United States
          Environmental Protection
                  Office of Water
                  Washington D.C.
EPA 503/9-92-001
  October, 1991
Workshop on the Water Quality-
based Approach for Point
Source and Nonpoint Source

Meeting  Summary

June 26-28, 1991
Chicago, Illinois
                                            Printed on Recycled Paper

Workshop on the Water Quality-based Approach for
     Point Source and Nonpoint Source Controls
                     Prepared by

                    Tetra Tech, Inc.
                     Fairfax, VA

              EPA Contract No. 68-C9-0013
                        for the

     Office of Wetlands, Oceans and Watersheds (WH-553)
         Office of Science and Technology (WH-585)
           U.S. Environmental Protection Agency
                Washington, D.C. 20460  • +

       This workshop was organized to assess the state of the science and chart future directions
for water quality-based pollution controls. Agencies concerned with water quality are beginning
to address new challenges such as nonpoint source pollution, storm water discharges, habitat
quality, and the condition of the resident ecological community.  Addressing these challenges
requires increasing the scales of analysis from river segments to watersheds, and from a simple
low  flow critical  period to low  and high flow  periods and in some cases multi-decade

       It is also apparent that we will need to integrate various  analytical tools,  in order to
comprehensively address all of the phenomena in a large geographical area such as a watershed.
Finally, there is a growing appreciation that attaining and maintaining water quality  standards
in many cases will require measures to control non-chemical stressors.  The capability  to predict
the ecological response to controls on non-chemical stressors is a critical need.

       The workshop participants discussed all of these topics and developed many  ideas for
improved approaches to water quality analysis. One particularly significant result is the dialogue
begun between  members  of  the different  disciplines brought  together for this workshop.  The
goals of the workshop were largely met and the  findings are expected to establish a blueprint for
future directions.
Bruce Newton                            Elizabeth Southerland
Chief, Watershed Branch                   Chief, Risk Assessment and Management Branch
U.S. Environmental Protection Agency      U.S. Environmental Protection Agency
Washington, D.C.                         Washington,  D.C.

      The Workshop on the Water Quality-Based Approach for Point Source and Nonpoint
Source Controls was jointly sponsored by the Office of Wetlands, Oceans and Watersheds
(OWOW)  and the Office of Science and Technology (OST).   Bruce Newton, Chief of the
Watershed Branch (OWOW) and Elizabeth Southerland, Chief of the Risk Assessment and
Management Branch (OST) are acknowledged for their role in initiating the  workshop and
overseeing its success.  Amy Sosin and Hiranmay  Biswas were the EPA project managers.
Karen Guglielmone of Tetra Tech is acknowledged for her help in coordinating the workshop
and preparing this summary.

                            TABLE OF CONTENTS




1.    Overview of the Workshop on the Water Quality-based Approach
      for Point Source and Nonpoint Source Controls	  1

2.    Opening Remarks	  2

3.    Section 303(d) of the Clean Water Act	  5

4.    Key Insights

         The Regional Perspective on Water Quality-based Point Source
         and Nonpoint Source Controls	9

         The Watershed Approach from a Point Source Perspective  	11

5.    Section Presentations

         Problem Diagnosis

         Top-Down and Bottom-up Approaches for Integrated Assessment
         of Point and Nonpoint Source Pollution	14

         Diagnosis and Assessment Tools/Models for Determining
         Nutrient TMDLs	16

         Watershed Modeling

         Watershed Modeling and TMDL Assessments	21

         Water Quality Diagnosis and Assessment Tools/Models	22

         Model Capabilities (Agricultural and Urban)	25

         Model Capabilities - A User Focus	26

         North Carolina's Basinwide Water Quality Management
         Approach to Developing TMDLs	28

         Modeling Integration and Data Access Tools	30

         Data Sources, Tools, and Investigations

         Multistage Remote Sensing Data Applications for CIS Data Base
         Development in Support of Nonpoint Watershed Modeling  	33

         GIS  for Nonpoint Source Watershed Modeling Applications	34

         Grass Waterworks - An Interface Between GIS and Models	35

         Remote Sensing of Agricultural Practices and Downstream Water
         Quality as Influenced by Sediment from Nonpoint Sources  	36

         The  Trials, Tribulations, and Successes of Using Remote Sensing,
         GIS, and Modeling  	37

         Predictive Modeling of Ecological Restoration

         An Overview  of Ecological Assessment and Restoration Tools	41

         Application of Tools for Ecological Restoration Predictive
         Modeling	42

         Ecosystem Assessment and Restoration:  The Role of
         Modeling and Predictability  . .  .•	44

6.    Workgroup Breakout Session Summaries and Recommendations

         Watershed Modeling	45

         Data Sources, Tools, and Investigations	51

         Predictive Modeling of Ecological Restoration	59

         Selected Point Source Issues	64

7.    Conclusions	72

APPENDIX A:        EPA Mainframe Data and Tools for Watershed
                      Assessments	 A-l

APPENDIX B:        Workshop Agenda	B-l

APPENDIX C:        List of Participants   	C-l


          1.  Overview of the Workshop on the Water Quality-based Approach
                    for Point Source and Nonpoint Source Controls
       The Workshop on the Water Quality-based Approach for Point Source and Nonpoint
Source Controls was held June 26, 27, and  28 in  Chicago, Illinois.  The  purpose  of the
workshop was to explore the state-of-the-science and determine technical guidance needs for
implementing section 303(d)  of the Clean Water Act. More specifically, the workshop served

•      Learn from  national  experts  and the user community what tools are available to
       implement the 303(d) Total Maximum Daily Load (TMDL) process on a watershed basis,
       and what national technical guidance EPA should provide to the user community.

•      Identify tools that can currently be incorporated  into national technical guidance.

•      Examine the trends of emerging science and technology, including ecological restoration,
       remote sensing,  and  geographic information  systems  (CIS) to identify promising
       approaches for future guidance.

•      Identify tools that can  integrate point source and nonpoint source pollution processes into
       a  single integrated approach for solving water quality  problems.

       Carl Myers,  Deputy  Director of U.S.  EPA's  Assessment  and Watershed Protection
Division  in the Office of Water, and William Diamond, Director of U.S.  EPA's Standards and
Applied Sciences Division in the Office of Science and  Technology, opened the meeting.  The
first day continued with presentations on problem diagnosis, watershed-scale modeling, and data
sources,  tools, and investigations.  Presentations on remote  sensing  and CIS, and predictive
modeling of ecological restoration continued through the first half of  the second  day; then the
workshop broke into four  concurrent workgroup sessions.   On the  final day,  the  four
workgroups reported their findings and recommendations.

       This meeting summary includes the opening  addresses,  abstracts for each  of the
presentations, and a summary of the work and recommendations for each of the  workgroup

       Included as appendices are the list of participants,  the agenda, and a summary of the
special EPA data base presentation.


                                2.  Opening Remarks
               Carl Myers, U.S. EPA, Deputy Director, Assessment and
                        Watershed Protection Division, OWOW

       As part of the Office  of Water's  recent reorganization, the Office  of Science  and
Technology (OST), the Office  of Wastewater Enforcement and Compliance (OWEC) and the
Office of Wetlands, Oceans and Watersheds (OWOW) have key responsibilities concerning the
water quality-based approach to  making point source and nonpoint source pollution  control
decisions. The Assessment and Watershed Protection Division (AWPD) of OWOW will be one
of the primary users  and promoters of this approach.  The Standards and  Applied Science
Division of OST will be working with AWPD to  incorporate it into the permitting process.

       The Workshop on the Water Quality-based Approach  to Point Source and  Nonpoint
Source Controls comes at a critical time. EPA has recently issued concise and useful guidance
that explains  the Total Maximum Daily Load  (TMDL) process, and how it can help water
quality managers  make better decisions; reauthorization of the Clean Water Act (CWA) has
created new language and a renewed emphasis on the entire TMDL process; and revised 303(d)
regulations that will re-emphasize and re-enforce several aspects of the TMDL process,  such as
targeting and prioritizing watersheds, and developing TMDLs on schedule will be issued later
this fall.  More pollution management activities are focusing on watersheds and ecosystems, as
opposed to individual permits or individual construction grants.  EPA administrator, Bill Reilly,
has shown great interest in advancing habitat protection and related issues, and is putting more
emphasis on the wider geographic influences that affect specific water bodies.  Much of this
emphasis is a direct result of the 1990 Science Advisory Board report, Reducing Risk, which
stated that as much importance should be  placed on reducing ecological risk as on reducing
human health risk. All this as EPA and the  States begin to use the water quality-based approach
more  extensively.

       Until now,  some states have hardly  used the water quality-based approach at all; some
have used it a little; and others have used  it extensively.  Even the states who have used the
process extensively,  however,  have  largely focused on  making specific decisions  involving
National Pollutant Discharge Elimination System (NPDES) permits for individual point sources -
- often just looking at one or two specific chemicals.  Although it is important and necessary to
make such decisions,  the TMDL process  must now be used  more broadly  to  address more
complex pollution control problems, including nonpoint sources and  multiple point  sources
within one watershed.  It is important  to move beyond the chemical-by-chemical, permit-by-
permit approach to encompass all threats to a given watershed.

       The Office of Water is working with the Regions and States on a Watershed Protection
Initiative which will focus water management efforts and martial resources to help States make
better, more comprehensive decisions and implement pollution control measures on a watershed
basis. Many OWOW programs, such as the Nonpoint Source Program, the Wetlands Program,
various coastal initiatives, and the National Estuary Program, already focus on watersheds and

require managers to make water quality-based decisions.  The goal of the WPI is to help re-
orient aU programs so that they operate on a targeted watershed basis.

      What else do we need in order to move ahead with the water quality-based approach?
This workshop should help to answer that very question.  While it is easy to say water managers
must  use TMDLs to  make decisions on  the pollution  controls necessary in a  particular
watershed, until now, we have not always been able to use available information beyond point
source decisions.  Recognizing why application of the TMDL process has been so limited, and
that we must gather and develop the technological tools and the approaches and the methods that
will let us use the TMDL process more broadly, is  the key.
             William Diamond, U.S. EPA, Director, Standards and Applied
                                Science Division, OST

       The reorganization of the Office of Water reflects the importance that is being placed on
total water quality  management.  The reorganization also serves  to focus more resources,
attention, time, and decisions on making TMDLs an integral part of the entire water quality
standards and permits process.

       This new direction has been influenced by several factors. First, although the recently
adopted  State  toxic water  quality  standards  were  controversial  during   Congressional
reauthorization, they mean little unless they can be quickly translated into real pollution controls.
The TMDL process can facilitate this translation. Second, the CWA Amendments on TMDLs
have focused political attention on the TMDL  process.   The Baucus/Chaffee bill has several
provisions  relating  to water quality-based decisions which, if enacted, would require rapid
adoption of more comprehensive TMDLs.  Third, over the last couple of years there has been
more TMDL-related litigation, indicating to EPA that it is time to enforce section 303(d) more
strictly. Negotiations on a suit in Alaska have recently closed; there are other suits pending; and
EPA expects  to see more as the courts are  seen as a successful route to 303(d) enforcement.
Finally, environmental programmatic needs are increasing and  require increasingly complex
decisions that can only be addressed through a comprehensive water quality-based approach.

       The explosive expansion of the water quality criteria and standards program will probably
continue through the early 1990s as we try to go beyond chemical specific criteria and deal with
such long-standing environmental  quality issues as contaminated sediment, nonpoint sources,
combined sewer overflows, and habitat destruction.

       This year EPA is proposing initial criteria to address contaminated sediment. A sediment
management strategy to deal with the range of issues concerning State adoption of sediment
standards is under development.  One question  that must be addressed is, how to model and
track the standards?  Another is, how to integrate the standards into the TMDL process?

       Combined sewer  overflows are a potential multibillion  dollar problem  that is being
addressed by the CWA and several  Agency programs.  In order to deal effectively with this
problem, wet-weather criteria must be developed.  Environmental professionals and modelers

must determine how to deal with nonpoint source pollutant loadings during wet-weather high
flows.  (Criteria for point sources and chemical specific toxics have traditionally been based on
critical low flow concentrations.)

       Habitat  destruction  is a  serious  threat to  many  watersheds.    Biologically-  and
ecologically-based  criteria must be established in order to address this problem.  However,
establishing such standards  and criteria, and then making them applicable as regulations and
policies is a tremendous challenge.   Through an integrated watershed approach, the TMDL
process offers a mechanism  to address both point  and  nonpoint pollution sources  that can
degrade valued habitat.

       TMDL assessment, monitoring, and modeling will play a crucial role in whether we meet
these challenges successfully; and the Office of Water recognizes that more time, resources, and
attention will be necessary.  This workshop will serve two broad purposes.  First, to learn from
technical experts and water  quality managers what technical issues must be addressed in order
to successfully  implement the TMDL process;  and second, to prioritize the recommendations
made by workshop participants.  Establishing priorities is critical to the management and success
of this  program not only in  the short term, but in the long term  as well. Over the next couple
of months we will be putting together a program plan based on the needs and priorities that we
establish here at this workshop.


                       3. Section 303(d) of the Clean Water Act
                 Bruce Newton, U.S. EPA, Chief, Watershed Branch,
                Assessment and Watershed Protection Division, OWOW

Section 303(d) of the Clean Water Act

       The water quality-based approach outlined in section 303 of the Clean Water Act has
three major elements. Subparagraphs a, b, and c of section 303 establish  State Water Quality
Standards, which include water quality criteria, as well as designated uses- and antidegradation
provisions. Section 303(d) establishes the process for setting pollution control requirements to
meet water quality standards.  Section  303(e) requires a  statewide water  quality planning
process.   These elements are the  essence  of  the  water quality-based  approach originally
envisioned in 1972.

       Section 303(d) itself is a three step process. The first step requires  States to identify all
waters that do not  meet water quality standards  or are not expected to  meet water quality
standards after the application of technology-based controls.  The second step requires them to
establish priorities among all identified water quality-limited waters. The third step is to develop
a control plan that is termed a total maximum daily load (TMDL).  [TMDL is a somewhat
antiquated term as EPA moves to address such topics as habitat protection and nonpoint source
controls; it is important not to be hindered by the limiting nature of this terminology.]

       Section 303(d) was enacted in 1972, but was  never fully  implemented. This is largely
because requirements for point sources were developed in conjunction with  permit issuance.
There are two principal reasons that EPA is now moving to more fully implement the water
quality-based management process originally envisioned.  The first is that the TMDL Process
is a logical way to address certain non-traditional pollution problems, such as nonpoint sources,
non-chemical stressors, and  cumulative effects  which are now recognized as major threats to
water quality. It is, therefore, time to broaden  the scope of decision making to include all
pollution sources in a watershed.  The ultimate goal  is to be able to consider the landscape as
a whole - looking at both terrestrial  and aquatic ecosystems together.  This change in the scope
of decision making fits well  with the requirements of 303(d).

       The second reason is to emphasize the several concepts inherent in the 303(d)  process
that will be necessary as we begin to address  more complex water management problems.  One
key concept of the 303(d) process is targeting and setting priorities.   Water quality managers
have come to the conclusion that progress can only be made on water quality  issues associated
with nonpoint source controls through  targeting.  Targeting will lead to better coordination
among more agencies and groups to develop pollution controls and protect critical habitat. Many
of the decisions that need to be made, are made at the local level.  This means that the public
must be engaged more than in the past;  and that is best done through targeting. There are also
specific public participation  requirements in 303(d)  that will  help to engage and educate the

Section 303 (d) Guidance

       EPA recently issued a guidance document on the water quality-based approach that
establishes Agency policy in this area.  The document outlines the programmatic requirements
of the TMDL process.   It is important to understand  these requirements before discussing the
technological tools that may be available to help decision makers. Specifically, the document
sets out EPA expectations concerning 303(d) - that States will do TMDLs, and that if they fail,
the Agency  will.  It establishes Agency  policy concerning nonpoint source waters -- that a
TMDL  must consider nonpoint sources, and that  we expect progress in  the area of nonpoint
source control despite uncertainties in data and uncertainties inherent in the use of models. The
Guidance also addresses policies concerning nonchemical stressors. EPA asserts that the TMDL
process  does apply to nonchemical stressors, to the extent that you can quantify the stressor and
tie it to water quality standards.  For example, if it is determined that designated uses are not
being attained due to lack of stream side vegetation and a measure such as percent canopy cover
can be tied to attainment of water quality standards,  then a TMDL for percent canopy cover
could be written.

      The Guidance details procedural aspects of the TMDL program, such as the submission
of lists -- which waters should be listed and which are  exempt -- and the identification of waters
targeted for TMDL development.  Both within the program guidance, and in the regulation that
we are in the process of finalizing, States are being asked to  identify, target, and list waters for
TMDL development on a two year cycle. This listing cycle corresponds closely with the process
under section 305(b) which also requires reporting of water  quality every two years.

       EPA's role in administrating the TMDL Process is to review and approve or disapprove
State lists, priorities, and TMDLs.  If EPA disapproves any of these, it is obliged to assume the
responsibility and do the work.  A "long series of court cases has also established that EPA is
obliged  to do the work if a State fails  to.

      The Guidance document also addresses the phased approach  for TMDL development.
This management approach applies where use of data and  models result  in a high degree of
uncertainty.  It establishes specific requirements for implementation schedules, monitoring, and
water quality standards attainment after the implementation of pollution controls.

      The phased approach strikes a balance between the need to move ahead and not wait for
the perfect model or the ideal amount of data,  and  the statutory requirement that where there is
uncertainty it must be accounted for with a margin of safety.  The process provides a  middle
ground to move ahead where we do not have as firm an understanding as we would like, but
also provides some safeguards in terms of follow-up monitoring and implementation  schedules.
If monitoring reveals that a TMDL is not  as stringent as it should be, for example, it can then
be revised.

The Workshop

       The heart of the 303(d) process is predictive modeling. We need to be able to determine
whether or not a control measure will result in the attainment and maintenance of water quality
standards.  We also need to be able to set defensible, realistic targets for allowable amounts of
pollution. Predictive modeling permits us to do this.

       Maintaining the diversity of participants at this workshop was important.  Our intention
was not to conduct a conference exclusively for modeling engineers to speak with each other
about models; instead,  we invited people from the remote sensing and geographic information
system (CIS) community, and others involved with a fairly  new area of research — predictive
modeling of ecological  restoration. About twenty-five percent of the workshop participants are
considered to be representatives  of the user  community — people who  can  speak for the
"customers" that we need to satisfy with regard to the level  of expertise and usability of these

The objectives of this workshop are threefold:

1.      To review the current state of the science to see exactly where we stand  in the face of
       these new challenges;

2.      To obtain expert recommendations about which approaches and tools can be used over
       the next three to five years; and

3.      To  obtain  prioritized recommendations about which tools are  most promising for
       development over the  next five to ten years.

       The workshop has been organized into key topic areas:  Problem Diagnosis; Watershed
Modeling;  Data Sources, Tools,  and Investigations; and Predictive Modeling of Ecological

       Problem  diagnosis is  important  for watershed managers.   When  dealing with point
sources it is fairly obvious which parameters need  to be modeled.  As we attempt to manage
watersheds more comprehensively, it will be more difficult to discern which parameters are most
important or indicative of the processes that need to be modeled.

       How to model all pollution sources and impacts over larger geographic scales is a key
technological issue of the TMDL Process.   It is closely tied to data management,  as well as
monitoring issues. Modeling on a larger scale and consideration of nonpoint sources along with
multiple point sources within a watershed requires the use of such technologies as remote sensing
to provide relevant data, and CIS to integrate the data.  A special demonstration has  been
provided on the current EPA  data systems that are available for use. Participants are asked to
learn  about them and make recommendations about what EPA should be doing  in the area of
data systems management.  Is the data management  approach that we have taken the last couple
of years appropriate?  How should we be directing  our resources with respect to this area?

       Predictive modeling of ecological restoration is a vital, emerging field of research that
has tremendous potential within the TMDL Process.  It is a high priority at EPA, and this
conference is the first step that EPA has taken to explore this new area.  In the future, when a
given stressor is identified in a watershed we expect to be able to make recommendations for
specific management actions that will result in an anticipated ecosystem improvement.

       The key to the success of this workshop is the breakout groups. These groups should
focus their discussion on the tools that can be used by State and local governments. We hope
the groups will also discuss how to establish a tiered hierarchy of applicable tools, from simple
screening  tools to more complex numeric  models.  Beyond models, the groups should identify
what  state and  local governments  need to manage watersheds,  as  well  as  barriers  to
implementation that EPA needs to address. Workgroup recommendations should be specific and
realistic. These recommendations  will be  used to develop an agenda to guide EPA  in guidance
development, research, and implementation support over the next three to ten years.


                                    4.  Key Insights
          The Regional Perspective on Water Quality-based Point Source and
                              Nonpoint Source Controls

              Jon Grand, Deputy Director, Water Management Division,
                                 U.S. EPA - Region V

       Personally and  professionally, I am completely committed to the watershed-based
approach to water management. It is the way I learned to deal with water resource issues and
it seems artificial to separate point and nonpoint sources and to address water pollution facility
by facility.  The watershed-based approach is a more effective way to address water quality
management. This workshop is an excellent opportunity to share information and experiences,
promote  the  watershed approach,  and develop a plan  to overcome existing  technical  and
institutional barriers to  implementation.

       With the  1987 amendments to the Clean Water Act (CWA), specifically the mandates to
control storm water and nonpoint source pollution, most of the tools necessary to assist States
with implementation now exist.  The States can target regulatory efforts to protect waters from
all sources of pollution.

       So, why are we here?

       The last 20 years have been a period of fairly significant environmental progress, but a
recent article pointed out that while the threat of a silent spring may be abated, the danger of
lifeless waterways looms ever larger.  Fish and other animals that live in the North American
waterways are disappearing  much  faster than land-based fauna.  Without broad measures to
protect water-dependent creatures from such threats as  pollution, unnatural competition,  and
drainage  and damming  of vital habitats, the rate  of extinction is liable to accelerate.

       So, I ask you again.  Why are we here?

       Because in the last 15 years the North American breeding duck population has gone from
45 million to 31 million -- a 30%  decrease.  Because in the last few years there has been a
dramatic  and mysterious decline in  amphibians.  One study in Oregon showed over an 80%
decline in frogs.  Nobody knows why.  Because worldwide  it is estimated that two to three
species become extinct  per day. This is a rate comparable to cataclysmic extinction periods in
the past.

       Why are  we here?

       Because in the last century, 40 species of North American fish have become extinct due
to habitat change and destruction, pollution, and overfishing. Because since  1986, poor quality
eutrophic lakes have increased by 10%, while good quality mesotrophic lakes have declined by
15%. Because 50% of the 22,000 lakes classified by trophic status are listed as poor or very
poor in water quality.

       As you can see, the task of restoring and protecting our waters is far from complete.
Clearly the more holistic watershed approach will be a factor, if we want to succeed at all.
Region V has placed a high priority on the Great Lakes for pollution control activities.  As
recognized in the Great Lakes Water Quality Agreement of 1978, entered into by the  United
States and Canada, and by sections 108 and 118 of the CWA, the Great Lakes are a uniquely
valuable freshwater resource.  Keeping this in mind, a Great Lakes Initiative which marshals
resources for use in a targeted watershed approach that is intended to protect the Great  Lakes,
has recently been established.

       In addition to the Great  Lakes' intrinsic importance, they are valuable as a national
laboratory for developing effective approaches to control nonpoint pollution sources.  First, the
Great Lakes reflect  a broad cross-section  of  common  nonpoint  source  problems,  including
agricultural  and  urban runoff, in-place contamination, and high quality waters threatened by
growth and  development.  Second, they  afford, the opportunity to provide a  model  for
institutional problem solving relationships among two countries, three EPA regions, and eight
states.   Third, they are addressed by a number of statutory programs and EPA initiatives that
control and  abate point and nonpoint pollution problems comprehensively, thus supporting an
opportunity  to demonstrate integrated approaches to watershed protection.

