EPA/600/R-03/081
Landscape and Watershed Influences on Wild Salmon
and Fish Assemblages in Oregon Coastal Streams
.-". V
. V
I"
P.J. Wigington, Jr., J.L. Ebersole, J.P. Baker, M.R. Church, J.E. Compton,
S.G. Leibowitz, D. White, and M.A. Cairns
August 26, 2003
National Health and Environmental Effects Research Laboratory-Western Ecology Division
Office of Research and Development
U.S. Environmental Protection Agency
200 SW 35th St.
Corvallis, OR 97333
-------
TABLE OF CONTENTS
Section Page
LIST OF TABLES v
LIST OF FIGURES vi
EXECUTIVE SUMMARY viii
1.0 INTRODUCTION 1
2.0 RATIONALE 4
3.0 RESEARCH QUESTIONS AND APPROACH 10
3.1 RESEARCH QUESTIONS AND CONCEPTUAL FRAMEWORK 10
3.2 RESEARCH APPROACH AND MAJOR COMPONENTS 14
3.3 PARTNERSHIPS WITH OTHER AGENCIES 16
4.0 INTEGRATED WATERSHED STUDY 19
4.1 APPROACH 21
4.2 SITE CHARACTERISTICS 22
4.3 FIELD STUDY DESIGN 26
4.3.1 Seasonal Distribution Surveys 26
4.3.2 Coho Salmon Growth and Survival 28
4.3.3 Physical and Chemical Habitat Characterization 29
4.3.4 Analyses 31
4.4 FUTURE DIRECTIONS 32
5.0 BROAD-SCALE ANALYSES 34
5.1 EMAP AND RELATED DATA 35
5.2 ODFW JUVENILE COHO SALMON SURVEYS 36
5.3 ODFW LIFE-CYCLE WATERSHEDS 38
5.4 STREAM FLOW 39
5.5 FUTURE DIRECTIONS 44
6.0 ROLE OF NUTRIENTS IN SALMON HABITAT 45
6.1 BACKGROUND 45
in
-------
6.2 APPROACH 46
6.2.1 Question 1. What is the relative importance of watershed-derived
versus marine-derived nutrients to fish nutrition? 47
6.2.2 Question 2. What are the major processes and landscape factors, both
natural and anthropogenic, which control spatial patterns and
concentrations of nutrients? 49
6.3 FUTURE DIRECTIONS 54
7.0 FISH MODELING 56
7.1 RATIONALE FOR MODELING APPROACH 56
7.2 STREAM NETWORK 57
7.2.1 Representation of the Stream Network 58
7.2.2 Environmental Properties of Segments 59
7.3 COHO SALMON MODEL 60
7.3.1 Approach 61
7.3.2 Model Evaluations and Experiments 65
7.3.3 Future Directions 66
7.4 SIMULATION OF FISH ASSEMBLAGES 67
7.4.1 Research Questions 68
7.4.2 Representation of Fish Species and Sub-populations 69
7.4.3 Simulation of Ecological Processes 72
7.4.4 Model Evaluations 74
7.4.5 Future Directions 75
8.0 PROJECT INTEGRATION 77
9.0 NOTICE 82
10.0 LITERATURE CITED 83
IV
-------
LIST OF TABLES
Table Page
3.1 Project Investigators: expertise and major research area 13
4.1 Characteristics of candidate integrated watersheds 23
4.2 Objectives and tasks for the integrated watershed study 27
6.1 Questions, goals and methods for the nutrient studies 48
-------
LIST OF FIGURES
Figure Page
3.1 Oregon Coast showing (a) research project overall study area,
(b) ODFW Life-Cycles Study Watersheds, (c) ODFW Juvenile Survey
sites, (d) EMAP/REMAP sites 11
3.2 Conceptual framework for the project. Hexagons are human-related
factors that affect fish. Solid lines indicate the primary linkages to be
addressed in the project 12
4.1 Distribution and relative abundance of coho spawners 2001-2002 in the
Winchester Creek and West Fork Smith River watersheds. Counts based
upon sighting of live fish and carcasses, thus do not account for loss or
movement offish between survey intervals 25
5.1 Stream gaging stations in coastal Oregon with historic stream flow data 41
5.2 Current stream gaging stations in coastal Oregon 41
5.3 Stream flow (discharge/watershed area) in six coastal Oregon rivers,
December 1, 1965-June 30, 1966 42
5.4 Estimated and measured stream flow of Smith River during
October 1, 1965 - September 30, 1966 43
6.1 Watershed N export as a function of broadleaf plus mixed (conifer-
broadleaf) cover, weighted by the slope coefficient for both cover types.
data are for 27 streams in the Salmon River basin in 2000.
(Compton et al., in press) 51
7.1 Schematic representation on interaction between stream network and
fish models. HC, is an array representing the habitat characteristics of
Stream segment /. Upland effects on HC, may be included in later years 58
7.2 Schematic representation of habitat-based life cycle model developed for
coho salmon (Nichelson and Lawson, 1998) 62
7.3 Schematic representation of assemblage model which will predict
presence/absence of native fish sub-populations by stream reach 68
8.1 Diagrammatic representation of project goal and research questions 77
VI
-------
EXECUTIVE SUMMARY
In the Pacific Northwest (PNW), many populations of wild anadromous salmonids are in
serious decline. Landscape change, water pollution, introduced predators, fishing, hydropower
development, hatcheries, disadvantageous ocean conditions, and other factors have led to the
extinction or decline and listing of many stocks under the Endangered Species Act.
