United States
^MRUpHQ Environmental Protection
^1	Agency
Application of the Riverine
Ecosystem Synthesis
(RES) and the Functional
Process Zone (FPZ)
Approach to EPA
Environmental Mission
Tasks for Rivers
RESEARCH AND DEVELOPMENT

-------

-------
EPA/600/R-11/089
Sept 2010
www.epa.gov
Application of the Riverine
Ecosystem Synthesis (RES) and the
Functional Process Zone (FPZ)
Approach to EPA Environmental
Mission Tasks for Rivers
Joseph E. Flotemersch1, James H. Thorp2, and Bradley S. Williams2
1	U.S. Environmental Protection Agency, National Exposure Research Laboratory, 26
W. Martin Luther King Dr., Cincinnati, OH 45268
2	Kansas Biological Survey, University of Kansas, Higuchi Hall, 2101 Constant Ave.,
Lawrence, KS 66047-3759
Notice: This document has been reviewed in accordance with the U.S.
Environmental Protection Agency policy and approved for publication. Mention of
trade names and commercial products does not constitute endorsement or
recommendation for use.
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460

-------
Acronyms
DEM
Digital elevation model
EPA
United States Environmental Protection Agency
EPSCoR
Experimental Program to Stimulate Competitive Research
FCL
Food chain length
FPZ
Functional process zones
GIS
Geographical information system
GPS
Global positioning unit
LIDAR
Light detection and ranging
NDH
Network dynamics hypothesis
NERL
National Exposure Research Laboratory
NGO
Non-governmental organizations
NS
Nutrient spiraling
NSF
National Science Foundation
ORD
Office of Research and Development
RCC
River continuum concept
RES
Riverine Ecosystem Synthesis
SpD
Species diversity
STAR
Science to Achieve Results
USGS
United States Geological Survey
-l-

-------
Table of Contents
Acronyms	i
1.0 Introduction and Background	1
2.0 Business Case	3
3.0 Problem Statement	3
3.1	Why Identifying the Hydrogeomorphic Character of Rivers is Important	3
3.2	Alternative Approaches to Characterizing and Classifying Rivers	4
3.3	Why Hierarchical Scale is Important	5
3.4	Lateral Perspectives, Classification Schemes, and Watershed Management	6
3.5	Linking Classification, Ecological Functioning, and Environmental Sensitivity	7
4.0 Proposed Solutions	7
4.1	Introduction to Solutions	7
4.2	Recommended Steps	9
4.3	Applications of Solutions	10
5.0 Future Directions	10
5.1	Government Training Course in FPZ Delineation	10
5.2	National Rivers Classification Manual	11
5.3	Documents on Applications of the RES/FPZ Approach to Other Mission Tasks	11
5.4	Integration with Previous and On-going EPA Analyses	11
6.0 Conclusions	12
Appendix A: Summary of Some Potential Applications of the FPZ Approach	13
A. 1 National Framework for River Classification	13
A.2 Monitoring Design and Study Reach Lengths	13
A.3 Reference Site Selection and Condition Assessment	13
A.4 Ecosystem Services	14
A.5 Asset Trading	14
A.6 River Rehabilitation	15
A. 7 Watershed Management	16
Appendix B: Summary of FPZ Delineation Techniques	18
References	19
List of Figures
Figure 1. A conceptual riverine landscape depicting various functional process zones (FPZs)
and their possible arrangement in the longitudinal dimension	2
Figure 2. Organizational hierarchies in river science	6
-li-

-------
1.0 Introduction and Background
The shift to watershed management of rivers from a more reach-based approach has had
far-reaching implications for the way we characterize and classify rivers and then use this
information to understand and manage biodiversity, ecological functions, and ecosystem services
in riverine landscapes. At the same time, we are faced with inherent challenges of how to best
take advantage of past studies (e.g., the many projects on river classification funded by the U.S.
Environmental Protection Agency's [EPA's] Science to Achieve Results [STAR] program) while
we shift to the higher hierarchical scale necessary to manage at the watershed level. To meet
these challenges, we require a model that links the physical structure of a river with its
ecosystem functioning and allows us to evaluate past, present, and future river conditions. Such a
model would ideally be cost effective, easy to employ, and capable of answering questions at
different hierarchical scales in river basins of varying sizes. One model that meets all these
criteria, while also accommodating many of the prominent approaches used by and/or developed
in collaboration with the EPA, is the Riverine Ecosystem Synthesis, or RES (Thorp et al. 2006,
2008).
Contrasting with earlier views of rivers as simple, continuous gradients in physical
conditions from headwaters to great rivers (i.e., river continuum concept [RCC]), research and
conceptual models in the last decade support the conclusion that rivers are more accurately
portrayed as downstream arrays of large hydrogeomorphic patches formed by factors such as
hydrologic patterns, geomorphic structure of the channel bed and valley, climate, and riparian
conditions (e.g., Montgomery 1999; Poole 2002; Thorns and Parson 2002, 2003; Thorp et al.
2006, 2008). These patches are described in the RES, at the critical valley-to-reach scale, as
functional process zones (FPZs). FPZs are named based on statistically-derived features of the
channel and surrounding valley along with geological and precipitation features, but some
widely known examples of channel types in different FPZs are constricted, meandering, braided,
anastomosing, and distributary. According to the RES, FPZs are repeatable along the
longitudinal dimension of rivers and only partially predictable in location (Fig. 1), especially at
scales above the ecoregional level. Because of physicochemical habitat differences, ecosystem
structure and function vary significantly (and predictably) among FPZs.
Use of the RES model in river management is just beginning to expand, especially as it
relates to tasks characteristic of EPA's mission. The FPZ approach is being applied at present to
the Kansas, Kanawha, and Neuse rivers in the U.S. and has previously been applied to dryland
rivers on other continents. However, this approach needs to be applied and evaluated for a fuller
spectrum of ecoregions, such as those characterizing the humid through arid regions and/or
northern through southern portions of the U.S. Starting with a foundation of ideas from "river
typing" work in Australia, we have now been able to accelerate the river typing process and are
starting to explore its use in multiple environmental tasks in the EPA mission. While we are now
involved in the planning and execution phases for some applications of the RES (e.g., the
physical classification of rivers), more research and development is needed to firmly establish
links between the physical and ecological portions of the RES. We propose to research, develop,
pilot, and implement the products necessary to successfully apply the RES concept and FPZ
approach to the mission tasks facing EPA.
-1-

-------
©!
FCL^^

/\

¦
Pool
Gorge
/
:l| ns jspp
ft
Braided
Meandering
Armored
Pool
Pool
NS Sp
Meandering
/
Meandering
Braided
Anabranch
Distributary
y&