       Since the vast majority of pollution in  the Great Lakes originates in Region V  States,
this Region has the lead for awarding Great  Lakes set-aside funds.  Set-aside funds will be used
exclusively to support nonpoint source implementation projects in the Great Lakes basin that will
demonstrate effective and innovative approaches that may be used nationally.  These Great Lakes
demonstration projects, as well  as other 319  projects,  willprovide information  that can be
utilized to  incorporate nonpoint source controls  into  total  nitrogen daily  load  processes.
Currently, the role of nonpoint source pollution control, which  is largely reserved to State
management regulation, is a main issue.

       While the Great Lakes basin is Region V Water Division's geographic priority, it is
important not to lose  sight of the need to have strong programs in other areas. The development
and implementation of strong watershed programs (a pollution prevention approach) will prevent
water quality degradation and, therefore, the need to do total maximum daily loads (TMDLs)
in the future.

       In conclusion, I  am  encouraged  by  the efforts  to enhance  our water  resources
management capabilities.  Successful water resource programs  require several key elements:
First, a strong state-local partnership; second, an integrated watershed plan for both watershed
management and water resources use;  and third, technical expertise.   This workshop  is one
means to provide that technical expertise to help States effectively manage their water resources.
Through discussions  here, I  hope that the basis for an ongoing  dialogue, and  a free exchange
of technical knowledge and experience  will be  established.

       Remember that what we do here does not affect us so much as the environmental legacy
we will leave for future generations. Given  your critical roles, you must ask yourself. .  . How
do you want that future to look?

               The Watershed Approach from a Point Source Perspective
                  "Give me a number and I'll follow you anywhere!"

                          William Brandes, U.S. EPA, Chief,
                     Water Quality and Industrial Permits Branch

       It is important to  remind each other that we are trying to establish a truly integrated
watershed approach that includes all pollution sources.  The point source program has got to be
part of this integrated approach.  While nonpoint sources are significant, and there is a necessity
to focus on them, control of point  source pollution and the National  Pollutant  Discharge
Elimination System (NPDES) program should not fall  to the wayside.

       The NPDES program consists of issuing permits to industrial and municipal point source
dischargers throughout the country.  It includes implementation of sewage sludge criteria,  and
also  the new storm water permitting program.  Currently,  storm  water  is perhaps the most
significant aspect of the NPDES program, representing a challenge that will loom large when
we try to address pollution from a watershed perspective.

       Before we  are able to discuss  the TMDL  process  and  its role in water  quality
management, it is important to consider the true scope of our water pollution problems.  For
traditional sources there  are  63,300 point sources,  of which 7,300 are considered  major.
Assuming that 25% of those major permits and 10% of the remaining minor permits will need
controls under the TMDL program, there will be approximately 1,825 major permits and 6,300
minor permits to include in the TMDL development process.

       There are about 115,000.point sources for industrial storm water. That does not include
approximately 215,000 oil and gas facilities, and 100,000 mining facilities.  There are also about
220 municipal separate storm water sewer systems located in the major population centers of the
country.  These contain thousands of pipes and conveyances; authorities have no idea how many.
There are about 1,000 systems in this country that have combined sewer overflows, and this is
estimated to represent about 10,000 point sources.

       All of these traditional and non-traditional point  sources will require NPDES permits; and
according to the CWA, they must meet water quality standards. The permit applications for
stormwater are due  sometime later  this year and in the middle of next  year.   Once the
applications are received, the CWA requires that a NPDES permit be issued within one year.
Dischargers must then comply with specified limitations within three years. So, within the next
four or five years,  each of these hundreds of thousands of pollution sources are supposed to be
in compliance with established water  quality standards.

       Setting priorities and targeting watersheds will be necessary in  order to  effectively
address so many pollution sources within such a  short time  period.  The NPDES program
currently does not  have a priority ranking system based on a watershed approach; the question
of how to set priorities while issuing permits must therefore be addressed.  Permits are generally
on a five year cycle, and they must be re-issued on that cycle.

       Scientifically defensible and implementable watershed scale TMDLs are a specific need
of the NPDES program.  This is the number permit writers would like to have and follow. Such
TMDLs are needed because they form the basis for waste load allocations, which in turn are the
basis of point source controls. Traditionally, scientifically defensible and implementable TMDLs
have meant numbers. When we understand the basis of a number and its associated uncertainty,
it is easier to put into a permit and enforce. There may be viable alternatives to numeric limits,
however, and these alternatives should be discussed during the workshop.  Whether they are
numeric or not, a successful watershed approach,  with a firm  TMDL base,  must be fully
enforceable.   That means that both NPDES dischargers and nonpoint  sources are enforceable,
Otherwise, all efforts in  the watershed arena are wasted.  The TMDL will not be effective.

       Certain aspects of the TMDL process could facilitate the NPDES permitting process.
Since the  TMDL process  considers all pollution sources  in a given watershed at one time,
permitting of multiple facilities  within the same watershed can be coordinated. Since, under the
TMDL process,  a water quality standard  is the focus for all pollutant dischargers,  the use of
water quality standards is further legitimized. The TMDL process opens the water quality basis
for permitting to public comment; a weakness of the program in the past.  The TMDL process
can better address antibacksliding or antidegradation  issues.  If a permit limit is based on a
TMDL and continued monitoring indicates that the  TMDL is too stringent, associated permit
limits  can be  loosened  without  violating antibacksliding and antidegradation requirements.
Finally, the TMDL process simplifies and speeds up permit limit development.

       So, give  me a number  and I'll follow you anywhere ... as long as the  process of
establishing TMDLs  on  a watershed basis is kept practical.  Many  States that  are trying to
establish TMDLs have indicated that the point source permitting program suffers while a TMDL
is being done.  This is largely a question of resources ~ there is only so much funding for State
environmental programs, and  when most of those  resources are being dedicated  to TMDL
development, other programs take a back seat. Other water quality programs that will ultimately
be influenced by the TMDL will stop until the watershed goals are determined.  Establishment
of TMDLs must therefore  be fast and effective.

       In summary, the point source permit program has specific needs  and questions that should
be addressed if it is to be an integrated part of the TMDL process.  They  are:

•      In  order  to  useful the  TMDL,  and  the  basis  for  that TMDL,  must  be  easily
       understandable and clearly expressed.  This will facilitate the translation of the TMDL
       into pollution controls.

•      A  reasonable amount  of  resources  must be dedicated  to  the  quick and effective
       establishment of a TMDL so that other water management programs for a watershed can
       proceed on schedule.

•      All TMDLs must be accompanied by a load allocation scheme; otherwise implementation
       will become a nightmare as dischargers haggle over who was given what.

Redefining or renaming "TMDL" would eliminate confusion with the traditional point
source view of it as being chemical and facility specific.

The use of ecological measures such as biocriteria, as TMDLs must be clarified.  Can
such measures be put into an NPDES permit?

Finally, a comprehensive "how to" document is badly needed.  There are numerous
documents on waste  load allocations and  thousands of modeling documents, but the
information is scattered.   A comprehensive document that covers all of the aspects of
TMDL  development  for  watersheds with  multiple  pollutants and  multiple pollutant
sources  would very useful.


                               5. Section Presentations
       This section provides abstracts of the presentations that were given at the workshop.  Any
questions that were asked or comments that were made at the end of a presentation are provided
at the end of each abstract.
                               PROBLEM DIAGNOSIS
           Top-Down and Bottom-Up Approaches for Integrated Assessment
                        of Point and Nonpoint Source Pollution

                                 Kenneth L. Dickson
                             Institute of Applied Sciences
                             North Texas State University
                                    Denton, Texas

       Effective assessment and abatement of the impacts of point and nonpoint source pollution
requires a watershed  perspective and  the  use of a combination of physical,  chemical, and
biological resource characterization techniques.  When looking at the watershed, using only
bottom-up techniques (i.e., in-situ sampling, monitoring) is not sufficient. Techniques, such as
remote sensing and geographic information systems (CIS), that look at the watershed from a
total systems perspective - from the top down -- must also be used.

       Application of remote sensing and CIS to evaluate land uses in basins and watersheds,
and to diagnose actual or potential water quality problems is more and  more common.  For
example, the city  of Dallas operates a number of reservoirs in the north central Texas region.
There is no public control of the watersheds and several years ago, the City became concerned
about nonpoint source impacts on the water quality  in their reservoirs.  Remote sensing and a
CIS was used  to evaluate the types of land uses in  the region, to determine land  use changes
over time.  This information was used with the universal soil loss equation to delineate areas
with the largest erosion and and nutrient loss,  and therefore, where the best locations for water
quality monitoring of nonpoint source  loadings might be.  Once known, those problem areas
were targeted for nonpoint pollution mitigation programs on a watershed basis.  The information
contained in the CIS for this project was also useful for siting  of best management practices
(BMP). The BMPs that were selected  were sediment retention basins.

       Bottom-up approaches  for  assessing  point  and nonpoint source pollution  are  also
important and  should  be examined, since development of watershed mitigation plans requires
knowledge of the stressor(s) impairing water quality. Ambient toxicity tests allow us to identify
that there is a problem, but do not specify exactly what that problem is.  Toxicity identification
evaluations could  help solve that limitation. In situ  biological evaluations, such as rapid

biological assessments,  show population  changes for species that are sensitive to specific
stressors.  These are  techniques that need further research,  but they are promising tools for
water quality impairment diagnosis.


Q:     How much did the remote sensing and GIS work for the project cost?
A:     They cost less  than $150,000.

Q:     Why have no management plans been completed for the watershed if all of the data to
       do them is available?
A:     The city of Dallas has not yet decided to commit the monetary resources to the public
       action/public involvement activities that would be required to complete such plans. The
       city, in fact, is a private utility where the reservoirs are concerned.  As such, it has no
       regulatory authority in the reservoir watersheds.  It must rely on extension services and
       cooperation from private landowners.

Q:     How long must you monitor a water  body in order to see the changes in water quality
       that are caused by the changes in land use shown by a GIS?
A:     GIS can give you excellent information,  when you are looking at long term trends in land
       use -- over years, or even seasonally for one year. It is possible to get satellite imagery
       for a watershed every sixteen days, but that may be overkill.

Q:     We often  see  pollutant  loading models where the fundamental driver of pollution is
       people, and the population is not included in the models, the analysis, or the presentation
       of the data. We use models, such as the Universal Soil Loss Equation, that pre-date the
       kind of population densities that we have today.  Why would you not favor GIS, for
       example, that can take into account the available data on population densities and develop
       models that relate to the true cause of many water quality problems?
A:     I agree.  This  issue is important and must be addressed.  Census data could easily be
       overlaid in a GIS system to better define who is creating the largest pollution loadings.
       That depends on the scope of your study, however.

                        Diagnosis and Assessment Tools/Models
                           for Determining Nutrient TMDLs

                                  Norbert Jaworski
                                   U.S. EPA - ERL
                             Narragansett, Rhode Island

       The focus of this presentation  is on techniques for estimating nutrient loadings, export
fluxes, and balances for  large watersheds.   This presentation is a  summary of four papers
currently being published which utilized the numerous data bases of the upper Potomac River
basin watersheds.  The objectives are  to:

1.     describe the monthly and annual variability of nitrogen loading in the upper Potomac
       River basin, and the possible impact of this variability on implementing and evaluating

2.     determine the input and output flux rates of nitrogen  and phosphorus for various  land
       uses within the watershed and  to compare these  rates to other coastal watersheds;

3.     determine various  methods  for establishing nitrogen mass loadings into coastal waters;

4.     review  input/output  mass balances of nitrogen and phosphorus in the upper Potomac
       River basin watershed;

5.     begin to determine the uncertainties in the mass  loading and balance estimates;

6.     examine how hydrology acts as a moderator of nitrogen processing within the watershed;

7.     suggest how science may add to the workshop agenda.

Study  Area

       The study area is  mainly the upper Potomac River basin (Figure 1).  This basin  was
selected for several reasons.  It is  the second largest watershed in the Chesapeake Bay system
and will be included in the December 1991 re-evaluation of nutrient  removal requirements for
the Chesapeake Bay.  In addition,  the Chesapeake Bay Watershed Model (Linker gj al., 1991)
is operational.   Finally,  numerous nitrogen management studies have been conducted in the
region  so that  there is considerable data on nutrient  sources, including annual basin nutrient
loadings  which were  estimated by the Metropolitan  Washington Council  of Governments
(MWCOG).  A detailed  nutrient inventory for the Potomac River basin was completed by J.
Lugbill of the MWCOG  in 1990.

Figure 1. The Potomac River basin

       Numerous other studies have been conducted on the nutrient fluxes of the Potomac River
basin.  Fisher and Oppenheimer estimated nitrogen input sources for the Chesapeake Bay in
1991 by two methods (Table  1).  Of the 628 x  106 kg N yr1 generated in  the basin, their
analysis suggested that 77% is retained in the watershed.  Recent estimates of atmospheric
nitrogen deposition have suggested similar rates.

Table  1.  Estimated nitrogen input sources for the Chesapeake Bay  (Fisher and Oppenheimer,
Point Sources
Atmospheric Nitrogen
Atmospheric Ammonium
Animal Waste
       A study by Jaworkski gt al. (1991) determined mass balances for nitrogen and phosphorus
for the upper Potomac River  basin.  A parallel study, conducted by Jaworkski and Linker in
1991, suggested that  an  "Input-Output Analysis Matrix" could be used to gain insight into
nutrient sources and sinks on a  watershed basis (Table 2).  Further study by Groffman and
Jaworski developed a linear relationship between nitrogen input and output fluxes for runoff and
storage for various land use types.

Brief Summary of Analysis Conducted

       Expanding on an nitrogen  balance for the Chesapeake Bay by the Environmental Defense
Fund (Fisher el aj.,  1988) and the MWCOG nutrient inventory for the Potomac River basin
Lugbill, 1990), nutrient mass  balances for the Potomac River basin were determined (Jaworski
Si al, 1991 and Jaworski and Linker,  1991).  Groffman and Jaworski (1991) continued the
analysis to determine the  nitrogen input and output flux rates (kg ha'1 yr1) for various land uses
within the watershed,  and compared these rates to the coastal watersheds.

       The high percentage of nitrogen (77%) estimated by Fisher and Oppenheimer (1991) in
the Chesapeake Bay watershed and the 65%  of nitrogen estimated by Jaworski, et al. for the
Potomac River basin to be either volatilized, denitrified, or stored in soil, biomass or ground
waters, suggest that nitrogen  flow or flux across the various ecotones within  the basin  is the
dominant factor that needs to  be  defined and quantified

       To the author,  the possible relationship between input and output flux rates suggests that
there was great potential to  further develop  this  relationship utilizing  the ecotone concept
suggested by Holland  gl al. (1991).

Table 2. The upper Potomac River basin, nitrogen mass balance calculations from 1983 to 1988
(kg x lOVyr).
Lost to
Edge of
River System Input
River System Loss

       An analysis of the upper Potomac River basin and four watersheds in the Georgia coastal
plain suggests that:

1.      Highest nitrogen export occurs during periods of high surface water flows.

2.      Monthly export was at a low of 0.01 kg month"1  for August 1966 with a high of 12.07
       kg month"1 for November 1985.

3.      Average annual nitrogen export form the basin varies over a factor of three from 10.6
       to 33.2 kg x 106 yrl.

4.      Export of nitrogen from the watershed  is coupled closely to the hydrologic cycle.

5.      Similarity of annual water yield and nitrogen yield time-series relationships indicate that
       processes which govern water yield also over nitrogen yields.

6.      Of the 143.5 x 106 kg yr1 total nitrogen input, about 42.6% came from  animal  waste.

7.      About 60% of the 28.9 x 106 kg yr1 of total phosphorus input came from animal waste.

8.      River exports of nitrogen and phosphorus were 17.2% and 7.2%,  respectively, of the
       total nitrogen and phosphorus inputs.

9.     About 94.2 x 106 kg  yr1 or 65.6% of the  nitrogen  inputs were retained, lost to the
       atmosphere or stored within the watershed in either biomass or groundwater.

10.    Of the 28.9 kg x  106 kg yr'1 total phosphorous inputs,  65.7% were retained in either the
       soil, biomass, or ground waters.

11.    The greatest  uncertainty in the nitrogen balance of the upper Potomac River basin  is
       quantifying the amount of nitrogen lost back to  the  atmosphere or change-in-storage
       including the amount going to ground water.

12.    For agricultural land use in the upper Potomac River basin, analysis suggests that  there
       is significant annual nitrogen loading of ground water.

13.    Current  BMPs do not strongly affect nitrogen storage mechanisms within agricultural

14.    Development of innovative within-field, as well as off site, BMPs are important.

15.    "First Principles" of science,  hydrology and engineering can be applied  to watershed
       pollutant management.  There is hope!!!


Fisher, D., J. Ceraso. T. Mathew and M. Oppenheimer. 1988.  Polluted Coastal Waters, the Role of Acid Rain.
Environmental Defense fund, Washington, DC.

Fisher, D.C. and M. Oppenheimer.  1991.  Atmospheric Nitrogen Deposition and  the
Chesapeake Bay Estuary. AMBIO.  No. 3-4.

Groffman, P.M. and N.A. Jaworski. 1991. Watershed Nitrogen Management:  Upper Potomac River Basin Case
Study.  Chesapeake Bay Res. Conference Proceedings.

Holland, M.M., P.G. Risser, and R.J.  Naiman.  1991.  Ecotones:  The role of Landscape
Boundaries in the Management and Restoration of Changing Environmental. U.S. Man and Biosphere Program and
the Ecol.  Soc. of Am., Washington, DC.

Jaworski, N.A., P.M. Groffman, A. Keller, and J.C. Prager (in revision, 1991).  A Watershed Scale Analysis of
Nitrogen Loading:  The Upper Potomac River.  Estuaries. Estuarine Research Federation.

Jaworski, N.A. and L. Linker, 1991. Uncertainties in Nitrgoen Mass Loadings in Coastal Watershed.  Chesapeake
Bay Research condreence Proceedings.

Jaworski, N.A.  1991.  The Application of the Ecotone Concept in Defining Nutrient Managment Requirements
for the Upper Potomac River Basin. Presented AMAB/UNESCO, IHP Workshop, Mikolajke, Poland, May 20-

Lugbill, J.  1990.  Potomac River Basin Inventory. Metropolitan Council of Government Washington, DC.

                             WATERSHED MODELING
                     Watershed Modeling and TMDL Assessments

                               Anthony S. Donigian, Jr.
                              AQUA TERRA Consultants
                              Mountain View, California

       This presentation will focus on the potential use of watershed modeling techniques for
Total Maximum Daily Load (TMDL) assessments.  A general overview of watershed modeling
will be provided, discussing both the nonpoint loading function and in-stream process components
included in watershed models.  The steps involved in the overall modeling process will be
presented to demonstrate the various tasks of model input development, calibration, validation,
and model  testing.  Available watershed models will be described through summary  tables of
model  capabilities and limitations, with emphasis on the deterministic simulation of watershed
(land surface and in-stream) processes at various levels of detail and complexity.  Nonpoint load
simulation  procedures appropriate for different land use categories (e.g., agricultural cropland,
urban,  forest) will be discussed in conjunction with in-stream processes that affect that fate and
transport of sediment, nutrients, and toxic contaminants.

       The potential use of watershed modeling techniques to satisfy the assessment needs of the
TMDL process will be explored in terms of the type of information generated by these models,
the issues involved in selection of appropriate models, and the relationship of user and data needs
for effective model application.  Research, data, and model development needs will be identified
to help focus program resources and objectives in areas that could most efficiently benefit future
model  use  in TMDL assessments.


Q:     Is uncertainty included in the model?
A:     No, it  is not.  Although, in the last few years  there have been more efforts  to use
       combined techniques, such as Monte Carlo simulation  along with deterministic models.

Q:     How much effort has there been to link  fate and transport models with the ecological
       effects  of pollutant  loadings?
A:     There is an EcoRisk  Research Program at  EPA ERL  in Athens, Georgia  that has
       developed a model called PIRANHA.  PIRANHA links pollutant loading,  fate and
       transport, and  food chain bioaccumulation in a geographic information system (GIS)
       context. It is in version 1.0 now,  but  it is not yet publically released.

Q:     When we attempt to model a watershed system, many stream miles are omitted in order
       to decrease the complexity (and hence the computing time) of the model.  There is a
       penalty for this simplification.   Has any analysis been done on what you gain when
       adding  complexity by  adding  higher order streams for better  resolution  within  a
A:     When you use  smaller scales  it is not clear that you are getting more detail;  the
       discretization of your data is smaller.  Although you may not lose detail when  modeling
       on a large scale, there are other factors, such as computing time that must be considered.
       When the Chesapeake Bay was modeled on the order of 68,000 square miles for detailed
       soil  and  in-stream transport processes,  modeling capabilities  were dictated  by  the
       feasibility of doing it at that scale.

Q/C:   It is important for modelers to clearly and concisely articulate the assumptions that are
       used  to develop the models.
A:     Extremely thick manuals that include all of the model assumptions are provided for many
       of the models I mentioned.  Unfortunately, many users do not read through the manuals.
       If they did, the assumptions -- sometimes pages and pages of them — would be evident.
       If one focuses in on a particular model application most of the time the information about
       important  assumptions is there;  sometimes -- many times - the information  is not as
       complete as it should be.  For the comprehensive document on TMDL assessments that
       will be developed, it makes sense to make sure that the assumptions used in the various
       analyses are laid out clearly.
                 Water Quality Diagnosis and Assessment Tools/Models

                                  Eugene D. Driscoll
                             Woodward-Clyde Consultants
                                 Oakland, New Jersey

       This presentation will provide an overview of screening methods that can be applied to
characterize nonpoint source pollutant discharges (primarily urban stormwater runoff) and their
receiving water impacts, and  will also compare point source and nonpoint source loads and

     •  In many cases,  data limitations, system complexity, and budget and time constraints
combine to preclude any meaningful load/impact predictions, which are usually the objective of
formal complex models. The objective of analysis in such cases is limited to diagnosis of water
quality and nonpoint source impacts based on the relatively limited information that is usually
available.  The diagnosis is then used to  (a) support an assessment of the relative significance
of different  sources, (b) guide decisions for management  plans,  and (c) identify  the  most
appropriate focus  for continuing monitoring  activities.

       This presentation will discuss the general nature of the simplified screening methods, the
types and availability of required input data, and the types of output results that can be
generated. An example of how output has been used to assess the relative significance of point

source versus nonpoint source pollutant  load  impacts  on water quality will  be presented.
Approaches for addressing the TMDL issue, considering both point and nonpoint sources, will
also be discussed.

       An essential aspect of screening, or simplified analysis approaches is the time scale on
which they are based. Normally, an average value is used for both the input parameters and the
projected outputs.  For example:

•      Rainfall is  converted  to runoff using an average runoff coefficient (Rv).  The  value
       assigned  for a catchment reflects the net effect of  the deterministic  elements  (soil
       moisture and permeability, slope, depression storage, storm size), and can be estimated
       from available rainfall-runoff monitoring data, deterministic model runs, or the general
       percent impervious relationships that have been developed.

•      Pollutant concentrations vary appreciably from event-to-event.  A  statistical analysis of
       available data is used  to determine an appropriate "average" value (for a site, a group
       of sites, or a land use category), which can then be combined with runoff volumes  to
       determine  pollutant loads.   Comparisons with  larger  data  bases,  organized  in  a
       comparable statistical  manner, are used to improve the level of confidence in runoff
       concentration estimates based on limited monitoring data.

