Practical_Guide_Eflows_for_Policy_(Meitzen)

A PRACTICAL GUIDE TO
ENVIRONMENTAL FLOWS FOR
POLICY AND PLANNING
WITH NINE CASE STUDIES IN THE UNITED STATES
Eloise Kendy, Colin Apse, and Kristen Blann
with selected case studies by Mark P. Smith and Alisa Richardson
m a y 2 0 1 2
Autumn view of the Roaring Fork river in the western slope area of Colorado. © Bill Grindle/TNC
2
CONTENTS
Acknowledgments .......................................................................................................................................................................3
Foreword ............................................................................................................................................................................................4
Executive Summary .....................................................................................................................................................................5
1.0 Introduction ...........................................................................................................................................................................7
1.1 Background .....................................................................................................................................................................7
1.2 Purpose .............................................................................................................................................................................8
1.3 Structure and Scope of this Report ........................................................................................................................9
2.0 Case Studies .......................................................................................................................................................................10
2.1 Michigan’s Water Withdrawal Assessment Process .....................................................................................13
2.2 Massachusetts Sustainable Water Management Initiative ........................................................................18
2.3 Connecticut River Basin Ecosystem Flow Restoration .................................................................................23
2.4 Colorado Watershed Flow Evaluation Tool .......................................................................................................26
2.5 Connecticut Statewide Environmental Flow Regulations ...........................................................................29
2.6 Middle Potomac River Basin Environmentally Sustainable Flows ...........................................................33
2.7 Susquehanna River Basin Ecosystem Flow Recommendations ...............................................................38
2.8 Ohio Thresholds for Ecological Flow Protection.............................................................................................42
2.9 Rhode Island Stream Depletion Method ...........................................................................................................45
3.0 Getting Environmental Flows to Scale: An Overview of the Process ..................................................50
3.1 Understanding Water Availability: Building a Hydrologic Foundation ...................................................50
3.1.1 What is the Hydrologic Foundation? ......................................................................................................50
3.1.2 Criteria for Hydrologic Model Selection ...............................................................................................51
3.1.3 Components of the Hydrologic Foundation ........................................................................................51
3.1.4 General Observations and Summary .....................................................................................................53
3.2 Classifying River Types.............................................................................................................................................55
3.2.1 Why Classify River Types? ..........................................................................................................................55
3.2.2 General Approaches to Classification ...................................................................................................55
3.2.3 Parameters Used for Classification ........................................................................................................56
3.2.4 General Observations ..................................................................................................................................56
3.3 Describing Flow-Ecology Relationships .............................................................................................................57
3.3.1 Hypothesis Development ...........................................................................................................................57
3.3.2 Quantitative Analysis ...................................................................................................................................58
3.3.3 Hybrid Approaches .......................................................................................................................................59
3.3.4 General Observations ..................................................................................................................................60
3.4 Making Flow-Ecology Relationships Operational:
Applying Environmental Flow Science at a Regional Scale .......................................................................60
3.4.1 Establishing Ecological Condition Goals and Defining Acceptable Risk .................................61
3.4.2 Implementation: Putting Flow Standards into Practice ...................................................................62
3.4.3 General Observations ..................................................................................................................................64
4.0 Conclusions .........................................................................................................................................................................65
5.0 References............................................................................................................................................................................67
3
This compilation would have been impossible without the forward thinking and hard work of many dedicated individuals.
LeRoy Poff, Angela Arthington, Brian Richter, Bob Naiman, Stuart Bunn, Mike Acreman, Brian Bledsoe, Mary Freeman,
Ken Rogers, Julian Olden, Rebecca Tharme, Dave Merritt, Jay O’Keefe, Rob Jacobson, Jim Henriksen, Jonathan Kennen,
and Andy Warner knew that environmental flow protection had to be accelerated, and had the vision to scale up what
previously had been a river-by-river endeavor to a practical framework for nationwide protection. Many of them, along with
Michele DePhilip, John Sanderson, Sally Palmer, Judy Dunscomb, John Stark, Kim Lutz, Julie Zimmerman, Thomas Wilding,
Nicole Rowan, Dave Armstrong, Stacey Archfield, Rodney Knight, Stephanie Flack, Claire Buchanan, Heidi Moltz, Paul
Seelbach, Troy Zorn, Rob Burgholzer, Carlton Haywood, Tara Moberg, Dave Hamilton, Josh Knight, Mark P. Smith, David
Sutherland, Chris Konrad, Rick Palmer, Austin Polebitski, John Hickey, Chris Hatfield, Rich Bowman, Todd Richards, and
more brave, trusting souls accepted their challenge and led the first applications of this framework on the ground. We want
to thank them not only for having the courage, resourcefulness, diplomacy, and tenacity to see these ambitious projects
through, but also for sharing with us the lessons they learned along the way. Most of all, we applaud their pivotal roles in
protecting and restoring environmental flows at a geographic scale never before imaginable.
The clarity and accuracy of this report greatly benefited from constructive reviews of earlier drafts by Kim Lutz, Michele
DePhilip, and Doug Shaw of The Nature Conservancy; Rick Eades of Nebraska Game and Parks Commission; Mike
Robertson of Wyoming Game and Fish Department; Mary Davis of the Southeast Aquatic Resources Partnership; Jackie
King of Water Matters; Alisa Richardson of the Rhode Island Department of Environmental Management; and Jennifer
Phelan of the Research Triangle Institute International.
Acknowledgments
4
Freshwater scientists around the world acknowledge a crisis in biodiversity loss and diminished ecosystem function in streams,
rivers, wetlands and lakes. One main culprit in this decline is human-caused change in the historical magnitude, frequency
and timing of river flows that have supported key ecological processes and native species in diverse aquatic systems.
The Ecological Limits of Hydrologic Alteration (ELOHA) framework arose as a response to the recognition that the rate
of global hydrologic change was outpacing science’s ability to develop environmental flow guidelines one site at a time.
A new scientific framework was needed, one to guide the development of environmental flow guidelines at a regional scale.
Regional environmental flow management is extremely challenging. Not only are data needs great for underlying hydro-
ecological models, but translation of science into policy and management necessarily occurs in a complex societal context
constrained by governance structures, regulatory authorities and competing political interests.
A comprehensive and scientifically sound framework was required, yet it had to be flexible enough to accommodate the
inevitable “experiments” that will result from its application to real-world settings.
This guidebook presents nine case studies in regional environmental flow management in the United States. The successes
described in these case studies clearly illustrate that innovative thinking and creative experimentation within the structured
ELOHA framework have significantly advanced the development of flow standards at regional scale. The accomplishments
are the fruit of the dedicated efforts of innumerable agency, academic and non-government scientists, engineers, water
managers, and policy makers, all of whom were willing to think outside the box and take up the challenge of pushing the
frontiers of sustainable environmental management of streams and rivers. Their stories are impressive and inspiring!
A key contribution of this guidebook is that it distills an excellent synthesis across the case studies and offers valuable
insight into transferable lessons learned from individual studies. Importantly, numerous tips and extended discourse are
offered on how to adapt the flexible ELOHA framework to streamline and strengthen the scientific process. While a more
robust social framework may still be needed to effectively translate science into policy and regulations, the information
presented in this guidebook serves as an indispensible resource for all those engaged in protection and management of
freshwater ecosystems. The authors of the guidebook are to be commended for their outstanding product, and more
generally for their dedication to the cause of science-based freshwater conservation.
N. LeRoy Poff, Ph.D.
Professor of Biology and Graduate Degree Program in Ecology
Colorado State University
May 10, 2012
Foreword
5
Environmental flows are gaining broad recognition across
the United States, and the underlying science is sufficiently
developed to support regional planning and policy
applications. This report explores how six states and three
interstate river basins are effectively developing and
applying regionalized environmental flow criteria to water
resource planning, water withdrawal permitting, and
multi-dam re-operation. The broad range of approaches
they relate clearly demonstrates the feasibility of integrating
science-based environmental flow needs into regional water
management in the absence of site-specific assessments.
The case studies embrace the following principles:
• Regionalized environmental flow criteria apply to all the
water bodies across a state or large river basin for which
site-specific criteria have not yet been established.
• Flow criteria link explicitly to the health of the entire
aquatic and riparian ecosystem, and are not limited to
specific species.
• Flow regimes mimic natural inter- and intra-annual flow
variability.
• The development of environmental flow criteria and the
policies for their implementation are closely linked.
Defining a clear path to policy implementation from the
onset ensures that the ensuing science answers the right
management questions.
• Flow criteria are developed through a transparent,
inclusive social process informed by sound science. A
structured social process for identifying, understanding,
and negotiating tradeoffs is critical.
To varying degrees, the Ecological Limits of Hydrologic
Alteration (ELOHA) framework influenced each of the case
studies. ELOHA is a flexible framework for determining and
implementing environmental flows at the regional scale
using existing hydrologic and biological information. Major
components of ELOHA include a hydrologic foundation of
streamflow data, classification of natural river types,
flow-ecology relationships associated with each river type,
and river condition goals.
A hydrologic foundation of streamflow data informs
water managers where, when, and how much water occurs
in all water bodies across the region. These data are used
to assess flow characteristics, classify river types, quantify
flow alteration, relate ecological responses to flow
alteration, and evaluate the status of sites relative to
environmental flow standards. They should be estimated
for every stream segment or sub-basin where
environmental flows will be managed. Without this
information, limits on flow alteration are very difficult to
permit, measure, or enforce without expensive site-specific
data collection.
Although a database of daily streamflow under naturalized
and current conditions remains an ideal, the lack of these
data need not hinder the determination of environmental
flow needs. In practice, few places have such a dataset in
place at the onset, and building it usually requires
considerable time and thought. Many water managers have
successfully advanced other parts of the ELOHA framework
while the hydrologic foundation is being developed, rather
than awaiting its completion before proceeding with
successive steps.
The basic components of a hydrologic foundation of daily
or monthly streamflow data are hydrologic simulation and
water use accounting. Hydrologic simulation is used to
estimate streamflow conditions, while water use accounting
estimates the impact of water use on streamflow conditions.
Two general approaches to hydrologic simulation are
regression modeling and process modeling.
Regardless of the modeling approach, the hydrologic
foundation can only be as accurate as the water-use
(withdrawal and return flow) data that go into it. Accurate,
spatially explicit water use reporting and improved water
use estimation methods and decision support systems are
greatly needed.
The treatment of interactions between groundwater and
surface water depends on the type of hydrologic model
used and on the hydrogeology. The approaches reported in
our case studies range from assuming direct and immediate
impacts of groundwater pumping on streamflow to using
linked groundwater and surface-water models to calculate
the time, place, and amount of depletion.
River Type classification can strengthen the statistical
significance of flow-ecology relationships based on rivers
that have been studied, and extends those relationships to
other rivers that have not been studied. However, recent
practice demonstrates that natural river type classification
is not always needed for setting scientifically defensible
environmental flow standards.
Where it has been used, river type classification tends to be
iterative and based on pre-existing classifications. Using an
Executive Summary
6
existing classification system helps build trust among
scientists who are familiar with it and accelerates the
overall process. By treating natural river type classification
and flow-ecology analyses as iterative processes, each one
strengthens the explanatory power of the other. Parameters
used to classify river types include hydrology, water
temperature, ecoregion, and watershed and macrohabitat
characteristics. Incorporating watershed and macrohabitat
variables from the onset is useful for assigning river types
to ungaged sites.
Flow-ecology relationships generalize the tradeoffs
between flow alteration and ecological condition for
different types of rivers. Successful projects follow a
progression from hypothesis development to data assembly
and analysis to build these relationships.
Several case studies used statistical analyses to isolate the
influence of flow alteration from that of other
environmental stressors, and then to identify the flow and
ecological metrics that best describe ecological response
to flow alteration. However, although strictly exploratory
data analysis may result in robust statistical relationships, if
the metrics used do not resonate with biologists and water
managers, then the results may be ineffective in supporting
environmental flow policy.
Even where large biological databases exist, rarely will
quantitative flow-ecology analyses represent all flow-
dependent taxonomic groups and ecological processes.
A holistic approach requires a combination of quantitative
and qualitative relationships to represent the ecosystems
and natural flow regimes across a state or basin. Structured
ecological literature review, historical streamflow analysis,
and facilitated expert workshops can build scientific
consensus around qualitative relationships of sufficient
rigor to quantify flow criteria.
Condition goals, also known as desired future condition,
underlie all environmental standards. Within any large
region, people value different rivers for different purposes.
They may strive to keep certain rivers nearly pristine, while
accepting that other, more developed rivers simply maintain
a basic level of ecological function. Stakeholder
negotiations that focus on applying different standards to
rivers with different ecological condition goals generally
lead to successful implementation.
The environmental flow standard is the degree of allowable
flow alteration associated with each condition goal,
according to the flow-ecology relationships. An ecological
risk-based framework associates ecological goals with
allowable flow alteration, and accounts for scientific
uncertainty by associating appropriate policy actions with
different levels of ecological risk. Decision support systems
greatly facilitate the integration of environmental flow
standards into state water allocation programs and
regional water resource planning.
Although ELOHA provides a useful framework to guide the
development of environmental flow criteria, there is no
one-size-fits-all approach. Regardless of the rigor of the
scientific analyses, expert judgment calls are required to
adapt the process and to interpret results. Facilitated expert
workshops and advisory panels improve the outcomes and
their credibility to the public. The level of sophistication
needed depends ultimately on what policy makers want and
stakeholders will accept. Early outreach to stakeholders,
the definition of shared guiding principles, the formation of
advisory committees, attention to political and economic
drivers, and process transparency are all likely to pay
dividends in environmental flows implementation.
7
1.1 Background
After decades of complacency, water managers not only in
the arid West, but increasingly in the humid eastern United
States, face chronic water shortages. Irrigation is on the
rise. Growing urban and suburban populations need secure
water supplies. Demands for domestic energy herald
hydropower intensification. Climate change, deteriorating
infrastructure, and shortsighted land developments
increase flood risk. At the same time, there is a rising
expectation that water resource development should not
degrade freshwater ecosystems (Acreman 2001, Postel
and Richter 2003).
Globally, flow alteration is among the most serious threats
to freshwater ecosystems. Natural, seasonal patterns of
rising and falling water levels shape aquatic and riparian
habitats, provide cues for migration and spawning,
distribute seeds and foster their growth, and enable rivers,
lakes, wetlands, and estuaries to function properly (Bunn
and Arthington 2002, Poff et al. 1997). Altering the natural
flow pattern—by damming, diverting or channeling
water—takes a serious toll on the plants, animals, and
people that depend on it. But how much change is too
much? When does “change” become “degradation” or
“unacceptable adverse impact”? These are the questions
that have spawned the interdisciplinary fields of
environmental flow science and management.
Environmental flows describe the timing and amount of
water to be retained in lakes, rivers, streams, and estuaries
to sustain seasonal patterns of high and low water levels
needed for natural functions, processes and resilience to
persist. While all natural flow provides some environmental
benefit, the need to allocate a portion of this water to meet
society’s needs for water supply, crop production, energy
generation, and flood management requires careful
evaluation and integration of competing uses.
For decades, environmental flow quantification has been
conducted at the scale of individual river reaches, with a
range of potential methodologies used to evaluate flow
requirements. Holistic methodologies that account for all
flow-dependent ecosystem needs are well-established
(Tharme 2003), but can take years to complete for just one
river reach. A more systematic approach applied at a
watershed, region, or state-wide scale is required if
freshwater ecosystem protection and recovery are to match
the pace and extent of water resource development.
Ultimately, this necessitates a scaling-up from site-by-site
environmental flow provisions to the state, provincial, or
national policy realm (Le Quesne et al. 2010). Only in this
way will environmental flows become integral to all water
management decisions from the onset, and not just as an
inconvenient afterthought.
Regionalizing environmental flow management means
making decisions that minimize ecological impacts of new
water developments, direct water development to least-
sensitive water bodies, and prioritize flow restoration
efforts. These decisions hinge on a scientific understanding
of how changes in the natural flow regime affect ecological
conditions.
The Ecological Limits of Hydrologic Alteration framework
(ELOHA; Poff et al. 2010) helps water managers meet this
challenge. ELOHA is a flexible framework for determining
and implementing environmental flows at the regional
scale using existing hydrologic and biological information.
ELOHA was developed specifically to meet the needs of
managing environmental flows through state, provincial,
basin, or national water policy: it addresses many rivers
simultaneously; explicitly links flow and ecology; and
applies across a spectrum of flow alteration, data
availability, scientific capacity, and social and political
contexts (Poff et al. 2010).
The ELOHA framework rests on the premise that although
every river is unique, many exhibit similar ecological
responses to flow alteration. Furthermore, within every river
type, or group of ecologically similar rivers, there exist
individual rivers under various degrees of hydrologic and
resulting ecological alteration. If, for example, within a group
of similar rivers, percent of water withdrawn in August is
plotted against the ecological condition of the remaining
fish populations, then a flow-fish ecology relationship can
be quantified for that type of river. ELOHA assumes that
this relationship holds for all rivers of that type.
The steps of ELOHA (Figure 1.1) may be carried out in a
number of different ways, ranging from professional
judgment to sophisticated statistical modelling, depending
on the available data and technical capacity. Likewise, the
sequence of steps is flexible. As later sections of this report
will detail, hydrologic modeling is used to create a
hydrologic foundation of streamflow data for every
stream segment in the project region. Stream
classification can be used to group the segments into
ecologically similar natural river types. Hydro-ecological
1.0 Introduction
8
analyses and/or structured scientific workshops generate
flow-ecology relationships (Figure 1.2) for different types of
rivers. This scientific process is depicted by the blue and
green boxes in Figure 1.1.
ELOHA’s “social process” (orange boxes in Figure 1.1) is
typically a policy-making effort driven by a new law,
regulatory rulemaking, or policy guidance. Such a process
is almost always led by a governmental authority and often
closely involves stakeholders. The determination of
“acceptable ecological conditions”, or ecological condition
goals, for river and stream segments has similarities to
many state water quality programs that classify water
bodies according to water quality attainment goals. This
goal classification is distinct from the scientific
classification of natural river types mentioned above.
Flow-ecology relationships link these socially-determined
ecological condition goals to differing degrees of
hydrologic alteration, enabling the establishment of
streamflow standards or criteria to meet different goals.
Thus, a river with a “good” ecological goal would be
required to meet a higher standard (less hydrologic
alteration) than the same type of river with a “fair”
ecological goal. The terms “standards” and “criteria”, as
used in this document, mean limits on hydrologic alteration
designed to achieve a set of management goals through a
planning or regulatory structure.
1.2 Purpose
Several authors have proposed conceptual frameworks
(Arthington et al. 2006, Poff et al. 2010, Richter et al. 2011)
as outlined above for establishing environmental flow
criteria for all waters across large jurisdictions, but stopped
short of fully applying the concepts themselves. Instead,
they challenged water managers to adapt the frameworks
to their own states, provinces, countries, and transboundary
river basins. River scientists and water managers from
around the world have expressed interest in adopting these
concepts, but most felt they needed more practical
guidance and on-the-ground examples to proceed.
This report is intended to fill that gap. The title of this report
notwithstanding, its main purpose is not to guide
practitioners stepwise through a structured process, but
rather to explain how others have adapted a general
Figure 1.1. Steps of the ELOHA framework (Poff et al. 2010).
9
Figure 1.2 Conceptual flow-ecology curves showing possible forms
of the relationship. A: linear, B: threshold, C: curvilinear. The graph
represents one river type. After Davies and Jackson (2006).
framework to different situations, and to provide references
for obtaining more in-depth information. By recounting the
experiences of nine states and river basins, this report
demonstrates a broad range of approaches for establishing
and implementing environmental flow criteria at the scale
needed to support regional water resource planning, water
withdrawal permitting, and multi-dam re-operation.
1.3 Structure and Scope of this Report
The main content of this report appears in the following two
sections. Section 2 presents case studies that illustrate how
six states and three interstate river basins are effectively
developing and applying regionalized environmental flow
criteria to water policy and planning. Because they are
written with the intent of standing alone, each case study
has its own, separate list of references and figure and table
numbering. Section 3 walks through the fundamental steps
of developing such a process. For each step, various
options are presented, referring to examples from the case
studies. Section 4 concludes with general guidelines,
reflecting on lessons learned from the case studies.
Each case study first explains its legal context for
environmental flow management, then outlines the project
management structure, steps through the relevant scientific
analyses conducted and tools developed, and culminates
with an account of how the results are being or will be
used for water management. While Poff et al. (2010) were
moderately prescriptive about the scientific process
supporting environmental flow management, they were
rather vague about the social process. Several case studies
begin to fill this gap by describing the sequence of
negotiations and stakeholder input that translated the
science into agreed-upon flow targets and institutional
procedures for achieving them.
10
The case studies described here span a broad range of
regional-scale approaches to environmental flow
management among completed or nearly completed
projects. The intent of the case study descriptions is to
convey the breadth of social and scientific processes
undertaken, and to provide references for readers to pursue
more detailed accounts of scientific methods and models.
We selected these case studies because to our knowledge
they illustrate the most advanced integration of
regionalized environmental flow science and policy in the
United States. They represent diverse approaches applied
over a range of geographic areas, from the 2,400-km2
Fountain Creek pilot basin in Colorado to the entire
254,000-km2 State of Michigan.
Of the case studies reported, however, only Michigan,
Rhode Island and Connecticut have fully translated
environmental flow criteria into statewide water management
programs. Michigan began its process many years before
the others. All of the case studies describe completed or
nearly completed scientific assessments and some degree
of implementation at the policy or planning level.
While reporting on partially completed projects is not ideal,
the immediate need for practical guidance compels us to
distribute this comprehensive report now. The broad range
of approaches it relates clearly demonstrates the feasibility
of integrating science-based environmental flow needs into
regional water management in the absence of site-specific
assessments. But the entire process from start to finish
takes time, and this report provides ample guidance for
new projects to launch while ongoing projects continue to
lead the way to implementation.
Table 2.1 summarizes how each project built a hydrologic
foundation, classified river types, related flow alteration to
ecological response, and applied (or is applying) the
science outcomes to water management policy. The
ELOHA framework influenced all of these projects from the
onset, except for Michigan, which evolved with the
framework. Rhode Island and Connecticut adopted
principles from ELOHA without systematically following the
framework. Massachusetts; Colorado; Ohio; and the
Susquehanna, Middle Potomac, and Connecticut River
basin projects were designed explicitly around ELOHA.
Although other applications of the ELOHA framework are
underway, the case studies profiled here were selected
because of their successful positioning at the intersection
between science and policy. In each case, scientists,
stakeholders, and decision makers worked together to
ensure that the scientific work supported a specific policy
need. Consequently, these case studies trace selected
environmental flow policy applications from their initiation
to their implementation of environmental flow management
across large geographic areas.
Most of our case studies describe the development of
biologically-based flow criteria for managing water
withdrawals. These applications take place in the eastern
US, where a few state governments have set flow standards
that reserve water for the environment; for example, by
incorporating streamflow protection into new water
withdrawal permit programs. In the eastern states where
government lacks that authority, policy reform to grant it is
feasible—if difficult—under the riparian doctrine of water
management. In these states, water resources are relatively
abundant and generally only over-allocated locally, but are
increasingly pressured by population growth and climate
change. Therefore, the focus of regional environmental flow
management tends to be on streamflow protection, rather
than on large-scale restoration, with notable exceptions
(e.g., see Connecticut Basin case study).
Principles Guiding Selection
of the Case Studies
• Development of environmental flow criteria and
policies for their implementation are closely linked.
• Regionalized environmental flow criteria apply to
all the water bodies across a state or large river
basin for which site-specific criteria have not yet
been established.
• Flow criteria link explicitly to the health of the
entire aquatic and riparian ecosystem, and are
not limited to specific species.
• Flow regimes mimic natural inter- and intra-
annual flow variability.
• Flow criteria are developed through a
transparent, inclusive social process informed
by sound science.
2.0 Case Studies
11
Table 2.1. Comparison between major process components of five state and three interstate river basin case studies described in this report.
Case Study Hydrologic Foundation River Type
Classification
Flow-Ecology
Relationships
Application
Michigan
(253,793 km2)
Median August flow, based on
multiple linear regression;
streamflow depletion model
(STRMDPL) estimates
groundwater pumping delay
11 by water
temperature,
catchment area
Fish community-flow models
based on large database of
fish species occurrence
Online tool for
screening proposed
water withdrawals in
relation to adverse
impact standard
Ohio
(116,096 km2)
Mean September flow based on
generalized least-squares
regression
5 by Aquatic Life
Use
Fish community-flow curves
using quantile regression
based on large fish and
habitat databases
Future thresholds for
permitting water
withdrawals
Massachusetts
(27,336 km2)
Daily flow based on duration-
curve regression model and water
accounting (Sustainable Yield
Estimator, SYE)
Not used to date Fish community-flow
alteration curves using
quantile regression and
generalized linear models
based on large fish
database
Environmental flow
criteria for permitting
new and existing water
withdrawals
Colorado
(3,700-km2 and
2,400-km2 pilot
basins)
Daily water accounting (Colorado
StateMod)
3 by ecoregion Fish, invertebrates,
vegetation, recreation
response to various flow
metrics using various
approaches based on data
found in literature
Risk-mapping tool for
water use planning
Connecticut
(14,357 km2)
Daily flow based on duration-
curve regression model and water
accounting (SYE)
Not used Flow needs of state river
species by bioperiod per
technical committee
recommendation
Statewide reservoir
release rules
Middle Potomac
River interstate
basin
(11,500 km2)
Daily flow based on process
model (HSPF), channel
morphology, flow routing, water
accounting, and non-linear
ground-water recession in a
Watershed Online Object
Oriented Meta-Model
(WOOOMM)
Not used; flow
and ecology
metrics
normalized to
account for
natural variability
Benthic invertebrate
response to 18 flow metrics
using quantile regression
based on large invertebrates
database
Interstate land and
water use planning;
potential to inform
water withdrawal
permitting in individual
states
Susquehanna
River interstate
basin
(71,000 km2)
Daily flow from minimally-altered
index gages; daily streamflow
estimator tool based on
duration-curve regression model
and water accounting (SYE)
5 by water
temperature,
catchment area,
hydrology
Nineteen hypotheses
relating various taxa and
ecological processes to flow
components, based on
literature review and expert
workshops
Water withdrawal
standards and dam
operations regulated
by Susquehanna River
Basin Commission
Connecticut River
interstate basin
(19,288 km2)
Daily flow based on duration-
curve regression model and water
accounting (SYE) and dam
operations model
Not used Conceptual models of full
range of taxa and flow
components, based on
literature review and expert
workshops
Collaborative decision
support tool to
integrate and optimize
operations of >60
dams
Rhode Island
(2,706 km2)
Regression-based (for 7Q10)Not used Fish-flow relations from
Georgia, supported by local
data analyses
Wetland permitting
and statewide
planning
12
Western states, in contrast, manage water under the prior
appropriation doctrine, and water users commonly have
appropriated most of the natural streamflow during much
of the year. In these states, it may be too late to reserve
water for nature; instead, environmental flow policy
initiatives focus on flow restoration by enabling and
financing senior water right transfers from offstream to
instream flows. These legal transactions require significant
commitments of time and money. Regional environmental
flow assessment helps prioritize these transactions.
Understanding environmental flow needs at a regional
scale opens up opportunities for efficiency by identifying
a suite of flow re-allocation, reservoir re-operation, and
conjunctive management strategies, and helps target
limited resources on acquiring the right amount of water at
the right times of year in the right places to gain the most
ecological benefit. The Colorado case study describes the
development of a decision support tool that starts
stakeholders down this path. The Connecticut River case
study, though not Western, also produced a decision
support tool for regional water management—in this case,
it is the integration of dam operations throughout the basin
to optimize both social and ecological outcomes.
13
The Michigan Water Withdrawal Assessment process
demonstrates an effective science-policy process with
user-friendly decision tools developed to support it.
Michigan’s process was established and fully implemented
before Poff et al. (2010) published their paper, yet in many
ways parallels the ELOHA framework. Hamilton and
Seelbach (2011), Ruswick et al. (2010) and Steinman et al.
(2011) provide detailed overviews of the entire scientific
and policy development.