INFORMATION KEY
Flow Flow
Pulse ii\ History
Flow
Regime
HYDROLOGY
Lateral
Longitudinal
¦
^ Vertical
CONNECTIVITY
FCL Food chain length
BgM Nutrient spiralling
MSlil Species Diversity
ECOSYSTEM FUNCTION
Figure 1. A conceptual riverine landscape depicting various functional process zones (FPZs) and
their possible arrangement in the longitudinal dimension. Not all FPZs and their possible spatial
arrangements are shown. Information contained in the boxes within the figure depicts the
predominant hydrological (i.e., flow pulse, flow history, and flow regime [Thorns and Parsons
2002]) and ecological (i.e., food chain length, nutrient spiraling, and species diversity) conditions
predicted for each FPZ. The ecological measures are scaled from long to short (i.e., low to high
for species diversity); the light bar indicates the expected median for each ecosystem function
and the shading estimates the range of conditions. The size of each connectivity arrow reflects
the magnitude of vertical, lateral, and longitudinal connectivity. [Revised from Fig. 1.1 in Thorp
etal. (2008).]
-2-

-------
This report presents some of the high points of the RES model, describes its uses in meeting
tasks in EPA's environmental mission, and integrates it with both current and past classification
and management techniques, as a way of improving implementation of mission tasks.Business
Case
Protection of riverine ecosystems, especially at the watershed level, requires accurate
knowledge of how differences in physical structure among sections of a river can alter the river's
fundamental ecological structure (e.g., species richness) and function (e.g., nitrogen processing
and carbon sequestration). This knowledge can greatly improve the ability to select reference
sites, contribute to the robustness of condition assessment, maximize ecological endpoints of
restoration, evaluate actual and potential ecosystem services, and establish a fair basis for asset
trading. A focus on hydrogeomorphic patches at the valley-to-reach scale (i.e., FPZs) will
improve EPA's ability to set study-reach lengths that are both mission-relevant and feasible
within a watershed approach to basin management. Furthermore, a statistically rigorous approach
to delineating FPZ that relies primarily on geospatial analysis will provide an efficient, national
framework for river classification at multiple spatial scales.
This report emphasizes both the "research and development" and "planning and
execution" needed to employ the FPZ framework for multiple components of the EPA mission.
In particular, it suggests a multi-component program over time to: (i) further develop and refine
applications of FPZs for specific EPA mission goals (e.g., reference site selection, condition
assessment, ecosystem services evaluation, etc.); (ii) test the efficacy of this approach for rivers
in multiple EPA regions by first statistically delineating FPZs from geospatial data and then
testing FPZ distributions against ecological variables previously generated by EPA and state
aquatic data sets; (iii) recommend specific, future plans for implementing these approaches at the
national and regional levels within EPA; and (iv) produce documents delineating the uses and
techniques of the FPZ approach for publication both within EPA and external to the Agency, in
refereed scientific literature. Plans for employing the FPZ framework in EPA mission tasks are
described in Section 5.0, Future Directions, and some of the major applications of the FPZ
approach are summarized in Appendix A.
3.0	Problem Statement
Federal and state agencies are increasingly faced with two daunting tasks. First, they
must distinguish between and evaluate sections of riverine ecosystems for multiple purposes,
such as reference site selection, rehabilitation, and asset trading (i.e., the dual process of river
characterization and classification). Second, they need to have a link between river classification
and the ecological functioning and environmental sensitivity of those sites. For both tasks, it is
vital that patterns be identified and processes be evaluated at correct hierarchical scales in a
quantitative, statistically rigorous fashion.
3.1	Why Identifying the Hydrogeomorphic Character of Rivers is Important
Identifying the hydrogeomorphic nature of a river section is vital for many reasons, as
illustrated in the following situations potentially facing government environmental agencies;
additional examples are given in Appendix A.
-3-

-------
•	Suppose a state environmental agency has to decide whether to allow a company to
degrade the quality of a 10-km section of a river in return for improving another 10-km
section. Is this a fair trade in ecosystem services from a regulatory perspective? How
much of an improvement would be required to at least balance the proposed degradation
elsewhere? The answers here require knowledge of the current FPZs of each section (and
original FPZs, if the river has been extensively altered) and an understanding of how each
section is likely to respond to the proposed changes from both hydrogeomorphic and
ecological perspectives.
•	In response to the need to identify reference sites, this same state agency selects site R-2
as a reference site based on water quality parameters. A later comparison of site R-2 with
"impaired" site 1-4 indicates, however, that the "impaired site" actually has greater
species richness. The reason for this anomaly could be based on the hydrogeomorphic
differences between the two sites. For example, if site 1-4 was a multi-channeled FPZ and
site R-2 was a constricted channel FPZ, their community compositions would likely vary
significantly, even if both were pristine.
•	After identifying various target areas for river rehabilitation, the agency must prioritize
their actions because of limited funds. Following evaluation of the essential
socioeconomic and environmental issues for multiple sites where dam removal, set-back
levees, or floodplain connections have been proposed, the agency could ask the following
questions: What are the original, present-day, and future FPZs (including likely FPZs if
rehabilitation efforts are undertaken) for each site? What would be the relative value
(e.g., ecosystem services, etc.) returned for every restoration dollar spent at each site? If
site A is a potential meandering, single channel FPZ while site B is a potential
anastomosing, multiple channel system, would the cost/benefit ratio be the same at each
site for a levee set back 100 m, 500 m, or 1 km?
Answers to some of the questions raised in these examples are discussed in a recent
manuscript in Bioscience entitled "Linking ecosystem services, rehabilitation, and river
hydrogeomorphology" (Thorp et al. 2010).
3.2 Alternative Approaches to Characterizing and Classifying Rivers
Recently, attempts to develop a "national river classification" system for use by
government agencies and non-governmental organizations (NGOs) have begun. An EPA-
sponsored workshop in Michigan in February 2009 contributed significantly to this process by
identifying the need to classify rivers at multiple scales, including the ecoregion, basin, valley
(i.e., the valley-to-reach), and reach levels.
River classification schemes rely on investigators to first measure a set of fundamental
and/or derived attributes for a river section and then place that section into a category that best
fits the set of attributes found. Fundamental attributes for rivers consist of physical habitat
features (i.e., principally geomorphic, hydrologic, and climatic attributes). Derived attributes are
biotic features (i.e., species composition, species abundance, etc.), which vary in response to
both natural and anthropogenic variables. Classification schemes based on fundamental attributes
can be used to answer many other questions (e.g., evaluating ecosystem services), while those
-4-