•      Receiving water impacts are usually examined in either of two ways.  For long-residence
       water bodies, such  as lakes or bays,  the  operative parameter is usually cumulative
       loadings.   Also, useful qualitative comparisons  between point source  and nonpoint
       source pollutant sources can often be made using comparative mass loadings.  Where
       intermittent concentrations, rather than longer term mass loads are important, as in the
       case of assessing stream impacts, the statistical distribution of the runoff concentrations
       and flows are used to project the probability distribution of in-stream  concentrations
       during storm events. Available models for performing this type of analysis include  direct
       analytical procedures and Monte Carlo methods.  In-stream storm concentrations can be
       expressed as return periods  for concentrations that exceed applicable  criteria.   The
       probabilistic  models provide a computational basis for combining  point source  and
       nonpoint source  impacts.

       Analysis can often effectively employ combinations that involve complex models and one
or more of the simpler screening analysis  elements identified above.  A recent study for  Santa
Clara County, California, illustrates this.  The program involved extensive monitoring, and the
development of model projections of nonpoint source loads to  South' San Francisco Bay, for
comparison with point source loads from publicly owned treatment works (POTWs).  The EPA
Storm Water Management Model (SWMM) was used to develop the runoff hydrology using rain
data from a fairly dense rain gage network, so that significant orographic differences across the
country could be properly reflected.  Runoff quality model projections were based on the mean
runoff concentration of the runoff quality.  Model calibration was based on a  comparison of
projected total seasonal loads with loads developed from observed stream flow and concentration

       A brief identification of some of the issues associated with the application of a TMDL
concept to intermittent and highly variable stormwater discharge loadings, will be presented to
serve as a reference for workshop discussions. One of the more important issues that should be
addressed in the workgroup breakout sessions is the need to develop a  consistent and useful
interpretation of the legislative term "Total  Maximum Daily Load", as it will be applied to
nonpoint source discharges.  Some of the difficulties associated  with a literal  interpretation
include the following:

•      MAXIMUM literally means "never, ever more than", which does not relate well to the
       substantial, inherent variability of rainfall driven nonpoint source discharges. We should
       perhaps consider interpreting "maximum" in terms of an appropriate return period, such
       as the "three years on average" that has been adopted for aquatic life toxic criteria for

•      DAILY LOAD applies reasonably well to POTW and  industrial waste discharges which
       are usually characterized on a daily average basis. For nonpoint source discharges the
       operative time scale is that for a storm "event," whose average duration  in the United
       States is on the order of six to twelve hours. Individual event durations range between
       one hour and 20 or 30 hours.

•      For lakes or bays, daily or event loads are less significant in  terms of  water quality
       impacts, than cumulative loads over an extended period.  Annual loads from nonpoint
       source differ markedly  in wet and dry years, so that the concept of a return period is
       appropriate to consider,  even in the case of annual loads.

       TMDL is a meaningful concept for characterizing the total acceptable load to a water
body segment, such that water quality objectives will not be violated under some selected design
condition (e.g., stream flow and temperature).  It relates well to continuous point sources and
provides a framework for allocating present and future allowable loadings among a number of
individual dischargers.  The basic concept can, at least theoretically, be extended to address the
regulation of pollutant load distributions between point source and nonpoint source discharges.
However, if the concept is to be applied in a useful and meaningful way, it will be necessary
to interpret TMDL in a manner that effectively reflects the important differences in the nature
of point and nonpoint source pollutant discharges.


Q:     Were the pollutant concentrations in the runoff that were reported as a mean value flow-
       weighted or time-weighted?
A:     The event concentrations were flow-weighted means.  The data that was developed for
       this example was either based on flow-in composite samples, or where it was a series of
       discrete individual samples they were integrated and then the event mean was estimated.

Q/C:   One  important advantage to working  with empirical data, where  it is available, is that
       you get a physical sense for how systems respond based on the data. It is an excellent
       reality check if you are involved with other complex or detailed modeling.

                      Model Capabilities (Agricultural and Urban)

                                 Leslie L. Shoemaker
                                   Tetra Tech, Inc.
                                   Fairfax, Virginia

       This presentation will discuss the suitability of available models for implementation of
the TMDL process in both agricultural and urban areas.   The TMDL process will first  be
described to highlight data requirements and the implementation process. Based on an evaluation
of the data requirements, the need for screening models are will be described based on their
ability  to provide screening-level analysis for the TMDL process.  A range of loading rate
models  will be evaluated,  including annual  load prediction techniques, spreadsheet models,
constant concentration models, and screening  models  that are used by municipalities.  Typical
applications of loading rate models will be briefly outlined.  The limitations of loading rate
models  will be evaluated and compared with the requirements for verification, accuracy,  ease
of use, and assessment of management effects.

       Other options for screening-level analysis will be described,  including "design storm"
modeling, extrapolation from detailed models to develop long-term characteristics, simplified
continuous simulation modeling,  and statistical techniques.   Example  model results will  be
presented to quantify the benefits of various model types.  A discussion of the requirements for
verification, accuracy, ease of use, and assessment of management effects will be included for
each option.

       The use of more detailed  models,  such as HSPF (Hydrological Simulation Program -
Fortran), SWMM (Storm Water Management Model), and CREAMS (Chemicals, Runoff, and
Erosion from Agricultural Management Systems) will be briefly discussed.  In  particular, the
application criteria will be evaluated in relation to the  associated level of effort.  The cost
benefits of model application will be described through examples.  The benefit of using detailed
models  to predict in-stream pollutant concentrations on a continuous basis will be discussed.
The associated costs for calibration  and verification will  be contrasted with the  benefits of
detailed simulation data.


Q:     By going into more  detail within a  sub-basin, or going into a  sub-sub-basin  to the third
       and fourth order streams,  do we really gain by it in terms of our predictive capabilities?
A:     You  gain predictive capability as you move to a smaller scale.  You can  do a more
       thorough  job  so far as modeling is concerned, allowing you to look at potential best
       management practices or management scenarios with a higher degree of accuracy. Going
       into a fourth order stream you can actually site a detention facility, for example, look at
       what is happening in that facility, and  determine its effects on the watershed  as a whole.
       That is good, but on the  other hand,  you have tremendously increased  the amount of
       expended effort. You may not need or want to go into that much detail, depending on
       your goal.  When modeling a very large area  — a basin such was the Chesapeake, for
       example -- a tiered approach should be taken.  Determine where your hot-spots are using

       a screening-level model; then, use a more detailed or complex model for problem areas.
       This is particularly feasible if a CIS is available.
                          Model Capabilities - A User Focus

                        Paul L. Freedman and David W. Diiks
                                  Limno-Tech, Inc.
                                Ann Arbor, Michigan

       From  a  practical perspective, the  development of a TMDL and a National Pollutant
Discharge Elimination System (NPDES) permit, and the calculation of noripoint watershed loads
are closely intertwined.  Regulatory  guidance states that the TMDL is equal to the sum of the
waste load allocation (WLA) and load allocation or nonpoint allocation (LA).  Hence the WLA
(and ultimate NPDES permit) cannot be developed consistent with the TMDL unless the LA is
defined.  This presentation will review practical considerations in developing LAs as they relate
to the TMDL and NPDES processes.

       From  a regulatory perspective, state staff typically need quick and simple approaches for
calculating TMDLs and LAs for the large number of systems they regulate.  Data, time, and
resources are almost always limited.  So far in the TMDL/LA process, we have seen  either zero
assumed  for  the LA,  or  the LA contribution has  been  set based  on  monitored in-stream
concentrations.  The Tualatin River  and Dillon Reservoir are two examples where data were
used (case presentation to follow). In another case, for the Columbia River dioxin TMDL, the
LA was  set after allocating the WLA to individual NPDES discharges.  All told,  the typical
current regulatory  approach  to  date has been  simplistic,  provides no  information on  the
attainability of target LAs, does not assess the significance of changes in basin land use, and is
typically  constrained by a shortage of data.

       At the other extreme from these simple approaches,  rigorous  predictive approaches are
available to calculate watershed  nonpoint loads and  integrate them  into the TMDL process.
Research in the Great Lakes and Chesapeake Bay Basins are two examples. The Great Lakes
studies probably demonstrate our earliest experience with TMDL type approaches.

       International Joint Commission target loads were established in the  1970s to meet water
quality objectives.  Initial efforts focused on wastewater effluent restrictions.  Since that time,
rigorous  modeling has been conducted in portions  of the Great Lakes, such as Saginaw Bay, to
evaluate  the importance of nonpoint contributions  and the tradeoff with additional point source
reductions. The Saginaw Bay basin case  study exemplifies how  the use of intensive edge-of-
field, minor drain and major tributary monitoring were combined with watershed modeling tools
(ANSWERS  [Areal Nonpoint Source Watershed Environment Response Simulation], and unit
area loading) to  calculate nonpoint loads  under different  scenarios.   In this  case,  cost
effectiveness  analysis was conducted to evaluate tradeoffs among LAs and WLAs. The benefit
and reliability of this approach is well proven, but the costs, skill level, and time commitment
are prohibitive for  typical regulatory use.

       If the state of the art in watershed tools for TMDL/LA use is examined, it is evident that
powerful tools are available, but involve heavy staff, schedule, and data burdens. Limited staff,
schedule,  and  data,  however, often  require regulatory staff to  seek out the most basic
approaches, which have little predictive capability.  There is, therefore, a need for easy-to-use
tools that link models with readily available data bases.

The key concepts in striving to meet this regulatory need are exemplified by these phrases:

       1)      "easy to use;"
       2)      "linked models and data;"  and
       3)      "readily available data."

Fortunately over the last  decade  we have  seen  substantial advances in simple,  user friendly
models, data linkages, and the use of  CIS  to better manage available data.  Case Studies are
presented  to show  the evolution  and availability of these technologies, and to highlight the
remaining need for  a simple watershed/water quality model that is linked to CIS data bases.  A
newly developed prototype is demonstrated.

       The presentation is closed with a brief overview of available  technology, gaps, and
regulatory  needs.  Key policy issues are  also outlined, relating to design flow and episodic or
cumulative effects which complicates the TMDL/LA analysis.


Q/C:   The Wallon Lake  Study involved looking  at forested  land  and residential land uses.
       Determining cover for such  land use practices is more simple than for agricultural uses
       where crop types and tillage practices must be considered.  The Saginaw basin, however,
       is highly agricultural, and alternative land use practices (such as fertilizer management
       and conservation tillage) were evaluated using a more sophisticated modeling tool - the
       ANSWERS  model.

Q/C:   The "solution" to nonpoint source pollution problems included CIS and nonpoint source
       models to get to nonpoint source controls, presumably.   I would like to point out that the
       Coastal  Zone Management Act (CZMA)  Guidance has an even  simpler solution to the
       nonpoint source problem; although it appears to digress  a bit from the TMDL approach:
       "Implementation of nonpoint  source management measures has  been intentionally
       divorced from identified water  quality problems because of  the enormous difficulty of
       establishing  cause and effect linkages between land use and water quality." Specifically,
       this says, "Hey, a TMDL isn't needed.   We're going straight to the implementation
       measures and we're going to make everyone apply them."

A:     (by Bruce Newton) We must recognize that there is a definite split in this country with
       respect to the regulatory approaches used  to clean up impaired waters.  One is  the water
       quality-based approach, which is  what this workshop is focusing on.  The other is the
       technology-based approach.  For the water quality-based approach, you develop pollution
       controls in order to attain specified  water quality standards, controlling only as much

       pollution as is needed.  The technology-based approach requires that all polluters apply
       controls; whether the control is needed or not.  The CZMA amendments of last year set
       forth some technology-based approaches for the nonpoint source arena.

Q:     If a State was to take a technology-based approach  for nonpoint source controls on an
       impaired water body, and then  submit that to EPA with a monitoring element and an
       explanation that it was intended to be an iterative approach would  that be accepted as a
A:     (by Bruce Newton)  It would have to be considered on a case-by-case basis; but it would
       probably not be accepted as a TMDL unless there was sufficient evidence to show that
       water quality was  considered.   If there  was sufficient evidence then  it  would be
       considered a TMDL under the phased approach.
           North Carolina's Basinwide Water Quality Management Approach
                                to Developing TMDLs

                                    Ruth Swanek
                      NC Division of Environmental Management
                               Raleigh, North Carolina

       The North Carolina Division of Environmental Management (NCDEM) has determined
that the requirements of Section 303(d) of the Clean Water Act will be fulfilled by the Division's
ongoing basinwide  water quality  management  initiative.  NCDEM's initiative involves the
development of management plans that address both point and nonpoint source pollution for each
of the seventeen major river basins in the State. These management plans will include strategies
for  controlling targeted pollutants based  upon the integration  of  chemical and biological
monitoring, modeling, state and federal rules and regulations, and best professional judgment.
Where appropriate,  TMDLs will be comprised of management practices or strategies and will
not necessarily include specific numeric loads.

       Consistent and effective management within a given river basin requires consideration
of pollutant sources and their  interaction from the headwaters of the basin to its mouth.
However, analyzing water quality on a basinwide scale poses a formidable challenge, particularly
for model development.  North Carolina surface waters cover more than 37,000 miles and are
comprised of a wide diversity of systems (i.e., streams, rivers, lakes, estuaries, wetlands). These
waterbodies are placed in a variety of geologic settings with multiple designated uses.   More
than  3000  NPDES  discharges exist  across the State, along with  several hundred ambient
monitoring  stations.  Aggregating the tremendous amounts of data for each basin into useful
analytical frameworks is, therefore, no small task.

       Increased access to sophisticated computer hardware and software (i.e. high speed/large
memory personal computers  and work  stations,  GIS  technology, and  relational  database
management software) along with the establishment of a well balanced and experienced Water
Quality program  staff, puts NCDEM in the position  to  truly  implement  Clean Water Act.
mandates for whole  basin management (i.e., Sections 303(d) and 303(e)).  Even with this well

developed infrastructure,  management tools  must remain  at  a practical level.   Resource
constraints prohibit the development of highly complex, mechanistic models for all surface
waters in the State; therefore, NCDEM has chosen to combine a moderate modeling approach
with an aggressive field monitoring approach to  develop and monitor the performance of its
TMDL strategies.

       For conventional wastes (e.g., BOD, NH3), the Division will focus on calibrating large-
scale QUAL2E and WASP (Water Quality Analysis Simulation Program) models along the main
stream of the basin and on tributaries with significant pollutant loads where the State's empirical
desk top model does not simulate the system accurately.  The focus of these models will be on
point source management with nonpoint source mechanisms  incorporated into background and
runoff parameters.

       Nutrients are being addressed through the use of nutrient budgets, empirical models, and
mechanistic models.  Both point and nonpoint sources are included in the analyses.  Modeling
analyses are  restricted  to watersheds that are found or suspected to be  sensitive to excessive
nutrient loading.

       NCDEM currently  limits toxics  using a simple mass  balance technique  in which
interaction among various point and nonpoint sources is usually not considered.  Since it is not
practical to perform complicated  mechanistic toxics modeling (e.g. TOXIWASP, HSPF) for an
entire basin, the State is developing a CIS-based mass balance model which accounts for all point
sources of a given toxicant within the entire basin. Nonpoint  source interaction is accounted for
in background and runoff concentration assumptions. More complex toxics models are developed
on a case by case basis where  local problems are  sufficient to warrant the expenditure of
substantial resources associated with this type of model development.

       CIS appears to be the key to successful basinwide data analysis. The GIS system allows
staff to readily aggregate information on pollutant loading, stream flow, stream hydrography,
stream classification, land use, and chemical and biological monitoring. This tool enhances staff
ability to assess  water  quality trends, prioritize areas of concern, determine  model input, and
display model results within the basin setting.

       While NCDEM's modeling  capabilities continue to improve, many issues remain to be
addressed. The State's technical needs include how to establish linkages between various model
types and GIS, how point and nonpoint sources should be addressed (since critical flows for
point sources occur during low flow conditions and critical flows for nonpoint sources typically
occur during  high flow conditions), and how to evaluate toxics data when most of the values are
less than the detection level.  Although these research needs remain, the State is moving forward
with its basinwide management approach  to developing TMDLs.


Q:     How will North Carolina incorporate biological data  into the modeling?
A:     Biological data  will be used in the GIS as an overlay to  determine where the problem
       areas  are.

                       Model Integration and Data Access Tools

                            Robert B. Ambrose, Jr., P.E.
                                  U.S. EPA - ERL
                                  Athens, Georgia

       During the  past twenty years, EPA has tackled the cleanup  of streams, lakes, and
estuaries through WLA of point source discharges of pollutants.  EPA's agenda for the future
concludes  that nonpoint source pollution is a  major contributor to remaining water quality
problems.   Urban  and combined sewer  overflow  issues  are now  urgent because of recent

       The Office  of Research and Development (ORD)  in  the past  has conducted research
efforts in developing models that can be  used on watershed or basin  scale problems.  These
models, notably HSPF, integrate the hydrology, chemical, and land management factors into
frameworks for developing exposure and water quality management strategies.  These research
efforts were concentrated during the middle to late 1970s. Nonpoint source research during the
1980s  has been limited mainly to model support and assistance in application.   Very little
research has been conducted to improve modeling capabilities or to take advantage of the ease
of use of modern personal computers and work  stations.

       A set of water quality and nonpoint source loading models are  available and supported
by the Center for  Exposure Assessment Modeling.  These include the water quality models
QUAL2E, WASP4, and HSPF, the nonpoint source models HSPF and PRZM (Pesticide Root
Zone  Model), the mixing  zone expert system CORNMIX,  and the probabilistic dilution model
DYNTOX. Other  specialized hydrodynamic and sediment transport  models are also available.
DBAPE, a national soils database and parameter estimation program, is available for rural
nonpoint source analysis.   These  models  and databases are distributed and  user assistance is
provided.  Training courses are provided with resources from the Office of Water.  •

       In addition  to these EPA-supported models, other models and techniques  have been
developed by Federal agencies such as United  States Department of  Agriculture (USDA) or
universities, and are used for particular tasks such as designing storm  series or estimating soil
erosion. For urban runoff, spreadsheets and regressions using National Urban Runoff Program
(NURP) data and United States Geologic Survey (USGS) data have been developed and are in
partial use.  For agricultural  runoff, the field  scale USDA model GLEAMS (Groundwater
Loading Effects of Agricultural Management Systems) is in use, and different basin-scale models
such as the SWRRB (Simulator for Water Resources in Rural Basins), AGNPS (Agricultural
Nonpoint Source Pollution), and ANSWERS may be promising.

       In the next five years, we need to make the transition from the existing WLA framework
to the more comprehensive TMDL framework, which integrates WLA  point source issues with
nonpoint source load allocation concerns.  To accomplish this, we need more comprehensive
models and databases that are easier to use for wider application.

       ORD should begin now to develop a  generation  of computational models that can
accurately represent the interactions of land and aquatic ecosystems within landscape mosaics.


These models should be capable of accepting land use patterns and landscape configurations that
occur in actual watersheds,  rather than relying on large-scale aggregation which neglects the
small-scale patch interaction important in real landscapes.  Further, these models must be easy
to use.

       Research needs can be divided into model integration, database development, watershed
processes (including both loadings  from land uses and in-stream effects of nonpoint source
loading), and groundwater/surface water interactions. These are briefly discussed below. A five
year program is needed with stable funding to complete this next generation TMDL modeling

Model Integration

       Several  existing models,  databases, and  computational techniques  address different
watershed issues.  The best  of these existing models should be modified and linked to address
EPAs watershed issues.  Software is needed to integrate improved models and databases with
modern pre- and post-coprocessors for ease of use.  The resulting products should incorporate
point and nonpoint source models, low and high flow  conditions, perennial and intermittent
loading, near field and far field effects, and chemical concentrations with biological response.

Database Development

       A limiting step in the application of current models is often the availability of data in the
proper form. Comprehensive databases of physical, chemical, and biological parameters must
be developed and made available.  Spatially variable data must be incorporated into a CIS that
is linked to the models.   Such data includes soil properties,  land  use,  vegetative cover,
predominant habitats and species, population density, meteorological and climatological patterns,
stream flow, drainage areas, and urban characteristics.

Watershed Processes

       More accurate models of land use and subsequent pollutant loading, and resulting in-
stream  effects  of  these loadings are needed  for watersheds  containing streams, lakes, and
estuaries.  To better estimate non-urban and mixed land use loadings, research is  needed on
BMP effectiveness. Such  as the effects of buffer strips on sediment, nutrient, and agrochemical
runoff.  The incorporation of channel erosion into watershed loading models is necessary, as are
improvements  in  algorithms that handle  runoff,  infiltration, and evapotranspiration.  For
urban/combined sewer overflow  (CSO) loadings,  screening estimates  would use  regression
equations based upon NURP and additional USGS data bases, where relationships are developed
with region, land use, dry weather deposition, climatological parameters. A data base for urban
runoff controls needs to be made available to the  user.  For detailed evaluation and design
purposes,  SWMM and  HSPF  need  to  be  enhanced with  more  physically-based  quality
procedures, and linked to the proper databases for ease of use.  Scour and deposition of sediment
and  associated pollutants on  the land surface and in the  sewer system  should be better

       In-stream effects  of watershed  loadings includes the ability to predict the effects of
sediment loading on habitat alteration, as well as the fate of persistent chemicals such as metals,
dioxin, and PCBs (polychlorinated biphenyls). The interplay between surface water and benthic
nutrients, the growth of benthic macrophytes,  and the resulting water quality and habitat
perturbations is a significant research need. The incorporation of benthic submodels, including
sediment oxygen demand, into water  quality models will give  the ability to predict stream
recovery following episodic loading events (such as CSOs) or upgrades in treatment of nonpoint
source controls.

Groundwater-Surface Water Interactions

       Chemicals that infiltrate may threaten groundwater resources directly, and surface water
resources through base flow. Linkage between existing groundwater quality and surface water
quality models must be developed and integrated with proper data bases.


Q:     You  seemed  to  indicate that you are against  tightly integrated models to  simulate
       processes that affect  watershed water quality.  Since a watershed is a tightly integrated
       system, why wouldn't you try to model its processes  in an integrated fashion?
A:     Yes,  instead of one main model that a user would have to learn entirely,  I would prefer
       to have three or four separate models in series.  There might be one person in a State
       agency who understands runoff processes.  He would be able to  use one model  to
       produce files of loadings to a stream.  The State agency might have a different person
       who  knows receiving water processes; he could take and use the loading files without
       understanding all  of  the details  of the runoff model.  He only needs to  understand his
       receiving water model.  There  are, however, some benefits from doing an integrated
       assessment automatically.  It is  more efficient, for example.

Q:     Does the Center in Athens have funds for technical support to people within the Agency?
A:     Yes.  The Center  does have funds for technical support that is available to anybody, not
      just Agency personnel.

Q:     Is SYMTOX available through the  Center?
A:     It was recently updated and we  are becoming familiar with  is in Athens.  It is not in a
       standard distribution yet; but,  it is a simple model and will not be hard to distribute soon.

            Multistage Remote Sensing Data Applications for GIS Data Base
               Development in Support of Nonpoint Watershed Modeling

                                   Ross S. Lunetta
                                U.S. EPA, EMSL-LV
                                 Las Vegas, Nevada

      With the proliferation of geographic information systems  (GIS) in both academia and
government, there has been a tremendous increase in demand for remote sensing as a data input
source to spatial data base development for nonpoint watershed modeling applications.  Products
derived from remote sensing are particularly attractive for GIS data base development because
they can provide cost-effective, large area coverage in a digital format that can be put directly
into a GIS.  Because  remote sensing data is collected in a raster format, the data can be cost-
effectively converted to a vector or quadtree format for subsequent spatial watershed  modeling

      The approach of using multiple sensors to create spatial data bases with variable levels
of spatial and categorical resolution is currently being applied to develop  cost-effective land
cover/land use (LC/LU) data bases for large area watersheds. Remote sensing satellite platform
sensor combinations  of particular  importance  for watershed modeling application  include
LANDS AT Thematic Mapper (TM) for land cover  mapping, and SPOT panchromatic data for
land use  delineation.   Resolution enhancements can be added  in a multistage approach over
"selected" areas using higher resolution aircraft scanner data and/or aerial photography to meet
specific model data input requirements.