In recent years, the U.S. Environmental Protection Agency (EPA) has become
increasingly involved in regulatory and policy issues related to habitat alteration. The Clean
Water Act has a goal to restore and maintain the physical, chemical, and biological integrity of
the Nation's waters, including the protection and propagation offish, shellfish, and wildlife.
Habitat alteration is a common cause for failures of aquatic systems to meet designated uses as
required by the Clean Water Act, and addressing these failures increasingly requires ameliorating
the cumulative impacts of diffuse stressors including nutrient loading, sedimentation, and altered
hydrologic regime. As required by the Endangered Species Act. EPA is being asked to
participate in interagency species protection and restoration efforts where habitat issues play a
key role.
Because of the national policy significance of Pacific salmon population declines, EPA's
National Health and Environmental Effects Research Laboratory (NHEERL) has assigned the
Western Ecology Division the responsibility of conducting habitat-related research that will
contribute to overall interagency efforts to restore viable populations of wild salmon and other
native fishes in the Pacific Northwest. This research is part of a larger nationwide NHEERL
research program to provide the scientific basis for assessing the role of essential habitat in
maintaining healthy populations of fish, shellfish, and wildlife and the ecosystems on which they
depend. For the research project described herein, our overall goal is:
To quantify the influence of human and natural disturbances at landscape and watershed
scales on salmon populations and native fish assemblages in Oregon coastal streams.
We have chosen to focus our research in Oregon coastal streams for a number of reasons.
First, salmon populations in Oregon coastal drainages have experienced declines similar to those
experienced in watersheds throughout the Pacific Northwest. Coho salmon (Oncorhynchus
kisutch) is the most notable salmonid species in coastal Oregon that has been listed as threatened
VI11
-------
8.2 Summary of contributions of research components to project research
questions 78
8.3 Project deliverables (annual performance measures - APMs) under
the NHEERL Aquatic Stressor Research effort 79
8.4 Timing of major research activities annual performance measures 80
8.5 Contribution of project components to annual performance measures
(APMs - See Figure 8.3) 81
vn
-------
by the National Marine Fisheries Service (NMFS), and the state of Oregon has developed a
major plan for the restoration of coastal coho salmon. Secondly, there appears to be good
potential in coastal streams, such as those in Oregon, to restore viable salmon populations.
Coastal streams are generally free flowing, with few dams and reservoirs. Therefore, past and
present watershed land use activities, fish harvest, and hatchery operations are the major human
influences on salmon populations.
This plan describes salmon and native fish research that we will conduct over a five-year
period. We have organized our research effort to address three major research questions.
1. How do coho salmon and other fish use the stream network during their freshwater
lifecycle? How important is the interplay among fish distributions, movement patterns,
and spatial - temporal patterns of habitat quality in sustaining coho populations and
native fish assemblages?
2. What roles do nutrients, temperature, and flow play, relative to physical habitat, in
determining coho salmon freshwater survival and growth? How do these factors
influence fish assemblage structure?
3. How does human land use interact with natural processes at watershed to landscape
scales to affect the long-term sustainability of coho salmon populations and native fish
assemblages in Oregon coastal streams?
The research described in this plan emphasizes work that will be conducted during the first two
years of a five-year research effort. After the first two years, we will evaluate our results and
refine both our research questions and our approach to addressing these questions.
Our major endpoints of concern are the long-term viability of coho salmon populations
and native fish assemblages in Oregon coastal streams. We also have adopted overall stream
habitat integrity as an endpoint. We selected coho salmon because of their historic abundance in
coastal streams, because of the listing of coastal coho stocks under the Endangered Species Act,
and because of the availability of coho monitoring data from the Oregon Department of Fish and
Wildlife (ODFW). We will expend somewhat greater effort on coho salmon than other fish,
reflecting the policy emphasis on salmon as well as the importance of salmon to aquatic
ecosystems in the region. However, we believe it is important to simultaneously consider
potential changes in the entire fish assemblage. Analyses of coho salmon response will address
population-level metrics (e.g., smolts produced per returning adult) and mechanisms of
population response. Fish assemblage metrics will be principally species richness and
IX
-------
assemblage composition (species presence/absence), and will be evaluated with less
mechanistic detail than for coho salmon.
Many factors affect salmon and other fish. Our focus is linking human land use to in-
stream habitat conditions to fish response. To interpret field data, we must account for the
influence of ocean/estuary conditions, harvest, and hatcheries, although these factors are not the
main thrust of our research. Four aspects of in-stream habitat are of interest: physical structure,
stream flow, temperature, and nutrients/productivity. Previous research in coastal Oregon has
demonstrated the importance of physical structure (e.g., large woody debris, pool depth, off-
channel habitats) to salmon populations. Our emphasis, therefore, is on the relative roles of
stream flow, temperature, and nutrients, and their interactive effects with physical structure.