-------
based on derived attributes, while useful in their own right, are more limited in their applications
to other questions. The process of classification can vary between qualitative (e.g., comparing
investigator measurements of a river channel pattern with photographs in a manual) and
quantitative and statistically rigorous. Acquiring data for these approaches can vary from
expensive, labor-intensive, bottom-up approaches to relatively cost-effective, top-down
approaches using geospatial data.
Fundamental attributes can be measured and evaluated at different spatial scales. The
most commonly used is the reach scale, where extensive measurements are made using bottom-
up approaches like the Rosgen Method (e.g., Rosgen 1994, 1996, 2006); extension of this
method to higher spatial scales requires the problematic merger of measurements and results
from multiple spatial scales over a large area. In contrast, the FPZ approach is employed at a
higher spatial scale (i.e., valley-to-reach scale) and can employ either the more efficient top-
down geospatial data (e.g., digital elevation model [DEM], remote aerial imagery) or more labor-
intensive bottom-up measurements.
Derived attributes typically consist of either taxonomic (e.g., species) or functional group
(e.g., cold-water vs. warm-water) compositions of biota (e.g., fish, macroinvertebrates, etc.). In
classification schemes based on these attributes, the investigator starts with a known distribution
of species and then correlates the distribution with certain natural (e.g., temperature, water
hardness, etc.) or anthropogenic variables (e.g., nitrate levels, land use, etc.) in an attempt to
infer causative mechanisms. Using this approach, the investigator can identify gaps in the
distribution range of a species (later seeking either new species distribution records or
explanations for its absence). Another goal of this approach is to make predictions on
distributions outside the known distribution range to other river sections or even different rivers.
This approach, while very useful, is difficult without adequate information on the
hydrogeomorphic structure of the rivers. GAP models and the National Fish Habitat Initiative are
examples of this widely-used approach in the U.S.
Earlier this decade, the STAR program funded a large number of studies seeking to
classify rivers; most used derived characters, but a few employed fundamental characteristics of
the channel or watershed to predict ecosystem structure or, on rare occasion, ecosystem function.
3.3 Why Hierarchical Scale is Important
Protecting, rehabilitating, and managing riverine ecosystems requires accurate knowledge
of the hydrogeomorphic nature of the river section(s) under consideration The hydrogeomorphic
nature of a river directly affects ecosystem structure and function by altering spatial and temporal
components of the habitat template (Frissell et al. 1986) within the riverine landscape (e.g.,
wetted channels, slackwaters, and floodplains). For some questions, knowledge of the stream
order/size and position downstream (as in the RCC) or the number of upstream tributary
connections (as in the network dynamics hypothesis [NDH]; Benda et al. 2004) provides
valuable information needed for river management. In other cases, however, investigators need
additional or alternative higher spatial scale data; this is particularly true when attempting to
manage rivers at the ecosystem level.
Matching the appropriate spatial scale of analysis with the ecological question or
environmental task being addressed is vital to obtaining accurate and relevant answers (Fig. 2),
as also discussed in a separate manuscript nearing submission (Thorp, Flotemersch, et al., In
Prep.). Mismatches of spatiotemporal scale and management goals are all too common around
-5-

-------
the world (Thorp et al. 2008). For tasks involving many aspects of watershed management, the
appropriate ecological level at which to evaluate ecosystem function or ecosystem services is the
valley-to-reach scale, or FPZ (see Appendix A-7). Evaluations at higher spatial scales are rarely
both mission-relevant and economically feasible for EPA and the states to employ. Analyses at
smaller spatial scales, such as the reach level, are particularly useful for providing detailed data
on questions involving, for example, point-source pollution and can also provide information on
mechanisms operating at higher spatial scales. However, from a purely economic perspective, it
is considerably more expensive and problematic to merge detailed data collected at the reach
scale to answer broader scale questions than it is to collect sufficiently accurate, but less spatially
precise data at the FPZ level.
_N
C
CTJ
CS)
i	
O
B
v>
¥
a>
Geomorphological Hydrological
organizational organizational
hierarchy	hierarchy
Drainage basin ¦
Functional process )
one
Ecological
organizational
hierarchy
Functional set )
Functional unit i
Micro habitat !
Community
Organism
100 -
10
+ ^
o
CA
1 -
0.1
Scale
extent
extent
grain
exlenl
gram
0.01
grain
Figure 2. Organizational hierarchies in river science. To use this framework, one must first
define the relevant spatiotemporal dimension for the study or question. Scales for each hierarchy
are then determined to allow the appropriate levels of organization to be linked. The scale at the
right demonstrates that linking levels across the three hierarchies may be vertical depending on
the nature of the question. [From Fig. 3.2 in Thorp et al. (2008).]
3.4 Lateral Perspectives, Classification Schemes, and Watershed Management
River ecosystems consist of complex riverine landscapes composed of the riverscape
(i.e., main channel and lateral slackwaters, such as bays, side channels, backwaters, etc.), and the
floodscape (e.g., isolated oxbows, lakes, wetlands, and usually dry alluvial floodplains; Thorp et
al. 2008). Consequently, watershed management of riverine ecosystems necessarily requires
-6-

-------
ecosystem models and classification schemes that incorporate both the riverscape and
floodscape. Reach-level analyses, such as those employing the Rosgen Method, and most
derivative models typically focus solely on the riverscape (i.e., the main channel only or
occasionally, the main channel and slackwater areas). For example, if a river is classified as
"warm-water fisheries," useful information can be gleaned about what species should live there,
but no information is provided on how hydrogeomorphically complex the system is or what kind
of interactions would be expected to occur between the main channel, slackwaters, and terrestrial
watershed. Likewise, both channel-oriented (e.g., RCC and NDH) and floodplains-oriented
models, such as the flood pulse concept (Junk et al. 1989; Junk and Wantzen 2004), only include
a portion of the riverine landscape. In contrast, the Riverine Ecosystem Synthesis is designed to
encompass (with FPZs) the entire longitudinal and lateral dimension of the riverine landscape. In
fact, the lateral extent of an FPZ is for all practical purposes, the distance from the hillslope on
one side of the river to the hillslope on the opposite side. It is within this area that the channel(s)
move and/or interact during floods, and these areas are directly pertinent to EPA's program goals
for Healthy Watershed Initiative through the Office of Water. FPZs are present from first-order
streams downstream to the mouth of the river, but their delineation using solely top-down
approaches is limited by the investigator's ability to distinguish the channel in a canopied
covered area and to gain access to high-quality elevation data, which may require LIDAR data
(see Appendix B).
3.5 Linking Classification, Ecological Functioning, and Environmental Sensitivity
The primary purposes for constructing classification schemes are to: (i) enable the
investigator or government regulator to compare and contrast different river sections; (ii) infer
ecological structure and function as well as ecosystem services; and (iii) predict effects of
anthropogenic change (i.e., disturbance or rehabilitation) to that river section. To do so, the
classification scheme needs to be quantitative, statistically rigorous, and either embedded in a
more comprehensive model or linkable a posteriori to its ecological components. A distinct and
rather unique advantage of the RES is its multi-faceted nature. That is, the RES incorporates a
physical model based on FPZs, serves as a hierarchically-scaled investigative framework, and
contains explicit ecological components linked to the physical model. The RES currently
contains a set of 17 ecological hypotheses that address issues ranging from species distributions
to landscape properties; as more is learned about rivers, these hypotheses can be expanded in
detail and number. Equally valuable from EPA's standpoint is that the RES is flexible enough
both to incorporate information obtained using many other approaches and to provide valuable
analytical tools to those other models. For example, the physical variables in many GAP-type
models would benefit from additional information on the arrangement of FPZs in relation to
species distributions in the river(s) being assessed. Likewise, the distribution of fishes from
habitat assessment models could be used to test the predictions of the RES' ecological
hypotheses and predict the effects of changing the local FPZ(s).
4.0	Proposed Solutions
4.1	Introduction to Solutions
Our proposed solutions to the issues described in Section 3.0, Problem Statement,
-7-