      Satellite  and aircraft multispectral data  will be presented to demonstrate the attributes of
each data type in terms of spatial and spectral data content.  Application(s) of each data type will
also be addressed.  An overview of the satellite data processing for the 68,000 square mile
Chesapeake  Bay watershed data base  construction project in support of the Agency's EMAP
(Environmental  Monitoring Assessment Program)-LC program and the Chesapeake Bay Program
Office, will be provided. An application of multistage satellite and multiple aircraft sensor data
for ecological process studies  being conducted at Green Bay, Wisconsin and at Saginaw Bay,
Missouri, will be presented in closing.


Q:    Remote  sensing and GIS currently provide  a static display of data. Is it possible to
      animate  the display to show  change over time?
A:    Yes.  There is a lot of research going  on in that area.  Currently, are two methods are
      available. One is image differencing.  The other involves doing a classification for two
      points in time, and then comparing the  results.  The  accuracy of your data must be
      known, however, so that the  comparison is of actual changes in the landscape, not error.
      Animating the display to show change  over time is an integral part of the EMAP-LC

      program.  EMAP involves updating the landscape characterization data bases for the
      entire country on a 10-year cycle to look at change and, eventually, trends.

Q:    What is the minimum space  you would need on your computer disk to generate CIS
      displays?  Is it at all feasible  on an IBM 386?
A:    I would recommend that you use a work station.  How much data you are going to enter
      and manipulate needs to be considered, however.  That will determine how much disk
      space you need.

Q:    The Chesapeake Bay classification  system  represents a lot of work.  Stage one was
      agriculture; stage two broke  out two categories,  one of which was crops; stage three
      broke the crop category  down into ten crops.   How much effort and expense was
      involved with doing the stage three breakout?
A:    Crops are difficult because their growth stages must be captured in the imaging in order
      to do a thorough job. This means that the data must be ordered up front.  The satellites
      that we have up now  -- LANDSAT 4 and LANDSAT 5 -- are not being operated unless
      a request is made, because the sensors are past their life stage.
              CIS for Nonpoint Source Watershed Modeling Applications

                                Mason J. Hewitt, III
                                U.S. EPA, EMSL-LV
                                 Las Vegas, Nevada

      The EMSL-LV Spatial Analysis Team is charged with the application of CIS to the
mission of the EPA.  During the last two years,  the Team has been addressing the use of GIS
to support nonpoint source pollution modeling.  Two projects will be addressed that will show
the use of various models to understand the effect of management practices on sediment delivery.
The two projects are summarized below.

Arizona Rangelands Water Quality Project

      The Arizona Rangelands project demonstrates the use of geographic information system
(GIS) technology to support nonpoint source pollution modeling. In particular, the Agricultural
Nonpoint Source Pollution (AGNPS) model was used to model sediment loading and delivery
in the surface waters of the Wet Beaver  Creek watershed of north-central Arizona.  The
objectives of the project were (1) to support Arizona Department of Environmental Quality in
its watershed management practices,  and  (2) to provide the  EPA with  standard operating
procedures  for linking the ARC/INFO GIS  with AGNPS.

Blackfoot River Nonpoint Source Pollution Modeling

      CIS techniques are being used to develop the basic data layers, analytical procedures, and
graphic  user interface necessary for  watershed-level modeling of nonpoint source pollution.
Because silviculture is the predominant land use activity in the watershed, the USEPA/ESFS
Water Resources Evaluation Nonpoint Sources Silvicultural (WRENSS) model was selected to
characterize nonpoint inputs to Blackfoot River subwatersheds.

      Some concluding remarks will be offered on use of models, data preparation, quality
control,  and management perceptions.


Q:.    Can you provide some feeling for the error involved with using  satellite imagery and
      CIS, and the magnitudes of alternatives that result?
A:    There is no readily available,  quantifiable method to tell us that the data derived from
      satellite imagery are plus or minus a certain accuracy at a certain confidence level. That
      is one area that must be explored.
             GRASS Waterworks - An Interface between GIS and Models

          Yung-Tsung Kang, Jon Bartholic, Chansheng He, and Baxter Vieux
              Institute of Water Research and Center for Remote Sensing
                               East Lansing, Michigan

       GRASS Waterworks is a tool box of utilities written at Michigan State University for use
by the U.S. Department of Agriculture Soil Conservation Service (SCS) and other organizations
involved in soil and water conservation and environmental planning and assessment.  It helps
to analyze the field and watershed-scale parameters needed to model hydrologic processes that
are affected by  the agricultural management decisions concerning water quality, erosion, and
sedimentation control.  This tool box can be thought of as a preprocessor of parameter
information for  use in water quality modeling.

       GRASS  Waterworks is also an interface between GRASS and AGNPS at this stage of
development.  GRASS, the Geographical Resources Analysis Support System, was developed
by the U.S. Army Corps of Engineers Construction Engineering Research  Laboratory  (US
CERL), in Champaign, Illinois.  The GRASS version used is the SCS version.  AGNPS is a
computer simulation model developed to analyze the water quality of watershed runoff.  This
interface is developed using the UNIX shell scripts on the AT&T 6386 computer.

       The goal of GRASS Waterworks is to use GIS techniques to derive model parameters.
It requires basic map layers, such as the digital elevation model  (DEM), soils, streams, and
LC/LU (frequently derived from remote sensing), and the soils data base to provide information.
The tool box then works interactively with users to derive all of the 22 parameters needed by


 the AGNPS model.  Output from GRASS Waterworks is an organized data input file which can
 be converted into DOS data format and put directly into to the AGNPS model.

       Future  development of GRASS Waterworks will focus on incorporating additional
 parameters so that it can be use with models, such as NLEAP (Nitrate Leaching and Economic
 Analysis Package),  GLEAMS (Groundwater  Loading Effects of Agricultural Management
 Systems), EPIC (Erosion-Productivity Impact Calculator), and NPURGE.


.Q:    What size watershed can GRASS Waterworks together with AGNPS handle optimally?
 A:    From a few acres to about 50,000 acres.

 Q:    Monte Carlo simulation is used often by modelers to determine where the sensitivity lies
       in the model variables.  Now, with CIS, can similar kinds of Monte Carlo simulation be
       done with the spatial arrangement of the cells to see if, in fact, shuffling the cells makes
       a difference?
 A:    We have  done some research in that area and it is daunting.  Depending on how many
       of those cells you take  and how you calculate the slopes, spatial sensitivity of the data
       can be very high.

 Q:    Does the  AGNPS user  need to know how to use GRASS, also?
 A:    No  extra  burden would be placed on the user to learn the model.  The idea of expert
       systems and shells is to have the shells as transparent as possible.

 Q:    Why did  the Soil Conservation  Service choose to  use GRASS  instead of ARC-INFO,
       which EPA uses?
 A:    ARC-INFO cannot be used on a field office personal computer; GRASS can and it is
       easier to use.
           Remote Sensing of Agricultural Practices, and Downstream Water
               Quality as Influenced by Sediment from Nonpoint Sources

            John G. Lyon, Andrew Ward, Keith Bedford, and Terry Logan
                              The Ohio State University
                 Gary Schaal, Ohio Department of Natural Resources
                                  Columbus, Ohio

       A series of experiments funded over ten years have addressed measurements of nonpoint
 sources of sediment in watersheds and tributaries of Lake  Erie.  The goals were to implement
 a combination of on-site sampling and remote sensor technologies for measurement of causal
 factors and transport mechanisms of sediments. The studies focused on the lake bed and glacial

till soils farmed in watersheds of the Sandusky River, and techniques to measure suspended
sediments that are transported through Sandusky Bay to nearshore Lake Erie.

       A detailed study of farmed fields in Seneca County, Ohio, was conducted from 1988-
1990. Fifty-five field sampling sites were visited every two weeks from March to October. We
evaluated  fields  with corn and  soybean crops,  lake bed and. till soils, conventional and
conservation tillage practices, and farms with and without tile drainage. The field data collection
was coordinated with overpasses  of the LANDSAT satellite, to evaluate the utility of satellite
data for measuring factors influencing nonpoint source pollution. Measurements were conducted
during the drought of 1988, the normal rainfall year of  1989, and the high rainfall year of 1990.

       Results will be reported with a special emphasis on parameters important to  303(d)
activities.  Results are related to characteristics of sedimentation, crop and residue cover (LAI,
wet biomass, actual yield), and general  capabilities of remote sensor data  that can  assist in
303(d) efforts.

       A  second experiment  demonstrated remote sensing  capabilities  for  measurement of
suspended sediments  in Lake  Erie waters.  A combination of on-site sampling, multiple-date
satellite data, and results of a water quality hydrodynamic model were  used  to  follow and
quantify the transport of sediments from Sandusky Bay into nearshore Lake Erie.  We focused
on a storm period in 1981 when the bulk of sediments  from nonpoint sources were  transported
into the open lake.  Remote sensor data were used to make thematic maps of surface suspended
sediments, and were used to validate results from a water quality hydrodynamic model.

       These experiments demonstrated  many of the capabilities of remote sensor  data for
evaluations of nonpoint sources,  and the results also present the "continuum" of the problem
from source farm lands to the impacted area -- eutrophic Lake Erie.
                    The Trials, Tribulations, and Successes of Using
                         Remote Sensing, CIS, and Modeling

                                   by Carol Russell
                formerly with the Department of Environmental Quality
                                   Phoenix, Arizona

       During my tenure  at the Arizona Department of Environmental Quality (ADEQ), I
experienced a number of trials and tribulations while using remote sensing, GIS, and modeling
to evaluate the potential impacts of nonpoint source pollution.  This article briefly relates not
only my problems, but also my  ultimate success  using these methodologies.   I hope my
experience will facilitate their use by others.

       To have a common basis for discussion, I will briefly describe my team's methodology
and define terms.  In the examples that follow, remotely sensed data originated from satellite
imagery and aerial photos.  "Remotely sensed"  data is  collected using sensors  not in direct


contact with the object. A GIS was then used to interpret our remotely sensed data and compare
it accurately with other available data.  A GIS is a computer system that stores, analyzes, and
displays spatially related data.  The GIS data layers were then used as input parameters for our
modeling efforts.  Our models were designed to simulate the transport of sediment within a
watershed in order to determine the effects of nonpoint source pollution, which results from
diffuse land uses.

       In order to gain perspective on  remote sensing,  GIS, and modeling,  it  is helpful to use
the imagination.  Picture  the galaxy, then increase the resolution until you can see our solar
system. Increase the resolution more to obtain a view of the earth as seen from  the moon.  Even
more, our continent; then Arizona, the Phoenix area; Sun City; a back yard; a  pond.  Although
the scene changes, it can be argued that we are looking at the same thing in each  instance.
Nevertheless,  it is clear  that the shift  in  perspective radically affects what we perceive.
Similarly, the usefulness of remote sensing tools in the decision-making process is dependent on
perspective and scale.

Remote Sensing, or "Can the answer be seen?11

       One of our first trials, and also our first success using remotely sensed data at the ADEQ
addressed a pesticide problem encountered in the western Phoenix metropolitan area.  Painted
Rock Reservoir is located at the terminus of a watershed where the pesticide residues were so
high in the  fish  tissue samples that the entire state park was closed.   Our efforts to employ
remote sensing to determine the origin of the pesticides were unsuccessful for three reasons:
(1) We were unable to correlate  visible land  uses with pesticide use; (2) we were unable to
visualize the contaminant transport mechanisms; and (3) the project area was enormous. Simply
put, we were using a picture to provide an answer,  when the answer could not be seen.

       Although  remote  sensing  did not yield  an answer to this pesticide problem, it  is
nevertheless a valuable tool that can address other types of problems successfully. For example,
in the Maryvale area near Sun  City, there was  a  leukemia  cluster  in  children  that,  we
hypothesized,  was caused by residual pesticides  in the soil.  Land  use in the small  area had
recently changed from agricultural to residential.  Using remote  sensing,  the Arid  Lands
Research Center in Tucson mapped the  land use changes over time using aerial  photographs and
a GIS.  Remote sensing and GIS  mapping then were used to stratify soil samples in order to
make a statistically verifiable and scientifically defensible evaluation that residual agricultural
pesticides coincided with the leukemia deaths.

       Note that when one maps changes in land use, it is necessary to begin  with a base map
that can then be modified for changes each year.  Mapping each year separately and then trying
to overlay them will not  work because of technical differences  in equipment,  shot angle, lens
distortion, etc., in the images taken over the years.

       How does this  example relate to  water quality? The data  generated  on the  pesticide
residues will be combined with geographically indexed data on water quality, crops grown, soil

types, and water usage to advise farmers on a site-by-site basis.  Farms will be targeted by the
calculation of risk potential to reduce the loading of sediment and associated absorbed pesticides
that may be moving downstream to Painted Rock Reservoir.  The intent is to "keep the soil
down on the farm," so to speak.

CIS and Time

       Tribulation and success number two involved Arizona's Nonpoint Source Assessment
program. A GIS was used on a statewide basis to help determine what types of nonpoint source
pollution were causing nonattainment of water uses.  Our intent was to maximize personnel
efficiency,  because site  visits could not be made to  each watershed to determine potential
pollution sources.  The methodology seemed reasonable since types of nonpoint source s are,
essentially,  land uses. We encountered, however, many problems. The latitudes and longitudes
of facilities permitted for water discharges to surface and ground water were not accurate, and
most were not even mapped. The maps that did exist were not current. For example, the most
recent  statewide map of irrigated  land was charted  in 1969!   Furthermore, we  made  the
assumption that all lands' in Arizona outside of cities and military reservations were used for
grazing, since in the state even wilderness areas were grazed.

       In retrospect, we could have visited every watershed in  Arizona in  the  time it took to
construct the data base.  Additionally, after one year of training  to become fluent in the use of
GIS  technology, our people became a prized commodity.  Most government agencies cannot
compete with private sector salaries, and ADEQ lost people for this reason.   The only tangible
success of  this endeavor was  a beautiful set of maps  that raised management's awareness of
nonpoint source  problems in Arizona, and potential uses of GIS technology.   Unfortunately,
GISs are a  very expensive and time-consuming method to make  pretty maps.

       An indirect benefit of the project was the development of more detailed GIS data bases.
For example,  the Department of Water Resources  annually maps  agricultural land in active
management areas for water conservation purposes.  This ensures that no additional land goes
under cultivation. Additionally, wetland and riparian vegetation in special areas of concern have
been mapped with the assistance of the U.S. Fish and Wildlife Service to determine acreage loss.

Modeling and Gross Estimates

       The third example, the  Arizona Rangeland Study, attempted to determine if water quality
standards would be consistently violated if forested lands were converted  to range land.   In
Arizona, where water is a premium resource, a proposal surfaced to augment the water supply
by vegetation  manipulation -- cutting trees.  At  ADEQ, we  were concerned with the water
quality implications of this proposal.  Choosing a model, however, was difficult. We needed
to know the event-related nonpoint source pollution loading, and  no current model provided the
information accurately on a watershed level; we anticipated, however, that the  Water Erosion
Prediction Project (WEPPS) under  development by  the Soil Conservation Service, the Forest
Service, and the Bureau of Land Management would provide the needed information.  We chose

the AGNPS because of its similarity to WEPPS regarding parameters that could be derived from
CIS covers.  We also conducted sensitivity analyses to determine the validity of our CIS and
other parameter inputs.

       The primary weakness of the AGNPS model was the rainfall parameter.  Precipitation
distribution is highly variable in Arizona, whereas the model would take only one measurement
of rainfall across the entire watershed.  We attempted to use Kriging programs to estimate more
accurately the distribution of rainfall based on real measurements, but the AGNPS program
would not accept the data.  Furthermore, the AGNPS model is field-based rather than watershed
based, and the very steep terrain we encountered resulted in questionable results.  We concluded
that  the  model could  only be  used  for  qualitative, and  not quantitative,  predictions.
Nevertheless, the program may be successful in the  future when final  decisions are made
regarding the water augmentation project. In addition, we anticipate success in using the model
derived for this one watershed on other watersheds for  prioritization purposes.


Q:     How  many people in the State water quality agency have worked using satellite imagery
       for their projects?  Were there other expert sources (e.g.,  universities) that could be
       tapped to help develop and use  the GIS.
A:     The ENSL lab in Las Vegas has been helpful.  Arizona has made a direct commitment
       to the utilization of GIS,  and the sharing GIS information across the board has made it
       more economical to use remote sensing as  a  tool in water quality and  wildlife
       management.  The Fish and  Game Department uses  it,  and also the State Land
       Department  for  fire prevention.  Unfortunately, there are not enough people within
       ADEQ who  know how to use the tools.  The technical expertise is not there.  It takes
       from  one to  three years to become truly comfortable not only on a technical level, but
       on a management level, too.   In  general state level  managers are weak in their
       understanding of how  much time data generation requires, whether their questions be
       answered using remote sensing, GIS, and/or modeling, and whether they can afford it?

Q:     What is the turnover rate among state "technical experts"  in ADEQ?
A:     An individual who has expertise in remote sensing and GIS can command almost double
       what  he can make with the state; therefore, turnover of experts at the State level is high
       -- about 1 per year.

             An Overview of Ecological Assessment and Restoration Tools

                                    Carl Richards
                         Natural Resources Research Institute
                            University of Minnesota-Duluth
                                  Duluth, Minnesota

       Restoration of aquatic ecosystems  requires knowledge of the biological physical, and
chemical components of these systems and identification of the spatial  scales at which  they
interact. '  Assessment  of aquatic  systems typically  involves  analysis  of functional (e.g.,
production, decomposition, nutrient dynamics) and structural (e.g., species richness, trophic
structure) elements that are important to maintenance of system health and stability.

       Anthropogenic influences that adversely effect functional and structural aspects of aquatic
systems occur at variety  scales  ranging  from  a few  square  meters to hundreds  of square
kilometers.  Assessment of these influences requires recognition of spatial and temporal scales
of importance.  Conventional means of measuring important ecosystem components along with
the development of biological indices that integrate both structural and functional information
can describe system health. Emerging technologies (geographic information systems [CIS]) for
quantifying watershed and land use characteristics that influence aquatic systems show much
promise for defining and relating spatial features to ecological end points. Several multivariate
statistical techniques are available for integration of complex data and determination of limiting
environmental features.

       The most effective restoration methodologies address limiting environmental features at
scales complementary with the processes  that cause disturbance.  Most often the focus is on
aquatic-terrestrial ecotones.  These ecotones moderate the exchange  of nutrients  and other
materials across the  landscape.  Knowledge of  the morphology of ecotones is particularly
important when attempting to restore wetlands and streams.  In streams, the dynamic nature of
stream channels  and  response of these channels  to  common perturbations must be  taken into
account before designing restorative measures. In a large river restoration project in Idaho, the
major restoration design feature revolved around creation and revegetation of a floodplain for
a meandering river.  Ecological restoration can be an effective means of reducing pollutant
loadings to aquatic systems.


Q:     How much did the Idaho restoration project cost?
A:     The approximate construction cost  of this project was one million dollars per mile.  In
       this case, the cost was justified as a one time expenditure that had a restorative effect on
       sixty kilometers of stream with endemic chinook salmon populations.

Q:     What scale should watershed planning/restoration occur at?
A:     It depends on the scale of the problem.  If the problem is a road crossing that causes
       erosion into a stream, a local solution is usually all that is required.  However, the most
       persistent and difficult nonpoint source problems are often the result of cumulative effects
       (many erosive road crossings) or large scale practices like agriculture. Large watersheds
       may be the only effective scale at which to deal with these problems.

Q:     What are the implications of linking biological integrity to the physical chemical status
       of a stream?
A:     When discussing pollutant control on a watershed basis, it is important to know what
       impacts pollutants have on the aquatic biological communities. As scientists, we would
       like to understanding or be able to predict the observable, quantitative linkage/change in
       a community when a  specific  pollutant loading is reduced by  twenty  percent,  for
                     Application of Tools for Ecological Restoration
                                  Predictive Modeling

                                    James A. Gore
                              The Center for Field Biology
                              Austin Peay State University
                                 Clarksville, Tennessee

       Although restoration techniques exist for wetland, lake, and running water ecosystems,
modeling efforts have been focused on predicting the effects of restoration on lotic ecosystems.
The recently introduced concept of hydraulic  stream ecology  suggests that a primary template
influencing the distribution of riverine organisms is the complex hydraulic conditions which exist
as the result of  the  interaction of  depth, velocity, and  substrate.   Thus, with  the  aid of
simulations which combine the physical habitat preferences of riverine organisms and predictions
of changes in hydraulic patterns after structure placement,  it is possible to weigh the estimates
of available habitat by response factors which indicate tolerances to certain conditions at each
mitigation site.  It must be remembered  that most of  these models  are  not at the stage of
predicting "true" ecological response (i.e., changes in production, biomass, recruitment, etc.),
but offer stream managers  a scale  to compare relative magnitudes of improvement or success
of the restoration effort.

       Examples of application of these models include the prediction of the effects of placement
of a single structure (a flow re-regulation dam) on trout fisheries with and without alterations
of release schedules from peaking hydropower facilities and an extensive physical restoration of
meander, substrate, and habitat  structure after coal surface  mining.  In both cases, components
of the Physical Habitat Simulation (PHABSIM)  [maintained by the U.S.  Fish  and Wildlife
Service as a portion of the Instream  Flow Incremental Methodology] were used to predict the
effects of habitat improvement.

       In combination with the Branched Implicit River Model (BIRM) to predict changes in
stage, a dynamic advective water quality model, RIV1, maintained by the U.S. Army Engineers'
Waterways Experiment Station, has been successfully used to predict the changes in basic water
quality parameters downstream of Wolf Creek Dam on  the Cumberland River in Kentucky.  A
subroutine, RTV1H allows the "placement"  of in-stream structures in the model.   In this case,
a reregulation dam affected both water quality and the amounts of available habitat for rainbow
and brown trout.  Trout habitat availability was predicted  by linking RIV1H water surface
elevation predictions and surveyed whole-valley profiles to the hydraulic simulation, IFG4.  The
predicted changes in hydraulic patterns were combined  with habitat preference curves for trout
species (juveniles and  adults) in the HABTAT simulation to indicate that reregulation (even with
turbine uprates which increase  maximum discharges) effectively increased habitat availability.
Since predicted water quality changes were  within optima provided by lexicological studies, no
additional weights were applied.   However, if the response curves are  available,  this same
technique can be used to predict combined effects of water quality and structural change in the
river system.

       In coal surface mine restoration projects on the Tongue River, Wyoming, PHABSIM was
used to predict the amount of available  macroinvertebrate and fish habitat prior to placement of
restoration  structures.  The model predicted that, after  implementation of restoration, changes
in water surface elevation and velocity  downstream of these structures  would  dramatically
increase  habitat  availability  and,  presumably,  diversity  and   density  of  the  fauna.
Macroinvertebrate densities and diversities were a full three-fold higher than unrestored sections
and fish densities (although primarily transient individuals) were three to five times higher than
unrestored  areas.  Monitoring of the fauna for two years indicated that a source-distance effect
must also be  incorporated into future models. That is, areas  further downstream (even as short
as one kilometer) may take 300 to 600 days longer to equal the community structure of areas
immediately adjacent  to sources of faunal colonizers. This research also revealed the need for
future research into the effects of changes in hydraulic pattern on colonization rates.  If  it were
possible to  construct "suitability for optimal colonization curves," PHABSIM could be used as
a predictive response  model for future  restoration projects.