A series of interstate compacts and Michigan water
management laws initially spawned the process. Annex
2001 to the Great Lakes Charter, ratified in 2008 in the
Great Lakes Compact, stipulates that signatory states may
cause no significant adverse individual or cumulative
impacts on the quantity and quality of the Waters and
Water-Dependent Natural Resources of the Great Lakes
Basin. Signatory states further commit to:
• establish programs to manage and regulate new or
increased withdrawals;
• implement effective mechanisms for decision making
and dispute resolution;
• develop mechanisms by which individual and cumulative
impacts of water withdrawals can be assessed; and
• improve the sources and applications of scientific
information regarding Waters of the Great Lakes Basin
and the impacts of withdrawals from various locations
and water sources on the ecosystems.
Michigan’s 2006 water law defined “Adverse Resource
Impact” as one that functionally impairs the ability of a
stream to support characteristic fish populations.
Occupying the top of the food chain, these fish are seen as
biological indicators of the overall health of Michigan’s
rivers and streams. The law also committed the state to
create an integrated assessment model to determine the
potential for any proposed water withdrawal to adversely
impact the state’s waters and water-dependent resources.
An Advisory Council composed of industry, advocacy, NGO,
agency, and academic stakeholders was convened and
given a 1-year timeline and strong bipartisan support to
recommend a process to the Michigan legislature to carry
out this mandate. The Council developed and operated
under Guiding Principles (see box), to which its success is
largely attributed. These Principles focused Council
members on their common interests, regardless of their
other differences. The process recommended by the Council
(Michigan Groundwater Conservation Advisory Council
2007) ultimately was adopted into state law (2008 Public
Act 189). The Michigan Department of Environmental
Quality was the primary implementing agency.
The first technical step was to delineate stream segments
for subsequent analysis and management of environmental
flows. Michigan’s 30,000 National Hydrography Database
Plus (NHD+) river reaches were grouped into about 6,800
segments believed to have characteristic and relatively
homogeneous hydrology, geomorphology, hydraulics, water
quality, water temperature, and biological attributes with
fish assemblages that are distinct from neighboring
segments (Brenden et al. 2008). Reviews by field scientists
further aggregated the number of stream segments to
about 5,400 for subsequent analysis.
Michigan’s hydrologic foundation is a database of the
median daily flow for the month of lowest summer flow
(typically August) for each stream segment. This can be
thought of as the typical low flow during the relatively dry
summer months. This “Index Flow” was chosen because it
represents the most ecologically stressful period of the
year. The amount of water that can be withdrawn is
expressed as a percent of Index Flow, as suggested by
Richter (2009). Multiple linear regression using landscape
and climate characteristics (aquifer transmissivity, forest
cover, average annual precipitation, and soil permeability)
was used to estimate the Index Flow for all ungaged stream
segments (Hamilton et al. 2008). Acknowledging model
uncertainty, these estimates were then adjusted by a
“safety factor” to ensure that estimated flow exceeds actual
flow only 10% of the time, further protecting rivers from
Adverse Resource Impacts due to excessive withdrawals.
In Michigan, groundwater discharge plays a significant role
in determining fish species assemblage. During the
summer low flow period, groundwater discharge into
Michigan’s rivers provides most of their flow, and regulates
their temperature and dissolved oxygen levels.
Groundwater withdrawals by pumping wells reduce natural
groundwater discharge to rivers. To account for
groundwater withdrawals, a computer model estimates
streamflow depletion from the nearest stream segments for
any proposed withdrawal based on well location, depth,
aquifer and riverbed characteristics, and the timing and
quantity of withdrawal (Reeves 2008, Reeves et al. 2009).
Michigan’s stream segments were classified according to
catchment size (streams, small rivers, large rivers) and
2.1 Michigan’s Water Withdrawal Assessment Process
14
Michigan Guiding Principles
March 20, 2007
1. Michigan has an abundance of water resources. There is no overall shortage of water in the State. Currently,
water withdrawals in Michigan do not present a crisis.
2. Not all water withdrawals are alike, and have differing levels and types of impacts. Certain water sources can
support a large amount of withdrawal without harm to other users or to the ecosystem. Other water sources
are more vulnerable to large withdrawals.
3. Some areas of the state have been identified as sensitive to groundwater withdrawals. Current and future
withdrawals in these areas require a higher degree of monitoring, scientific research, and understanding.
4. Water is a valuable asset, and if used efficiently, can provide the basis of a strong economy and high quality of
life in Michigan.
5. Ground and surface water are strongly interrelated and cannot be viewed as separate and distinct.
6. In order to protect basic ecological function, adequate stream base flow must be maintained.
7. Water use by type of user or by purpose of use is not prioritized.
8. The amount of water withdrawn from a hydrologic system must be sustainable. Water resource sustainability
involves the use of scientific analysis to balance the economic, social and environmental demands placed on
the resource to ensure that the needs of current and future generations are not compromised by current usage.
9. Indicators of sustainability are important to assessing Michigan’s water use.
10. The accuracy and effectiveness of water management is an evolutionary, long-term process that must be
continually enhanced with scientific information. Additional monitoring of stream flows, water levels, aquatic
ecosystems, and related mapping and analysis is essential to protecting water resources.
11. Any water management process must be consistent with applicable statutory and common law in Michigan,
neither abrogating nor expanding the law absent specific legislative action.
12. Consistency of regulation and predictability between state and local units of government are essential to
managing the resource.
13. Education is critical for all water users, private and public, to understand their responsibilities for water
conservation and efficient use.
14. Local, voluntary problem-solving approaches for resolving water use disputes and withdrawal impacts are the
desirable starting point for conflict resolution. Michigan has a role in disputes involving impacts on
environmentally sensitive areas. Legal action by any party should be seen as the last option.
15. Withdrawals presenting the greatest risk of causing an adverse impact to natural resources should be the
primary focus of a water management process.
16. Information gathered and provided for the purpose of preliminary evaluation of water withdrawal projects
must be simple and understandable in the most accurate and represented manner possible.
17. Mitigation of adverse resource impacts is a reasonable alternative for new and expanding water withdrawals
where deemed appropriate.
18. Conservation of water resources includes the efficient use and protection of quality.
19. Preliminary evaluation of potential adverse resource impacts on fish populations and other existing water
users caused by new water withdrawal must have value to new and existing water users, is important prior to
significant economic investment and is critical to determining the need for further analysis.
20. The goals of a water use assessment tool are to provide a better understanding of withdrawal impacts, to
minimize water use conflicts, to facilitate water planning among stakeholders, and to assess long-term
conservation strategies.
15
thermal regime (cold, cold transitional, warm transitional
and warm)—the dominant variables previously shown to
influence fish assemblages in Michigan (Lyons et al. 2009,
Wehrly et al. 2003, Zorn et al. 2008). This classification
yielded 11 river types (Brenden et al. 2008, Seelbach et al.
2006), which were mapped onto the Michigan NHD+
stream segment data layer.
For each of the 11 river types, Zorn et al. (2009) modeled
fish response curves that relate population and density
changes in fish communities to percentage reductions in
Index Flow (Figure 1). The curves are based on a
representative subset of samples collected from about
1,700 locations over 30 years (about one sample per year
per site from about 20 sites per river type) and housed in
three databases. Curves for thriving species (those
expected to be especially abundant) can be considered
“early warning flags” of Adverse Resource Impact, which
the legislature defined in terms of characteristic species
(expected to be more abundant than the state mean
abundance). Michigan’s ecological response curves are
unique because they summarize in a single model the
response of the entire fish community to flow alteration in
a given river type. Other taxonomic groups (invertebrates,
vegetation, etc.) were not assessed.
To compensate for uncertainties in the models, the 2008
Michigan law created “management zones” representing
increasing levels of risk to the environment (Figure 1), and
prescribed a suite of water management actions for each
level. Because the curve for each river type is different, the
flow removal associated with a given change in fish
assemblage—and therefore the boundaries between
management zones—differs by river type (Figure 2).
Prospective water users employ an online Water Withdrawal
Assessment Tool (WWAT; Michigan Department of
Environmental Quality 2009) to determine the level of risk
associated with their proposed withdrawals. Users enter
the location, timing, quantity, and if relevant, the screen
depth of their proposed groundwater or surface water
withdrawals. Using the hydrologic foundation and
groundwater model, the WWAT calculates flow depletion of
the nearest stream segment during summer low flow due
to the proposed withdrawal, added to the cumulative
withdrawals from upstream segments. Using the stream
types and fish response curves, the WWAT associates that
depletion with its risk level for that type of river. If the risk
level is low, then the withdrawal may be registered online
with no further analysis. If the risk level is high, meaning
the withdrawal would likely cause an Adverse Resource
Impact, then site-specific review by Department of
Environmental Quality staff is required, using local flow and
fish data and expert opinion instead of the less accurate
Figure 1. Typical fish-response curves. ARI indicates Adverse
Resource Impact, depicted here as 90% of characteristic fish species
remaining, as recommended by the Advisory Committee. Light lines
indicate thresholds between water management zones associated
with different degrees of ecological change. A = register water use,
B = notify local water users, C = form a water user committee.
statewide model. After site review, the withdrawal will be
registered, registered with modifications, or rejected.
Outcomes of the process are:
• Withdrawals are capped at the volume that risks
adversely impacting fish communities during the
most-sensitive time of year. This volume applies all
year long. Therefore, flow variability is maintained
and low-flow thresholds are avoided.
• New withdrawals registration is expedited when
environmental risk is low.
• Government staff time focuses on withdrawals that
pose the most risk and stream segments that are most
highly valued by society (because anyone can request
a site review).
• Future water withdrawals will likely be taken from
least-sensitive rivers.
ACKNOWLEDGMENT
The authors thank Paul Seelbach of the U.S. Geological
Survey for describing this project to us and reviewing this
case study.
REFERENCES
Brenden, T. O., Wang, L., and Seelbach, P. W. 2008. A river
valley segment classification of Michigan streams based on
fish and physical attributes. Transactions of the American
Fisheries Society 137:1621-1636. http://www.tandfonline.
com/doi/abs/10.1577/T07-166.1.
16
Hamilton, D. A., and Seelbach, P. W. 2011. Michigan’s water
withdrawal assessment process and internet screening
tool. Michigan Department of Natural Resources, Fisheries
Special Report 55, Lansing.
Hamilton, D. A., Sorrell, R. C., and Holtschlag, D. J. 2008.
A regression model for computing index flows describing
the median flow for the summer month of lowest flow in
Michigan. U.S. Geological Survey Scientific Investigations
Report 2008-5096 http://pubs.usgs.gov/sir/2008/5096/.
Lyons, J., Zorn, T., Stewart, J., Seelbach, P., Wehrly, K., and
Wang, L. 2009. Defining and characterizing coolwater
streams and their fish assemblages in Michigan and
Wisconsin, USA. North American Journal of Fisheries
Management 29:1130-1151. DOI: 10.1577/M08-118.1.
Michigan Department of Environmental Quality. 2009.
Water Withdrawal Assessment Tool http://www.miwwat.org/.
Michigan Groundwater Conservation Advisory Council.
2007. Report to the Michigan Legislature in response to
Public Act 34, 37 p. http://www.michigan.gov/documents/
deq/Groundwater_report_206809_7.pdf.
Poff, N. L., Richter, B. D., Arthington, A. H., Bunn, S. E.,
Naiman, R. J., Kendy, E., Acreman, M., Apse, C., Bledsoe, B.
P., Freeman, M. C., Henriksen, J., Jacobson, R. B., Kennen, J.
G., Merritt, D. M., O’Keeffe, J. H., Olden, J. D., Rogers, K.,
Tharme, R. E., and Warner, A. 2010. The ecological limits of
hydrologic alteration (ELOHA): a new framework for
developing regional environmental flow standards.
Freshwater Biology 55:147-170. http://www3.interscience.
wiley.com/cgi-bin/fulltext/122588390/PDFSTART.
Reeves, H. W. 2008. STRMDEPL08 — An extended version
of STRMDEPL with additional analytical solutions to
calculate streamflow depletion by nearby pumping wells.
U.S. Geological Survey Open-File Report 2008-1166, 22 p.
http://pubs.usgs.gov/of/2008/1166/.
Figure 2. Michigan’s fish response curves, showing how each type of river has different curves, and therefore different water withdrawals
associated with each management zone. X- and y-axis labels are as shown in Figure 1. From Zorn et al. (2008).
17
Reeves, H. W., Hamilton, D. A., Seelbach, P. W., and Asher,
A. J. 2009. Ground-water component of the Michigan Water
Withdrawal Screening Tool. U.S. Geological Survey
Scientific Investigations Report 2009-5003 http://pubs.usgs.
gov/sir/2009/5003/pdf/sir2009-5003_web.pdf.
Richter, B. D. 2009. Re-thinking environmental flows:
from allocations and reserves to sustainability boundaries.
River Research and Applications 22(8):1052-63. DOI:
10.1002/rra.1320.
Ruswick, F., Allan, J., Hamilton, D., and Seelbach, P. 2010.
The Michigan Water Withdrawal Assessment process:
science and collaboration in sustainaing renewable natural
resources. Renewable Resources Journal 26:13-18.
Seelbach, P. W., Wiley, M. H., Baker, M. E., and Wehrly, K. E.
2006. Landscape-based identification and classification of
ecological river segments: concepts, approach, and
application across Michigan’s Lower Peninsula. Pages
18-38 in R. Hughes, L. Wang, and P. W. Seelbach, eds.
Landscape influences on stream habitats and biological
communities, American Fisheries Society.
Steinman, A. D., Nicholas, J. R., Seelbach, P. W., Allan, J. W.,
and Ruswick, F. 2011. Science as a fundamental framework
for shaping policy discussions regarding the use of
groundwater in the State of Michigan: a case study. Water
Policy 13:69-86.
Wehrly, K. E., Wiley, M. H., and Seelbach, P. W. 2003.
Classifying regional variation in the thermal regime based
on stream fish community patterns. Transactions of the
American Fisheries Society 132:18-38.
Zorn, T. G., Seelbach, P. W., Rutherford, E. S., Wills, T. C.,
Cheng, S.-T., and Wiley, M. J. 2008. A regional-scale habitat
suitability model to assess the effects of flow reduction on
fish assemblages in Michigan streams. Michigan
Department of Natural Resources, Fisheries Research
Report 2089, Ann Arbor, Michigan, 46 pp. http://www.
michigan.gov/documents/dnr/RR2089_268570_7.pdf.
18
The Massachusetts Sustainable Water Management
Initiative demonstrates the use of (1) a duration-curve
regression approach to build a hydrologic foundation,
(2) bioperiods as a temporal basis for setting flow criteria,
(3) quantitative flow-ecology response curves to inform
decision-making, and (4) a management framework that
associates implementation actions with different condition
goals. It is a work in progress.
Responding to water quality and quantity concerns, the
1987 Massachusetts Water Management Act (WMA)
established a water withdrawal permitting system. Twenty
years later, implementation of the Act was falling short of
its objectives, as evidenced by persistent impacts from
stream depletion. Consequently, environmental groups
appealed permit decisions for not adequately protecting
rivers and streams from excessive water withdrawals, and
filed legislation requiring the development of environmental
flow protection standards. In 2009, responding to
continuing controversy, the state1 launched the
Massachusetts Sustainable Water Management Initiative
(SWMI). Both the social and scientific processes of SWMI
closely follow the ELOHA framework.
An Advisory Committee representing water suppliers,
conservationists, agriculture, state agencies, and other
stakeholders was established to develop a comprehensive
approach to water management, including water
withdrawals. A Technical Committee representing the same
stakeholders and state and Federal agencies was formed to
help inform and scientifically ground this effort. To date,
these committees have met formally many dozens of times
over the course of two years to design and carry out the
criteria development process.
The hydrologic foundation is the Massachusetts
Sustainable-Yield Estimator (SYE), a statewide, interactive
decision-support tool developed by the U.S. Geological
Survey (USGS) (Archfield et al. 2010). SYE first estimates
the 1960-2004 series of unregulated (baseline), daily
streamflow at ungaged sites using a duration-curve
regression approach (Figure 1). Quantile regression is used
to estimate the flow-duration curve for the ungaged site,
based on climate and physical parameters. A minimally
altered reference gage is then selected systematically as
described by Archfield et al. (2010) and used to transform
the flow-duration curve into a daily time series of baseline
flows. Armstrong et al.’s (2004) analyses of streamflow and
fish populations at the reference sites confirmed that they
are minimally altered.
Current-condition flows are calculated by adding and
subtracting water withdrawal and return flow data provided
by the Massachusetts Department of Environmental
Protection (DEP) for the period 2000-2004. Monthly water
use data are divided evenly by the number of days in the
2.2 Massachusetts Sustainable Water Management Initiative
1 Massachusetts Department of Conservation and Recreation (MDCR), Massachusetts Department of Environmental Protection (MDEP), and the Massachusetts Department of Fish and Game (MDFG)
Figure 1. Duration-curve regression approach used to estimate baseline daily flow series in the Massachusetts Sustainable Yield Estimator
(SYE). From Massachusetts Executive Office of Energy and Environmental Affairs (2012).
19
month; however, the user has the option of substituting
more accurate information. Stream depletion due to
groundwater withdrawals is assumed to occur
instantaneously with the withdrawal. Alternatively,
STRMDEPL (Barlow 2000) may be used to distribute the
depletion over time, given well locations and basic aquifer
characteristics. Weiskel et al. (2010) added several
enhancements to SYE, including distributed flow models to
simulate groundwater discharge into rivers in certain parts
of the state. Additional, detailed dam operation data are
needed to simulate reservoir storage (Archfield et al. 2010).
The Technical Committee of stakeholders, guided by state
resource agencies, identified four seasonal bioperiods
necessary to support life histories and biological needs of
resident fish communities and fluvial-dependent diadromous
species: overwintering and salmonid egg development,
spring flooding, rearing and growth, and fall salmonid
spawning. The Technical Committee confirmed that January,
April, August, and October adequately represent the four
bioperiods to allow for simplified criteria development.
Weiskel et al. (2010) delineated 1,395 nested,
topographically defined sub-basins draining to National
Hydrography Database (NHD) stream reaches. For each
sub-basin, current-condition flows were calculated as
described above, adding back in estimated domestic well
withdrawals and septic system discharges. This allowed for
the calculation of a range of flow statistics, including the
baseline and current-condition median flow during January,
April, August, and October. Flow alteration was calculated
by comparing baseline to current-condition data.
Flow-ecology relations are described in two USGS
Scientific Investigation Reports by Armstrong et al. (2010.
2011), the more recent of which used data from 669
fish-sampling sites in the Massachusetts Division of
Fisheries and Wildlife fish-community database. Literature
review guided the selection of a set of flow-sensitive fish
metrics, including species richness, abundance of
individual indicator species, and abundance of species
grouped based on life history. A set of environmental
variables was calculated for the contributed watershed
area for fish sampling sites, including land cover (e.g., %
buffer in wetland), land use (e.g., % impervious cover), and
fragmentation (e.g., dam density) variables. In addition,
dozens of streamflow alteration variables were calculated
for all the fish sampling sites, using the general approach
of Weiskel et al (2010).
Most of the streamflow alteration variables are monthly
median flow statistics, and a subset of those was calculated
by separating out different types of withdrawals and return
flows. For example, a variable was created for the percent
alteration of August median flow due to groundwater
withdrawals (excluding return flows and other types of
withdrawals). Estimated indicators of flow alteration were
found to be highly correlated among the twelve months of
the year. Ordination and cluster analysis facilitated the
grouping of fish species into habitat use classifications
such as fluvial specialist.
The number of environmental variables in the study was
reduced using principal components analysis and
Spearman rank correlation. It was found that estimated
August median flow alteration resulting from groundwater
withdrawals was both significant and not highly correlated
with other non-streamflow variables in the model. Once
this step was completed, quantile regression (Cade and
Richards 2005) and generalized linear modeling were used
to quantify the response of stream fish to the reduced set
of environmental variables.
As illustrated in Figure 2, quantile regression revealed that
increases in the percent alteration of August median flow
from groundwater withdrawal was associated with
decreases in the 90th quantile for relative abundance of
brook trout and blacknose dace (both classified as “fluvial
fish”). A similar declining pattern relative to increasing flow
alteration was seen for fluvial fish relative abundance and
fluvial fish species richness.
The generalized linear modeling (GLM) approach led to a
small set of somewhat strong models for describing the
relations between fish-response and environmental/
anthropogenic variables. The study found that, relative to
eight chemical and physical covariates, diminished flow
magnitudes are primary predictors of biological integrity
for fish and wildlife communities. The August median flow
alteration due to groundwater withdrawals variable is
significant in the subset of GLM equations describing the
relationship between fluvial-fish relative abundance (as
measured by catch per unit effort) and environmental
factors. In other words, the key resulting equation revealed
that a “1 percent increase in percent alteration of August
median streamflow from groundwater withdrawals is
associated with a 0.9 percent decrease in fluvial-fish
relative abundance” if other variables in the model, such
as impervious surface, are held constant (Armstrong et al.
2011). A graphic of this relationship is shown in Figure 3.
As part of the Massachusetts Sustainable Water
Management Initiative Technical Committee process, the
Massachusetts Department of Fish and Game (DFG) led a
process of categorizing the current condition of the state’s
flowing waters using fish community metrics as a surrogate
for ecological integrity. More specifically, DFG used the
GLM and quantile regression results to assign ranges of
alteration in fluvial fish relative abundance values to five
condition classes, or “Biological Categories.” This approach
20
required use of the analyses described above, along with
best professional judgment, and was built upon the
biological condition gradient concept (Davies and Jackson
2006). These condition classes were designed to reflect the
combined effects of withdrawals and impervious surfaces
on fluvial fish relative abundance, and are in tiers. As a
result, all sub-basins in the state can be placed into a
Biological Category based on known natural watershed
characteristics along with estimates of impervious surface
and percent alteration of the August median flow due to
groundwater withdrawals.
A set of draft streamflow criteria (limits on hydrologic
alteration) and a larger water management framework have
been proposed by Massachusetts resource agencies for
comment by participants in the SWMI process. The draft
streamflow criteria (Table 1) were developed by using the
GLM equation for fluvial-fish relative abundance in
Armstrong et al (2011), defining the amount of flow
alteration that corresponds with the boundaries between
Biological Categories when impervious surface is held
constant at 1% (as agreed to by the SWMI Technical
Committee). Specifically:
• 3% alteration of August median flow leads to a shift from
Biological Category 1 to 2;
• 10% alteration leads to a shift from Category 2 to 3;
• 25% alteration leads to a shift from Category 3 to 4; and
• 55% alteration leads to a shift from Category 4 to 5.
For any river or stream segment, the set of streamflow
criteria that would apply is based on that water body’s
existing level of flow alteration due to groundwater
withdrawal (i.e. “Flow Level”). Those water bodies with
relatively low existing hydrologic alteration would have
stricter criteria apply in August than those that have more
significant current use of groundwater.
Figure 2. Examples of flow-ecology curves for Massachusetts showing quantile-regression relationships between fish-community metrics and
August flow depletion. From Armstrong et al. (2011).
Figure 3. Generalized linear modeling result linking alteration in
August median flow from groundwater withdrawal to fluvial fish
relative abundance. From Armstrong et al. (2011).
21
Draft streamflow criteria also were proposed for October,
January, and April, in order to protect the overall pattern of
the natural flow regime across the year. Each of these
months is linked to a biological need in Massachusetts
rivers and streams through the bioperiod concept. Based
on best professional judgment, and lacking a GLM
equation for these months, the non-summer criteria were
generally set at one “Flow Level” below (more stringent
than) each of the August flow criteria. October’s flow
criteria are somewhat less strict, recognizing that flow
alteration across the state during the fall is larger than
during winter or spring. Given the relatively significant
impacts to fluvial fish communities—and by proxy
ecological integrity—from August alterations due to
groundwater withdrawals greater than 25 percent, no
quantitative criteria were proposed for Flow Levels 4 or 5.
Withdrawals from rivers and streams within these sub-
basins would be required to minimize existing impacts to
the greatest extent possible and to conduct feasible
mitigation and improvement in a way that is commensurate
with their impact.
These draft streamflow criteria would be used to guide
allowable withdrawals under the Massachusetts Water
Management Act permitting process2. For high-quality
streams, defined as either those with documented cold
water fisheries or those in Biological Categories 1, 2 or 3,
additional review and minimization of impacts would be
required. The framework also is designed to prevent stream
degradation from an existing to a lower Biological Category.
ACKNOWLEDGMENTS
The authors thank Alison Bowden and Mark P. Smith of The
Nature Conservancy for describing this project to us and
reviewing this case study. Thanks also to the participants in
the Massachusetts Sustainable Water Management
Initiative process, including Massachusetts resource state
agencies, and the U.S. Geological Survey.
REFERENCES
Archfield, S. A., Vogel, R. M., Steeves, P. A., Brandt, S. L.,
Weiskel, P. K., and Garabedian, S. P. 2010. The
Massachusetts Sustainable-Yield Estimator: A decision-
support tool to assess water availability at ungaged stream
locations in Massachusetts. U.S. Geological Survey
Scientific Investigations Report 2009-5227, 41 p. plus
CD-ROM http://pubs.usgs.gov/sir/2009/5227/.
Armstrong, D. S., Richards, T. A., and Levin, S. B. 2011.
Factors influencing riverine fish assemblages in
Massachusetts. U.S. Geological Survey Scientific-
Investigations Report 2011-5193, 58p http://pubs.usgs.gov/
sir/2011/5193/.
Armstrong, D. S., Richards, T. A., and Brandt, S. L. 2010.
Preliminary assessment of factors influencing riverine fish
communities in Massachusetts. U.S. Geological Survey
Open-File Report 2010-1139, 43 p. http://pubs.usgs.gov/
of/2010/1139/.
Table 1. Massachusetts draft streamflow criteria structure (February 3, 2012 version). Flow-ecology curves relating fluvial fish communities to
percent flow alteration in August from groundwater withdrawal informed the development of these proposed criteria. From Massachusetts
Executive Office of Energy and Environmental Affairs (2012).
2 http://law.onecle.com/massachusetts/21g/index.html
Seasonal Streamflow Criteria
Flow Level
August Flow Level
(Range of % alteration due to
groundwater withdrawal)
% allowable alteration of estimated
unimpacted median flow
Aug Oct Jan April
1 0 to < 3% 3% 3% 3% 3%
2 3 to < 10% 10% 5% 3% 3%
3 10 to < 25% 25% 15% 10% 10%
4 25 to < 55%
feasible mitigation and improvement
5 55% or greater
Streamflow Criteria Narrative: Existing sources in subbasins with alteration levels higher than those shown will be required to minimize
existing impacts to the greatest extent feasible and mitigate additional withdrawal commensurate with impact.
22
Armstrong, D. S., Parker, G. W., and Richards, T. A. 2004.
Evaluation of streamflow requirements for habitat
protection by comparison to streamflow characteristics at
index streamflow-gaging stations in southern New
England. Pages 108. U.S. Geological Survey Water-
Resources Investigations Report 03-4332 http://water.usgs.
gov/pubs/wri/wri034332/.
Barlow, P. M. 2000. Documentation of computer program
STRMDEPL — a program to calculate streamflow depletion
by wells using analytical solutions in P. J. Zarriello and K. G.
Ries, III, eds. A precipitation runoff model for analysis of the
effects of water withdrawals on streamflow, Ipswich River
Basin, Massachusetts: U.S. Geological Survey Water
Resources Investigations Report 00-4029, p. 77-89.
Cade, B. S., and Richards, J. D. 2005. User manual for
Blossom statistical software. U.S. Geological Survey
Open-File Report 2005-1353, 124 p. http://www.fort.usgs.
gov/products/publications/pub_abstract.asp?PubID=21536.
Davies, S. P., and Jackson, S. K. 2006. The biological
condition gradient: a descriptive model for interpreting
change in aquatic ecosystems. Ecological Applications
16(4):1251-1266.
Massachusetts Executive Office of Energy and
Environmental Affairs 2012. Sustainable Water
Management Technical Subcommittee, internal working
documents http://www.mass.gov/eea/air-water-climate-
change/preserving-water-resources/sustainable-water-
management/swm-mgmnt-tech-subcommittee/swm-
technical-subcommittee-resources.html
Weiskel, P. K., Brandt, S. L., DeSimone, L. A., Ostiguy, L. J.,
and Archfield, S. A. 2010. Indicators of streamflow
alteration, habitat fragmentation, impervious cover, and
water quality for Massachusetts stream basins. U.S.
Geological Survey Scientific Investigations Report 2009-
5272, 70 p. plus CD–ROM. http://pubs.usgs.gov/
sir/2009/5272/.
