River flows and water wars: emerging science for
environmental decision making
Author
Poff, NL, Allan, JD, Palmer, MA, Hart, DD, Richter, BD, Arthington, AH, Rogers, KH, Meyers,
JL, Stanford, JA
Published
2003
Journal Title
Frontiers in Ecology and the Environment
DOI
https://doi.org/10.1890/1540-9295(2003)001[0298:RFAWWE]2.0.CO;2
Copyright Statement
© 2003 Ecological Society of America. The attached file is reproduced here in accordance
with the copyright policy of the publisher. Please refer to the journal's website for access to the
definitive, published version.
Downloaded from
http://hdl.handle.net/10072/5989
Link to published version
http://www.frontiersinecology.org/
Griffith Research Online
https://research-repository.griffith.edu.au
298
www.frontiersinecology.org © The Ecological Society of America
C
onflicts based on the perceived needs of ecosystems
versus humans for fresh water are increasingly seen in
the news. In the US, a fiery debate has erupted in the
Klamath basin of Oregon and California, where farmers
have protested the loss of irrigation water to protect
endangered fish, and where over 30 000 chinook salmon
and other fish recently died, perhaps due to insufficient
water quantity and/or quality (Levy 2003; Figure 1). The
states of Georgia, Alabama, and Florida have been
engaged for over a decade in contentious negotiations
over water allocation in the Apalachicola-Chatta-
hoochee-Flint River basin, with demands coming from
the growth of metropolitan Atlanta, agricultural irriga-
tion, and the Apalachicola Bay oyster fishery (Richter et
al. 2003b).
In New Zealand, a debate rages over how to allocate
enough water to maintain the ecological needs of the
Rangitata River while addressing the water demands of the
dairy industry (Robson 2002). Similar conflicts between
water requirements for irrigation and environment along
the Lower Balonne River system in Australia have
prompted an independent review of the science underlying
river condition assessments and environmental flow rec-
ommendations (Cullen et al. 2003). As human population
growth and climate change impose new constraints on the
spatial and temporal distribution of water (Postel et al.
1996; Vörösmarty et al. 2000), we may expect more such
water conflicts, and even environmental water “wars”,
such as in the Klamath basin. Increased human demand
will compete with the real needs of freshwater ecosystems
(Baron et al. 2002; Poff et al. 2002; Arthington and Pusey
in press).
The challenge now facing river scientists is to define
ecosystem needs clearly enough to guide policy formula-
tion and management actions that strive to balance com-
peting demands and visions. Is the current science up to
this challenge? Recent events in the Klamath basin sug-
gest a number of problems. An interim National Research
Council (NRC) report evaluating the science underlying
the federal management of water to protect endangered
fishes in the Klamath highlighted the uncertainty of avail-
able science to support decision making (NRC 2002c).
The Klamath conflict, and many others around the world,
REVIEWS REVIEWS REVIEWS
River flows and water wars: emerging
science for environmental decision making
N LeRoy Poff
1
, J David Allan
2
, Margaret A Palmer
3
, David D Hart
4
, Brian D Richter
5
, Angela H Arthington
6
,
Kevin H Rogers
7
, Judy L Meyer
8
, and Jack A Stanford
9
Real and apparent conflicts between ecosystem and human needs for fresh water are contributing to the
emergence of an alternative model for conducting river science around the world. The core of this new para-
digm emphasizes the need to forge new partnerships between scientists and other stakeholders where shared
ecological goals and river visions are developed, and the need for new experimental approaches to advance
scientific understanding at the scales relevant to whole-river management. We identify four key elements
required to make this model succeed: existing and planned water projects represent opportunities to conduct
ecosystem-scale experiments through controlled river flow manipulations; more cooperative interactions
among scientists, managers, and other stakeholders are critical; experimental results must be synthesized
across studies to allow broader generalization; and new, innovative funding partnerships are needed to
engage scientists and to broadly involve the government, the private sector, and NGOs.