       Projected research on the effects of in-stream structures to control nonpoint sources from
agricultural areas, combinations of in-stream habitat improvements and cascades of low head re-
regulation weirs, and  long term studies of habitat improvements and recruitment of endangered
mussel fauna and  host fish will allow the  "fine tuning" of existing models to better  predict
effective management strategies to stressed  ecosystems.

             Ecosystem Assessment and Restoration: The Role of Modeling
                                   and Predictability

                                   Michael Johnson
                               Kansas Biological Survey
                                 Kansas City, Kansas

       Modeling the effects of stresses on ecosystems  should be conducted with the objective
of assessing both current and future status of the ecosystem.  This objective can be achieved by
placing the ecosystem into a framework of structure and function, and modeling the ecosystem
with a structural equation modeling procedure called LISREL.

       Using LISREL, it is possible to develop empirically-based models that combine aspects
of simulation modeling with statistical analyses and hypothesis testing. Modeling with LISREL
generates a measure of the stability of the ecosystem, thereby generally  predicting the ultimate
fate of the ecosystem.  These models can be analyzed further to provide additional information
concerning the future status  of the ecosystem including  (1)  ecosystem  sensitivity -  what
functional pathway within the ecosystem, when perturbed, results in the greatest change in the
stability of the ecosystem, (2) ecosystem projection —  given a set of parameter values for the
structural components of the ecosystem, what happens to the structural components at some time
in the  future, (3) perturbation analysis  -  what happens to the structural components and the
overall integrity of the ecosystem if an anthropogenic stress  affects the ecosystem, and (4)
hierarchial structure — at what level of taxonomic/tropic/functional resolution does the system
gain/lose stability when impacted by an anthropogenic stress. This information can be used to
develop risk assessment protocols for ecosystems, and as an aid in restoration  of ecosystems
known to be impacted by anthropogenic stresses.


Q:     What  was the time step used for the calculations; and  how much data collection is
       necessary to validate LISREL?
A:     The data was collected each summer; although, it may have been better to shorten that
       time step to early-summer, mid-summer, and late-summer. Since the model is empirical,
       the classical validation techniques that were discussed yesterday during  the watershed
       modeling presentations, are not applicable.


            Topic Chair:
Lee Mulkey, U.S. EPA-Athens
James Giattina, U.S. EPA-Region V
David Dilks, Limno-Tech, Inc.

Ambrose, Robert
Baumgartner, Robert
Beyerlein, Douglas
Biswas, Hira
Burt, John
Dannel, Mimi
Davenport, Tom
Donigian, Anthony
Driscoll, Eugene
Fontenot, Wildon
Goggin, Mike
Greenfield, James
Harper, Warren
Jawson, Mike
Jaworsky, Norbert
Knighton, Dean
Livingston, Erik
Lynch, Robert
Paluzzi, Jeanna
Pellegro, Marianne
Pepin, Robert
Richardson, William
Schmelling, Stephen
Shoemaker, Leslie
Suzukida, Irene
Swanek,  Ruth
Terstriep, Mike
Weatherbee, Donald
U.S. EPA-Athens
OR Department of Environmental Quality
Snohomish County, WA
U.S. EPA-Region VI
U.S. EPA-Region V
AQUA TERRA Consultants
Woodward-Clyde Consultants
USDA Forest Service
U.S. EPA-Region IV
USDA Forest Service
USDA Forest Service DC
FL Department of Environmental Regulation
OK Conservation Commission
Wayne County, MI
U.S. EPA-Region V
U.S. EPA-Region V
U.S. EPA-ORD Grosse Isle, MI
Tetra Tech, Inc.
NC Division of Environmental Management
IL State Water Survey
Provence of Ontario, Canada


       The breakout session focused upon one simple question: "How (if at all) should models
be used in the Total Maximum Daily Load (TMDL) process?" There was significant debate and
many differences of opinion  among  those  responsible  for  developing TMDLs and  those
responsible for complying with them.   A compromise approach  was generally agreed  upon,
although two unresolved issues were identified.


State of the Science

       Watershed models are used to estimate loadings from both point sources and nonpoint
sources of pollution from urban and rural land use activities, and  are particularly useful  in the
pollution  allocation  phase of the TMDL process.  EPA currently distributes two  relatively
complex watershed models ~ Hydrologic Simulation Program  - Fortran (HSPF) for urban and
rural mixed land  use activities, and Storm Water Management Model (SWMM) for urban land
use activities.  EPA also provides some limited technical support for these models, but such
support needs to be expanded  if they are to  be  used on a wider scale.  U.S. Department of
Agriculture Soil Conservation Service (USDA SCS) has developed six simpler and more user-
friendly models that may be useful to the TMDL program.  They are Agricultural  Nonpoint
Source Pollution Model (AGNPS), Simulator for Water Resources  in Rural Basin-Water Quality
model (SWRRBWQ), Erosion-Productivity Impact Calculator (EPIC), Groundwater Loading
Effects of Agricultural Management Systems (GLEAMS), Chemicals, Runoff and Erosion from
Agricultural Management Systems (CREAMS), and Nitrate Leaching and Economic Analysis
Package (NLEAP).

GLEAMS, CREAMS, EPIC, and NLEAP are field scale models and therefore have limitations
with regard to the TMDL process. The SWRRBWQ model can be applied on a watershed basis
for continuous simulation, however, additional development of model components for nutrient
and pesticide transport is underway. The AGNPS model is watershed based, but is currently
limited by application to design storms only. USDA is currently upgrading the  model to include
continuous simulation as well.

       One limitation that was identified by the Workgroup is a lack of available data to permit
proper validation of nonpoint source model  applications.  Use  of  these state-of-the-science
models with that data that is currently available was deemed appropriate only for determining
relative impacts,  not for predicting absolute  loadings.  A second concern  was that there are
multiple Federal agencies each supporting different nonpoint source models.  It would be helpful
to States, local governments, and other model users if the Federal agencies coordinated their

Limitations of Tools and Data

        Other federal agencies, the Regions, and States have been mandated to develop TMDLs
as soon as possible using whatever data are available. A primary concern among these groups,
as well as EPA, is the availability of models that can be quickly and easily applied to the TMDL
process. Use of "state of the science"  models was not considered to be necessary or even
desirable given the limited expertise and funding among most users.  Use of simple  models is
considered acceptable as  long as the approach to  develop the TMDL is reasonable from a
scientific point of view.

        Those responsible for complying with TMDLs (e.g.,  U.S. Forest Service, USDA) do
not believe that the predictions of currently available models are sufficiently reliable to base
management decisions that may require expensive pollution control measures.  Although models
are  useful  in  determining relative  impacts,  their  ability  to predict  absolute  pollutant
concentrations is highly suspect.

TMDL Applicability


        A loose consensus on the use of modeling in  the TMDL process was eventually reached.
The consensus approach closely matched the "Phased Approach" described  in EPA's  Guidance
for Quality Based Decisions: The TMDL Process (EPA, 1991), even though the great majority
of workgroup participants had not yet seen this document.  The steps of the Phased Approach

1.      Perform simple, screening level modeling  using existing data to  define a "Phase I"
        TMDL in the timely fashion required.

2.      Implement the pollution controls identified in the Phase I TMDL.

3.      Conduct a monitoring program to:  a) determine if the Phase I controls are appropriate
        for compliance with applicable water quality standards, and b) provide a database for
        more rigorous modeling.

4.      Perform additional modeling  and develop a Phase II TMDL, if monitoring data show
        that the Phase I TMDL was inappropriate.

Unresolved Issues

        Two  issues were identified by the group that remain unresolved:

1.      How  quantitative must the Phase I  TMDL be?

2.      By what process should a TMDL consider  the range  of potential  flow conditions that
        might affect pollutant loading to a water body?

       Discussion of development and use of the Phase I TMDL lead to a stalemate between
the regulated community and the regulators.  To facilitate the permitting process, regulators have
an obligation to define specific, quantitative load allocations (LAs) and waste load allocations
(WLAs) for a Phase I TMDL. The regulated community believes that Phase I TMDL modeling
should be used only to screen the relative effectiveness of various pollution control strategies,
and that quantitative LAs based on screening level modeling would be too unreliable to base a
permit requirement on them.

       The second unresolved issue pertains to flow conditions for which the TMDL will be
applied.  Point source NPDES (National Pollutant Discharge Elimination System) permit limits
have traditionally been calculated based on  low flow conditions, since pollutant concentrations
are highest, and therefore  most critical, during low flow periods.   Most  nonpoint  source
pollutant contributions  are  highest during high flow, or storm events, when runoff from
construction sites, and agricultural or silvicultural areas is most likely.  However,  point sources
are still a consideration at high flow, and  some nonpoint sources are important at lower or
average flows. The specific process by which the TMDL should consider the range of potential
flow conditions was unclear to the group.


       Seven categories of recommendations based on both short-term (S) and long-term (L)
needs were identified.


•      Convene a small group of experts from each facet of the TMDL process to explore the
       existing science, provide immediate short-term technical guidance, and develop a long-
       term strategy for developing TMDLs. (S)

•      Implement the long-term strategy identified by the expert group. (L)

•      Continually update and refine TMDL guidance. (L)

Case Studies

•      Compile  and  evaluate existing watershed and TMDL modeling efforts. (S)

•      Build a long-term record of case studies among Agencies, Regions, and States, focusing
       upon model validation and uncertainty. (L)

•      Conduct  more work on an inter-Agency level. (L)


Geographic Information Systems (GIS)/Data Management

•      Develop a working prototype linking CIS to TMDL modeling, concentrating on how
       GRASS and ARC-INFO should or should not be integrated, how to integrate CIS and
       remote sensing data into both static and dynamic models, and how national data bases
       can be made more accessible with a long term view of the eventual development of
       operational systems. (S)

•      Develop operational systems linking CIS to TMDL. (L)

Design Storm versus Continuous Simulation

•      Prepare a monograph  ~ a description of the pros and cons, dos and don'ts, whys and
       wherefores,  and alternative views - critically exploring this subject in order to develop
       fast, simple, easy, standardized ways of looking at watershed modeling for nonpoint
       sources.  (S)

•      Begin directed studies of this subject.  (S)

•      Continue directed studies. (L)

User Support/Modeling Usability

•      Improve documentation of existing models, making sure that they are available. (S)

•      Create a centralized model support center that will be run jointly by all federal Agencies
       in order  to facilitate model support to State and  local governments, and consultants. (S)

•      Increase the amount of targeted training for use of models within the TMDL process.

•      Standardize  the presentation  shell for more user friendly access to various models. (S)

•      Provide easy to use screening models. (S)

•      Continued evolution of models. (L)

Ecological Links

•      Examine linkages between  available  models  and ecological  habitat, i.e.  velocity,
       sediment transport, stream morphology.  (S)

•      Examine linkages between models and existing  ecological assessment tools (i.e., Index
       of Biotic Integrity [IBI], Rapid Bioassessment [RBA]). (S)

•      Develop linked watershed-water quality-ecology models. (L)

Monitoring to Support Models

•      Monitoring guidance on the proper procedures and techniques required to collect data
       that can be useful in the available models. (S)

•      Description of current monitoring activities.  (S)

•      Application of other technologies, i.e. remote sensing to take advantage of the Earth
       Observation System (EOS)/NASA  program, to get the kind of data we need to do
       TMDL assessments. (L)

•      Develop a long term  monitoring strategy. (L)

      Topic Chair:
Ross Lunetta, U.S. EPA EMSL-Las Vegas
Mason Hewitt, U.S. EPA EMSL-Las Vegas
Michael McCarthy, Research Triangle Institute

Atkinson, Samuel F.
Bartholic, Jon
Bondelid, Tim
Goay, Gary
Gray, Ted
Greenwood, David
Gumtow, Robert
Higgins, John
Jaynes, Dan
Kuehn, Mike
Luman, Donald
Pitt, Jerry
Russell, Carol
Swanson, Ed
Taylor, Phill
Northern Texas State University
Michigan State University
Research Triangle Institute
LA Department of Environmental Quality
Northeastern IL Planning Commission
LA Department of Environmental Quality
WY Department of Environmental Quality
Tennessee Valley Authority
USDA Forest Service Alaska
Northern Illinois University
U.S. EPA-Region VII
AZ Department of Environmental Quality
AZ Department of Environmental Quality
Tetra Tech, Inc.


      The Workgroup focused its discussion on the following questions:

1.     How can remote sensing information and geographic information systems (CIS) be used
      in the Total Maximum Daily Load (TMDL) process?

2.     What are the  limitations of using remote sensing and CIS with watershed models?

3.     How can improvements be made to remote sensing and GIS as data tools?

4.     Where are there gaps in knowledge?

5.     What are recommended actions for research and guidance?


State of the Science

Determining Land Use/Land Cover (LU/LC)

      As nonpoint source pollution controls are integrated into the TMDL process, accurate
information of LU/LC will be essential.  For example, nutrient, sediment, and pesticide loads
from  agricultural  watersheds cannot be estimated without data on the extent and  type  of
cropland, pasture land, livestock operations, etc.  Furthermore, ecological considerations (i.e.,
habitat extent, juxtaposition, interspersion, etc) may become important components of TMDLs.
Remote sensing technology can provide LC/LU data in a format that can be input directly into
watershed models or converted to vector based GIS data base coverage.

      The workshop presentations illustrated the use of remote sensing technology to provide
detailed LU/LC information (i.e., determine the crop cover type of individual agricultural fields,
delineate wetland boundaries, perform species specific measurements of forest crown closure or
size class).  The level of detail that can currently be obtained using remote sensing is dependent
on the sensor being  used, prevailing atmospheric conditions at the time of data collection, and
the temporal variability of the target(s) of interest.  In general,  aerial photography can provide
the greatest level of LC/LU detail and is the most cost effective method for small study areas
(i.e., less than 200 square miles); and satellite imagery has the potential to provide similar levels
of LC/LU detail over large study areas (i.e., greater than 200 square miles).  Satellite derived
data has the added advantage of direct input into GIS without costly intermediate data transfer

Modeling and Model Verification

      A few point  source and nonpoint  source models are now being operated in the GIS
environment.    This is especially  relevant  where  distributed  input data  (as  in  AGNPS
[Agricultural Nonpoint Source model], WRENSS [Water Resources Evaluation Nonpoint Sources

Silvicultural model] and some other nonpoint source models) are used and where outputs are best
displayed in graphical format.

      Currently, LANDSAT  imagery is being used in conjunction with CIS for watershed
modeling and  even multi-basin level  planning.  Examples of watershed applications  were
presented by Hewitt, Lunetta, Bartholic, and Lyon during the first part of the Workshop.

TMDL Applicability

      Session participants identified several ways  in which remote sensing data and CIS could
be used in the TMDL process.

Integrating Information

      CIS is a valuable tool for integrating the many types of data with LU/LC information to
State  and local officials,  including  TMDL  developers.  (The  Assessment and  Watershed
Protection Division [AWPD] in EPA's Office of Wetlands, Oceans and Watersheds is currently
compiling case studies on the  use of CIS in water quality and watershed planning.)

Monitoring of Best Management Practices (BMPs), Wetlands, LU/LC

      The phased approach will require the installation of BMPs, and possibly the protection
and restoration  of  sensitive areas  such as wetlands and riparian zones.   At the basin level,
remote sensing could reduce the cost of monitoring  both the installation of BMPs and LU/LC
changes within sensitive areas.

Sampling Network Design

      CIS systems can aid in the design of sampling networks.  As an example, EMSL-LV has
designed  an Reach  File 3 (RF3)-GIS data bridge which has been implemented by OIRM.  This
bridge will promote data sharing  between RF3 and  GIS users.  The functionality of  GIS will
provide a system for sampling network design  using RF3 as a base information layer.   The
EMAP Program has already implemented this plan for the Surface Waters Resource Group.

Monitoring the Water Quality of Lakes,  Bays and Estuaries

      As a check  on the  overall  effectiveness of TMDLs in  a  watershed or basin, remote
sensing data for certain parameters (e.g., sediment,  chlorophyll a) may be used to  track  long-
term water quality.  For example, after  calibrating satellite measurements to field data from
synoptic  surveys,  satellites might  monitor estuarine water quality much  more frequently and
extensively than is  possible in situ.

      Other uses for remote sensing data that were  mentioned by  the Workgroup were model
verification and two-dimensional mixing zone analysis.

Limitations of Tools and Data

Remote Sensing

      Current LANDSAT sensors are cost effective for large-area (i.e., greater than 200 square
miles)  applications, while  TMDLs may require higher resolution  data for small watershed
studies. In most watershed studies, a multistage remote sensing approach will provide the best
results.  Multistage remote sensing provides complete watershed coverage at the coarsest spatial
and categorical resolution (i.e., one hectare/EMAP [Environmental Monitoring and Assessment
Program]-LC stage two), with selective higher spatial and/or categorical resolution for a selected
geographic area or categorical  element of a  watershed to meet modeling requirements.   (The
spatial resolution of remote sensing data can be to the sub-meter level  and categorical resolution,
down to the species level.)  LU/LC is becoming available for much of the country at about one
hectare resolution using LANDSAT Thematic Mapper (TM) imagery, but use classifications are
fairly gross for some LU  types.  For example, LANDSAT imagery is being  processed for
EMAP-LC to lump all forest land cover types into three categories ~  evergreen, deciduous, and

      A potential problem of significant concern is the status of the existing LANDSAT sensors
and the United States near-term commitment to the LANDSAT program. Currently, LANDSAT
4 and 5 are in operation. However, both satellites  are long past their designed operational life.
LANDSAT 5, the principal TM data  acquisition platform is  beginning to  degrade, and its
replacement,  LANDSAT 6, is not scheduled for launch until the spring-summer of 1992.  We
are currently seven to ten years from deploying the next generation of satellite sensors being
developed under the Mission-to-PIanet-Earth, Earth Observation System (EOS).

    Obtaining detailed crop type delineation data using  remote sensing typically requires multiple
remote sensing  data  collections to  coincide with  local crop calendars for a high degree of
accuracy.  Detailed crop type delineation using satellite imagery is hit and miss due to the lack
of multiple growing season data collections for most locations.

      Other limitations of remote sensing/GIS  tools and data include:

•     Uniform standards for remote sensing and CIS data quality are not in  place (e.g., for data
      referencing requirements);

•     Merging data from multiple platforms (e.g., SPOT [pan.] and LANDSAT TM) is routine,
      but it is the most cost effective means of providing high resolution large area LC/LU data
      at this time.

•     Aerial photography can provide highly detailed  LU/LC information, but is  expensive and
      can require multiple flyovers for certain applications (e.g., determining crop cover)

Geographic Information Systems

      Widespread use of GIS has been limited by the resources (hardware and personnel costs)
required to support them. The systems in current use for state-wide or regional use tend to be
complex and  resource demanding.  Many States lack staff that are trained to use them,  and
therefore,  when a potential application is identified, environmental planners must often wait
months to obtain expert help.  The fact  that different Federal agencies use different systems
(EPA supports only ARC-INFO; the Department of Agriculture Soil Conservation Service (SCS)
supports GRASS) complicates data transfer decisions.

      The situation is changing rapidly.  Many PC-based GIS systems are currently on the
market and the hardware evolution will soon bring high-performance work stations within the
budget of many State agencies.  The Federal agencies and vendors are in the process of agreeing
upon  a spatial data  transfer standard  which will promote data sharing among many different
platforms.  Finally, GIS technical  support  centers,  such as EMSL-LV, are pioneering GIS
applications that are directly applicable to watershed process modeling and management.

      Given  the evolution of the technology and technical support,  use of GIS in  a TMDL
framework is a realistic goal. States should  begin to seek guidance and support from the EPA
Regional GIS Teams.  These Teams have been established in each Region to, in part, promote
EPA-State data sharing.  The Regional GIS  Teams,  in turn, can call upon the EMSL-LV GIS
Research  and Development  Center for assistance  in planning and implementing systems to
support the TMDL process.

      In the past, watershed protection  and water quality planning have not  been priority
programs for State GIS centers. Some State governments have been developing GIS capabilities
for years (e.g., Illinois, Michigan), but do not yet use the systems widely for watershed or water
quality planning.  This too is beginning  to  change.  Both Montana and South Carolina have
embarked  on  ambitious plans to begin  state-wide watershed planning supported  by GIS.
(Contacts  for these States are Gary Ingman,  Montana Department of Health and Environmental
Services, or Anne Marie Hale, South Carolina Water Resources Commission).

State of the Science

      The overall lack of technical  guidance  and  understanding about integrated  PS/NPS
TMDLs is a current obstacle. Remote sensing and GIS experts need to know detailed program
requirements  and objectives so they can improve on the tools that exist.

      To  develop useful tools, EPA needs to know present and future State/local capabilities,
especially  regarding data analysis platforms  like  GIS.  The goal is to develop tools that can be
widely used.  GIS technology is forging ahead, while State and local governments tend to have
equipment from the early 1980's, if they have any  at all. Different types of users should be
profiled.   (Note:  AWPD is pulling together  information and case studies of State/local GIS use
for watershed planning in support of this recommendation.)

      Since many State water agencies are not close to having GIS capability, there is a need
to know how other platforms (e.g., the EPA/NCC mainframe computer) can be used to support
the data needs of TMDL development.

      Specific technical and institutional issues about remote sensing and GIS integration are
discussed in a series of articles in the June 1991 issue of Photogrammetric Engineering and
Remote Sensing.  Two excellent sources of information for potential GIS users are: Ceo Info
Systems published by Aster Publishing, and GIS World published by GIS World Inc.



Geographic Information Systems

•    Develop GIS/model interfaces so that models operating outside a GIS platform can access
      GIS  data layers; modelers then do not need to learn ARC-INFO programming.

•    For special applications, develop models and tools  within the  GIS environment that can
      make intensive use of distributed GIS data, with user-friendly macros for the modelers;
      an example is the nutrient budget program  being developed by the Research  Triangle
      Institute for basin planning in  the Albemarle-Pamlico Estuary  Study Area.

•    Establish data quality standards for TMDL modeling.

•    Develop raster-to-vector data conversions (and vice-versa). The power of GIS in TMDL
      modeling may be best realized with raster-based systems; such systems can best process
      distributed data  like LU/LC  and  soil  types.  SCS's GRASS system  is  raster-based;
      ARC/INFO is vector-based, but will be available as a raster-based system within the next

Remote sensing

EPA should:

•    Participate in the development of new sensors (e.g., for  the EOS sensors).

•    Explore opportunities to utilize multiplatform remote sensing data (e.g., LANDSAT TM
      and SPOT [pan.] data in combination for priority watersheds).

•    Perform urban and agricultural TMDL pilots to demonstrate the use of  remote sensing,
    ' GIS,  and  models in  concert with State and local agencies.  Projects should cover  the
      range of capabilities, tiers of complexity, and costs.  They should evaluate State/local
      capabilities and support development of implementation guidance.

      Develop standardized classification systems for CIS and remote sensing-oriented data such
      as LU/LC, soil type, and topography.

•     Technical guidance about integrated point and nonpoint source TMDLs is needed.  This
      guidance should include case studies;  however, since the number of integrated point and
      nonpoint source TMDLs is extremely limited, the guidance should also include experts'
      judgement about remote sensing and CIS data requirements for future TMDLs.

•     With the above in mind, an ongoing forum  should be established for GIS and remote
      sensing experts,  modelers, and others  responsible  for TMDL development.    A
      workgroup, for example, could deal with some of the technical and administrative issues
      identified above (e.g., data needs, data  quality, sensor design,  standardized remote
      sensing and GIS products, user interface improvements, information exchange).

•     Guidance on common spatial data references and on scale and resolution issues is needed
      (e.g., which base maps to use).

•     Guidance on GIS quality assurance and quality control issues is needed.

