A key note speech given at the ASCE Wetlands Engineering & River Restoration Conference 2001
Reno, Nevada
August 28, 2001
INTRODUCTION
The explosion of interest in river restoration that has occurred in the last decade has come about because
there has been a fundamental change in the way society values its rivers and watersheds. A new societal
demand for managing and restoring rivers to sustain ecologic and environmental benefits has qualified
our earlier support for the economic utilitarian exploitation of a river’s resources. This shift poses an
unprecedented challenge –and opportunity –to both civil engineers and public works agencies that have
previously been given responsibility and until now have assumed the role of de facto river managers.
I believe that we, the engineering profession, will not be successful in achieving society’s goals for
restoring rivers until we have reconciled the conflict between two philosophies or paradigms: that of river
River engineers who were trained to facilitate exploitation of a river’s resources are now being asked to
restore them, but are not equipped with the tools, concepts, or institutional infrastructure to play this role
management paradigm, explain how conflict between them is impeding restoration and how we might
The Industrial Revolution brought not only technological advances but unleashed the power of an idea
that nature could be conquered and its resources utilized and exploited for the benefit of humanity. A
new profession was required to achieve this –civil engineering. The founding statement of the institute of
civil engineers in 1830 was “to harness the great sources of power in nature for the use and convenience
of man”.
Engineers such as General Humphries [Figure 1], architect of the plan to embank the Mississippi river,
Figure 1.
Figure 2.
River engineers used the new discoveries of science to developed new methods to control, divert,
channelize or dam rivers in order to utilize floodplains, prevent floods, irrigate fields, generate power or
improve navigation.
Our training was derived from the science of fluid mechanics tempered with practical experience in the
hydraulics of irrigation, drainage and shipping channels. Inevitably when we turned our attention to
natural rivers we viewed them as disorganized systems that needed simplifying to replicate the channels
we had studied. An example of this legacy is the ubiquitous use of Manning’s equation to define river
hydraulics –without understanding its historical context and derivation from the experience of land
To support this new applied science the societal infrastructure required for its implementation was created
largely in the 19th century. New institutions were invented--including public works agencies, new
university departments, and new legal constructs. Over the last century these institutions have played the
dominant role in determining the education, expertise, and attitudes of the men who made decisions
The main task engineers were asked to address was how to design and construct river engineering works
in a sound cost effective manner to achieve the exploitation of a river’s resources. These structural works
Engineers were not asked to be river or watershed managers, nor were they asked to consider the
ecologic, or even the long-term consequences of their actions. Nevertheless, for at least the last 150 years
we engineers have sought and been given authority to take actions drastically affecting all aspects of a
river system. We have had the authority but not the responsibility of river managers.
THE CONSEQUENCES OF RIVER ENGINEERING
Within the last century and particularly in the last 40 years, most of the world’s rivers have been
transformed by river engineering works. An example is what happened to the Sacramento River. Even as
late as the 1930s this is what large parts of the river looked like [Figure 4]. After the big multipurpose
dams and the levees this view is more typical [Figure 5]. The change in habitats was devastating. Figure
6 shows that the area of floodplain wetlands in the Sacramento Valley has been reduced to a few percent.
Figure 4.
Figure 5.
Increasingly, scientific evidence is demonstrating that major river engineering works have been a main
cause of environmental degradation of river and estuarine ecosystems, and a significant factor in the loss
of global biodiversity.
The regulation of river flows, the isolation of floodplains from their river channel, the interruption of
sediment flows and the elimination of riparian corridors has not only led to ecologic degradation but the
Figure 7.
Here you see the decline in the salmon fishery of the Sacramento River in the era of big water projects.
We now realize these changes have also resulted in unanticipated increasing natural hazards losses as we
start to understand how flood control stimulates inappropriate land use, as shown here in the town of
Olivehurst [Figure 8] inundated from a levee failure on the Yuba River in the 1986 flood. The cumulative
effect of the failure of flood control as a substitute for an integrated flood hazard reduction strategy is
Figure 9.
