. Updates are
universal. No problems with
version control.
Same as Web-Only.
Automatic. Users
likely involved with
updating their
programs.
Mail out CDs to
users.
Maintenance
Keep up with browser changes
and web specifications.
Information/application easily
backed-up and recovered by
computer support center.
Keep up with browser
changes and web
specifications.
Keep up with browser
changes, web
specifications, and
operating system (OS)
changes.
Keep up with OS
changes
Analytical
Power
Less dependent on user's
computer, more on server
performance and number of
concurrent users. If the
application is running on
shared server, it may compete
for processing power with
other applications.
Dependent on server
performance and
number of concurrent
users.
Dependent on user's
computer.
Dependent on user's
computer.
Internet
Security
U.S. EPA security
requirements may limit
functionality.
U.S. EPA security
requirements may
limit functionality
Should not be an
issue.
Should not be an
issue.
Capacity
Dependent on server. If
limiting, could be improved
with more CADDIS resources.
Dependent on server
and user's system. If
server is limiting,
could be improved
with more CADDIS
resources.
User-dependent.
User-dependent.
Compatibility
Browser issues.
Browser issues.
Browser and OS
issues.
OS issues.
Resources
Requires resources for
computational power and data
storage on a CADDIS server.
The user does not store the
application locally, saving
disk space.
Requires resources for
computational power
and data storage on a
CADDIS server.
Requires server
resources for
updating.
Resources for CD
distribution.
4-20
-------
Table 1. (Continued)
Issue
Web only
Web with applets
Local application
with wph arrpss
Local application
without wph arrpss
Compliance
The application must comply
with the Agency standard web
templates; web functions
permitted in the U.S. EPA
environment and with Agency
security standards.
Same as Web only.
Fewer Agency
requirements.
Fewer Agency
requirements.
Control
User has less control over
data. Version changes may
limit reproducibility of results.
User has less control
over data.. Version
changes may limit
reproducibility of
results.
User has control over
data.
User has control over
data and software.
User Support
Browser issues.
Browser, version and
OS issues.
Version and OS
issues.
Version and OS
issues.
Training
Incorporate into website.
Incorporate into web-
site.
Incorporate into
website.
CD based training,
not easily updated.
Capabilities
More difficult to incorporate
existing analytical tools.
More difficult to create
complex, interactive graphical
user interfaces.
More difficult to
incorporate existing
analytical tools.
Easier to incorporate
existing analytical
tools.
Easier to incorporate
existing analytical
tools; no access to
information sources
on the web.
Customization
Difficult. Users would submit
suggestions to CADDIS
project for incorporation.
Difficult. Users
would submit
suggestions back to
CADDIS project for
incorporation.
No limitations.
No limitations.
Licensing
Varies. If the application uses
a commercial database or
other commercial software,
the owner must purchase an
"unlimited" license for web
use. If simple and non-
commercial software is used,
there is no additional cost.
Same as web only.
Commercial software
may have a
distribution fee.
Commercial software
may have a
distribution fee.
Design and Implementation Strategy for a Long-lived System
Group C came to agreement that CADDIS would be a long-term commitment for both
developers and users. As such, commitments from the developers (U.S. EPA) and from the
entity operating and maintaining the system (possibly U.S. EPA or an NGO) would need to be
long term; otherwise, the system would not survive. The group identified that long-term
financial support would be critical to long-term survival. The group also recognized that
long-term buy-in by U.S. EPA management to fund the development and operation of the system
4-21
-------
would be necessary. The group recommended that U.S. EPA management be kept informed of
the system development, products, uses, financial needs and user response. Briefings should be
provided to U.S. EPA management in the short term to justify support for the development of the
system, and commitments must be made to keep management informed over the long term.
Clear milestone products and successes would help over the long term to maintain U.S. EPA
commitment to the system.
From the user perspective, the group concluded that users need to find the system helpful,
user-friendly, and available, in timely fashion, to be a long-term success. From this standpoint,
the group felt that during development, the system should release the most useful features first
and soon. The most useful features could be determined from the user input at the outset of
system design. The group thought there would be huge implications regarding user longevity
depending on the content of the early released features, whether they be literature, data or tools.
If users get system components quickly and are satisfied with their usefulness and capability, this
will help to ensure their long-term acceptance of the system. Some members of the group
suggested that a modular design of the system would separate the user interface from the
databases and from other pieces, making the design more flexible and more able to change over
the long term. The group also suggested that if a web-based design is used for the system, it
would be easier to implement a hierarchical design that might serve a larger user base.
4.2.2 Session 2: QA Requirements and Issues
During the second session, Group C was charged with answering four questions:
1. What models are available for dealing with QA issues in a system like this?
2. Which QA issues are most important to the various user groups?
3. How do we strike the balance between access to all relevant information versus
access to high-quality information? Is the balance different for various user groups?
4. What are the experiences of the STORET, WATERs, or TMDL tracking system
developers?
When considering these charge questions, members of the group noted that CADDIS QA
requirements will depend on what components eventually comprise the system. Group C
identified all the potential CADDIS components and then discussed their highly variable
potential QA needs. These components included:
4-22
-------
System
- Design
Programming
- Rule base
Literature
- Refereed literature
Gray literature
Data
Data within CADDIS
- Data referred to by CADDIS
- Data input to CADDIS by users
Tools
External
Internal
Members of Group C were able to offer insights on QA requirements and potential issues based
on experience with U.S. EPA and non-U.S. EPA information systems such as WATERS,
ECOTOX, MIMS, STORET, BASINS, and American Society of Civil Engineers' Best
Management Practices effectiveness database. This report section provides the first steps toward
outlining a CADDIS QAPP, discusses issues affecting each CADDIS component, and concludes
by offering general principles that may help guide CADDIS QA in a generic sense.
System OA Considerations
Two major categories of system QA considerations were discussed: QA of the program
logic (i.e., the basic system design), and QA of the rule base. Basic system design QA is certain
to dominate CADDIS QA early on. Customarily, peer review is advisable for the system design
document and any related prototypes. System programming QA requirements will vary with the
functionality and tools that eventually comprise CADDIS. These may be incremental, and, thus,
compatibility with pre-existing CADDIS components and functions will be an important part.
The group advised the development of automated unit tests and system tests that could
mechanize some of the QA on an ongoing basis. Configuration management, or how
components interrelate, would also need to be addressed. Paired programming was mentioned as
one development technique that integrates some of the QA at the programming stage by
assigning two programmers to write and continually review designs and code as they are created.
4-23
-------
The second category of system QA involves the rule base, or decision logic, that drives
the expert systems, if any, utilized by CADDIS. This may rank among the most complicated QA
tasks, both in peer reviewing the causal analysis assumptions and decisions (see also "Tools"
below) and in assessing program functionality. Use of existing expert system shells may help
the latter.
Literature OA Considerations
One major component of CADDIS may be causal analysis literature search and retrieval.
For the most part, the QA responsibilities lie with the originators of this information, and
CADDIS QA role is more in characterizing than performing the QA. Literature QA
considerations affect inclusion/rejection of general categories of literature as well as individual
documents. Also, future decisions about how CADDIS handles its literature functions (e.g.,
citation retrieval only; external key words; CADDIS-assigned key words; external search engine;
annotated bibliographies, etc.) will have additional QA implications that would fall among U.S.
EPA's QA responsibilities.
As a rule, peer-reviewed sources will likely be the core of CADDIS literature, but
numerous gray literature sources were also considered valuable by Group C due to their often
greater detail, earlier availability and difficulty of obtaining. The group noted that peer review
and publication are no guarantee of relevance or quality if the causal analysis elements of the
study were not specifically QA'ed. Gray literature, on the other hand, despite lacking refereed
publication, may still contain well QA'ed causal information. In short, general guidelines for
literature quality are advisable, but some rules for characterizing QA appear necessary as well.
Data OA Considerations
Data and literature represent more of a gradient than two distinct categories, but Group C
did separately consider data QA for information ranging from unpublished case studies and
assessments through raw data measurements. As a rule, raw data was considered beyond the
scope of CADDIS and probably unusable, but use of case studies, assessment results, and such
materials as stressor-response relationships will likely appear in CADDIS. The group addressed
the general needs associated with data and studies within CADDIS, referred to by CADDIS, and
newly input by users.
Data encompassed within CADDIS could include U.S. EPA-designed case studies (for
demonstrating CADDIS), externally designed case studies adapted for CADDIS demonstration,
and any test cases run to assess system performance. Significant QA responsibility accompanies
this category. The QA metadata for these kinds of studies should be provided. External studies
4-24
-------
might require different levels of QA effort depending on how they are used as demonstrations of
the system, or in relation to the rule base; in any case, the QA procedures followed in all case
studies internal to CADDIS should be reviewed and made available. Data referred to by
CADDIS should meet a stated QA standard, and users should be able to access information on
what the QA of a given data set is. QA considerations are similar to those discussed under the
literature section, above; again the QA burden lies primarily with the originator, with CADDIS
responsible for characterizing its QA. Requirements are less clear for data input by users due to
the variety of possible circumstances CADDIS may encounter. For the time being, Group C
assumed that locally input and locally used data QA is the full responsibility of the user but an
increasing level of QA verification and characterization would become necessary where input
data become accessible to other CADDIS users or may become part of the CADDIS knowledge
base and decision logic.
Tools OA Considerations
As initially discussed under system QA, the decision support tools utilized by CADDIS
will also present a QA challenge. External tools such as BASINS or MIMS may again follow
the pattern of originator responsibility for QA, CADDIS review and acceptance of that level, and
CADDIS communicating the QA level and type. Internal tools, on the other hand, connote
significant QA responsibility in the soundness of their science as well as their functionality.
Clearly the standards set for the supporting science in the CADDIS internal tools must be
consistent with, but also will aid in the QA of, the CADDIS literature and data components.
Summary: Generic OA Principles to Consider
Group C offered the following general statements as useful guidance for the continued
development of CADDIS QA elements:
Every function designed into CADDIS should be periodically verified as performing
according to plan.
CADDIS has a responsibility to characterize and effectively communicate the type
and level of QA of the information a user may encounter through CADDIS and/or use
in an analysis.
The level of effort devoted to QA should be commensurate with the intent of
the analysis and the consequences of encountering error.
4-25
-------
Clearly articulated criteria that determine use or rejection of information sources in
CADDIS can objectify the process and help avoid the appearance of Agency
endorsement and favoritism or their inverse.
Data quality objectives selected for CADDIS components can strongly influence the
eventual magnitude, complexity, functional capabilities and reliability of CADDIS.
4.2.3 Session 3: Interaction With and Among Users
When considering the development of a system from the perspective of the user
community, it is important to think about how the system will facilitate communication among
the users, as well as between the developers and users. As such, Group C spent time in Session 3
discussing four questions:
1. How can we optimize the communication among users, CADDIS and researchers?
2. What benefits are associated with interactions within the user group? How can we
minimize the risks?
3. Which website "contact us" information would attract the feedback we want?
4. How would a CD handle interactions with and among users?
The group focused on three main types of communication: outreach, technical support,
and researcher interactions. Outreach to the user groups includes needs assessment, marketing,
and training. Technical support was considered to include both system support and process
support. Researcher interactions focused on the communication of research needs from the users
to the researchers, in addition to providing direct technical assistance to the users. The group
discussed not only how CADDIS could optimize these three types of communication, but also
the risks to facilitating this communication.
Optimizing Communication
Outreach
Outreach to the user groups needs to happen at all phases of CADDIS development.
During the early phases of development, communication with the users involves needs
assessments as well as a marketing component. Later, once a system is established, while needs
assessment and marketing continue, a training component should also be added. Group C
recognized the importance of utilizing various forms of communication, from electronic to face-
4-26
-------
to-face interactions, to get access to a broad cross-section of the user community. There was
also an interest in capitalizing on existing methods of communication. These included:
Speaking and holding training sessions at professional and coordinator meetings (e.g.,
U.S. EPA Regional meetings, Water Coordinator groups, the Council of State
Governments, Society of Environmental Toxicology and Chemistry, North American
Benthological Society, American Water Resources Association)
Writing articles in newsletters and trade journals (e.g., U.S. EPA Office of Water's
Watershed Events and Nonpoint Source Newsnotes, Ohio U.S. EPA Newsletter)
Posting messages on an established listserv or e-mail distribution list (e.g., U.S. EPA
techloops, STORET and NHD listservs, Water Quality Standards listserv)
Speaking on established conference calls held by various groups
Creating a training module on the U.S. EPA, OW Watershed Academy website
Creating a feedback form on the current U.S. EPA SI Guidance website.
In addition to noting these existing paths of communication, members of Group C
expressed interest in creating links with the industries and corporations that are likely to be
impacted by the outcomes of SI. These are groups that the CADDIS developers will want to
engage to make the best use of existing information databases have been created by the regulated
community. Additionally, industry groups are considered potential users of CADDIS and are
likely to use the system in their defense when named as the source of an impairment.
Technical Support
Once CADDIS is developed, users will need to have access to technical support.
Members of Group C were adamant about the need for a long-term technical support
commitment by the Agency or group that houses the system. The group considered a three-
pronged approach to technical support. The first, and most resource intensive, type of support
would be a dedicated technical support team that could work with users on a one-on-one basis.
In addition to individual support, the use of a listserv, with a searchable archive function, was
suggested as a way to allow the users to help themselves. Since there is a great deal of
experience within the user community, Group C felt that facilitating user communication with a
listserv would be a cost-effective method of tapping into considerable expertise. However, there
would need to be some level of monitoring of the listserv by the technical support team in case
4-27
-------
questions cannot be answered by the user community, and to ensure that the help offered by the
users is correct. The third type of technical support mentioned by Group C was allowing users
access to lists of bugs in the system and planned system improvements. By making these types
of information readily available, the developers demonstrate their commitment to open
communication with the users.
Researcher Interactions
The primary type of researcher interactions that Group C discussed was how CADDIS
could be used to facilitate the planning of research in the ORD. First and foremost, members of
Group C recognized the importance of ensuring that the development of CADDIS and related
diagnostic research was included in the existing mechanisms of ORD planning: multi-year plans,
the Research Coordination Teams (RCT) and SPRC. One way to help promote CADDIS in the
planning process is to demonstrate how CADDIS can be used as a vehicle for ORD diagnostic
information transfer, specifically from the Environmental Monitoring and Assessment Program
(EMAP) and the Science to Achieve Results (STAR) program.
In addition to promoting research planning for CADDIS within ORD, it was stated that a
mechanism for user feedback should be developed in the initial CADDIS designs. For example,
as an individual uses CADDIS to do a case, he or she would have the option to fill out a form
identifying research gaps. This information could be automatically placed into a database of
research needs, which could then be used by the research planners. Other avenues of
communication of research needs from the users to the research planners include the use of
listservs, one-on-one meetings, and e-mail lists.
While discussing design options for user feedback, members of Group C also mentioned
the need to design the system to accept data and information produced by researchers. In this
vein, some members of the group expressed the need for synthesis and review papers, instead of
single-study papers, since it is difficult and time consuming for investigators to sift through large
amounts of literature on one individual subject. Also, bringing information from synthesis
papers into CADDIS would be more cost-effective than adding data from single study articles.
Additionally, CADDIS is another way to distill information and could build upon the
information presented in review articles. Finally, as CADDIS identifies research gaps, the
writing of synthesis papers may be a cost-effective method to satisfy these needs.
Specific user and researcher interactions were also discussed by the break-out group.
The encouragement of direct technical assistance to the users by researchers was seen by the
group as a possible role of CADDIS. CADDIS could help link technical experts in specific areas
4-28
-------
to the users by providing them with a database of experts and their contact information,
categorized by technical expertise. While this was seen as a great benefit to the users, it was also
recognized that the researchers themselves may show some resistance to this contact. However,
since technical assistance is being promoted by U.S. EPA management as one role of ORD
scientists, there may be increasing managerial support for this function of CADDIS.
Risks
While the benefits to promoting communication lines between researchers, system
developers and the user community are easy to perceive, it is important to consider the possible
risks that are associated with this communication as well. Group C discussed the risks
associated with the communication in outreach, technical support and researcher interactions.
Outreach
Members of Group C noted that U.S. EPA has had problems in the past with inconsistent
messages being given to user groups and to the States in particular. Thus, it was felt that a
communication strategy that demonstrated to the States how OW programs and ORD programs,
such as EMAP and CADDIS, all fit together. In fact, CADDIS might be used as a vehicle for
showing the connections between various U.S. EPA programs.
There was also a recognized risk in reaching out to the possible sources (e.g., the
regulated community). Specifically, there is the possibility of misuse of the system by industries
that have been identified as sources of impairments. A source could use data unavailable to the
States that affects the results and points to another source. It is also feasible that industry groups
may take the system and enhance it to the point that it surpasses the capabilities of the system as
it is used by the States. Such enhancements might give the industries greater confidence,
certainty and power in litigation to defend themselves when identified as sources.
When communicating with all user groups, Group C also identified the importance of
managing expectations. Particularly when engaging in marketing activities, there is danger of
raising the expectation of the user community and then not delivering the product in the end.
There is also the danger of raising the U.S. EPA managers' expectations that the system, or part
of it, will be available sooner than is technically feasible.
4-29
-------
Technical Support
The risks associated with technical support are limited mostly to the difficulty in
managing versions of the system. There is a danger of users having varying versions or
accepting only some of the system updates. This would make technical support a difficult
endeavor, and, as such, it is important to consider version control when designing the system.
Researchers Interactions
Group C identified the primary risks in this category as being those associated with
losing intellectual property. From the perspectives of the researchers, there is little incentive to
put their data and information up into the system prior to publication for fear of being "scooped."
Thus, there needs to be a mechanism for encouraging researchers to get information into
CADDIS as soon as possible and in a way that does not compromise their intellectual property.
Summary
Group C discussed the importance of communication with and among user groups, once
those groups have been identified. While there were recognized risks to fostering this
communication via CADDIS, the benefits appeared to far outweigh perceived risks. The group
noted the importance of keeping communication in mind when designing the system, as well as
building effective communication pathways into the system from the beginning. As long as
effective management of expectations and protection of scientists' intellectual property are
incorporated into the communication strategy, CADDIS should prove to be an effective tool for
identifying research needs and for ORD research planning.
4-30
-------
5. FINAL PLENARY DISCUSSION
At the final plenary discussion, participants provided their recommendations on the
system development approach, high-priority functions, information and research needs, and
timing and dates for system development and use. At the close of this discussion, each
participant was asked to express a key thought about CADDIS. Upon review, several themes
emerged. These are presented next along with a synthesis of the participants' remarks.
Focus on the Users
"Work toward building a community of satisfied users."
Many of the participants emphasized the need to be in touch with potential user groups to
fashion the design and implementation of CADDIS. This was expressed as a "must do" to make
CADDIS a valuable system.
State and Tribal scientists working in TMDL programs were identified as a core user
community. Investigators in other organizations (e.g., watershed groups) and programs (e.g.,
Superfund) may also use the system. While there are many potential user groups, caution should
be exercised so that CADDIS does not try to do too much and be everything for everyone.
Although professional society- and U.S. EPA-sponsored meetings are good venues to
reach potential users, not all the right people may be present, and incomplete distribution of
materials may occur back at work. Other suggestions for identifying and engaging users
included:
Add a box for feedback and suggestions to the SI web page. The 26,000 hits on this
page indicate that it is an important information source for scientists working in this
area.
Link to the Watershed Academy and have a training module to introduce SI and the
development of CADDIS to the user community.
Distribute a fact sheet that would point people to a prototype where they try out
different functions and provide feedback.
Arrange one-on-one meeting with users who have varying levels of familiarity with
SI and CADDIS.
Implement or encourage Region-sponsored forums.
5-1
-------
All potential users and user groups do not have to be contacted at the same time during
development. Rather, user needs assessment and engagement should be a continuing part of
system development.
Users should help define functions of CADDIS that should receive high priority by
identifying what they cannot do or find difficult to do now. Users could assist in the formation
of electronic list services, give suggestions for the presentation of synthesis information on such
things as stressor-response relationships, and provide feedback on specific program needs such
as TMDL reviews. Such user engagement was also seen as a way to demonstrate how CADDIS
might deal with the diversity within the States with respect to state-specific approaches, state-
specific impairments, state-specific needs and wants, the availability of various state databases
and other state-specific sources of information, and the information technology requirements
across States.
Provide Functions That Guide Users Through the SI Process
"Walk this way."
Participants viewed CADDIS as providing an electronic, step-by-step guide through the
SI process. The step-by-step guide was seen as important to help users document the decision
process and improve the defense of causal relationships.
System functions in the early phases of development can include tips on getting started,
guidance and prompts for users as they move through the process, and advice on avoiding
pitfalls. Forms and formats for tables needed for the processes can be provided. A progress
tracking system can help users organize the information they use during the process and create
bibliographies. More advanced functions can help users gather and organize information in one
place, increase the automation of table creation, and provide easy access to particularly useful
analytical tools (e.g., data visualization and correlation). Specialized features might target
requirements for TMDL reporting and review.
Participants cautioned that a computer cannot substitute for professional judgment:
system operations and output functionality should be designed to avoid a "black box" effect. For
example, CADDIS could simply provide users with tables, or it could generate complete reports
based on information supplied by users. However, the CADDIS design should prevent the
possibility of automating report generation to the extent that users do not know what went into
that report.
5-2
-------
Provide Functions That Offer Helpful Information at Key Steps in the SI Process
"Knowledge is power."
A number of participants remarked that if CADDIS was nothing more than an SI guide
and tracking system it might fall short. They strongly recommended that CADDIS provide key
information (i.e., a knowledge base) for the user, and that the development and synthesis of a
knowledge base be a primary function of CADDIS. Providing information that is difficult to
obtain could serve as incentive for getting users engaged with CADDIS.