These areas of emphasis reflect both research gaps as well as the expertise of project scientists.
In the first years of the project, the majority of our effort will deal with habitat-fish
linkages, to ensure that subsequent landscape-habitat research addresses the habitat attributes
most important to fish. Ultimately, our goal is to identify how human land use affects
biologically important habitat attributes, against a backdrop of natural landscape and watershed
dynamics.
Stream ecosystems in coastal Oregon are temporally dynamic and spatially diverse, and
these spatial and temporal variations may be important determinants offish response. They also
are major themes in our research. We are interested particularly in how spatial patterns of habitat
condition within stream networks and across watersheds affect salmon survival and growth and
fish assemblage structure, and the influence of seasonal, annual, and longer-term dynamics of
habitat and landscape change on the long-term viability of salmon and other fish in coastal
streams.
Thus, three main features distinguish our research project: (1) focus on the interplay
between fish life history strategies and spatial and temporal variations in habitat quality within
watersheds and across the landscape; (2) consideration of stream flow, temperature, and nutrients
as additional habitat factors of potential importance; and (3) evaluation of effects not just on
salmon populations but the entire fish assemblage in Oregon coastal streams. Our major research
questions reflect these areas of emphasis.
We will address our three research questions through a combination of (1) an integrated
watershed study, (2) broader scale analyses, and (3) simulation modeling offish responses.
-------
The largest project component will be the integrated watershed study in which all aspects
of the project will be addressed within a given watershed: coho salmon and native fish
assemblages; fish responses to habitat and habitat responses to landscape/watershed processes;
and all four habitat attributes of interest (physical structure, stream flow, temperature, and
nutrients). Through intensive field sampling, we will characterize spatial and seasonal patterns in
habitat quality within each watershed and relate these patterns to spatial and seasonal patterns in
fish assemblages and the abundance, movement, survival, and growth of juvenile coho salmon.
Biologically important habitat attributes will then be related to landscape/watershed processes
and historical records of human land use. In 2002 and 2003, we propose to concentrate our
efforts in a single watershed, West Fork of the Smith River, with limited sampling in Winchester
Creek watershed. Both are life-cycle watersheds that ODFW monitors to quantify numbers of
returning coho adults and outmigrating smolts. We hope eventually to expand the project,
contingent on funding, to encompass 3-4 watersheds. Watershed selection criteria include (1)
existing monitoring and ongoing studies (in particular, smolt trap monitoring), (2) diversity of
habitat types and patterns within the watershed (lowland as well as upland habitats), and (3)
covering the range of watershed characteristics important in coastal Oregon (in particular,
geology and land use).
Broader scale analyses of among-watershed patterns will rely principally on existing data
or data being collected by others, with limited supplemental data collected in this project.
Multicollinearity among potential causal factors and substantial unexplained "noise" are likely to
make it difficult to distinguish individual causal relationships based solely on large-scale
correlative analyses. Such analyses, however, still provide an important check on the consistency
of mechanisms observed in the integrated watershed study across the larger region. They can
also help generate hypotheses for testing in the integrated watershed study. Broad-scale analyses
to relate landscape features to regional patterns of salmon habitat quality and juvenile salmon
abundance are already underway for coastal Oregon as part of the Coastal Landscape Analysis
and Modeling Study (CLAMS), a cooperative venture of the USDA Forest Service, Oregon State
University, and the Oregon Department of Forestry. We will coordinate with CLAMS to
compare their results with data we collect in the integrated watershed study. Broad-scale
analyses that we will conduct include: (1) variations in nutrient concentrations as an additional
factor influencing regional patterns in juvenile salmon abundance, by adding measurements of
XI
-------
nutrients to ODFW coast-wide juvenile salmon surveys and expanding CLAMS analyses to
include nutrients; (2) associations between fish assemblage structure, in-stream habitat, and
landscape features using data collected as part of EPA's Environmental Monitoring and
Assessment Program (EMAP) and related studies; (3) empirical models to predict stream flow,
derived from existing flow records with limited supplemental data collection; and (4)
comparisons of freshwater coho survival among ODFW life-cycle watersheds and across years,
with supplemental habitat and watershed characterization in this project. All these analyses will
be conducted collaboratively with the scientists responsible for original data collection.
As part of previous EPA research, field studies began in 2000 to assess watershed
processes controlling stream nutrient concentrations in the Salmon River watershed. We propose
an additional 1-2 years of sampling in this watershed to complete these studies within this
project. Data from the Salmon River will then be used in conjunction with the integrated
watershed study and broad-scale analyses to identify the watershed characteristics that control
nutrient loadings and the relative importance of watershed- and marine-derived nutrient sources.
Data from the integrated watershed study and broad-scale analyses will be used to
calibrate simulation models offish response. The fish simulation models, in turn, will help
interpret and extrapolate results from the field studies to other watersheds and over longer time
frames. Simulation modeling is particularly useful for exploring the role of long-term dynamics
and interactions between life history strategies and spatial-temporal habitat patterns, issues that
cannot be easily addressed with field data alone because of the long time frame of response and
complexity of interacting factors. We will develop separate simulation models for coho salmon
populations and native fish assemblages. Ultimately, both models will be run using a common
set of scenarios to assess, for example, how management strategies targeted to restore coho
salmon are likely to affect the overall fish assemblage. Initially, modeling will deal only with
fish responses to in-stream habitat patterns. We hope eventually to combine these fish-habitat
models with model components dealing with landscape-habitat relationships, developed by
others or in later years of the project, to address Research Question 3.