In the U.S. in the 1960s environmental activists started challenging traditional river engineering projects
on their economics and rationale, not just their environmental costs. River engineers began increasingly to
Figure 10.
This challenge by environmental advocates was increasingly persuasive because it relied on evidence
from two bodies of science that had been excluded from standard river engineering practice and were
outside the typical training of an engineer. These were: fluvial geomorphology, the science of the form
and evolution of rivers that matured in the 1960s--in almost complete isolation from hydraulics
engineering--and the science of ecology that also started to develop rapidly in the 1960s.
By the 1990s, aided by increasing taxpayer skepticism over the huge cost of river engineering works,
environmentalists were becoming successful in halting traditional river engineering projects and seeking
remedies for past environmental damage. This in turn stimulated a rethinking of the relationship between
The increasing importance of river restoration has presented a new challenge to river managers: to
develop a rigorous coherent methodology that would guide appropriate cost effective restoration actions
Restoration science is still in its infancy and restoration practice largely untested –with few major
initiatives that have been implemented. Nevertheless a consensus has emerged on what this methodology
should be.
There are three important concepts underlying this methodology that are not reflected in the river
engineering paradigm.
First, it is founded on a concept of physical and ecologic integrity that is intrinsic to a river ecosystem. I
Succinctly stated, it is the idea that a river is a dynamic evolving system; that a river’s form is an
expression of its watershed, climatology and geomorphic history; that rivers have an inherent tendency to
recreate that form, and that the rivers ecology has adapted to and takes advantage of both the form of the
Our need to take into account this inherent form and self correcting tendency is vividly illustrated by the
response of the Walla Walla River to channelization in the 1963 flood [Figure 12]
Figure 12.
management system that has clear measurable objectives. River management is multi-objective, has a
long term commitment, is accountable, draws on interdisciplinary expertise, has at its disposal a
continuum of management interventions from minimal maintenance to massive river engineering works,
and is adaptive through monitoring its performance. In addition river management is concerned with a
range of spatial and temporal scales, from the watershed to the coastal ocean, from the immediate
Because it addresses sustainable long-term management the river management paradigm can anticipate
and plan for the geomorphic and ecologic response to human interventions that take many decades to take
effect. We now recognize we are no longer 19th century pioneers who can exploit our rivers and then
Increasing scientific evidence indicates that the preferred way to restore the river environment is not to re-
engineer it but to encourage the river to restore itself through natural processes. This ‘self healing’ is best
done by removing or reducing human interventions, particularly those implemented by river engineers of
earlier generations. For example as was done by the breaching of levees on the Cosumnes River to
Figure 13.
Nationwide there are now increasingly ambitious proposals to restore river flows, setback flood control
levees, or remove dams to restore fisheries such as has already been done in France with demolition of
While some elements of the new paradigm have become accepted, we have not acknowledged that they
have created contradictions that result in a bias against implementing effective restoration.
At the beginning of the 21st century we are stalled by the conflict between the river engineering and the
river management paradigms. We need to be clear what the differences are between these paradigms
[Figure 15]:
S tu ck b e tw e e n tw o p a ra d ig m s
R IV E R E N G IN E E R IN G R IV E R M A N A G E M E N T
-S ingle purpose -M ulti-objective
-E ngineering expertise -Interdisciplinary
-C onstruction focus -C ontinuum of interventions
-R each scale -W atershed scale
-H y draulics tim e scale -G eom orphic tim e scale
-M onitoring externalized -M onitoring internalized
-L im ited accountability -L ong term com m itm ent
-M aintenance divorced -M aintenance is m anagem ent
from design activity
R iver E ngineering versus R iver R estoration
Figure 15.
Wherever river restoration is a goal of river management, it is the crux of this conflict.
I have identified the following consequences that bias us against implementing effective restoration
Over-engineered restoration
Many river engineers and some agencies are responding to the new paradigm as if river engineering was
itself a form of river management, rather than one tool in a spectrum of management interventions. This
mistakenly leads them to advocate ‘structural’ engineered river restoration when less expensive and more
effective management options exist. An example shown here are the massive siphons constricted over the
Rhine River levees to emulate the role of natural floodplains in storing flood peaks [Figure 16].