Stressor-response relationships were foremost on the list of needs for the knowledge
base. Other needs included targeting the literature and experiences regarding remediation and
restoration efforts, and synthesizing information on how consistently stressors are associated
with different impairments. Access to this information might involve providing one- to two-
page synopses, longer and more thorough synthesis papers, as well as facilitating access to the
literature. CMs can serve as an organizing framework for the information, and categorization
schemes can be used to facilitate access to information by relevant spatial units (e.g., ecoregion
or watershed).
Participants emphasized the need to initiate development of the knowledge base with the
highest priority stressors and responses. Our current knowledge of the stressors influencing
TMDLs can serve as a foundation for prioritization. Information on responses to physical
stressors such as sediment and habitat alteration were seen as urgent needs. Beyond stressor-
response information, participants noted that our clientele also need source-stressor relationship
information. Although this stretches the current scope of CADDIS, its importance for effective
management (e.g., allocation in a TMDL) strongly advocates its inclusion in the knowledge
base.
Provide a Means of Sharing Information and Lessons Learned
Participants indicated that CADDIS should foster communication among users, as well as
between users and the research community. This function was emphasized because there is still
limited experience within the scientific community on addressing unknown causes in aquatic
systems. Sharing case studies and useful sources of information would enable the community as
a whole to improve its expertise in causal evaluation. In addition, our understanding of certain
stressor-response relationships is incomplete. Given an effective feedback mechanism, research
needs and data gaps identified by CADDIS users could be fed into existing research planning
5-3
-------
and coordination efforts such as the OW/ORD SPRC Workgroups, the ORD multi-year research
plans and the research planning committees.
Two commercial systems, S-Plus and GoldSim, were mentioned as models of possible
approaches for sharing information and experiences. S-Plus maintains a listserv where users can
post questions and S-Plus programs and routines. The listserv is mediated by the system
developers so that they can personally reply to the more difficult questions, and it is archived to
facilitate searching for relevant postings. GoldSim is a simulation engine that also encourages
user interaction. The developers maintain a website where users can upload case studies. The
website approach may allow more flexibility in sharing information.
Decide on System Platform and Home
"There's no place like home."
Early decisions on system platform and home need to be made to initiate development of
CADDIS. However, final decisions should follow from a more extensive probing of user needs
and technological capabilities.
Group C (see Section 4.2.1) developed a table summarizing strengths and limitations of
different platform options. Different homes for the system reflect different strengths and
limitations; it may prove best to pursue different homes for different functions. For example,
guidance about the SI process, forms, and case studies, can all easily be done using an U.S. EPA
server. In contrast, functions that involve uploading and sharing information from users external
to U.S. EPA were seen as a potentially difficult issue, based on concerns for security, privacy
and maintenance of server functionality. While it may be possible to provide these functions
within the U.S. EPA-server constraints, a mechanism outside of U.S. EPA (e.g., via an NGO,
academic institution, or professional society) may provide easier, earlier, and more flexible
functionality.
Money, Resources and Marketing
"What's the BUZZ?"
It was clear that the success of CADDIS as a project will require time and resources.
Support and commitment of resources are needed within the Agency. More and fresher legs are
needed to carry out the long-term needs related to design, development, revision, and marketing.
Improving the scientific underpinning of CADDIS will also require planning and resources.
This is an ambitious project that has the potential for greatly improving the States' and Agency's
5-4
-------
abilities to identify problem areas and to guide environmental improvements. To accomplish
this, participants emphasized that a team will be needed.
Achieving Agency recognition that CADDIS is a long-term project will require
developing and articulating a clear vision. Compelling short-form descriptions for distribution
within the Agency will help solidify a vision of CADDIS - its structure, function and value.
Developing a strategic plan for CADDIS "the project" as well as the plan for CADDIS "the
system" will help in the planning for longer term science and development needs while
providing valuable functions in the shorter term.
Early successes will also help solidify commitment for the longer term vision. The
planning and development process should aim for delivery in the short term, with additional
functionality refined through user interaction. The scope of CADDIS should be periodically
revisited with user input, but kept sufficiently narrow to insure success.
Effective marketing of CADDIS is another requirement for long-term success.
Components of a marketing strategy should include engaging internal managers for budgeting
support, the consulting community to stimulate their development of SI and CADDIS as an
application for solving environmental problems, and scientists involved with the TMDL
program.
Developing links to the TMDL program was seen as a particularly critical element to
effective marketing. TMDL user groups can be engaged by provide links to CADDIS on
TMDL-related sites. The familiar "figure eight" diagram of listing and TMDL development
under the CWA can be marked up to show where CADDIS fits (Figure 3). The conceptual
modeling component of CADDIS can help link SI with TMDL efforts. Demonstrations or
prototypes of CADDIS can be provided at key meetings at which state biologists and others
involved with TMDL are expected. Case studies can be used to illustrate how CADDIS can
serve as the "front end" of a TMDL problem, can demonstrate how CADDIS can save time when
doing SI, and can underscore the value added of CADDIS for dealing with complex situations.
Approaches
"Be happy at small steps and gains, as this is likely to be the reality."
A number of participants emphasized the value of an iterative approach to developing
CADDIS. An iterative approach can keep the focus on users' needs, and enable the developers
to learn quickly from successes and failures.
The overriding theme of these comments was to get started immediately, but to develop
CADDIS in little steps guided by user input. The "agile development" user-guided approach to
5-5
-------
development (see plenary abstract Section 3.5) was seen as a promising approach. The
development team was encouraged to bite off small, doable chunks (phased scope) so that they
can document progress, successes and failures. CADDIS may best be developed as a set of
modules so that users can select from some or all of these modules. For example, a CM builder
function might be attractive to a broad variety of Agency and State programs. A modular
approach would also be consistent with a phased development of CADDIS.
Participants recommended identifying a minimum set of functions for the early
development phases. These appear to include: a system to step the user through the SI process,
access to information about stressor-response relationships (and perhaps other critical
information) when needed, a CM support system (at least at a simple level), and a recording and
reporting mechanism. Analytical capabilities within CADDIS are seen as less essential in the
early phases, although user group inputs may refine this concept over time if an agile
development approach is employed.
Guided by user input, CADDIS might grow in at least two ways: growth in sophistication
and complexity (i.e., include analytical tools, better accounting systems, etc.) and growth in the
supporting knowledge base (e.g., information about more stressors or more types of habitats).
Realizing the full potential of CADDIS will require a long-term commitment.
Broader Vision
"I can see for miles and miles and miles."
Two key points were raised with respect to looking beyond the applications of SI to date
and beyond the TMDL program:
There are lots of different aquatic and wetland habitats besides freshwater streams.
While the TMDL writers represent a core group of users, there are many other
potential users and we should be careful not to miss those opportunities. Examples
include applications for Superfund and ecosystem management programs (water,
wetlands, and terrestrial).
Timing and Getting Started
"Time's a-wasting."
Many participants emphasized that the project needs to be on a fast track, at least for the
early phase. States are moving forward on how to deal with TMDL-related issues. The need is
5-6
-------
there now. Since that need will be filled one way or another, timing is critical for success.
Participants provided the following suggestions for near-term actions:
1. Get a prototype up and running. Something needs to be "out there" in a short period
of time, even if only as a demonstration. At a minimum, make forms available. A
second priority would be uploading case studies into the system. Shoot for early
successes with low-technology options, and then repackage them for the web. ECOTOX
is willing to provide downloads as samples to show how that database can provide
information useful to SI.
2. Start engaging users. The upcoming Water Environmental Federation meeting in
November and the U.S. EPA biocriteria clinics planned for March 2003 are two good
opportunities. Provide training in both SI and the CADDIS prototypes. The number and
type of case studies should be expanded to include more Regions and non-point-source
issues. For example, it would be helpful to have something that is more non-point source
related, and in the western part of the U.S.
5-7
-------
6. CONCLUSIONS
There was a clear consensus among workshop participants that CADDIS should be
developed, because it has the potential to greatly aid investigators in States, Tribes, and Regions
to identify the causes responsible for ecological degradation. Workshop participants generated
the following recommendations:
High priority functions include a step-by-step guide through the SI process, a
knowledge base that would help users find and interpret relevant information,
assistance in organizing and reporting results, and a means for sharing information
and lessons learned.
The system should be developed incrementally and iteratively.
Frequent user input and feedback will be essential to the system's success.
Different system homes have different strengths and limitations; it may prove best to
pursue different homes for different system functions.
A clear vision and early successes will help garner the resources and commitment
necessary for the long-term viability of CADDIS
CADDIS has many potential uses beyond the TMDL program, including applications
in terrestrial environments and ecosystem management. The system should be
developed in a way that could support future expansion.
Time's a-wasting. The early phases of CADDIS should be pursued immediately
because States are already moving forward on many TMDL-related issues.
This workshop report will be used as the basis of a strategic plan for system development
and for more specific recommendations for the system's design, platform and architecture. It
will also be used to identify longer term research projects that will be useful for filling data gaps
and populating the system. By helping investigators identify the stressors responsible for effects,
CADDIS will provide a critical step toward identifying sources, taking action and improving the
nation's environmental quality.
6-1
-------
APPENDIX A
AGENDA, PARTICIPANTS AND
BREAK-OUT GROUP ASSIGNMENTS
-------
Agenda
Causal Analysis/Diagnosis Decision Information System
August 26-28 2002
Mt. Sterling, OH
Workshop Objectives:
1. To conceptualize an information system (decision support system, expert system) that
will help investigators find, access, organize and share information useful for causal
evaluations in aquatic systems.
2. To identify critical research needs for system implementation and population.
Monday August 26
Morning
Travel to Deer Creek
11:30-12:45 Registration
Afternoon
1:00 Welcome/meeting objectives
1:15 Stressor Identification History and SPRC Bill
1:30 Stressor identification process
2:15 The CADDIS concept, Meeting Objectives
2:35 Discussion of objectives and agenda
Mike Slimak
Swietlik
Glenn Suter
Sue Norton
Susan Cormier
3:00 Break
3:30 Options papers, Prototype
4:30 Developing Innovative Software
5:00 Overview of Tuesday's agenda
Tom Stockton
Steve Fine
Leela Rao
6:00 Poster social
A-l
-------
TuesdayAuaust27
8:00
Intro to the Little Scioto Case Study
Susan Cormier
8:30
Intro to the Mid-Atlantic Case Study
Barbara Brown
9:00
Session 1
11:30
Lunch
12:30
Report out from Session 1
Charlie Menzie (Facilitator)
1:30
Session 2
4:30
Break
5:00
Report out from Session 2
Charlie Menzie (Facilitator)
6:00
Overview of Wednesday's agenda
Leela Rao
Wednesday Auaust 28
8:00
Session 3
10:30
Break
11:00
Report out from Session 3
Charlie Menzie (Facilitator)
12:00
Lunch
1:00
Final plenary
Charlie Menzie (Facilitator)
3:00
Adjourn
A-2
-------
Participant List
Name
Phone
Affiliation
E-mail
Address
Brad Autrey
513-569-7368
SoBran
Autrey.Brad@epa.gov
26 W. Martin Luther King Dr., Cincinnati, OH 45268
Sharon Batterman
218-529-5220
ORD/NHEERL
Batterman.Sharon@epa.gov
6201 Congdon Boulevard, Duluth, MN 55804
Paul Black
Neptune & Co.
Barbara Brown
401-782-3088
ORD/NHEERL
Brown. Barbara@epa.gov
27 Tarzwell Dr., Narragansett, Rl 02882
Cheri Butler
Menzie-Cura
Susan Cormier
513-569-7995
ORD/NERL
Cormier.Susan@epa.gov
26 W. Martin Luther King Dr. [642], Cincinnati, OH 45268
Tod Dabolt
202-566-1186
OW/OWOW
Dabolt.Thomas@epa.gov
1200 Pennsylvania Ave., NW [4503T], Washington, DC 20460
Naomi Detenbeck
218-529-5025
ORD/NHEERL
Detenbeck. Naomi@epa.gov
6201 Congdon Boulevard, Duluth, MN 55804
Don Dorsey
240-777-7729
Montgomery Co.
DEP, Watershed
Mgmt. Division
don@askdep.com
255 Rockville Pike, Suite 120, Rockville, MD 20850
Chris Faulkner
202-566-1185
OW/OWOW
Faulkner.Chris@epa.gov
1200 Pennsylvania Ave., NW [4503T], Washington, DC 20460
Steve Fine
919-541-0757
ORD/NERL
Fine.Steven@epa.gov
USEPA Mailroom [E243-04], RTP, NC 27711
Jeff Frithsen
202-564-3323
ORD/NCEA
Frithsen.Jeff@epa.gov
1200 Pennsylvania Ave., NW [8601D], Washington, DC 20460
Laura Gabanski
202-566-1179
OW/OWOW
Gabanski.Laura@epa.gov
1200 Pennsylvania Ave., NW [4503T], Washington, DC 20460
Ken Galluppi
919-541-3306
ORD/NERL
Galluppi. Kenneth@epa.gov
USEPA Mailroom [D305-01], RTP, NC 27711
Jeroen Gerritsen
410-356-8993
Tetra Tech, Inc.
jeroen.gerritsen@tetratech.com
10045 Red Run Blvd., #110, Owings Mills, MD 21117
Gretchen Hayslip
206-553-1685
EPA/Region 10
Hayslip.Gretchen@epa.gov
1200 Sixth Ave. Seattle, WA 98101
Cheryl Itkin
202-564-3357
ORD/NCEA
ltkin.Cheryl@epa.gov
1200 Pennsylvania Ave., NW [8601D], Washington, DC 20460
Jan Kurtz
850-934-9212
ORD/NHEERL
Kurtz.Jan@epa.gov
Sabine Island Drive, Gulf Breeze, FL 32561-5299
Jackie Little
T.N & A.
Charlie Menzie
Menzie-Cura
Scott Minamyer
513-569-7175
ORD/NRMRL
Minamyer.Scott@epa.gov
26 W. Martin Luther King Dr. [G75], Cincinnati, OH 45268
Bruce Mintz
919-541-0272
ORD/NERL
Mintz.Bruce@epa.gov
USEPA Mailroom [D305-01], RTP, NC 27711
Pete Nolan
617-918-8343
EPA/Region 1
Nolan.Peter@epa.gov
11 Technology Drive [OEME], North Chelmsford, MA 01863-2431
Doug Norton
202-566-1221
OW/OWOW
Norton.Douglas@epa.gov
1200 Pennsylvania Ave., NW [4503T], Washington, DC 20460
Sue Norton
202-564-3246
ORD/NCEA
Norton.Susan@epa.gov
1200 Pennsylvania Ave., NW [8623D], Washington, DC 20460
Lynn Papa
513-569-7587
ORD/NCEA
Papa.Lynn@epa.gov
26 W. Martin Luther King Dr. [190], Cincinnati, OH 45268
Bruce Peirano
513-569-7185
ORD/NRMRL
Peirano.Bruce@epa.gov
26 W. Martin Luther King Dr. [290], Cincinnati, OH 45268
Dave Pfeifer
312-353-9024
EPA/Region 5
Pfeifer.David@epa.gov
77 West Jackson Boulevard [WT-16J], Chicago, IL 60604-3507
Ed Rankin
740-517-2275
Ohio University/
CABB
rankin@ilgard.ohiou.edu
9666 State Route 37 East, Sunbury, Ohio 43074
Leela Rao
202-564-3362
ORD/NCEA
Rao.Leela@epa.gov
1200 Pennsylvania Ave., NW [8623D], Washington, DC 20460
Chris Russom
218-529-5218
ORD/NHEERL
Russom.Chris@epa.gov
6201 Congdon Boulevard, Duluth, MN 55804
Mike Slimak
202-564-3324
ORD/NCEA
Slimak.Michael@epa.gov
1200 Pennsylvania Ave., NW [8601 Dl, Washington, DC 20460
-------
Name
Phone
Affiliation
E-mail
Address
Jonathan Smith
202-564-5181
ORD/NCER
Smith.Jonathanh@epa.gov
1200 Pennsylvania Ave., NW [8725R], Washington, DC 20460
Robert Spehar
218-529-5123
ORD/NHEERL
Spehar.Robert@epa.gov
6201 Congdon Boulevard, Duluth, MN 55804
Tom Stockton
Neptune & Co.
Bhagya Subramanian
513-569-7349
ORD/NERL
Subramanian.Bhagya@epa.gov
26 W. Martin Luther King Dr. [642],Cincinnati, OH 45268
Glenn Suter
513-569-7808
ORD/NCEA
Suter.Glenn@epa.gov
26 W. Martin Luther King Dr. [117], Cincinnati, OH 45268
Bill Swietlik
202-566-1129
OW/OST
Swietlik.William@epa.gov
1200 Pennsylvania Ave., NW [4304T], Washington, DC 20460
Marge Wellman
202-566-0407
OW/OST
wellman.marjorie@epa.gov
1200 Pennsylvania Ave., NW [4305T], Washington, DC 20460
James Wickham
919-541-3077
ORD/NERL
Wickham.James@epa.gov
USEPA Mailroom [E243-05], RTP, NC 27711
Chris Yoder
740-517-2274
Ohio University/
CABB
voder(®jlaard.ohiou.edu,
db8177 @dragonbbs. com
P.O. Box 21561, Columbus, OH 43221-0561
Lester Yuan
202-564-3284
ORD/NCEA
Yuan.LesterOeDa.aov
1200 Pennsylvania Ave.. NW I8623D1. Washinaton. DC 20460
-------
Breakout Group Assignments
Breakout Group
Role
Participant/Affiliation
Group A:
Little Scioto Case Study
Facilitator
Jackie Little
T.N. & Associates
Tech Lead
Susan Cormier
U.S. EPA / NERL- Cincinnati
Tech Lead
Paul Black
Neptune & Co. (subcontractor to TN & A)
Notetaker
Bhagya Subramanian
U.S. EPA / NERL - Cincinnati
Sharon Batterman
U.S. EPA / NHF.F.RI. - Duluth
Chris Faulkner
U.S. EPA / Office of Water - OWOW
Ken Galluppi
U.S. EPA /NERL- RTP
Gretchen Hayslip
U.S. EPA /Region 10
Bruce Mintz
U.S. EPA /NERL -RTP
Bruce Peirano
U.S. EPA / NRMRL - Cincinnati
Bob Spehar
U.S. EPA / NHF.F.RI. - Duluth
Jeroen Gerritsen
Tetra Tech
Brad Autrey
SoBran
Group B:
Cabin John Case Study
Facilitator
Charlie Menzie
Menzie-Cura (subcontractor to TN & A)
Tech Lead
Barbara Brown
U.S. EPA / NHEERL-Narragansett
Tech Lead
Glenn Suter
U.S. EPA / NCEA - Cincinnati
Notetaker
Cheri Butler
Menzie-Cura (subcontractor to TN & A)
Naomi Detenbeck
U.S. EPA/ NHF.F.RI.- Duluth
Jeff Frithsen
U.S. EPA / NCEA - Washington
Laura Gabanski
U.S. EPA / Office of Water - OWOW
Jan Kurtz
U.S. EPA / NHEERL- Gulf Breeze
Pete Nolan
U.S. EPA /Region 1
Don Dorsey
Montgomery County
A-5
-------
Breakout Group
Role
Participant/Affiliation
James Wickham
U.S. EPA/NERL-RTP
Chris Yoder
Ohio University / CABB
Lester Yuan
U.S. EPA / NCEA - Washington
Group C: System Development/
User Perspectives
Facilitator
Scott Minamyer
U.S. EPA / NRMRL - Cincinnati
Tech Lead
Doug Norton
U.S. EPA / Office of Water - OWOW
Tech Lead
Tom Stockton
Neptune & Co. (subcontractor to TN & A)
Notetaker
Leela Rao
U.S. EPA / NCEA - Washington
Tod Dabolt
U.S. EPA / Office of Water - OWOW
Steve Fine
U.S. EPA/NERL-RTP
Cheryl Itkin
U.S. EPA / NCEA - Washington
Lynn Papa
U.S. EPA / NCEA - Cincinnati
Dave Pfeifer
U.S. EPA /Region 5
Ed Rankin
Ohio University / CABB
Chris Russom
U.S. EPA / NHF.F.RI. - Duluth
Jonathan Smith
U.S. EPA/NCER
Bill Swietlik
U.S. EPA / Office of Water - OST
Marge Wellman
U.S. EPA / Office of Water - OST
Cross-session
facilitator
Sue Norton
U.S. EPA / NCEA - Washington
A-6
-------
APPENDIX B
BACKGROUND MATERIALS
-------
Design Options for the
Causal Analysis and Diagnosis
Decision Information System
(CADDIS)
Workshop Draft
Prepared by
Tom Stockton and Paul Black
of
Neptune and Company, Inc.
Prepared for T N & Associates, Inc.
under Contract Number 68-C-98-187
March 2002
revised July 2002
This document has been reviewed and approved for use as background for
the CADDIS Workshop by the EPA Work Assignment Manager.
B-l
-------
Introduction
This report outlines options for the Causal Analysis and Diagnosis Decision Information System
(CADDIS). CADDIS is envisioned as an interactive web-based decision support tool that will help
States/Tribes, Federal agencies, and other water resource managers identify causes of biological
impairments or ecological stress in water bodies. CADDIS will provide investigators access to
relevant information sources and decision-based tools that can be used to process information to
support consistent, defensible and scientifically-based causal evaluations of impaired aquatic
systems. This can lead to implementation of better water quality management controls, and more
useful and effective expenditures of valuable and often limited resources. The system is based on
EPA's Stressor Identification (SI) guidance (USEPA, 2000).
Three system design options for CADDIS are presented here in a hierarchical fashion from less to
more complex and resource intensive to develop and implement (CADDIS1, CADDIS2, and
CADDIS3). These options build on each other sequentially to define a resource efficient path forward
for developing CADDIS. The move to more complex approaches is in large part a move from simply
streamlining access to information in a static web-based environment, towards a more holistic
framework based on decision analytic approaches aimed at adaptive decision-making in which
uncertainties and decision consequences are taken into account. As CADDIS is implemented,
examples can be developed and added to the system, so that CADDIS can be dynamic and iteratively
update with each application.