23
The Connecticut River basin project demonstrates (1)
coordinating diverse stakeholders to assess the costs and
benefits of re-operating more than 70 dams within an
interstate basin and (2) pursuing basin-wide environmental
flow and water use objectives through collaborative
simulation and optimization modeling.
This case study describes a major component of The Nature
Conservancy’s Connecticut River Program3 to restore
important river processes, thereby improving the health of
declining native species and diverse habitats along the river
and its tributaries. With 44 major tributaries, approximately
70 large dams, more than 2,600 smaller dams and 44,000
road-stream crossings spanning 4 states (Connecticut,
Massachusetts, New Hampshire, and Vermont) within the
19,288-km2 watershed, coordinated basin-scale planning
and management clearly is needed. Currently, management
of the 70 dams—including 14 owned and operated by the
US Army Corps of Engineers (Corps)—is not integrated as
a system.
The objective of the ecosystem flow restoration component
is to modify management of dams and water supply systems
to provide environmental benefits while continuing to
supply water, reduce flood risk, and generate hydropower
(Zimmerman et al. 2008). It is collaboratively managed and
funded by the Corps New England District Office through
a congressionally authorized (in the Water Resource
Development Act, or WRDA) study budgeted at $3 million.
The Nature Conservancy is a cost-share partner and has
raised its $1.5 million share through a private donation.
Preliminary technical studies by The Nature Conservancy
established the spatial extent, distribution, and scope of
flow alteration. Zimmerman (2006a) documented how
streamflow patterns influence physical processes and the
native species and communities of the Connecticut River
basin, based primarily on literature review. Zimmerman
(2006b) rigorously analyzed hydrologic alteration
downstream from flood-control dams on two tributary rivers.
Zimmerman and Lester (2006) mapped the potential degree
and extent of such alteration across the basin. Zimmerman
et al. (2009) modelled sub-daily flows and analyzed hourly
flow variability downstream of multiple dams across the
basin. Additionally, Gannon (2007) inventoried permitted
withdrawals and discharges to gain insight into each
state’s water resource management policies and their
relative contributions to hydrologic alteration within the
watershed. These studies laid a sound technical foundation
and helped focus and engage stakeholders.
Local, state, and federal stakeholders were convened on
numerous occasions and in a variety of formats, beginning
with a 2008 kick-off meeting. In 2009, the non-profit
Consensus Building Institute interviewed all key
stakeholders across the four states. One constituency that
was crucial to the project’s success was the private large
dam owners. A 2009 workshop and one-on-one onsite
visits with dam owners over 1.5 years proved essential for
understanding their operational constraints and for gaining
their involvement in the process. A 2010 workshop
introduced stakeholders to the modeling that was
underway and a 2011 workshop made initial environmental
flow recommendations for the basin.
The modeling team is building a hydrologic model and set
of decision–support systems (DSS) (Figure 1) for integrated
water resource management. Water managers and
stakeholders will be able to use the system to evaluate
environmental and economic outcomes of various water
management and climate change scenarios. The DSS also
will be useful for upcoming Federal Energy Regulatory
Commission (FERC) relicensing actions, for setting
individual dam operations in their regional context. Model
construction began in 2009 and is nearing completion.
The DSS includes two simulation models, one built by
University of Massachusetts Amherst (UMass) modelers
and the other by the US Army Corps of Engineers (Corps).
The UMass model uses STELLA system-dynamics software
to directly represent current reservoir operations and
economic outcomes in sub-basins. This model readily
solicits and synthesizes feedback from stakeholders. The
more operationally detailed Corps model generates
essentially identical output to the UMass model, but in the
Res-Sim format with which Corps dam engineers—who will
attempt to implement the recommendations that result
from the project—are most comfortable. Both models input
a hydrologic foundation of unimpaired (baseline) daily
streamflow hydrographs developed by the U.S. Geological
Survey, using duration-curve regression modelling (see
Massachusetts case study) for the middle 90 percent of
flows, and a modified drainage area ratio method for
extreme high flows. Resulting hydrographs are accessed
through the new Connecticut River UnImpacted Streamflow
Estimator4 (CRUISE) tool.
The simulation models are linked to a multi-objective
optimization model built by UMass modelers using Lingo
programming language. The optimization challenge was to
find daily releases from 70 reservoirs that meet flood
2.3 Connecticut River Basin Ecosystem Flow Restoration
3 http://conserveonline.org/workspaces/ctriver
4 http://webdmamrl.er.usgs.gov/s1/sarch/ctrtool/index.html
24
control, hydropower, water supply, recreation and
ecosystem requirements over time periods ranging from
one season to many years. The Connecticut River Wiki5
Page tracks model development progress.
The inclusion of environmental flow regime targets is novel
to water resource optimization modeling. In 2010,
environmental flow scientists convened a workshop with
UMass modelers to better understand modeling
constraints. Together, they devised a way to tailor the 2011
expert workshop to express flow needs in “model”
language: environmental flows are modeled as optimization
targets accompanied by penalty functions that describe
their flexibility. A steep penalty function indicates that the
target must be met; a shallow penalty function implies less
urgency in meeting that target.
At the 2011 expert workshop, The Nature Conservancy
provided participants with a list of preliminary flow
recommendations organized by species and biological
communities, based on extensive literature review. Each
flow recommendation was expressed in terms of season,
environmental flow component, flow range, and the
underlying flow-ecology hypothesis. In breakout sessions,
participants grouped according to their scientific
disciplines to review and refine the preliminary flow
recommendations. Then, all participants reconvened to
resolve differences in recommendation terms and
inconsistencies between the discipline-specific flow
recommendations.
Although the workshop participants were comfortable
recommending flow targets, they felt unprepared to define
penalty functions needed for optimization modeling. The
project team therefore has proposed environmental flow
penalty functions based on the “presumptive standard”
described by Richter et al. (2011).
The workshop attendees have agreed to reconvene to learn
how the optimization and operations models represent the
environmental flow recommendations, examine how their
flow recommendations and associated penalty functions
Figure 1. The basic structure of multi-agency water management decision support system for the Connecticut River basin, which includes an
optimization routine with environmental flow targets.
5 http://ctriver.ecs.umass.edu/wiki/index.php/Main_Page http://www.nature.org/ourinitiatives/regions/northamerica/unitedstates/connecticut/connecticutriver/index.htm
25
perform in terms of maintaining key aspects of unregulated
hydrographs and flow duration curves, and show how often
these recommendations can be achieved under various
water management scenarios. HEC-RAS, a hydraulic model
that calculates stage-discharge relations, will facilitate this
conversation. Participants will be asked to refine their initial
recommendations based on these results.
In 2012, project efforts will focus on refinement of
environmental flow targets and use of the models for
decision support. The Nature Conservancy will actively
involve stakeholders in exploring opportunities for dam
re-operation to provide environmental flows. At least one
experimental environmental flow release from an Army
Corps’ dam is anticipated in 2013. Monitoring will
document ecological conditions before and after flow
implementation and strengthen flow-ecology relationships.
Already, baseline mapping of vegetation at 91 floodplain
sites and hydraulic modeling (in HEC-RAS) have been
completed for the entire mainstem in Connecticut and the
mainstem in the vicinity of Northampton (MA), and may be
expanded to other upper mainstem or tributary reaches.
This information will be used to assess future benefits of
any dam re-operations.
Efforts to engage diverse stakeholders are paying off.
Owners of one of the large dams plans to use the project
models to support its FERC relicensing request, and the
National Atmospheric and Oceanographic Administration
has already been using the CRUISE tool to support its
participation in upcoming relicensing reviews. Having all
stakeholders agree on a common scientific foundation is
essential for integrating dam operations across the basin.
ACKNOWLEDGMENT
The authors thank Kim Lutz of The Nature Conservancy for
describing this project to us and reviewing this case study.
REFERENCES
Gannon, C. 2007. Hydrologic alteration of the Connecticut
River watershed: authorized water withdrawals/diversions
and discharges in Massachusetts and Connecticut. The
Nature Conservancy, Connecticut River Program,
Northampton, MA, 36 p. http://conserveonline.org/
workspaces/ctriver/documents/papers.
Richter, B. D., Davis, M., Apse, C., and Konrad, C. P. 2011.
A presumptive standard for environmental flow protection.
River Research and Applications. DOI: 10.1002/rra.1511
Zimmerman, J. K. H. 2006a. Response of physical
processes and ecological targets to altered hydrology in
the Connecticut River basin. The Nature Conservancy,
Connecticut River Program, Northampton, MA, 53 p. http://
conserveonline.org/workspaces/ctriver/documents/papers.
Zimmerman, J. K. H. 2006b. Hydrologic effects of flood
control dams in the Ashuelot River, New Hampshire, and
the West River, Vermont. The Nature Conservancy,
Connecticut River Program, Northampton, MA, 30 p. http://
conserveonline.org/workspaces/ctriver/documents/papers.
Zimmerman, J. K. H., and Lester, A. 2006. Spatial
distribution of hydrologic alteration and fragmentation
among tributaries of the Connecticut River. The Nature
Conservancy, Connecticut River Program, Northampton,
MA, 19 p. http://conserveonline.org/workspaces/ctriver/
documents/papers.
Zimmerman, J. K. H., Lester, J. A., Lutz, K., Gannon, C., and
Nedeau, E. J. 2008. Restoring ecosystem flows in the
Connecticut River watershed. The Nature Conservancy,
Connecticut River Program, Northampton, MA
Zimmerman, J. K. H., Nislow, K. H., Letcher, B. H., and
Magilligan, F. J. 2009. Determining the effects of dams on
subdaily variation in river flows at a whole-basin scale.
River Research and Applications. DOI: 10.1002/rra.1324
26
The Colorado ELOHA project demonstrates (1) using
flexible approaches to develop flow-ecology curves based
on studies reported in the literature and (2) using flow-
ecology curves to inform basin-scale water-resource
planning. Sanderson et al. (2011) provide a more detailed
discussion of the project.
In 2005, the Colorado Legislature passed the Colorado
Water for the 21st Century Act, launching a statewide
water planning effort. The Act mandated that
representatives of cities, farms, and other water users join
conservation and recreation interests at “basin
roundtables” to assess future water supply needs for their
watersheds. These assessments are framing discussions
about future water allocations and must address both
consumptive and non-consumptive (recreation and
environmental flows) water needs.
The Colorado Water Conservation Board funded this
1-year, approximately $200,000 project to help two pilot
basin roundtables—the Roaring Fork watershed in western
Colorado and the Fountain Creek watershed in eastern
Colorado—understand tradeoffs between consumptive and
non-consumptive water uses. In 2008, the consulting firm
Camp Dresser & McKee (CDM) worked with scientists
from Colorado State University and The Nature
Conservancy, staff from the Colorado Water Conservation
Board, and representatives of the Colorado Basin
Roundtable to apply ELOHA to estimate flow-related
ecological risk at the basin scale. The Watershed Flow
Evaluation Tool (WFET) displays the results under various
water management scenarios.
A hydrologic foundation existed for the Roaring Fork
watershed before this project began. The State of
Colorado’s water supply model, StateMod (CDWR and
CWCB 2009), is a monthly water accounting program that
begins with gaged streamflow data under current
conditions. Reservoir storage changes, water diversions,
and return flows are added or subtracted to obtain
baseline flows. Simple water accounting, weighted by
drainage area and precipitation, is then used to calculate
baseline flows at ungaged sites. Monthly flows are
disaggregated into daily flows using one of several
techniques, most commonly by emulating the daily flow
patterns recorded at selected gages. Baseline flows at
ungaged sites are calculated by apportioning flows across
watersheds according to their drainage areas and mean
annual precipitation rates. Current-condition flows at
ungaged sites are calculated by adding or subtracting
reservoir storage change, water diversions, and return
flows to the baseline flows. Groundwater withdrawals and
return flows are similarly added and subtracted from
streamflow, allowing for an aquifer-dependent time delay.
Several options are available for distributing monthly water
use data to daily time steps. StateMod has not yet been
calibrated for the Fountain Creek watershed, so analyses
for that pilot were restricted to gaged sites.
Indicators of Hydrologic Alteration (IHA) software (The
Nature Conservancy 2009) used output from StateMod to
calculate changes in five ecologically relevant flow
statistics—mean annual flow, mean August flow, mean
September flow, mean January flow, and mean annual
peak daily flow—after major water-supply projects had
started. Because StateMod was developed for purposes
other than ecological assessments, engineers analyzed its
assumptions and output to determine that these particular
IHA metrics could be calculated with sufficient accuracy.
River type classification was straightforward. As an
informal framework for organizing information about flow
and ecology, rivers were designated as Interior West, Rocky
Mountains, or Great Plains, according to the Level-1
ecoregion (CEC 1997) in which they are located.
Geomorphologic sub-classification limited the application
of resulting flow-ecology relationships to appropriate river
reaches. For example, relationships pertaining to floodplain
function were not applied to steep canyon reaches.
A Colorado State University Ph.D. student, Thomas Wilding,
developed relationships between streamflow and
warmwater and coldwater fish, riparian vegetation,
invertebrates, and white-water rafting and kayaking, based
on his review of 108 studies reported in the literature.
Quantitative approaches varied, depending on the form
and abundance of relevant information, and ranged from
statistical analysis using quantile regression (Cade and
Noon 2003) to categorical relationships and expert
consultation (Camp Dresser & McKee Inc. et al. 2009).
Figure 1 shows some examples.
The technical team then identified 3-5 risk classes for each
ecological attribute, based on expert opinion, if the
flow-ecology relationships were not already categorical.
Using the flow-ecology relationships, they determined the
range of flow values associated with each ecological risk
class. Then, using the StateMod and IHA output, they
associated each river segment with its level of ecologic risk.
For each river reach, the resulting map (Figure 2) indicates
2.4 Colorado Watershed Flow Evaluation Tool
27
Figure 1. Selected flow-ecology relationships for Colorado (Camp Dresser & McKee Inc. et al. 2009). Upper left: Response of riparian plant
communities to peak flow alteration. Upper right: Response of Rocky Mountain invertebrate species diversity to flow depletion on the day of
sampling, using data from two studies. Lower left: Response of flannelmouth sucker, a warmwater fish, to peak flow. Flow data were divided by
watershed area to compare different sized rivers. Lower right: Response of brown trout recruitment success to mean July flow. Regression lines
indicate percentiles advised by expert committee.
the risk that flow alteration has compromised ecological
values. The WFET allows basin roundtables to similarly
analyze the spatial distribution of ecological risk associated
with different potential future water use scenarios.
As mentioned earlier, unlike the Roaring Fork watershed,
the Fountain Creek watershed lacks streamflow data for
ungaged sites where biological data have been collected.
The researchers found that without a hydrologic
foundation, they were unable to formulate flow-ecology
relationships with sufficient certainty to warrant the
development of a WFET for that watershed.
Following the successful deployment of the Roaring Fork
WFET, the Basin Roundtable chose to expand the WFET to
the entire mainstem of the Colorado River within Colorado
and its tributaries. Stakeholders are now using the results of
the WFET to assess where flow restoration may be feasible,
to estimate the quantities of flow that may be needed for
restoration, to identify areas where additional study is
needed, and to prioritize protection actions for areas with
little flow-related ecological risk. The WFET is emerging as
a valuable tool in the development of a basinwide plan for
the protection and restoration of river health.
ACKNOWLEDGMENT
The authors thank John Sanderson of The Nature
Conservancy for describing this project to us and reviewing
this case study.
28
REFERENCES
Cade, B. S., and Noon, B. R. 2003. A gentle introduction to
quantile regression for ecologists. Frontiers in Ecology and
the Environment 1(8):412-420. doi:
10.1890/1540-9295(2003)001[0412:AGITQR]2.0.CO;2
Camp Dressser & McKee Inc., Bledsoe, B. D., Miller, W. J.,
Poff, N. L., Sanderson, J. S., and Wilding, T. K. 2009.
Watershed Flow Evaluation Tool pilot study for Roaring Fork
and Fountain Creek watersheds and site-specific
quantification pilot study for Roaring Fork watershed
(draft). Colorado Water Conservation Board, Denver, CO
http://www.conservationgateway.org/file/watershed-flow-
evaluation-tool-pilot-roaring-fork-and-fountain-creek.
CDWR and CWCB. 2009. State of Colorado’s water supply
model (StateMod) Version 12. Colorado Division of Water
Resources and Colorado Water Conservation Board:
Denver, Colorado ftp://dwrftp.state.co.us/cdss/swm/sw/
StateModManual_20090216.pdf.
CEC. 1997. Ecological regions of North America — toward
a common perspective. Commission for Environmental
Cooperation, Montreal
Sanderson, J. S., Rowan, N., Wilding, T., Bledsoe, B. P.,
Miller, W. J., and Poff, N. L. 2011. Getting to scale with
environmental flow assessment: the watershed flow
evaluation tool. Published
online in Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/rra.1542.
The Nature Conservancy. 2009. Indicators of Hydrologic
Alteration, Version 7.1 User’s Manual, 76 p.
http://conserveonline.org/workspaces/iha.
Figure 2. Watershed Flow Evaluation Tool output detail, showing level of flow-related ecological risk for each river reach. Risk levels: red = high,
orange = moderate, yellow = minimal, green = low. Blue = not modelled (no flow data).
29
This case study demonstrates the development of
statewide regulations to protect environmental flows,
including a process to establish condition goal classes for
every river and stream in the state and the development of
reservoir release rules based on their estimated natural
inflows. These reservoir release rules feature several
innovative approaches to mimic natural conditions, to
ensure water reliability for communities, and to be flexible
during drought events.
In 1971, the Connecticut legislature passed a law requiring
the protection of environmental flows for the “stocked
streams” of Connecticut—that is, those rivers and streams
stocked with fish by the Division of Wildlife. In 2005, at the
request of environmental advocates and with the
concurrence of water users, the legislature updated this
statute to require the Department of Environmental
Protection (DEP) to develop environmental flow regulations
for all rivers and streams in the state while also providing
for the other uses of water.
The statute requires that these regulations be developed
with the participation of an advisory group representing a
broad range of stakeholders. The DEP Commissioner
subsequently established three advisory groups—a
technical advisory committee, a policy advisory committee,
and a Commissioner’s advisory committee—to inform the
process. These groups met and negotiated for six years
until final regulations were agreed to in November 2011.
This process included a formal public comment process
and high-visibility advocacy campaigns, both for and
against the new regulations. In addition, in Connecticut all
final regulations must be reviewed by the Regulations
Review Committee of the legislature to ensure they comply
with the legislative intent. The first of these committee
reviews sent the initial regulations back for further
negotiation and narrowing of the scope before the final
regulations were produced. In the end, all sides advocated
for approval of the regulations as something they could
“live with.”
The new regulations6 contain three primary components:
(1) a set of narrative streamflow standards that apply to all
streams; (2) a goal classification process through which
every stream segment in the state will be associated with
one of four environmental flow standards it needs to meet;
and (3) a detailed set of flow release requirements for
reservoirs and impoundments, with different requirements
for small and large reservoirs. Each of these three aspects
introduces novel approaches to environmental flow
2.5 Connecticut Statewide Environmental Flow Regulations
By Mark P. Smith, The Nature Conservancy
Excerpt from the Connecticut Statute
“(1) Apply to all river and stream systems within
this state (2) preserve and protect the natural
aquatic life, including anadromous fish, contained
within such waters; (3) preserve and protect the
natural and stocked wildlife dependent upon the
flow of such water; (4) promote and protect the
usage of such water for public recreation; (5) be
based, to the maximum extent practicable, on
natural variation of flows and water levels while
providing for the needs and requirements of public
health, flood control, industry, public utilities, water
supply, public safety, agriculture and other lawful
uses of such waters; and (6) be based on the best
available science, including, but not limited to,
natural aquatic habitat, biota, subregional basin
boundaries, areas of stratified drift, stream gages
and flow data, locations of registered, permitted,
and proposed diversions and withdrawal data
reported pursuant to section 22a-368a, locations
where any dams or other structures impound or
divert the waters of a river or stream and any
release made therefrom, and any other data for
developing such regulations or individual
management plans.”
management. The regulations also include the typical
requirements related to rights of appeal, public
participation, and due process.
Narrative Streamflow Standards: The streamflow
regulations provide for four stream condition classes:
Class 1 streams “shall exhibit, at all times, the depth,
volume, velocity and variation of stream flow and water
levels necessary to support and maintain habitat conditions
supportive of an aquatic biological community
characteristic of that typically present in free-flowing river
or stream systems of similar size and geomorphic
characteristics under the prevailing climatic conditions.”
Class 2 streams have conditions that can support an
aquatic biological community minimally altered from a
free-flowing stream of a similar type. Class 3 streams have
conditions that support communities moderately altered
from systems of a similar type.
6 Connecticut Department of Energy and Environmental Protection Stream Flow Standards and Regulations Section 26-141b-1 to 26-141b-8, inclusive, of the Regulations of Connecticut State Agencies, Effective December 12, 2011, http://www.ct.gov/dep/lib/dep/regulations/26/26-141b-1throughb-8.pdf.
30
Finally, Class 4 streams allow for substantial alteration of
flow conditions to meet human water needs, but must
“exhibit to the maximum extent practicable the depth,
volume, velocity and variation of stream flow and water
levels consistent with the narrative standard for Class 3
river and stream segments.” This class was added in
recognition that many existing dams and reservoirs will
not be able to comply with class 3 standards but are
required to meet them to the best of their ability. The
detailed reservoir release requirements for such systems
are covered below. The regulations do make it difficult
(but not impossible) for any new reservoirs to be built to
a class 4 standard.
An important feature of the classification system is that it
does not require those biological communities to actually
be present (because they could be affected by stressors
other than flow), but rather that the flow conditions must
be sufficient to support such communities. As discussed
below, the level of alteration allowed within each condition
class is based on fish and flow data that were available to
the technical committee.
Stream Condition Goal Classification: The regulations
prescribe a detailed public process by which the
Department of Environmental Protection will assign a
condition goal class to each stream or river reach. DEP
must consult with the state Department of Public Health to
ensure the classification accounts for both environment
and public health needs.
The regulations stipulate 18 different considerations that
the DEP will take into account when classifying streams.
These range from environmental considerations like the
presence of sensitive species, to the presence of existing
water withdrawals, to existing and planned development
upstream of the reach. They also provide that no river
segment below an existing public water supply system
(not all dams are for public water supplies) will be
classified as the highest levels (class 1 or 2), which
provides current water suppliers some assurance of
continued use of their supplies.
The regulations also outline a public review process,
including a public hearing, whereby the draft goal
classification, by major basin, is released for public
comment before a final classification is made. Additionally,
the regulations provide for a process by which a
classification can be changed in the future—based either
on new information that may become available or based on
future needs that cannot be anticipated. However, the
burden of proof is higher for streams that petition to move
to a lower class than to move to a higher class of protection.
Reservoir Release Rules: The regulations provide
detailed release requirements for reservoirs and other
waters impounded by a dam or diversion based on stream
condition class and the size of the impoundment relative to
its watershed or catchment.
The release requirements are based on existing studies that
developed flow-response curves for fluvial-dependent
species, as studied in Georgia (Freeman and Marcinek
2006) and subsequently confirmed for the northeastern
United States (Vokoun and Kanno 2009, 2010; Armstrong
et al. 2010, 2011). Building upon these and other
environmental flow studies, the technical committee used a
“weighted evidence” approach (Norris et al. 2012) to
recommend stipulations that dams on class 1 streams
cannot actively manipulate the storage of the reservoir (in
effect making the dam “run of river”) and dams on class 2
streams must release at least 75% of their reservoir inflows
at all times. These were developed to meet the narrative
standards for class 1 and 2 rivers, as discussed above.
An innovative approach was developed for the class 3
releases. Almost all dams of any significant size fall into this
category (see below for more on reservoir size). For these
structures, the release requirements include two additional
factors to improve their ability to mimic natural flows.
First, the volume of the release required depends on the
bioperiod in which it occurs. Bioperiods are biologically-
based seasons lasting between one and four months. The
regulations define six bioperiods based on the flow needs
of the range of river species typically found in Connecticut
(see Table 1). Varying the release requirements according
to bioperiod improves their accuracy in mimicking natural
seasonal flows.
Second, larger releases are required during high flows than
during normal flows typical for the bioperiod. Although the
technical committee discussed the application of this
two-level release system year-round, the final regulation
only applies it during the rearing and growth bioperiod
(summer).
To implement this two-level release framework, the
technical committee developed a system for defining flow
levels based on average inflows over the preceding two
weeks. If inflow to the reservoir during the preceding two
weeks exceeded the bioperiod Q25 exceedance value, then
the flow level is considered “high”; if inflow was less, then it
is considered “normal.” This was a practical compromise
between aquatic ecologists, who would have preferred
daily flow adjustments, and reservoir operators, who have
operational and manpower issues to consider. This
approach adequately addresses one of the difficult issues
regarding release policies: how to account for wet and dry
periods as they occur to prevent either overly augmenting
streamflow that would naturally be low or unduly depleting
streamflow that would naturally be high. By tying releases
31
to immediately preceding conditions, they mimic the
natural flow regime.
Hydrologic Foundation: To quantify the daily unaltered
flows for all streams in Connecticut, the DEP is investing in
developing the hydrologic foundation component of the
Safe Yield Estimator, which was originally constructed for
Massachusetts (Archfield et al.2010), for the entire state.
Once the baseline flows are determined, they will be used
to calculate exceedance probabilities.
To ensure that releases are typical of the particular stream
or river on which a dam occurs, the release requirements
are expressed as bioperiod exceedance probabilities (Table
1). This addresses the challenge of using a consistent
approach for the entire regulated community while
acknowledging the different flow characteristics of different
types of streams. The exceedance probability of any stream
is based on the period of record of that particular stream or
river; so, the Q95 flow (a flow rate that is exceeded 95% of
the time) of a high-baseflow river differs substantially in
volume from the Q95 of a flashy river. In addition, when
reading Table 1 it is useful to take into account the
seasonality of flow in Connecticut rivers and streams, since
the relative size of releases associated with particular
exceedance probability values may not be immediately
obvious. For example, the required releases associated with
a “habitat forming” (spring) bioperiod Q99 is always much
larger than the discharge rate associated with a “rearing
and growth” (summer) bioperiod Q80.
Balancing People and the Environment: The regulations
also include several important provisions that ensure the
reliability of public water supplies (but do not apply to
dams used for other purposes). One is that during drought
conditions, the release requirements can be reduced
systematically (see Table 2). These reductions in releases
are tied to the water suppliers’ required drought
contingency plans so that the reduced releases to streams
occur somewhat simultaneously with the implementation
of water-use restrictions that suppliers impose as they
approach an emergency declaration. This includes a
provision for zero releases during water supply drought
emergencies, which ensures that the “last drop” of water
goes to people rather than to the stream. These reduced
releases during droughts significantly decrease the impact
of the release requirements on the security of water
supplies for human use.
The regulations also allow for extended, but time-limited,
release reductions to ensure that public water suppliers
maintain an adequate margin of safety. In Connecticut, the
Department of Public Health strongly encourages public
water suppliers to maintain 15% more water than their
typical daily or monthly demand. Suppliers who cannot
meet this margin of safety can reduce their reservoir
releases as they undertake other measures—either
conservation or development of new supplies—to ensure
they have an adequate margin of safety under the new
reservoir release requirements.
Negotiations of release requirements were substantially
aided by the use of the Safe Yield Wizard tool that the
Stockholm Environment Institute developed for the Water
Evaluation and Planning (WEAP)7 model (Vogel et al.2007).
This tool calculates changes in the safe yield, or the
amount of water consistently available for human use, that
would result from different release requirements. Water
suppliers and aquatic ecologists used the tool iteratively to
design release requirements that would not unacceptably
impact the amount of water available to supply customers.
Reservoir size: The regulations impose different release
rules for the smallest impoundments. The technical
committee recognized that very small impoundments
(defined either by small volume relative to the catchment
Bioperiod Effective Dates
Minimum Required Release
Antecedent Period Dry Antecedent Period Wet
Overwinter Dec 1- Feb 28/29 Bioperiod Q99
Habitat Forming Mar 1 – Apr 30 Bioperiod Q99
Clupeid Spawning May 1 – May 31 Bioperiod Q95
Resident Spawning June 1 – June 30 Bioperiod Q90
Rearing and Growth July 1- Oct 31 Bioperiod Q80 Bioperiod Q50
Salmonid Spawning Nov 1 – Nov 30 Bioperiod Q90
Table 1. Effective dates and minimum release requirements by bioperiod, as stipulated by Connecticut’s environmental flow regulations.