A major consideration in designing this research project was to complement other related
research in coastal Oregon. As a new organization entering the salmon research arena, we
recognize that it is essential to complement ongoing research and, to the extent possible, build
collaborative research partnerships. Five other organizations have major research efforts in
xn
-------
coastal Oregon: (1) ODFW, (2) CLAMS (3) EMAP conducted cooperatively by EPA and the
Oregon Department of Environmental Quality (DEQ), (4) NMFS, and (5) U.S. Geological
Survey (USGS). As a research team, our established strengths are in biogeochemistry,
hydrology, and landscape-scale analysis and modeling, and we are building our fisheries
expertise. By working closely with investigators involved in other research programs, we can
best apply our expertise to policy-relevant questions about salmon and aquatic integrity in
coastal streams.
xin
-------
1.0 INTRODUCTION
In the Pacific Northwest (PNW), many populations of wild anadromous salmonids are in
serious decline. Landscape change, water pollution, introduced predators, fishing, hydropower
development, hatcheries, disadvantageous ocean conditions, and other factors have lead to the
extinction or decline and listing of many stocks under the Endangered Species Act (Bauer and
Ralph, 1999; CENR, 2000). In January of 1999, the President of the United States initiated a new
partnership designed to reverse the dramatic declines in Pacific Coast salmon and to restore
salmon as an integral element of the region's ecology, culture and economy. A critically
important element of the Pacific Salmon Recovery Initiative is a commitment to strengthening
and coordinating Federal science to build an effective and lasting recovery of salmon. In the
report, From the Edge, Science to Support Restoration of Pacific Salmon, the Committee on
Environment and Natural Resources (CENR) identified science needs for Pacific salmon and
related species (CENR, 2000). They concluded that a comprehensive life-cycle approach that
addresses both variability in natural conditions and human impacts on physical, chemical, and
biological processes that affect salmon is needed to define relationships between habitat and
salmonid productivity. Furthermore, they recognized that restoration and recovery efforts must
proceed with due consideration of consequences for other native species. CENR recognized that
habitat for salmonids and all native aquatic species, and hence their populations, are strongly
influenced by watershed conditions at a landscape scale.
In recent years, the U.S. Environmental Protection Agency (EPA) has become
increasingly involved in regulatory and policy issues related to habitat alteration. The Clean
Water Act's primary goal is to restore and maintain the physical, chemical, and biological
integrity of the Nation's waters, including the protection and propagation offish, shellfish, and
wildlife. Although the chemical integrity of aquatic resources is much improved, physical and
biological integrity remains a concern. Habitat alteration is a common cause for failures of
aquatic systems to meet designated uses as required by the Clean Water Act, and addressing
these failures increasingly requires ameliorating the cumulative impacts of diffuse stressors
including nutrient loading, sedimentation, and altered hydrologic regime. An integrated approach
to environmental protection and to improving riverine condition is perhaps best provided by
habitat-based criteria. As required by the Endangered Species Act, EPA is being asked to
-------
participate in interagency species protection and restoration efforts where habitat issues play a
key role. Because one of EPA's core ecological regulatory authorities is the Clean Water Act, the
species endpoints for which habitat alteration is of greatest concern are aquatic species. By
focusing on aquatic ecosystems and habitats supporting species of combined ecological,
economic, and societal importance, EPA can advance broad environmental protection goals
while directly addressing issue-driven stakeholder concerns.
Because of the national policy significance of Pacific salmon population declines, EPA's
National Health and Environmental Effects Research Laboratory (NHEERL) has assigned the
Western Ecology Division the responsibility of conducting habitat-related research that will
contribute to overall interagency efforts to restore viable populations of wild salmon and other
native fishes in the Pacific Northwest. This research is part of a larger nationwide NHEERL
research program that is designed to provide the scientific basis for assessing the role of essential
habitat in maintaining healthy populations offish, shellfish, and wildlife and the ecosystems on
which they depend (NHEERL, 2002). For the research project described herein, our overall goal
is:
To quantify the influence of human and natural disturbances at landscape and watershed
scales on salmon populations and native fish assemblages in Oregon coastal streams.
We have chosen to focus our research in Oregon coastal streams for a number of reasons.
First, salmon populations in the Oregon coastal drainages have experienced declines similar to
those experienced throughout the Pacific Northwest. Coho salmon (Oncorhynchus kisutch) is the
most notable salmonid species that has been listed as threatened by the National Marine Fisheries
Service (NMFS), and the state of Oregon has developed a major plan for the restoration of
coastal coho salmon (Nicholas and O'Mealy, 2000). Secondly, there appears to be good potential
in coastal streams, such as those in Oregon, to restore viable salmon populations. Coastal streams
are generally free flowing, with few dams and reservoirs. Therefore, past and present watershed
land use activities, fish harvest and hatchery operations are the major human influences on
salmon populations. In contrast, the goal of restoring wild salmon populations in the Columbia
River basin appears to be much more difficult because of the added influence of numerous dams
and competing societal uses for water. There is currently much less investment in salmon
research in Pacific Northwest coastal streams than in the Columbia River basin, but there are
-------
great opportunities for collaborative research with agencies and organizations including the
Oregon Department of Fish and Wildlife (ODFW), the USDA Forest Service (FS), the NMFS,
the Bureau of Land Management (BLM), Oregon State University (OSU), South Slough
National Estuary, private landowners such as Roseburg Resources, Inc., and others.