In the description of the three options that follows, examples of the look, feel, and the technical
approach for implementing the desired functionality are available at www.neptuneandco. com/caddis.
The prototype CADDIS web site is only intended as a tool for exploring how a user might be guided
through the SI process and for exploring the technical tools needed to implement CADDIS. This site
is not intended to be a model for the aesthetics of CADDIS or to provide complete functionality. In
general the web tools used in this prototype are Open Source and freely available. A table of the
functionality CADDIS could provide is presented in Appendix B-l. For each function or feature, this
table indicates which of the three option levels applies, and the anticipated level of effort and
resources needed to implement the feature.
The descriptions of the options below focus on implementation in a web-based environment.
However, most of the CADDIS functions could be served over the web but actually implemented
on a CD-ROM. For example, if CADDIS1 is implemented solely in HTML it could easily be
distributed on a CD-ROM and viewed using a browser such as Internet Explorer or Netscape.
CADDIS2 and CADDIS3 include dynamic components that could be implemented on a CD-ROM
using a "localhost" server that, in essence, creates a local web server on the user's computer,
potentially providing all the functionality of the web served version. There are several disadvantages
of a CD-ROM based approach. It is difficult to keep the CD-ROM version up to date, especially if
each user is allowed to contribute to the building of CADDIS. The CD-ROM based user could not
take advantage of other user's input, gain access to external data sources or GIS systems, or
contribute their input to the continued development of the system. Also, if internet access is not
B-2
-------
available then hyperlinked sites would not be available, limiting access to information that might be
available on States, EPA, or other water resources web sites . Potential advantages are that access
to the Internet will not be required, and performance will not be based on the performance of the
CADDIS server.
Option 1 (CADDIS1):
CADDIS1 is envisioned as a guide to both general and stream specific information sources and data
relevant to stressor identification. The core of CADDIS1 would be built on hyperlinking components
of the Stressor Identification Guidance Document. This could be somewhat similar to the Adobe
Acrobat verison of the Guidance Document but with a wizard rather than Table of Contents
interface. The wizard approach could step the user through the SI process with hyperlinks to web
sites with more information regarding a particular topic in the Guidance Document. This could
include web sites for:
State, Federal and local governments, local watershed groups (e.g., Ohio Environmental
Protection Agency Division of Surface Water),
mapping (e.g., EnviroMapper),
metadata (e.g., EIMS),
observational data (e.g., STORET), and
decision and data analysis (e.g., StatWiz).
CADDIS1 could present the user with hyperlinks to information and help pages describing each
component of the SI process. Each description could contain further hyperlinks to examples for the
particular component of the SI process. The examples section of the Guidance Document could be
extensively hyperlinked to provide information at several different levels or scales of the technical
analysis. Links to downloadable SI worksheets in the form of spreadsheet templates could also be
provided to allow the user to complete the SI process steps off-line. In turn, this capability could
allow users to submit examples and completed worksheets. The submitted information and examples
would then be made available to other users in the form of static information or examples and case
studies of implementing the SI process.
A secondary component of CADDIS1 could tailor the information accessible to each user by
providing a map of the United States that leads to a web page specific to each State. Then, each
State's web page could provide links to relevant EPA sites, non-profit organizations sites, local
watershed groups, as well as the web site of the state's water quality management agency.
Figure B-l shows an example of the web interface that could be used in CADDIS1. Links from the
SI process components and from a map of the U.S. can be made available directly from this page.
These links could access help files, information sources, databases, or State's web-pages. Additional
links to allow downloading and uploading worksheets could also be made available. Databases could
be downloaded for external analysis, but data analysis capabilities will not be available in CADDIS1.
B-3
-------
31 ADIMS Minosiift-1ril'^i 1 xpli-tirr
File Edit 'Jiqv =av writes Help
¦J'Back - J _] l3 J^ea-ci __j =avwrites J-Kedia .J -JjT ^ __si ' \ &~
oj h~irr«:'/"f-?.^'25.1 ^.2.7.4,c-ri'~di:~/ndex h-rrl
~ ^oea'ch *A/eb search at* p^':Rar"; Qpaqelnfo *• [£_|Up
I- n|xj
,V
-------
Technical Tools
The features of CADDIS1 are limited to allow CADDIS to be implemented solely in static HTML.
This minimize the resources required to built CADDIS and limits the resources required to serve
CADDIS on the internet. This also limits the dynamic interactivity of CADDIS and doesn't allow
for dynamic updating or saving user information. This system can help inform the user, but will not
directly assist with analysis and implementation of the SI process.
Resources Required
The steps envisioned for building CADDIS1 include:
write an HTML wizard to the Stressor Identification Guidance Document, and
build and populate each State specific web page.
Writing the HTML wizard would be relatively straightforward and could probably be accomplished
in a few months. Building and populating each State specific web page would be more resource
intensive, although this information may already be compiled and available. Alternatively, CADDIS
Option 2 (CADDIS2) provides an approach that allows the State specific pages to be built gradually
by the CADDIS developer within the CADDIS system, while augmented by the CADDIS users. In
this way CADDIS2 could be more resource efficient alternative than CADDIS1.
Option 2 (CADDIS2):
The move from CADDIS1 to CADDIS2 is largely a move from a static to a dynamic interactive tool
that could have a user updating component. The interactive tool will allow the user to implement the
SI process within CADDIS and will provide direct access to some analysis capabilities to facilitate
the implementation.
The core component of CADDIS2 is a stream specific wizard that allows the user to enter, store, and
retrieve information for their stream application. The wizard could step the user through each phase
of the SI process, building a record for a stream in a database of all streams for which CADDIS has
been applied. Each application can be stored and retrieved at will, allowing a database of
applications to be developed on-line. This database will effectively provide a resource tool for other
potential users of the system, who will be able to review case studies and examples, or even copy
them as a template for their own stream specific or water management problem.
In addition to allowing the user to complete the SI process electronically within CADDIS2, the user
will also be provided access to databases and other information sources. This option of CADDIS
will also include some data analysis functionality, so that data that are retrieved from external
sources can be analyzed on-line, and the results can be used to populate that SI process worksheets
and spreadsheets, which will be stored in an internal database for future review and editing.
B-5
-------
CADDIS2 will provide access to the information sources specified in CADDIS1, and will also
provide on-line implementation with analysis functionality. Analysis capabilities will be limited to
individual components, so that retrieved external data, for example, can be analyzed directly for the
user to input the results in the SI process pages, but will not also include a capability for taking a
holistic view of the SI process in which all of the inputs are integrated into a single model. CADDIS2
will provide access to information as specified in CADDIS1, as well as databases, GIS overlays, data
analysis functionality and a presentation tool.
In application the wizard is a series of HTML forms presented to the user in web pages. The
following prototype web tool presents one approach to providing this capability and provides a
means for exploring and comparing the potential of the various web tools available. The first step
or page allows the user to either enter a new stream name in a text form or to chose a previously
entered stream from a select form (Figure B-2). The streams made available in the select form could
be dynamically generated from the database of previously entered streams. Continually updating a
stream database and providing access to the database through the stream name would allow the user
to stop and return to the wizard at the most recently completed step without starting over. This page
could be preceded by a user login screen. This could serve to monitor the use of CADDIS2.
Once a stream name is entered the next step could be to provide the stream a geographical context
If the stream does not already exist in the database the user could be presented with a map of the
United States with clickable States, Watersheds, and/or EPA Regions. Figure B-3 provides an
example based on a State map. Clicking on a State adds a "State" field to the stream database and
allows CADDIS to guide the user to State specific information and data sources. If the stream
already exists in the stream database the geographical context could be provided by the State field
in the database. In this manner each CADDIS user always has a geographical context that CADDIS
can use to refine and tailor itself to the user. A State database would be access by CADDIS that
provides links to State specific, as well as, regional and nationwide information sources. If the user
discovers information sources not originally provided by CADDIS, the user could add this hyperlink
to the State or US database via a text form. Subsequent users of CADDIS then have the advantage
of these updated links databases. This allows CADDIS to be continually updated with each
application.
Once the geographical context of the CADDIS user has been identified the user moves to the Define
the Impairment page (Figure B-4). This page could provide a text field form into which the stream
impairment information can be entered. If impairments for the stream were previously entered, the
text field could be dynamically filled from the stream database and edited as necessary. The new or
edited impairment information could then be stored in the stream database. On the Define the
Impairment page with the impairment form would be links to aid the user in defining the
impairments if they are have not already been identified. One of these links could be a mapping and
data retrieval tool. The user could be presented with a map of the relevant State with streams,
watersheds, towns, roads, gaging stations, water quality monitoring stations, NPDES point sources,
land use, and other relevant GIS data layers. Ideally in such a tool, each GIS data layer would provide
clickable access to raw data that could be imported into a data analysis wizard.
B-6
-------
HE]
~
•» tVGo Lnki "
Water
Causa! Ara'ysis/Diagnosis Decision Information Svslen
<£',ntis4
© Enfcsr Stfsam
The first step in CADDIS is to enter information for your stream or retrieve
information previously entered information for your stream. If your stream has been
entered in the database it will appear in the select form below otherwise enter the
stream name in the text field below
Enter a new stream:
Pick a previously entered
stream:
Note. The above selac! form ir
(stream, xmi). Entering a new
a>ng built dynamic ly from an XML til
earn starts a new record in stream.)
Ny.J->
M^nu rftaH1"1 -r,r i ;ci=> Jj [nrs-T.^r
Figure B-2. Initial CADDIS wizard page for entering the stream name. The stream name
select form is populated from a database of previously entered streams.
B-7
-------
^ ^'oo Lirls f>
Water
Cause! Analysis/Diagnosis Decision Information Svsten
i':!@ ®l Wat* pi Sfffct i Rcsaa
-------
UiAIMMS *
Fila Fiiih F^'nte-*-. Tn.il--. Hpjp
•!¦¦¦¦ Bad ~ _! _] llii ji'.^ar.:}- [-avorites /Media j3 ...jj' _p _/' *¦ j -*-> -
AdJicii -j.'] //tiJ.iij1:. 1 . 124;c =icdiv.ir de."itr>"::
» &fexch'#eb Site »*"•««»• Qpojclnfe - f?™|Ut *
•'Ijl. Li'lfi
Water
Causal «ralysis/Biaynusis Decision information System
W CADDIS Wizard
© Dtfirri Ir:;.p.s;f»vints
©Description ©Information ©Tools
-
List Little Scioto River impairments:
Loss of fish and benthic invertebrate species
Decrease in the number of individual fish
Increase in the relative weight of fish
Note; The above-1eui form was filled dynamicly from stream.>:ml based on trie previous!)'
chosen stream nam* Change5 he^e would te* *a-/ecf back to stream -{mi
Care, id ate L c.Ubd:- ••••>
a r^eiu ready fur t_se J Irr.erre:
Figure B-4. Define the Impairment page. This page includes a link to a Define the Impairment
Help page and a text field for entering and editing impairments. The impairments are then stored
in a database referenced by stream name.
B-9
-------
Currently EPA's EnviroMapper (maps.epa.gov/enviromapper) provides some of the desirable
functionality. The data layers EnviroMapper provides that would be most relevant to CADDIS
include streams (apparently Reach File 1), towns, watersheds, counties, and point sources. Data
layers not apparently provided by EnviroMapper include detailed streams (Reach File 3), Federal and
State water quality monitoring stations, stream gaging streams, local government or watershed group
monitoring stations. EnviroMapper also does not provide a direct data retrieval mechanism. EPA's
EnviroFacts web page provides access to some of the data relevant to the CADDIS user that would
be identified through EnviroMapper, such as some NPDES point source information, however, direct
query of EnviroFacts databases by the public is currently restricted. EPA's Environmental
Information Management System (EIMS) provides metadata for data sets, databases, documents,
models, projects, and spatial data. However, in general EIMS does not appear to provide access to
the underlying data. EIMS could be useful to CADDIS as a guide to data sources other than data
sources that already have been identified. An alternative to these existing tools would be
development of a mapping and data retrieval tool specifically designed for CADDIS (discussed
further in Options 2 and 3 with example available on the CADDIS web site).
STORET could be a very useful data source but it may be difficult to provide a wizard to access the
data. STORET is split between the STORET Legacy Data Center and Modernized STORET. The
STORET Legacy Data Center includes data prior to January 1, 1999 while Modernized STORET
includes post January 1, 1999 data. Neither of these data sources appear to allow data query other
than through forms on the STORET web page. Data retrievals are run overnight and requesters are
notified via email of a web page where the data can be "picked up".
After the impairments are entered and saved to the Stream database the wizard sends the user to
Candidate Causes page with a similar look and feel to the Impairment page (Figure B-5). Candidate
Causes can be entered into a text field form that is saved to the Streams database. Alternatively, help
on developing a list of candidate causes can be accessed. This could provide links to examples,
maps, local, state, regional, nationwide information data sources, as well as a link to a data analysis
wizard. Ideally data and information would be accessed visually with a map interface. Of particular
importance would be links to facilitate conceptual model building. Conceptual models are often built
in graphical form. Although Open Source free software is available (e.g., dia) there doesn't appear
to be a clear path forward for developing a web-based conceptual model drawing tool. One
possibility may lie in scalable vector graphics (SVG).
B-10
-------
hie bck View havcntes lools heto
*j-> back ~ _' _] lii search _JhaVOf:tes pMecia J? __j- ~ | ~
A--j:] -ittp: lb j^i':6dJis/-nde,x.ht-rii
- fi^eaitWet ""I*"-* Qpjsemfo * r«-p
l"V »-U _|i iki
Causal Analysis/Diagnosis Decision Information SystBn ,:::EE,
| tpooM Df; lit Bins ti Wm tni OffiM il tostifi t«l tedp* i •« ' :"".; : 7
@ CADDIS Wizard "= \
® Candidate Causes
Description ® Information © Tools
List Little Scioto River impairment candidate causes:
IHabitat alteration: embeded stream and deepened channel *1
{Exposure to FAHs
IHetal contamination
!Ammonia toxicity
Ilow DO
{Nutrient enrichment
Note' The above t?nt firm was filled dynamidy from lirearo.^ml based on the previous
chosen stream name Changes here would be saved hack to stream «ml
"riaEviUer-uti —>
.y] "¦"'1-ri il i •sd'Jy ful u>e.
J I: ilsm-l
Figure B-5. List Candidate Causes page.
B-ll
-------
The Analyze Evidence page then follows the List Candidate Causes page (FigureB-6). This page
could provide a link to a data analysis wizard, as well as, guidance documents for interpreting the
data analysis results. These guidance documents could include:
1. general statistics guidance,
2. the Analyze Evidence Chapter in the Stressor Identification Guidance Documents, and
3. State and EPA guidance on interpreting data and water quality criteria.
A web-based water quality data analysis wizard could be built with modules designed, for example,
to facilitate the comparison of up and down stream biological data, analyze data for trends, or
compare data to water quality criteria or thresholds. A data analysis wizard could also provide for
visualization and mapping of the stream and data. An important piece of a data analysis wizard
would be a data import wizard to facilitate moving data from a water quality database (e.g.,
STORET) or from a standard software package (e.g., Excel). The data analysis wizard should also
provide summary results that facilitate completing each step of the SI process. An example of such
as data analysis wizard that implements some of these features is available at
www.neptuneandco.com/StatWiz. Analysis and interpretation of data could be a component in each
step of the SI process. Therefore the links available on the Analyze Evidence page could be viewed
as a compilation of all the links available on the help pages for the previous pages.
After the Impairments and Candidate Causes are defined and the Evidence Analyzed the CADDIS
wizard provides a Characterization of Causes page that allows all Candidate Cause and Impairment
combinations to be scored in several causal considerations categories (Figure B-7, the causal
considerations presented in this figure are not meant to be exhaustive but only serve as an example).
Scores are input in a matrix of text forms and saved to the Stream database. These scores are
synthesized and presented in the Identify Cause page. This page could provide a summary
visualization of the Cause characterization table to aid in Identifying causes.
Overall, CADDIS2 will provide access to information sources similar to CADDIS1, but will also
provide access to databases with retrieval capabilities for standard database structures (Oracle,
ACCESS, Excel, text files), and access to GIS information and coverages. If data are pulled or are
provided from external sources, a data analysis wizard can be made available that is tailored to the
needs of the SI process steps. The user will be able to use and manipulate information that is
gathered, and will be provided a capability to implement the SI process directly on-line. Examples
will be stored in an internal database format for future retrieval and editing (with some password
protection probably), and these examples will add to the list of case studies and examples that can
be made available to all potential users. In this way, CADDIS will be continually updated by new
applications. Once the user-supplied information is input into the CADDIS forms and worksheets,
the stressors can be scored following the SI guidance and the most likely stressors using this
mechanism can be identified. This stops short of providing an integrated modeling capability that
is deferred to Option 3.
B-12
-------
U'AM"'IS Mi rnlt-1 rxpliin-i
File Ed t View =avDrit?r Tods Help
-I—: Fa"k - 1 ] '1'; ^.v ci=i=ih"h j =.5i'vnr,ihfl"- f Vl^ri^ J} Ji ,.| T j
Address J1] -rfitp:,,1't.:.L^b.lb.^J4/caddis/incec.hcrrl
Search /vcb Search Site "=q*F*nk r'_^)r»-.g= irf0 » ^jLlp "
I- ~ !*!
* Links.
Water
Causal Decision Intormation Systeoi
' Description @ Information W Tools
m-
Little Scioto Kiwar
Ohio
Ohio EPA Division of Surface Water
EPA Region 5
USGS
Note: A text field four- could So added here to aiiow the user to add useful fnks to
a mi H^s.anlj tile such that when, for example, anj» Ohio stream is entered
11 ¦1 :i nhs section is dynamic ly augmented (this wnuiri need to ha cleaned
u l)
Develop a conceptual model
Search for relevant data
EPA ErwiroMapper of Little Scioto River
SVG Map of Little Scioto River
Cha-octer13 Ccuss; >
•jJ Ment ready for use Internet
Figure B-6. Analyze Evidence page.
B-13
-------
EJ
~ f jo L n<5 yy
Causal Analysis/Diagnosis Decision informa:ion Systsm
I iBwsf l»; Till Iffct i< iiw ml Ufa if fieswdt mi toriepriwu ol tte Unied She® Eflviiiftfliflftt Protacio«t Aacncy f \ '.!!"
:=:
® CADDIS Wizard
© Characterize Causes
Description m Information m Tools
CHARACTERIZATION OF CAUSES
Little Scioto River
Loss of fish and benthic invertebrate species
1 Habitat i
i alteration: j
embeded i
stream i
i and i
Causal deepened i Exposure
Metal
Ammonia
Low
Nutrient
Consideration channel j to PAHs
contamination
toxicity
DO
enrichment
cooccurence
p__
|ne
,Nb
p__
|ne
temporality
|NE
|ne
|ne
|ne
|ne
Decrease in the number of individual fish
cooccurence
|ne
|ne
|ne
|ne
|ne
|ne
temporality
|ne
|ne
|ne
|ne
|ne
|ne
Increase in the relative weight of fish
|NE
cooccurence |NE
|ne
|ne
|ne
|ne
temporality |NE
|ne
|NE
|ne
|ne
|ne
p |pvt rvPp=; —>
Scoring
NE: no evidence
NA: not applicable
i itfj Menu ready for use 15$ Internet
Figure B-7. Characterization of Causes page. Web based worksheet for organizing
impairment and causal relationships. The causal considerations list here are not meant to be
exhaustive but are provide for illustrative purposes.
B-14
-------
Technical Tools
CADDIS2 requires web tools that enable CADDIS to be dynamic and interactive. The dynamic
components of the prototype CADDIS were built using HTML, Common Gateway Interface (CGI)
Perl scripting, and Perl XML (Extensible Markup Language) parsing. Alternative approaches to CGI
include Java Server Pages (JSP), Extensible Server Pages (XSP), and Active Server Pages (ASP).
The example data analysis wizard, StatWiz (www.neptuneandco.com/statwiz), uses CGI Perl as a
web interface to R, an Open Source statistical programming language. R provides a powerful
statistical package that can be tailored to the needs of CADDIS including probabilistic simulation.
The prototype CADDIS uses XML for data storage and retrieval. Alternatively a Relational Database
Management System (RDMS) such as Oracle could be used. A RDMS would likely complicate
CADDIS and limit portability without providing much value added, since XML is well-suited for
storing and managing data and information including features such as text, figures and drawings, and
numerical information.
The missing functionality from CADDIS2 that would be highly desirable includes data retrieval
wizards for some of the standard water quality databases including STORET and USGS water
quality and stream flow databases. Providing CADDIS with this functionality does not appear to be
achievable in the near future. Links to these data sources could be included in CADDIS and data
import wizards could be developed that facilitate import of the data into a data analysis wizard. The
EIMS system is designed to allow remote queries and could provide a valuable information resource
for CADDIS. The information in EIMS is largely metadata thus does not appear to provide a source
for data. In general, current sources of the type of information that CADDIS would use are not
organized in a way that allows easy or suitable retrieval. Access to standard databases (e.g., Oracle
or ACCESS) can be provided through the data analysis wizard or more directly.
Resources Required
The resources required to build CADDIS2 would include the resources required to build the
hyperlinked Stressor Identification Guidance Document and the State specific pages outlined in
CADDIS1. Much of the dynamic capabilities of CADDIS2 are exemplified in the prototype of
CADDIS. Fully implementing these features, as well as, fully developing a surface water data
analysis wizard would require additional resources.
The main functions that are needed to make the system functional (e.g., linking to information
sources, obtaining data from external data bases, GIS functionality, data analysis functionality,
interactive implementation of the SI process) can probably be implemented as version zero in
approximately 1 year after initial design with sufficient resources. The initial system will require peer
review, beta testing and iteration before it can be released for more general use. The system will also
be updated from examples that are implemented and recorded. System management will be
necessary, although the system can be built such that it is self-managed to some extent. When the
design steps are undertaken, some consideration will need to be given to system management,
B-15
-------
hosting, centralization and distribution across networks, and hardware requirements (storage capacity
and operational functionality) for the server.