•     An evaluation  of the ability of  current GIS systems to  meet TMDL users'  needs is
      recommended.  Guidance should also address basic infrastructure needs and promote the
      States' efforts to acquire this infrastructure.

•     GIS and remote sensing considerations should be added to the Appendix E list of models
      in the April  1991 TMDL policy guidance.


•     Improve access to data from EPA databases.  The  current Office of Information and
      Research Management effort to create data access methods should be encouraged (Project
      GATEWAY). Other opportunities should be explored, including data sources  from the
      Departments of Agriculture and the Interior.

•     Models should  be designed  to use remote sensing data  and GIS; remote sensing  data
      collection should be designed to fit the best models.

•     In  anticipation of national availability  of  remote  sensing data,  data  storage/data
      management issues should be addressed early.

•     Standard remote sensing products should be developed and distributed. If EPA expects
      local and State  use of remote sensing  data, the agency needs to recommend and produce
      standard products; users at the State and local level usually cannot process raw satellite
      data, for example.

To effectively use remote sensing/GIS tools, a new regulatory framework for TMDLs
may be needed.  For example, the technology supports the incorporation of riparian zone
protection/restoration into TMDLs, while the regulatory framework appears  to be a

            Topic Chair:
      Nelson Thomas, U.S. EPA-Duluth
      Ron Carlson, U.S. EPA-Duluth
      Amy Sosin, U.S. EPA-HQ

Cleland, Bruce
Cooter, Bill
Creager, Clayton
Dettmen, Ed
Dicksen, Kenneth
Droppo, Ian
Germann, Sandy
Herricks, Edwin
Hubbard, Robert
Gagler, Jeff
Gore, James
James, William
Johnson, Michael
Laufer, Sue
Lyon, John
Michell, Peggy
Bruce Newton
Owen, Robert
Painter,  Bill
Peterson, Spence
Richards, Carl
U.S. EPA-Region X
Research Triangle Institute
Western Aquatics
University of North Texas
National Water Research Institute, Canada
University of Illinois
U.S. EPA - Region V
Austin Peay State University
University of Guelph, Ontario
KS Biological Survey
Tetra Tech, Inc.
Ohio State  University
Natural Resources Research Institute


       Ecological  restoration and predictive modeling  of ecological response is a rapidly
emerging  area of research.   In order to effectively integrate consideration of ecological
restoration into the Total Maximum Daily Load (TMDL)  process, several questions must be
addressed. The Workgroup discussion focused on some  of these questions:

1.     How should we define ecological restoration? What  is the goal of ecological restoration?

2.     What are our  current capabilities to predict ecological response?  What are existing
       empirical relationships that can be used in predictive ecological response models?

3.     How can predictive modeling  of ecological restoration be integrated into  the TMDL

4.     What are our needs? (near term and long term)

5.     What are recommended actions?


State of the Science

       After some discussion, the Workgroup developed  consensus definitions  of ecological
restoration and the purpose of predictive modeling of ecological response:

•     The goal of ecological restoration is to establish  a  viable,  functioning and sustainable
       ecosystem by providing an appropriate physical, chemical, and biological environment.
       The resulting ecosystem should be similar in its natural processes to adjacent ecosystems
       and/or meet appropriate biological criteria.

•     The goal of predictive modeling of ecological restoration  is to be able  to accurately
       predict community changes (e.g., changes in the abundance of a key species)  due to
       changes in the  environment (stressors).

       The Workgroup discussed the desired endpoint of an ecological restoration effort. The
definition  above states that restoration should result in an  ecosystem similar in characteristics to
adjacent areas. This immediate goal — to mimic an existing nearby system  ~ is based on the
necessity for a source  of colonizers. Restored ecosystems should also be  stable, and be able to
deal, to some degree, with various stressors.  The realistic endpoint of an  ecological restoration
project  may vary with the  size and scope of  the project; site-specific characteristics must be
taken into consideration.  Workgroup  members concluded, however, that the ultimate goal of
restoration is a return  to pristine,  undisturbed  conditions.

Limitations of Tools and Data

       One barrier to successfully predicting ecological response through models is the current
lack of appropriate ecological data (especially long term data).  Indeed, decisions must often be
made with little or no data.  However,  the Workgroup concurred that the role of predictive
models  should not necessarily be  to precisely quantify or to establish causality, but  to
communicate the relationships between ecological restoration efforts and the resulting benefits.

       With this role in mind, the Workgroup acknowledged the importance of simple empirical
relationships  for use in  modeling (as  opposed to detailed, deterministic  approaches).  Many
empirical  relationships that can be incorporated into predictive models have already been
identified.  Examples of these include the relationship between distance from  a  colonization
source and colonization rate, and between stream flow and macroinvertebrate abundances.  The
Workgroup agreed that identification of new empirical relationships is also necessary in order
to link various instream  characteristics to riparian, out-of-stream components.

       Further, the Workgroup recognized that models must be consistent with existing policy
requirements. There is a need for models with varying levels of complexity and resolution since
site-specific conditions may call for different levels of accuracy.

TMDL Applicability

       To successfully integrate ecological restoration into the TMDL process requires a direct
link between restoration and management.  The  Workgroup concluded that TMDLs must be
redefined to include habitat and ecological and restoration criteria.

       In terms of models, the Workgroup concluded that there is not necessarily a need for new
models. Rather, there is a need for appropriate linkages among existing models and validation
of these linkages  for the TMDL process. Further, as part of the TMDL  process,  improved
monitoring programs will  help to test  model effectiveness, as well as provide information, for
model refinement.

       A hierarchical approach to TMDL development can be used in order to effectively link
existing models.  This could be accomplished by initially using screening level models to target
watersheds;  more complex and site-specific models could then  be  used  to focus on targeted
areas.  Such a tiered approach would enable us to deal with complex ecological systems.

       While improved  data as a result of research and ongoing  projects allow the use  of
increasingly complex models, uncertainty - in the accuracy and predictive capabilities of models,
and in ecological criteria  that are developed - is still an issue.  The Workgroup,  however,
maintained that the establishment of ecological criteria, despite some uncertainty, is critical for
the development of TMDLs.


      The following  is a short  list of near-term and long-term needs that the Workgroup

•     Identify models that can be used and linked together for predicting ecological response
      (as part of this  effort, develop a compendium of models to summarize their capabilities
      and availability).

•     Validate and assess linked models to test the models and relevant criteria (feedback loop).

•  .   Develop a compendium of case studies that can be used to demonstrate  successful (and
      unsuccessful) approaches for TMDLs.

•     Develop a  compendium  of best  management practices  (BMP) and document BMP
      successes and failures.

•     Develop simple empirical relationships and identify response and community  metrics
      (e.g., regionally specific) that can be incorporated into predictive models for ecological

•     Develop regionally (spatial/temporal) relevant biological criteria.
•     Develop ecological criteria.

•     Classify waterbodies by their recovery/restoration potential.

•     Develop public education/outreach programs to communicate (e.g., through case  studies)
      the benefits of ecological restoration at the local level.

•     Develop a regulatory  framework that will allow enforcement of ecological restoration.

•     Develop mechanisms  to integrate restoration  projects into watershed programs (move
      away from a program-by-program approach).

      From the needs described above,  the  following were established as priorities that will
serve as recommendations for immediate action:

•     Develop ecological criteria (spatial and temporal).

•     Develop a compendium of models.

•     Develop a compendium of case studies.

•     Develop a classification scheme  for waterbodies based on  their restoration/recovery

•     Develop a targeted monitoring feedback loop to evaluate models and criteria.

            Topic Chair:
      Betsy Southerland, U.S. EPA - HQ
      Bruce Zander, U.S. EPA-Region VIII
      Paul Freedman,  Limno-Tech, Inc.

Boynton, King
Cavacas, Al
Howard, Tom
Jackson, Wayne
Liu, Ed
Pagenkopf, James
Sullivan, Nancy
Brandes, William
Rossman, Lew
Tetra Tech, Inc.
CA State Water Resources Control Board
U.S. EPA-Region II
U.S. EPA-Region IX
Tetra Tech, Inc.
U.S. EPA-Region I
U.S. EPA-HQ Permits


       Water quality-based effluent permits for individual point sources  have been used  to
control impacts on receiving water quality for decades.  The Total Maximum Daily Load
(TMDL) process proposes to change this permit-by-permit approach.  TMDLs must consider all
of the point and nonpoint pollution sources in a watershed that contribute to a water quality
problem, and then allocate allowable loads to each source. Integrating point source waste load
allocations (WLAs) with nonpoint load allocations (LAs) to establish a TMDL, however, creates
various management and technical challenges.  Identifying and evaluating these challenges from
a point source perspective was the objective of the Point Source Workgroup.


Urban Runoff and Overflow

       Urban runoff can  severly impact water quality.   Overflow from combined sewer and
storm water drainage systems has long been an acknowledged water quality problem.   Several
recent developments, however, have placed new emphasis the evaluation and control of these
pollution sources.  First, TMDL development will require evaluation and control of all pollution
sources, including combined sewer overflows (CSOs) and storm  water.  Second, new storm
water permitting regulations will create a new data base characterizing  storm water discharge
throughout  the nation.  Third, a National CSO Policy is being  implemented state  by state,
requiring very specific actions for CSO characterization and control.

State of the Science

       Methods for collecting urban runoff and overflow data have been refined and reliable
equipment and methodologies now exist to characterize and quantify these pollutant loads. The
data that has been collected using these methodologies is abundant, but highly variable.  To
date, the National Urban  Runoff Program has collected the largest  nationwide data  base, but
countless other engineering studies have been or are being conducted.  Deficiencies in data on
receiving water  impacts may be the only major information gap.  This need not hinder TMDL
development, however, since  water  quality impacts, which  tend to be site specific,  can be
assessed during  the TMDL analysis.

       There are numerous methods and models that quantify combined sewer and storm water
overflows.   Fairly  sophisticated,  yet reliable models are also available to simulate the time
variable nature of wet weather urban overflows (e.g.  SWMM, ILLUDAS, DR3M, STORM).
Many of these tools can be used to provide long term continuous simulation, and are well suited
to address most of the parameters of concern.  Simple methods, such as the Rational Method
and the EPA simplified Storm Water Managment Model (SWMM),  that are somewhat less
accurate are also available, but they are most appropriate for storm water runoff.

       Concentration criteria are available for most media and pollutants of concern.  The only
exception is for clean and contaminated sediments.  The criteria are well suited for evaluating
overall impacts, but computational methods and/or models are required to  evaluate transient

impacts.  Fortunately, numerous methods and models are available to calculate water quality
impacts, but the detailed temporal data on loads and system hydrodymamics that are necessary
to support the modeling are typically not available.

TMDL Applicability

       Urban, wet weather pollution loads must be considered by TMDLs because they impact
water quality and often impair designated uses.   However, the analysis must be tailored for
different pollutants.  Impacts from nutrients, sediments, and persistent toxicants are generally
not episodic and should be analyzed on a seasonal, yearly, or period-average basis.  In contrast,
transient impacts from oxygen demand, metals, toxic organics, and bacteria need to be evaluated
in an episodic context to identify transient violations in water quality criteria.

       Unfortunately, the  differences among these  problems pose difficulties  for  TMDL
analysis. For example, no simplified methodology exists to quantify the loadings and impacts
of parameters with cumulative  impacts.  Complex methods are available,  but  they  are too
complex and  costly for routine TMDL analyses.  For episodic impacts, methods are available
but there is no guidance  on how to select suitable design conditions for an episodic TMDL
analyses, and there is no guidance on to how to integrate critical drought flow TMDL analysis
for continuous  loads with the analysis for episodic storm impacts.  A simple framework is
needed to consider both continuous and transient wet weather loads for  TMDLs.

Limitations of Tools and Data

       Models and data  are available to incorporate  urban runoff as a pollutant source in a
TMDL analysis. Several important limitations were noted, however:

•     Simple loading and impact analysis procedures that are targeted for rapid implementation
       by regulators are not available.  Available methods may be applicable, but need testing
       and demonstration.

•     A  simple flow criteria  or design condition is  not  available to regulatory analysts for
       calculation of urban wet weather TMDLs.  The equivalent to a mean 7-day,  10-year low
       flow and/or rainfall criteria needs to be developed for wet weather urban conditions.

•     No  methods  or  policies  exist  for integrating  episodic and  continuous  impact
       considerations in one TMDL or one TMDL framework.

Sediment  Criteria

       To protect beneficial uses, water pollution regulation has historically focused on meeting
water quality concentration criteria.   This may have protected  sediments indirectly, but no
explicit evaluation was ever performed. EPA will soon complete a multiyear research effort to
develop explicit sediment quality criteria for protection of beneficial uses. The existence of new
sediment criteria will place new demands on  the TMDL process.

State of the Science

       Data, methods and models for evaluating sediment quality are being developed.  Data
collection methods for sediment are well developed, and a substantial data base is being created;
therefore, once quality criteria are defined, established, procedures are available to evaluate
compliance. Sediment criteria for six organic pollutants are targeted for promulgation by EPA
by fall 1991.  Metals criteria are  targeted for late  1992.

       A computational methodology or a mathematical model is needed to relate pollutant loads
to sediment conditions.  Researchers have  developed  several models  that relate pollutant
loadings, overlying water concentrations, and sediment concentrations (e.g., WASP, MICHRIV,
SMPTOX3).  These models are very effective for simulating dynamic equilibrium conditions
among sediments, water quality,  and loading. The equilibrium model approach  assumes that
substance partitioning between water and solids is always in instantaneous equilibrium.  The
models, however, are best suited for relatively steady conditions.  They are not suited for highly
transient conditions, or conditions with large changes in loading.

TMDL Applicability

       The development of sediment criteria will impose new quality objectives on TMDL
analysis. Linking pollutant loading to sediment quality is an appropriate consideration and is
relevant for persistent toxicants.  If implemented, however, these criteria  will compete with
water quality-based  concentration criteria when establishing the TMDL.   In practice, it i.s
assumed  that both will need be calculated and the  most stringent used.

       Use of sediment criteria in the TMDL process creates two immediate concerns. First,
sediment quality reflect the cumulative effects of pollutant loading over a long period. A TMDL
intended  to meet sediment criteria,  however, will probably need to be expressed as a period-
averaged condition.   Unfortunately, there are no established procedures that define how to
compare a period-averaged TMDL with an episodic or a drought flow TMDL.   Second,  the
TMDL approach does not address the significance  of historical contamination. In  water quality
based analyses,  historical conditions are less important because of flushing (except in large lakes
and reservoirs).  Sediment contamination has a much longer residual memory that complicates
the TMDL analysis.  Will past contamination prevent standard compliance despite zero loading?

Limitation of Tools and Data

       The role of sediment quality criteria in TMDL analysis has not yet been unexplored.
Theoretically,  the equilibrium modeling approach used  in many EPA  supported models is
capable of relating  pollutant  loading to sediment concentrations, but for relatively  steady
conditions.   Nonetheless, the exploration and demonstration  of the approach  for TMDL
development is only now being undertaken.  The procedures are promising, but require specific
case study demonstrations to identify technical deficiencies and policy needs.

       On another  level, the equilibrium  model approach  supported  by EPA models is
significantly deficient in several areas. First, the approach is weak in addressing historical

contamination.  The models have not proven effective for simulating large changes in sediment
conditions, an area requiring more research. Second, available models are not well suited for
evaluating dramatic changes caused by highly variable hydraulic conditions which greatly alter
sediment settling,  transport,  resuspension, and  erosion.  Finally, long term  sediment  bed
transport is not well  incorporated into  available quality models.  Considerable research on
sediment transport has been done, but there has been very little done to link the hydraulics with
water and sediment quality modeling research.


       One objective of the Clean Water Act is to restore and maintain the biological integrity
of waterways.  Various physical and chemical criteria have been used as an indirect measure of
water quality.  Recent research, however, is developing  ecological criteria to directly measure
the health of water bodies.

State of the Science

       Techniques  to measure  and quantify aquatic  populations are  well developed  and
documented.  However, a consistent relationship between pollutant discharge  and  ecosystem
health as measured by population, biomass, and  species diversity or the  presence of indicator
species has not yet been established. Appropriate and consistent indices of ecosystem health are
not yet defined.   One  obstacle is natural variations in habitat that can impact population
dynamics and can exacerbate or minimize pollution impacts, making it difficult to define specific
aquatic health criteria.

       Whole effluent toxicity testing is one approach that has been heavily researched in the
laboratory, and applied  successfully to  regulate  point source discharges.  Its use as a direct
measure of ecosystem health is appealing, but there is little to indicate that the adverse impacts
observed under controlled laboratory conditions  reflect  the adverse impacts  may or may not
occur in an  uncontrolled, field  environment.   Furthermore, to use ecological criteria in a
regulatory context, reliable methods to relate decreases  in effluent  loading or urban runoff to
improvements in the  field  are needed.  The technology is not currently available.

TMDL Applicability

       Although ecological criteria are not well  suited for direct use in TMDL calculations at
this time, such criteria can play an important,  secondary role in the overall TMDL processs.
Two applications were identified as promising:

1.     Biocriteria can be  useful as an objective test of the effectiveness of an existing  TMDL
       to protect ecosystem health; they may  be used to justify relaxation or tightening of a
       WLA or LA in an existing TMDL calculation.

2.     Biocriteria may  also  be  well  suited as  a screening tool to  examine  attainability of
       designated uses in water bodies targeted for TMDL analysis. The result would be better
       selection of applicable chemical  specific criteria  for TMDL computations.

Limitation of Tools and Data

       Without methods to quantitatively link pollutant source and environmental processes to
effects (i.e., biological measure), direct use of biocriteria for calculating pollutant allocations
within a TMDL is not possible.   Specific criteria are undeveloped at this time, and  the
comprehensive data needed  to quantitatively support such criteria are lacking.  As a result,
available data, criteria and models are considered to be too limited, as  they exist today,  for
applicability to the TMDL process.


       The Point Source Workgroup evaluated the significance of the TMDL process to point
source WLAs and control.  Numerous technical and policy issues were identified that available
data, technology, guidance, and policy could not immediately resolve.

Marketing TMDLs

       The TMDL process is not necessarily familiar to State regulatory staff; therefore, in the
near-term EPA needs to provide technical support, funds, and guidance which encourage TMDL
development. The  Workgroup recommends that EPA should:

•      Target funds to support state efforts in TMDL development.

•      Develop and support a TMDL "SWAT" Team of experienced practitioners who would
       travel from state to state (or Region to Region) helping regulators develop a system for
       TMDL implementation and trouble shooting individual problems with TMDL analysis.

•      Expand support for hands-on workshops conducted by experienced TMDL practitioners
       to train state staff in each EPA Region in implementing TMDLs.

•      Develop institutional and technical guidance that promotes  the merits of the TMDL
       process. Specifically, the concept of WLA and LA tradeoffs must be better defined and
       marketed as an attribute of the TMDL process to encourage states and regulated NPDES
       (National Pollutant Discharge Elimination System) permit holders to invest their effort
       and money. Tradeoffs between load reductions and habitat improvement also need to
       be explored in order to identify the  most cost effective  means  to meet  water quality

Technical Guidance

       "Guidance for Water Quality-based Decisions: The TMDL Process" is the only TMDL-
related guidance which addresses both point and nonpoint issues in an integrated fashion. This
document, although valuable from a programmatic perspective,  was not designed to address
technical issues.  A TMDL "technical primer" is needed to outline integration of point source
waste  load allocations  with  nonpoint source load allocations  and the technical  steps and
considerations involved in TMDL development. Case studies that will serve as TMDL templates

would be included.  Particular attention should be focused on episodic versus cumulative and
continuous  impacts, plus drought flow versus wet weather, high flow TMDLs.  This primer
would precede a more comprehensive Technical Support Document  that is recommended to
follow in two to three years.

       More specifically, future  technical guidance on  the  following  subjects  should be

•      Institutional and technical aspects of trading point source WLA with nonpoint source

•      Case study documentation of TMDLs that have been implemented,

•      TMDLs for multiple  pollutants and whole effluent toxicity  (WET).   (This should
       consider synergistic and antagonist effects, as well as calculating a representative return
       frequency for toxic conditions and comparing it to the one in three year EPA guidance

•      Simplified methods of analyses to characterize CSOs and storm water.  Also, simple
       flow and  rainfall "design" criteria (or procedures)  are needed to allow rapid  and
       widespread TMDL calculations.

•      Technical  guidance to  explore and define how existing data, methods, and models can
       be used to calculate TMDLs consistent with new sediment criteria.

•      Guidance on allocation of point and nonpoint source loads.

•      Re-evaluate the Technical Support Document for Water Quality-based Toxics Control
       with respect to wet-weather issues, including  wet weather criteria, effluent and permit
       statistics, storm load whole effluent toxicity, and mixing zones.


       The following four areas were identified as  long-term needs:

•      Multiple Pollutants: There is little information about multiple pollutant interactions and
       whole effluent toxicity from multiple sources. Research is needed on additivity, and
       procedures for calculating  toxicity recurrence intervals  when all toxic parameters and
       pollutant sources are considered in an integrated fashion.

•      Biological and WET Criteria:  TMDLs have traditionally focused on specific chemical
       concentrations.  The TMDL process must be evaluated with respect to the role of whole
       effluent toxicity criteria and biological criteria.

Best  Management  Practices  (BMPs):    Comprehensive  information  on  BMP
effectiveness is lacking.   Simple models are needed to calculate the effectiveness of
structural  BMPs, and a  comprehensive review of available  performance data for
nonstructural BMPs must be compiled. In addition, the significance of ground water and
surface  water  interactions must be explored for  BMPs that  rely on infiltration for
pollutant control.

Sediment Contamination and Transport:  Technical research and guidance are needed
to evaluate the transport and fate of historically contaminated sediment.  Understanding
the dynamics  of these sediments is a  key to matching TMDL calculations to the
attainability of beneficial uses.


                                   7.  Conclusions
      The objectives of this workshop were to review the current state of the science with
respect to watershed modeling, predictive modeling of ecosystem restoration, and remote sensing
and geographic information  systems to see exactly where we  stand in face of these new
challenges; to obtain expert recommendations about which approaches and tools can be used to
help implement the 303(d) Total Maximum Daily Load  (TMDL) process over the next three to
five years; and to obtain prioritized recommendations about which tools are most promising for
development over the next five to ten years.

      Overall, workgroup discussions and subsequent recommendations covered a wide variety
of topics.  Many were highly technical  and topic specific, but there were several cross-cutting
themes that related directly to the TMDL process and integration of technology to address the
specific  needs generated by implementation of the TMDL program.


      Workgroup #1  , #4,  and #2 each indicated in their recommendations that EPA should
develop and support a core team of experts who are experienced using a particular technology
within the TMDL framework.   This idea was also promoted during the  closing Workshop
discussion. The role of this TEAM would be to provide technical support to the Regions, States,
and local governments who will be developing and implementing TMDLs. Not only would they
trouble shoot individual problems with  TMDL analysis, but this small group of experts would
continue to explore the potential use of  existing science and  technology for TMDL analyses, to
provide  immediate short-term technical guidance,  and to  develop a long-term strategy for
developing TMDLs.


      Throughout the course of the Workshop, the desire for case studies demonstrating both
the successful and unsuccessful development and/or implementation of TMDLs was discussed.
Other, more specific documentation would include case studies involving model validation and
uncertainty, -and use of predicted ecological  response to drive pollutant management decisions.
It  was considered important to build a long-term record of such case studies that would be
known of and accessible to all Federal  agencies, EPA Regions, and States.


      Most of the Workgroups discussed the need to develop technology interfaces to existing
link tools,  making their use more amenable to the TMDL process. This was an especially
relevent point among the remote  sensing, CIS, and modeling experts who need to develop a
working prototype linking CIS land use/land cover information to watershed modeling pollutant
transport and load estimates.