Figure 16.
Because river engineering has not been conducted as part of a management system, it has never been held
accountable for its performance in the way that river restoration actions now are.
To make way for the new paradigm, we have to understand clearly what is wrong with the old one.
We still do not know how effective river engineering projects have been in fulfilling their promises to
No comprehensive independent audit of our massive investment in river engineering has ever been carried
out.
This failure to learn from experience has created a double standard of accountability that biases us
towards preserving and maintaining environmentally destructive river engineering works, frustrating
Because river engineering concepts have dominated historically, river management objectives are miscast
in the old terminology. For example, we know the societal goal is not to control floods but to reduce flood
hazards and that floodplain restoration might achieve this more effectively than higher flood walls. But
emergency flood control actions, such as shown here on the Mississippi River in the flood of 1993
[Figure 17], are too often unquestioned even though they often foreclose opportunities for more effective
Figure 17.
Reductionist solutions favored
In many instances restoration actions can result in cumulative benefits such as the example of the
restoring floodplains for both ecologic restoration and reduction of flood hazards. These are frequently
not recognized from the perspective of river engineering that usually considers individual problems in
isolation rather than their interrelationship at the watershed scale. This is captured in this cartoon entitled
‘an engineers view of the river basin–a series of hydraulic problems [Figure 18].
Figure 18.
Flawed design procedures undercut restoration rationale
We now know many instances where river engineering design procedures are flawed or out of date but
have persisted as standard design practice because there has been so little interest in monitoring
performance. This can perpetuate a structural intervention where multi-objective restoration solutions
For example, the experience of channelization projects like Corte Madera Creek [Figure 19], a concrete
flood control channel designed to convey the 200-year flood assuming clear water hydraulics based on
Manning’s equation. Figure 20 shows what happened in the 1986 flood –a 6-year event that,
unremarkably, mobilized enough bed load to significantly increase flow resistance. Such experiences still
Figure 19.
Figure 20.
Some river engineering methodologies accepted as standard practice are biased against recognizing
restoration benefits. A good example is the standard use of the classic steady state HECRAS model still
This model ignores–and hence considers worthless–the beneficial effect of floodwater storage in upstream
floodplains.
Long term maintenance, life cycle and remedial costs are usually underestimated, externalized or not
recognized in river engineering projects but are internalized in river management solutions. Life cycle
and the costs of decommissioning are almost never considered in river engineering design.
Too often we see scenes like this where unachievable prescriptive maintenance is imposed on a river
manager, a result of maintenance being considered in isolation from hydraulic design. Too often in river
River engineering actions that favor simplification and predictability are considered ‘good’ and more
likely to be approved than an ecologic restoration program that restores complexity and values variability.
This is a result of our engineering culture (Figure 23] where deterministic solutions are favored and
Research agendas dictated by river engineering priorities have been slow to focus on new key restoration
questions one example: 15 years ago when we first started designing multi-objective flood channels we
needed to know the effect of vegetation on flow resistance. We found the best research review coming
not from the engineering labs but from freshwater biologists interested in challenging ecologically
University teaching is still focused on the computer and flume not on understanding real rivers.
Hydraulics engineers are still graduated without ever seeing a river in flood.
The bias for the status quo and against change
We have been bequeathed a legacy of ageing river engineering projects whose objectives were
simplistically defined, whose effectiveness are uncertain and which were planned in ignorance of their
We know that many of these projects constitute major interventions in the river system whose persistence
or continued operation can result in continued massive ecologic decline. For example:
every time we decide to repair an eroding levee we are preventing the river from reestablishing a
healthy floodplain;
every time we decide to store all flood flows, large or small, in reservoirs we eliminate flow
pulses essential for ecologic processes. The drastic effect this has can be seen in this post project
comparison of flood frequency below Oroville Dam, California [Figure 25]. The 2-year flood has
been reduced by an order of magnitude but the reduction in larger damaging flows has been
proportionately less.