7 http://www.weap21.org/
32
area or by a very small catchment area) regulate flows less
than larger reservoirs. This is primarily because small
reservoirs are full for all but the driest times of the year and
most of the river water spills over their dams; therefore,
their natural hydrographs remain mostly intact. Also, for
very small watersheds, the releases required would be so
small that they would be difficult to implement or measure.
Therefore, the rule requires only a single minimum release
from most small reservoirs (the rearing and growth Q80)
and exempts reservoirs whose calculated release would be
less than 0.1 cubic feet per second.
The regulations include some important exemptions and
exceptions. For example, water diverted for emergency
purposes such as fire suppression is not explicitly exempt.
Similarly, dams that are under Federal Energy Regulatory
Commission (FERC) jurisdiction are not covered in order to
avoid duplicate or potentially conflicting requirements
(states have strong authority under the Clean Water Act’s
401 certification process to provide conditions for FERC
licensed facilities). Also, water diversions by agriculture and
golf courses are included only in that the regulations
require that they comply with other, existing best
management practices for those uses. Finally, critical to the
regulations’ acceptance was the provision that those
regulated by this rule have ten years from the time from
which a stream is classified until they must fully comply
with the regulations. This gives water managers, both
public and private, the security of having sufficient time to
make structural modifications to dams and/or to find
additional supplies if necessary.
REFERENCES
Archfield, S. A., Vogel, R. M., Steeves, P. A., Brandt, S. L.,
Weiskel, P. K., and Garabedian, S. P. 2010. The
Massachusetts Sustainable-Yield Estimator: A decision-
support tool to assess water availability at ungaged stream
locations in Massachusetts. U.S. Geological Survey
Scientific Investigations Report 2009-5227, 41 p. plus
CD-ROM http://pubs.usgs.gov/sir/2009/5227/.
Armstrong, D. S., Richards, T. A., and Brandt, S. L. 2010.
Preliminary assessment of factors influencing riverine fish
communities in Massachusetts. U.S. Geological Survey
Open-File Report 2010-1139, 43 p.
http://pubs.usgs.gov/of/2010/1139/.
Armstrong, D. S., Richards, T. A., and Levin, S. B. 2011.
Factors influencing riverine fish assemblages in
Massachusetts. U.S. Geological Survey Scientific-
Investigations Report 2011-5193, 58p
http://pubs.usgs.gov/sir/2011/5193/.
Freeman, M. C., and Marcinek, P. A. 2006. Fish assemblage
responses to water withdrawals and water supply reservoirs
in Piedmont streams. Environmental Management
38(3):435-450. DOI: 10.1007/s00267-005-0169-3
Norris, R.H., Webb, J.A., Nichols, S.J, Stewardson, M.J. and
Harrison, E.T. 2012. Analyzing cause and effect in
environmental assessments: using weighted evidence from
the literature. Freshwater Science. 31(1): 5–21
Vogel, R. M., Sieber, J., Archfield, S. A., Smith, M. P., Apse,
C. D., and Huber-Lee, A. 2007. Relations among storage,
yield and instream flow. Water Resources Research 43
(W05403). doi:10.1029/2006WR005226; http://www.weap21.
org/downloads/RelationsAmongStorage.pdf.
Vokoun, J. C., and Kanno, Y. 2009. Evaluating effects of
water withdrawals and impoundments on fish assemblages
in Connecticut streams. University of Connecticut,
Department of Natural Resources and the Environment
College of Agriculture and Natural Resources
Vokoun, J. C., and Kanno, Y. 2010. Evaluating effects of
water withdrawals and impoundments on fish assemblages
in southern New England streams. Fisheries Management
and Ecology 17:272-283. http://www.nrme.uconn.edu/
pages/people/bios/documents/KannoVokoun.
WaterWithdrawals.FME.2010.pdf.
Water Supply Plan Trigger
Percentage of Required Dry-Period Release
Rearing & Growth Bioperiod All Other Bioperiods
Drought Advisory 100%75%
Drought Watch 50%50%
Drought Warning 25%25%
Drought Emergency No Release Required No Release Required
Table 2. Reduced release requirements for public water suppliers during drought.
33
The Middle Potomac River basin project demonstrates
(1) the determination of environmental flow needs for rivers
and streams that are generally more impaired by land use
change than by withdrawals or impoundments; (2)
engagement of multiple water resource agencies and other
stakeholders across jurisdictional boundaries; and (3) a
structured, iterative approach for selecting flow and
ecology metrics and refining river types to strengthen
flow-ecology relationships.
The project began in May 2009 and is slated for completion
in spring 2012. Its approximately $1 million ELOHA project
budget is funded mainly by the U.S. Army Corps of Engineers
(75% Federal cost-share through the Corps’ Section 729
Watershed Assessment program) and The Nature
Conservancy (25% non-Federal cost-share), with additional
support from the National Park Service, the Interstate
Commission on the Potomac River Basin (ICPRB), and other
basin jurisdiction agencies. Boundaries of the 11,500-mi2
Middle Potomac project area were determined by
Congressional designation of the Corps’ study authority, but
the project analyses extended upstream to allow for system
connectivity. The project area includes parts of Pennsylvania,
Maryland, Virginia, West Virginia, and all of Washington, DC.
ICPRB is the project’s technical lead. The Commission was
created by interstate compact in 1940, primarily to provide
technical support and expertise to the watershed
jurisdictions. ICPRB lacks authority to regulate streamflow.
Therefore, the project was designed to support efforts of
the five watershed jurisdictions to protect and restore
environmental flows.
The project team has informed and involved the watershed
jurisdictions throughout the project development and
analytical process. A seven-part webinar series, technical
advisory group meetings, a technical workshop, agency
consultative meetings, and a project website8 have
maintained watershed states’ involvement throughout the
project, from inception to completion. Through these
interactions, stakeholders have reviewed the technical
approach, discussed potential policy applications, and
considered how to use the flow-ecology relationships to
inform water and land use management decisions.
Because the Potomac River basin project area has few
large dams and flow is relatively unimpaired by major
impoundments, this assessment was not oriented towards
changing dam operations. In fact, the analysis is finding
that land use change is having a greater impact on the
river’s hydrologic regime than dams or impoundments.
The project goals are to:
• Estimate current and future water withdrawals, given
population, land use, and climate change projections;
• Determine impacts of water withdrawals, discharges,
impoundments, land use, and climate change on flow;
• Characterize flows needed to support healthy biotic
communities in smaller streams and rivers; and
• Provide data, information, and analyses to support water
and land use planning and decision making at the state
level.
A modified version of the site-specific “Savannah” process
(Richter et al. 2006) was used to determine flow needs for
selected segments of the Potomac River mainstem and
selected large tributaries (Cummins et al. 2011), while the
regional-scale ELOHA framework was used for smaller
tributary streams and wadeable rivers. Here we describe
only the ELOHA process.
Figure 1 shows how the project technical team modified
the original ELOHA framework. The major modification is
their exploratory approach to the biological data analyses,
rather than first describing flow-ecology hypotheses based
on available literature. They iteratively refined river types,
flow metrics, and biometrics to determine flow-ecology
relationships, which they then presented to watershed
scientists for review. The project website—particularly the
archived webinar series—documents the iterative analytical
process in detail.
The project’s hydrologic foundation consists of 21 years of
daily flow data at biological monitoring sites under seven
scenarios—modelled baseline (or relatively unaltered),
modelled current, and five modelled future alternative flow
scenarios—simulated by the Virginia Department of
Environmental Quality’s decision support system,
WOOOMM9 (Watershed Online Object Oriented Meta-
Model). Input to WOOOMM includes edge-of-stream flows
generated by the Chesapeake Bay Program’s Phase 5.2
HSPF (Hydrological Simulation Program FORTRAN) model
(enhanced to include non-linear groundwater recession
and re-segmentation at major impoundments), a U.S.
Geological Survey (USGS) channel morphology model, and
a channel routing routine. The model was calibrated to
measured flow at 56 USGS gages. The 747 subwatersheds
in the final model capture 869 biological monitoring sites
2.6 Middle Potomac River Basin Environmentally Sustainable Flows
8 http://www.potomacriver.org/2012/projects/middle-pot-assess
9 http://sifn.bse.vt.edu/sifnwiki/index.php/WOOOMM_Modeling
34
and representative distributions of bioregions, land cover
and catchment areas. Both baseline and current-condition
flow series use 1984-2005 climate data.
To identify baseline flow conditions, modelers used a
Category and Regression Tree (CART) analysis of 105
gaged watersheds in the Potomac and adjacent
Susquehanna River basins. The CART analysis determined
thresholds when flows were significantly impacted by
anthropogenic land use, withdrawals, discharges, and
impoundments. For each anthropogenic factor, thresholds
were defined above or below which flows are considered
altered. The analysis found that watersheds with greater
than or equal to 78% forest cover and less than or equal to
0.35% impervious cover and no impoundments,
withdrawals, or discharges have the least altered flows.
Therefore, in the modeled baseline scenario, land use in
every watershed was adjusted to have at least 78% forest,
less than or equal to 0.35% impervious cover, and no
withdrawals, impoundments, or discharges.
Current conditions were represented in the models using
land use data for 2000, withdrawal and discharge data for
2005, and significant impoundments. Surface-water
Figure 1 Modified ELOHA framework for the Middle Potomac River basin sustainable flows study.
35
withdrawal data were obtained from the individual states.
Groundwater withdrawals were not modeled due to
incomplete data, insufficient understanding of complex
groundwater flow systems, and limitations of the hydrologic
foundation models. Permitted point-source discharge data
were obtained from the Chesapeake Bay Program’s
discharge database. The Chesapeake Bay model simulates
four large impoundments in the study area. Twelve smaller
impoundments were added to the Middle Potomac project
model because they are located near biological monitoring
sites and contain significant storage or are used for
hydropower production.
Eighteen flow metrics were selected for flow-ecology
analysis from 256 metrics initially calculated by Indicators
of Hydrologic Alteration (The Nature Conservancy 2009)
and Hydrologic Integrity Tool (Henriksen et al. 2006)
software (Figure 2). Analysis of flow alteration reduced the
initial set to those that have changed the most from
baseline to current conditions and are expected to change
the most from current to future conditions. Metrics that
correlate strongly with other metrics were then removed.
The selected subset of hydrologic metrics represents all
parts of the hydrograph (Table 1).
River type classification initially was based on watershed
size and percent karst geology. This first-cut classification
was later abandoned in favor of an iterative statistical
approach aimed at increasing sample sizes and
strengthening flow-ecology relationships. Ultimately,
Flow Range Magnitude Duration Frequency Other
High Mean high flow
volume
High flow duration High pulse count,
High flow frequency,
Flood frequency
Skewness in
annual maximum
flows
Medium Median annual flow
volume
Flood-free season Fall rate,
Flashiness
Low 4-day harmonic
mean low,
Seasonal Q85,
7Q10
Low pulse duration,
Extreme low flow duration,
Coefficient of variation in
low flow pulse duration
Low pulse count,
Extreme low flow
frequency
Table 1. Subset of flow metrics selected for the Middle Potomac Sustainable Flows Project after screening. Italics indicate metrics exhibiting
strong relationship to Chessie BIBI, a benthic index of biotic integrity.
Figure 2. Process for selecting flow metrics for flow-ecology analysis.
36
selected hydrologic metrics were normalized by watershed
area and biological metrics were normalized to a comparable
scale, accounting for much of the basin’s natural variability
and thus obviating the need to classify rivers.
Biometric selection began with multiple exploratory
analysis of an extensive basin-wide benthic invertebrate
database to identify those metrics that are most responsive
to anthropogenic stress and habitat degradation.
Correlations between those biometrics and candidate flow
metrics then were tested. Although a range of biometrics
was examined, one particularly revealing metric was the
Chesapeake Bay benthic index of biotic integrity database
(“Chessie BIBI”). Chessie BIBI combines macroinvertebrate,
habitat, and water quality data from 23 federal, state, local,
and ICPRB monitoring programs in a uniform database
structure. Starting with 50 family-level benthic invertebrate
metrics, scientists selected the overall index metric
(Buchanan et al. 2011) and 19 other metrics that indicate
community status, are not correlated, and are expected to
respond to flow alteration. These metrics represent
taxonomic composition, pollution tolerance, functional
habitat group, and functional feeding groups. Biological
data for 2000-2008 were used.
Flow-ecology relationships were determined using
quantile-regression. Examples are shown in Figure 3. The
biological samples used represent status at a single point
in time, but are being used to represent status over a
longer time period. To account for uncertainty in the true
biological status around the value calculated from a single
point, flow-ecology curves were defined as the 90th
percentile regression rather than as the maximum values of
biological metrics reported. The regression curves
(calculated with the Blossom program (Cade and Richards
2005)) represent the best possible biological score (with
10% allowance for uncertainty) for a given degree of flow
alteration.
At the expert workshop in November 2011, local and
regional scientists reviewed preliminary flow alteration-
ecological response statistical relationships. The project
technical team members explained that they chose to focus
exclusively on benthic macroinvertebrates because of the
availability of a basinwide dataset, and they gave an
overview of their study design and initial findings. After
reviewing the findings, the workshop participants
recommended that the study team consider other taxa,
particularly for evaluating ecological responses to low flow
alteration—a primary management concern for which
benthic macroinvertebrates proved to be poor indicators.
The experts also questioned whether the relationships
found for macroinvertebrates adequately captured
flow-dependencies of other ecosystem components, such
as fish and riparian vegetation. They also expressed a
Figure 3. Quantile-regression plots of selected flow-ecology
relationships in the Middle Potomac River basin project area.
Red is 90th percentile. Additional quantile regressions at 10th,
25th, 50th, and 75th percentiles are also shown.
37
concern that water managers and users would have
difficulty understanding the hydrologic metrics and relating
to the biological metrics that were used, and they
recommended providing alternative presentations of the
study findings that would be more intuitive and
understandable by resource managers.
The Potomac project was designed as a holistic, interstate
environmental flow needs assessment for the entire
watershed, using a shared hydrologic foundation and
biological dataset. However, state agencies regulate water
withdrawals in the Potomac watershed, and local
authorities make land use decisions that affect flows. For
this reason, flow recommendations emerging from this
regional analysis will need to be implemented at the
individual state or local level. The Potomac project team is
sharing flow alteration-ecological response relationships
with state-level resource managers and teams to support
their technical assessments and recommendations for
protecting and restoring environmental flows and stream
health throughout the watershed.
ACKNOWLEDGMENT
The authors thank Stephanie Flack of The Nature
Conservancy and Heidi Moltz and Claire Buchanan of the
Interstate Commission on the Potomac River Basin for
describing this project to us and reviewing this case study.
REFERENCES
Buchanan, C., Foreman, K., Johnson, J., and Griggs, A. 2011.
Development of a basin-wide benthic index of biotic
integrity for non-tidal streams and wadable rivers in the
Chesapeake Bay Watershed. Interstate Commission on the
Potomac River Basin, ICPRB Report 11-1
http://www.potomacriver.org.
Cade, B. S., and Richards, J. D. 2005. User manual for
Blossom statistical software. U.S. Geological Survey
Open-File Report 2005-1353, 124 p. http://www.fort.usgs.
gov/products/publications/pub_abstract.asp?PubID=21536.
Cummins, J., Buchanan, C., Haywood, C., Moltz, H., Griggs,
A., Jones, R. C., Kraus, R., Hitt, N., and Bumgardner, R. V.
2010. Potomac basin large river environmental flow needs.
Interstate Commission on the Potomac River Basin ICPRB
Report 10-03 http://www.potomacriver.org.
Henriksen, J. A., Heasley, J., Kennen, J. G., and Nieswand,
S. 2006. Users’ Manual for the Hydroecological Integrity
Assessment Process Software (including the New Jersey
Assessment Tools). U.S. Geological Survey Open-File
Report 2006-1093, 71 p. http://www.fort.usgs.gov/products/
publications/21598/21598.pdf.
Richter, B. D., Warner, A. T., Meyer, J. L., and Lutz, K. 2006.
A collaborative and adaptive process for developing
environmental flow recommendations. River Research and
Applications 22:297-318.
The Nature Conservancy. 2009. Indicators of Hydrologic
Alteration, Version 7.1 User’s Manual, 76 p.
http://conserveonline.org/workspaces/iha.
38
The Susquehanna River basin project demonstrates (1) the
systematic organization of relevant information sources,
including published and gray literature and existing data, to
facilitate expert input on flow-ecology relationships and
environmental flow recommendations, and (2) a novel
expression of environmental flows for maintaining long-
term hydrologic variability.
The 1972 Susquehanna River Basin Compact between New
York, Pennsylvania, Maryland, and the Federal government
established the interstate Susquehanna River Basin
Commission (SRBC). SRBC’s mission is to manage the
basin’s water resources under comprehensive watershed
management and planning principles, and it has authority
to regulate water withdrawals within the three basin states.
SRBC facilitated this science-based process to determine
environmental flow needs throughout the basin. Because
the SRBC has interstate regulatory authority, the resulting
recommendations are expected to be used to revise water
policy, inform basin planning, and improve water releases
from reservoirs within the basin starting in 2012.
This project was completed under Section 729 authority of
the Water Resource Development Act, which authorizes the
U.S. Army Corps of Engineers (Corps) to assess water
resource needs of river basins, including needs related to
ecosystem protection and restoration and water supply.
SRBC provided the non-federal cost share and contracted
with The Nature Conservancy (TNC), which was the
technical lead. The project began in early 2009 and was
completed in 18 months.
The project’s success hinged on the ability to synthesize
diverse sources of information, to present it in formats that
facilitate group discussion, and to convene and use expert
knowledge effectively. Box 1 outlines the project schedule,
organized around three pivotal meetings.
Through consultations with experts, the technical team
assembled a broad list of ecological indicators, including
flow‐sensitive taxa groups, vegetation community types,
and physical processes. The technical team then surveyed
scientific literature to find dependencies between these
indicators and specific flow components and, where
possible, to extract relationships between flow alteration
and ecological response. Using species distribution data
and expert consultations, they associated species groups
with major habitat types and described common traits and
microhabitat preferences for each species group.
2.7 Susquehanna River Basin Ecosystem Flow Recommendations
Box 1. Susquehanna River Basin
Ecosystem Flows Study Outline
Orientation meeting (9 March 2009)
Meeting outcomes
Engaged stakeholders
Nominated ecological indicators
Suggested information sources
Identified potential data gaps
Technical team work (Mar 2009-Oct 2009)
Reviewed and synthesized literature
Identified functional species groups
Delineated preliminary river types
First expert workshop (14-15 Oct 2009)
Materials provided to participants before the
meeting
Hydrographs showing 1960-2008 inter- and
intra-annual flow variability and the timing of
life-history stages for each species group
Table of detailed information associated with each
species and life stage
Meeting outcomes
Drafted flow-ecology hypotheses by river type
Prioritized additional information for summary
report
Suggested analyses to help develop flow
recommendations
Technical team work (Oct 2009-Apr 2010)
Reviewed literature and consulted experts to
support hypotheses
Drafted flow recommendations
Drafted summary report
Second expert workshop (7-8 Apr 2010)
Materials provided to participants before the
meeting
Draft flow-needs diagram for each major habitat
type (Figure 1)
Draft flow recommendations (Figure 2)
List of literature cited
Meeting outcomes
Peer-reviewed major products
Technical team work (Apr 2010-Sept 2010)
Analyzed effects of flow recommendations on
streamflow under different water withdrawal
scenarios; obtained further expert consultation
Finalized recommendations
Final report (Sept 2010)
DePhilip and Moberg (2010)
39
A basic habitat classification based on watershed size,
temperature, and flow stability was developed for organizing
and synthesizing information. Three existing classification
systems were tapped to assign river reaches to five major
habitat types. The Northeast Aquatic Habitat Classification
(Olivero and Anderson 2008) defined “major tributaries”
and “mainstems” as reaches with drainage areas exceeding
200 mi2. Hydrologic classification using the USGS
Hydrologic Integrity Process (HIP) software (Henriksen et
al. 2006) defined “high-baseflow streams.” Water-quality
designations from state regulatory programs defined “cool
and coldwater streams” and “warmwater streams.”
Long-term data for 45 minimally-altered (baseline) stream-
flow gages indicate that the flow volume on any day of the
year varies considerably from year to year. To capture this
variability, the technical team defined monthly high, seasonal,
and low flow components for each major habitat type.
Representative hydrographs juxtaposing these flow
components to life-history stages of native species
prepared participants for the first expert workshop.
Workshop participants used this information to identify the
most sensitive periods and life stages for each habitat type,
and to formulate flow-ecology hypotheses. Following the
first workshop, the technical team further compiled and
synthesized diverse information, using the flow-ecology
hypotheses to focus their research.
Ecosystem flow needs were then summarized graphically
by season in relation to high, seasonal, and low flows for
each major habitat type (Figure 1). These graphs and
supporting narratives describe the role of inter-annual as
well as seasonal hydrologic variability in forming channels
and floodplains; maintaining water quality; and supporting
life stages of fish, aquatic insects, mussels, reptiles,
amphibians, birds, and mammals.
The vast array of ecosystem flow needs convinced the
project team that it needed to develop environmental flow
recommendations for many different taxa for each major
habitat type—even those that lack large databases. Rather
than assume that a single species or group of species can
represent all ecosystem needs, the team took a novel
Figure 1. Graph showing ecological functions that depend on typical low, seasonal, and high flows during fall, winter, spring, and summer for
one habitat type (Major Tributaries) in the Susquehanna River basin. A similar graph for each habitat type greatly facilitated development of
flow recommendations in an expert workshop.
40
approach. The resulting flow recommendations are based
on (a) existing literature and studies that described and/or
quantified relationships between flow alteration and
ecological response, (b) expert input, (c) the analysis of
long-term flow variability at minimally-altered gages, and
(d) results of water withdrawal scenarios that tested the
sensitivity of various flow statistics.
Ten types of flow statistics were selected to describe the
magnitude and frequency of large and small floods, high
flow pulses, median monthly flow, and monthly low flow
conditions in the Susquehanna River basin: magnitude and
frequency of 20-year (large) flood, 5-year (small) flood, and
bankfull (1-2 year high flow) events; frequency of high flow
pulses in summer and fall; high pulse magnitude (monthly
Flow Need Flow Statistic and Recommendation Supporting Literature and Studies
SUMMER
Promote/support
development and growth of
all fishes, reptiles, and
amphibians — Summer and
fall flows needed to maintain
high velocity riffles, low
velocity pools, and backwaters
and stream margins.
All habitat types
Seasonal Flow May-Oct
• Monthly median between 45th and
55th percentile; and
• Less than 20% change to monthly
range
In a large river, availability and persistence of
shallow-slow water habitats were directly correlated
with fish abundance, particularly percids, catostomids
and cyprinids (Bowen et al. 1998).
Reductions of streamflows during this period have
had measurable impacts on size of adult brook trout
(Hakala and Hartman 2004, Walters and Post 2008)
On headwater and small streams, a simulated removal
of 8% of Aug median (p50), predict 10% shift in fish
assemblage; On large rivers removal of 10% in of the
Aug median (p50) predict 10% shift in fish
assemblage (Zorn et al. 2008).
Baseflows in a large river were augmented by an
estimated 100% under regulated conditions resulting
in an estimated 40% reduction of shallow slow water
habitat patch size during normal baseflow periods
(summer-fall-early winter) (Bowen et al. 2003).
Young-of-year abundance most correlated with
shallow-slow habitat size and persistence. Suitable
conditions predicted by statistics including seasonal
median daily flow, high pulse magnitude, duration and
rate of change (Freeman et al. 2001).
A comparison of large warmwater streams along a
withdrawal index gradient finds a shift in fish
assemblages from fluvial specialists to habitat
generalists as withdrawals increase above 50% of
7Q10 (Freeman and Marcinek 2006).
Longitudinal connectivity is important as map turtles
migrate to nesting locations. Stream migrations of 1-3
km have been documented on the lower Susquehanna
River (Richards and Seigel 2009).
Low Flow - Mar-July
Headwaters
• No change to monthly Q75; and
• No change to monthly low flow
range
Streams > 50 square miles
• No change to monthly Q95; and
• <10% change to monthly low flow
range
Maintain connectivity
between habitats and
refugia for resident and
diadromous fishes — resident
and diadromous fish need
seasonal flows to maintain
thermal refugia and maintain
connectivity among habitats
All habitat types
Seasonal Flow - Jun-Oct
• Monthly median between 45th and
55th percentile; and
• Less than 20% change to monthly
range
Elimination of longitudinal connectivity (simulated
barriers) prevented upstream migration of brook trout
and led to extinction of local brook trout populations
within 2 to 6 generations. Extinction of source
populations increased the probability of
metapopulation extinction (Letcher et al. 2007).
Figure 2. Format of Susquehanna River basin flow recommendations, associating ecological function with ranges of associated flow statistics,
and information that supports the recommendation. Colors indicate flow components (low, seasonal, or high (not shown)).
41
Q10); monthly median (Q50); typical monthly range (area
under monthly flow duration curve between the Q75 and
Q10); monthly low flow range (area under monthly flow
duration curve between Q75 and Q99); monthly Q75 and
monthly Q95. In addition, monthly range and monthly
low-flow range statistics were used to quantify changes in
flow-duration curve shape (Vogel et al. 2007),
complementing analyses of changes in individual flow
metrics to assess seasonal impacts of water use on
ecological flow regimes. DePhilip and Moberg (2010)
explain how to process output from Indicators of
Hydrologic Alteration software (The Nature Conservancy
2009) to calculate flow alteration as differences between
flow duration curves. Flow recommendations are expressed
in terms of acceptable ranges of these flow statistics.
The team systematically documented the flow needs that
each recommendation supports, and the literature and
studies on which the recommendation is based (Figure 2).
Structuring the flow recommendations in this way
facilitated the review process and provides a framework for
adding or refining flow needs, substituting flow statistics,
revising flow recommendations, and documenting
additional supporting information. This structure also
focuses future research on relationships between specific
types of flow alteration and specific ecological responses.
To further understand the sensitivity of each flow component
and to help translate the flow recommendations into policy,
the team analyzed a suite of future water withdrawal
scenarios and compared alternative flow thresholds. In
March 2012, SRBC released a draft Low Flow Protection
Policy for public comment, based in part on the flow
recommendations generated by this project.
TNC currently is extending the work described in this case
study to the Ohio and Delaware River basins in Pennsylvania
and adjoining states. At the same time, USGS is developing
a Virtual Gage Tool similar to Massachusetts’ Sustainable
Yield Estimator (Archfield et al. 2010) to estimate minimally-
altered (baseline) daily time series for ungaged sites in
Pennsylvania. Adding water-use data to these time series
will enable comparison between flows under baseline and
current conditions. Pennsylvania Department of
Environmental Protection (PA DEP) and SRBC plan to use
this tool to help review proposed water withdrawals and to
ensure that future water use maintains the environmental
flows recommended in this and future studies.
ACKNOWLEDGMENT
The authors thank Michele DePhilip and Tara Moberg of
The Nature Conservancy for describing this project to us
and reviewing this case study.
REFERENCES
Archfield, S. A., Vogel, R. M., Steeves, P. A., Brandt, S. L.,
Weiskel, P. K., and Garabedian, S. P. 2010. The
Massachusetts Sustainable-Yield Estimator: A decision-
support tool to assess water availability at ungaged stream
locations in Massachusetts. U.S. Geological Survey
Scientific Investigations Report 2009-5227, 41 p. plus
CD-ROM http://pubs.usgs.gov/sir/2009/5227/.
DePhilip, M., and Moberg, T. 2010. Ecosystem flow
recommendations for the Susquehanna River basin. Report
to the Susquehanna River Basin Commission and U.S. Army
Corps of Engineers. The Nature Conservancy. 96 p +
appendices http://www.nature.org/media/pa/tnc-final-
susquehanna-river-ecosystem-flows-study-report.pdf.
Henriksen, J. A., Heasley, J., Kennen, J. G., and Nieswand, S.
2006. Users’ Manual for the Hydroecological Integrity
Assessment Process Software (including the New Jersey
Assessment Tools). U.S. Geological Survey Open-File
Report 2006-1093, 71 p. http://www.fort.usgs.gov/products/
publications/21598/21598.pdf.
Olivero, A. P., and Anderson, M. G. 2008. Northeast aquatic
habitat classification system. The Nature Conservancy
Eastern Regional Office, Boston, Massachusetts, 88 p.