       Other, more programmatic conclusions were also reached and general comments made
during the course of the workshop.  They are stated below.

•      The States and EPA should make a concerted effort to integrate the TMDL process into
       the 314, 319, and 305(b) programs.

•      Section 404 permits require a certification under 401 that water quality standards will be
       met; it appears that a TMDL calculation would be needed for such actions to demonstrate
       compliance  with standards, including loadings due to erosion.

•      EPA should work more closely  with the U.S.  Department of  Agriculture Soil
       Conservation Service to institute the TMDL process in agricultural programs to better
       address nonpoint source issues.

•      Additional guidance is needed on how States can prioritize TMDL needs using available

•      EPA should identify and address disincentives  that may exist to applying the  TMDL
       process to nonpoint sources.

•      Nonpoint source best management practice guidance was recently developed in response
       to  the  Coastal  Zone  Act  Amendments  of 1990 (Proposed  Guidance  Specifying
       Management Measures for Sources of Nonpoint Pollution in Coastal Waters. U.S. EPA,
       Office of Water, May 1991.)


                 Exhibit and Demonstrations for Workshop Attendees

                  Phillip Taylor1, Tim Bondelid2, Jon Harcum', and Cindy Angell1

1   Tetra Tech, Inc., 10306 Eaton Place, Suite 340, Fairfax, Virginia 22030
2   Research Triangle Institute, 30-40 Comwallis Road, Research Triangle Park, NC 27709

Work stations with direct access to the EPA National Computer Center were available during
the workshop.  The work stations provided an opportunity for attendees  to learn more about
many national water quality data files and systems available to state and EPA programs and to
immediately see data and  graphics based on specific questions.  A number of displays  were
provided covering trend analysis,  spatial analysis of water quality in streams  and lakes, and
screening and prioritization of drainage basins using STORET water quality data. The Reach
File,  Water Quality Analysis, and STORET  Systems were  used extensively  to prepare the
displays and were accessed on the mainframe when responding to an attendee's inquiry.

Significant achievements have been made in addressing water quality problems over the past 20
years.  Despite these achievements, significant work still remains to address key environmental
issues such as nonpoint source pollution.  Over the past two years, EPA has developed program
guidance that combines existing Clean Water Act mandates (Section 303(d)) with input from
Regional and selected State representatives to develop an integrated approach for addressing
point and nonpoint source pollution problems (Guidance for Water Quality-based Decisions:
The TMDL  Process).

The development of total maximum daily loads  (TMDLs) for watersheds often involves the
evaluation of both point and nonpoint pollution sources.  As a result, data must be integrated
from numerous sources.  An approach taken by many investigators is to proceed using a staged
effort;  that is,  to compile and analyze data that are readily available and then collect additional
data to fill in the gaps.  This approach is particularly useful when targeting limited resources for
further  work,  prioritizing watersheds  for detailed  assessments,  performing  preliminary
assessments, or assembling data for transfer to a geographic information system  (CIS) or PC

While  some States have  made significant progress in building GIS  capabilities for watershed
analyses, many do not have the expertise or funding to acquire this capability within the next
few years.  An alternative is to use nationally available hydrologic and water quality data and
applications  software on EPA's mainframe. Recent and ongoing system enhancements provide
significant capability to conduct  integrated  point and  nonpoint source  assessments  on  a
watershed, regional,  or national scale, taking advantage of national data bases on stream flow,
hydrologic routing, land use, point sources, and water quality.  The purpose of this exhibit and
demonstration was to stimulate discussion on whether the EPA mainframe can meet the needs
and goals discussed at this workshop.



EPA's NCC/IBM  mainframe provides numerous tools (applications) for integrating  a wide
variety of data that are nationally available.   A detailed  listing of available conversational
procedures and data files  is presented at the end of this paper.  This exhibit demonstrated
selected tools  that can be readily used with limited local  watershed knowledge and  limited
hardware/software capabilities.  A summary of mainframe data sources and tools that may help
meet  the needs and goals of this workshop is presented  in  Table  A.I.  Access to EPA's
mainframe is summarized in Table A.2.
Table A.I.  Summary of selected data sources and tools available on the mainframe
Data  sources and tools demonstrated at
this exhibit
Other selected data sources and tools
available on the mainframe
   Permit Compliance System (PCS)
   Reach File(RFl, RF3)
   Industrial Facilities File (IFD)
   Environmental Data Display Manager (EDDM)
   Mapping and Data Display Manager (MDDM)
   Downloading to PCs (dBASE,  LOTUS 1-2-3,
   Waterbody System (WBS)
   Digital Line Graph (DLG)
   Digital Elevation Model (DEM)
   Land Use/Land Coverage
   Ground water data
   Toxic Release Inventory (TR1)
Table A.2.  Work station access to EPA's mainframe
Dial up
   PC  -  1200/2400  bps modems/XTALK,
   PC -- 2400 bps modem emulator 3270
   PC -- 4800 bps modem emulator mainframe
   APA/GDDM graphics
Direct connect (EPA and States)

•  Cluster controller — terminal or PC with coax
•  LANs - Novell netware 286/386 (EPA now has
   147 file servers connecting approximately 7,000
   work stations)

Clean Lakes System

The Clean Lakes System has been
developed to provide water quality
reports, analyses, summaries, and
graphical  displays  for lakes that
are  part  of  the  Clean  Lakes
Program.  Currently, a prototype
system  is  operating  on EPA's
mainframe   computer  at   the
National    Computer   Center
(Research Triangle Park, North
Carolina).  Figure A.I represents
many of the lakes already entered
into the location data  base.   The      Figure A. 1.  Locations of lakes in the Clean Lakes Program
Clean  Lakes  System  provides
access  to   the   Reach   File,
STORET, the Permit Compliance System, the Industrial Facilities Discharge File, and the Clean
Lakes Program data base.
                           3 MIES
 Figure A.2. Detailed 1.5 minute map using
             Reach File 3
By providing an integrated analysis tool,

•   Analysts can quickly gather information and data
    to  identify  existing  or  potential   sources  of

•   Decision makers can assess diagnostic information
    to define procedures for controlling the sources of

•  , Managers   can   assess   the   post-restoration
    improvements in  lake water quality.

Graphical  displays,   analyses,  and   reports   are
immediately available by selecting  a lake  from  the
directory listing displayed on the terminal.  Initially, a
detailed 7.5 minute quadrangle map such as Figure A.2
for Lake Minnetonka, Minnesota using Reach File 3,
would be useful to identify the location of tributaries,
water  quality  monitoring stations,  and  point source

 Figure A.3.  Location of NPDES Facilities and Water Quality
                  Monitoring Stations
   In Figure A.3, only the shoreline of
   the  selected lake is displayed along
   with   nearby   cities,   NPDES
   dischargers   (P),   ambient   stream
   monitoring  stations  (S), and  ambient
   lake  monitoring stations (L).   Data
   from  specific stations  can then  be
   displayed interactively.  In the figure
   below (Figure A.4), the user selected
   a time series  plot  of Secchi  Depth.
   Within   seconds,   the  Clean  Lakes
   System accessed data from STORET
   and graphically displayed ten years of
   data at the users terminal.
In many cases, assessing and summarizing
all of the available data at each lake can
be rather time consuming.  To support
screening level analyses, the Clean Lakes
System produces a Manager's Summary
of the  water quality monitoring data.  In
Figure A.5 the number of observations
outside  desirable  levels  for  indicator
variables are summarized for  all ambient
lake monitoring stations.
        . SECCMI DISC cntlERSI
        ZIMIHNL  27-OIJJ-02
                                            Figure A.4. Graphical display of water quality data
6.0 <• PH <• 9.0




17:31 THURSDAY. FEBRUARY 28, 1991
400 CTS/100ML

905 0

                             Figure A.5. Manager's summary report

Reach File

The Reach File provides hydrologic connectivity between geographic locations and historical data
created for the express purpose of performing hydrologic routing for modeling programs. Reach
File, Version 1 (RF1) contains over 68,000 stream reaches covering 100% of the continental US
and is  indexed to STORET, IFD, drinking water supplies, stream gages, and fish kills.
In Figure  A.6, one  watershed in western
Michigan  is depicted.    This  figure  was
created  by using the  mapping  procedure in
STORET.  In this  case, the user  requested
that   cataloging   unit   04050006,   which
corresponds to the lower reaches of the Grand
River, Michigan, be  plotted with the state
boundary as a background.  Figure A.7 is a
closer look  at  the  same  watershed using
MDDM.    Various  options  available in
MDDM  allow one to  flag nearby  cities,
dischargers, and monitoring stations. Figure
A.3  is an EDDM example with these flags
turned on  for Lake Minnetonka, Minnesota.
                                                Figure A.6. RF1 rivers in sample watershed
 lU = 0405000b
 MI = 07.89
  REflCH DflTfl
   = 0025
 r|_flT R
          Figure A.7. Cataloging unit 04050006 using MDDM
The   reader  should  also
notice the  shaded portion
of the watershed in Figure
A.7.  This subwatershed is
a user-defined boundary of
latitudes  and  longitudes.
Perhaps  this could  be  a
priority  area  within  the

Using Reach File 3 (RF3),
detailed  hydrography  is
displayed   for   the
subwatershed    (Fig-ure
A.8).   With  more  than
3,000,000   new   reaches
added to  create Reach File
3, the level of detail will
allow   more    detailed
evaluation.      RF3
information can also be

downloaded to ARC/INFO
compatible  files.   Unlike
ARC/INFO,  the   Reach
File  is  coded  to  know
which  way   the  water
flows.  As a result,  one
can   search   up    and
downstream   for   the
location  of water quality
monitoring  stations, water
supplies, and dischargers.
In Figure A.9, a two mile
by  two mile  window  is
displayed.    The  lake  is
from the western  side  of
the example subwatershed
(see  Figure  A. 8).     A
monitoring   station   is
identified as well.

By integrating many data
sources (at many different
scales),  RF3  can  support
basin   planning.      This
potential is summarized in
Figure A. 10.
                          JflTER SURVEY REPOR I

  Prea..| I   ''4-85
qi. qq
     i n Sq Mi les|
                            Jist  i n Mi les|
                               [COLOR KtT
                            ijet lands]
                            Ji tch Lana I
                       nichigan Demo Vafershed
                                            Figure A.8.  Subwatershed using RF3
                  2 Ml x  Z ninfip. it CUKNtR I SI
                                              DH FH|
       L 5m LOG
 •XD003.1 SO. MI.
                                              Figure A.9. Two-mile window


i *

State &

I ,


Planning Scenarios

Land use


wd other

                          Figure A. 10. Basin-wide planning with RF3
Trend Analysis with STORET

National (or watershed)  level trend analyses are also possible on the mainframe.  In less than
20 minutes at a work station, we performed a crude trend test on dieldrin at all USGS NASQAN
monitoring stations.  (Crude in the sense that there are more powerful and robust  trend tests
available in the literature; however, they have not been incorporated on the mainframe to date.)
Figure A. 10 depicts the  locations of all NASQAN monitoring stations  and Figure A. 11 shows
increasing or  decreasing trends (corresponding to  the plotted arrow direction) in the 85th
percentile dieldrin concentration.

Screening and Prioritization of River Basins
The upper Mississippi river drainage area was. divided into 14 basins as summarized in Table
A.3 and Figure A. 13.  STORET was  used to retrieve and  summarize  all mercury,  PCB,
phosphorus, and nitrogen data for each basin.  Figures A. 14 through A. 19 are LOTUS 1-2-3
plots that were created from the downloaded STORET  reports.   For the  purposes of finding
"hot-spots," we elected to plot the 85th  percentile concentration for each  river basin.  (Other
summary statistics  can be as  easily plotted; however, the 85th  percentile tends to avoid the
problems of erroneously high  values [outliers] as well as detection levels.) More than 10,000
observations are represented by each nutrient figure.  By reviewing these figures, one may hope
to find basins that require  more thorough evaluation. For our example, we arbitrarily elected
to review phosphorus concentrations in Rock River with more detail. Figure A.20 is a plot of
total phosphorus as a function of river mile above the confluence with the Mississippi River.
Table A.3. River basin code

 1   Mississippi Headwaters
 2   Minnesota River
 3   St. Croix
 4   Black-Root
 5   Chippewa River
 6   Maquoketa
 7   Wisconsin
 8   Iowa-Skunk-Wapsipinicon
 9   Rock River
10   Des Moines River
11   Salt
12   Upper Illinois River
13   Lower  Illinois River
14   Kaskaskia-Meramec
15   All Rivers
                                     Figure A. 13.  River basins for comparison of mercury, PCBs,
                                                    phosphorus, and nitrogen

   <  0.1

     0 0
          1  2   3
                   5  6  7  8  9  10 11  12 13  14  15

                   RIVER BASIN CODE

                                              2.5  6
                                            '•*  2

                                            1.0  O
                                            0.5  LU

                                            oo  .-
           Figure A. 14. Mercury concentration summary (rivers)
O  0.2

I-  0.1
     • o o


                                              05  -
                                               0 2

          1  2  3  1  5  S  7  9  9  10 11  12 13  14 15

                   RIVER BASIN CODE
         • TOTAL Hg   HIT ISSUE Hg

                                 (SEDIMENT  Hg
            Figure A. 15. Mercury concentration summary (lakes)

           TOTAL    PCBs
  3.0 -
   0 -
   0 -
       1  2  3  4  5  6   7  8  9 10  11  12 13  14 15
                 RIVER BASIN CODE
1 2
z 1
                  F ISH TISSUE

       Figure A. 16. PCB concentration summary (rivers)
           TOTAL    PCBs
       123456789 10  11
                 RIVER BASIN CODE

                • FISH TISSUE

                                   13  14 15
        Figure A. 17. PCB concentration summary (lakes)


  or  1.0
  Z  0.5

                       RIVER BASIN CODE

                                        IORG N
    Figure A. 18.  Phosphorus and nitrogen concentration summary (rivers)
    0 8
    a s
    0 4
    0 0
         1  2   3   4   5   6   7   9   9  10  11  12  13  14  15

                       RIVER BASIN CODE
               ITOTAL P^NH3 + NH4 B8SJNO3-N

     Figure A. 19.  Phosphorus and nitrogen concentration summary (lakes)

              TOTAL    PHOSPHORUS
                                ROCK RIVER
            1 5
         LU  0.5
              0     50     100     150     200     250    300     350

                              _»_ TOTAL P
          B5t.n PERCENT ILE POR ALL RIVER DATA FROM 1980-1991
                Figure A.20. Phosphorus concentration profile for Rock River
Facility Information

In addition to identifying problem
areas based on receiving water
quality, it may also be necessary
to examine individual dischargers.
In  Figure  A.21,  STORET  was
used  to retrieve  and plot  all
facility  locations  from  the
Industrial  Facilities  Discharge
(IFD) file that discharge to major
rivers  in the  upper Mississippi
River drainage area.
                                       Figure A.21.  NPDES facilities

In some cases, the information in
Figure A.21 is more than what is
needed.   For example, the water
quality analyst may already know
which discharge to examine.  In
this  case,  MDDM  is  used  to
display a two mile  by two mile
window   of  a  facility  on  the
Wisconsin River (Figure A.22).

Using  EDDM,   one  may  then
interactively  plot  data  from  a
facility's   discharge   monitoring
report (DMR).  Figure A.23 is a
plot of  flow, total  nonfilterable
residue, BOD5, and total ammonia
nitrogen accessed from the Permit
Compliance  System   (PCS)  via
EDDM.  (Not all  mainframe users
have clearance for this option.)
 Figure A.22. Specific facility
               O*. la caouu o» i»
                      pet C
             -ILL • tOtlST P PIPti )
              800. •> OM. ;e KG c
                      PCS CM u[(M2(t«
          OUCHS-ILl • FOKST 0 PIPti }
   «SICn*. TOI«. HOI* IL««».£ mo/L)
           PCS cm uieewaie
   Ninraccii. maun. loin, mo/i m HI
           PCS cm
   -iiL • Fo»csr p PIPC: )
                              Figure A.23. Data retrieved from PCS

Spatial Analysis of Selected jCoastaLAreas
In a similar analysis to the upper Mississippi River drainage basin example, six coastal areas
were selected for nutrient screening analysis.   Table  A.4 and Figure A.24 summarize  the
selected areas.

                               Table A.4. Coastal code

                                 109  Boston Harbor/Bay
                                 204  Delaware Bay
                                 206  Norfolk, VA
                                 206  Norfolk, VA
                                 301  Pamlico Sound, NC
                                 302  Pamlico Sound, NC
                                 310  Tampa Bay, FL
                                1203  Galveston Bay, TX
                                1204  Galveston Bay, TX
                                   Figure A.24.  Coastal map

In this case, polygons defining the boundary of the selected coastal zones were entered by the
user. The phosphorus and nitrogen 85th percentile concentration for each zone is summarized
in Figure A.25. Figure A.26 is a monitoring location map for Galveston Bay, Texas and was
generated by STORET. Figure A.27 is a two dimensional plot of pH and a surface plot of
chlorophyll a for Galveston Bay using UNIMAP.  (It looked a whole lot better in color ~ trust
                            SELECETED  COASTAL AREAS
         2 5
          2 0
      Z   1.5


      CC   1.0
                           206   208   301   302   310  1203
                             COASTAL CODE SUMMARY
                 [TOTAL PMNH3
                                            iORG  N
                 Figure A.25. Phosphorus and nitrogen concentration summary

                                                           JTOflCT IY9TCH
DF :
    Figure A.26. Monitoring locations in Galveston Bay
          Figure A.27.  Example UNIMAP output

                                             OWOW/AWPD Data Files

Drinking Water Supply File - The Drinking Water Supply File contain* information on 8,000 water supplies that utilize surface waters. Data
         covers FRDS number, utility name, city, state, basin, latitude and longitude, stream reach, population served, water volume used,
         and locations for plant, intakes, and sources.

Gage - The Stream Gage  Dau File contains information on approximately 36,000  stream gaging locations throughout the United  States.
         Information stored  includes locations of gaging streams, stream reach identification, types of data collected, frequency of data
         collection, media in which the daU are stored, identification of the collection agency, mean annual flow and 7 day/10 year flow.

City - This file contains data on 53,000 cities, towns, and villages located in the United States and its possessions.!  Information on each city
         includes the city's unique identification code; the county, state, major and minor river basin, and congressional district within which
         it is geographically located; stream reaches associated with the city; its latitude and longitude; and its census population data.

Dam - Inventory of 68,000 dams produced y U.S. Army Corps of Engineers which provides type, ownership, purpose, height, volume,  surface
         area, latitude-longitude, stream reach.

Industrial Facilities - The Industrial Facilities  Discharge File contains information on 128,000 NPDES industrial and municipal facilities (active
         and inactive) useful for environmental analyses. DaU consist! of NPDES, DUNS, and Needs A/F numbers with name, address, basin
         latitude and longitude, stream reach, flow, SIC codes, discharges type for facility and pipe level, and industrial category.  Indirect
         discharges to POTW systems are also included.

CETIS - The Complex Effluent and Toxicity  Information System contains Bioassay results for NPDES discharges loxicity tests.

ASIWPCA - Streams reported in the Americas' Clean Water - STEP report by the Associations of State and Interstate Water Pollution Control
         Administrators (ASIWPCA) covering water quality impairments for 1*972, 1982, and 1984 and indexed to the version 1 reach file.

ICAT - Industrial categories used in effluent guidelines studies are grouped with standard  industrial classification (SCI) indexes.

STORET FARM - Information on 13,000 STORET parameters indicating reporting units, media, CAS registry number, and chemical/biological

ODES - The Ocean Data Evaluation System is an extensive system of software for managing and analyzing marine environmental monitoring

PCS - The Permit Compliance System contains NPDES permit compliance, tracking and discharge monitoring reports for active permitted

Reach File - The River Reach File provides hydrologic connectivity between geographic locations and historical data  created for (he express
         purpose of performing hydrologic routing for modeling programs. The Reach File, Version 1, contains 68,000 stream reaches covering
          100% of the continental- U.S. and is indexed with STORET, IFD, drinking water supplies, stream  gages, and fishkills,  Version 3
         covers 80% of the U.S. and provides hydrologic linkages for 3.S million reaches based ion the USGSW DLG data.

STORET-BIOS - A component of STORET containing distribution, abundance, physical condition, and habitat description of aquatic organisms.
         These are integrated with the water quality file and linked to the reach file,  PCS, IFD, and Gage files.

STORET-USGS Flow - Contains daily stream observations of stream flow and miscellaneous water quality at USGS gaging stations. These data
         represent more than 695,000 water years for over 29,000 gages.

STORET - WQ - The agency's water quality system containing physical,  chemical and biological parameters.  More than 800 monitoring
         organizations have provided 175  million parametric observations from 700,000 sampling locations for surface water, ground water,
         fish tissue, and sediment. The sample locations are indexed to the reach file IFD, GAGE, and drinking water file with a PCS interface.

STORET - Tissue - Tissue sample results cover over 530 parameters including metals,  organics, and pesticides specific to species, tissue types,
         length, weight, and sex.  These data and stations are integrated with the water quality file and indexed to the reach file with  indexes
         to IFD, Gage,  drinking water files  and the PCS interface.

STORET Form 2C - Priority pollutant data reported by NPDES second round permits are in  STORET referenced by  the permit number and
         can be integrated with the water quality data and PCS data with the PCS/STORET interface.

WBS - The Watcrbody System data base contains water quality assessment information collected by states for 305(b) reporting. These data serve
         as an inventory of each state's navigable waters that have been assessed.

                           OW Mainframe Procedures Through TSO
ASIWPCA      Interactive program providing information on stream use impairment
CITY          Interactive program providing overview information on cities
DAMR                Interactive program providing information on dams in the U.S.
DFLOW        Interactive program providing assessing daily flows at gages and providing selected flow statistics
DXLIST        Interactive program providing information on dioxin
EDDM         Interactive program for graphically displaying locations of monitoring activities, PCS and WQ data
FLOW         Interactive program providing information on gage mean flow, 7Q 10 low flows, and daily flows
ICAT          Procedure to list industrial categories and related SIC codes in IFD File used by FTD
IFDPLOT      Interactive program to setup graphical displays of Facility data
IFDRET        Interactive program for generating standard tables of information from IFD
IPS5                  Interactive procedure for generating reports from STORET, PCS and IHS files
ISR                   Interactive generates selected STORET and IFD reports and D.O. model using PCS DMR
MDDM        Interactive mapping system for the Reach File, Facility, WQ and WBS data
PARM         Interactive program for providing information on STORET parameters
PATHSCAN    Retrieval of information on hydrological streampaths from NPDES  discharge locations
RCHDAT      Interactive program for retrieving reach, streamflow & discharger data
RCHRET      Interactive procedure for generating reach trace Auxfiles for plotting or  export
RPA3          Reach pollutant assessment software providing reports using STORET, PCS, TRIS and IHS data
SIC                   Interactive program to obtain SIC codes and descriptions used in the IFD File
SITEHELP     Interactive graphical and text retrieval of stream referenced  data using  IHS, STORET, & PCS
STRAUX      Procedure which generates a STORET AUXFILE for selected reach number
USE                  Interactive program for summarizing  WQAB procedure usage for given user

Info. Systems
Compendium *
Drinking Wate
Supply File
Gage, City,
Dam Files
File i
Reach File
- Dally Flow
- Fish Kill
- WQ

an lMt!
/CAT !'"•
/SS • IMt
304(1) \\ °
PARM ii'"»
ODES if"
Trace \
Trace )



                              Workshop Agenda

8:00 a.m.          Welcome & Introductory Remarks
                  Carl Myers, U.S. EPA, Deputy Director, Assessment and Watershed
                  Protection Division

                  William Diamond, U.S. EPA, Director, Standards and Applied Science
8:30 a.m.          Regional Perspective  on Water Quality-based Point  Source  &  NPS

                  Jon Grand,  Deputy Director, Water Management Division, U.S. EPA -
                  Region V
8:50 a.m.          Overview of 303(d) & Workshop Objectives

                  Bruce Newton, U.S. EPA, Chief, Watershed Branch

9:10 a.m.          The Watershed Approach from a Point Source Perspective

                  William Brandes, U.S. EPA, Chief, Water Quality and Industrial Permits

9:30 a.m.          BREAK

Section 1:  Problem Diagnosis
Nelson Thomas, Topic Chair

9:45 a.m.          Top-Down & Bottom-up Approaches for Integrated Assessment of Point
                  & Nonpoint Source Pollution

                  Kenneth L. Dicksen, Institute of Applied Sciences
10:30 a.m.         Diagnosis & Assessment Tools/Models for Determining Nutrient TMDLs

                  Norbert Jaworski, U.S. EPA ERL - Narragansett

Section 2:  Watershed Modeling
Lee Mulkey, Topic Chair

11:15 p.m.         Watershed Modeling & TMDL Assessments

                  Anthony S. Donigian, AQUA TERRA Consultants

12 noon           LUNCH

1:15 p.m.          Water Quality Diagnosis & Assessment Tools/Models

                  Eugene D. Driscoll, Woodward-Clyde Consultants

2:00 p.m.          Model Capabilities (Agricultural & Urban)

      :            Leslie L.  Shoemaker, Tetra Tech, Inc.