A CADDIS tailored GIS approach would provide a much more functional system than can be
provided by EnviroMapper, however significant resources may be required to create the State
specific maps and data organization. This would be somewhat similar to the approach used by
BASINS and could require computer server resources to store, manage and serve.
Another resources consideration is that CADDIS2 places a larger burden on the server since the
dynamics of CADDIS are based on some form of server based scripting (CGI, JSP, etc.). CADDIS2
also includes a data storage and retrieval component, which further complicates the web server
configuration. The data analysis wizard component could provide a substantial computational burden
on the server as well, depending on the level of use CADDIS sees. These issues can be managed by
increasing hardware capabilities, but might need to be considered early in the planning and design
process. Given that CADDIS2 will not be ready for about 2 years, and hardware capabilities continue
to improve at an alarming rate, hardware specifications might become less of an issue. Similar
considerations apply for web-based software tools that might be applicable. The capabilities of these
tools has increased dramatically in the last year or so, allowing different software to be linked to
produce a fully operational integrated system. For example, web-based tools such as Perl/CGI and
JSP now can work hand-in-hand with HTML and link with data bases, GIS programs, statistical
software and graphical software. It will be important to stay on top of hardware and software issues,
so that the most up to date and necessary computer technology is used when CADDIS2 is built, but
it should be expected that the appropriate computer resources will exist.
Option 3 (CADDIS3)
In addition to the functionality provided by Options 1 and 2, CADDIS3 would encompass analytical
tools that would facilitate modeling of aquatic systems in the context of the SI guidance. That is,
fully coupled systems could be developed in which relationships among stressors can be identified
and characterized simultaneously. This would be a more holistic approach to stressor identification
than provided in CADDIS2. Stressor models could be developed using deterministic or probabilistic
approaches, either of which could be extended to accommodate decision analysis if desired.
Probabilistic modeling will allow uncertainty to be propagated through each step of the SI process.
Using this approach uncertainty can be managed effectively and sensitivity analysis can be performed
to identify the major sources of uncertainty in the system. Steps that would need to be included in
CADDIS3 to facilitate this type of model-based approach (beyond those already specified in
CADDIS2) would include building of conceptual models, translating conceptual models into
numerical models, specifying the numerical models, simulation, uncertainty and sensitivity analysis,
and model updating as more information is collected. CADDIS3 would hence encompass analytical
tools that allow uncertainty to be propagated through each step of the SI process. Ideally this would
conducted within a Bayesian paradigm incorporating prior probabilities and data. Prior probabilities
B-16
-------
could be generated from expert judgment, prior data collection and analysis efforts as various scales
ranging from EMAP at a national scale to studies performed in nearby watersheds.
In any probabilistic modeling system such as this, propagating uncertainty, whether in a Bayesian
approach or in a classical probabilistic simulation approach, generally requires some form of Monte
Carlo sampling. Influence diagrams are a visually and logically appealing way to implement a
probabilistic simulation of a conceptual model. Figure B-8 presents the example conceptual model
from the Stressor Identification Guidance Document (EPA 2000). Conceptual models are typically
represented in this fashion as flow diagrams. The relationships defined by the flow diagram can be
interpreted as defining a joint probability distribution for all components of the conceptual model.
Thus if the conceptual model is developed and simulated as an influence diagram the results are
probability distributions for all the components of the conceptual model. Using an appropriate
simulation algorithm (e.g., Markov Chain Monte Carlo, MCMC) allows the joint probability
distribution to be integrated to define the relationship between the uncertainty in stressors and the
associated uncertainty in the responses. The real appeal to this approach is that all available
information, regardless of its source (primary data, secondary data, expert opinion, non-expert
opinion), can be appropriately incorporated into the analysis by expressing the level of confidence
in the data source through the variance parameter of the information probability distribution. If the
input distributions are objectively parameterized then the analysis will represent the state of
knowledge qualitatively described by the conceptual model.
A secondary appeal of this approach is it provides a rigorous mechanism and guide for collecting
more information and iteratively updating the conceptual modeling. An ideal approach is to initially
develop a conceptual model influence diagram that places minimal resources on collecting new data
and information collection. Input probability distributions are specified that reflect the current state
of knowledge and the conceptual model influence diagram is simulated. The sensitivities of
conceptual model responses to inputs provides a rigorous means of evaluating the cost and value of
collecting more information and iteratively updating the conceptual model.
Ideally CADDIS3 would provide a mechanism for drawing, storing, specifying, and solving
conceptual model influence diagrams. There are threes basic issues to be resolved in implementing
such a web-based tool
drawing and storing the conceptual model influence diagrams,
providing guidance and data analysis tools to the user for specifying input probability
distributions, and
solving a conceptual model influence diagram.
There exists standalone software that allow influence diagrams to be drawn, specified, and solved,
including Analytica and WinBUGs. Analytica is a commercial Windows package that facilitates
model building through formal decision-based influence diagrams that can then be solved using a
classical Monte Carlo simulation approach. WinBUGs (www.mrc-bsu.cam.ac.uk/bugs/) is a free
(currently free but may not be free in the future) Windows package that implements a Bayesian
B-17
-------
u~y ,n \
^ \ Migration /
Adult Salmon
Detritus
Fry
Macmirwartotntes
Firigeriings
^ Survival ^
w/ Out \
Migration /
Juveniles
Figure B-8. Example conceptual model diagram from the Stressor Identification
Guidance Document (EPA 2000).
B-18
-------
MCMC approach to solving complex probabilistic models. One advantage of WinBUGs is that it
has built in functionality for Bayesian probability theory, which would facilitate updating of a model
once new information is collected. Other software such as GoldSim (Golder Associates) is also
available for building environmental models and performing Monte Carlo simulations. Some of the
same functionality could be developed using R, although some interaction between platforms is also
possible (for example, R can call WinBUGs programs). These software products are continually
being updated to add functionality, and between them can be expected to provide the necessary
capabilities for CADDIS3 in the time frame in which this system might be developed.
The path forward is not clear on how to implement a similar web-based tool. An approach that would
seem to have a lot of potential is to use Extensible Markup Language (XML), Extensible Stylesheet
Language (XSL), and scalable vector graphics (SVG) for the web-based drawing and storing of
conceptual model influence diagrams. Since SVG is a text-based graphic description language
implemented in XML both the graphical description and the probability distributions of the influence
diagram can be stored and manipulated as an integrated fashion. Using XML/XSL to draw, store,
and present the conceptual model also allows the conceptual model to be integrated with the stream
information and data analysis described under CADDIS2 in a text based non-proprietary standardized
form that is transportable between operating systems. An SVG approach could also have potential
for visually eliciting and specifying input probability distributions for a conceptual model. Solving
the conceptual model influence diagram could in one of a number of programming languages
including R, Java, or C++. Alternatively a SVG based conceptual model influence diagram could
generate input files for BUGS (the command line verison of WinBUGS) to provide the influence
diagram solution engine. Also, a new version of Analytica is being written using XML and Java.
The XML and SVG approach could be mixed with drawing tools built in Java or simulation tools
such as Hydra (http://software.biostat.washington.edu/statsoft/MCMC/Hydra). Hydra is an
open-source, platform-neutral library programmed in Java for performing MCMC sampling.
Technical Tools
CADDIS3 would require development of web-based interactive conceptual model influence diagram
drawing and solution software. How this might be achieved is not entirely clear at this time, but it
is clear that such an approach will be feasible in the time frame of interest because of the rapid
advancements that are being made in web-based software and linking to analytical software.
Potential solutions that are available now include XML, XSL, SVG, and Java for developing and
managing the web-based influence diagram drawing software, as well as, for a user interface to
MCMC software such as Hydra and BUGS. MCMC sampling could also require significant
computational resources depending on the level of use of CADDIS. Other potential solutions include
using a web-based version of Analytica. Given the time frame of interest, what will be important is
to keep track of software developments as they occur so that once CADDIS3 is designed, the
appropriate software is identified.
B-19
-------
Resources Required
Given the uncertainty in the development of CADDIS3, determining the resources required at this
time is very difficult. The web-based tools that would be relied upon are developing rapidly and may
provide a clearer path forward in the near future. One might expect that a system such as CADDIS3
could be built within 3 years after completion of Option 2 (give or take a couple of years), depending
on the features required, and depending on the nature of Option 2 when it is implemented. Also, it
would be possible to add functionality of this type in a piecemeal fashion, highlighting each step of
the SI process in turn towards development of a complete system. This way, each component could
be tested for feasibility before the next component is attempted.
Summary
Three Options for a path froward of the design of CADDIS have been presented. CADDIS1 provides
a simple interface to information sources relevant to the SI guidance along with a capability to
download SI worksheets, and to upload completed worksheets into a database that is continually
updated. CADDIS1 will be essentially a static system. CADDIS2 will provide dynamic capabilities
in terms of on-line implementation of CADDIS, interaction with databases, GIS coverages, and data
analysis functionality. Scoring rules can also be built into CADDIS2, so that stressor identification
will have a numerical basis, but a complete holistic system is reserved for CADDIS3. This final
option for CADDIS will also include modeling capabilities so that SI problems can be solved in a
fully integrated and coupled system. Option 2 requires that the user solve their SI problem outside
of CADDIS, whereas, Option 3 will provide the capability for CADDIS to numerically solve the
problem, or will provide a numerical tool that will help managers solve SI applications on-line.
CADDIS1 and CADDIS2 provide valuable SI functionality that could be rapidly developed while
laying a foundation for continually improving CADDIS in a resource efficient manner. The
development steps for CADDIS3 are unclear at this time but as web-based tools continue to rapidly
develop the vision for the path forward should clarify in the near future, or in the time frame in
which CADDIS2 would be built.
References
U.S. Environmental Protection Agency (USEPA). 2000. Stressor Identification Guidance
Document. EPA-8220B-00-025.
B-20
-------
Appendix B-l. Table of potential CADDIS functionality.
Function
CADDIS
Option
Comp-
lexity
Level of
Effort
Comments
STRESSOR IDENTIFICATION PROCESS
Stressor Identification Guidance Document online
CADDIS1
low
low
Could be easily done.
CADDIS Wizard
CADDIS2
low
medium
Partially implemented in
CADDIS prototype.
DEFINE THE IMPAIRMENT
Links to guidance on stakeholder involvement
CADDIS1
low
high
Requires locating appropriate
documents and web sites.
Provide regulatory/programmatic context of impairment
Links with regulations and water management programs
CWA: 305(b), 303(d) TMDLs links, 319, NPDES, ...
State and Local watershed management links
CADDIS1
low
high
This type of functionality is
demonstrated in the CADDIS
prototype.
Biological data visualization
This will require the user to upload their data implying
some pre-specified form
Plotting the data will require a statistical graphics
package (R)
Underlay GIS coverages
Provide a GIS base map search (EIMS,
EnviroMapper, EnviroFacts)
Merge and load GIS data with the impairment
data into a statistical graphics package (R)
CADDIS2
CADDIS2
CADDIS2
CADDIS2
low
high
high
high
low
medium
high
medium
Demonstrated in StatWiz
and the CADDIS prototype.
-------
Function
CADDIS
Option
Comp-
lexity
Level of
Effort
Comments
Gather other biological data
Local geographic search for other relevant biological
data
National geographic search for other relevant biological
data
Identify
Merge and load biological data into a statistical
graphics package
CADDIS1
CADDIS1
CADDIS2
low
low
high
high
high
medium
Demonstrated in CADDIS
prototype.
Demonstrated in StatWiz.
Gather other biological information
• Internet search for relevant biological information
CADDIS1
low
high
There are likely to be many
sources for this type of
information.
LIST CANDIDATE CAUSES
Map data on sources and stresses
Provide a GIS base map search (links to EIMS,
EnviroMapper, EnviroFacts)
Merge and load maps, sources, and stresses with the
impairment data into a statistical graphics package
CADDIS1
CADDIS2
low
high
high
high
Demonstrated in
EnviroMapper and CADDIS
prototype.
Demonstrated in StatWiz.
Identify common sources and stressors in impairment area
Link to State/local water management programs
Query State/local water management
databases/information
CADDIS1
CADDIS2
low
high
high
high
Demonstrated in CADDIS
prototype.
Develop Conceptual Model
Links to guidance on developing conceptual models
Links to library of existing conceptual models
Wizard for web-based Conceptual model drawing tool
CADDIS1
CADDIS1
CADDIS3
low
low
high
low
medium
high
Requires locating sources.
Requires building library
Difficult but a SVG base
approach looks interesting.
-------
Function
CADDIS
Option
Comp-
lexity
Level of
Effort
Comments
Gather other biological information
Internet search for relevant stressor information
CADDIS1
low
high
Requires access to
State/Tribe, Federal, citizen
watershed groups and other
web sites.
ANALYZE EVIDENCE
Links to EPA and other guidance on statistics and modeling.
This could include links to EPA's Quality program, such as
DQOs, DQA, etc.
CADDIS1
low
low
Providing links to guidance is
simple.
Case studies and Examples links
CADDIS1
low
medium
Building the capability is
straightforward. But, effort
will be required to maintain it
as it grows.
Analyze Data
Implement conceptual model
Associate candidate causes with observed effects
Associate observed effects with mitigation or
manipulation of causes
• Combine data from site with other information
CADDIS2
CADDIS3
medium
high
medium
high
Possibilities demonstrated in
StatWiz.
Implies a Bayesian approach,
difficult
Causal Mechanisms
Decision Logic depicting candidate causes by dimming
eliminated causes, and highlighting remaining causes
based on mitigation/manipulation analysis
CADDIS2
high
high
Possibilities demonstrated in
the StatWiz decision logic
approach
-------
Function
CADDIS
Option
Comp-
lexity
Level of
Effort
Comments
Search/link and implement TIE methods & results
CADDIS1
low
high
High level of maintenance as
more examples are made
available.
Search/link to remediation and restoration projects
CADDIS1
low
high
High level of maintenance as
more examples are made
available.
CHARACTERIZE CAUSE
Eliminate
Links to philosophy of approach
• Examples links
Upstream/downstream
Time precedence
Site/reference comparison
CADDIS1
CADDIS1
CADDIS1
CADDIS1
CADDIS1
low
low
high
high
high
low
high
medium
medium
medium
Examples will prove to
require a high level of effort
in the long term.
Statwiz functions.
Diagnose
Links to philosophy of approach
• Examples links
Links to specific diagnostic tools and keys
CADDIS1
CADDIS1
CADDIS1
low
low
low
low
high
low
Much of CADDIS1 is simple
to implement. The high LOE
is because of maintenance.
Strength of evidence analysis
Links to philosophy of approach and examples
Link to Causal Considerations
Modeling Wizard
Weighting Wizard
CADDIS1
CADDIS1
CADDIS3
CADDIS2
low
low
high
high
low
low
high
medium
Links are simple once the
sources are identified.
Modeling is difficult.
Weighting schemes are not
difficult but are prone to
decision biases.
-------
Function
CADDIS
Comp-
Level of
Comments
Option
lexity
Effort
Identify cause
Link to suggested report outline and examples
CADDIS1
low
medium
Requires developing material
Generate organized material for report
CADDIS2
medium
medium
XML will facilitate
implementation
Requires defining reporting
format and requirements
Evaluate uncertainties and sensitivities
Guidance links
CADDIS1
low
low
Sources of information are
well known to us.
• Examples links
CADDIS1
low
high
High maintenance.
Probabilistic modeling wizard
CADDIS3
high
high
Bayesian approach with web-
based influence diagram
drawing and simulating tool.
NEXT STEPS
Iteration Options
Links on combining sensitivity analysis and data
CADDIS1
low
low
Simple links to sources of
collection
information.
Link: "Reconsider the impairment: False positives"
CADDIS1
low
low
Adding functionality is more
• Value of information links
CADDIS1
low
low
complex and will require a
greater LOE.
-------
U.S. EPA Information Sources
Prepared by
MENZIE • CURA & ASSOCIATES, INC.
Environmental Consultants
One Courthouse Lane, Suite Two • Chelmsford, Massachusetts 01824
Phone (978) 453-4300 • Fax (978) 453-7260
Prepared for T N & Associates, Inc.
under Contract Number 68-C-98-187
This document has been reviewed and approved for use as background for
the CADDIS Workshop by the EPA Work Assignment Manager.
B-26
-------
This memo summarizes some of the key features of several information sources developed
and/or maintained by Environmental Protection Agency (EPA) that may be useful to users of the
Stressor Identification (SI) Guidance Document (USEPA, 2001). For each source we provide:
The information contained in each source that might be useful in the SI process;
The platforms/structure of the database/system; and
Whether the information has a locational component and if so, whether it is
georeferenced.
These characteristics are summarized in Table B-l and are described in more detail in the text
below. The information sources we reviewed are listed in alphabetical order, not by their
usefulness or application to particular steps in the Stressor Identification Process.
AQUATOX
http://www.epa.gov/ost/models/aquatox/
The U.S. EPA has developed a freshwater ecosystem simulation model called AQUATOX that
predicts the fate of various pollutants, such as nutrients and organic chemicals, and their effects
on the ecosystem, including fish, invertebrates, and aquatic plants. The AQUATOX model can
help someone using the SI process to establish the cause and effect relationships between
chemical water quality, the physical environment, and aquatic life. It can test sensitivity of
biological responses to individual stressors. It may also help to determine the most important of
several environmental stressors, e.g. where there are both nutrients and toxic pollutants.
AQUATOX is a process-based, or mechanistic, ecosystem model that simulates the transfer of
biomass and chemicals from one compartment of the ecosystem to another. It does this by
simultaneously computing important chemical and biological processes over time. AQUATOX
can predict not only the fate of chemicals in aquatic ecosystems, but also their direct and indirect
effects on the resident organisms. It can represent a variety of aquatic ecosystems, including
lakes, reservoirs, ponds, rivers and streams. It also simulates the fate and effects of multiple
environmental stressors, such as nutrients, organic toxicants, temperature and turbidity, and
changing depths and flow regimes. Understanding these cause and effect relationships can help
determine effects that have been shown to be diagnostic of particular stressors (as part of
Characterizing Causes in the Stressor Identification Process).
B-27
-------
Table B-l. Summary of Information Sources for Use in the Stressor Identification Process
Information sources
Use in SI process
Web Address
Database
Platform
Mode of Access
Georeferenced?
AQUATOX
Model cause and effects
in water body of concern
http://www.epa.gov/ost/rn
odels/aquatox/
Windows-based
software
Order from the National
Service Center for
Environmental Publications or
download from web site
No
ASTER
Identify chemical
stressors associated with
particular effects
http://www.epa.gov/med/
databases/aster.html
NA
Request search via
e-mail
No
BASINS
Identify stressors,
quantify impact of
multiple stressors on
watershed
http://www.epa.gov/ost/ba
sins/index.html
NA
Free software
Yes, may be used
with Arc View GIS
software.
ECOTOX
Identify chemical
stressors associated with
particular effects
http://www.epa.gov/
ecotox/
NA
Internet
No
Ecotox Thresholds
Identify chemical
stressors
http ://www. epa.gov/super
fund/resources/
ecotox/index.htm
DOS-based
software
Download from website
No
Envirofacts
Identify candidate
causes (point sources)
http://www.epa.gov/envir
o/index java.html
ORACLE
Internet
Yes
EIMS
Identify candidate
causes and links
between cause and effect
based on case studies
http://www.epa.gov/eims/
eims.html
ORACLE
Internet
Yes
EMAP
Identify candidate
causes and links
between cause and effect
based on case studies
http://www.epa.gov/
emap/
ASCII text files and
ARC/INFO export
files
Internet
Yes
EVISTRA
Identify effects
associated with
particular stressors
http://www.epa.gov/med/
databases/evistra. html
NA
NA
NA
-------
Table B-l. (Continued.) Summary of Information Sources for Use in the Stressor Identification Process
Information sources
Use in SI process
Web Address
Database
Platform
Mode of Access
Geo referenced?
FIELDS
Delineate spatial
distribution of stressor
(includes watershed
mapping feature)
http://www.epa.gov/regio
n5fields/text/pages/
index.html
EQuIS by Earthsoft
& database
developed by
FIELDS Team
Software
Yes
Inventory of Ecological
Restoration Projects
Within the Mid-Atlantic
Integrated Assessment
(MAIA) Region
Identify wetland
stressors (and possibly
links to effects)
http://www.epa.gov/owo
w/wetlands/restore/
http://yosemite l.epa.gov/
water/restorat.nsf/PYPMa
pPg6?0penPage
NA
Internet
No
National Water Quality
Standards Database
(WQSDB)
Identify levels of
stressors associated with
potential effects
http ://www. epa.gov/wqsd
atabase/
NA
Internet
Yes. Linked to on-
line mapping tool,
EnviroMapper.
River Corridor and
Wetland Restoration
Project Directory
Identify wetland
stressors (and possibly
links to effects)
http://www.epa.gov/owo
w/wetlands/restore/
http://yosemite Lepa.gov/
water/restorat. nsf/rpd-
2a. htm?OpenPage
NA
Internet
No
STORET
Identify stressors in
water body of concern
(chemical, nutrient,
other)
http://www.epa.gov/
storet/
UNIX/ ORACLE
database server
Internet or free software
Yes
Toxicity /Residue
Database
Identify chemical
concentration in tissue
associated with effects
http://www.epa.gov/med/
databases/toxresidue.
html
NA
NA
NA
WATERS
Determine status of
waterbodies of concern
http://www.epa.gov/
waters/
Microsoft Access
Internet
Yes. Link to
ArcView or use
EnviroMapper
Notes:
The designation "NA" indicates that we were unable to find information on the World Wide Web to fully describe the information source. The desired
information could most likely be obtained by talking with the persons who developed or maintain these information sources.