The Nature Conservancy. 2009. Indicators of Hydrologic
Alteration, Version 7.1 User’s Manual, 76 p. http://
conserveonline.org/workspaces/iha.
Vogel, R. M., Sieber, J., Archfield, S. A., Smith, M. P., Apse,
C. D., and Huber-Lee, A. 2007. Relations among storage,
yield and instream flow. Water Resources Research 43
(W05403). doi:10.1029/2006WR005226; http://www.weap21.
org/downloads/RelationsAmongStorage.pdf.
42
This case study demonstrates (1) the integration of
flow-ecology relationships and proposed streamflow
protection standards with existing water quality standards
using a tiered Aquatic Life Use (ALU) approach and (2)
the development of flow-ecology response curves from
statistical analysis of flow-habitat and habitat-ecology
relationships using extensive habitat and biological
databases.
This project is a work in progress. The process described
here (Rankin et al. 2012) was carried out independently
by a non-profit research institute, the Midwest Biodiversity
Institute (MBI), with funding from The Nature Conservancy
(TNC). A coalition of environmental groups is using the
results to secure ecologically-based low flow protection
in the ongoing Ohio Great Lakes Compact Implementation
process. Additionally, the Ohio Department of Natural
Resources (ODNR) has expressed interest in using the
flow-ecology response curves developed during this
process to evaluate proposed water withdrawals after
a regulatory program is in place.
Ohio’s development of ecological flow protection standards
stems from Ohio’s commitment to comply with the Great
Lakes Compact (see Michigan case study). The Ohio
Legislature’s ratifying language and allotted time of one
year to develop implementation language constrained the
initial focus to low flows, which represent the most
ecologically stressful period of the year. Given the time
limit, water users’ resistance to new regulatory programs,
and the highly altered condition of many of Ohio’s streams,
the approach was designed to mesh with the Ohio
Environmental Protection Agency’s existing ecological
monitoring and tiered ALU framework. Because the
Compact drove the process, initially it was developed only
for the Ohio streams that are tributary to the Great Lakes.
The hydrologic foundation is a database of mean daily flow
for the month of lowest flow (historically September) over a
20-year period, housed in the U.S. Geological Survey
StreamStats system (Koltun et al. 2006). Flow regression
modeling (Koltun and Whitehead 2002) was used to
estimate this flow statistic for ungaged sites. Because
pre-development flows are not determined, the hydrologic
foundation implicitly sets the current condition as the
baseline. Groundwater-surface water interactions were not
considered during this initial phase because almost all
Ohio streams in the Lake Erie basin are runoff-dominated.
2.8 Ohio Thresholds for Ecological Flow Protection
Class Description % of Waters
Warmwater Habitat (WWH)Principal restoration target for most of Ohio’s rivers and streams in Ohio, with
“typical” warmwater species assemblages.
77.4
Exceptional
Warmwater Habitat
(EWH)
Protection goal for Ohio’s best water resources, which support “unusual and
exceptional” assemblages of aquatic organisms, with a high diversity of species,
particularly those which are highly intolerant and/or rare, threatened, endangered,
or special status (i.e., declining species).
10.2
Modified Warmwater Habitat
(MWH)
Streams and rivers that have been subjected to extensive, maintained, and
essentially permanent hydromodifications such that the biocriteria for the WWH
use are not attainable, with species that tolerate low dissolved oxygen, siltation,
nutrient enrichment, and poor quality habitat.
3.8
Limited Resource Water
(LRW)
Small streams (usually less than 3 mi2 drainage area) and other water courses
that have been irretrievably altered to the extent that no appreciable assemblage
of aquatic life can be supported; includes small streams in extensively urbanized
areas, those that lie in watersheds with extensive drainage modifications, those
that completely lack water on a recurring annual basis (i.e., true ephemeral
streams), and other irretrievably altered waterways.
6.2
Coldwater Habitat
(CWH-N and CWH-F)
Waters that support assemblages of native coldwater organisms (CWH-N) and/or
those that are stocked with salmonids with the intent of providing a put-and-take
fishery on a year-round basis (CWH-F).
2.4
Table 1. Ohio’s Aquatic Life Use Classes (Ohio EPA 2004).
43
To classify river types, researchers considered base flow
index, upstream catchment size, biotic assemblage, water
quality, temperature, and other ecoregional characteristics.
They found that none of these could better explain
ecological response (number of sensitive fish species
supported) than does the existing ALU classification (Table
1). Furthermore, adopting an existing classification avoids
creating a new regulatory framework. The Ohio ALU
classification stratifies on the basis of ecological condition,
existing flow alteration, and thermal regime (coldwater or
warmwater habitat). Note that the Ohio Aquatic Life Use
classification is not a river type classification in the sense
described in the ELOHA framework because it considers
current ecological condition.
Ohio’s flow-ecology curves (Figure 1) relate number of
sensitive fish species to mean daily flow in September for
each river type. These relationships are derived from
fish-habitat and habitat-flow relationships, and are based
on the premise that water withdrawals reduce available
habitat niches, which reduces the number of sensitive fish
species that a stream can support. The habitat portion of
this relationship was developed from Ohio’s Quantitative
Habitat Evaluation Index (QHEI; Rankin 1989, 1995), which
includes a measure of niche availability at ecological
sampling locations.
Pollution-sensitive fish species were selected for use as a
regulatory target for two reasons. First, Ohio maintains a
list of pollution-sensitive fish which, according to expert
opinion and literature review, are also sensitive to flow
alteration. Second, responses of the species on this list to
low-flow depletion are strongly related to those of other
sensitive aquatic species such as freshwater mussels,
macroinvertebrates, and other fish species. This was
determined by analyzing long-term (1990-2009) ecological
sampling data at the 3,070 sample points from the Ohio
Ecological database (Ohio Environmental Protection
Agency, Ecological Assessment Section) paired with
corresponding gaged and synthesized low-flow data.
Quantile regression (Cade and Noon 2003, Konrad et al.
2008), using USGS Blossom software, quantified the
flow-ecology relationships at the 95th percentile (Figure 1).
Currently, ODNR registers but does not otherwise regulate
large consumptive withdrawals. To comply with the Great
Lakes Compact, withdrawals will need to be managed
actively to prevent “adverse resource impact.” Through a
stakeholder small workgroup process, MBI and TNC
proposed to representatives of regulated industries and the
Ohio Chamber of Commerce that adverse impact be
defined in terms of percent loss of sensitive fish species.
The proposal specifies allowable losses of sensitive fish
species for each river type. For streams of highest
ecological quality (EWH, CWH), which contain the largest
number of sensitive fish species—including several that are
rare in Ohio—MBI and TNC proposed an allowable loss of
2%. For warmwater streams (WWH), which generally have
fewer sensitive species with less sensitivity to flow
alteration, they proposed an allowable loss of 10%. For
altered streams (MWH), which have few sensitive and no
rare fish, they proposed a 50% species loss threshold. A
proposed withdrawal rate that would approach the level at
Figure 1. Flow-ecology relationships for two of Ohio’s river types, EWH (exceptional warmwater habitat) and CWH (coldwater habitat).
Dotted lines indicate 95-percent quantile regressions.
44
which adverse species loss occurs would trigger agency
review of a water withdrawal application.
In this way, the flow-ecology curves allow the ODNR to
determine the cumulative amount of low-flow depletion
that would cause these predetermined unacceptable
losses. Only withdrawals that maintain cumulative flows
above threshold levels would be permitted automatically.
Proposed withdrawals that trigger the permit process
would be reviewed individually.
To calculate cumulative flow depletion, the ODNR, MBI,
and TNC recommend modifying the existing Ohio Stream
Withdrawal Evaluation Tool (OSWET). Currently, OSWET
calculates streamflow depletion due to an individual
withdrawal. In the future, OSWET also could calculate
cumulative depletion during September due to all local and
upstream withdrawals.
The proposed Ohio thresholds would provide the benefit of
protecting ecologically sensitive freshwater ecosystems,
while allowing future development in more resilient
ecosystems. Because the process developed for Ohio uses
existing river condition to classify river types and uses
current conditions as the baseline, it “grandfathers in”
existing water uses and sets no restoration goals at
present. Moreover, future withdrawals that cause
thresholds to be exceeded still could be approved after
agency review. Even so, regulated interests rejected the
proposal and sought alternative legislation to exempt most
withdrawals from regulation. Although the legislature
passed the industry-backed bill in spring 2011, Ohio’s
Governor vetoed it due to technical and legal shortcomings.
As of February 2012, it is not clear how the Ohio
Department of Natural Resources will comply with the
Great Lakes Compact.
ACKNOWLEDGMENT
The authors thank John Stark of The Nature Conservancy
for describing this project to us and reviewing this case study.
REFERENCES
Cade, B. S., and Noon, B. R. 2003. A gentle introduction to
quantile regression for ecologists. Frontiers in Ecology and
the Environment 1(8):412-420. doi:
10.1890/1540-9295(2003)001[0412:AGITQR]2.0.CO;2
Koltun, G. F., Kula, S. P., and Puskas, B. M. 2006. A
streamflow statistics (StreamStats) web application for
Ohio. U.S. Geological Survey Scientifc Investigations Report
2006-5312, 62 p.
Koltun, G. F., and Whitehead, M. T. 2002. Techniques for
estimating selected streamflow characteristics of rural,
unregulated streams in Ohio. U.S. Geological Survey
Scientific Investigations Report 02-4068, 50 p.
Konrad, C. P., Brasher, A. M. D., and May, J. T. 2008.
Assessing streamflow characteristics as limiting factors on
benthic invertebrate assemblages in streams across the
western United States. Freshwater Biology.
doi:10.1111/j.1365-2427.2008.02024.x
Ohio EPA. 2004. Summary of Ohio’s Beneficial Use
Designations (OAC 3745-1-07) - April 2004. http://www.
epa.state.oh.us/dsw/wqs/index.aspx#Beneficial%20Use%20
Designations
Rankin, E. T. 1989. The Qualitative Habitat Evaluation Index
(QHEI): Rationale, methods and application. Ohio
Environmental Protection Agency, Division of Water Quality
Planning and Assessment, Ecological Assessment Section,
Columbus, Ohio.
Rankin, E.T. 1995. Habitat indices in water resource quality
assessments in W.S. Davis and T.P Simon (editors).
Biological assessment and criteria: Tools for water resource
planning and decision making. Lewis Publishers, Boca
Raton, Florida, p. 181-208.
Rankin, E. T., Mueller, R., and Yoder, C.O. 2012. Ecological
low flow protection process for Ohio streams and rivers
of the Lake Erie basin. Prepared by the Midwest Biodiversity
Institute, Columbus, Ohio for The Nature Conservancy
in Ohio.
45
This case study demonstrates (1) the use of flow-ecology
relationships to determine groundwater withdrawal limits,
(2) ecological goal setting through river basin classification,
and (3) the adoption of an ecologically-based presumptive
flow standard.
The Rhode Island Stream Depletion Method (SDM) was
developed in 2010 from a series of initiatives that began in
1999, driven by an increasing need for water supply in the
state. In 1999, the RI General Assembly granted the Water
Resources Board (WRB) sole authority to devise a fair and
equitable allocation of water resources among users and
uses to ensure that long-range considerations of water
supply prevail over short-term considerations (Rhode Island
Gen. Laws §46-15.7). In 2002, the Water Resources Board
(WRB) formed the Water Allocation Program Advisory
Committee (WAPAC) and launched an inclusive water
allocation planning effort, which brought together 150
people from 66 participating organizations. The Committee
recommended, and in March 2004 the Board approved,
the establishment of the Streamflow Working Group, a
partnership between WRB and the Rhode Island
Department of Environmental Management (DEM) to
address streamflow issues such as aquatic base flow
and the further development of a statewide streamflow
gaging network.
The SDM stemmed from this recommendation. At the time,
most state instream flow methods in the Northeast were
focused on releasing water from storage reservoirs to
provide environmental flows. Very quickly it became clear
that methods such as U.S. Fish and Wildlife Service’s New
England Aquatic Base Flow (ABF) (U.S. Fish and Wildlife
Service 1981; Lang 1999) and the Rhode Island Aquatic
Base Flow (Richardson 2005) approach could not work for
groundwater withdrawal permitting, from a water supplier
perspective, because it could prohibit groundwater
withdrawals adjacent to a river with flows at or below
natural August median (i.e. much of the summer).
Water suppliers needed a reliable and predictable amount
of water to withdraw, particularly in summer, while the
resource agencies needed to define environmentally
sustainable limits on those withdrawals. Rhode Island has a
large reservoir, the Scituate, which supplies 60% of the
population; however, newer suburban development was
occurring outside of Scituate’s service area. Economic
pressure to continue developing water sources was forcing
water suppliers to find supplemental groundwater in places
where water was already being withdrawn for agricultural
needs, often at a substantial rate relative to the summer
baseflow that these aquifers discharge to RI streams.
As DEM struggled to set environmentally protective
withdrawal limits from permit to permit, criticism began
mounting that water permits were unpredictable and costly
and the process took too long. Under existing RI law,
wetland permits are needed from DEM for any new or
increased withdrawal of 10,000 gallons per day (gpd) or
more. Additionally, the permit conditions might mandate
water shutoffs during the lowest streamflow periods, which
were concurrent with the highest demand periods, an
unsupportable outcome for water suppliers. At the end of a
long, expensive process, applicants might ultimately receive
permits with limits that they did not anticipate or were
infeasible to implement.
DEM began looking at various frameworks to improve their
permitting process to meet the needs of both the water
suppliers and the ecosystem. Two journal articles and an
application concept by the Connecticut DEP (2009) played
a major role in the development of the SDM. First, Freeman
and Marcinek (2006) evaluated fishery response to surface
and groundwater withdrawals. Second, the ELOHA
framework (Poff et al. 2010) incorporated the concept of
balancing human and ecological needs for water by
differentiating the degree of flow alteration (i.e., allowable
depletion) according to ecological condition goals.
The SDM was developed by DEM and a “Streamflow
Subcommittee” of the Water Allocation Program Advisory
Committee (WAPAC) as part of their eighteen-month public
process. This Subcommittee included water suppliers,
scientists, federal agency, and state agency representatives
and outlived the WAPAC process. It was designed to
answer the question of what is “sustainable” water use
under Rhode Island law. In this way, DEM acted as a
technical advisor to the Water Resource Board, which is the
water allocation body.
METHODOLOGY
The SDM provides resource agencies with a withdrawal or
streamflow depletion allowance that establishes the volume
of water that can be extracted from a stream even during
dry conditions (whether as direct stream withdrawals or as
indirect groundwater withdrawals), while still leaving
sufficient streamflow to maintain habitat conditions
essential to a healthy aquatic ecosystem. This methodology
is currently being applied to all new or increased
groundwater withdrawals, and is being applied in the
2.9 Rhode Island Stream Depletion Method
By Alisa Richardson, Rhode Island Department of Environmental Management
46
Rhode Island Water Resources Board’s Statewide planning
process to assess build-out scenarios and the potential for
future water resource availability.
In order to arrive at a set of watershed condition goals for
water management, RI watersheds were classified based
on landscape and site characteristics (Figure 1). Each
sub-watershed was scored using characteristics defined
in a geographic information system. Higher scores were
assigned to higher degrees of human alteration. By using
a scoring approach, the state takes into consideration the
fact that all watersheds are not of equal ecological value
due to watershed characteristics and existing human
influences that may alter habitat and/or natural streamflow
characteristics within a watershed. Note that if a watershed
contained a coldwater fishery, then 10 points were
subtracted from the overall score, thereby allowing less
water to be withdrawn and providing a higher level of
protection via goal setting.
The total metric score was calculated for each sub-
watershed and then grouped into larger watersheds to
simplify management application. Each watershed was
assigned to one of the goal classifications based on its
total metric score, as shown in figure 2—the higher the
score, the higher the class, and the higher the presumed
degree of deviation from natural conditions. By applying
the biological condition gradient (Davies and Jackson,
2006), appropriate management practices could be applied
to each watershed.
The methodology allows for a fairly simple calculation of
allowable streamflow depletion by considering existing
withdrawals and returns, their locations within the
Figure 1. Representation of metrics and associated scores used in Watershed Goal Classification.
Figure 2. Final Watershed Goal Classification for each of the major
watershed in Rhode Island.
watershed, the time of year, the watershed characteristics,
and the natural low-flow conditions of the potentially
affected river or stream. The allowable depletion calculation
under the SDM takes into consideration the ecological
importance of seasonal flow variations. Since it was well
understood that the best management practices would
maintain the natural flow regime (Poff and Zimmerman
2009; Poff et al. 1997), seasonal ecological needs were
linked with natural streamflow patterns. Six hydroperiods
were defined to represent the biological need for seasonal
variability and four flow ranges were described to clarify
the typical flow range in each hydroperiod (Figure 3).
The final step in the methodology is defining the specific
streamflow depletion that is allowable and likely to
maintain the natural streamflow variations described by the
hydroperiods. The SDM is based upon studies conducted
in the southern Piedmont area of Georgia by the U.S.
47
Geological Survey and the U.S. Fish and Wildlife Service.
Freeman and Marcinek (2006) evaluated fish assemblage
responses to surface water withdrawals and instream
reservoirs, and documented that the richness of fluvial
specialists declined as permitted monthly withdrawal rates
increased above the volume equivalent to 50% of 7Q10.
They also found that increasing withdrawal rates increased
the odds that a site’s Index of Biotic Integrity (a measure of
the macroinvertebrate population health) score would fall
below a regulatory threshold of biological impairment. They
concluded that significant losses of river fish are associated
with withdrawal rates greater than 50% of the 7Q10.
DEM gleaned from Freeman and Marcinek (2006) that 50%
of the 7Q10 during the summer months (lowest flows) is
the greatest depletion that could protect ecological
function. DEM made the judgment that during non-
summer months, more water could be depleted through
withdrawals since there is almost always more water in the
system. Yet it is important to note that summer low flows
represent the major constraint to groundwater supply
development.
Table 1 shows allowable streamflow depletions (as a
percentage of 7Q10) linked to a set of Watershed Goal
Classes that apply statewide. The largest depletion
allowances apply to the highest classes. For example,
watersheds that are the most heavily altered (e.g., greatest
existing withdrawals, highly impervious, and no cold-water
fisheries) were assigned a Class 5 protection goal (Figures
1 and 2) and are allowed a reliable depletion of 50% of the
natural 7Q10 in summer months. Those watersheds that
are the least altered (e.g., cold-water fisheries, low
withdrawals, conservation land) were assigned a Class 1
protection goal and are allowed a reliable depletion of 10%
of the natural 7Q10 (or a de minimus amount) in summer.
APPLICATION
When an applicant applies for a new wetlands permit, the
SDM is used by the Rhode Island DEM wetlands permitting
group as a presumptive approach. The DEM has prepared
guidance for groundwater withdrawals (Richardson 2006)
and fills out a worksheet by which DEM can quickly
determine if the proposed depletion will be allowed without
further study. If the proposed increased withdrawal is at or
below the allowable depletion level (taking into account
existing uses), then the withdrawal request is reviewed for
other freshwater wetlands impacts and National Pollutant
Discharge Elimination System (NPDES) flow needs.
Figure 3. Annual Natural RI Hydrograph with associated hydroperiods and life stages of fish found in Rhode Island.
48
Wetlands biologists evaluate the analysis, and if it is
acceptable, then the permit is issued as requested. Impacts
to wetlands are assessed and typically approved if there
are no impacts to vernal pools and if the wetland
vegetation is mostly surface-water driven. If the applicant’s
increased withdrawal exceeds the allowable depletion, or
wetland impacts are unacceptable, then the applicant must
conduct a site-specific study, reduce the request, or be
subject to shutoffs. All decisions take into account
cumulative impacts of existing use to the extent possible.
This method is also being applied at the statewide planning
level. The WRB is evaluating future resources and
projections of build-out demand by comparing future
demand to the SDM for each watershed in Rhode Island.
This method has helped to identify locations where water
resources may constrain future development.
REFERENCES
Connecticut Department of Environmental Protection (DEP)
2009. Stream flow: the next two decades balancing human
use and ecological health. Hartford, CT. 18 p. http://www.
ct.gov/dep/lib/dep/water/watershed_management/
flowstandards/streamflow_next_2_decades.pdf.
Davies, S.P., and Jackson, S.K. 2006. The biological
condition gradient: a descriptive model for interpreting
change in aquatic ecosystems. Ecological Applications
16:1251-1266.
Freeman, M.C., and Marcinek, P.A. 2006. Fish assemblage
responses to water withdrawals and water supply
reservoirs in Piedmont streams. Environmental Management
38(3):435-450.
Lang, V. 1999. Questions and answers on the New England
flow policy. Concord, NH, U.S. Fish and Wildlife Service, 9p.
http://southeastaquatics.net/uploads/category/NE%20
Streamflow%20Policy%20by%20USFWS.pdf.
MONTH BIOPERIOD
NEED
HYDROPERIOD CLASS 1 CLASS 2 CLASS 3 CLASS 4 CLASS 5
OCTOBER Spawning &
Outmigration Medium - Low 20%40%60%80%100%
NOVEMBER
DECEMBER
Overwinter Medium 40%80%120%160%200%
JANUARY
FEBRUARY
Overwinter &
Channel
Forming
High 60%120%180%240%300%
MARCH
APRIL
Anadromous
Spawning High 60%120%180%240%300%
MAY Anadromous
Spawning Medium 40%80%120%160%200%
JUNE Peak
Resident
Spawning
Medium-Low 20%40%60%80%100%
JULY
AUGUST
SEPTEMBER
Resident
Spawning
Rearing &
Growth
Herring &
Shad Out
Low 10%20%30%40%50%
Table 1. Allowable streamflow depletion as a percent of 7Q10.
49
Poff, N. L., Richter, B. D., Arthington, A. H., Bunn, S. E.,
Naiman, R. J., Kendy, E., Acreman, M., Apse, C., Bledsoe, B.
P., Freeman, M. C., Henriksen, J., Jacobson, R. B., Kennen, J.
G., Merritt, D. M., O’Keeffe, J. H., Olden, J. D., Rogers, K.,
Tharme, R. E., and Warner, A. 2010. The ecological limits of
hydrologic alteration (ELOHA): a new framework for
developing regional environmental flow standards.
Freshwater Biology 55:147-170. http://www3.interscience.
wiley.com/cgi-bin/fulltext/122588390/PDFSTART.
Poff, N.L., and Zimmerman, J. 2009. Ecological responses to
altered flow regimes: a literature review to inform the
science and management of environmental flows.
Freshwater Biology doi:10.1111/j.1365-2427.2009.02272.x
Poff, N.L., Allan, J.D., Bain, M.B., Karr, J.R., Prestegaard, K.L.,
Richter, B.D., Sparks, R.E., and Stromberg, J.C. 1997. The
natural flow regime: a paradigm for river conservation and
restoration. Bioscience 47: 769-784.
Richardson, A.R. 2005. Modified Aquatic Base Flow
(RI-ABF) for Rhode Island. Rhode Island Department of
Environmental Management. Providence, RI.
Richardson, A. R. 2006. Pre-application guidance for those
seeking a groundwater withdrawal permit for >10,000 gpd.
Rhode Island Department of Environmental Management.
Providence, RI. http://www.dem.ri.gov/programs/benviron/
water/withdraw/pdf/guidance.pdf.
US Fish and Wildlife Service 1981. New England Aquatic
Base Flow. Concord, NH.
50
Figure 1.1 reflects the original conceptualization of the
ELOHA framework (Poff et al. 2010; Apse et al. 2008). This
framework can be summarized in four not necessarily
sequential steps:
• Building a Hydrologic Foundation and Assessing Flow
Alteration
• Classifying River Types
• Developing Flow-Ecology Relationships
• Defining Goals and Setting Standards
In practice, as our case studies show, the up-scaling of
environmental flows for policy and planning has taken
many forms. The ELOHA framework has guided many of
these processes, in others it has had less bearing, and in
all cases the processes are adapted to local circumstances
and data availability. This section of the report outlines
each step of the framework, reviews how different users
have adapted these steps to conform to their unique needs
and constraints, and provides general observations
regarding each step.
3.1 Understanding Water Availability:
Building a Hydrologic Foundation
Informed regional water management decisions require
knowledge of where, when, and how much water occurs in
all water bodies across the landscape. Thus, the ELOHA
framework is built on a “hydrologic foundation” of
streamflow data10 within a region. This information is used
to assess flow characteristics, classify river types, quantify
flow alteration, evaluate ecological responses to flow
alteration, and evaluate the status of sites relative to
environmental flow standards.
True to its name, the hydrologic foundation is indeed the
foundation of ELOHA and as such its construction is
considered the first step of the ELOHA framework. In
practice, few if any places have such a dataset in place at
the onset, and building it usually requires considerable
time and thought. To maintain the momentum generated
at the early stages of the project, many water managers
have successfully advanced other parts of the framework
while the hydrologic foundation is being developed, rather
than awaiting its completion before proceeding with
successive steps.
3.1.1 What is the Hydrologic Foundation?
The hydrologic foundation envisioned by Poff et al. (2010)
consists of two databases of daily streamflow time series
representing baseline and current conditions for every
analysis node over a common time period of at least 20
years to represent climate variability. For planning
purposes, databases of future streamflow scenarios also
may be created.
Analysis nodes ideally are located where ecological data
have been collected, where flow management actions such
as water allocation may be taken, where streamflow will be
monitored to ensure compliance with flow standards, and
above and below major river confluences.
Baseline conditions typically refer to minimally altered
conditions before major dams and diversions affected
hydrology. Understanding baseline flow conditions and
their natural range of variability is fundamental to
understanding ecological flow needs and the response of
ecosystems to hydrologic changes (Apse et al. 2008, Poff et
al. 1997). Baseline conditions also may represent prior land
cover and drainage conditions, depending on management
and restoration goals, data availability, feasible restoration
options, and political expediency.
Current conditions account for cumulative effects of dams,
surface-water diversions, groundwater withdrawals, return
flows, and other existing causes of flow alteration. Current-
condition flow data can be compared with baseline flow
data to calculate flow alteration at any analysis node.
Future conditions also may be modeled. For example, in the
Middle Potomac River basin, scientists are modeling
streamflow under five potential future scenarios, in addition
to baseline and current conditions. This way, stakeholders
can evaluate how different water management policies,
population growth patterns, and climate change are likely
to affect environmental flows.
Some adaptations of ELOHA use only current conditions,
relying on spatial variation across watersheds to infer the
range of potential conditions from least to most altered.
This allows for development of flow-ecology curves, but
limits the ability of researchers and decision makers to
understand the degree to which flows have already been
altered. Michigan and Ohio based their flow-ecology
curves exclusively on current flow conditions.
3.0 Getting Environmental Flows to Scale: An Overview of the Process
10 This report refers to streamflow data and flow alteration. Although our case studies used flow data to assess rivers and streams, ELOHA is equally applicable to lakes and wetlands, in which case water level data would substitute or supplement flow data, and water-level and hydroperiod alteration would be assessed.
51
3.1.2 Criteria for Hydrologic Model Selection
Regardless of the application, developing a hydrologic
foundation requires modeling. At many gaged sites,
existing time series need to be extended and baseline
conditions need to be simulated. At ungaged sites, the
entire time series for both baseline and current conditions
need to be simulated. Hydrologic modeling fills in these
gaps and provides a powerful foundation for management.
Various approaches can be used to model hydrology. The
model chosen depends on the project budget and schedule,
data availability, hydrology, ecology, and modeler expertise.
The finer the temporal and spatial scale, the more flexible
the model is likely to be for management. The databases
generated by hydrologic modeling need to have enough
spatial detail to resolve reaches with different streamflow
characteristics (e.g., because of an intervening tributary)
and small streams that nonetheless provide significant
habitats. The reach scale of the U.S. Geological Survey’s
National Hydrography Dataset (NHD+) meets these
criteria; moreover, the NHD+ provides a consistent spatial
platform for routing flows and for compiling and processing
other relevant data. Ideally, the model generates daily or
even sub-daily flow data.
Daily flow data allow for the calculation of ecologically
relevant flow statistics (Henriksen et al. 2006, Mathews and
Richter 2007). Where daily flow synthesis is impractical, the
model may generate weekly or monthly time series. If
groundwater discharge provides significant baseflow to
surface water, then it must be accounted for in the model.
Likewise, if rivers discharge into estuaries, then estuarine
inflows also should be modeled. As a general rule, the
period of record modeled should be at least 20 years to
account for normal climate variability. Kennard et al. (2009)
provide more rigorous guidance on selecting a period of
record for estimating hydrologic metrics for ecological
studies.