-------
involve: (i) use of the RES as the comprehensive model for ecosystem management at the
watershed level, including rehabilitation projects; (ii) inclusion of the FPZ approach as one of the
hierarchical levels in developing classification schemes for rivers throughout the U.S.; (iii)
incorporation of data from other EPA sampling programs to improve predictions of the RES; and
(iv) employment of the FPZ approach to improve predictions of larger-scale derivative models.
Within the RES, the FPZ approach is hydrogeomorphically-based, of direct ecological relevance
(Fig. 1; Thorp et al. 2008), and scaled to be economically feasible and mission relevant. In
addition, the FPZ approach has seven other characteristics especially important to EPA:
1.	Delineation of FPZs is quantitative and statistically rigorous. [See Appendix B for a brief
description of how FPZs are delineated.]
2.	The FPZ approach can easily be merged with many previous approaches used by EPA
over the last two decades. For example, the approach fits easily with ecoregional
classifications; it can provide information helpful for data analysis in monitoring studies
that have employed random or stratified random approaches; and it can use data from
other studies in its own RES model to link FPZs with ecosystem function.
3.	FPZs can be delineated using top-down, geospatial approaches, thereby greatly reducing
personnel costs and time delays (see Appendix B). Alternatively, FPZs can also be
delineated using previously collected, bottom-up data and some geospatial information
on the watershed.
4.	Although FPZs are most easily delineated from current conditions, it is possible in some
cases to evaluate past and future conditions in relationship to anthropogenic
modifications of the channel and watershed, thereby aiding mission tasks, such as
rehabilitation actions.
5.	Once a river's FPZs have been delineated, the FPZ composition is relatively permanent
and is subject to change only with major changes to the watershed and channel (addition
or removal of dams, levees, etc.). Therefore, the FPZ delineation can be used for many
future tasks without periodic re-analysis.
6.	The FPZ approach can be applied anywhere within the U.S. or world, even if a given
river has not been sampled previously and is relatively inaccessible or difficult to sample
by traditional ground methods, as long as the needed data layers (principally geospatial
data, but some geologic and precipitation data as well as remote sensing imagery of
channels) are available.
7.	Finally, the FPZ approach is not limited in its application to one or two tasks in the EPA
mission, but instead can be employed in a wide variety of ways with past and future data.
Some of the potential uses of this approach are summarized in Appendix A.
The concept of linking hierarchically-scaled components of fluvial geomorphology to
ecosystem structure and function in longitudinal, lateral, vertical, and temporal dimensions of
riverine ecosystems first began to coalesce following a 2003 plenary session talk by J.H. Thorp
at a regulated rivers meeting in Australia organized by M.C. Thorns. This led to development of
a journal article on the riverine ecosystem synthesis (Thorp et al. 2006). The original model was
based on fundamental theory and pristine systems; however, plans to expand the model to
modern, regulated rivers and apply it to the environmental missions of governments and NGOs
were underway before 2006, culminating in the 2008 Thorp et al. book. The hydrogeomorphic
-8-

-------
patchiness of rivers has been recognized for decades by fluvial geomorphologists, but the
division into repeatable patches (FPZs) at the valley-to-reach scale was refined by Thorns and
Parsons (2002, 2003).
Use of the RES model in river management is in its infancy, especially as it relates to
tasks characteristic of EPA's mission. There are two reasons for this. First, while the FPZ
approach to watershed management has been successfully applied in dryland rivers in Australia
(i.e., the Murray-Darling River system) and South Africa (i.e., rivers in Kruger National Park),
the approach needs to be applied and evaluated for a fuller spectrum of ecoregions, such as those
characterizing the U.S. Details of these applications are described in Thorp et al. 2008, as are the
conclusions that many assessments of river condition use data collected at an inappropriately low
level or scale (e.g., the reach or site level) to infer catchment-scale condition and manage river
ecosystems. This prior work on other continents and our current work on the Kansas, Kanawha,
and Neuse rivers in the U.S. will enable us to accelerate analyses of ecoregions in the U.S. using
the FPZ approach. Second, the link within the RES between the physical model (the nature and
distribution of FPZs) and ecosystem function is primarily conceptual at this point, although there
is a strong body of aquatic literature supporting the concept and likely links. Therefore, while we
can begin planning and execution of some applications of the RES (e.g., the physical
classification of rivers), more research and development is needed to firmly establish links
between the physical and ecological portions of the RES.
4.2 Recommended Steps
The following are suggestions (in recommended chronological order) for how to refine
and employ the FPZ approach for EPA's use in completing its mission to protect riverine
environments:
1.	Refine techniques and protocols for rapid delineation of FPZs using computer-based, top-
down geospatial approaches for integrating geomorphic, climatic, and hydrologic data
and employing statistical clustering and analysis techniques. Develop a user manual for
employing these ArcGIS-based techniques. [Note: We are currently refining and applying
the approaches for EPA's use, but we have not yet written and tested a user manual.]
2.	Test the efficacy of this approach with field data from rivers in one or more EPA regions
by: (i) statistically delineating FPZs from geospatial data; and (ii) conducting a pilot
study testing the FPZ distribution against ecological variables in aquatic data sets
generated previously by EPA. [This task began for the Kanawha River in the summer of
2010, but it needs to be complete there and extended to other types of rivers.]
3.	Recommend specific plans for implementing these approaches in future mission tasks at
the national and regional levels within EPA; begin these task activities once funds are
available.
4.	Determine how the RES concept/FPZ approach can be best integrated with other past and
present EPA river management approaches, classification schemes, and field data.
5.	Produce documents describing, in greater detail, the uses for the RES and FPZ approach
and methods for delineating FPZs; publish the information in EPA documents and
refereed scientific literature.
-9-