-------
AQUATOX is designed to provide a realistic representation of aquatic ecosystems with a
minimal amount of detailed site specific information or site calibration. The following are
required input data:
Loadings to the waterbody (environmental loadings can be from multiple sources)
General site characteristics
Chemical characteristics of any organic toxicant
Biological characteristics of the plants and animals
AQUATOX comes bundled with data libraries that provide default data. This is of particular
importance for the biological data, which are probably the most difficult for a user to obtain.
AQUATOX also allows the user to evaluate model uncertainty by varying the value or
distribution of input parameters, or to test the sensitivity of a given endpoint to specific inputs.
AQUATOX has many potential applications to water management issues and programs,
including water quality criteria and standards, total maximum daily loads (TMDLs), and
ecological risk assessment. AQUATOX can be used to predict ecological responses to proposed
management alternatives. AQUATOX can be helpful where the user needs to understand the
processes relating the chemical and physical environment with the biological community and the
following conditions exist:
Where ecological and biological processes are complex
Where indirect effects are important but difficult to monitor
When one needs to articulate linkages between nutrients and biotic community
Where the environmental conditions may change appreciably
AQUATOX is a Windows-based software program that was developed for EPA by Richard A.
Park and Jonathan S. Clough of Eco Modeling. The model, users guide, technical documentation
and validation reports may be ordered from the National Service Center for Environmental
Publications, or they can be downloaded from the web site. The results of several validation
studies that have been conducted using AQUATOX are also available on the website.
ASTER (Assessment Tools for the Evaluation of Risk)
http://www.epa.gov/med/databases/aster.html
ASTER is a Unix-based computer program that is useful when looking for toxicity information
for pollutants that are less commonly found in the environment or whose toxicity is not well
studied. This database integrates the toxic effects information from Aquatic Toxicity Information
Retrieval (AQUIRE) and the Quantitative Structure Activity Relationships (QSAR) system to
develop toxicity profiles for compounds for which no empirical toxicity information is available.
B-30
-------
"When empirical data are not available, mechanistically-based predictive models are used to
estimate ecotoxicology endpoints, chemical properties, biodegradation, and environmental
partitioning. The QSAR system includes a database of measured physicochemical properties such
as melting point, boiling point, vapor pressure, and water solubility as well as more than 56,000
molecular structures stored as Simplified Molecular Input Line Entry System (SMILES) strings
for specific chemicals." The results found in the Fathead Minnow database compiled by MED-
Duluth are incorporated into the ASTER database.
These toxicity data in this database represent concentrations at which specific biological effects
have been observed. The information in the ECOTOX database can help someone using the SI
process to identify effects that are associated with exposure to particular chemical stressors.
ASTER is currently not available to the public, but affiliates of the government (local, state,
federal, international) or people who are cooperating with or contracted with a government
agency can ask technical support staff to conduct a search for them. To request a search, you can
e-mail technical support at ecotox.support@epa.gov. Be sure to include the CAS # of the
chemical you are interested in. You will receive a toxicity profile that will include the following
types of information:
Chemical Identification - chemical name, CAS number, SMILES and chemical formula
Environmental Exposure Assessment-molecular weight, melting point, boiling point,
vapor pressure, heat vaporization, solubility in water, log P, pKa, adsorption coefficient,
Henry's Law constant, hydrolysis half-life, BOD half life and fugacity
Ecotoxicological Hazard Assessment-species-specific acute toxicity data, chronic
toxicity data, bioconcentration data
Ecological Risk Characterization-some interpretation of values given in exposure
assessment and ecotoxicological hazard assessment sections.
• References
EPA-Duluth hopes to make the ASTER database public sometime in 2002. For more information
on ASTER contact the Scientific Outreach Program at (218) 529-5225 or e-mail to
ecotox. support@epa.gov.
BASINS (Better Assessment Science Integrating Point and Nonpoint Sources)
http://www.epa.gov/ost/basins/index.html
BASINS is a system developed (1) to facilitate examination of environmental information, (2) to
provide an integrated watershed and modeling framework, and (3) to support analysis of point
and nonpoint source management alternatives. It supports the development of TMDLs, which
require a watershed-based approach that integrates both point and nonpoint sources. BASINS can
support the analysis of a variety of pollutants at multiple scales, using tools that range from
simple to sophisticated. It has been useful in identifying impaired surface waters from point and
B-31
-------
nonpoint pollution, wet weather combined sewer overflows (CSO), storm water management
issues, and drinking water source protection. It has also been used in urban/rural land use
evaluations, animal feeding operations, habitat management practices and as a free resource of
GIS and surface water data for the U.S. for schools and educational institutions.
BASINS can be used in the stressor identification process to georeference stressor data (e.g.,
point source and water quality information) and to quantify the impact of multiple stressors on a
watershed.
BASINS is a software package that can be installed on a personal computer. BASINS allows the
user to assess water quality at selected stream sites or throughout an entire watershed. In
BASINS 3.0, the stream network within the watershed may be determined from Digital Elevation
Modeling or preexisting digitized stream networks (e.g., Reach File version 1 (RF1), RF3, NHD,
or the Census Tiger Files) may be overlain and burned in. Its databases and assessment tools are
directly integrated within an Arc View GIS environment. By using GIS, a user can fully visualize,
explore, and query to bring a watershed to life. The simulation models run in a Windows
environment, using data input files generated in Arc View. Training courses and on-line tutorials
are available to help you learn how to use the BASINS software. The BASINS version 3 system,
documentation, and GIS data of watersheds in each of the ten EPA Regions are available through
National Service Center for Environmental Publications (NSCEP).
ECOTOX
http://www.epa.gov/ecotox/
The ECOTOXicology database is a source for locating single chemical toxicity data for aquatic
life, terrestrial plants and wildlife. ECOTOX integrates three toxicology effects databases:
AQUIRE (aquatic life), PHYTOTOX (terrestrial plants), and TERRETOX created by the U.S.
EPA, Office of Research and Development (ORD), and the National Health and Environmental
Effects Research Laboratory (NHEERL), Mid-Continent Ecology Division, Duluth, MN (MED-
Duluth).
The U.S. EPA Office of Pesticide Program's Environmental Effects Database (EEDB) of toxic
effects data for registered pesticides is also included within ECOTOX via periodic updates from
the Office of Pesticide Programs. The EEDB includes toxicity data for aquatic and terrestrial life.
These data have been reviewed and categorized as acceptable for fulfillment of pesticide
registration and re-registration guideline requirements as explained under FIFRA Subdivision E,
Parts 158.145 and 158.150.
These toxicity data represent concentrations at which specific biological effects have been
observed. The information in the ECOTOX database can help someone using the SI process to
identify effects that are associated with exposure to particular chemical stressors. Users of
ECOTOX should consult the original scientific paper cited in the database to ensure an
understanding of the context of the data retrieved.
The AQUIRE (AQUatic toxicity Information REtrieval) database was established in
1981. Scientific papers published both nationally and internationally on the toxic effects of
B-32
-------
chemicals to aquatic organisms and plants are collected and reviewed for AQUIRE.
Independently compiled laboratory data files that include AQUIRE parameters and
meet the quality assurance criteria are also included. The majority of aquatic literature that has
been reviewed and entered into the system describes studies conducted between 1970 and the
present. Toxicity test results and related testing information for any individual chemical from
laboratory and field aquatic toxicity tests are extracted and added to AQUIRE. Lethal, sublethal
and bioconcentration effects are recorded for freshwater and marine organisms. All AQUIRE
data entries have been subjected to established quality assurance procedures. (See description of
EVISTRA database and its accompanying guidance document.)
The PHYTOTOX database is a computerized information resource that permits the rapid
retrieval and comparison of data pertaining to lethal and sublethal responses, excluding residue
effects, of terrestrial plants to the application of chemicals. Both natural and synthetic organic
compounds administered to native, crop, or weed species have been considered. Independently
compiled laboratory data files that include PHYTOTOX parameters and meet the quality
assurance criteria are also included. The PHYTOTOX database was developed as a PC-based
program through a joint effort by the University of Oklahoma and the U.S. EPA, Western
Ecology Division (formerly Environmental Research Laboratory-Corvallis). PHYTOTOX is
currently maintained at MED-Duluth.
TERRETOX is a terrestrial wildlife toxicity database established to provide data linking
quantified chemical exposures with observed toxic effects. TERRETOX includes results for
lethal, sublethal and bioaccumulation effects. TERRETOX identifies sources of alternative data
(domestic or laboratory animal toxicity and bioaccumulation information) when there is a paucity
of information on wildlife species. Animals associated with the aquatic environment that do not
breathe using gills (e.g., ducks, whales) are included in the TERRETOX database. Results from
exposures of terrestrial life stages of amphibians are included in the TERRETOX database.
Exposures to the aquatic life stages of amphibians are included in the AQUIRE database
component of ECOTOX. Independently compiled laboratory data files that include TERRETOX
parameters and meet the quality assurance criteria are also included. The TERRETOX database
was developed as a PC-based system by the U.S. EPA, Western Ecology Division and is
currently maintained by MED-Duluth.
AQUIRE, PHYTOTOX and TERRETOX may be accessed through the ECOTOX Search
Homepage. Access to these databases is open to the public without restriction. Reports can be
retrieved as a browser viewable report or as an ASCII delimited file. For more information on
any of these databases, contact the Scientific Outreach Program at (218) 529-5225 or E-mail:
ecotox.support@epa.gov.
The ECOTOX database has several useful features:
It is accessible via the World Wide Web
It is easy to use and includes a detailed Help section
It is updated quarterly and you can select "view recent update information" at the
ECOTOX Search Homepage. Therefore, you have access to up-to-date information.
B-33
-------
Each study is reviewed to see if it meets data quality requirements before it is entered into
the database.
It states the limitations of the data in the database as well as the limitations of the web
version of the database itself.
Ecotox Thresholds
http://www.epa.gov/superfund/resources/ecotox/index.htm
EPA has developed software that calculates Ecotox Thresholds (ETs) for selected chemicals and
can print out a table of ETs and their sources. Ecotox Thresholds (ETs), are defined as "media-
specific contaminant concentrations above which there is sufficient concern regarding adverse
ecological effects to warrant further site investigation." ETs are designed to provide Superfund
site managers with a tool to efficiently identify contaminants that may pose a threat to ecological
receptors and focus further site activities on those contaminants and the media in which they are
found. ETs are meant to be used for screening purposes only; they are not regulatory criteria, site-
specific cleanup standards, or remediation goals. For those chemicals with the potential to
bioaccumulate to toxic levels in upper trophic wildlife (e.g., methyl mercury, PCBs,
DDT, dioxins, and lead), these benchmarks may not be low enough at some sites.
The ET software calculates site-specific ETs by adjusting for pH and hardness in surface water
and total organic carbon in sediment. The software can also compare the site-specific ETs to the
concentrations detected at the site. The January 1996 ECO Update on ETs (EPA 540/F-95/038)
describes, in detail, how to calculate and use ETs. This issue of ECO Update is available as a
Portable Document Format (PDF) file that may be downloaded from the Ecotox Thresholds
website. Someone using the SI process can use these Ecotox Thresholds to develop a list of
chemical stressors of concern in the water body he or she is evaluating.
Envirofacts
http://www.epa.gov/enviro/indexjava.html
The Envirofacts Warehouse allows you to retrieve environmental information from EPA
databases on Air, Chemicals, Facility Information, Grants/Funding, Hazardous Waste, Risk
Management Plans, Superfund, Toxic Releases, and Water Permits, Drinking Water, Drinking
Water Contaminant Occurrence, and Drinking Water Microbial and Disinfection Byproduct
Information [Information Collection Rule (ICR)]. You may retrieve information from several
databases at once, or from one database at a time. Online queries allow you to retrieve data from
these sources and create reports, or you may generate maps of environmental information by
selecting from several Mapping Applications. The locational information in Envirofacts contains
all of the latitude and
longitude coordinate data for EPA related facilities.
The Envirofacts database is helpful in creating a list of candidate causes of impairment at your
site because it allows you to identify nearby pollution sources that may influence the biological
conditions at your site.
B-34
-------
The Envirofacts database is only available for online querying via the Internet. To access selected
data tables from Envirofacts, you can use the online Query forms. To access to all the available
data, you can connect to the database directly over the Internet by following the instructions on
the "Accessing the Envirofacts Database" webpage. Access to the Envirofacts from the Internet is
restricted to registered users. You can complete the Envirofacts Registration Form and e-mail it
to enviromail@epa.gov to obtain an EPA UserlD and password. Registration is free. Envirofacts
is an ORACLE database and can be accessed by any software package that can connect to an
ORACLE database.
Environmental Information Management System (EIMS)
http://www.epa.gov/eims/eims.html
EPA's Office of Research and Development (ORD) has developed EIMS to organize descriptive
information (metadata) for data sets, databases, documents, models, projects, and spatial data.
The EIMS design also provides a repository (giant card catalog) for scientific documentation that
can be easily accessed with standard Web browsers. EIMS is revised periodically and
information about the latest updates is available at the website. Users can search within EIMS to
find information sources of interest based upon topic or defined criteria related to types of
environmental resources, geographical extent, date, or content origin.
Each metadata entry in EIMS describes the actual data object, and it includes a contact name, e-
mail address, and phone number for someone who is knowledgeable about the information
resource. Contacts can provide details on how to get an information resource, and may even be
able to provide it themselves.
EIMS establishes intra-agency partnerships with projects and Program and Regional
offices to build the content of its database. These partnerships promote data and information
sharing across EPA. The following partners are storing their metadata in the growing EIMS
collection:
ORD Organization Partners
National Center for Environmental Assessment (NCEA)
National Center for Environmental Research (NCER)
National Exposure Research Laboratory (NERL)
National Health and Environmental Effects Research Laboratory (NHEERL)
National Risk Management Research Laboratory (NRMRL)
Office of Resources Management and Administration (ORMA)
ORD Project Partners
Environmental Monitoring and Assessment Program (EMAP)
Human Exposure Database System (HEDS)
Regional Environmental Vulnerability Assessment (ReVA)
Science Information Management Coordination Board (SIMCorB)
EPA Program Offices
Office of Environmental Information
Office of Water Surf Your Watershed Program
B-35
-------
EPA Regional Office Partners
EPA Region 2
EPA Region 5
EPA Region 8
EPA Region 9
EPA Region 10
Intra-Agency Partnerships
Global Change Research Program
Great Lakes National Program Office (GLNPO)
Inter-Agency Partnerships
Enviro Science e-Print
EIMS also became EPA's node of the National Spatial Data Infrastructure (NSDI)
(http://www.fgdc.gov/) as of August 24, 2000. EIMS is a registered Geospatial
Clearinghouse Node of NSDI, and as such allows the national network of NSDI Gateway servers
to search EPA-EIMS for geospatial metadata at EPA. The EIMS database supports the full
Geospatial Metadata Content Standard developed by the Federal
Geographic Data Committee (FGDC). NSDI is an excellent opportunity for scientists to publish,
share, and use digital geographic information over the Internet.
This database may be used in several steps in the Stressor Identification Process. It can be used to
identify common stressors found in estuaries systems by accessing metadata from EPA's
Environmental Monitoring and Assessment Program (EMAP). EIMS may also help you analyze
the evidence you've collected for each candidate cause to determine which one is likely causing
the observed impairment at your site. For example, EIMS can be used to locate other studies that
found certain associations between candidate causes and effects or suggest ways that you can
analyze your site-specific data to establish associations. Different EPA regions, programs or
offices may also post results from studies of the effect of deliberately eliminating or reducing a
candidate cause (e.g., a field experiment, lab experiment or regulatory or remedial action) to
establish cause and effect relationships. Finally, EIMS can be used to analyze the strength of
evidence for each candidate cause at your site, for example, in terms of whether the hypothesized
cause and effect relationship at your site is similar to any well-established cases.
The EIMS application was developed to operate with the web browser Netscape. Netscape and
ORACLE, the database platform used for EIMS, follow the open standard for use of Java. The
proprietary Java standard employed by Microsoft Internet Explorer is not compatible for all
operations of EIMS.
If you want to add new metadata to EIMS, you can register with EIMS by providing your name,
phone number, and email address. A simple on-line form steps you through the registration
process and on-line forms guide you through the data input process.
B-36
-------
Environmental Monitoring and Assessment Program (EMAP)
http://www.epa.gov/emap/
"The Environmental Monitoring and Assessment Program (EMAP) is a research program to
develop the tools necessary to monitor and assess the status and trends of national ecological
resources (see EMAP Research Strategy). EMAP's goal is to develop the scientific understanding
for translating environmental monitoring data from multiple spatial and temporal scales into
assessments of ecological condition and forecasts of the future risks to the sustainability of our
natural resources. EMAP's research supports the National Environmental Monitoring Initiative of
the Committee on Environment and Natural Resources (CENR). " [A database of Federal
Environmental Research and Monitoring Programs in the Mid-Atlantic Region can be queried at
the national Environmental Monitoring Initiative website at:
http://www.epa.gov/monitor/owsquery/]
One of the goals of EMAP is to develop and demonstrate indicators to monitor the condition of
ecological resources. EMAP funds academic research into ecological indicators and other topics
through STAR grants (see http://es.epa.gov/ncerqa/grants/). Therefore, the information on this
website may be used identify common stressors found in aquatic systems that are affecting the
condition of ecological resources (Step?/Chapter 2 of the Stressor ID process). It may also help
you characterize causes of an observed impairment at your site (Step?/Chapter 4 of the Stressor
ID process) by considering the results from EMAP projects.
You can access EMAP metadata from the EIMS website described above, but the EMAP website
contains additional information such as program information, data and metadata files, and
publications. Data are formatted in two EPA standards: ASCII text files and ARC/INFO export
files. Many files available for download from the EMAP website are compressed or ZIP'ed to
save disk space. The EMAP website was created for best viewing with Netscape 4+ and
Microsoft Internet Explorer 4+ web browsers; however, it should still be reasonably functional
and readable with other browsers and earlier versions of Netscape and Internet Explorer.
EVISTRA (Evaluation and Interpretation of Suitable Test Results in AQUIRE)
http://www.epa.gov/med/databases/evistra.html
The purpose of this database is to present results that were obtained from aquatic toxicity tests on
selected chemicals and were evaluated for suitability and quality and interpreted when necessary.
"Supplementary information (e.g., results of quality evaluations, taxonomic and geographic
information concerning test species, etc.) and software can be accessed to facilitate use of test
results in EVISTRA." Also as part of the EVISTRA effort, a draft guidance document titled
"Guidance for Evaluating Results of Aquatic Toxicity Tests" has been written at MED-Duluth.
The database and guidance document can help the user identify effects associated with particular
stressors as well as learn how to evaluate the quality of aquatic toxicity test data obtained from
other sources (e.g., journal articles, technical reports).
B-37
-------
FIELDS (Fully Integrated Environmental Location Decision Support) system
http://www.epa.gov/region5fields/text/pages/index.html
The FIELDS system is a cooperative project between the USEPA Water Division and Superfund
Division. The Fully Integrated Environmental Location Decision Support (FIELDS) Team's
mission is to identify, assess, communicate and help solve priority environmental problems in
specific geographic areas. To achieve their mission, the FIELDS team has developed a process to
assist environmental organizations who are seeking better and more efficient ways to investigate,
communicate and help solve environmental problems.
Software has also been developed by the FIELDS team to aid site investigators in the following
areas:
Base Mapping Help-Obtaining a good base map is an important first step to be able to
display spatial data and provide a geographical reference for the area under investigation.
This is also listed as a key first step in the Stressor Identification Guidance Document.
Sampling Design Module-This part of the software allows the user to create sampling plans
in GIS and then export the latitude and longitude information to a GPS unit in order to
conduct the field sampling for the project.
Data Entry and Storage-FIELDS uses two relational databases for environmental data. One is
a database developed by the FIELDS team for sediment and fish. The other is the EquIS
database developed by Earthsoft which is used for groundwater well, surface water and soil
data.
Modeling-"The FIELDS modeling tools allow discrete sampling data points to be
interpolated into a surface. Important uses of these interpolated surfaces include delineating
hot spots, calculating average concentrations, estimating contamination mass and volumes,
and developing post-remediation scenarios."
Analysis-The Analysis Tools FIELDS extension currently consists of three tools: Average of
Grid, Estimation Error, and Mass and Volume Calculation. These tools are used in
conjunction with the modeling results to allow decision-makers to isolate any particular
contaminant, level of contamination, or any other characteristic of the site they may wish to
analyze. The tools for quantifying these analyses include 2D and 3D imaging as well as
traditional text-based data reports.
The FIELDS system could be particularly helpful in visualizing available stressor data at very
large sites or at sites with multiple stressors. The FIELDS system has been implemented at more
than 20 sites in the U.S. to-date. A Beta version of the FIELDS system software is now available
to Federal, State and Tribal agencies.
B-38
-------
National Water Quality Standards Database (WQSDB)
http://www.epa.gov/wqsdatabase/
This website provides access to several WQS reports that provide information about designated
uses, waterbody names, State numeric water quality criteria, and EPA recommended numeric
water quality criteria. The WQSDB allows users the ability to compare WQS information across
the Nation using standard reports. The website contains EPA's compilation of the State, Territory
and authorized Tribal water quality standards effective as of May 30, 2000. The site is updated
periodically to reflect EPA approved revisions and additions to State and Tribal water quality
standards.
Currently, the database includes designated use information for 16 states and numeric water
quality criteria for one pilot state (Missouri). Work is progressing to include this information for
the remaining States, Tribes, and Territories in the U.S.
This database can be used when analyzing the available evidence to establish cause and effect
relationships in the SI process (Step?/Chapter 3). "Water quality criteria developed under section
304(a) are based on data and scientific judgments on the relationship between pollutant
concentrations and environmental and human health effects"
http://www.epa.gov/waterscience/standards/critsum.html).