2:45 p.m.          BREAK

3:00 p.m.          Model Capabilities - A User Focus

                  Paul L. Freedman, Limno-Tech, Inc.

3:45 p.m.          North Carolina's Basinwide  Water Quality Management Approach to
                  Developing TMDLs

                  Ruth Swanek, North Carolina Department of Environmental Management
4:15 p.m.          Modeling Integration & Data Access Tools

                  Robert B. Ambrose, U.S. EPA ERL - Athens

5:00 p.m.          ADJOURN

5:30 to 7:00 p.m.   Informal Reception


Section 3:  Data Sources, Tools,  & Investigations
Ross Lunetta, Topic Chair

8:00 a.m.          Multistage  Remote Sensing  Data  Applications for GIS  Data Base
                  Development in Support of Nonpoint Watershed Modeling

                  Ross S. Lunetta, U.S. EPA  EMSL - Las Vegas

8:45 a.m.          GIS for Nonpoint Source Watershed Modeling Applications

                  Mason J. Hewitt, III, U.S. EPA EMSL - Las Vegas

9:30 a.m.          Grass Waterworks - An  Interface Between GIS & Models

                  Jon F. Bartholic, Institute of Water Research and Center for Remote

10:15 a.m.   BREAK

10:30 a.m.          Remote .Sensing .of. Agricultural Practices & Downstream Water Quality
                   as Influenced by Sediment from Nonpoint Sources

                   John G.  Lyon, The Ohio State University
11:15 a.m.          The Trials. Tribulations. & Successes of Using Remote Sensing. GIS and

                   Carol Russell, (formerly with)  Arizona Department of Environmental

12 noon            LUNCH
Section 4:  Predictive Modeling of Ecological Restoration
Nelson Thomas, Topic Chair

1:00 p.m.          An Overview of Ecological Assessment and Restoration Tools

                   Carl Richards, Natural Resources Research Institute

1:45 p.m.          Application of Tools for Ecological Restoration Predictive Modeling

                   James A. Gore, Austin Peay State University

2:15 p.m.          Ecosystem  Assessment  &  Restoration:   The Role  of  Modeling and

                   Michael L.  Johnson, Kansas Biological Survey

3:00 p.m.          BREAK

Workgroup Breakout Sessions

3:15 p.m.          All workgroups should address the following:

                         the state-of-the-science
                         the quality of available tools
                         the use of each tool in the TMDL process
                         limitations/needed improvements of tools

                         recommendations for research and development
                         recommendations for technical guidance

                   1.  Watershed Modeling

                         Chair:          Lee Mulkey, U.S. EPA - Athens
                         Facilitator:      James Giattina, U.S. EPA - Region V
                         Recorder:       David Dilks, Limno-Tech

                   2.  Data Sources. Tools. & Investigations

                         Chair:          Ross Lunetta, U.S. EPA - Las Vegas
                         Facilitator:      Mason Hewitt, U.S. EPA -  Las Vegas
                         Recorder:       Michael McCarthy, RTI
                   3.  Predictive Modeling of Ecological Restoration

                         Chair:          Nelson Thomas, U.S. EPA - Duluth
                         Facilitator:      Ron Carlson, U.S. EPA - Duluth
                         Recorder:       Amy Sosin, U.S. EPA - HQ

                   4.  Selected Point Source Issues

                         Chair:          Elizabeth Southerland, U.S. EPA - HQ
                         Facilitator:      Bruce Zander, U.S. EPA - Region VIII
                         Recorder:       Paul Freedman,  Limno-Tech
5:00 p.m.           ADJOURN


8:00 a.m.           Continuation of Workgroup Breakout Sessions

12 noon            LUNCH

1:00 p.m.           Watershed Modeling - Workgroup Summary

1:20 p.m.           Data Sources. Tools.  & Investigations - Workgroup Summary

1:40 p.m.          Predictive Modeling of Ecological Restoration - Workgroup Summary

2:00 p.m.          Selected Point Source Issues - Workgroup Summary

2:20 p.m.          BREAK

2:30 p.m.          Plenary Discussion - Priorities
                  Bruce Zander, U.S. EPA Region VIII - Moderator

3:15 p.m.          Wrap-up:  Where Do We Go From Here?

                  Bruce Newton, U.S. EPA, Chief, Watershed Branch - Moderator

3:30 p.m.          ADJOURN

                                    List of Participants
Robert Ambrose
U.S. Environmental Protection Agency
Center for Exposure Assessment Modeling
960 College Station Road
Athens, GA  30613
(404) 546-3130, FTS/250-3130, FAX/546-3340

John Anagost
CIS Coordinator - Region V
U.S. Environmental Protection Agency
230 South Dearborn Street
Chicago,  IL  60604
(312)886-0143, FAX/886-1420

Samuel F. Atkinson
Institute of Applied Science
Northern  Texas State University
P.O. Box  13078
Denton, TX  7620-3078
(817) 565-2145

Jon Bartholic
Director,  Institute of Water Research
334 Natural Resources Building
Michigan State University
East Lansing, MI  48824
(517) 353-3742, FAX/353-1812

Robert P.  Baumgartner
OR Department of Environmental Quality
811 S.W. 6th Avenue
Portland,  OR  97223
(503) 229-5877

Douglas Beyerlein
Snohomish County Department of Public Works
County Administration Building, 5th Floor
Everett, WA  98201
(206) 388-3464

Hiranmay Biswas
U.S. Environmental Protection Agency
Office of Science and Technology (SASD)
401 M Street, S.W.
Washington, DC   20460
(202) 260-7012, FTS/260-7012, FAX/260-7024

Tim Bondelid
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC  27709
(919) 541-6796, FAX/541-5945

D. King Boynton
U.S. Environmental Protection Agency
Office of Science and Technology
401 M Street, S.W. (WH-553)
Washington, DC  20460
(202) 260-7013, FAX/260-7024

Donald J. Brady
U.S. Environmental Protection Agency
Assessment and Watershed Protection Division
401 M Street, S.W. (WH-553)
Washington, DC  20460
(202) 260-5392, FTS/260-5392, FAX/260-5394

William (Rick) Brandes
U.S. Environmental Protection Agency
Permits  Division
401 M Street, S.W. (EN-336)
Washington, DC  20460
(202) 260-9537, FTS/260-9537

John Burt
U.S. Department of Agriculture
Soil Conservation Service
P.O. Box 65670
Washington, DC  20013-2890
(202) 245-5008

A. Ron Carlson
U.S. Environmental Protection Agency
Environmental Research Laboratory
6201 Congdon Avenue
Duluth,  MN  55804
(218) 720-5523, FTS/780-5523

Alan Cavacas
Tetra Tech, Inc.
10306 Eaton Place, Suite 340
Fairfax, VA  22030
(703) 385-6000, FAX/385-6007

Bruce Cleland
TMDL Coordinator - Region X
U.S. Environmental Protection Agency
1200 Sixth Avenue
Seattle, WA  98101
(206) 553-2600, FTS/399-2600, FAX/553-0119

William Cooler
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC   27709
(919) 541-5918, FAX/541-7155

Clayton Creager
Western Aquatics, Inc.
Executive Park,  Suite 220
1920 Highway 54
Durham, NC 27713
(919) 544-9454, FAX/544-9453

Joel Cross
IL Environmental Protection Agency
2200 Churchill
Springfield, IL   62704
(217) 782-3362

Mimi Dannel
TMDL Coordinator - Region VI
U.S. Environmental Protection Agency
1445 Ross Avenue (6W-QT 6DT)
Dallas, TX   75202-2733
(214) 655-7145, FTS/255-7145, FAX/655-6490

Thomas Davenport
Nonpoint Source Coordinator - Region V
U.S. Environmental Protection Agency
2305 South Dearborn Street (5WQS TUB-8)
Chicago, IL 60604   .
(312) 886-0209, FTS/886-0209, FAX/886-1420

Wayne Davis
Monitoring Coordinator - Region V
U.S. Environmental Protection Agency
230 South Dearborn Street
Chicago, IL 60604
(312)386-6233, FTS/886-6233

Edward Dettman
U.S. Environmental Protection Agency
Environmental Research Laboratory
South Ferry Road
Narragansett, RI 02881
(401) 782-3039, FTS/838-6000, FAX/782-3030

William R.  Diamond
Director, Standards and Applied Science Division
U.S. Environmental Protection Agency
401 M  Street, S.W. (WH-585)
Washington, DC 20460
(202) 260-7301, FTS/260-7301

Ken Dickson
Director, Institute of Applied Sciences
North Texas State University
P.O. Box 13078
Denton, TX  76203
(817) 565-2228

David Dilks
Limno-Tech, Inc.
2395 Huron Parkway
Ann Arbor, MI  48104
(313) 973-8300, FAX/973-1069

Anthony Donigian
AQUA TERRA Consultants
2672 Bayshore  Parkway, Suite 1001
Mountain View, CA  94043
(415) 962-1864, FAX/964-2027

Eugene D. Driscoll
Woodward Clyde Consultants
101 Manito Avenue
Oakland, NJ 07436
(201) 337-2217, FAX/337-2218

Ian Droppo
National Water Research Institute
687 Lakeshore  Road
P.O. Box 5050
Burlington, Ontario
(416) 336-4701, FAX/336-4989

Wildon Fontenot
U.S. Department of Agriculture
Soil Conservation Service
South Agricultural Building
Washington, DC  20250
(202) 475-5250, FAX/447-2646

Paul Freedman
Limno-Tech, Inc.
2395 Huron Parkway
Ann Arbor, MI  48104
(313) 973-8300, FAX/973-1069

Jeff Gagler
U..S. Environmental Protection Agency
Region V (WQS TUB-8)
230 South Dearborne Street
Chicago,  IL  60604
(312) 886-6679

Sandy Germann
U.S. Environmental Protection Agency
Office of Wetlands, Oceans and Watersheds
401 M. Street, S.W. (WH-556F)
Washington DC 20460
(202) 260-7118, FAX/260-6294

James Giattina
U.S. Environmental Protection Agency
Region V (SWQS)
230 South Dearborn Street
Chicago, IL   60604
(312) 886-0139, FTS/886-1420

Gary Goay
LA Department of Environmental Quality
Office of Water Resources
P.O. Box 82215
Baton Rouge, LA  70884-2215
(504) 765-0550, FAX/765-0635

Mike Goggin
U.S.D.A. Forest Service at
U.S. Environmental Protection Agency
401 M Street, S.W. (WH-551)
Washington, DC  20460
(202) 260-7010, FAX/260-7024

James Gore .
Center for Field Biology
Austin Peay State University
P.O. Box 4718
Clarksville, TN  37044

Jon Grand
Deputy Director, Water  Division - Region V
U.S. Environmental Protection Agency
230 South Dearborn Street
Chicago, IL   60604
(312) 353-2000, FTS/353-2000

Ted Gray
Northeastern IL Planning Commission
400 West Madison Street
Chicago, IL   60606
(312) 454-0400, FAX/454-0411

Jim Greenfield
TMDL Coordinator - Region IV
U.S. Environmental Protection Agency
Water Division
345 Courtland Street, N.E.
Atlanta, GA   30365
(404) 347-2126, FTS/257-2126, FAX/347-3269

David Greenwood
LA Department of Environmental Quality
7920 Bluebonnet Road
Third Floor Engineering
Baton Rouge, LA  70884-2215
(504) 765-0557, FAX/765-0635

Robert Gumtow
WY Department of Environmental Quality
Water Quality Division
Herschler Building, 4 West
Cheyenne, WY  82002
(307) 777-7098

Warren Harper
U.S. Department of Agriculture
Forest Service
P.O. Box 96090
Washington,  DC  20090
(202) 453-9475, FAX/382-6066

Edwin E. Herricks
3215 Newmark Civil Engineering Lab
University of Illinois
205 North Mathews Avenue
Urbana, !L  61801
(217) 333-0997, FAX/359-9593

Mason Hewitt
U.S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
P.O. Box 93478
Las Vegas, NV  89108
(702) 798-2377, FTS/798-2692

John M. Higgins
Tennessee Valley Authority
270 Haney Building 2C-C
1101 Market Street
Chattanooga, TN  37343
(615) 751-7299, FAX/751-7479

John Houlihan
TMDL Coordinator - Region VII
U.S. Environmental Protection Agency
726 Minnesota Avenue
Kansas City, KS  66101
(913) 551-7475, FTS/276-7432

Tom Howard
CA State Water  Resources Control Board
901 P Street
Sacramento, CA  95814
(916) 324-7970
Robert K. Hubbard
U.S.D.A. Agricultural Research Services
Southeast Watershed Research Laboratory
P.O. Box 946
Tifton, GA   31794

Wayne Jackson
U.S. Environmental Protection Agency
Region II
26 Federal Plaza
New York,  NY   10278
(212) 264-5685, FAX/264-2194

William James
School of Engineering
University of Guelph
Guelph, Ontario
(519) 824-4120, FAX/836-0227

Norbert Jaworski
U.S. Environmental Protection Agency
Environmental Research Laboratory
27 Tarzwell Drive
Narragansett, RI  02882
(401) 782-3000, FTS/782-3030

Dan Jaynes
U.S. Department of Agriculture
Agricultural Research  Service NSTL
2150 Pammel Drive
Ames,  IA   50011
(515) 294-8243, FAX/294-8125

Marshall Jennings
U.S. Geological Survey
8011 Cameron Road, Building 1
Austin, TX   78753
(512) 832-5791

Michael Johnson
KS Biological Survey
2041 Constant Avenue
Lawrence, KS  66047
(913) 864-7724, FAX/864-7724

Don Josif
U.S. Environmental Protection Agency
Region V, GISMO
536 South Clark
Chicago,  IL   6060S
(312) 886-0838

Dean Knighton
U.S. Department of Agriculture
Forest Service
201 14th  Street, S.W.
Washington DC  20250
(202) 453-9524, FAX/382-0530

Bill Kooter
Research  Triangle Institute
3040 Comwallis Road
Research  Triangle Park, NC   27709
(919) 541-6796, FAX/541-5945

Mike Kuehn
U.S. Department of Agriculture
Forest Service
Federal Office Building
P.O. Box 21628
Juneau, AK  99802-1628
(907) 586-7864, FTS/871-7864, FAX/586-7843

Dan Lawson
U.S. Environmental Protection Agency
Region V (5WQS-TUB-8)
230 South Dearborn Street
Chicago,  IL   60604
(312) 886-3017, FAX/886-1420

Sue Laufer
Tetra Tech,. Inc.
10306 Eaton Place,  Suite 340
Fairfax, VA   22030
(703) 385-6000, FAX/385-6007

Larry Lehrman
U.S. Environmental Protection Agency
Region V (GISMO)
536 South Clark
Chicago,  IL   60605
(312) 886-0836

Ed Liu
Monitoring Coordinator - Region IX
U.S. Environmental Protection Agency
1235 Mission Street
San Francisco, CA  94103
(415) 744-2006, FTS/484-2006

Eric H. Livingston
PL  Department of Environmental Regulation
Nonpoint Source Management Section
2600 Blair Stone Road
Tallahassee, FL  32301
(904) 488-0782, FAX/488-6579

Donald Luman
Geography Department
118 Davis Hall
Northern  Illinois University
DeKalb, IL  60115-2854
(815) 753-6840, FAX/753-0198

Ross S. Lunetta
U.S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
P.O. Box 93478
Las Vegas, NV   89193-3478
(702) 798-2175, FTS/545-2175, FAX/798-2692

Robert Lynch
OK Conservation Commission
2800 North Lincoln Boulevard, Suite 160
Oklahoma City, OK  73105
(405) 521-2384

John Lyon
Civil Engineering Department
The Ohio State University
2070 Neil Avenue
Columbus, OH  43210
(614) 292-6039

Michael M'Carthy
Research  Triangle Institute
P.O. Box 12194
Research  Triangle Park, NC 27709
(919) 541-6796, FAX/541-5945

Carl Meyers
Deputy Director, AWPD
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington,  DC  20460

Peggy Michell
U.S. Environmental Protection Agency
Assessment and Watershed Protection Division
401 M Street, S.W. (WH-553)
Washington, DC   20460
(202) 260-5378, FAX/260-7024

Lee Mulkey
U.S. Environmental Protection Agency
Environmental  Research Laboratory
College Station Road
Athens, GA 30613
(404) 546-3358, FTS/250-3129

Romy Myszka
U.S. Environmental Protection Agency
Region V (GLNPO 5GL)
230 South Dearborn Street
Chicago,  IL 60605

Bruce Newton
Chief, Watershed Branch
U.S. Environmental Protection Agency
Assessment and Watershed Protection Division
401 M Street, S.W. (WH-553)
Washington, DC   20460
(202) 260-7076, FTS/260-7074, FAX/260-7024

Rosella O'Connor
TMDL Coordinator - Region II
U.S. Environmental Protection Agency
26 Federal Plaza
New York, NY  10278
(212) 264-2059, FTS/264-8479

Robert Owen
CO Department of Health
Water Quality Control Division
4210 East llth Avenue
Denver, CO  80227

James Pagenkopf
Tetra Tech, Inc.
10306 Eaton Place, Suite 340
Fairfax, VA  22030
(703) 385-6000, FAX/385-6007

William G. Painter
U.S. Environmental Protection Agency
Office of Policy Analysis
401 M  Street, S.W. (PM-221)
Washington, DC  20460
(202) 260-5498, FTS/260-4332

Jeanna  M. Paluzzi
Division Compliance Coordinator
Wayne County Division of Public Works
415 Clifford
Detroit, MI  48226
(313) 224-3634, FAX/224-0045

Brad Parks
U.S. Environmental Protection Agency
ERL Large Lakes Research Station
9311 Groh Road
Grosse He, MI  48138
(313) 692-7633, FAX/692-7603, FTS/692-7503

Marianne Pellegro
U.S. Environmental Protection Agency
Region V
111 West Jackson, Suite 900
Chicago, IL   60604
(312) 886-1764

Robert Pepin
TMDL Coordinator - Region V
U.S. Environmental Protection Agency
230 South Dearborn Street
Chicago, IL   60604
(312) 886-1505, FTS/886-1420

Spence Peterson
U.S. Environmental Protection Agency
Environmental Research Laboratory
200 S.W.  35th Street
Corvallis,  OR   97333
(503) 757-4457, FTS/420-4457, FAX/757-4338

Dave Pfeifer
U.S. Environmental Protection Agency
Region V (5WQS TUB-8)
230 South Dearborn Street
Chicago IL  60604
(312) 353-9024

David Pincumbe
TMDL Coordinator - Region I
U.S. Environmental Protection Agency
John F. Kennedy Federal Building
Boston, MA  02203
(617) 565-3546, FTS/835-3546

Jerry Pitt
U.S. Environmental Protection Agency
726 Minnesota Avenue
Kansas City, KS  66101
(913) 551-7766, FAX/551-7765

Carl Richards
Natural Resources Research Institute
University of Minnesota .
5013 Miller Trunk Highway
Duluth, MN  55811

William Richardson
U.S. Environmental Protection Agency
ERL Large Lakes Research Station
9311 Groh Road
Grosse He, MI  48138
(313) 692-7611, FTS/378-7611, FAX/692-7603

Walter Rittall
U.S. Department of Agriculture
Soil Conservation Service
South Agriculture Building, Rm. 6036
Washington, DC  20250

Lew Rossman
U.S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
26 West Martin Luther King  Drive
Cincinnati, OH  45268
(513) 569-7603, FTS/684-7603

Carol Russell
17866 West Lunnonhaus Drive, #12
Golden, CO   80401
(303) 279-9614

Stephen Schmelling
U.S. Environmental Protection Agency
Robert S. Kerr ERL
P.O. Box 1198
Ada, OK  74820
(405) 332-8800, FTS/743-2334, FAX/332-8800

Leslie Shoemaker
Tetra Tech, Inc.
10306 Eaton Place, Suite 340
Fairfax, VA  22030
(703) 385-6000, FAX/385-6007

Amy Sosin
U.S. Environmental Protection Agency
Assessment and Watershed Protection Division
401 M Street, S.W. (WH-553)
Washington, DC   20460
(202) 260-7058, FAX/260-7024

Elizabeth Southerland
Chief, Risk Assessment & Management Branch
U.S. Environmental Protection Agency
Office of Science & Technology
401 M Street, S.W.
Washington, DC   20460

Nancy Sullivan
U.S. Environmental Protection Agency
Region I
John F. Kennedy Federal Bldg.,  WQP425
Boston, MA  02203
(617) 565-3546, FTS/835-3546,  FAX/565-4940

Irene Suzukida
U.S. Environmental Protection Agency
Assessment and Watershed Protection Division
401 M Street, S.W. (WH-595)
Washington, DC   20460
(202) 260-7059, FAX/260-7024

Peter Swenson
U.S. Environmental Protection Agency
Region V
230 South Dearborn Street
Chicago IL  60604
(312) 886-0209

Ruth Swanek
NC Division of Environmental Management
P.O. Box 27687
Raleigh, NC  27611
(919) 733-5083, FAX/733-9919

Ed Swanson
AZ Department of Environmental Quality
2655 East Magnolia, Suite 2
Phoenix, AZ  85034
(602) 392-4043, FAX/392-4017

William Swietlik
U.S. Environmental Protection Agency
401 M Street, S.W. (EN-336)
Washington, DC   20460

Phill Taylor
Tetra Tech, Inc.
10306 Eaton Place
Fairfax, VA  22030
(703) 385-6000, FAX/385-6007

Mike Terstriep
IL State Water Survey
2204 Griffith Drive
Champaign, IL  61820
(217) 333-4959

Nelson Thomas
U.S. Environmental Protection Agency
Environmental Research Laboratory
6201 Congdon Boulevard
Duluth,  MN  55804
(218) 720-5702, FTS/785-5702, FAX/720-5539

Donald Weatherbe
1352 Safeway Crescent
Mississauga, Ontario
(416) 896-4759, FAX/846-7954

Louise Wise
U.S. Environmental Protection Agency
401 M Street, S.W. (WH-556F)
Washington DC  20460
(202) 260-7166, FAX/260-6294

Bruce Zander
TMDL Regional Coordinator - Region VIII
U.S. Environmental Protection Agency
Water Division
999  18th Street, Suite 500
Denver, CO  80202-2406
(303) 293-1580, FTS/293-1580