Although a daily flow time series is ideal, many states and
basins are having success using monthly or other time
steps. In addition, in some applications (e.g., Michigan and
Ohio), ecologically relevant flow statistics have substituted
for time series; however, this approach limits the ability to
analyze hydrologic alteration and to compare various
management scenarios.
Individual and cumulative impacts on streamflow of
surface-water diversions, groundwater withdrawals, and
return flows should be included in the hydrologic
foundation of current conditions. In states and basins with
adequate reporting, this accounting is not particularly
difficult or costly (e.g., Massachusetts). More difficult is the
inclusion of dam operations and land-use changes into
current condition hydrologic estimates. Both should be
modeled if feasible, but it is worth noting than none of our
case studies effectively incorporates both land use and
dam operation impacts. Finally, the ability to simulate
hydrologic impacts of climate change is useful for planning
applications.
To summarize, an ideal hydrologic model creates a
hydrologic foundation that will:
• be spatially comprehensive to capture regional-scale
hydrologic variability and to include locations where
water management decisions will be made and where
ecological data have been collected;
• have the smallest time step possible for management
needs;
• represent baseline (minimally altered), current, and
potentially future streamflow conditions;
• address groundwater and estuarine flows where
appropriate;
• be able to simulate new water uses and reservoir
operations;
• allow for the calculation of the range of ecologically
relevant flow characteristics; and
• simulate individual and cumulative effects of water use,
reservoir operations, and potentially land use and climate
change.
Every model has limitations: some excel at modeling high
flows, others simulate low flows better, still others capture
annual variability particularly well, and so on. The choice of
hydrologic model depends largely on the flow components
to which the subject ecosystems are most sensitive. For
example, Michigan and Ohio modeled only August or
September flows because those are the months when their
aquatic ecosystems are most sensitive to water
withdrawals, and their water withdrawal permitting
programs are designed to address these sensitive periods.
In contrast, for the Susquehanna and Connecticut River
basins, the purpose of the hydrologic foundation is to
evaluate the relative difference in flows between scenarios.
In these cases, consistency and accurate water accounting
may be more important than obtaining absolute flow values.
3.1.3 Components of the Hydrologic Foundation
The basic components of a hydrologic foundation of daily
streamflow data are hydrologic simulation and water use
accounting. Below we provide brief overviews of each
component and examples of their application to ELOHA.
52
Hydrologic Simulation
Hydrologic simulation is used to estimate streamflow
conditions. Three general approaches to hydrologic
simulation are drainage area ratio method, regression
modeling and process modeling. Drainage area ratio method
and regression modeling tend to be faster, simpler, and
therefore less expensive, whereas process modeling enables
evaluation of land use and climate change scenarios.
Drainage Area Ratio Method
One of the simplest approaches to developing minimally
altered streamflow data is the drainage area ratio method.
This method derives a daily flow time series for an ungaged
site by scaling the ratio of drainage area above that site to
that above a gaged index site with minimally altered flows
(Stedinger et al. 1992). This method is greatly limited by the
spatial and temporal availability of gage data from
minimally altered rivers that have similar watershed
characteristics to the ungaged site of interest. However,
this approach can be useful in areas of the country in
which the density of stream gages on unregulated streams
is relatively high and watershed characteristics do not vary
greatly. Linking an ungaged site to a gaged index site can
be as simple as using proximity and as complicated as
kriging (Kitanidis 1992). Connecticut River basin modelers
found the drainage area ratio method to estimate flood
flows more accurately than regression techniques.
StateMod, the hydrologic foundation for ELOHA in
Colorado, also uses the drainage area ratio method,
weighted by precipitation.
Regression Modeling
Various regression techniques have been used to estimate
flow statistics and examples abound. For decades, USGS
hydrologists have developed and published simple
regression equations for estimating selected hydrologic
statistics of local interest for water management. USGS
StreamStats is an online application that computes flow
statistics with regression equations for the user at any
location. StreamStats is being developed on a state-by-
state basis as funding becomes available. In Minnesota,
for example, StreamStats uses regression equations
(Lorenz et al. 2009) to estimate instantaneous peak flows
with recurrence intervals of 1.5, 2, 5, 10, 25, 50, 100, and
500 years.
Sanborn and Bledsoe (2005) developed a modified
regression approach that generates ecologically relevant
flow statistics for highly heterogeneous regions under
unaltered conditions. To account for heterogeneity, rivers
are first stratified into similar hydrologic groups before
developing regression equations specific to each group.
Their regression parameters include climate and watershed
characteristics. The method predicts flow magnitude,
timing, and rate of change metrics better than streamflow
variability metrics. This approach is being used to develop a
hydrologic foundation for predicting risk of invasive species
spread in river networks under various climate and dam
management scenarios (LeRoy Poff, Colorado State
University, written communication Feb.13, 2012).
Generally, regression modeling is a relatively inexpensive
approach for reliably estimating baseline conditions
statewide. It can generate a wide range of flow statistics,
often with low standard errors of prediction. However, its
versatility is limited. Apse et al. (2008) discuss caveats
regarding simple regression models, including their limited
ability to simulate extreme high and low flows and
extremely large and small catchments. Furthermore,
regression alone cannot generate daily flow series;
regression only calculates certain statistics that
characterize flow over a long time period. In contrast, with
a time series of data, hundreds of flow statistics can be
calculated and systematically reduced to those with the
most ecological relevance for a particular river type.
To overcome this limitation, Archfield et al. (2010)
developed a method that uses parameter-based regression
to estimate flow-duration curves, which are then
transformed into daily flow series for ungaged sites. A map
correlation method (Archfield and Vogel 2010) is used to
select reference gages for the regressions. The parameters
used for regression include precipitation, air temperature,
geologic substrate, percent of basin occupied by open
water, and other basin characteristics. Because physical
and climate processes affect portions of the flow-duration
curve differently, different variables are used to estimate
different streamflow quantiles. The Sustainable Yield
Estimator (SYE) used this technique to generate baseline
flow series for Massachusetts, Pennsylvania, and the
Connecticut River basin. As mentioned above, current-
condition flows are calculated by adding or subtracting
water use and, in the case of the Connecticut River basin,
reservoir release data to the baseline flows.
Michigan and Ohio used multiple linear regression and
quantile regression, respectively, to estimate low-flow
statistics for their flow-ecology models. Michigan added
flow routing and a program that calculates time-varying
streamflow depletion due to groundwater pumping (Barlow
2000), creating an online decision support system for water
withdrawal permitting. Apse et al. (2008) describe other
statistical approaches used to estimate ecologically
relevant streamflow statistics in Pennsylvania, Tennessee,
western United States, the United Kingdom, and elsewhere.
53
AFINCH (Analysis of Flows in Networks of CHannels) is a
new computer application that uses regression and water
accounting to generate monthly time series of current-
condition flows at the National Hydrography Dataset Plus
(NHD+) reach scale. Flows are accumulated and
conserved downstream through the NHD+ streamflow
network (Holtschlag 2009). Although AFINCH has not yet
been used for flow-ecology analysis, its fine spatial
resolution is amenable to coupling flow data with biological
sampling sites. A model using the AFINCH application is
currently being developed for the Great Lakes basin
through Great Lakes Aquatic GAP11. Like all regression-
based approaches, AFINCH is limited in its ability to model
land-use and climate changes, to represent areas with
karst or mined hydrogeology, and to simulate intermittent
headwater streams.
Process Modeling
Physical process modeling, also known as rainfall-runoff,
watershed, or hydrologic process modeling, tracks the flux
of water through the entire hydrologic cycle, accounting for
surface and subsurface watershed properties and weather.
Although these models can be complicated to construct
and calibrate, they can be used to simulate many different
types of scenarios, including climate and land-use change.
However, because of their complexity, process models
typically are applied to sub-watersheds that are smaller
than ELOHA’s intended geographic scope or, when they are
applied to large regions, their spatial discretization may be
too coarse for ELOHA. Hydrological Simulation Program
FORTRAN (HSPF), Precipitation Run-off Modeling System
(PRMS), Soil and Water Assessment Tool (SWAT) and
MIKE SHE12 are commonly used hydrologic process models.
Scale issues notwithstanding, process models have
generated hydrologic data for some ELOHA applications.
The Middle Potomac River basin project and the
Commonwealth of Virginia (see Commonwealth of Virginia
Flow-Ecology website13) built their hydrologic foundations
from an existing HSPF model, the Chesapeake Bay
Program Watershed Hydrology Model. Kennen et al. (2008)
used a process model called TOPMODEL to simulate daily
streamflow under baseline and current conditions for 856
mostly ungaged biological monitoring sites in New Jersey.
An empirically-based algorithm was added to improve
simulation of runoff from impervious surfaces. The
biological and hydrologic databases are now poised for
analyzing flow-ecology relationships.
Water Use Accounting
A full hydrologic foundation that includes baseline and
current-condition hydrographs employs water use
accounting, regardless of the hydrologic simulation
approach. By adding and subtracting water withdrawals and
return flows to streamflow, water use accounting estimates
the impact of water use on streamflow conditions.
The Middle Potomac River basin and Virginia14 used
process modeling to estimate current-condition flows, then
added and subtracted withdrawals and discharges to
generate baseline-condition hydrographs. Coming from the
other direction, Massachusetts and the Susquehanna River
basin used regression to estimate baseline flows, then
subtracted and added withdrawals and discharges,
respectively, to generate current-condition hydrographs.
Moreover, the routing function of their water accounting
module enables regression-based models to calculate
cumulative effects of upstream water uses at any site.
Hydraulic flow routing and reservoir operation
modeling improve model accuracy by accounting for the
time delays of downstream water movement due to
channel characteristics and dams, respectively. The
WOOOMM model, which provides the hydrologic
foundation for the Middle Potomac River basin, includes
hydraulic flow routing. The Corps’ HEC-ResSim model
incorporates reservoir operations and flow routing into the
Connecticut River basin model.
3.1.4 General Observations and Summary
Table 3.1 summarizes some of the main strengths and
limitations of various approaches for developing a
hydrologic foundation.
A hydrologic foundation need not be completed at the
onset of the project. For the State of Connecticut and the
Susquehanna and Connecticut River basins, scientists and
managers recommended environmental flow ranges based
on conceptual models extracted from literature,
professional judgment, and analysis of flows at existing,
minimally-altered gages to determine baseline flow
variability. However, modeled daily flow data will be needed
to implement their recommendations. In the Susquehanna
River basin, proposed withdrawals and dam operations will
be evaluated to determine whether they could alter
streamflow beyond the recommended ranges. In the
Connecticut River basin, streamflow and environmental
flow targets feed into a model that compares different
multi-dam operation scenarios. Therefore, both projects
currently are building hydrologic foundations of baseline
and current-condition daily streamflow series.
Regardless of the modeling approach ultimately selected,
the hydrologic foundation can only be as accurate as the
water-use (withdrawal and return flow) data that go into it.
Ideally, all major surface and groundwater withdrawals are
reported accurately. In practice, many states lack water use
11 http://cida.usgs.gov/glri/projects/accountability/watershed_modeling.html
12 http://www.crwr.utexas.edu/gis/gishyd98/dhi/mikeshe/Mshemain.htm
13 http://sifn.bse.vt.edu/sifnwiki/index.php/Commonwealth_of_Virginia_Flow-Ecology
14 Ibid.
54
reporting programs and many more have incomplete or
inaccurate water use data. Withdrawal and discharge
permits can be used as surrogates, but with caution because
the difference between permitted and actual water use can
be substantial. Therefore, there is a great need to develop
and improve state water use reporting programs while also
improving remote techniques for estimating the locations
and timing of water withdrawals and return flows. Advocacy
for state water use reporting laws along with federal and
state investments in water use estimation continues to be
necessary, particularly in states challenged by significant
irrigation use or limited water management programs.
Periodic USGS water use reports (e.g., Kenny et al. 2009)
are insufficient for these purposes, as they compile
reported monthly or annual data at the county level, despite
the lack of actual reporting requirements in many states.
Furthermore, disaggregating these data by day and by
stream reach requires assumptions about water use
patterns. For regional policy and planning applications,
one approach in the absence of accurate water use
reporting is to research individual large water uses to
obtain the most accurate data possible, and to estimate
the smaller water uses.
Likewise, reservoir operation rules and actual releases can
be obtained directly from willing dam owners and
operators. Yet the time required to get this information can
be considerable; in the case of the Connecticut River basin,
researchers spent 1.5 years meeting with dam owners to
understand their operations.
The treatment of interactions between groundwater and
surface water depends on the type of hydrologic model
and on the hydrogeology. Most process models incorporate
groundwater flow, and do not require additional
programming to simulate interaction with surface water.
Regression-based and simple water-accounting models
may warrant additional programming. In bedrock-
dominated systems, where runoff is the main control on
streamflow patterns and groundwater is not a significant
water supply, groundwater may not need to be modeled.
In narrow alluvial valleys, groundwater withdrawals may
be assumed to deplete nearby streamflow directly and
immediately. That is the default assumption of the
Massachusetts SYE and of regression-transform
approaches in Pennsylvania and the Connecticut River
basin. Between these two extremes, the surface-water
hydrologic model can be linked to a groundwater model
Approach Examples Strengths Limitations
Drainage-area ratio method StateMod (Colorado), flood
flows (Connecticut River basin)
Low cost, easy to generate.Limited accuracy if index
gages are sparsely located or
do not represent the natural
range of flow regimes.
Regression-generated monthly
statistic
Median August flow
(Michigan), mean September
flow (Ohio)
Low cost, easy to generate,
widely accepted.
Current-condition only. Not a
time series. Represents only
one environmental flow
component.
Regression with water
accounting and flow routing
U.S. Geological Survey (USGS)
AFINCH (No ELOHA case
study)
High spatial resolution; linked
to NHD+.
Monthly time series only. Has
not been tested outside Great
Lakes basin.
Duration-curve regression plus
water accounting
USGS Sustainable Yield
Estimator (SYE)
(Massachusetts, Pennsylvania)
Relatively low cost, easy to
generate. Daily time step.
Difficult to simulate flows at
hydrograph and basin-size
extremes. Has not been
applied outside eastern US.
Duration-curve regression plus
dam operations model
USGS SYE plus US Army
Corps of Engineers HEC-DSS
(Connecticut River basin)
Same as above, with ability to
model dam releases.
Relatively time-consuming
(several years) to develop;
example required two federal
agencies.
Hydrologic process model plus
water use accounting and
channel routing
WOOOMM (Watershed Online
Object Oriented Meta-Model)
(Potomac River basin)
Can model land-use and
climate change.
Resolution typically too coarse
or area too small for regional
application without
modification.
Table 3.1 Strengths and limitations of selected approaches for developing a hydrologic foundation, listed in approximate order of effort and
expense. All approaches listed include water accounting. Case studies (section 2) elaborate on the examples listed.
55
as simple as the STRMDPL program used in Michigan or
as comprehensive as MODFLOW (Harbaugh 2005).
Massachusetts SYE users have the option of linking to
existing STRMDPL models instead of using the default
assumption in certain parts of the state.
3.2 Classifying River Types
This section summarizes a variety of river system
classification approaches for regional environmental flow
assessment, and defines the key parameters used for
classification. We refer here to natural system classification
(ELOHA’s science process), rather than to goal classification
for purposes of management or establishing standards
(ELOHA’s social process). Ideally, river type classification
should result in a relatively small number of river types that
capture the major dimensions of streamflow-related
biological variability within a region. Section 3.4.1 covers
goal classification.
3.2.1 Why Classify River Types?
Conceptually, river type classification extrapolates
understanding of ecohydrologic conditions at sites that
have been studied to similar sites that have not. The first
reason to classify river types is to strengthen the statistical
significance of flow-ecology relationships by combining
available information from many rivers. The second reason
is to extend those flow-ecology relationships to other rivers
of the same type in order to define their environmental flow
needs. A third reason is to direct future monitoring efforts
to improve the strength of initial flow-ecology relationships
or to extrapolate site-specific monitoring results. Apse et al
(2008) covers classification for environmental flow
standards in some detail.
Researchers have developed classification systems for
Australia (Kennard et al. 2010, Pusey et al. 2009),
Washington (Reidy Liermann et al. 2011), Canada (Monk et
al. 2011), New Jersey (Hoffman and Rancan 2007, Kennen
et al. 2007), Missouri (Kennen et al. 2009), Texas (Hersh
and Maidment 2007), Pennsylvania (Apse et al. 2008),
southeastern US15, and elsewhere, all intended to meet the
needs of ELOHA. In other places, such as South Africa,
researchers have been classifying rivers for similar
purposes since the 1980s.
Recent practice has demonstrated that river type
classification is not always needed for setting scientifically-
defensible environmental flow standards. In Massachusetts,
for example, a statewide regression relationship links
relative abundance of fluvial fish to watershed
characteristics (area, gradient, etc.), obviating the need to
classify aquatic system types according to those
characteristics. In the Connecticut River basin, the project
team decided that small differences between rivers within
the project area did not warrant their being subdivided by
type. Analyses by the Middle Potomac project team indicate
that segregating rivers by type does not significantly
strengthen their statistical relationships, and in fact could
weaken them by reducing the number of data points per
analysis. Classifying watersheds may help reduce
variability, but classification also reduces sample size,
which increases uncertainty.
Other researchers have found river type classification to
be useful. In New Zealand, Snelder et al. (2011) report
that flow-ecology relationships (represented by habitat
availability) vary among major river types defined by
morphology and flow regime. In Michigan, classifying
rivers according to water temperature and catchment size
protects the fish communities that are most sensitive to
streamflow depletion.
3.2.2 General Approaches to Classification
Poff et al. (2010) envisioned rather sophisticated, time-
intensive river classification systems being fully developed
for ELOHA before being tested by flow-ecology analysis.
For example, Reidy Liermann et al. (2011) used Bayesian-
mixture modeling, a recursive partitioning algorithm,
random forests, and a geomorphic classification to create
a 14-tier hydrogeomorphic classification for Washington,
in preparation for flow-ecology analysis, which is
currently underway.
In practice, river type classification for ELOHA tends to
be iterative or to use pre-existing classes. The iterative
analytical approach is well-illustrated in the Middle
Potomac River basin project. The first iteration, based on
hydrologic analysis and habitat type, classified river
reaches according to watershed size and karst geology.
This informed the flow-ecology analyses, which in turn
informed re-classification. In the end, biological and
hydrologic metrics were normalized so that data from all
sites could be combined, thereby maximizing the size of
the datasets used to quantify flow-ecology relationships.
The Colorado project used a pre-existing classification.
Rivers were classified by ecoregion, using a classification
system that was already established. Literature review and
flow-ecology analyses confirmed that this simple typology
sufficiently captures eco-hydrologic variability of Colorado’s
river systems, especially considering the very limited
databases with which the analysts had to work.
Using an existing classification system not only saves time,
but also may help link streamflow management to regulatory
programs that already are in place. By adopting Aquatic
Life Use classes from an existing water quality program, the
15 http://sifn.bse.vt.edu/sifnwiki/index.php/Main_Page
56
Ohio project team deflected water users’ concerns that
biological flow criteria would create another layer of
regulation. Moreover, the Ohio researchers were able to use
extensive biological databases associated with the existing
water quality program to develop flow-ecology relationships.
Two of the river types for the Susquehanna River basin also
borrowed from a water quality regulatory program.
3.2.3 Parameters Used for Classification
River classification for ecohydrologic analysis is becoming
increasingly sophisticated, a trend apparently accelerated
by publication of Poff et al. (2010). Olden et al. (2011)
provide an excellent review of the full spectrum of
approaches and their respective applications. Here, we
focus on the parameters and approaches that these case
studies have used to support flow-ecology analyses and
flow criteria development.
Hydrology
As shown in Figure 1.1, Poff et al. (2010) suggested
classifying rivers initially according to their hydrology.
This is accomplished using hydrologic statistics calculated
from daily streamflow data with Indicators of Hydrologic
Alteration (IHA; The Nature Conservancy 2009),
Hydroecological Integrity Assessment Process’ (HIP)
Hydrologic Index Tool (HIT; Henriksen et al. 2006), or similar
software. The HIT software calculates statistics and the HIP
process lays out an approach for using them to classify
river types. Briefly, principal components analysis
eliminates redundant statistics, and cluster analysis then
groups the remaining data. In this way, the HIP process
determines which hydrologic parameters are most
appropriate for classifying river types.
A HIP classification was conducted for Pennsylvania (Apse
et al. 2008). Of the five river types that HIP delineated, the
Susquehanna project team incorporated one (baseflow-
dominated streams) in the final classification because it
represented an ecologically important hydrologic type that
other classifications did not capture. In fact, none of our
case studies adopted classifications based strictly on
hydrology, although, like the Susquehanna project,
preliminary hydrologic classifications did inform final river
types in several.
Typically, hydrologic statistics that are used to classify river
types differ from the metrics used to express environmental
flow criteria, as described in section 3.3.2.
Water Temperature
Olden and Naiman (2010) argue for using water
temperature in environmental flow assessments, especially
where reservoir releases greatly alter natural temperature
regimes. Water withdrawals, too, can affect water
temperature to the extent that biological communities
transform. This is certainly the case for coldwater streams
in the upper Midwest; both Michigan and Ohio captured
this phenomenon by incorporating water temperature into
their river type classifications. Michigan’s new water
withdrawal permitting system is designed intentionally to
keep coldwater streams cold by maintaining sufficient
(cold) groundwater discharge into their channels. Many
state water quality programs routinely monitor water
temperature, so ample data may be readily available.
Ecoregion
Freshwater ecoregional classification seeks to identify
critical areas for conservation by capturing representative
components of freshwater biodiversity (Higgins et al. 2005).
Although not developed expressly for flow management,
ecoregional classification is based on many of the same
factors that influence flow-ecology relationships. In Colorado,
a coarse, high-level ecoregional classification (CEC 1997)
proved adequate for distinguishing river types for flow-
ecology relationships and accelerated the project timeline.
Watershed and Macrohabitat Characteristics
The influence of river size on flow-ecology relationships is
well established (Higgins et al. 2005). Michigan,
Susquehanna, and Middle Potomac River basin
classifications incorporate catchment area, which strongly
correlates with river size. Other useful watershed
characteristics for defining river types include land cover,
geology, climate, geomorphology, topography, water quality,
and elevation. The value of using watershed characteristics
to distinguish river types is determined by the ability of
resulting river types to strengthen the statistical
significance of flow-ecology relationships. In addition,
watershed characteristics may be used to classify rivers for
which flow data are unavailable.
In the Susquehanna River basin, the Northeast Aquatic
Habitat Classification (Olivero and Anderson 2008)
informed river type classification. This classification system
incorporates watershed characteristics such as geology
and river size, as well as macrohabitat variables such as
modeled stream temperature and gradient. Using an
established, credible classification system accelerated the
Susquehanna project timeline by precluding the need to
develop and defend a new system.
3.2.4 General Observations
In the literature and in our case studies, several aspects of
river type classification differ in practice from the
conceptual framework proposed by Poff et al. (2010). Some
observations are:
57
• River type classification is intended to strengthen the
statistical strength of flow-ecology relationships;
therefore, it may be more effective if conducted in
tandem with, rather than before, flow-ecology analyses.
• Natural river type classification is not always necessary
to achieve management objectives, even if it may
improve scientific defensibility of flow-ecology
relationships. The Massachusetts case study illustrates
this point.
• Given the goal of environmental flow standard
development, hydrologic regime classification can clearly
play an important role. But as the case studies
demonstrate, other factors such as water temperature,
stream geomorphology, ecoregion, and river size are
important determinants of flow-dependent ecosystem
composition and functions that should also be
considered in classification.
• Incorporating watershed and macrohabitat variables
from an existing framework such as the Northeast
Aquatic Habitat Classification can improve the strength
of flow-ecology relationships while building credibility
among river science experts who recognize it.
• Many of the hydrologic classification applications that
have been used to date lack a spatial extrapolation step
for mapping hydrologic types to ungaged stream
reaches. Integrating hydrologic classification within an
aquatic habitat classification framework can enable such
extrapolation and facilitate the classification’s use in
developing flow-ecology relationships, selecting
monitoring sites, and managing water use.
• Regardless of the classification approach used, field
scientists who are familiar with the rivers should review
the resulting river types before the classification is
finalized.
Overall, these variations on the ELOHA framework
demonstrate the value of being flexible in how we approach
natural system classification for developing flow-ecology
relationships and setting environmental flow standards.
3.3 Describing Flow-Ecology Relationships
Relationships between flow alteration and ecological
response are grounded in the biological condition gradient
(Davies and Jackson 2006), which recognizes that
increasing degrees of anthropogenic stress lead to
decreasing ecological condition. Flow–ecology relationships
may be expressed in various forms, depending on the
information available and the results required: as an
ecosystem attribute (E) as a function of the change in
hydrologic condition (Q) from baseline (Q/E), as a change
in ecosystem attribute from a reference condition as a
function of change in hydrologic condition (Q/E), or as
an expected status of an ecosystem attribute as a function
of the value of a hydrologic metric (Q/E). Sanderson et al.
(2011) used the latter two forms to build the Watershed
Flow Evaluation Tool for Colorado. In some cases, these
relationships include habitat as an intermediate variable,
as seen in the site-specific DRIFT (King et al. 2003) and
PHABSIM (Milhous et al. 1989) approaches and in the
regional-scale Ohio case study.
Poff et al. (2010) envisioned a progression from hypothesis
development to data assembly and analysis to build these
relationships (Figure 1.1). In practice, successful projects
appear to follow this progression, with the information
available and the implementation mechanism influencing
the relative emphasis on quantitative versus qualitative
approaches.
3.3.1 Hypothesis Development
Regardless of the analytical approaches ultimately used,
development of flow-ecology relationships begins with
hypotheses derived from the literature and expert input
about how each flow metric or environmental flow
component (Mathews and Richter 2007) influences
physical, chemical, and particularly biological processes
within a river type. Subsequent quantitative analyses are
then designed specifically to test these hypotheses.
Well-supported flow-ecology hypotheses also can lead
directly to policy development if time, budget or data
constraints prohibit the development of quantitative
flow-ecology relationships. The State of Connecticut based
statewide reservoir release criteria on best professional
judgment and the extrapolation of quantitative relationships
from another state.
The Susquehanna River basin project team introduced a
structured approach for developing consistently worded
hypotheses in an expert workshop setting (DePhilip and
Moberg 2010). The Connecticut River basin project
adopted a similar approach. The objective in both cases
was to capture systematically the entire spectrum of
taxonomic groups and physical processes across the entire
flow regime. In each case, experts were asked to express
hypotheses that answer the questions:
• Who (species or group of species)
• What (flow magnitude or event)
• When (month or season)
• Where (river type and habitat)
• Why/how (ecological response)
58
For example, “If summer (when) low-flow magnitude
(what) decreases in baseflow-dominated streams (where),
then water temperature will increase (why) and salmonid
populations (who) will decline (how).” To facilitate
hypothesis development, project scientists displayed
flow-dependent life stages of native species for each river
type superimposed on “typical” hydrographs. An example is
shown in the Susquehanna case study.
In the Susquehanna, as in every large region, insufficient
quantitative data were available to test every hypothesis.
Yet, the literature conveyed that every ecosystem and flow
component in the Susquehanna is important for
maintaining ecological integrity, and no single data-rich
species or guild could adequately represent the others.
The experts agreed that to be scientifically defensible, their
recommendations had to preserve both the inter- and
intra-annual flow variability needed to protect the entire
ecosystem. The only way to do that in the limited time
allotted was to base environmental flow recommendations
primarily on the literature review and their best professional
judgment. The resulting recommendations are linked
explicitly to their underlying hypotheses so that they may
be tested quantitatively in the future.
In Massachusetts and in the Middle Potomac River basin,
hypothesis development was not a formal part of the
process. Yet, flow-ecology relationships developed for
Massachusetts are directly informing environmental policy as
intended. In the Middle Potomac, policy adoption of these
relationships is yet to be seen. It appears that the main
factor affecting policy adoption is whether a policy nexus
exists and is seized upon from the onset, regardless of the
hypotheses (or lack thereof) driving flow-ecology analyses.
3.3.2 Quantitative Analysis
The Michigan, Ohio, Massachusetts, and Middle Potomac
project teams decided that scientific defensibility of their
policy applications required quantitative analysis of
extensive databases. All four had large biological
databases with which to work, and carried out systematic
processes for selecting the parameters that ultimately
would define their flow-ecology relationships.