This plan describes salmon and native fish research that we will conduct over a five-year
period. In the sections that follow, we present the rationale for our research (Section 2) and our
overall research approach (Section 3). Sections 4-7 describe the major components of our
project, and Section 8 provides a synthesis of how results from this work collectively will
address our goal and research questions. The research described in this plan emphasizes work
that will be conducted during the first two years of a five-year research effort. After the first two
years, we will evaluate our results and refine both our research questions and our approach to
addressing these questions.
-------
2.0 RATIONALE
The temperate forest basins of coastal Oregon historically supported diverse and
abundant runs of anadromous Pacific salmon. Kostow (1997), for example, estimated pre-
harvest abundance of coho salmon along the Oregon coast north of Cape Blanco in 1900 was 1.7
million adults. Oregon coastal basins also supported spring and fall run chinook salmon
(Oncorhynchus tshawytschd), steelhead trout (Oncorhynchus mykiss), chum salmon
(Oncorhynchus keta), and Pacific lamprey (Lampetra tridentata) in addition to resident
salmonids, cyprinids, cottids and catostomids. These species used a broad array of aquatic
habitats within Oregon coastal basins, including headwater streams, river mainstems, floodplain
wetlands, and estuarine marshes. Not all habitats and not all coastal basins, however, were
equally productive at any give time. Unpredictable natural disturbances within coastal basins
such as flooding, fire, and landslides created a dynamic environment (Reeves et al., 1995) that
has been likened to a mosaic of suitable habitats shifting in space and time across the landscape
over decades and centuries. In the near shore marine environment, climatic regimes that shift
over decades caused cycles of marine productivity due to changes in ocean currents and
upwelling conditions (Beamish et al., 2000). The impacts of such dramatic variations can cause
local extinctions offish populations. Yet fish, especially salmon, in the PNW have not only
survived but have flourished historically despite large-scale natural disturbances and periodic
local extinctions.
Salmon have colonized and persisted in geomorphically-active and hydrologically
dynamic PNW rivers by successfully employing a variety of life history strategies including high
fecundity, asynchronous spawning timing, a low but persistent rate of adult straying, and various
patterns of freshwater habitat rearing duration (Groot and Margolis, 1991). Such life history
diversity, in conjunction with high fidelity of salmon for spawning locations and the strong
selective pressures of the coastal environment, has allowed continual adaptation of localized
populations to the particular environments of coastal basins, resulting in population-level
genotypic and phenotypic differences among salmonid stocks (Healey and Prince, 1995). In
addition to life history diversity, there are a number of ecological concepts that may help explain
how populations can be maintained in fluctuating or marginal habitats; e.g., source-sink
dynamics (Pulliam, 1988), metapopulation dynamics (Levins, 1970), and the concept of refugia
-------
(Brown and Lomolino, 1998). All these concepts view dispersal as an important mechanism for
recolonizing or maintaining local populations, and thus a characteristic that has allowed fish to
adapt to dynamic environments. Cooper and Mangel (1999) and Young (1999) suggest that
metapopulation dynamics may be important for salmonids, and Sedell et al. (1990) examined the
role of refugia in the recovery of PNW fish communities from disturbance.
Although fish may have been adapted to the PNW environment historically, over the last
century this natural disturbance regime has been significantly modified by chronic human
impacts, especially fish harvesting and land use changes. These human impacts have affected
significantly almost all of the region's anadromous and native fish, and nearly decimated some
populations - most notably coho salmon, which have declined precipitously during the last
century to approximately 10% of historic abundance (Kostow, 1997). This trend reflects regional
declines of anadromous salmonids attributed to extensive modification of aquatic habitats, high
rates of salmon harvest, deleterious genetic change to wild salmon populations, and shifts in
ocean productivity (Weitcamp et al., 1995).
Freshwater habitats for native fishes along the PNW coast have been extensively affected
by human activities over the past century (Lichatowich, 1987; Reeves et al., 1997). Logging,
agriculture, road-building and urbanization have altered the supply, routing and storage of water,
wood, sediment, and nutrients. Effects of these changes on stream ecosystems include
simplification of stream channel structure through losses of large wood and channel straightening
(Mclntosh et al., 1994; Bilby and Bisson, 1998). Channel simplification can interact with high
and low discharge events to increase the magnitude of adverse impacts of discharge extremes on
aquatic species by reducing the size, frequency and accessibility of refugia from flood or drought
(Sedell et al., 1990). Simplification of channel and riparian structure is also associated with
increases in thermal extremes in streams due to losses of atmospheric buffering provided by
riparian canopy closure and stream-floodplain-hyporheic interactions (Poole and Berman, 2001).