-------
4.3 Applications of Solutions
Almost all rivers have multiple FPZs, but the types, diversity, total number, average
downstream expanse, and distribution will vary among rivers. As rivers increase in size
downstream, the length of an individual FPZ generally increases and the diversity of FPZ types
decrease. The ability to predict the types and distribution of FPZs for a river decreases above the
ecoregional level because of inherent changes in climate, geology, and topography, all of which
impact a river's hydrogeomorphic characteristics. River regulation frequently changes the local
FPZs, primarily through alteration in the channel form and numbers, bed characteristics, flow
patterns, and interactions with the riparian zone and watershed.
Success in delineating FPZs for any particular river, once the procedures are refined, will
depend primarily on access to geospatial data (see also Appendix B). DEM data of at least 10-m
pixel size are sufficient in most cases; 10-m DEM data is currently available for most of the U.S.
and can be obtained from the U.S. Geological Survey (USGS). While finer-resolution, remotely-
sensed data, such as light detection and ranging (LIDAR) data, are more precise (and thus could
produce more accurate FPZ delineations, especially for headwater streams), these data have
some disadvantages. First, evaluating LIDAR data for large watersheds demands much greater
computer processing speeds, huge data storage capacity, and software that can handle these
monumental data sets. Second, the spatial scale of the LIDAR data is much more precise than is
needed for all but the smallest streams at the valley-to-reach scale of FPZs. However, if both the
LIDAR data and computing capabilities are available, then it is a good option. Much of the
precipitation data required for FPZ delineation on most rivers of the conterminous U.S. is
available at no cost from multiple sources (e.g., National Climate Data Center, or the PRISM
Group at Oregon State University). For the channel planform parameters used in the model,
access to remote sensing imagery is very useful. Such data can be obtained for some areas and
for growing seasons for free from the National Agriculture Imagery Program (NAIP) or be
purchased in bulk from sources such as the commercial company DigitalGlobe®; this company
provides geo-referenced satellite images, aerial photographs, and maps for sites throughout the
U.S. Success in determining the original FPZ of a site (prior to regulation) will depend on the
historical/archive data available; however, reasonable estimates can sometimes be made from the
valley characteristics depicted in the site's current imagery.
5.0	Future Directions
Once the techniques for rapidly delineating FPZs are refined in the first year of the
project and the approach has been field tested using current EPA data sets, it would be possible
to move forward simultaneously on tasks listed in Section 4.0 and those described below. The
first two "directions" below have relatively defined objectives, while the third and fourth are
more diffuse and could involve many avenues of pursuit.
5.1	Government Training Course in FPZ Delineation
As soon as the techniques for FPZ delineation are refined and simplified for rapid, but
statistically rigorous use and a user manual has been written, a five-day introductory training
course could be developed for federal employees, other scientists, and river managers. This
course could: (i) briefly familiarize participants with the principles of river science that are
-10-

-------
fundamental to understanding the ecological and hydrogeomorphic bases of using FPZs; (ii)
present a background introduction to the nature and availability of environmental data necessary
to delineate FPZs; (iii) discuss uses of the FPZ approach for accomplishing the environmental
mission of EPA, other government agencies, and NGOs; and (iv) provide extensive, hands-on
training in delineating FPZs. Parts i-iii would require much of Day 1, while Part iv would occupy
the remainder of the week. Interactive, PowerPoint-based lectures could be supplemented by a
printed manual (developed by EPA) and computer software on FPZs and their delineation. In
addition, optional textbooks could be made available for use, including The Riverine Ecosystem
Synthesis (Thorp et al. 2008) and a primer on fluvial geomorphology.
5.2	National Rivers Classification Manual
One recommended goal of this report is the development of a national river classification
framework that employs the FPZ approach for classifying rivers at multiple spatial scales. As a
corollary to this, EPA should consider publishing a National Rivers Classification Manual,
generically comparable in scope to the ecoregional manuals for terrestrial (Ricketts et al. 1999)
and aquatic systems (Abell et al. 2000). Such a manual could include an introductory chapter
describing the scientific basis for this framework and its integration with other ecoregional and
watershed approaches, followed by individual chapters on major river systems of the U.S.
Individual chapters would include information on FPZs of the river ecosystem and any additional
environmental information, as desired. Depending on the legal ramifications, it might be possible
to extract and reprint maps and other desired information from the rivers manual by Benke and
Cushing (2005) with agreement from the editors and publisher; alternatively, EPA could
independently produce similar river basin maps and information.
The steps to producing a National Rivers Classification Manual include: (i) developing
the national framework for classification (as discussed in this report); (ii) selecting target rivers;
(iii) statistically delineating FPZs for the main channels and as many tributaries as is cost- and
time-effective; and (iv) preparing, publishing, and distributing the manual.
5.3	Documents on Applications of the RES/FPZ Approach to Other Mission Tasks
The FPZ approach within the RES could be employed to help address other EPA mission
tasks, such as challenges related to ecosystem services, river rehabilitation, and asset trading. We
recommend the development of document(s) addressing the use of the RES/FPZ approach in
these specific tasks. As an initial step in the process, a workshop on ecosystem services and river
hydrogeomorphology was held at the University of Kansas' Kansas Biological Survey in
December 2008 under the sponsorship of the EPA's Office of Research and Development (ORD)
National Exposure Research Laboratory (NERL) in Cincinnati and a grant from the State of
Kansas' National Science Foundation (NSF) Experimental Program to Stimulate Competitive
Research (EPSCoR). The first product from this workshop was a manuscript linking ecosystem
services, rehabilitation, and river hydrogeomorphology (Thorp et al. 2010); a follow-up book is
currently being considered.
5.4	Integration with Previous and On-going EPA Analyses
In addition to applying the FPZ approach to future mission tasks at the EPA, this
approach can be applied to past and on-going studies, with the former involving re-analysis of
-11-

-------
existing biological data trends. For example, if an assessment study was previously conducted
using a random or stratified-random design, the data could be sorted by FPZs separate from or
within ecoregions, and then the data re-analyzed to see if predictability improved. An advantage
of this would be that the prior sampling design would not be lost, but merely integrated with the
FPZ approach. Depending on the flexibility of on-going regional or national studies (e.g., the
national survey of non-wadeable streams), the FPZ approach could also be integrated into the
sampling design or later statistical analyses of these various projects.
6.0 Conclusions
This report recommends adoption of both an internationally-proven approach to river
characterization and classification based on hydrogeomorphically-defined sections of rivers at
the valley-to-reach scale (i.e., functional process zones) and a watershed-level management of
rivers based on the Riverine Ecosystem Synthesis concept. Functional process zones are at the
appropriate hierarchical scale for feasibly assessing catchments, and the FPZ approach is
applicable to a wide diversity of mission tasks and can be easily integrated, using the RES, with
current approaches to analyze present-day, future, and past environmental data.
Five initial steps for adoption of this approach are recommended: (i) refine the protocols
and produce a user manual for delineating FPZs for EPA's use (partially underway); (ii) test the
FPZ approach with environmental data previously collected by EPA for rivers in multiple EPA
regions; (iii) analyze the diversity of uses of the FPZ approach for EPA's mission to protect
riverine ecosystems; (iv) determine how the RES/FPZ approach can be best integrated with other
past and present EPA river management approaches, classification schemes, and field data; and
(v) publish EPA and externally refereed documents on the application of this approach to
environmental protection and management of rivers (begun with Thorp et al. 2010).
-12-