Front Ecol Environ 2003; 1(6): 298–306
1
Dept of Biology, Colorado State Univ, Fort Collins, CO 80523
(poff@lamar.colostate.edu);
2
School of Natural Resources and
Environment, Univ of Michigan, Ann Arbor, MI 48109;
3
Dept of
Entomology, Univ of Maryland, College Park, MD 20742;
4
Patrick
Center for Environmental Research, Academy of Natural Sciences,
Philadelphia, PA 19103;
5
The Nature Conservancy, Charlottesville,
VA 22901;
6
Cooperative Research Centre for Freshwater Ecology,
Griffith Univ, Nathan, QLD 4111, Australia;
7
Centre for Water in the
Environment, Univ of the Witwatersrand, Johannesburg, South Africa;
8
Inst of Ecology, Univ of Georgia, Athens, GA 30602;
9
Flathead Lake
Biological Station, Univ of Montana, Polson, MT 59860
In a nutshell:
• The ecological sustainability of river ecosystems is threatened
by the extensive hydrologic alterations carried out by humans
• Despite the strong conceptual basis for sustainable river man-
agement, scientists are challenged to define ecosystem needs
clearly enough to guide policy formulation and management
actions that balance competing demands and goals
• An alternative model of collaboration between scientists,
managers, and other stakeholders to perform large-scale river
experiments is emerging around the world
• Innovative funding partnerships between government agen-
cies, not-for-profit foundations, and the private sector are
required to advance the scientific basis of water management
NL Poff et al. River flows and water wars
299
© The Ecological Society of America www.frontiersinecology.org
provide dramatic evidence that
improvements are needed in the sci-
ence, the decision-making process, or
probably both.
Society is clearly willing to invest in
flow management for ecological
restoration. For example, the
Northwest Power Planning Council
and Bonneville Power Administration
have spent several billion dollars over
the past decade on salmon restoration
in the Columbia River system, with
considerable emphasis on flow and
habitat restoration (ISG 2000). The
Tennessee Valley Authority has spent
more than $44 million modifying dam
operations to increase flows and
improve water quality (Bednarek 2002;
Figure 2). The 7-day experimental
flood of the Grand Canyon in 1996
cost $2.5 million in lost hydropower revenue (Patten and
Stevens 2001). Australia’s plans to provide an environ-
mental flow equal to 28% of the mean annual flow for the
Snowy River, to restore ecosystem integrity and recre-
ational opportunities, will cost at least US$216 million
(Pigram 2000).
Already, scientists in North America, South Africa,
Australia, New Zealand, and Europe are actively advising
the public and river managers on the necessary quantity
and timing of river flow needed to maintain desired eco-
logical characteristics. Such advice is being provided in
diverse political contexts, ranging from the cooperative to
the controversial. In the US, scientists have helped guide
efforts to recover endangered species in rivers like the
Klamath, the Colorado, and the Missouri (NRC 2002b),
to restore degraded flagship ecosystems such as the Florida
Everglades (Holling et al. 1994), and to define the flows
needed to protect ecological integrity on federally-owned
lands (NPS 1996). In addition, scientists are now being
asked to project the ecological responses to dam removal
in the US (Hart et al. 2002), or in Australia’s Murray-
Darling River Basin, to advise on how to “press dams into
environmental service”, to provide so-called “environ-
mental flows” (D Blackmore, pers comm). However, the
calculus of environmental, economic, and social costs and
benefits in many of these restoration efforts is complex,
and scientific uncertainty further complicates matters
(Pigram 2000; Bunn and Arthington 2002; Stanley and
Doyle 2003).
Although society is often willing to invest in the
restoration and protection of rivers, there are also high
expectations for measurable ecological returns. Over the
past decade, scientists have developed a solid conceptual
understanding of the importance of natural flows for river
ecosystems (Naiman et al. 1995; Poff et al. 1997; Puckridge
et al. 1998), and this can provide a strong foundation for
large-scale water manipulations and environmental flow
restoration strategies (Stanford et al. 1996; Arthington
Figure 1. Dead salmon in the Klamath River, September 2002.