Nutrient concentrations in Oregon coastal streams are influenced strongly by forest stand
composition, particularly red alder (Abuts nibra), and may have shifted due to changes in the
spatial extent of alder in coastal forests (Compton et al.. In review). In addition, reductions in
adult salmon escapement to natal streams due to harvest and population declines have altered
aquatic food webs and nutrient dynamics (Gresh et al., 2000). The cumulative effects of these
habitat alterations have been pervasive, touching every aspect of the ecology of Pacific salmon
-------
(Gregory and Bisson, 1997). Although human effects on aquatic habitats for anadromous fishes
may have less obvious effects during optimal marine conditions, it is during the most stressful
portions of the natural disturbance regime that human activities have the potential to have the
greatest adverse effects on populations and risk of extinction (Lawson, 1993).
The suitability of habitats for stream fishes is broadly defined by attributes of stream
flow, thermal regimes, cover from predators, as well as opportunities for feeding (Matthews,
1999). In Oregon coastal streams, stream flow can be critical to fish populations during winter
and spring peak flow periods when the considerable metabolic energy expenditures required to
maintain stream position might negatively affect growth and survival (Bustard and Narver, 1975;
McMahon and Hartman, 1989; Cunjak, 1996). Summer low flows can also impose limitations on
growth and survival for species such as coho salmon, particularly where deeper pool refugia are
not available (Kruzic et al, 2001). Water temperature as a habitat component is a key factor
affecting growth and survival of all aquatic organisms, regulating many physiological and
behavioral processes (McCullough, 1999; Sullivan et al., 2000). Temperature also may interact
with other factors to influence fish survival and growth. For example, as temperatures rise, the
amount of dissolved oxygen that is generally available to fish decreases, disease-related
mortality increases, and competition for limited food supplies increases. Availability of light,
nutrients, and other chemical constituents provide additional constraints on the suitability of a
particular habitat to support individuals of a particular species and their prey or forage base
(Matthews, 1999). The trophic basis for fish productivity of PNW streams may be strongly
influenced by relative contributions of both in-stream productivity (controlled by nutrients and
light) and allochthonous inputs (Bilby and Bisson, 1992). Opportunities for foraging strongly
constrain the suitability of habitats for stream fishes (Fausch, 1984) and are particularly relevant
to young-of-year fishes that must attain a minimum size to survive winter conditions (Shuter and
Post, 1990). Additionally, survival of coho salmon smolts has been shown to be strongly
dependent upon timing of ocean entry and size at time of ocean entry (Holtby et al., 1990). Both
timing and size of coho smolts leaving freshwater habitats are influenced by rearing conditions
within the stream network (Quinn and Peterson, 1996), thus linking ocean survival and
recruitment success back to the freshwater habitat during the early life history of the fish.
Although the quality and distribution of freshwater habitats are primary determinants of
salmonid recruitment success in the physically dynamic coastal landscape and have received the
-------
primary focus of research efforts, the fish assemblage of which salmon are members provides an
additional aspect of the environment of salmon (Warren, 1971; Li et al., 1987). Quantifying the
community context for salmonid life histories (e.g., the composition, distribution and trophic
status of freshwater fish assemblages within coastal drainages) will contribute to understanding
other potential biotic interactions, such as predation and competition among native fish species,
that may be particularly relevant in habitats altered by human activities (Fresh, 1997).
The multiple habitat requirements of stream fishes are often life-stage specific. For
example, the physical space required for an individual salmon changes with age and season, in
response to shifting bioenergetic requirements and behavioral needs (Keeley and Grant, 1995).
Because the diverse and often life-stage specific habitat needs of salmon and other species can
seldom be met in any single portion of a stream network, populations of salmon and other stream
fishes may use different portions of the stream network at different times during the life cycle
(Baxter, 2002). For example, several distinct life history types among chinook salmon (Reimers,
1973), steelhead trout (Everest, 1973), and coho salmon (Miller and Sadro, 2000) have been
described in coastal Oregon rivers. Variation in timing of spawning, and duration and location of
freshwater residency may effectively "spread the risk" of extinction across space and time,
buffering salmon populations from localized catastrophes (Weavers, 1993).
Many studies of stream fish-habitat relationships have focused on the site (reach) scale.
Reach-scale fish-habitat relationships may not necessarily provide sufficient knowledge
regarding important demographic and physical habitat processes operating at larger scales,
processes that ultimately drive productivity and persistence of stream fishes (Schlosser and
Angermeier, 1995; Labbe and Fausch, 2000). Persistence of stream fishes is dependent upon
adequate overlap in space and/or time of quality habitats for multiple life-stages (Mobrand et al.,
1997). In some small coastal streams, where spawning, early rearing and overwintering habitats
for juvenile salmonid occur in close spatial proximity with few barriers to movement between
them, dispersal characteristics (opportunities) of juvenile fishes may be less critical than the
volume or capacity of the available habitat or other limiting factors (Nickelson and Lawson,
1998). Where quality habitats for specific juvenile life stages are spatially discrete (e.g.,
Peterson, 1982), success of individuals, measured as growth and/or survival, or the persistence of
a population may be highly dependent upon the ability of some minimal number of individuals to
find and use suitable habitats. Human disturbances have the potential to affect long-term
-------
sustainability directly, by drastically reducing abundance. They may also affect long-term
sustainability indirectly by increasing the rate of local extinctions and by reducing the ability of
fish to disperse into and recolonize suitable habitats.