-------
Appendix A: Summary of Some Potential Applications of the FPZ Approach
Below are summaries of some major applications of this hydrogeomorphic approach to
EPA's mission of protecting riverine ecosystems.
A.l National Framework for River Classification
One of the initial uses of the RES/FPZ approach is its contribution to the development of
a national river classification framework. To solve the primary problems described in Section
3.0, we recommend that EPA incorporate the quantitatively and statistically rigorous FPZ
approach to classify rivers of the U.S., with an initial focus on delineating FPZs in major and
otherwise important rivers. This FPZ approach would be integrated with hierarchical levels
larger (e.g., ecoregion) and smaller (e.g., reach level) the valley-to-reach scale at which the FPZs
are delineated. The FPZ approach, which emphasizes the hydrogeomorphic patchiness of rivers,
will benefit greatly from more economical top-down approaches that rely primarily on geospatial
data and can be integrated with more traditional bottom-up approaches at the reach level, such as
the Rosgen Method, to delineate smaller spatial areas (e.g. very small headwater streams whose
channels are obscured by canopies through most of the year), thereby developing a broad river
classification scheme for the entire river network. The FPZ delineation of major U.S. rivers can
proceed rapidly in an assembly-line fashion once initial procedures are refined, the approach is
tested on a few rivers (e.g., the Kansas and Kanawha Rivers have already been delineated),
priorities are set for river selection, and funding sources are identified.
A.2 Monitoring Design and Study Reach Lengths
Monitoring designs are typically: (i) unit-based (e.g., samples per a set distance); (ii)
stratified by some natural or anthropogenic feature of the river (e.g., above, between, or below a
chain of reservoirs) or land (e.g., ecoregions or political boundaries); and (iii) either stratified-
random or statistically random within the river network. Most of these sampling designs would
benefit from the simple inclusion of information on the hydrogeomorphic nature of the river
being assessed.
Because FPZs are ecologically relevant, statistically delineated, and intermediate in size
between reaches and entire watersheds, their use would enable EPA to set study-reach lengths
that are mission-relevant, logistically feasible, and economically flexible.
A.3 Reference Site Selection and Condition Assessment
The ideal reference site would be pristine and comparable in size, hydrogeomorphic
nature, and community composition to that originally present in streams now considered
impaired. Pristine or even near-pristine streams are difficult to locate throughout much of the
U.S.; instead, states are often forced to use least-impaired systems for comparison. As discussed
in Section 3.1, however, a problem arises when the quasi-reference sites differ in the nature of
their FPZ from comparative streams. In those cases, hydrogeomorphic conditions could be so
different in the natural state that valid assessment of impairment could be difficult to detect or a
challenge to defend in court. The delineation of FPZs offers a scientifically defensible method
for the characterization of river sections that facilitates comparison to other sections of river that
-13-

-------
are equivalent in both structure and function. These "comparable" sections may be within the
same river or in other rivers. With the ability to account for an increased amount of the natural
variability inherent to a system, EPA's ability to accurately assess the condition of rivers, and
sections within rivers, will be greatly enhanced.
A.4 Ecosystem Services
The ecological services provided by a river section in the past, present, and future are
linked directly to ecosystem structure and function, both of which are directly influenced by the
hydrogeomorphic nature of that section of the river and how it has been impacted by natural and
anthropogenic influences. Ecological services are to some extent dependent on both temporal
and spatial scales of the ecosystem, as described in Thorp et al. (2008). As a general relationship,
the greater the hydrogeomorphic complexity (and thus habitat complexity) of the FPZ, the
greater the biodiversity and functional complexity in that FPZ (Thorp et al. 2008). Moreover,
"The levels of ecosystem services provided by riverine landscapes are an increasing function of
the hydrogeomorphic complexity of the local functional process zone" (Thorp et al. 2010). For
example, the hypoxia zone off the coast of Louisiana results from anthropogenic changes in both
nitrogen inputs and nitrogen processing. The former is affected by the amount of nutrients
entering the river from upstream agricultural lands and non-point source pollution, while the
latter is affected by the vast levee system in the Mississippi River (especially the lower
Mississippi). In the latter case, the river's natural ability to decrease nutrient spiraling lengths
(i.e., the distance a nutrient atom must travel to complete one nutrient cycle from inorganic to
organic and back to inorganic form) and increase nitrogen processing are related to the amount
of lateral slackwater that is present. By understanding the original, current, and future FPZs for a
river, the ecosystem services can be evaluated under different scenarios of river complexity. This
also provides the empirical and conceptual bases for guiding processes to improve the
environmental quality of the river through activities such as river rehabilitation (Section A. 6) and
asset trading (Section A.5). Other ecosystems services (provisioning, regulating, supporting, and
cultural services; Limburg 2008) are also intimately affected by current FPZ complexity.
At present, and in support of the Ecosystem Research Program, the FPZs of the Neuse
River Basin are currently being delineated. This information will also be useful for the
characterization of ecosystem services basin-wide. As stated earlier in this document, because
FPZs differ substantially in hydrogeomorphic characteristics, FPZs are also likely to vary
significantly in community structure, ecosystem function, and response to nutrient loadings, and
thus will respond differently to efforts at river rehabilitation. For this project, the FPZs will be
delineated for the entire Neuse River Basin using 10-m DEM data supplemented by some
vertical LIDAR data. Our focus on the Neuse River provides an opportunity to test the ArcGIS
river delineation model on a river that is hydrogeomorphically distinct from both Kansas River
of the Great Plains and the Kanawha River of the mountainous East.
A.5 Asset Trading
The use of FPZs to improve asset trading was briefly discussed in Section 3.0. Two of the
components necessary for a fair basis of trading are: (i) a regional or national framework for
classifying rivers at the appropriate scale (including at least the valley-to-reach scale of FPZs);
-14-