Courtesy of Tim McKay, Northcoast Environmental Center
Figure 2. Improvements in Tennessee Valley Authority dam operations include (left) surface water pumps in the Douglas Reservoir
(French Broad River, NC), designed to push oxygenated water from the surface to deepwater turbines, and (right) infuser weirs
downstream of Chatuge Dam (Hiawassee River, NC), designed to provide both minimum flow and dissolved oxygen.
Courtesy of AT Bednarek, Earth Institute at Columbia University
River flows and water wars NL Poff et al.
and Pusey in press). Indeed, the “Natural Flows
Paradigm” (Poff et al. 1997) is already becoming the blue-
print for river corridor restoration and flow management
in several countries, particularly Australia and South
Africa. However, water managers and other stakeholders
are now demanding more than just a strong conceptual
understanding to guide the management of individual
rivers such as the Klamath. They are asking, how much
flow restoration is necessary to ensure ecological sustain-
ability? How natural do flow quantity, seasonal timing,
and water quality need to be to achieve the desired eco-
logical outcomes?
New approaches to research, management, and policy
development are needed to answer these critical ques-
tions. Elements of such approaches are emerging from var-
ious countries, with considerable progress being reported
at international conferences in the US (Managing River
Flows for Biodiversity, Fort Collins, Colorado, July 2001),
South Africa (Environmental Flows for River Systems,
Cape Town, March 2002), Cambodia (Second Large
Rivers Symposium – LARS 2, Phnom Penh, February
2003), and Australia (The Nature, Causes, and
Consequences of Variability in Large Rivers, Albury
NSW, July 2003). Our growing understanding of two
major shortcomings of past scientific investigations has
stimulated innovations in the practice of river science.
First, conventional research methods (eg small-scale
experimentation and large-scale comparative studies) are
300
www.frontiersinecology.org © The Ecological Society of America
often insufficient for gaining adequate eco-
logical understanding to support effective
decisions for river-specific restoration and
management. Second, to achieve desirable
ecological outcomes, scientists must also
see themselves as partners at the table with
resource managers and other stakeholders
in a collaborative process, so that scientific
understanding, management strategies, and
societal goals are effectively integrated.
Strengthening roles
Based on the growing recognition that
more effective approaches are needed, we
propose four steps to strengthen the roles
of science and society in managing rivers
(and other fresh waters) to meet human
and ecosystem needs. The conceptual
framework linking these four steps is pre-
sented in Figure 3.
Step 1: Implement more large-scale
river experiments on existing and
planned water management projects
We have achieved major theoretical
advances in our understanding of how
streams and rivers function. Nevertheless,
fundamental problems of uncertain knowledge and lim-
ited predictive capability continue to beset the science
underlying river ecosystem management. This uncer-
tainty arises both from irreducible ecosystem complexity,
and from the limited transferability of general ecological
understanding to site-specific situations. A learning-by-
doing approach therefore becomes a prerequisite for
the effective management of complex river ecosystems
(Rogers in press). Because extrapolating results from tra-
ditional small-scale experiments to the much larger scales
relevant to river management requires untested assump-
tions of transferability in scaling up (Walters and Korman
1999), well-designed, large-scale experiments and moni-
toring arguably offer the best approach to learning in the
long run (Walters 1997). This, of course, is the core prin-
ciple of adaptive management. Even so, a major limita-
tion in advancing scientific knowledge to guide ecologi-
cal flow management is the lack of opportunities to
conduct large-scale experiments, where whole-system
responses can be evaluated at scales that match manage-
ment actions (Kingsford 2000; Bunn and Arthington
2002). Currently, far too many opportunities are being
missed to learn from river flow manipulations at these
larger scales.
We argue that water development and river restoration
projects should be routinely established as scientifically
credible, ecosystem-scale experiments, where water man-
agement activities represent treatments to test specific
Figure 3. Conceptual flow diagram illustrating interactions and feedback loops
between science, stakeholders, and funders in the pursuit of improved science-
based policy and management of river ecosystems. This is accomplished by using
existing water management structures as opportunities to conduct large-scale
learning experiments.