Recognition of widespread alteration of aquatic habitats has led to restoration efforts
aimed at reversing declines in Pacific salmon and other native fishes. Unfortunately, localized
habitat restoration efforts focusing on individual stream reaches, while sometimes successful at
addressing site-specific habitat needs, largely have failed to match the spatial scale and intensity
of habitat-shaping processes that have been altered by human activity. As a result, these well-
intentioned efforts have been ineffective at restoring stream habitats at the scales necessary to
sustain salmon populations (NRC, 1996). Subsequently, recovery strategies for imperiled aquatic
species have recognized increasingly the need to facilitate habitat restoration at watershed and
landscape scales (Bisson et al., 1997; Frissell and Ralph, 1998). Additionally, recovery strategies
will best be guided by understanding the biotic interactions among stream fish communities and
the trophic basis for stream productivity (Gresh et al., 2000). In coastal Oregon basins, this will
require quantification of the roles of watershed derived nutrients and marine derived nutrients
and how these spatially and temporally variable factors interact with physical habitats to
influence fish productivity and persistence.
Critical to understanding how fish populations respond to the combination of human and
natural stress regimes is knowledge of how fish use freshwater habitats (including movement
among various habitat types) during different life history stages (Roni et al., 2002). In
collaboration with our partner agencies and researchers (Section 3), we have identified three
distinct research needs pertaining to fish-habitat relationships in coastal basins. A primary
research need is to understand how fish use and move between habitats at different life history
stages and how they disperse and recolonize areas after local extinctions. Specifically, we need
to understand the interplay between life history diversity and spatial - temporal patterns in habitat
quality in sustaining coho populations and native fish assemblages. Secondly, the fitness benefits
or costs of individuals using and moving among habitat patches of varying quality must be
assessed. Such information would require understanding the relative roles of multiple,
interacting, habitat factors influencing fish fitness such as temperature, stream flow, nutrients,
and physical habitat. Data pertaining to fish growth rate, condition and survival as influenced by
temporally dynamic local habitat conditions and the costs of movement between habitats would
-------
be of particular value. Lastly, research is needed to quantify the interacting effects of human land
use and natural disturbances on the sustainability of a range offish populations. Although much
is known about the habitat requirements of salmonids, and the effects of land use on aquatic
habitats, predicting the effects of individual human activities on salmonids and entire fish
assemblages is daunting. This is because stream fishes integrate the influence of multiple habitat
factors at a variety of spatial and temporal scales, and the dynamic nature offish assemblages
and stream habitats provides a continually-shifting target for quantitative analysis. Modeling
responses of salmonids and fish assemblages to habitat change is needed to assess relationships
that cannot be effectively addressed in field studies.
The information gained from such research would contribute toward understanding
recovery potential and extinction risks offish populations in stressed environments. Such
knowledge could help direct the restoration of processes that enhance connectivity between
critical habitats, and could contribute to understanding life-stage and habitat-specific limiting
factors. Information is especially critical for coho salmon because of its listing under the
Endangered Species Act (ESA) and subsequent litigation. Research on other native species
comprising the fish assemblage of coastal stream networks is also needed because of potential
ESA concerns, the need to protect the biotic integrity of the Nation's waters under the Clean
Water Act, and perhaps most importantly because salmon exist within aquatic communities of
which they are integral members, interacting in complex ways.
-------
3.0 RESEARCH QUESTIONS AND APPROACH
3.1 RESEARCH QUESTIONS AND CONCEPTUAL FRAMEWORK
As noted in Section 1, the overall goal of the project is to quantify the influence of human
and natural disturbances at landscape and watershed scales on salmon populations and native fish
assemblages in Oregon coastal streams. More specifically, we will focus on the following
research questions, consistent with the research needs identified in Section 2:
1. How do coho salmon and other fish use the stream network during their freshwater
lifecycle? How important is the interplay among fish distributions, movement patterns,
and spatial - temporal patterns of habitat quality in sustaining coho populations and
native fish assemblages?
2. What roles do nutrients, temperature, and flow play, relative to physical habitat, in
determining coho salmon freshwater survival and growth? How do these factors
influence fish assemblage structure?
3. How does human land use interact with natural processes at watershed to landscape
scales to affect the long-term sustainability of coho salmon populations and native fish
assemblages in Oregon coastal streams?
Figure 3.1 defines the study area along the Oregon coast. We want to address the above
questions over the range of conditions within this region.
Our major endpoints of concern are the long-term viability of coho salmon populations
and native fish assemblages in Oregon coastal streams. We also have adopted overall stream
habitat integrity as an endpoint. We selected coho salmon because of their historic abundance in
coastal streams, because of the listing of coastal coho stocks under the Endangered Species Act,
and because of the availability of coho monitoring data from the ODFW (see Section 3.3). We
will expend somewhat greater effort on coho salmon than other fish, reflecting the policy
emphasis on salmon as well as importance of salmon to aquatic ecosystems in the region.