-------
and (ii) an understanding of the link between river classes and resulting differences in ecosystem
structure and function for the river sections being compared. Thorp et al. (2008) provided a
framework for both classifying the relevant scale of rivers and for understanding the likely
ecological responses to different actions. However, many of the predicted ecological differences
between FPZs presented in Thorp et al. (2008) are hypothetical, because river typing using
hydrogeomorphic classification into FPZs is just beginning in the U.S. through a joint venture
between EPA-ORD (i.e., NERL-Cincinnati) and the Kansas Biological Survey at the University
of Kansas. It should be possible, however, to refine these hypotheses using current EPA
ecological data and develop general guidelines for differences in biodiversity patterns and
ecosystem services among FPZs to enhance the basis for fair asset trading.
A.6 River Rehabilitation
Many attributes are factored into decisions on river rehabilitation/restoration, some of
which were alluded to in Section 3.0. Rehabilitation can have many objectives, including
removal of dams, reconnecting the main channel with floodplains, improvement of channel bed
structure (inorganic and organic), enhancement of riparian/channel exchange, naturalization of
the flow regime, removal of exotic species, and the addition of formerly native species.
Decisions on many of these objectives would be improved by knowledge of the past, current, and
future FPZs likely in the affected area.
Some types of river rehabilitation are relatively straight-forward and focused on a
specific site and rehabilitation object. For example, dams on several rivers in the state of Maine
were considered for removal based on various socioeconomic, political, and environmental
reasons. In this case the environmental action involved a simple decision - removal or non-
removal - with no reasonable, intermediate position. The old Edwards Dam, the most
downstream dam on the Kennebec River, was selected for removal. Some obvious ecological
responses to the dam's removal were predicted (as discussed by Casper et al. 2006), such as
increased activity of migratory fish. Had the FPZ composition of the river been delineated,
however, it would have been possible to evaluate the ecological benefits likely to accrue from
dam removal (in terms of the past, present, and near-future nature of the FPZs) for this river and
others in Maine. Knowledge of the future FPZs is not always clear when the past state is not
known. For example, removal of mill dams in the eastern U.S. did not immediately produce
meandering streams (as was expected) or the original anastomosing channels characteristic of the
region (which was initially unexpected); these both may develop over time (Walter and Merritts
2008).
More complicated decisions in river rehabilitation involve: (i) how far to set back a levee
in order to develop favorable cost/benefit ratios; and (ii) where to locate controllable breaches in
a levee to connect with wetlands, how many should be present, and how they should be operated
(e.g., amount of flow and the frequency, length, and seasonality of connection). These decisions
involve a balance of costs (e.g., construction, operation, and purchase of land) and returns from
ecosystem services. Clearly, the farther the levee is set back laterally and the more populated the
region, the more expensive the process; however, the ecological endpoints will not directly track
with economic costs (Thorp et al. 2010), but rather will depend on what type of FPZ develops in
the restored area (which is influenced strongly by what was there in the beginning). For example,
if a channelized section of river was originally characterized by a simple meandering FPZ with
-15-

-------
nominal lateral movement, then minimal set-back levees would be required as larger lateral areas
produced by increased levee set-backs would not produce many additional benefits. In contrast,
if the section of river was originally characterized by side channels, parallel channels, and
forested islands (a braided or possibly anastomosing FPZ), then, in terms of ecosystem services,
more extensive set-backs would be warranted. In addition to the differences among FPZs,
cost/benefit ratios for rehabilitation are also influenced in a non-linear fashion by the type of
ecosystem services highlighted (Thorp et al. 2010). To maximize ecological endpoints of
rehabilitation, the future FPZ for the area in question (which is strongly influenced by the
original FPZ present) needs to be known, as well as the time needed to obtain that state.
A.7 Watershed Management
"Management" is a hierarchical process in rivers, just as it is in most human endeavors,
and thus is subject to a variety of often vague definitions. When the meaning of this and other
critical terms are not specified, communication is confounded and environmental action is
impeded. River management may involve activities in the main channel, full riverscape (main
channel plus lateral slackwaters), or riverscape and floodplains. If the geographic coverage
extends laterally into the floodplains and surrounding valley basin or catchment, the process is
sometimes called watershed management. Except in very limited cases, river management
should most effectively encompass processes operating in both aquatic and terrestrial
components of the riverine landscape.
The appropriate hierarchical level of management activities depends on the human
concerns/targets and their spatial extent. At the highest spatial scale, river management involves
the entire drainage basin or watershed (i.e., from the highest-elevation first-order stream to the
lowest-elevation river section where the river enters a larger river, ocean, lake, or dry basin). In
contrast, river management may operate at a much lower scale, such as a small reach. The
management level is affected by river network size, political boundaries, and availability of
management funds.
EPA research should be conducted at multiple hydrogeomorphic levels and
spatiotemporal scales in order to address various tasks in the agency's environmental mission.
The hierarchical level and spatial scale of the research will vary among and within tasks;
generally speaking, the larger the spatial scale of the stressor (e.g., non-point source pollution
commonly operates at a greater scale than point-source pollution) or effect to be achieved, the
higher the appropriate focal level in the hierarchy. Moreover, the hierarchical level and spatial
scale appropriate for a study tend to increase with river size and are affected by the number of
FPZs per linear length of the river. For example, a focus at the valley-to-reach level is more
appropriate when FPZs change frequently along the length of the river than when they are
relatively constant in type over long distances.
To manage at any given level and scale, the controls exerted at the next higher level and
mechanisms operating at the next lower level are especially relevant to consider. For example, if
you wished to manage the hydrogeomorphic structure, the ecology, or the inputs to an entire
river, or at least a very large section within a state, the appropriate level for understanding
mechanisms would be the valley-to-reach level. In contrast, if you are concerned with more
spatially-limited stressors, you might focus instead at the reach level within an FPZ (realizing
that comparisons among reaches in different FPZs will require knowledge of the differential
-16-

-------
impacts of the various FPZs). Management at any hierarchical level is better guaranteed success
when the scale of field assessment activities carefully account for the hierarchical level of the
management target or research question. Thorp et al. (2008) provide a framework for different
aspects and scales of river management.
-17-