Our case studies support Apse et al.’s (2008) conclusion
that to be useful for management ecological metrics
selected for analysis should be:
• sensitive to flow;
• meaningful indicators of river health;
• broadly distributed spatially in a variety of watershed
types and sizes, along a gradient of flow alteration; and
• recently sampled (to pair with current flow conditions).
Hydrologic metrics should:
• represent natural variability in the flow regime;
• be sensitive to change and have explainable behavior;
• be easy to calculate and replicable;
• have conceptual and empirical linkages to ecological
response
• be easy for non-hydrologists to understand; and
• be non-redundant.
The Susquehanna River basin project developed an
innovative set of hydrologic metrics to meet these criteria.
Ten flow statistics to represent the frequency, duration, and
magnitude of ecosystem-dependent flow components were
based on monthly exceedance values and magnitude and
frequency of high flow and events. Additional metrics to
maintain the temporal distributions of flow were based on
seasonal flow-duration curves.
Ecological condition of a river is the result of many factors,
of which flow regime is only one. A focus of recent
research is to isolate the influence of flow alteration from
that of other environmental stressors, and then to identify
the flow and ecological metrics that best describe
ecological response to flow alteration. Several statistical
techniques facilitate this analysis.
Multivariate statistical analysis can identify the
environmental parameters that most strongly correlate with
observed variation in ecological indicators. When indicators
of hydrologic alteration are among the parameters
analyzed, their importance relative to other stressors can
be evaluated. This preliminary analysis greatly reduces the
universe of flow statistics for subsequent flow-ecology
analysis and builds confidence that these relationships will
be minimally obscured by other factors. It also can be used
to identify hydrologic metrics for defining environmental
flow recommendations or standards. The Middle Potomac
River basin scientists used multivariate statistical analyses
to select non-redundant hydrologic and ecological metrics
that define statistically significant flow-ecology relationships.
Kennen et al. (2010) used multivariate methods to identify
a subset of eight ecologically relevant hydrologic variables
describing streamflow magnitude, frequency, duration,
timing, and rate of change to explaine variation in
macroinvertebrate assemblage composition across the
339,290-km2 northeastern United States. The study used
physical, chemical, and biological data collected as part of
the National Water-Quality Assessment Program and
landscape characteristics from the National Land Cover
Database. Principal component analysis (PCA) and partial
collinearity assessment reduced 527 environmental and
59
land-use variables initially analyzed to a subset of 52
variables that accounted for the most variance in
macroinvertebrate assemblage, while minimizing
redundancy and reducing the effects of natural variation.
Conditional multiple linear regression was then used to
quantify relationships between the remaining 52 variables.
From this analysis, significant bivariate relationships were
developed to depict relationship between macroinvertebrate
assemblage structure and 8 hydrologic variables.
Several other studies have similarly used multivariate
statistical analyses to select hydrologic statistics. Using
generalized linear modeling, Armstrong et al. (2010)
quantified fish response to several hydrologic statistics in
Massachusetts. Using multiple regression analysis, Kanno
and Vokoun (2010) showed that water withdrawal rate was
more important than other natural and anthropogenic
factors (e.g. land cover and stream size) in explaining
several fish assemblage metrics. After using multivariate
analysis to eliminate hydrologic parameters associated with
anthropogenic disturbance, Kennen and Riskin (2010)
found significant linear and curvilinear bivariate flow-
ecology response relationships for fish and invertebrate
assemblages in the New Jersey Pinelands. Kennen et al.
(2007) combined watershed modeling and indirect
ordination techniques to identify components of the
hydrologic regime that most significantly affect aquatic-
assemblage structure across a disturbance gradient.
Important variables included the average number of annual
storms producing runoff, ratio of 25% to 75% exceedance
flows (flashiness), diversity of natural stream substrate, and
the percentage of forested land near the stream channel
(forest buffer). Knight et al. (2008) analyzed hydrologic time
series to identify three hydrologic metrics essential to
habitat suitability and food availability for insectivorous fish
communities in streams of the Tennessee River Valley:
constancy (flow stability or temporal invariance), frequency
of moderate flooding (frequency of habitat disturbance),
and rate of streamflow recession. Roy et al. (2005)
quantified relationships among fish assemblage metric
response, hydrologic variables, and imperviousness in small
streams and their subcatchments in Georgia.
Classification and Regression Tree (CART) and
Boosted Regression Tree (BRT) are statistical methods
that identify threshold values for explanatory variables that
separate groups of response variables. Carlisle et al. (2010)
used CART to relate two indicators of altered hydrology—
streamflow depletion and streamflow surcharge—and
aquatic biological community impairment across the
conterminous US compared to eight other covariates
(water temperature, specific conductance, pH, total
nitrogen, total phosphorus, channel gradient, agricultural
land cover, and urban land cover of the riparian buffer).
The degree of alteration (depletion and surcharge) was
estimated based on regression models using landscape
and watershed variables to predict flows at reference
gages versus gages with highly modified upstream
conditions.
Quantile-regression and other modeling can be used
to quantify bivariate flow-ecology relationships from large
datasets that represent sites affected by multiple stressors.
The premise is that scattered flow-ecology data are
bounded by “floors” and “ceilings” that represent the
maximum ecological condition that could be achieved at
any given flow value if all other stressors were absent (Cade
and Noon 2003, Konrad et al. 2008). Regression is used to
quantify the decline in maximum ecological condition as
flow alteration increases. Using the 90th instead of the
100th percentile accounts for some uncertainty. Quantile-
regression splines may be used to characterize changes
along nonlinear response curves (Anderson 2008). The
Colorado, Massachusetts, Middle Potomac, and Ohio case
studies illustrate the use of quantile-regression modeling to
define flow-ecology relationships.
In Michigan, scientists studied large fish and flow
databases, along with other habitat suitability information
(catchment size, base flow yield, July mean temperature),
to develop predictive models of fish assemblage structure
under a range of base flow reductions (Zorn et al. 2008).
These models then generated flow-ecology curves for
water withdrawal permitting.
3.3.3 Hybrid Approaches
Relying on existing biological databases limits flow-ecology
analyses to a subset of a complex ecosystem. Likewise,
relying on a single flow metric limits analyses to a subset
of a complex hydrologic pattern. Conversely, basing flow
recommendations on conceptual models may pose
credibility issues in a controversial political milieu.
In Colorado, scientists blended the best of both. Lacking
large ecological databases, the withdrawal thresholds that
populate Colorado’s Watershed Flow Evaluation Tool are
based on literature review and expert input. In this case,
the experts not only helped develop flow-ecology
hypotheses, but they also suggested how to use the very
limited data found in the literature to test those hypotheses.
Analytical approaches ranged from categorical threshold
delineation to quantile regression, depending on the form
and quantity of data available. Ultimately, flow-ecology
relationships were quantified for warmwater and coldwater
fish, invertebrates, riparian vegetation, and recreation.
Many of those were based on only a handful of sites, which
are assumed to represent their river type. Camp Dresser &
McKee Inc. et al. (2009, Appendix B) document the specific
approach used to quantify each flow-ecology relationship
that they generated.
60
3.3.4 General Observations
Developing flow-ecology hypotheses (or conceptual
models) is an important step in developing regionalized
environmental flow standards or targets. Well-vetted
flow-ecology hypotheses guide the selection of both
hydrologic and ecological metrics. These hypotheses
benefit subsequent data analyses by targeting the
relationships that are examined and by increasing the
transparency of the scientific process to stakeholders.
The flow-ecology hypotheses themselves can be used in
a structured expert workshop setting to arrive directly at
environmental flow recommendations. For example, in
the Susquehanna River basin, workshop leaders used
qualitative flow-ecology hypotheses and quantitative
analyses of the natural flow variability to facilitate expert-
driven flow recommendations.
The scientific literature is replete with statistical analyses
that relate flow alteration to ecological response. Yet very
few of these have been used to support water management.
Water managers and other stakeholders are more likely to
respond to science that uses ecological and hydrologic
metrics that they can explain and understand their causal
links. A weighted-evidence approach (Norris et al. 2012)
can help stakeholders understand the strength of causal
links supporting different flow-ecology relationships.
Environmental flow projects in the northeastern U.S. and
Australia are piloting this approach to assess its application
to watershed management.
The case studies demonstrate that because biological and
hydrological databases are limited, a combination of
quantitative and qualitative approaches is necessary to
create flow-ecology relationships that are truly holistic. By
holistic, we mean relationships that incorporate a range of
environmental flow components (e.g., small floods as well
as summer low flows) and ecological target groups (e.g.,
invertebrates and vegetation as well as fish). Collaborative
and interdisciplinary teams, engaged from the start of the
scientific process, enhance the likelihood that the
information policymakers need is delivered through a
balance of quantitative analysis and expert workshop-
driven products informed by flow-ecology hypotheses.
3.4 Making Flow-Ecology Relationships
Operational: Applying Environmental Flow
Science at a Regional Scale
Sections 3.1 - 3.3 discussed the development of flow-
ecology relationships. Figure 3.1 illustrates the use of such
relationships to translate an ecological condition goal
(y-axis) into an environmental flow criterion (x-axis). In this
section, we discuss ways to establish those ecological
condition goals, and then how such goals are being
pursued in policy and on the ground.
Environmental flow criteria cannot be defined by science
alone. Scientists can quantify the tradeoffs (flow-ecology
relationships) that underlie their definition, but the criteria
themselves are social decisions about the desired
ecological condition of various water bodies.
Consider water quality standards as an analogy. Scientific
analyses determine the concentration of a pollutant that
has a risk of killing one in a million people who ingest it,
the concentration with a risk of killing one in a thousand,
and so forth. Statutes or rules state the allowable risk
associated with ingesting pollutants based on societal
tolerance, feasibility of removing the pollutant, economic
costs, and other factors. The water quality standard is the
allowable concentration associated with that socially
determined risk.
Now consider environmental flow criteria. Scientific
analyses determine the degrees of flow alteration
associated with various levels of ecological degradation.
These relationships are expressed as flow-ecology
response curves. Statutes, rules, or perhaps guidelines
state the allowable level of ecological degradation based
on societal tolerance, existing water uses, and other
factors. The environmental flow standard is the degree of
flow alteration associated with that level.
Figure 3.1 depicts the social outcome with a simple
flow-ecology curve that relates mid-summer water
withdrawals (flow alteration) to fish community structure
(ecological condition) for one type of river in Michigan.
A technical advisory committee recommended, and the
legislature then codified, a state map showing all the water
bodies expected to achieve “acceptable” ecological
condition, which they defined as maintaining a certain
percentage (90% in Figure 3.1) of their native fish species.
Based on the flow-ecology curve, the flow standard, or
criterion, then, is 45% of natural mid-summer flow. Water
managers tasked with achieving the ecological goal now
manage water withdrawals and dam operations such that
no more than 45% of natural mid-summer flows are
diverted from these rivers. Scientists periodically monitor
the fish community to ensure that the flow standard
achieves its ecological goal.
Based on our case studies, two useful steps help put
flow-ecology relationships into practice for environmental
criteria. First, define ecological condition goals, or risk
levels, in terms of the ecological metrics used in the
flow-ecology response models. For example, what range of
invertebrate richness indicates a high level of risk of
ecological degradation? What range represents low risk?
This decision could begin with ecologists proposing
threshold levels. After public consultation, it could
culminate with formal adoption through an appropriate
61
legal process. In this conception, the flow-ecology curves
act to translate ecological condition goals into
environmental flow criteria. Hydrologists also may use
models to help water users understand implications of the
proposal on current or future water availability. As with
water quality standards, plans can be made for monitoring,
periodic review, and risk level revision as new information
becomes available.
Second, determine policy actions associated with each
ecological risk level. For example, if a proposed water
withdrawal has a low risk of harming the ecosystem,
can it be approved immediately? If its risk is high, will the
proposed withdrawal automatically be denied? Or, if the
risk is already high, will that river reach be prioritized for
water right transactions?
Sections 3.4.1 and 3.4.2 further explain how these policy
decisions are made, and how they are being implemented
in various water management contexts.
3.4.1 Establishing Ecological Condition Goals and
Defining Acceptable Risk
Underlying all environmental standards is the concept of
condition goals, also known as desired future condition.
Although some jurisdictions may choose to have a single
goal and associated environmental flow standard, most will
follow the example of most state water quality standards
and have a tiered set of goals and standards. This is
because, for practical reasons, not every aquatic system
can be managed to maintain outstanding ecological
qualities; for some heavily used water bodies, a simplified
functioning ecosystem is the best condition attainable.
Some state flow management programs, such as Maine’s,
have adopted condition goal classes from existing water
quality programs16. Others have defined new condition
goals that apply explicitly to water quantity, based on
existing conditions and stakeholder input. For example,
Connecticut’s 2011 streamflow protection regulations
prescribe a condition goal class system for each of the
state’s river reaches. The implementing agency will map
the state’s water bodies by goal class, facilitate a formal
public comment process, and revise the map accordingly.
Rhode Island created a watershed goal classification based
on natural characteristics (e.g., presence of cold water
fish), land use condition (e.g., impervious surface), land use
status (e.g., development zones, conservation land), water
withdrawals, and water quality. The five resulting ecological
condition goals range from “natural streams” to “significant
human influence.”
The process for defining condition goals, or “acceptable
ecological conditions” (Figure 1), may depend on the form
Figure 3.1. Using a simple flow-ecology curve to set an
environmental flow standard. For an ecological goal of maintaining
90% of the original fish species, the environmental flow standard is
that no more than 45% of the mid-summer streamflow may be
withdrawn (or, conversely, 55% of the streamflow must remain in the
stream). If withdrawals exceed that amount, then this model predicts
that more than 10% of the original fish species will be extirpated
from a stream of this river type.
16 http://www.maine.gov/dep/water/swup/index.htm
of the flow-ecology relationship (Figure 1.2). When a
threshold response exists, the benchmarking approach of
Arthington et al. (2006) can be used to establish an
ecologically and societally acceptable level of risk. When
the response is linear or curvilinear, “a consensus
stakeholder process may be needed to determine
acceptable risk. One possible process for setting such risk
levels is to use expert panels to identify ‘thresholds of
potential concern’ (Biggs and Rogers 2003; Acreman et al.
2008), which establish where along the flow alteration
gradient there is agreement among stakeholders (including
scientists and managers) that further hydrologic change
carries with it unacceptably high ecological risk” (Poff et al.
2010).
In either conception, the ecological risk levels linked to
condition classes can be informed by ecologists, who
describe the specific ecosystem outcomes associated with
each ecological condition class for each river type (if a
natural river system classification is used). These ecological
outcomes may describe key attributes of river ecosystem
health, such as physical habitat, water quality, flow
connectivity, biological composition and ecosystem
services. If a natural river system classification is used,
it is unlikely to overlap with the condition goal classes.
For example, Figure 3.2 depicts four different condition
goals for a single river type.
Figures 3.2 and 3.3 illustrate conceptually the translation
from river condition classes to environmental flow criteria
for one river type. In this example, a technical committee of
62
scientists and water managers described the ecological
outcomes associated with each river condition class (A-D)
in terms of physical habitat, water quality, flow connectivity,
biological composition, and ecosystem services (Figure 3.2,
left). Some states, such as Connecticut, combine these
ecological outcomes with human use goals into formal
narrative standards.
From these descriptions, hydrologic and ecological
indicators that are important to each river type can be
defined, ideally ensuring that the indicators resonate with
decision makers. As described above, a technical
committee can develop flow-ecology functions that relate
these indicators to each other, again by river type. Next,
they assign each river condition class to a range of
ecological indicator values along the y-axis (Figure 3.2,
center). They obtain the environmental flow ranges for
each river condition class from the x-axis of the flow
alteration-ecological response functions (Figure 3.2, right).
In this example, environmental flows are expressed as a
percent alteration from baseline condition, and these
percentages happen to be the same for high and low flows.
Figure 3.3 shows the environmental flow ranges in
hydrograph form for a particular river type and condition
class. The blue line represents the baseline hydrograph.
For the example shown, the environmental flow matrix
indicates that both x and y = 50% for condition class C.
This clearly is not the only way to link flow-ecology
relationships to ecological condition goals. Depending on
political circumstances, quantitative flow-ecology analyses
may not even be necessary to define environmental flow
criteria. In the Susquehanna River basin, experts based
quantitative flow recommendations primarily on conceptual
models. In the Connecticut River basin, initial flow
recommendations are based on a sustainability boundary
approach (Richter 2009, Richter et al. 2011). In both cases,
however, the flow metrics themselves (as opposed to their
values) were rigorously identified as those that best
represent flow variability and ecosystem dependence for
their respective river types. This assists communication of
results to water managers and water users, and facilitates
adaptive management.
3.4.2 Implementation: Putting Flow Standards
into Practice
The ability to estimate environmental flow needs for every
water body in a large region unlocks a broad range of
opportunities for implementation. Our case studies provide
examples of integrating ecosystem health into water
withdrawal permitting, multi-reservoir re-operation, and
water supply planning at the policy level. Elsewhere, the
ELOHA framework is guiding the estimation of
environmental flow needs for integrated water resource
management across large river basins. Although not all of
our case studies have reached the implementation stage,
some generalizations emerge regarding how to get there.
The first two sub-sections below highlight the importance
of having a structured, social process for defining
implementation policy. The third highlights the value and
use of decision support systems.
Setting Water Withdrawal Standards
Michigan used flow-ecology relationships and a well-defined
social process to guide two major policy decisions regarding
water withdrawals. First, the state legislature defined the
threshold for “adverse resource impact,” culminating a
science-driven stakeholder process. Second, condition
classes were reframed in terms of ecological risk levels,
and water withdrawal permitting policies were designed to
address each risk level. These policies were then
Figure 3.2. Process for translating condition classes (left) into environmental flow criteria expressed as degree of allowable flow alteration from
baseline (right) for two flow components (high and low flows) for a hypothetical river type.
63
incorporated into an online screening tool for prospective
water users to determine which policy would apply to their
proposed withdrawal. The Michigan Water Withdrawal
Assessment process, as it is called, has won three national
awards for streamlining government programs.
Beyond strong science, one of the key ingredients in
Michigan’s success was its clear, separate “social process”
centered around a set of guiding principles that focused a
diverse group of public, private, and non-profit stakeholders
on common water management interests. The final water
withdrawal standards were informed by a scientific
assessment of ecological risk, but defined through a
distinct political and social decision-making step that
weighed social needs against ecological protection goals.
The result is a set of standards that focuses water
management agency effort on those rivers that are most
vulnerable to impact and/or most valued by society, while
steering future water development toward other systems.
Setting Dam Management Standards
Dams are designed, built, and operated to achieve specific
objectives, including water supply, hydropower, recreation,
and flood control. Only very recently have dam design and
operation begun to consider downstream ecological health.
Therefore, it is critical to work with dam owners and operators
from the onset to find opportunities and understand
constraints on re-operating existing dams to provide
environmental benefits. This rings as true for regional dam
management as it does for individual dam re-operations.
In Connecticut, a state-sponsored technical committee
modeled numerous simulations to test and evolve a set of
general dam operating rules that would protect natural
flow variability (and presumably minimize ecological risk) at
minimal cost to reservoir safe yield. Yet, despite these
extensive analyses, it was critical to move from a technical
phase to a socio-political phase in which the form and
substance of the final reservoir release rules could be
negotiated. This stakeholder negotiation allowed for
trade-offs between ecological risk and socioeconomic
costs. The resulting release rules will improve
environmental flow releases to 156 river reaches.
Building a Decision Support System
A decision-support system (DSS) is an extremely useful
tool for implementing environmental flow management and
planning at the regional scale. According to Georgakakos
(2004), DSSs are technical tools intended to provide valid
and sufficient information to decision makers, comprised of
five main components: data acquisition system, user-data-
model interface, database, data analysis tools, and a set of
Figure 3.3. Sustainability boundaries (Richter 2009) shown in red, around natural hydrograph (blue) resulting from environmental flow criteria
depicted in Figure 3.2 (right), where both x and y = 50% for condition class C.
64
interlinked models. A useful DSS makes the decision
variables and results accessible to water managers and
stakeholders, and hides the complex models, equations,
and databases behind them.
DSSs for managing water withdrawals can readily
incorporate environmental flow criteria. For any site, the
Massachusetts SYE and Michigan WWAT can calculate the
streamflow depletion that would result from a proposed
new withdrawal, combined with the cumulative impacts of
all upstream withdrawals and return flows, and compare it
to environmental flow targets to determine the availability
of water for additional withdrawals. Water managers use
these tools to ensure that their water allocation decisions
protect ecological values.
DSSs also can be designed to integrate environmental
flows into regional water management. In the Connecticut
River basin, federal and state agencies have come together
to integrate operations of 70 large dams. One of the most
significant challenges was to constructively involve
individual dam owners. Through a series of workshops and
one-on-one meetings, the dam owners themselves helped
build a DSS to optimize basin-scale efficiency and provide
environmental flows. With the owner of 14 of the largest
dams playing a central role in the project, and at least one
major dam owner committing to use the tool for federal
relicensing, the likelihood of implementing a solution
informed by the optimized scenario is high.
Colorado is building DSSs to integrate environmental flows
into basin water planning. In the United States, water
resource planning typically accounts for water supply and
demand, hydropower, flood control, and perhaps other
economically-driven factors. Only recently have
environmental flows begun to be considered. Defining
environmental flow needs and integrating them with other
water demands at the large basin scale creates opportunities
for efficiency; for example, the same water that sustains
environmental flows upstream can be used for irrigation
downstream. In Colorado, scientists created a pilot DSS that
calculates cumulative streamflow depletion and associates
it with ecological risk levels for any location. Color-coded
basin maps indicate the degree of ecological risk to which
each river segment would be subject under various
scenarios, thus helping basin stakeholders understand
tradeoffs between water management options. In this case,
the majority of river reaches are already under some
degree of stress due to flow alteration, so environmental
objectives are geared toward flow restoration. Under
Colorado water law, re-allocation of water to the
environment is possible, but each re-allocation requires
extensive research, relationship-building, and often a lot of
money. The DSS helps target flow restoration to river
reaches that would most benefit the basin overall.
In the interstate Middle Potomac River basin, water
managers are analyzing how the combined impacts of land
use change, water withdrawal, and impoundments affect
low flows and stream health in small streams and rivers. A
basin-scale DSS is being considered to inform each state’s
future land and water management and to benefit the
basin overall. The DSS would be built from the existing
process-model-based hydrologic foundation, so it could
factor in future water use, land use, and climate change
projections.
The next big challenge for all projects is to design DSSs
that integrate information to support streamflow
management decisions with other management actions.
For example, decision tools that simultaneously assess both
water quantity and water quality management are feasible.
Likewise, if a DSS for environmental flows is built from a
process-model-based hydrologic foundation, then options
can be added to assess land use and drainage impacts.
However, if development has drastically changed spatial
drainage patterns and stream channel morphology, then it
may be difficult to model impacts on ecosystem health.
3.4.3 General Observations
Implementation of science-based environmental flow policy
and planning is still very much in its infancy. Nonetheless,
there are lessons to glean from progress to date:
• Defining a clear path to policy implementation from the
onset ensures that the ensuing science answers the right
management questions.
• An ecological risk-based framework associates
ecological goals with allowable flow alteration, and
accounts for scientific uncertainty by associating
appropriate policy actions with different levels of
ecological risk.
• A distinct and structured social process for identifying,
understanding, and negotiating tradeoffs is critical. To
launch this process, a set of “guiding principles”, or
agreed-upon common objectives, focuses diverse
stakeholders on common interests, and increases the
likelihood of achieving implementable outcomes.
• Decision support systems greatly facilitate the integration
of environmental flows into state water allocation
programs, integrated dam management, and regional
water resource planning.
65
Environmental flow protection is gaining broad acceptance
in the United States, and the science is sufficiently devel-
oped to support regional planning and policy applications,
including water withdrawal permitting, multi-reservoir
operation, and regional water resource planning. Remark-
able advances are being made in the use of structured
scientific literature review, hydrologic modeling, river
system classification, flow-ecology analysis, and condition
goal designation to inform water management.
Hydrologic foundation development is a common goal of
most, if not all, environmental flow management efforts.
While a database of daily flow data under naturalized and
current conditions remains an ideal for most applications,
the lack of these data need not hinder the determination of
environmental flow needs.
• If only current-condition hydrologic data are available,
then flow-ecology relationships can still be discerned.
However, without distinguishing altered from natural
ranges of hydrologic and ecological conditions, flow
restoration targets are more difficult to quantify.
• If only flow statistics are available—that is, if simple
regression is used to generate the hydrologic
foundation—then actionable flow-ecology relationships
can still be determined if a limited set of statistics is
deemed sufficiently protective of expected flow
alteration. For example, Michigan decided to limit
withdrawals all year to the amount needed to maintain
fish communities during their most stressful (lowest
flow) month. Given Michigan’s hydrology, restricting
water development in this way will prevent adverse
impacts of withdrawals on river ecosystems all year.
• If only monthly (not daily) flow series are available,
meaningful measures still can be taken to protect the
magnitude of seasonal flows. However, monthly flow
data do not capture freshets and other short-duration
events that are important for maintaining certain
ecological functions. Nevertheless, monthly flow data
can be more than adequate in water bodies with very
stable hydrologic regimes or for managing groundwater
pumping, which generally does not impact short-
duration flow events.
• No matter how estimates of unregulated hydrologic
conditions are derived, a major constraint will almost
always be accurate, spatially explicit water use data.
Advocacy for state water use reporting laws along with
federal and state investments in water use estimation
and decision support systems is a high priority,
particularly in states challenged by significant irrigation
use or limited water management programs.
Regardless of the data or statistics used, they should be
estimated for every stream segment or sub-basin where
environmental flows will be managed. Without this
information, limits on flow alteration are very difficult to
permit, measure, or enforce without expensive site-specific
data collection.
River type classification, as envisioned in Poff et al.
(2010), is not being used in the majority of cases we
reviewed. Where it has been used, it tends to be iterative
and based on pre-existing classifications. By treating natural
river type classification and flow-ecology analyses as
iterative processes, each one strengthens the explanatory
power of the other. Therefore, it makes sense to begin
analyzing flow-ecology relationships before investing
deeply in any detailed river type classification. Using an
existing classification system helps build trust among
scientists who are familiar with it. Moreover, incorporating
watershed and macrohabitat variables from the onset is
useful for assigning river types to ungaged sites.
Flow-ecology relations, whether in the form of
hypotheses or the results of quantitative analyses, are also
tremendously useful tools. They may be developed after or
in tandem with the hydrologic foundation.
• Flow-ecology hypotheses guide quantitative analyses,
focus expert workshops, and help engage stakeholders.
These qualitative relationships typically are based on a
combination of literature review and expert input. They
can drive the selection of hydrologic and ecological
metrics for subsequent analysis and flow management,
and can also ensure that stakeholders have a common
conception of the likely impacts of flow alteration.
• If large biological databases and a hydrologic foundation
are available, then statistical analyses can be framed to
test flow-ecology hypotheses. Strictly exploratory data
analysis may result in robust statistical relationships, but
if the metrics used do not resonate with biologists and
water managers, then the results may be ineffective in
supporting environmental flow policy.
• Even where large biological databases exist, rarely will
the results of flow-ecology analyses represent all
flow-dependent taxonomic groups and ecological
processes. A holistic approach requires a combination of
quantitative and qualitative relationships to represent the
4.0 Conclusions
66
ecosystems and natural flow regimes across a state or
basin. In the Susquehanna River basin, scientists based
flow recommendations on the historic natural variability
of the flow components that flow-ecology hypotheses
suggest are important. This approach can supplement
biological data analyses to generate flow
recommendations that protect the entire ecosystem.
Condition goal classification is proving extremely
valuable for driving stakeholder negotiations that lead to
policy implementation. Within any large region, people
value different rivers for different purposes. They may strive
to keep certain rivers nearly pristine, while accepting that
other, more developed rivers simply maintain a basic level
of ecological function. If stakeholders can agree to apply
different flow standards to rivers with different tiers (i.e.,
classes) of ecological condition goals, then they will likely
protect the most valuable ecosystems, while encouraging
development of less sensitive ones. This approach has
worked fairly well with water quality standards in the
United States, and is emerging for water management.
Condition goals may be linked to ecological risk levels
resulting from flow alteration.