NL Poff et al. River flows and water wars
ecological hypotheses. The results of
such experiments will greatly advance
our scientific capacity for ecological pre-
diction, and promote more efficient and
ecologically sustainable water manage-
ment. For example, extensive monitor-
ing in conjunction with the planned
1996 Grand Canyon flood evaluated
specific hypotheses and produced new
understanding about sediment availabil-
ity and habitat forming processes, infor-
mation that is now being used to guide
future flow manipulations (Rubin et al.
2002). A 5-year research program on the
Green River below Utah’s Flaming
Gorge Dam included experimental flow
releases to better establish the specific
requirements of endangered fish species,
leading to a set of flow and temperature
recommendations to guide management
of this large river (Muth et al. 2000).
Many river basins have multiple
impoundments or other types of water
developments that cumulatively alter
flow regimes (Pringle 2001). We should
therefore also seek opportunities to con-
duct experimentation at the scale of
entire river basins, taking into account
spatially distributed water control struc-
tures (Figure 4). For example, by coordinating critical com-
ponents of the natural flow regime (such as high flow pulses
at specific times) across multiple water control structures,
we could assess ecological benefits relative to existing con-
ditions in a powerful, integrated manner across many eco-
logical scales. Current efforts in the Yakima River in
Washington State (Stanford et al. 2002) provide an encour-
aging example, as do some governmental programs aimed at
coordinating restoration activities for large river basins. For
example, in Australia’s Murray-Darling River Basin, a fed-
eral initiative has established a community–government
partnership “to promote and coordinate effective planning
and management for the equitable, efficient, and sustain-
able use of the water, land, and other environmental
resources”, using adaptive management (MDBI 2003). The
California Bay–Delta Program represents a similar initiative
in the US (CALFED 2001).
Experiments are powerful agents for learning, but even
the best designed experiment cannot resolve all the scien-
tific uncertainty associated with an ecosystem’s response
to human manipulations. Indeed, scientists have recently
criticized ecosystem management for its overemphasis on
experimentation (Holling and Allen 2002). Furthermore,
the use of the word “experiment” may raise concerns
among managers and stakeholders who are uncomfortable
implementing actions with uncertain outcomes. Neutral
language, such as “researching ecosystem response to a
change in driving variables”, may more effectively moti-
vate them. Clearly, other valid modes of scientific infer-
ence, such as comparative and correlative studies (Pickett
et al. 1994), represent essential elements in advancing river
science (see Step 3 below). Nonetheless, experiments pro-
vide learning at the appropriate scale with the potential to
discriminate between competing hypotheses (Holling and
Allen 2002) and must therefore be pursued if we are serious
about fine-tuning river management for ecological sustain-
ability. Dismissing large-scale experiments as too socially
intrusive or “risky” represents a misunderstanding of sci-
ence’s role in sustainable management.
A medical analogy illustrates this important point.
Adopting experiments as a mode of learning is analogous
to the widely accepted practice of using clinical trials with
human subjects (also complex systems) in the hope of
improving medical therapies. Experiments in resource
management should not reflect the engineering model of
a quick fix to the problem, but rather the medical model of
engaging the problem in all its complexity, with the inten-
tion of achieving more efficient pathways towards sustain-
ability (M Healey, pers comm).
Step 2: Engage the problem through a collaborative
process involving scientists, managers, and other
stakeholders
Clearly, an effective study design is essential to maximize
the lessons learned from large-scale ecosystem manipula-
301
© The Ecological Society of America www.frontiersinecology.org
Figure 4. Dam operations in the upper Colorado River basin have altered natural
river flow patterns to varying degrees. Richter et al. (1998) assessed alterations in
natural hydrologic conditions important to river ecosystem structure and function for
the Colorado and Green rivers. River segments shown in blue were not evaluated.