-------
Appendix B: Summary of FPZ Delineation Techniques
Functional process zones can be identified with either top-down techniques (e.g.
geospatial-based analyses) or bottom-up field methods (e.g., traditional on-site methods in
fluvial geomorphology). Because the former is sufficiently accurate and much less costly, labor
intensive, and time consuming, it is usually the recommended approach. General techniques for
delineating FPZs are described in Thorp et al. (2008).
The following 14 independent and dependent variables are used to derive FPZs:
geological conditions, mean annual rainfall, elevation, valley width, valley floor width, valley
side slope, down-valley slope, ratio of valley to valley floor width, wavelength of the channel
belt, sinuosity of the channel belt, width of the river channel belt, sinuosity of the river channel,
number of channels, and channel planform (Thorp et al. 2008). Other variables can be added, but
the data in this list has proven sufficient for FPZ delineation using an ArcGIS model. Using one
of several multivariate clustering techniques, a dendrogram of sites is produced with an
appropriate threshold level. Groups of sites can be ordinated with semi-strong, hybrid multi-
dimensional scaling and then tested to see whether they occurred by chance. To determine which
physical variables were most important in separating clusters, a principal axis correlation can be
conducted, followed by a Monte Carlo permutation test; only variables with an R2 greater than
the 75th percentile are recommended as being significant. Once the significant clusters are
identified and named using standard terminology for river channel types (e.g., braided FPZ) as
modified for other important contributors (e.g., upland or lowland), these clusters can then be
added to maps at specific coordinates to depict the spatial arrangement of FPZs along the river.
While FPZs can theoretically be delineated from the smallest headwater stream to the
largest great river, there can be practical limitations in headwater regions. The primary issue is
the ability of the investigator to determine the channel planform (number and type of channels)
by remote sensing, which can be limited by the presence of riparian cover (ecoregional and
seasonal) and the precision (i.e., pixel size) of available data. Consequently, FPZs can be
determined more easily for smaller streams in prairies than in forested areas, because the riparian
canopy of the forested areas tends to cause errors in the elevation data. Extensive coniferous
canopies cause more trouble than deciduous canopies, of course, because the former is closed
throughout the year. However, in most cases FPZs can be determined for at least third-order
streams in most ecoregions using LIDAR, 10-m DEM, or 30-m DEM data and down to first
order streams in many prairie watersheds. Where canopies obscure the channel form, on-site
reach data (derived by traditional bottom-up approaches) can be used profitably to delineate the
FPZ. The advantage of the more precise data decreases proportionately with stream size, while
the large computer processing demands stay the same. Hence, LIDAR is useful in very small
streams, while 30-m DEM data may be sufficient for most other stream sizes.
Analyzing FPZs requires a moderately-fast, memory-rich computer (especially if using
LIDAR data) and the appropriate software to download the necessary data and extract the needed
variables. The geospatial data are combined with the other variables in the model to produce the
needed clusters for FPZ delineation. Some knowledge of fluvial geomorphology is needed, but
this is minimal compared to the ability to process the DEM or LIDAR data and analyze the data
statistically. Efforts are currently underway to allow data processing and analysis to be
conducted with minimal time and effort by the user through semi-automation of the process.
-18-

-------
References
Abell, R.A., D.M. Olson, E. Dinerstein, P.T. Hurley, J.T. Diggs, W. Eichbaum, S. Walters, W.
Wettengel, T. Allnutt, C.J. Loucks, and P. Hedao. 2000. Freshwater Ecoregions of North
America: A Conservation Assessment. Washington, D.C.: Island Press. 368 pp.
Benda, L., N.L. Poff, D. Miller, T. Dunne, G. Reeves, G. Pess, and M. Pollock. 2004. The
network dynamics hypothesis: how channel networks structure riverine habitats. BioScience
54(5): 413-427.
Benke, A.C. and C.E. Cushing (eds.). 2005. Rivers of North America. Burlington, MA: Elsevier
Academic Press. 1168 pp.
Casper, A.F., J.H. Thorp, S.P. Davies, and D.L. Courtemanch. 2006. Ecological responses of
zoobenthos to dam removal on the Kennebec River, Maine, USA. Archiv fur Hydrobiologie
(Large Rivers Supplement) 16(4): 541-555.
Frissell, C.A., W.J. Liss, C.E. Warren, and M.D. Hurley. 1986. A hierarchical framework for
stream habitat classification: viewing streams in a watershed context. Environmental
Management 10(2): 199-214.
Junk, W.J., P.B. Bayley, and R.E. Sparks. 1989. The flood-pulse concept in river-floodplain
systems. Canadian Special Publications of Fisheries and Aquatic Sciences 106: 110-127.
Junk, W.J. and K.M. Wantzen. 2004. The flood pulse concept: new aspects, approaches, and
applications - an update. In: Welcomme, R.L. and T. Petr (eds.), Proceedings of the Second
International Symposium on the Management of Large Rivers for Fisheries, Volume 2. Food
and Agriculture Organization Regional Office for Asia and the Pacific, Bangkok, Thailand.
RAP Publication 2004/17, pp. 117-149.
Limburg, K.E. 2009. Aquatic ecosystem services. In: G.E. Likens (ed.), Encyclopedia of Inland
Waters. Oxford: Academic Press, pp. 25-30.
Montgomery, D.R. 1999. Process domains and the river continuum. Journal of the American
Water Resources Association 3 5 (2): 3 97-410.
Poole, G.C. 2002. Fluvial landscape ecology: addressing uniqueness within the river
discontinuum. Freshwater Biology 47(4): 641-660.
Ricketts, T.H., E. Dinerstein, D.M. Olson, C.J. Loucks, W. Eichbaum, D. DellaSala, K.
Kavanaugh, P. Hedao, P.T. Hurley, K.M. Carney, R. Abell, and S. Walters. 1999. Terrestrial
Ecoregions of North America: A Conservation Assessment. Washington, D.C.: Island Press.
508 pp.
Rosgen, D.L. 1994. A classification of natural rivers. Catena 22: 169-199.
Rosgen, D.L. 1996. Applied River Morphology. Second edition. Pagosa Springs, CO: Wildland
Hydrology. 390 pp.
Rosgen, D. 2006. Watershed Assessment of River Stability and Sediment Supply (WARSSS). Fort
Collins, CO: Wildland Hydrology. 246 pp.
Thorns, M.C. and M. Parsons. 2002. Ecogeomorphology: an interdisciplinary approach to river
science. In: Dyer, F.J., M.C. Thorns, and J.M. Olley (eds.), Structure Function and
Management Implications of Fluvial Sedimentary Systems. The International Association of
Hydrological Sciences, Wallingford, UK. Publication 276, pp. 113-119.
Thorns, M.C. and M. Parsons. 2003. Identifying spatial and temporal patterns in the hydrological
character of the Condamine-Balonne River, Australia, using multivariate statistics. River
-19-

-------
Research and Applications 19(5-6): 443-457.
Thorp, J.H., J.E. Flotemersch, M.D. Delong, A.F. Casper, M.C. Thorns, F. Ballantyne, B.S.
Williams, B.J. O'Neill, C.S. Haase. 2010. Linking Ecosystem Services, Rehabilitation, and
River Hydrogeomorphology. BioScience 59(1): 67-74.
Thorp, J.H., M.C. Thorns, and M.D. Delong. 2006. The riverine ecosystem synthesis:
biocomplexity in river networks across space and time. River Research and Applications
22(2): 123-147.
Thorp, J.H., M.C. Thorns, and M.D. Delong. 2008. The Riverine Ecosystem Synthesis: Towards
Conceptual Cohesiveness in River Science. San Diego, CA: Elsevier Academic Press. 208
pp.
Walter, R.C. and D.J. Merritts. 2008. Natural streams and the legacy of water-powered mills.
Science 319(5861): 299-303.
-20-

-------
Office of
Research
and
Development
(8101R)
PRESORTED
STANDARD
POSTAGE & FEES
PAID
United States
Environmental Protection
Agency
If you do not wish to receive these reports CHECK HERE
~ ; detach, or copy this cover, and return to the address in
the upper left-hand corner.
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-11/089
Sept. 2010
www.epa.gov
V
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free

-------