River Corridor and Wetland Restoration
http://www.epa.gov/owow/wetlands/restore/
There are two directories/databases of restoration projects in the United States that are accessible
from this website:
Inventory of Ecological Restoration Projects Within the Mid-Atlantic Integrated
Assessment (MAIA) Region
http://yosemitel.epa.gov/water/restorat.nsf/PYPMapPg670penPage
River Corridor and Wetland Restoration Project Directory
http://yosemitel.epa.gov/water/restorat.nsf/rpd-2a.htm70penPage
The USEPA's Office of Research and Development, working jointly with the Office of Water,
has developed an Internet-accessible inventory of ecological restoration projects within the Mid-
Atlantic Integrated Assessment (MAIA) Region. The inventory encompasses individual
restoration projects conducted in both aquatic and terrestrial environments by federal, state, and
local government entities and non-government organizations. Each project listing includes such
information as a description of the restoration site, the problem being addressed, goals of the
project, critical ecosystem/environmental factors, technical approaches, costs, monitoring
activities, and a contact for more detailed information. The purpose of the inventory is to provide
a central repository of restoration project information specific to the MAIA Region. It is intended
to be useful to researchers, restoration practitioners, environmental resource managers, policy
makers, consultants, and communities within the Mid-Atlantic area. All project information is
provided by the organizations conducting the restoration activities.
B-39
-------
This inventory has been integrated with the Office of Water's River Corridor and Wetland
Restoration database and website. The River Corridor and Wetland Restoration Project Directory
lists all ecological restoration projects by State (including projects in the MAIA Region) and is
also accessible via the Internet. The following types of information are provided for each project:
Location (City/Town and country)
Abstract
Date Provided
Is it a MAIA project?
Project Identification
Lead Restoration Contact
Project Status
Project Characterization
Principal Funding Organizations
Type of Funding Organization
Main Monitoring Focus
Invasive Species
Someone using the stressor identification process could use these databases to identify some
typical wetland stressors and how these stressors were mitigated to restore wetland function.
STORET (STOrage and RETrieval)
http://www.epa.gov/storet/
Project Number
Adjacent Land Use
Greater Surrounding Area Land Use
Descriptive Characteristics of Project Area
Restoration Goals
Restoration Technique/Approach
Project Outlook
Restoration Site Description
Restoration Projects/Results, Reports and Articles
Lessons Learned and Assistance Desired
General Comments on Lessons Learned
STORET (data STOrage and RETrieval system) is a repository for water quality, biological, and
physical data. This database uses reach numbers to help users query, analyze, and map water
quality related issues. The U.S. Environmental Protection Agency (EPA) maintains two data
management systems containing water quality information for the nation's waters: the Legacy
Data Center, and STORET.
The Legacy Data Center, or LDC, contains historical water quality data dating back to the early
part of the 20th century and collected up to the end of 1998. STORET contains data collected
beginning in 1999, along with older data that has been properly documented and migrated from
the LDC.
Both systems contain site-specific raw biological, chemical, and physical data on surface and
ground water collected by federal, state and local agencies, Indian Tribes, volunteer groups,
academics, and others. Data are submitted by many different organizations and data posting is
voluntary. All 50 States, territories, and jurisdictions of the U.S., along with portions of Canada
and Mexico, are represented in these systems. The data in LDC and STORET is not reviewed by
EPA, so the agency makes no warranty regarding the accuracy of the environmental data stored
in those databases. There are no data quality requirements for submittal, however all data kept on
the LDS/STORET web site is fully documented with respect to field and laboratory methods
employed, monitoring technologies used, project goals and plans, and QA/QC practices of the
submitting organizations. Questions concerning the specific data of any organization are best
directed to the organization from whom the data were obtained.
B-40
-------
Each sampling result in the LDC and in STORET is accompanied by information on where the
sample was taken (latitude, longitude, state, county, Hydrologic Unit Code and a brief site
identification), when the sample was gathered, the medium sampled (e.g., water, sediment, fish
tissue), and the name of the organization that sponsored the monitoring. In addition, STORET
contains information on why the data were gathered; sampling and analytical methods used; the
laboratory used to analyze the samples; the quality control checks used when sampling, handling
the samples, and analyzing the data; and the personnel responsible for the data.
These databases can be used in the Stressor Identification Process to help identify any existing
stressors in the waterbody of concern. Users can, for instance, input a reach number and "ask"
STORET to return all point source dischargers, water quality monitoring stations, drinking water
intakes, etc. upstream of that Reach. This provides users with a powerful tool for identifying
what environmental data is available and also the location of that data relative to their Reach of
concern.
LDC or STORET data in the EPA Data Warehouse are located on a UNIX/Oracle database
server that may be browsed or downloaded by anyone using a web browser such as Netscape® or
Microsoft IE® (Internet Explorer) by clicking on "Obtaining Water Quality Data" on the main
STORET homepage. No user registration is required for this type of access, and with the
exception of data marked by its owner as "preliminary", all data in STORET will be freely
available to the public. You can query the LDC or STORET databases on the Internet and
generate data reports for download in html, pdf or txt format. The data from a STORET query
can also be downloaded in "Tilde Delimited" flat ASCII text format.
You may also order a free copy of STORET software to create your own data base of water
quality data. (Call 1-800-424-9067 or send an email to
STORET@epa.gov for your free copy of STORET software). The STORET software along with
ORACLE database engine (which must be obtained separately) will allow you to create and
update your own STORET database and, if you choose, export your data to the EPA Data
Warehouse, where it will be made available to the general public.
Toxicity/Residue Database
http://www.epa.gov/med/databases/tox_residue.html
This database contains concentrations of inorganic and organic chemicals in tissue of aquatic
organisms that are associated with effects to the organism. This database can be used to
investigate hypotheses related to effect/residue relationships (e.g., body burdens of certain
chemicals shown to be associated with particular effects). "The database contains more than
3,000 effect and no-effect endpoints for survival, growth and reproductive parameters for
invertebrates, fish and aquatic life-stage of amphibians. Data were abstracted from approximately
500 literature references on approximately 200 chemicals and 190 freshwater and marine test
species. Survival endpoints account for about 74% of the total amount of data, with growth and
reproduction accounting for 19 and 7% respectively."
B-41
-------
U.S. EPA Mid-Continent Ecology Division
The U.S. EPA's Mid-Continent Ecology Division, part of the Office of Research and
Development and located in Duluth, MN (MED-Duluth), has developed the following databases:
AQUIRE, ASTER, ECOTOX, EVISTRA, Fathead Minnow, PHYTOTOX, TERRETOX,
Toxicity/Residue. All of these databases provide information to link exposure concentrations
with biological effects. Each database is described separately in this memo.
WATERS (Watershed Assessment, Tracking & Environmental Results)
http://www.epa.gov/waters/
WATERS links several state agency and EPA databases in one place to provide information on:
The uses that waterbodies have been designated for, such as drinking water
supply, recreation, or fish protection. These designated uses are part of a state's water
quality standards, provide a regulatory goal for the waterbody and define the level of
protection assigned to it. This information is available in the National Water Quality
Standards Database (see description above).
Those waterbodies listed by the state as impaired under Section 303(d) of the Clean
Water Act, and for those waterbodies, the status of control actions known as TMDLs
(calculations of the maximum amount of a pollutant that a waterbody can receive and still
meet water quality standards, and an allocation of that amount to the pollutant's sources).
This information is also available in the TMDL Tracking System.
The next release of WATERS (Fall 2001) will include:
Drinking water intakes
Recreational beaches
• No discharge zones
Water quality monitoring stations (STORET LDC)
Water quality assessments from Section 305(b) of the Clean Water Act
Other water information that may eventually be included in WATERS include:
319 Non-point source proj ects (GRTS)
Facility outfall locations
Source water protection boundaries
Fish Consumption Advisories
B-42
-------
Water related National Priority List (NPL) locations
Total Maximum Daily Load delineation boundaries
Combined sewer overflows
State and federal water quality managers, as well as interested citizens, can use WATERS to
quickly identify the status of individual waterbodies of interest to them. It can also be used to
generate summary reports on all waters of a state. WATERS uses EnviroMapper, the EPA
standard mapping application, to display this information. EnviroMapper for Water is a web-
based Geographic Information System (GIS) application that dynamically displays information
about bodies of water in the U.S. This interactive tool allows you to create customized maps
portraying the nation's surface waters along with a collection of environmental data using your
web browser. The application can be used to view environmental information from the national
level down to community level (within one mile), and provides the ability to pan, zoom, label
and print maps. You also can link to text reports after identifying a specific waterbody of
interest. A remaining question is whether the user can access any of the raw data displayed by
EnviroMapper (e.g., georeferencing data or water quality data) directly from WATERS database.
The National Hydrography Dataset, or NHD, is the nationally consistent waterbody network that
serves as the foundation and common "language" for WATERS. In the 305(b) database that was
created before the broad use of GIS, waterbodies were often defined as small watersheds. This
ended up being problematic because water quality assessment data stored in the database often
only applied to portions of these watershed waterbodies. An advantage of the WATERS database
is that it uses the National Hydrography Dataset Reach Indexing Tool (NHD-RIT), which is an
interactive GIS application that allows users to georeference surface water data using the
National Hydrography Dataset (NHD). This allows all water quality data to be georeferenced
based on a consistent standard (the NHD). The NHD-RIT can be used by any organization that
has a mandate to monitor and track surface water information. Users may use the NHD-RIT with
Arc View to visualize the data or use WATERS as a web-based alternative for displaying the data
in the NHD.
Currently there are three databases that are used in conjunction with the Reach Indexing Tool
(RIT) and the National Hydrography Dataset (NHD):
TMDL Tracking System: this database is used by EPA to maintain information about the
States TMDL lists.
Contact: Chris Laabs
Download: http://www.epa.gov/owow/tmdl/trcksys.html
Water Quality Standards Database (WQSDB): contains designated use information as it was
assigned by the States.
Contact: Bill Kramer
WQS Reach Indexing Website: WQS Review Site
B-43
-------
305(b) Assessment Database (ADB): is a database used by the States to maintain their surface
water under the 305(b) Clean Water Act.
Contact: Dwane Young
Download: http://water305b.rti.org
Currently only the TMDL Tracking System and the Water Quality Standards Database are linked
to WATERS, but the Assessment Database (ADB) will be integrated into the next release of
WATERS in the Fall of 2001. All states assess their individual waterbodies for degree of
designated use support (e.g., "fully supporting aquatic life; not supporting primary contact
recreation"). If a waterbody's uses are impaired, the stressors and sources of impairment are also
determined (e.g., "causes/stressors are nutrients and sediment; sources are urban runoff and row
crop agriculture"). The ADB is designed to make tracking and reporting this data accurate,
straightforward and user- friendly for participating states, territories, tribes and basin
commissions. It also enables users to meet the requirements of Section 305(b) of the Clean Water
Act. All three databases can help develop a list of candidate causes in the SI Process. The
National Water Quality Standards Database can also be used when analyzing available evidence
to determine the cause of observed impairment using the SI process.
Summary:
Table B-2 summarizes how these information sources can be used in each step of the stressor
identification process. The information sources we have reviewed are most useful in developing
a list of stressors, visualizing available data, obtaining toxicological effects data for chemical
stressors, and locating case studies to which the user can compare his or her site. Additional
resources are needed for identification of physical stressors (e.g., hydrological changes) and the
effect of nutrients on aquatic systems. It would also be useful to have a software program that
may be used in conjunction with the Stressor Identification Guidance to construct a site
conceptual model. We have provided a list of additional, non-EPA websites that may prove
useful in the stressor identification process (Table B-3). However, it was beyond the scope of this
task to evaluate each of them as thoroughly as we have for the EPA information sources in this
memo.
B-44
-------
Table B-2. Uses of EPA Information Sources in the Stressor Identification Process
Information sources
List
candidate
causes
Analyze the
evidence
Characterize
Causes
AQUATOX
X
ASTER
X
BASINS
X
ECOTOX
X
Ecotox Thresholds
X
Envirofacts
X
EIMS
X
X
X
EMAP
X
X
EVISTRA
X
FIELDS
X
Inventory of Ecological
Restoration Projects Within the
Mid-Atlantic Integrated
Assessment (MAIA) Region
X
National Water Quality
Standards Database (WQSDB)
X
River Corridor and Wetland
Restoration Project Directory
X
STORET
X
Toxicity/Residue Database
X
WATERS
X
X
B-45
-------
Table B-3. Internet Sources of Stressor Identification Information
URL
Subject
http://www.nal.usda.gov/wqic/
Agriculture Dept Water Quality Info Center
http://www.wes.army.mil/
Army Corps of Engineers' Waterways
Experimental Station (WES)
http://www.wes.army.mil/el/arams/arams.html
Army Risk Assessment Modeling System
http://atsdrl.atsdr.cdc.gov:8080/atsdrhome.html
ASTDR Home Page
http://www.atsdr.cdc.gov/toxfaqf.html
ASTDR ToxFAQs
http://www.atsdr.cdc.gov/mrls.html
ATSDR's Minimal Risk Levels (MRLs) for
Hazardous Substances
http://www.epa.gov/ost/basins/index.html
BASINS (Better Assessment Science Integrating
Point and Nonpoint Sources)
http://www.chemfinder.com/
Chemfinder
http://info.cas.org/
Chemical Abstract Service
http://www.wes.army.mil/el/t2dbase.html
Environmental Residue-Effects Database (ERED),
Biota Sediment Accumulation Factor (BSAF)
Database, Ocean Disposal Database
http://www.chesapeakebay.net/bayprogram/
EPA Chesapeake Bay Program
http://www.epa.gov/superfund/resources/ecotox/index.htm
EPA Ecotox Thresholds
http://www.epa.gov/emap/
EPA Environmental Monitoring and Assessment
http://www.epa.gov/wqsdatabase/
EPA National Water Quality Standards Database
(WQSDB)
http ://www. epa.gov/ow/soft. html
EPA Office of Water Databases and Software
http://www.epa.gov/region5fields/text/pages/index.html
EPA Region 5 - FIELDS (Fully Integrated
Environmental Location Decision Support) system
http://www.epa.gov/med/databases/tox residue.html
EPA Toxicity /Residue Database
http://www.epa.gov/ost/models/aquatox/
EPA's Aquatox Release 1 - A Simulation Model
for Aquatic Ecosystems
http://www.epa.gov/med/databases/aster.html
EPA's ASTER Database for Ecotoxicological
Profiles
http://www.epa.gov/medecotx/
EPA's ECOTOX Database
http://www.epa.gov/enviro/indexJava.html
EPA's Envirofacts Data Warehouse and
Applications
http://www.epa.gov/eims/eims.html
EPA's Environmental Information Management
System (EIMS)
http://www.epa.gov/med/databases/evistra.html
EPA's EVISTRA (Evaluation and Interpretation
of Suitable Test Results in AQUIRE) Database
http://yosemitel.epa.gov/water/restorat.nsf/PYPMapPg670penPage
EPA's Inventory of Ecological Restoration
Projects Within the Mid-Atlantic Integrated
Assessment (MAIA) Region
http://yosemitel.epa.gov/water/restorat.nsf/rpd-2a.htm70penPage
EPA's River Corridor and Wetland Restoration
Project Directory
http://www.tera.org/
Toxicity Excellence for Risk Assessment
Table 3 continued.
URL
Subject
http://es.epa.gov/ncerqa/grants/
EPA's Science to Achieve Results (STAR)
Program
B-46
-------
Table B-3 cont.
URL
Subject
http://www.epa.gov/storet/
EPA's STORET (STOrage and RETrieval)
Database
http://www.epa.gov/waters/
EPA's WATERS (Watershed Assessment,
Tracking & Environmental Results)
http://www.epa.gov/watertrain/
EPA's Watershed Acadmemy Web - Distance
Learning Modules on Key Watershed
Management Topics
http://irptc.unep.ch/irptc/invent/igo.html
Inventory of Information Sources on Chemicals
http ://www 1. nature. nps. gov/datasci/
National Park Service - Data on Park Resources
http://wwwl.nature.nps.gov/toxic/
National Park Service Environmental
Contaminants Encyclopedia
http://response.restoration.noaa.gov/cpr/sediment/squirt/squirt.html
NOAA Screening Quick Reference Tables for
inorganic/organic contaminants in environmental
media
http://www.opsd.nos.noaa.gov/tp4days.html
NOAA Tide Predictions
http://www.esd.ornl.gov/programs/ecorisk/ecorisk.html
Oak Ridge National Laboratory (ORNL)
Ecological Risk Analysis: Tools and Applications
http ://risk.lsd. ornl. gov/homepage/eco tool, shtml
ORNL- Risk Assessment Information System-
Ecological Benchmarks
http://www.riskworld.com/NEWS/99ql/nw9aa081 .htm
Risk World Index of Endocrine Disruptor
Resources
http://www.toxnet.nlm.nih.gov/
TOXNET
http://www.usgs.gov/
USGS
http://www.best.usgs.gov/
USGS Biomonitoring of Environmental Status and
Trends (BEST) Program Databases
http://www.nbii.gov/
USGS National Biological Information
Infrastructure
http://www.nwrc.gov/index.html
USGS National Wetlands Research Center
http://www.nwrc.gov/publications/specindex.html
USGS National Wetlands Research Center -
Species Profiles
http://www.pwrc.usgs.gov/
USGS-Patuxent Wildlife Research Center
http://fwie.fw.vt.edu/WWW/macsis/index.htm
VA Tech-Marine and coastal species information
system-fish and wildlife information exchange
http://www.vims.edu/env/research/risk/software/other software.htm
VIMS Environmental Toxicology and Risk
Assessment - Software and Documentation
http://www.ecy.wa.gov/services/as/iip/eim/index.html
Washington State Department of Ecology's
Environmental Information Management (EIM)
System
http://www.terrene.org/wirsdata.htm
Watershed Information Resource System (WIRS)
Database (in partnership with EPA's Clean Lakes
Program)
Key:
Shading indicates that we have described this resource in our memo.
B-47
-------
DRAFT
Additional U.S. EPA Information Sources:
Mid-Atlantic Focus
Prepared by
MENZIE • CURA & ASSOCIATES, INC.
Environmental Consultants
One Courthouse Lane, Suite Two • Chelmsford, Massachusetts 01824
Phone (978) 453-4300 • Fax (978) 453-7260
Prepared for T N & Associates, Inc.
under Contract Number 68-C-98-187
This document has been reviewed and approved for use as background for
the CADDIS Workshop by the EPA Work Assignment Manager.
B-48
-------
DRAFT
This memo summarizes some of the key features of several information sources for the mid-
Atlantic region that may be useful to users of the Stressor Identification Guidance Document
(USEPA, 2001). For each source we provide:
The information contained in each source that might be useful in the Stressor
Identification (SI) Process
A website or other means to access the information
The type of information and forat in which it is available
• Whether the information is georeferenced.
These characteristics are summarized in Table B-4 and are described in more detail in the text
below. The information sources we reviewed are listed in alphabetical order, not by their
usefulness or application to particular steps in the SI Process.
Alliance Citizen Monitoring Database
http://www.acb-online.org/monitoring/site.cfm
The Alliance Citizen Monitoring program is a regional network of more than 145 trained
volunteers who perform weekly water quality tests that help track the condition of waters flowing
toward the Chesapeake Bay. The data collected by these volunteers has been compiled in a
database containing monitoring site information, monitoring event data and water quality data
collected from 8/17/1985, to the present for just under 200 waterbodies, across the Chesapeake
region in Pennsylvania, Maryland and Virginia. The database also contains jellyfish data from
1/14/2001, to the present. The database is updated continuously and includes data for
waterbodies ranging from small creeks to large rivers. This data can be queried on the website
via a series of prompts. The user can then download data as text, tab-delimited files or view
graphical representations of the data for a particular monitoring location.
The information in this database can be used to find common sources and stressors in the mid-
Atlantic region as part of the "List Candidate Causes" step in the Stressor Identification Process.
More information about the Alliance Citizen Monitoring program is available at http://www.acb-
online.org/citmon.htm.
B-49
-------
DRAFT
Table B-4. Summary of Information Sources for the Mid-Atlantic Region For Use in the Stressor Identification Process
Information source
Use in SI Process
Web Address
Data Type and Format
Georeferenced?
Alliance Citizen Monitoring
Database
List Candidate Causes
http://www.acb-
online. org/monitoring/site. cfm
Can select data to download in
text, tab-delimited format or graph
data from the webpage.
Yes
Chesapeake Information
Management System
(CIMS)
List Candidate Causes
http ://www. chesapeakebay .net/cims/index.
htm
Can view metadata and watershed
profiles on-line.
Yes
Delaware Department of
Natural Resources and
Environmental Control
Environmental Navigator
List Candidate Causes
htto://www.dnrec.state.de.us/enweb/
Access to base map of DE and
web-based interactive mapping
application linked to a database of
information on contaminant source
sites and ambient environmental
monitoring sites.
Yes
Delaware Stream Watch
List Candidate Causes
http://www.delawarenaturesociety.org/
swtechmon.htm
Excel spreadsheets & hard copies
available upon request.
Yes
Delaware Surface Water
Quality Management
List Candidate Causes &
Analyze Evidence
lUtD://\vw\v.dnrcc. state.dc.us/dnrcc2000/Li
brarv/Water/305brer>ort.htm and
http://www.dnrec.state.de.us/dnrec2000/Di
visions/Water/WaterQuality/Standards.htm
Can download State of Delaware's
305(b) reports and Surface Water
Quality Standards as PDF files.
No
Environmental Monitoring
and Assessment Program
(EMAP) Data
Define the Impairment &
List Candidate Causes
http://www.epa.gov/emap/html/datal/surfw
atr/data/
Can download data (ASCII) and
metadata (PDF) from the web
Yes
Maryland Biological Stream
Survey
Define the Impairment &
List Candidate Causes
lUtD://mddnr. chcsaDcakcbav.net/mbss/
search, cfm
Can view data on the website or
download 1995-1997 data as text,
tab-delimited files
Yes
Maryland Department of
Natural Resources -Streams
List Candidate Causes &
Analyze Evidence
http://www.dnr.state.md.us/streams/pubs/p
ub_list.html
Some technical reports are
available for download in PDF
format.