Institutional inertia
Technical challenges pall in the face of a larger problem; overcoming the institutional and conceptual
legacy of the era of resource exploitation. One of the biggest obstacles to implementing river restoration
is the lack of appropriate executive agencies that have the interdisciplinary expertise, mission and
resources.
Figure 25.
The new river management paradigm is being imposed on the technological infrastructure developed to
sustain the old river engineering paradigm it is replacing. Flood control agencies established 100 years
ago to build levees are now being asked to protect endangered species; a symptom of this institutional
inertia is that federal funding is available for constructing river restoration projects but very little money
is available for land acquisition or maintenance and almost none for monitoring.
In recent years there have been a variety of imperfect responses to this problem. Some public works
agencies like the U.S. Army Corps of Engineers have expanded their mission to encompass
environmental restoration and watershed planning, in other instances ad-hoc temporary groups of single
purpose resource management agencies are formed like the Cal-Fed Bay-Delta program in California.
Elsewhere, lay ‘stakeholder’ led watershed management planning groups are attempting to fill the
vacuum.
TOWARDS EFFECTIVE RIVER MANAGEMENT
We now have the opportunity to create a synthesis of ideas that uses the best of our past river engineering
experience in the service of new societal goals. To initiate this, we need to take the following seven
actions.
We need to examine the impacts, performance and effectiveness of not just restoration projects but
past engineering interventions. This means analyzing the long term, large scale, ecologic and
geomorphic changes that have been and will be caused by river engineering projects. It means
identifying their societal objectives and determining whether these objectives have been met. It also
means revising engineering design procedures and methodologies in the light of this new experience.
River restoration encompasses those actions required to change a river from its present degraded state
to one whose management better achieves multiple societal objectives. We can do this best if we
have a consensus of what this end point looks like. An example where this has been done well is the
community led initiative that resulted in the Napa flood management plan, now being implemented
Implementing the river management vision usually requires incremental steps taken opportunistically
as funding or land becomes available –often in the wake of floods. To realize the cumulative
integrative benefits over longer time periods we need proactive templates to guide this action. A
model for such a template is the river management strategy being developed by U.S. Fish and
Wildlife Service to guide appropriate emergency flood response actions after future floods [Figure
27].
This means monitor and periodically review whether their performance or persistence meets multi-
objective river management goals. We need to understand what the costs and benefits are of
maintaining the no-action alternative, and this should be determined in the same way as for a river
restoration option. The FERC re-licensing process for private hydro dams gives an idea of what can
Figure 27.
Researchers need to address not only important technical questions such as what really is the affect of
riparian trees on channel conveyance, but examine the broader interdisciplinary questions as well,
Consulting firms engaged in river management need graduates who have interdisciplinary and
systems training. It is rare to find someone with a degree in hydraulics engineering and fisheries
We need government agencies that have the mission, mandate, resources authorities and skills to
effectively manage rivers. This means rethinking the role of flood control agencies as well as natural
resource management agencies. For example, the Santa Clara Valley Water District is currently
attempting to transform itself from a flood control and water supply agency to a watershed
stewardship agency.
We need to understand the interaction of ecologic and physical processes to develop informed river
management solutions. This will require a full integration of the sciences of fluvial geomorphology,
fluid mechanics and sediment hydraulics as river science. In the 1960s this was the intent of
An example of the new kind of challenge and achievement of the river manager is this log jam on the
Stillaguamish River, WA [Figure 28]. What is shown here is actually an engineered structure that
has been designed to take advantage of natural processes to recreate fish habitat in a degraded river.
We need to establish a new professional practice that incorporates the best of the established
pragmatic professional traditions of the engineer including qualities such as accountability, liability,
quality control, rigorous design standards and professional training; and redirects it towards restoring
If river restoration is seen only through the lens of the past, as a new form of river engineering; we are
unlikely to achieve success. Instead we need to create a new profession of river manager to meet the 21st
century demands for river restoration–just as river engineering was created to meet the needs of the
industrial revolution.