Implementation is still very much in its infancy, as is the
ELOHA framework itself. Although ELOHA provides a useful
framework to guide and describe the scientific process
associated with environmental flow criteria development,
there is no one-size-fits-all approach to either the scientific
or the policy process. Regardless of the rigor of the
scientific analyses, expert judgment calls are required to
guide the process and to interpret results. Facilitated expert
workshops and advisory panels improve the outcomes and
their credibility to the public. The level of sophistication
needed depends ultimately on what policy makers want and
stakeholders will accept. No matter the situation, before
developing scientific products, it is essential to consider the
social and policy context in which environmental flow
management would take place. Early outreach to
stakeholders, the definition of shared guiding principles, the
formation of advisory committees, attention to political and
economic drivers, and process transparency are all likely to
pay dividends in environmental flows implementation.
Although much of this report focuses on scientific tools and
approaches, social and policy processes, as evidenced
by the case studies, are becoming more sophisticated in
their incorporation of scientific information and definition
of ecological goals linked to water management. Poff
et al.’s (2010) ELOHA framework lacks clear guidance on
development and implementation of these processes.
Given the importance of social and political drivers to
environmental flow implementation, the most important
advance needed to guide future environmental flow policy
and planning is to build and test a robust social framework
that complements ELOHA’s scientific framework.
67
Acreman, M. C. 2001. Ethical aspects of water and
ecosystems. Water Policy 3:257-265. http://cat.inist.fr/?aMo
dele=afficheN&cpsidt=1107221.
Acreman, M. C., Dunbar, M. J., Hannaford, J., Wood, P. J.,
Holmes, N. J., Cowx, I., Noble, R., Mountford, J.O., King, J.,
Black, A., Extence, C., Crookall, D. & Aldrick, J. 2008.
Developing environmental standards for abstractions from
UK rivers to implement the Water Framework Directive.
Hydrological Sciences Journal, 53, 6, 1105-1120.
Anderson, M. J. 2008. Animal-sediment relationships
re-visited: Characterizing species’ distributions along an
environmental gradient using canonical analysis and
quantile regression splines. Journal of Experimental Marine
Biology and Ecology. 366:16-27.
Apse, C., DePhilip, M., Zimmerman, J., and Smith, M. P.
2008. Developing instream flow criteria to support
ecologically sustainable water resource planning and
management. Final report to the Pennsylvania Instream
Flow Technical Advisory Committee, 196 p. http://www.
portal.state.pa.us/portal/server.pt/document/440033/
pa_instream_flow_report-_tnc_growing_greener-_final.pdf.
Archfield, S. A., Vogel, R. M., Steeves, P. A., Brandt, S. L.,
Weiskel, P. K., and Garabedian, S. P. 2010. The
Massachusetts Sustainable-Yield Estimator: A decision-
support tool to assess water availability at ungaged stream
locations in Massachusetts. U.S. Geological Survey
Scientific Investigations Report 2009-5227, 41 p. plus
CD-ROM http://pubs.usgs.gov/sir/2009/5227/.
Armstrong, D. S., Richards, T. A., and Levin, S. B., 2011.
Factors influencing riverine fish assemblages in
Massachusetts. U.S. Geological Survey Scientific
Investigations Report 2011–5193, 59 p. http://pubs.usgs.
gov/sir/2011/5193/.
Armstrong, D. S., Richards, T. A., and Brandt, S. L. 2010.
Preliminary assessment of factors influencing riverine fish
communities in Massachusetts. U.S. Geological Survey
Open-File Report 2010-1139, 43 p. http://pubs.usgs.gov/
of/2010/1139/.
Armstrong, D. S., Parker, G. W., and Richards, T. A. 2004.
Evaluation of streamflow requirements for habitat
protection by comparison to streamflow characteristics at
index streamflow-gaging stations in southern New
England. Pages 108. U.S. Geological Survey Water-
Resources Investigations Report 03-4332 http://water.usgs.
gov/pubs/wri/wri034332/.
Arthington, A. H., Bunn, S. E., Poff, N. L., and Naiman, R. J.
2006. The challenge of providing environmental flow rules
to sustain river ecosystems. Ecological Applications
16(4):1311-1318. doi: 10.1890/1051-0761(2006)016[1311:TC
OPEF]2.0.CO;2, http://intranet.iucn.org/webfiles/doc/
Environmental_Flows/Challenge_of_providing_EFlows.pdf.
Barlow, P. M. 2000. Documentation of computer program
STRMDEPL — a program to calculate streamflow depletion
by wells using analytical solutions in P. J. Zarriello and K. G.
Ries, III, eds. A precipitation runoff model for analysis of
the effects of water withdrawals on streamflow, Ipswich
River Basin, Massachusetts: U.S. Geological Survey Water
Resources Investigations Report 00-4029, p. 77-89.
Biggs, H. C., and Rogers, K. H. 2003. An adaptive system to
link science, monitoring, and management in practice. In:
The Kruger Experience. Ecology and Management of
Savanna Heterogeneity (eds. J. T. duToit, K. H. Rogers, and
H. C. Biggs). Island Press, Washington, p. 59-80.
Brenden, T. O., Wang, L., and Seelbach, P. W. 2008. A river
valley segment classification of Michigan streams based on
fish and physical attributes. Transactions of the American
Fisheries Society 137:1621-1636. http://www.tandfonline.
com/doi/abs/10.1577/T07-166.1.
Buchanan, C., Foreman, K., Johnson, J., and Griggs, A. 2011.
Development of a basin-wide benthic index of biotic
integrity for non-tidal streams and wadable rivers in the
Chesapeake Bay Watershed. Interstate Commission on the
Potomac River, ICPRB Report 11-1
http://www.potomacriver.org.
Bunn, S. E., and Arthington, A. H. 2002. Basic principles and
ecological consequences of altered flow regimes for aquatic
biodiversity. Environmental Management 30:492-507.
Cade, B. S., and Noon, B. R. 2003. A gentle introduction to
quantile regression for ecologists. Frontiers in Ecology and
the Environment 1(8):412-420. doi:
10.1890/1540-9295(2003)001[0412:AGITQR]2.0.CO;2.
Cade, B. S., and Richards, J. D. 2005. User manual for
Blossom statistical software. U.S. Geological Survey
Open-File Report 2005-1353, 124 p. http://www.fort.usgs.
gov/products/publications/pub_abstract.asp?PubID=21536.
5.0 References
68
Camp Dressser & McKee Inc., Bledsoe, B. D., Miller, W. J.,
Poff, N. L., Sanderson, J. S., and Wilding, T. K. 2009.
Watershed Flow Evaluation Tool pilot study for Roaring Fork
and Fountain Creek watersheds and site-specific
quantification pilot study for Roaring Fork watershed
(draft). Colorado Water Conservation Board, Denver, CO
http://www.conservationgateway.org/file/watershed-flow-
evaluation-tool-pilot-roaring-fork-and-fountain-creek.
Carlisle, D. M., Wolock, D. W., and Meador, M. R. 2010.
Alteration of streamflow magnitudes and potential
ecological consequences:a multiregional assessment.
Frontiers in Ecology and the Environment
doi:10.1890/100053.
CDWR and CWCB. 2009. State of Colorado’s water supply
model (StateMod) Version 12. Colorado Division of Water
Resources and Colorado Water Conservation Board:
Denver, Colorado ftp://dwrftp.state.co.us/cdss/swm/sw/
StateModManual_20090216.pdf.
CEC. 1997. Ecological regions of North America — toward a
common perspective. Commission for Environmental
Cooperation, Montreal.
Connecticut Department of Environmental Protection (DEP)
2009. Stream flow: the next two decades balancing human
use and ecological health. Hartford, CT. 18 p. http://www.
ct.gov/dep/lib/dep/water/watershed_management/
flowstandards/streamflow_next_2_decades.pdf.
Cummins, J., Buchanan, C., Haywood, C., Moltz, H., Griggs,
A., Jones, R. C., Kraus, R., Hitt, N., and Bumgardner, R. V.
2010. Potomac basin large river environmental flow needs.
Interstate Commission on the Potomac River Basin
ICPRB Report 10-03 http://www.potomacriver.org.
Davies, S. P., and Jackson, S. K. 2006. The biological
condition gradient: a descriptive model for interpreting
change in aquatic ecosystems. Ecological Applications
16(4):1251-1266.
DePhilip, M., and Moberg, T. 2010. Ecosystem flow
recommendations for the Susquehanna River basin. Report
to the Susquehanna River Basin Commission and U.S. Army
Corps of Engineers. The Nature Conservancy. 96 p +
appendices http://www.nature.org/media/pa/tnc-final-
susquehanna-river-ecosystem-flows-study-report.pdf.
Freeman, M.C., and Marcinek, P.A. 2006. Fish assemblage
responses to water withdrawals and water supply
reservoirs in Piedmont streams. Environmental Management
38(3):435-450.
Gannon, C. 2007. Hydrologic alteration of the Connecticut
River watershed: authorized water withdrawals/diversions
and discharges in Massachusetts and Connecticut.
The Nature Conservancy, Connecticut River Program,
Northampton, MA, 36 p. http://conserveonline.org/
workspaces/ctriver/documents/papers.
Georgakakos, A., 2004. Decision support systems for
integrated water resources management with an
application to the Nile basin. IFAC Workshop on Modelling
and Control for Participatory Planning and Managing Water
System, Venice, Italy, 2004.
Hamilton, D. A., and Seelbach, P. W. 2011. Michigan’s water
withdrawal assessment process and internet screening
tool. Michigan Department of Natural Resources, Fisheries
Special Report 55, Lansing.
Hamilton, D. A., Sorrell, R. C., and Holtschlag, D. J. 2008. A
regression model for computing index flows describing the
median flow for the summer month of lowest flow in
Michigan. U.S. Geological Survey Scientific Investigations
Report 2008-5096 http://pubs.usgs.gov/sir/2008/5096/.
Harbaugh, A. W. 2005. MODFLOW-2005, the U.S.
Geological Survey modular ground-water model -- the
ground-water flow process. U.S. Geological Survey
Techniques and Methods 6-A16 http://pubs.usgs.gov/
tm/2005/tm6A16/.
Henriksen, J. A., Heasley, J., Kennen, J. G., and Nieswand, S.
2006. Users’ Manual for the Hydroecological Integrity
Assessment Process Software (including the New Jersey
Assessment Tools). U.S. Geological Survey Open-File
Report 2006-1093, 71 p. http://www.fort.usgs.gov/products/
publications/21598/21598.pdf.
Hersh, E. S., and Maidment, D. R. 2007. An integrated
stream classification sysem for Texas. Center for Research
in Water Resouces Online Report 2007-02, University of
Texas, Austin http://www.crwr.utexas.edu/reports/2007/
rpt07-2.shtml.
Higgins, J. V., Bryer, M. T., Khoury, M. L., and Fitzhugh, T. W.
2005. A freshwater classification approach for biodiversity
conservation planning. Conservation Biology 19(2):1-14.
http://science.natureconservancy.ca/centralinterior/docs/
ERAtoolbox/7/Higgins%20et%20al.%202005.pdf.
Hoffman, J. L., and Rancan, H. L. L. 2007. The
hydroecological integrity assessment process in New
Jersey. New Jersey Geological Survey, Trenton, NJ
http://www.state.nj.us/dep/njgs/pricelst/tmemo/tm09-3.pdf.
69
Holtschlag, D. J. 2009. Application guide for AFINCH
(analysis of flows in networks of channels) described by
NHDPlus. U.S. Geological Survey Scientific Investigations
Report 2009-5188, 106 p. http://pubs.usgs.gov/
sir/2009/5188/.
Kanno, Y., and Vokoun, J. C. 2010. Evaluating effects of
water withdrawals and impoundments on fish assemblages
in southern New England streams, USA. Fisheries and
Management Ecology 17:272-283. http://nre.uconn.edu/
pages/people/bios/documents/KannoVokoun.
WaterWithdrawals.FME.2010.pdf.
Kennard, M. J., Mackay, S. J., Pusey, B. J., Olden, J. D., and
Marsh, N. 2009. Quantifying uncertainty in estimation of
hydrologic metrics for ecological studies. River Research
and Applications DOI: 10.1002/rra.1249.
Kennard, M. J., Pusey, B. J., Olden, J. D., Mackay, S. J., Stein,
J. L., and Marsh, N. 2010. Classification of natural flow
regimes in Australia to support environmental flow
management. Freshwater Biology 55(1):171-193.
Kennen, J. G., Henriksen, J. A., Heasley, J., Cade, B. S., and
Terrell, J. W. 2009. Application of the Hydroecological
Integrity Assessment Process for Missouri Streams. U.S.
Geological Survey Open File Report 2009-1138, 57 p.
http://pubs.usgs.gov/of/2009/1138/.
Kennen, J. G., Henriksen, J. A., and Nieswand, S. P. 2007.
Development of the Hydroecological Integrity Assessment
Process for determining environmental flows for New
Jersey streams. U.S. Geological Survey Scientific
Investigations Report 2007-5206, 55 p. http://pubs.er.usgs.
gov/usgspubs/sir/sir20075206, http://www.fort.usgs.gov/
Products/Software/NJHAT/Default.asp.
Kennen, J. G., Kauffman, L. J., Ayers, M. A., and Wolock, D.
M. 2008. Use of an integrated flow model to estimate
ecologically relevant hydrologic characteristics at stream
biomonitoring sites. Ecological Modelling 211:57-76.
doi:10.1016/j.ecolmodel.2007.08.014.
Kennen, J. G., and Riskin, M. L. 2010. Evaluating effects of
potential changes in streamflow regime on fish and
aquatic-invertebrate assemblages in the New Jersey
Pinelands. U.S. Geological Survey Scientific Investigations
Report 2010-5079, 34 p. http://pubs.usgs.gov/
sir/2010/5079/.
Kennen, J. G., Riva-Murray, K., and Beaulieu, K. M. 2010.
Determining hydrologic factors that influence stream
macroinvertebrate assemblages in the northeastern US.
Journal of Ecohydrology, DOI 10.1002/eco.99.
Kenny, J. F., Barber, N. L., Hutson, S. S., Linsey, K. S.,
Lovelace, J. K., and Maupin, M. A. 2009. Estimated use of
water in the United States in 2005. U.S. Geological Survey
Circular 1344, 52 p.
King, J., Brown, C., and Sabet, H. 2003. A scenario-based
holistic approach to environmental flow assessments for
rivers. River Research and Applications 19(5-6):619-639.
Kitanidis, P. K. 1992. Geostatistics in D. R. Maidment, ed.
Handbook of Hydrology. McGraw-Hill, Inc., chapter 20.
Knight, R. R., Gregory, M. B., and Wales, A. K. 2008. Relating
streamflow characteristics to specialized insectivores in the
Tennessee River Valley: a regional approach. Ecohydrology
1:394-407, doi: 10.1002/eco.32.
Koltun, G. F., Kula, S. P., and Puskas, B. M. 2006. A
streamflow statistics (SteamStats) web application for Ohio.
U.S. Geological Survey Scientifc Investigations Report
2006-5312, 62 p.
Koltun, G. F., and Whitehead, M. T. 2002. Techniques for
estimating selected streamflow characteristics of rural,
unregulated streams in Ohio. U.S. Geological Survey
Scientific Investigations Report 02-4068, 50 p.
Konrad, C. P., Brasher, A. M. D., and May, J. T. 2008.
Assessing streamflow characteristics as limiting factors on
benthic invertebrate assemblages in streams across the
western United States. Freshwater Biology.
doi:10.1111/j.1365-2427.2008.02024.x.
Lang, V. 1999. Questions and answers on the New England
flow policy. Concord, NH, U.S. Fish and Wildlife Service, 9p.
http://southeastaquatics.net/uploads/category/NE%20
Streamflow%20Policy%20by%20USFWS.pdf.
Le Quesne, T., Kendy, E., and Weston, D. 2010. The
implementation challenge: Taking stock of government
policies to protect and restore environmental flows. WWF
and The Nature Conservancy http://conserveonline.org/
workspaces/eloha/documents/wwf-tnc-e-flow-policies-
report.
Lorenz, D. L., Sanocki, C. A., and Kocian, M. J. 2009.
Techniques for estimating the magnitude and frequency of
peak flows on small streams in Minnesota based on data
through water year 2005. U.S. Geological Survey Scientific
Investigations Report 2009-5250, 54 p.
Lyons, J., Zorn, T., Stewart, J., Seelbach, P., Wehrly, K., and
Wang, L. 2009. Defining and characterizing coolwater
streams and their fish assemblages in Michigan and
Wisconsin, USA. North American Journal of Fisheries
Management 29:1130-1151. DOI: 10.1577/M08-118.1.
70
Mathews, R., and Richter, B. 2007. Application of the
Indicators of Hydrologic Alteration software in
environmental flow-setting. Journal of the American Water
Resources Association (JAWRA) 43(6):1400-1413.
Michigan Department of Environmental Quality. 2009. Water
Withdrawal Assessment Tool http://www.miwwat.org/.
Michigan Groundwater Conservation Advisory Council.
2007. Report to the Michigan Legislature in response to
Public Act 34, 37 p. http://www.michigan.gov/documents/
deq/Groundwater_report_206809_7.pdf.
Milhous, R. T., Updike, M. A., and Schneider, D. N. 1989.
Physical Habitat Simulation System reference manual -
version II. U.S. Fish and Wildlife Service, Instream Flow
Information Paper No. 26, Biological Report 89(16) http://
www.fort.usgs.gov/products/publications/3912/3912.pdf.
Monk, W. A., Peters, D. L., Curry, A., and Baird, D. J. 2011.
Quantifying trends in indicator hydroecological variables
for regime-based groups of Canadian rivers. Hydrological
Processes 25(19):3086-3100. DOI: 10.1002/hyp.8137.
Norris, R.H., Webb, J.A., Nichols, S.J, Stewardson, M.J. and
Harrison, E.T. 2012. Analyzing cause and effect in
environmental assessments: using weighted evidence from
the literature. Freshwater Science. 31(1): 5–21.
Olden, J., and Naiman, R. J. 2010. Incorporating thermal
regimes in environmental flows assessments: modifying
dam operations to restore freshwater ecosystem integrity.
Freshwater Biology 55(1):86-107.
Olden, J. D., Kennard, M. J., and Pusey, B. J. 2011. A
framework for hydrologic classification with a review of
methodologies and applications in ecohydrology.
Ecohydrology. DOI: 10.1002/eco.251.
Olivero, A. P., and Anderson, M. G. 2008. Northeast aquatic
habitat classification system. The Nature Conservancy
Eastern Regional Office, Boston, Massachusetts, 88 p.
Poff, N. L., Richter, B. D., Arthington, A. H., Bunn, S. E.,
Naiman, R. J., Kendy, E., Acreman, M., Apse, C., Bledsoe, B.
P., Freeman, M. C., Henriksen, J., Jacobson, R. B., Kennen, J.
G., Merritt, D. M., O’Keeffe, J. H., Olden, J. D., Rogers, K.,
Tharme, R. E., and Warner, A. 2010. The ecological limits of
hydrologic alteration (ELOHA): a new framework for
developing regional environmental flow standards.
Freshwater Biology 55:147-170. http://rydberg.biology.
colostate.edu/~poff/PoffPublicationsPDF.htm.
Poff, N.L., and Zimmerman, J. 2009. Ecological responses to
altered flow regimes: a literature review to inform the
science and management of environmental flows.
Freshwater Biology doi:10.1111/j.1365-2427.2009.02272.x,
http://rydberg.biology.colostate.edu/~poff/
PoffPublicationsPDF.htm.
Poff, N. L., Allan, J. D., Bain, M. B., Karr, J. R., Prestegaard,
K. L., Richter, B. D., Sparks, R. E., and Stromberg, J. C. 1997.
The natural flow regime: a paradigm for river conservation
and restoration. BioScience 47:769-784.
Poff, N.L., Allan, J.D., Bain, M.B., Karr, J.R., Prestegaard, K.L.,
Richter, B.D., Sparks, R.E., and Stromberg, J.C. 1997. The
natural flow regime: a paradigm for river conservation and
restoration. Bioscience 47: 769-784.
Postel, S., and Richter, B. 2003. Rivers for Life: Managing
Water for People and Nature. Island Press, Washington, D.C.
Pusey, B., Kennard, M., Hutchinson, M., Sheldon, F., Stein, J.,
Olden, J., and McKay, S. 2009. Ecohydrological
regionalisation of Australia: a tool for management and
science. Land and Water Australia Technical Report
PN22591, 253 pp. http://lwa.gov.au/files/products/
innovation/pn22591/pn22591.pdf.
Reeves, H. W. 2008. STRMDEPL08 — An extended version
of STRMDEPL with additional analytical solutions to
calculate streamflow depletion by nearby pumping wells.
U.S. Geological Survey Open-File Report 2008-1166, 22 p.
http://pubs.usgs.gov/of/2008/1166/.
Reeves, H. W., Hamilton, D. A., Seelbach, P. W., and Asher,
A. J. 2009. Ground-water component of the Michigan Water
Withdrawal Screening Tool. U.S. Geological Survey
Scientific Investigations Report 2009-5003 http://pubs.usgs.
gov/sir/2009/5003/.
Reidy Liermann, C. A., Olden, J. D., Beechie, T. J., Kennard,
M. J., Skidmore, P. B., Konrad, C. P., and Imaki, H. 2011.
Hydrogeomorphic classification of Washington State rivers
to support emerging environmental flow management
strategies. River Research and Applications Published
online in Wiley Online Library, wileyonlinelibrary.com DOI:
10.1002/rra.1541.
Richardson, A.R. 2005. Modified Aquatic Base Flow
(RI-ABF) for Rhode Island. Rhode Island Department of
Environmental Management. Providence, RI.
71
Richardson, A. R. 2006. Pre-application guidance for those
seeking a groundwater withdrawal permit for >10,000 gpd.
Rhode Island Department of Environmental Management.
Providence, RI. http://www.dem.ri.gov/programs/benviron/
water/withdraw/pdf/guidance.pdf.
Richter, B. D. 2009. Re-thinking environmental flows:
from allocations and reserves to sustainability boundaries.
River Research and Applications 22(8):1052-63. DOI:
10.1002/rra.1320.
Richter, B. D., Davis, M., Apse, C., and Konrad, C. P. 2011.
A presumptive standard for environmental flow protection.
River Research and Applications. DOI: 10.1002/rra.1511.
Richter, B. D., Warner, A. T., Meyer, J. L., and Lutz, K. 2006.
A collaborative and adaptive process for developing
environmental flow recommendations. River Research and
Applications 22:297-318.
Roy, A. H., Freeman, M. C., Freeman, B. J., Wenger, S. J.,
Ensign, W. E., and Meyer, J. L. 2005. Investigating
hydrologic alteration as a mechanism of fish assemblage
shifts in urbanizing streams. Journal of the North American
Benthological Society 24:656-678. doi: 10.1899%2F04-022.1.
Ruswick, F., Allan, J., Hamilton, D., and Seelbach, P. 2010.
The Michigan Water Withdrawal Assessment process:
science and collaboratin in sustainaing renewable natural
resources. Renewable Resources Journal 26:13-18.
Sanborn, S. C., and Bledsoe, B. P. 2005. Predicting
streamflow regime metrics for ungaged streams in
Colorado, Washington, and Oregon. Journal of Hydrology
325:241-261.
Sanderson, J. S., Rowan, N., Wilding, T., Bledsoe, B. P.,
Miller, W. J., and Poff, N. L. 2011. Getting to scale with
environmental flow assessment: the watershed flow
evaluation tool. River Research and Applications Published
online in Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/rra.1542.
Seelbach, P. W., Wiley, M. H., Baker, M. E., and Wehrly, K. E.
2006. Landscape-based identification and classification of
ecological river segments: concepts, approach, and
application across Michigan’s Lower Peninsula. Pages
18-38 in R. Hughes, L. Wang, and P. W. Seelbach, eds.
Landscape influences on stream habitats and biological
communities, American Fisheries Society.
Snelder, T., Booker, D., and Lamouroux, N. 2011. A method
to assess and define environmental flow rules for large
jurisdictional regions. Journal of the American Water
Resources Association (JAWRA) 47(4):828-840. DOI:
10.1111 ⁄ j.1752-1688.2011.00556.x.
Stedinger, J. R., Vogel, R. M., and Foufoula-Georgiou, E.
1992. Frequency analysis of extreme events in D. R.
Maidment, ed. Handbook of Hydrology. McGraw-Hill, Inc.,
chapter 18.
Steinman, A. D., Nicholas, J. R., Seelbach, P. W., Allan, J. W.,
and Ruswick, F. 2011. Science as a fundamental framework
for shaping policy discussions regarding the use of
groundwater in the State of Michigan: a case study. Water
Policy 13:69-86. http://www.iwaponline.com/wp/up/
wp2010047.htm.
Tharme, R. E. 2003. A global perspective on environmental
flow assessment: emerging trends in the development and
application of environmental flow methodologies for rivers.
River Research and Applications 19:397-441. http://www.
eflownet.org/downloads/documents/Tharme2003.pdf.
The Nature Conservancy. 2009. Indicators of Hydrologic
Alteration, Version 7.1 User’s Manual, 76 p.
http://conserveonline.org/workspaces/iha.
US Fish and Wildlife Service 1981. New England Aquatic
Base Flow. Concord, NH.
Vogel, R. M., Sieber, J., Archfield, S. A., Smith, M. P., Apse,
C. D., and Huber-Lee, A. 2007. Relations among storage,
yield and instream flow. Water Resources Research 43
(W05403). doi:10.1029/2006WR005226; http://www.weap21.
org/downloads/RelationsAmongStorage.pdf.
Vokoun, J. C., and Kanno, Y. 2009. Evaluating effects of
water withdrawals and impoundments on fish assemblages
in Connecticut streams. University of Connecticut,
Department of Natural Resources and the Environment
College of Agriculture and Natural Resources.
Vokoun, J. C., and Kanno, Y. 2010. Evaluating effects of
water withdrawals and impoundments on fish assemblages
in southern New England streams. Fisheries Management
and Ecology 17:272-283. http://www.nrme.uconn.edu/
pages/people/bios/documents/KannoVokoun.
WaterWithdrawals.FME.2010.pdf.
Wehrly, K. E., Wiley, M. H., and Seelbach, P. W. 2003.
Classifying regional variation in the thermal regime based
on stream fish community patterns. Transactions of the
American Fisheries Society 132:18-38.
Weiskel, P. K., Brandt, S. L., DeSimone, L. A., Ostiguy, L. J.,
and Archfield, S. A. 2010. Indicators of streamflow
alteration, habitat fragmentation, impervious cover, and
water quality for Massachusetts stream basins. U.S.
Geological Survey Scientific Investigations Report 2009-
5272, 70 p. plus CD–ROM. http://pubs.usgs.gov/
sir/2009/5272/.
72
Zimmerman, J. K. H. 2006a. Hydrologic effects of flood
control dams in the Ashuelot River, New Hampshire, and
the West River, Vermont. The Nature Conservancy,
Connecticut River Program, Northampton, MA, 30 p. http://
conserveonline.org/workspaces/ctriver/documents/papers.
Zimmerman, J. K. H. 2006b. Hydrologic effects of flood
control dams in the Ashuelot River, New Hampshire, and
the West River, Vermont. The Nature Conservancy,
Connecticut River Program, Northampton, MA, 30 p. http://
conserveonline.org/workspaces/ctriver/documents/papers.
Zimmerman, J. K. H., and Lester, A. 2006. Spatial
distribution of hydrologic alteration and fragmentation
among tributaries of the Connecticut River. The Nature
Conservancy, Connecticut River Program, Northampton,
MA, 19 p. http://conserveonline.org/workspaces/ctriver/
documents/papers.
Zimmerman, J. K. H., Lester, J. A., Lutz, K., Gannon, C., and
Nedeau, E. J. 2008. Restoring ecosystem flows in the
Connecticut River watershed. The Nature Conservancy,
Connecticut River Program, Northampton, MA.
Zimmerman, J. K. H., Nislow, K. H., Letcher, B. H., and
Magilligan, F. J. 2009. Determining the effects of dams on
subdaily variation in river flows at a whole-basin scale.
River Research and Applications. DOI: 10.1002/rra.1324.
Zorn, T. G., Seelbach, P. W., Rutherford, E. S., Wills, T. C.,
Cheng, S.-T., and Wiley, M. J. 2008. A regional-scale habitat
suitability model to assess the effects of flow reduction on
fish assemblages in Michigan streams. Michigan
Department of Natural Resources, Fisheries Research
Report 2089, Ann Arbor, Michigan, 46 pp. http://www.
michigan.gov/documents/dnr/RR2089_268570_7.pdf.
73