No
Maryland Water Monitoring
Council
Define the Impairment
httr>://www.mes.md.eov/mwmc/
Can download MS Access
database of water monitoring
projects.
Yes (some)
Mid-Atlantic Highlands
Coordinating Council;
Potomac River Basin Pilot
Proiect
Define the Impairment &
List Candidate Causes
lUtD://acl. cr. usas. aov/aroiiDs/ais/Doto mac/
mahaw.htm
Can view metadata online & save
as HTM or TXT file
No
-------
DRAFT
Table B-4. Summary of Information Sources for the Mid-Atlantic Region For Use in the Stressor Identification Process
Information source
Use in SI Process
Web Address
Data Type and Format
Georeferenced?
Mid-Atlantic Integrated
Assessment (MAIA) Project
List Candidate Causes
htto://md.uses.eov//maia/ and
htto://www.er>a. eov/maia/assets/odf/
MAHAStreams.pdf
Can download data in ASCII, PDF
and Arclnfo format from the web.
Yes
National Environmental
Monitoring Initiative - Mid-
Atlantic Inventory
Define the Impairment &
List Candidate Causes
http://www.epa.gov/cludygxb/site-
mid.html
List of available monitoring data
and contacts for monitoring
programs
No
PA Department of
Environmental Protection -
Water Quality Assessment
and Standards
List Candidate Causes &
Analyze Evidence
htto ://www. dea state .Da.us/dcD/dcDutatc/
watermet/wcro/wastandards/wastandards.
htm and
lUtD://w\vw.dcD. state, oa. us/dcD/siibicct/cci
b/2000/June20/fch93Annex.r)df
Can download State of
Pennsylvania's 305(b) reports in
PDF format
No
Pennsylvania Senior
Environmental Corps
(PaSEC)
List Candidate Causes
htto://www.environmentaleducation.ore/
default, lasso
Searchable on-line by "registering".
Can view data tables in your web-
browser
Yes
Pennsylvania Spatial Data
Access (PASDA)
Define the Impairment &
List Candidate Causes
lUtD://www.Dasda.Dsu.cdu/acccss/do\Ynlo
ad. shtml
Free, downloadable data from web
or all data on PASDA can be
obtained on a CD-ROM for a fee
(multiple formats).
Yes (most)
SRBC Nutrient Assessment
Program
List Candidate Causes
lUtD://\vww. srbc.nct/nutricntdata.lUm
htto ://www. srbc. net/trends. htm
lUtD://\vww. srbc.net/nutricntloads. htm
Can view summarized data on web
or download data in comma
delimited text file, PDF or Excel
format
Yes
USGS National Water
Quality Assessment
(NAWQA) Program
Define the Impairment &
List Candidate Causes
htto://orxddwimdn.er.uses.eov/servlet/r>a
se? r>aeeid=543& dad=oortal30& sche
ma=PORTAL30
Can view or download data in
several formats (XLS, WKS, TXT,
CSV, HTM)
Yes
Virginia Department of
Environmental Quality
List Candidate Causes
httos://www.dea.state.va.us/webar>r>/wam
. homeraee
Can query data by county/city or
stream and view it on web
Yes
West Virginia DEP,
Division of Water
Resources, Watershed
Assessment and Strategic
Planning Section
Define the Impairment &
List Candidate Causes
htto://www.der>.state.wv.us/wr/index.cfm
?rase=OWR Website/C and D/WAP.
htm
Can download data summaries in
PDF format
?
-------
DRAFT
Chesapeake Information Management System (CIMS)
http://www.chesapeakebav.net/cims/index.htm
"The Chesapeake Information Management System (CIMS) is an organized, distributed library of
information and software tools designed to increase basin-wide public access to Chesapeake Bay
information. The Internet sites in CIMS are maintained by CIMS Partners who have enacted
Memoranda of Agreement (MOA) to follow certain guidelines for assembling and publishing
Chesapeake Bay related information." Internet applications that can be accessed from the CIMS
website and may be useful in the Stressor Identification Process include:
Watershed Profiles
"Watershed Profiles assembles maps, charts and information that portray the
environmental condition of Chesapeake Bay watersheds. It operates at a variety of scales
from the entire 64,000 square mile Chesapeake Bay watershed to small tributary
watersheds. Information on landscape changes, Bay Program activities, other organization
activities, and places to visit are some of the information displayed in easy to read charts,
maps and tables." This application would be very useful to gather information about
common sources and stressors (landscape, nutrients and toxins) in large and small
watersheds (e.g., watersheds for individual creeks and rivers). This information could
help "List Candidate Causes" in the Stressor Identification Process.
CIMS Search
The website listed above also provides a link to a Search Engine for searching the CIMS
metadata database and COMET (CIMS Online Metadata Entry Tool) which is used to
maintain the database. Metadata is available for water bodies of all sizes and data types
include databases, data sets, GIS coverage, maps, spatial data bases, spatial datasets, and
reach files to name a few. This web page can be used to locate data sources that may be
useful to map your data or list stressors and sources in the "List Candidate Causes" step
of the Stressor Identification Process.
Delaware Department of Natural Resources and Environmental Control (DNREC)
Environmental Navigator
http: //www. dnrec. state. de .us/enweb/
"The Environmental Navigator is a web-based interactive mapping application linked to a
database of information on potential, possible, and known contaminant source sites and on
ambient environmental monitoring sites in the state of Delaware". It is very similar to EPA's
Enviromapper application, but is specific to the state of Delaware. It will eventually include
additional information such as aerial photos, wetlands, land use, biodiversity indices, soil types,
air and water ambient monitoring results, and facility violations. A base map of Delaware is
available from this web page. (Note: This application does not work in Netscape (something to
do with Java.)
B-52
-------
DRAFT
"The Environmental Navigator includes summary information on over 8000 sites of more than
twenty types. Information on each site includes site name, ID number, status, location (XY,
county, basin, watershed), and a hyperlink to more information about that site if it exists.
Potential contaminant source sites also include a Contaminant Source Potential Rating and
monitoring status information by media and contaminant class."
The information in the environmental navigator can be used to find and map common sources
and stressors in Delaware as part of the "List Candidate Causes" step in the Stressor
Identification Process. Maps showing user-selected data can be saved as JPG files.
Delaware Stream Watch
http ://www. delawarenaturesocietv. org/ swtechmon.htm
The Technical Monitoring program of Delaware Stream Watch is a collaboration between the
Delaware Nature Society and the Delaware Division of Natural Resources and Environmental
Control. Volunteers monitor assigned sites on a monthly basis, testing for dissolved oxygen, pH,
alkalinity, nitrate, conductivity, salinity (in tidal reaches) and temperature. Additional procedures
ensure quality control in sampling techniques. The data collected by Stream Watch volunteers is
put into the state's database and copies are maintained (in Excel spreadsheets) at the Delaware
Nature Society. Data collected by Stream Watch in the past 6 years is for the Christina Basin,
which includes the watersheds of the Brandywine Creek, Red Clay Creek, White Clay Creek and
the Christina River. To access this data, individuals can contact Julie Smith, the Technical
Monitoring Coordinator for Stream Watch at 302-239-2334 ext. 52 or send email to
stream@dnsashland.org. This information may be useful in to "List Candidate Causes" in the
Stressor Identification Process.
Delaware Surface Water Quality Management
The Watershed Assessment Section of the Delaware Department of Natural Resources and
Environmental Control (DNREC) has developed a website to provide information about Surface
Water Quality Management in the State. The State of Delaware's 1998 305(b) report can be
downloaded from: http://www.dnrec.state.de.us/dnrec2000/LibrarvAVater/305breport.htm. The
305(b) report identifies water quality problems and sources for watersheds in Delaware, which is
useful in the "List Candidate Causes" step in the Stressor Identification Process.
The most recent copy of the State of Delaware's Surface Water Quality Standards can be
downloaded in a PDF file from the following web page:
http://www.dnrec.state.de.us/dnrec2000/Divisions/Water/WaterOualitv/Standards.htm. These
criteria are based on scientific judgments about stressor levels that, if met, will protect designated
uses of water bodies in the State. Thus, a user can compare the stressor levels (e.g., chemical
concentrations) at his or her site to these criteria in the "Analyze Evidence" step of the Stressor
Identification Process.
B-53
-------
DRAFT
More information about Surface Water Quality Management in Delaware is accessible at:
http://www.dnrec.state.de.us/dnrec2000/DivisionsAV ater/W aterOualitv/W OM.htm
Environmental Monitoring and Assessment Program (EMAP)
http ://www. epa. gov/emap/html/datal/ surfwatr/data/
Surface waters data and metadata for Mid-Atlantic streams can be accessed from the EMAP web
page listed above. Benthic, water chemistry, sediment, fish and periphyton data collected from
thirteen Mid-Atlantic streams between 1993 and 1998 can be downloaded in ASCII format and
metadata can be downloaded in PDF format. This biological and chemical data may be useful to
"Define the Impairment" & "List Candidate Causes", respectively, in the Stressor Identification
Process at your site.
Maryland Biological Stream Survey (MBSS)
http://mddnr.chesapeakebav.net/mbss/search.cfm
The MBSS collects data to look at the physical, chemical and biological status of a stream, the
geography (landscape) surrounding the stream and the relationships between the living resources
(fish and bugs) and their physical surroundings to try to define the problem of acid deposition in
Maryland. The MBSS is a program led by the Monitoring and Nontidal Assessment Division of
the Maryland Department of Natural Resources.
In cooperation with U.S. EPA's Mid-Atlantic Integrated Assessment Program the Maryland
Department of Natural Resources (DNR) recently completed the first-ever, comprehensive
survey of the almost 9,000 miles of small and medium-sized freshwater
streams. Sample sites were randomly selected 75 meter segments of first, second, and third order
streams (Strahler; 1:250,000 scale USGS maps) in each of 18 major river basins in Maryland,
sampled in both spring and summer where possible. Parameters measured included physical
habitat characteristics, water chemistry, fish, and benthic macroinvertebrates. Herps, submerged
aquatic vegetation, bivalves and crustaceans are also measured when present. Data collected
between 1995 and 1997 is available for download as text, tab-delimited files and data sets from
1995-1997 and 2000 may be searched on the MBSS web page by county, stream name, river
basin or watershed.
This biological and chemical data may be useful to "Define the Impairment" & "List Candidate
Causes", respectively, in the Stressor Identification Process at your site. Also, the report
summarizing the survey findings, "From the Mountains to the Sea - The State of Maryland's
Freshwater Streams," may be a useful case study for identifying associations between candidate
causes and effects in the "Analyze Evidence" step of the Process for sites in Maryland. This
report can be downloaded from EPA's Mid-Atlantic Integrated Assessment (MAIA) web site at:
http://www.epa.gov/maia/html/reports.html. Other MBSS publications are available at:
http://www.dnr.state.md.us/streams/mbss/mbss pubs.html.
B-54
-------
DRAFT
Maryland Streams - part of the DNR website
http://www.dnr.state.md.us/streams/pubs/pub list.html
Monitoring and Non-Tidal Assessment (MANTA) does much work in Maryland streams. The
Monitoring and Non-Tidal Assessment Division (MANTA) of the Maryland Department of
Natural Resources (DNR) is organized into three interactive programs - Atmospheric Deposition
(AD), Ecological Assessments (EA) and Monitoring (M). The Division is responsible for
assessment of status and trends of biological communities in the non-tidal portions of tributaries
in Maryland. A bibliography of reports and technical documents written by MANTA, the Stream
Corridor and Management Division of MD DNR and the Maryland Water Monitoring Council
can be found at: http://www.dnr.state.md.us/streams/pubs/pub list.html. Examples of
publications that may be useful in the Stressor Identification Process are:
Relationship between effects on fish stocks with hydrographic conditions and pollution
loadings;
Relating Nutrient Loading to Production of Selected Fish Populations;
Assessment of Chesapeake Bay Benthic Macroinvertebrate Resource Condition in
Relation to Water Quality and Watershed Stressors;
Status and Temporal Trends in Benthic Macroinvertebrate Communities as Indicator of
Water Quality at Maryland's Core Monitoring Stations, 1976-1992.
Basin Fact Sheets (describing the current status of wadeable streams)
Some reports are available on-line in PDF format and others are available upon request from the
MD Department of Natural Resources (see web page for contact information). The Maryland
2000 305(b) Report is available for download in PDF format from:
http://dnrweb.dnr.state.md.us/download/bavs/MD2000 305b.pdf
These or other publications on the Maryland DNR web site may provide useful information in
order to "Identify Candidate Causes" and "Analyze Evidence" (e.g., search for associations
between causes and effects from other sites).
Maryland Water Monitoring Council
http://www.mgs.md.gov/mwmc/data.html
The Maryland Water Monitoring Council is an umbrella organization that includes
representatives from professional and volunteer monitoring organizations. "The MWMC is an
organization created in 1995 to foster cooperation among groups involved in all types of water
monitoring Activities. The Council is interested in physical, chemical, and biological monitoring,
B-55
-------
DRAFT
as well as the evaluation of those land use factors that affect changes in aquatic habitat quality
and quantity."
From this website, you can download a database of non-tidal tributary water quality monitoring
programs in the Chesapeake Bay Watershed compiled by The Chesapeake Information
Management System (CIMS). The database is available for download as a zipped Microsoft
Access file. The database provides information about over 330 monitoring programs conducted
by over 160 organizations throughout the Chesapeake Bay watershed. The database indicates
whether data sets gathered by different monitoring programs are available to the public and, if so,
how to obtain them. Thus, this database can potentially be used to locate monitoring data from
locations in your project area to "Define the Impairment" or "List Candidate Causes" in the
Stressor Identification Process.
Mid-Atlantic Highlands Coordinating Council; Potomac River Basin Pilot Project
http:// ael. er .uses. gov/ groups/gi s/potomac/mahaw.htm
At the request of the Mid-Atlantic Highlands Coordinating Council (MAHCC), a coalition of
federal and state natural resource agencies, personnel at USGS compiled a preliminary inventory
of biological, chemical and physical data for the upper Potomac River drainage basin (in MD,
PA, VA and WV). Detailed information was identified for the Cacapon River, Evitt's Creek, the
North Branch of the Potomac River, Opequon Creek and Sideling Hill Creek. Inventory tables
accessible on this web page list several data sets that were collected for these waterbodies
between 1950 and 1997. The following information is provided for each data set: the data format,
the agency that collected the data and a contact person at that agency. Obtaining the types of data
listed on this web page could be helpful to "Define the Impairment" or "List Candidate Causes"
in the Stressor Identification Process. [Note: Upon contacting USGS, I was told that no further
work has been done on the Potomac River Basin Pilot Project and the MAHCC no longer exists.
Mid-Atlantic Integrated Assessment (MAIA) Project
http://md.usgs.gOv//maia/
http://www.epa.gov/maia/assets/pdf/MAHAStreams.pdf
The Mid-Atlantic Integrated Assessment (MAIA) is a research, monitoring, and assessment
initiative. From the MAIA website (http://www.epa.gov/maia/), there are links to estuaries,
surface waters, landscape ecology, forest and agro-ecosystems data. The U.S. Geological Survey
(USGS) has recently completed analyses of the occurrence of nitrate and pesticides in waters of
the Mid-Atlantic Region as part of the Mid-Atlantic Integrated Assessment (MAIA) project of
the U.S. Environmental Protection Agency (USEPA). The MAIA project encompasses USEPA
Region III and adjacent parts of major river basins in New Jersey, New York, and North
Carolina. Information about this USGS project for MAIA can be found at:
http://de.usgs.gov/publications/ofr-98-158/.
B-56
-------
DRAFT
Data collected from 463 sites between October 1973 and March 1997 were used to characterize
the occurrence of pesticides in surface water of the Mid-Atlantic Region in the report: Pesticides
in Surface Water of the Mid-Atlantic Region, by M.J. Ferrari, S.W. Ator, J.D. Blomquist, and
J.E. Dysart (1998). Compiled nitrate and pesticide data for the Mid-Atlantic Region and
supporting information are contained in four tab-delimited ASCII files and may be obtained
along with the report listed above from http://md.usgs.gov//maia/. This data could be helpful to
"List Candidate Causes" in the Stressor Identification Process.
The report Mid-Atlantic Highlands Streams Assessment can be downloaded from the MAIA
website at http://www.epa.gov/maia/assets/pdf/MAHAStreams.pdf. One of the purposes of the
report is to identify and rank the relative importance of stressors affecting the conditions of
streams in the Mid-Atlantic Highlands region. Therefore, it could be helpful to "List Candidate
Causes" in the Stressor Identification Process as well.
National Environmental Monitoring Initiative Mid-Atlantic Inventory
http: //www, epa. gov/ cludv gxb/site-mi d. html
"This inventory presents 128 federal research and monitoring programs that operate in the mid-
Atlantic region. Programs are organized by sampling design into three tiers recognized by the
National Environmental Monitoring Initiative: spatially continuous monitoring over a large
region 910,000 km2 or more), spatially sub-sampled surveys and monitoring designed to evaluate
the status of a large region by sampling a subset of the total area, intensive research sites that are
selected due to their known ecological condition or suitability for experimental manipulation."
Metadata (e.g., name of monitoring program, contact information, data available) for the Mid-
Atlantic region can be accessed using user-defined queries. Data sets identified from this source
may be useful to "Define the Impairment" or "List Candidate Causes" in the Stressor
Identification Process.
Pennsylvania Department of Environmental Protection - Water Quality Assessment and
Standards
http://www.dep.state.pa.us/dep/deputate/watermgt/wqp/wqstandards/wqstandards.htm
http://www.dep.state.pa.us/dep/subiect/eqb/2000/June20/fch93Annex.pdf
The State's 2000 Section 305(b) - Water Quality Assessment Report and 2001 Section 305(b)
Report - Update as well as older copies of the 305(b) reports are accessible through the water
quality assessment and standards web page
(http://www.dep.state.pa.us/dep/deputate/watermgt/wqp/wqstandards/wqstandards.htmy These
reports discuss causes of observed impairment in the state's waters and thus may be useful in the
"List Candidate Causes" step in the Stressor Identification Process.
The latest water quality amendments to Chapter 93 of DEP's Regulations (Title 25, PA Code)
include modifications to Pennsylvania water quality standards. The amendments can be
B-57
-------
DRAFT
downloaded at:
http://www.dep.state.pa.us/dep/subiect/eqb/2000/June20/fch93Annex.pdf.
These standards can be used in the "Analyze Evidence" step of the Stressor Identification Process
for sites in Pennsylvania.
Pennsylvania Senior Environmental Corps (PaSEC)
http://www.environmentaleducation.org/default.lasso
The PaSEC program began in 1997 and is the result of a partnership between Pennsylvania's
DEP, Department of Aging and the Environmental Alliance for Senior Involvement (EASI). The
EASI Water Monitoring Database has two major components:
¦ One component is for the members of the Pennsylvania Senior Environmental Corps
(PaSEC) and any other group using the standardized protocols. This component has fields
for seven physical and chemical indicators of water quality including pH, conductivity, water
temperature, dissolved oxygen, nitrates, total phosphates and sulfates. A habitat assessment
and a water quality rating based on benthic macroinvertebrate sampling are also part of this
component.
¦ The second component is an "open" component available for use by any volunteer monitors
who are using their own protocols and quality control plans. This part of the database has
fields for the seven physical and chemical indicators listed above as well as for hardness,
turbidity, chlorides, chlorophyll a, total dissolved solids, total suspended solids, aluminum,
manganese, iron, fecal coliform and Secchi depth.
Both of these databases are searchable on-line and users can view the results of their data queries
in table format in their web browser. This monitoring data is another source of information that
could be helpful to "List Candidate Causes" in the Stressor Identification Process for sites in
Pennsylvania.
Pennsylvania Spatial Data Access (PASDA)
http://www.pasda.psu.edu/access/download.shtml
"PASDA is Pennsylvania's official geospatial information clearinghouse and the
Commonwealth's node on the National Spatial Infrastructure (NDSI). The PASDA clearinghouse
provides for the widespread sharing of geospatial data, eliminates the creation of redundant data
sets, and serves as a resource for locating data throughout the Commonwealth through its data
storage, interactive mapping/webgis applications, and metadata/documentation efforts. The data
available on PASDA is the result of efforts by state, local, regional, and federal governments,
non-profit organizations, and academic institutions who wish to make their data or metadata
available through the clearinghouse."
Data can be downloaded for free from the PASDA Data download web page
(http://www.pasda.psu.edu/access/download.shtmn or all data on PASDA can be obtained on
CD-ROM through a cooperative arrangement with the Pennsylvania State Data Center for a
B-58
-------
DRAFT
nominal charge. The following types of data can be downloaded from the PASDA website for the
municipal, county, watershed or state level:
¦ Water quality sampling records for 622 sites throughout Pennsylvania (Alliance for
Aquatic Resource Monitoring (ALLARM) data).
¦ Drainages, streams, road, and population coverages from the Chesapeake Bay
drainage are (compiled by Chesapeake Bay Program)
¦ Maps with locations of natural heritage sites classified according to biological
diversity and ecological integrity (not georeferenced).
¦ Physiographic provinces and glacial extents.
¦ Aerial photographs
¦ Various coverages from the EPA concerning Acid Mine Drainage, forest cover types,
sub-ecoregions, toxic resource inventory, CERCLIS, and reach files.
¦ Historical fish species database from the Pennsylvania Fish and Boat Commission
¦ State-wide and county-wide floodplain coverages.
¦ Urban, county and state-wide watersheds.
¦ Map of landform subdivisions of Pennsylvania (preliminary).
¦ Natural Resources Conservation Service soil survey maps
¦ Census block data, ecoregion boundaries, Pennsylvania hydrology, and zip codes.
¦ State-wide boundary and county-wide roads coverages from the Pennsylvania Dept.
of Transportation.
¦ Multiple use coverages from the Pennsylvania Department of Environmental
Protection
¦ Satellite imagery
¦ County-wide streams coverages from Penn State's Environmental Resources Research
Institute (Streams Networks).
¦ State- and county-wide surface geology coverages.
¦ Digital elevation models for Pennsylvania available from USGS National Elevation
Database (NED).
One of the data sets that can be downloaded from the PASDA web site is the ALLARM
database. ALLARM is a project of the Environmental Studies Department at Dickinson College,
who partners with Pennsylvania communities and individuals who are working to protect and
restore watersheds. ALLARM's partner watershed groups include:
¦ Conodoguinet Creek Watershed Association (CCWA), Cumberland County
¦ Codorus Creek Monitoring Network, York County
¦ Pine Creek Headwaters Protection Group, Tioga County
¦ Kiski-Conemaugh Alliance, Armstrong, Indiana, Westmoreland, Cambria, and Somerset
Counties [Three groups: Roaring Run Watershed Association, Conemaugh Valley
Conservancy, Stony Creek - Conemaugh Rivers Improvement Project]
¦ Coplay Creek Monitoring Project - Lehigh Valley Chapter of the Sierra Club, Lehigh
County
¦ Ridge & Valley Streamkeepers, Bedford County
B-59
-------
DRAFT
¦ Shermans Creek Conservation Association, Perry County
¦ Schuylkill Office, Delaware Riverkeeper Network
¦ Powells and Armstrong Creek Conservation Association, Dauphin County
¦ Fishing Creek Watershed Association, Columbia County
ALLARM's database contains more than 33,466 records on water quality from 1986 to the
present from 622 testing sites throughout Pennsylvania. Information in records includes at least
alkalinity and pH and includes nitrates and phosphates for some sites since 1996. A ZIP file with
ALLARM's full database in Microsoft® Access© format with records linked through SQL to
Arc View© shapefiles can currently be downloaded from the PASDA data download web page.
The PASDA database is updated every six months. Data summaries are also available from the
ALLARM website (http://www.dickinson.edu/storg/allarm/mainpages/database.html).
Also, on PASDA's "PA Explorer" web page (http://www.pasda.psu.edu/explore/index. shtml),
there are several links to web sites that allow users to create their own maps of PASDA data over
the Internet. This section is ideal for users who are unfamiliar with GIS or lack GIS software to
map their data.
The data sources accessible from the PASDA website can help "Define the Impairment" or "List
Candidate Causes" in the Stressor Identification Process for sites in Pennsylvania.
Susquehanna River Basin Commission (SRBC) Nutrient Assessment Program
http ://www. srbc.net/nutrientdata.htm
http://www.srbc.net/trends.htm
http://www.srbc.net/nutrientloads.htm
"The Susquehanna River Basin Commission implemented a five-year nutrient-monitoring
program in October 1984 to establish a database for estimating nutrient and suspended sediment
loads in the Susquehanna River Basin. The initial five-year program was concluded at the end of
December 1989, and five of the twelve original sites were selected for continued long-term
monitoring." Thus, the data set available at this web site includes nutrient and sediment data
collected between 1984 and 2000 at various waterbodies in the Susquehanna River Basin. The
raw data as well as data summaries (trends and loads) can be viewed on the web or downloaded.
The nutrient and sediment data web page (
http://www.srbc.net/nutrientdata.htm) summarizes the sample collection methods, parameters
analyzed and types of data that can be downloaded. The current data files are formatted as
comma delimited text files. Nutrient and sediment trends and loads statistics and maps can be
downloaded in PDF or Excel format from the following web pages:
http://www.srbc.net/trends.htm and http://www.srbc.net/nutrientloads.htm. This monitoring data
could be helpful to "List Candidate Causes" in the Stressor Identification Process for sites in the
Susquehanna River Basin.
B-60
-------
DRAFT
USGS National Water Quality Assessment (NAWQA) Data Warehouse
http://orxddwimdn.er.usgs.gov/servlet/page? pageid=543& dad=portal30& schema=P0RTAL3Q
The U.S. Geological Survey (USGS) began its NAWQA (National Water Quality
Assessment) program in 1991, systematically collecting chemical, biological, and
physical water quality data from study units (basins) across the nation. This data
warehouse contains data from the first 36 study units. Data from the 15 study units
started in 1997 will be incorporated later. The data warehouse currently contains and
links the following data:
¦ Chemical concentrations in water, bed sediment, and aquatic organism tissues for about 500
chemical constituents;
¦ Site, basin, well and network characteristics with many descriptive variables;
¦ Daily stream flow information for fixed sampling sites;
¦ Ground water levels for sampled wells;
¦ 2,800 stream sites and 5,000 wells;
¦ 26,000 nutrient samples and 15,000 pesticide samples as well as 5,000 VOC samples;
¦ 1200 samples of bed sediment and aquatic organism tissues.
There are 6 study units in the mid-Atlantic region: Delaware River Basin, Lower Susquehanna
River Basin, Delmarva Peninsula, Potomac River Basin, Allegheny and Monongahela Basins and
Kanawha-New River Basin. Data collected between 1991 through 1997 is available for download
from the data warehouse for a few of these study units. More information about each study unit
(e.g., sampling efforts, summary reports) can be accessed at the web pages listed below.
Allegheny-Monogahela River Basin NAWQA data
http://pa.water.usgs.gov/almn/almn main data.html
Delaware River Basin
http://ni.usgs.gov/delr/sampling.html
Kanawha-New River Basin
http://www-wv.er.usgs.gov/nawqa/
This study unit was one of the second set of study units which began investigations in
1994. The period of the study has ended and the data are no longer available online. For
information on the Kanawha-New River Basin study unit or to request available reports
contact:
District Chief
U.S. Geological Survey Water Resources
11 Dunbar Street Charleston, WV 25301
(304)347-5130
B-61
-------
DRAFT
Lower Susquehanna River Basin
http://wwwpah2o.er.usgs.gov/proiects/lsus/
Further information on the survey and on the specific data for each sampling site can be
obtained from Kevin J. Breen or Michael D. Bilger (New Cumberland, PA) at (717) 730-
6900.
Potomac NA WQA Datasets
http: //md. water. usgs. gov/pnawqa/ datarpt/
Water quality and spatial data collected as part of the Water-Quality Assessment of the
Potomac River Basin from 1992 to 1996 is available for download from the web page
listed above. The data set includes chemical, physical, streamflow, bed-sediment
contaminants, aquatic-tissue contaminants, fish community, and selected stream habitat
data. These data were collected at a variety of water bodies in the Potomac River Basin.
The data from the data warehouse can be viewed on-line, printed, or exported in several
formats (e.g., Excel, Lotus 1-2-3, HTM, tab delimited, comma delimited). Data on the web
pages specific to a study unit can be viewed on-line or downloaded in HTM, PDF or TXT
format from the web page (there is no export function). This monitoring data may be useful to
"Define the Impairment" & "List Candidate Causes" as part of the Stressor Identification
Process for sites located within the study units mentioned above.
Virginia Department of Environmental Quality -Water Quality Program
https://www.deq.state.va.us/webapp/wqm.homepage
Virginia DEQ monitors water quality in rivers, lakes and estuaries at some 1,100 locations to
detect pollutants and to assess pollution prevention efforts. DEQ staff in each of the regional
offices collects water samples on a routine schedule at more than 1,000 locations across the
Commonwealth. The tens of thousands of samples and chemical test results generated each year
are kept in a computer database that is updated daily. In addition to recent data, the database
contains historical STORET data collected by the VA DEQ as well as other historical data going
back as far back as far as 1941. This water quality data can be queried by County/City or Stream
at the Water Quality Monitoring Homepage (
https://www.deq.state.va.us/webapp/wqin.hoinepage). Sample location information (e.g., stream
name, type of waterbody, watershed, latitude & longitude), sample date and time, and chemical
results are given as a result of the query. Biological data (e.g., macroinvertebrate data) is not
included in this database. Rather, this type of data is compiled in databases at 7 regional DEQ
offices. DEQ also supports a cooperative arrangement with hundreds of citizen volunteers who
collect water quality data throughout the Commonwealth, but the data collected by these
volunteers are not compiled in one central database. This monitoring data may be useful to "List
Candidate Causes" as part of the Stressor Identification Process for sites located in Virginia.
Virginia's 2000 305(b) Water Quality Assessment Report can be downloaded from
http://www.deq.state.va.us/water/305b.html. This report may also be helpful to "List Candidate
B-62
-------
DRAFT
Causes" at sites in Virginia. Also, Virginia's water quality criteria and standards can be
downloaded in PDF format from http://www.deq.state.va.us/wqs/#vawqs. These criteria can be
used to "Analyze Evidence" at sites in Virginia.
West Virginia Department of Environmental Protection, Division of Water Resources,
Watershed Assessment and Strategic Planning Section
http://www.dep.state.wv.us/wr/index.cfm?page=OWR Website/C and D/WAP.htm
"The Watershed Assessment Program collects and compiles water quality data on waters
statewide. The Strategic Planning group uses the water quality data to compile the state's list of
impaired streams and write and implement water quality improvement plans known as total
maximum daily loads, or TMDLs. " Activities undertaken by these programs that may provide
useful information for the "Define the Impairment" and "List Candidate Causes" steps of the
Stressor Identification Process include:
Watershed Assessments
The results of chemical, habitat, and biological monitoring at many of the states 9000+ streams
are summarized in Watershed Assessment Reports that can be downloaded from this website.
Citizens Volunteer Monitoring
WV Save Our Streams is a program empowering citizens to monitor the water quality of their
local streams via chemical and/or biologic methods. Volunteers collect information about: water
quality (pH, DO, temperature, conductivity), water conditions (clarity, color, odor),
macroinvertebrate counts, characteristics of the stream bed, characteristics of surrounding
riparian zone and surrounding land uses. A few groups also collect water and sediment samples
for chemical analysis. There are links to the survey forms used by the volunteers at the WV Save
Our Streams website
(http://www.dep. state. wv.us/wr/index.cfm?page=OWR_Website/wvsos/wvsos.htm). Most of the
stream survey data (collected since the early 1990's) are in paper format. However, in the past
few years, an access style database has been set-up to log data collected by volunteers. Each year
a report is produced from this database. These annual reports are available on the WV DEP
website. To obtain data for a particular waterbody or watershed, you can contact:
Timothy Craddock, Citizens Monitoring Coordinator
WVDEP, Office of Water Resources
1201 Greenbrier Street
Charleston, WV 25311
Phone: 304-558-2108
Fax: 304-558-2780
tcraddock@mail.dep. state, wv.us
Large Stream Ambient Monitoring Network
Quarterly monitoring for chemical constituents at 26 locations on West Virginia's larger rivers
and streams. A map of and information about ambient stations is available from this website.
B-63
-------
DRAFT
Special Biological Studies
Analysis of point source effluents on biological stream quality at selected sites statewide.
305(b) Reporting
Assessment of the condition of the state's waters, released every two years in a report. The 1998
305(b) Report and 2000 305(b) Report are available on-line. The 2002 305(b) report is expected
to be released in April of 2002. The following information is provided in the 305(b) report for
each watershed in the surface water assessment section: Storet sampling locations, overall
designated use report, use support matrix summary, relative assessment of causes, relative
assessment of sources, and 303(d) listed streams. There are also sections on Lake Water Quality
Assessment, Groundwater Quality and Wetlands.
303(d) Listing
Listing of streams which are not meeting water quality standards, often referred to as the
impaired streams list. Work is underway to compile the 2002 303(d) list and is expected to be
finalized in April of 2002. The 1996 303d List, 1998 303d List and 2002 303d List Call for Data
are available on-line.
The Environmental Quality Board of the WV DEP is responsible for issuing rules that set the
water quality standards for West Virginia's surface and ground waters. West Virginia's surface
water quality standards can be accessed by following the links at:
http://www.dep.state.wv.us/eqboard/index.cfm?page=waterqualitv.htm
B-64
-------
APPENDIX C
CASE STUDY MATERIALS AND REFERENCES
-------
CADDIS Workshop
Mid Atlantic Case Study
Identify Impairment and List Candidate Causes
The Premise of the Case Study: Stressor identification sufficient to support effective water
quality management can be performed at the scale of small watersheds in addition to reach/site
specific scales.
Biological impairment is often the result of multiple processes occurring over the
drainage area. There may be multiple "pollutants" (i.e., proximal stressors) affecting the biota.
However, these stressors in many cases may derive from a single source or human activity. As
an example, biota in the receiving waters of a small watershed may be impaired due to the
multiple insults of excess nutrients and suspended solids, increased temperature regimes, and
increased sedimentation. However, all these may be the result of conversion of riparian forest in
the watershed to agriculture. The common source of these stressors is therefore the agricultural
fields along the streams.
The current framework for water quality standards focuses on identifying the causative
stressors/pollutants, identifying the contributing sources, then allocating maximum loading for
the stressors among those sources. However, the larger-scale activity (e.g., the areal pattern of
agriculture in the drainage basin relative to its distance to the receiving waters) may be the
reason for a particular pattern of multiple site-specific instances of biological impairment. It
may therefore be more efficient and cost-effective to recognize and deal with the principle cause
of the impairments (the pattern of sources) as opposed to the site-specific, stressor-specific
cause(s) of any particular biological impairment.
At this scale of attention, the focus on causality is one step back in the pathway. The
presence, co-occurrence, etc. of a particular suite of stressors is supporting evidence for a
particular source or land use pattern. Conversely, if evidence for a particular source or land use
pattern is easily identified, and the cumulative effects of such a pattern are already scientifically
supported, resources need not be devoted to identifying specific stressor-response relationships
for each individual stressor at each specific site, but can instead be devoted to determining the
most effective management action to take to reduce the overall biological impairment.
The Challenge: The Office of Water is increasing emphasis on the states' Continuing Planning
Process, supporting increased use of watershed management practices. To deal with the huge
volume of TMDLs, OW is also strongly recommending "batching" or grouping TMDLs in the
same watershed, with the recognition that many of the TMDLs are interrelated. Solutions for
those impairments will require effective holistic management action to get at the root source or
land use pattern. States need methods and case studies to implement OW requirements.
Obviously, there are significant gaps in our knowledge. As one example, quantification
of the correlations between particular land use patterns and the resultant environmental effects
(particularly biology) is in its infancy. However, so is quantification of stressor-response
relationships, particularly in terms of understanding multiple, cumulative and synergistic
stressor-response. The Stressor Identification Guidance articulates a flexible structure to
evaluate evidence, using well-known principles for strength of evidence analysis. It recognizes
gaps in both current knowledge and in data availability, and provides an approach for drawing
C-l
-------
conclusions anyway.
Using two kinds of case studies at the CADDIS workshop (one focused on a reach-
specific analysis, one focused on a small watershed approach) will help us to design a more
useful decision support information system, and, as added benefits, may help refine the method
for this type of case as well as provide an example for states struggling to implement the new
OW directions. The Mid Atlantic case study walk through the structured SI approach for the
Cabin John watershed with the intent of conceptualizing an information system that will help
investigators find, access, organize and share information useful for evaluation of the cause(s) of
impairment at a small watershed scale.
The Situation: (Van Ness and Haddway, 1999)
The Cabin John Creek watershed is within the Potomac River drainage in southern
Montgomery County (Map 1). The watershed, located in the Piedmont ecoregion, occupies
approximately 16,022 acres. Cabin John Creek originates within the city of Rockville and flows
in a southerly direction to a confluence with the Potomac River between the Little Falls dam and
Great Falls. The Watershed is bounded by Rockville Pike (Rt. 355) and Old Georgetown Pike
(Rt. 187) to the east and Falls Rd.(Rt. 189) to the west.
The watershed has been significantly impacted by suburban development patterns
centered around the County's main transportation corridors. The Interstate 495/270 corridor
passes through the central part of the watershed, and commercial and high density residential
development are common along this corridor, particularly affecting the eastern tributaries.
Rockville Pike and the City of Rockville occupy the headwaters of Cabin John. In contrast, the
western tributaries transition to lower density residential communities with far less commercial
development. On-site stormwater runoff controls are uncommon in Cabin John. Like many of the
watersheds in the lower part of the county, this area developed before environmental regulations
for stream buffers and stormwater management went into effect.
C-2
-------
Cabin John Creek Watershed
fftureicm
Upper Mainskni
Map I
Land Cover
Srakedsri
ifc-jiidi
ddFarm
Branch
Buck Brawh
Middle MUfflOn
Urns-
Btids
Branch
Lower
-------
Define the Scope of the Investigation:
Montgomery County uses biomonitoring to assess the cumulative environmental impacts
of economic development on the health of its local streams and ecosystems. This data has
allowed them to develop a countywide stream protection strategy to use in their planning
process. The local-level biomonitoring data feed into the state's NPDES point source
permitting, and the state's CWA 305(b) and 303(d) reporting processes, as well as identifying
county priorities for stream restoration. The county adopted the state's MBSS protocols in 2001,
thus ensuring full compatibility with state-level assessment.
The county collects information on fish, benthic macroinvertebrates, and habitat, and has
developed IBIs based on the MBSS Eastern Piedmont indices. Site-specific IBI results are used
to estimate the extent of streams in excellent, good, fair and poor condition with respect to the
biotic integrity of the fish and benthic communities. Table C-l contains detailed descriptions for
each of the IBI categories. The highest scores were designated as good recognizing that
reference sites may not represent the highest attainable condition
Table C-l: Example of Biological Integrity Classification Method Using Total Index of Biological Integrity Scores, Narrative
Biological Integrity Class Descriptions, and Characteristics of Each Class
IBI Score
Narrative
Integrity Class
Characteristics
46 to 50
Excellent
Comparable to the biological community found in reference streams. Exceptional
assemblage of species with a balances community composition
34 to 45
Good
Decreased number of sensitive species, decreased number of specialized feeding groups,
some intolerant species present
22 to 33
Fair
Intolerant and sensitive species are largely absent; unbalanced feeding group structure
10 to 21
Poor
Top carnivores and many expected species absent or rare; general feeders and tolerant
species dominant.
C-4
-------
Describe the Impairment:
Analysis of the 1996 Stream Monitoring and Habitat Data
Ten stations were monitored in Cabin John Creek in 1996 (Map 3). The data is shown in Table
C-2. The overall resource condition of Cabin John Creek was determined by assessing the
cumulative impacts that occurred in the watershed as indicated by the use of an interim Index of
Biological Integrity (IBI) for freshwater fish and benthic macroinvertebrates (Map 3). A yearly
assessment of "excellent", "good", "fair", or "poor" resource condition was made by examining
the trends expressed by the two IBI's. This is not the same as averaging the two scores. Seasonal
trends were examined and a yearly stream condition was established for the subwatersheds.
Resource conditions were evaluated for 13 subwatersheds. Buck Branch and Ken Branch
received an overall resource condition of good. Snakeden Branch, Thomas Branch, and Booze
Creek received an overall resource condition rating of poor. The entire mainstem received an
overall resource condition assessment of fair (Map 3).
Table C-2
Monitoring Data, Cabin John Creek Watershed, 1996
STATION NUMBER
IBI (benthic)
IBI (fish)
CJBB201 (w.br.Buck Br.)
22 - FAIR
4.3 - GOOD
CJBC202 (Booze Cr.)
18-FAIR
1.2-POOR
CJCJ302 (mainstem)
16 - POOR
2.8 -FAIR
CJCJ303 (mainstem)
14 - POOR
3.2 -FAIR
CJCJ305 (mainstem)
24 - FAIR
3.7 - GOOD
CJCJ307 (mainstem)
14 - POOR
3 - FAIR
CJKB201 (w.br. Ken Br.)
24 - FAIR
3.7 - GOOD
CJSB101 (Snakedon Br.)
24 - FAIR
1- POOR
CJTBlOla (Thomas Br.)
14 - POOR
1.7-POOR
CJTBlOlb (Thomas Br. trib)
14 - POOR
1.2-POOR
C-5
-------
Cabin John Creek Stream Condition
Bogley
Old Farm Branch
Based on biobgieal indicators.
See Chapter 2 for details.
Stream Biological (VmdiDun
['AlA'Iklli
ttaxi
. Fair
j No (irrent lim
Fish Species Colkxtod
Alien car Eel
GokUish
Ganmn cap
Gdden Airier
RtKyside dace
Creek chub
t-alltisii
Central sunbrotkr
Qjtlips niinow
BJatkmse
-------
REFERENCES
Van Ness, K; Hadway, M. (1999) Cabin John Creek Watershed Study. Montgomery County Department of
Environmental Protection, Watershed Management Division. Rockville, MD. Available for download at
http://www.co.mo.md.us/services/dep/SPA/pdf%20files/cabinReport99.pdf (accessed August 20, 2002).
C-7
-------
LITTLE SCIOTO CASE STUDY
REFERENCES
Cormier, SM;.Norton, SB; Suter, GW; et al. (2002) Determining the Causes of Impairments in the Little Scioto
River, Ohio, USA: Part 2. Characterization of Causes. Environmental Toxicology and Chemistry, 21(6): 1125-
1137.
Norton, SB; Cormier, SM; Suter, GW; et al. (2002) Determining Probable Causes of Ecological Impairment in the
Little Scioto River, Ohio, USA: Part 1. Listing Candidate Causes and Analyzing Evidence. Environmental
Toxicology and Chemistry, 21(6): 1112-1124.
Suter, GW; Norton, SB; Cormier, SM. (2002) A Methodology for Inferring the Causes of Observed Impairments in
Aquatic Ecosystems. Environmental Toxicology and Chemistry, 21(6): 1101-1111.
C-8
-------
Please make all necessary changes on the below label, PRESORTED STANDARD
detach or copy, and return to the address in the upper left- POSTAGE & FEES PAID
hand corner.
United States EPA
Fnwirnnmental Prntertinn Anpnro/nPn If you do not wish to receive these reports CHECK HERE ~; PERMIT No. G-35
tnvironmeniai rroieciion Agency/UKU detach, or codv this cover, and return to the address in the
National Center for upper left-hand corner.
Environmental Assessment
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-02/078
October 2002
------- |