However, as noted in Section 2, we believe it is important to simultaneously consider potential
changes in the entire fish assemblage. Analyses of coho salmon response will address
population-level metrics (e.g., smolts produced per returning adult) and mechanisms of
population response. Fish assemblage metrics will be principally species richness and
10
-------
Freshwater Habitat
Study Area
H Cjtirs and (owns
— Rhtrs
(b)
N. Pk NchaJcm River
Oregon Department
of Fwh and Wildlife
life Cycle Watersheds
Mill Creek, Silra Rn*r
Mill Creek, Yaquina River
I'jMjilr f'rcck. Alw-j River
W. Fk. Smith
^X^n^:hettc^ Ciwk :
(c)
(d)
«',"•
.,*l-t-
* * *«
Oregon Ll!cpaTtmrnt of
Hsh and Wildlife
Juvenile Survey Si i ci
O Rearing only
^ Rearing and vpawning
Rearing and physical
hahitat
. Rearing, yawing.
and jihviical hibiui
EPA Sites
• 1LMAP
• REMAP
•r.
Figure 3.1 Oregon Coast showing (a) research project overall study area.
(lj) ODFW Life-Cycle Study \\ atersheds, (c) ODFW
Juvenile Suney sites, (d) EMAP/REMAP site§.
i:
-------
assemblage composition (species presence/absence), and will be evaluated with less mechanistic
detail than for coho salmon.
Many factors affect salmon and other fish. Our focus is linking human land use to in-
stream habitat conditions to fish response (Figure 3.2). To interpret field data, we must account
for the influence of ocean/estuary conditions, harvest, and hatcheries, although these factors are
not the main thrust of our research.
Natural
Landscape / Watershed
Aquatic Habitat
Physical Structure Flow
Temperature Nutrients
Fish Response
Coho Salmon
Native Assemblages
Ocean and
Estuary
Conditions
Figure 3.2. Conceptual framework for the project. Hexagons are human-related factors
that affect fish. Solid lines indicate the primary linkages to be
addressed in the project.
12
-------
Four aspects of in-stream habitat are of interest: physical structure, stream flow,
temperature, and nutrients/productivity. Previous research in coastal Oregon has demonstrated
the importance of physical structure (e.g., large woody debris, pool depth, off-channel habitats)
to salmon populations. Our emphasis, therefore, is on the relative roles of stream flow,
temperature, and nutrients, and their interactive effects with physical structure (Research
Question 2). These areas of emphasis reflect both research gaps as well as the expertise of
project scientists (Table 3.1).
Table 3.1. Project investigators: expertise and major research areas.
Project Investigator Expertise
Major Research Area
Jim Wigington
Joan Baker
Michael Cairns
Robbins Church
Jana Compton
Joe Ebersole
Steve Klein
Scott Leibowitz
Denis White
Hydrology
Fisheries biology
Water quality
Biogeochemistry
Biogeochemistry
Fisheries biology
Forestry
Landscape ecology
Biogeography
Project Leader; Stream flow and
temperature
Fish assemblage responses and modeling
Stream flow and temperature
Nutrients
Nutrients
Coho salmon responses
Historical land use
Coho salmon modeling
Fish assemblage modeling
In the first years of the project, the majority of our effort will deal with habitat-fish
linkages, to ensure that subsequent landscape-habitat research addresses the habitat attributes
most important to fish. Ultimately, our goal is to identify how human land use affects
biologically important habitat attributes, against a backdrop of natural landscape and watershed
dynamics (Research Question 3).
As discussed in Section 2, stream ecosystems in coastal Oregon are temporally dynamic
and spatially diverse, and these spatial and temporal variations may be important determinants of
fish response. They also are major themes in our research. We are interested particularly in how
13
-------
spatial patterns of habitat condition within stream networks and across watersheds affect salmon
survival and growth and fish assemblage structure (Research Question 1), and the influence of
seasonal, annual, and longer-term dynamics of habitat and landscape change on the long-term
viability of salmon and other fish in coastal streams (Research Question 3).
Thus, three main features distinguish our research project: (1) focus on the interplay
between fish life history strategies and spatial and temporal variations in habitat quality within
watersheds and across the landscape; (2) consideration of stream flow, temperature, and nutrients
as additional habitat factors of potential importance; and (3) evaluation of effects not just on
salmon populations but the entire fish assemblage in Oregon coastal streams. Our major research
questions reflect these areas of emphasis.
3.2 RESEARCH APPROACH AND MAJOR COMPONENTS
We will address the above research questions through a combination of (1) an integrated
watershed study, (2) broader scale analyses, and (3) simulation modeling offish responses. The
expertise of EPA scientists working on this project is summarized in Table 3.1. During fiscal
year 2002, EPA provided $ 491,500 to cover equipment, travel, field crew, and analytical
chemistry costs. A significant portion of the field work described in Section 4 will be
accomplished via a contract with the Dynamac Corporation. We anticipate that total funding will
increase somewhat during the course of the five year study period, but future funding levels have
not been established at this time.
The largest project component will be the integrated watershed study, described in
Section 4. All aspects of the project will be addressed within a given watershed: coho salmon
and native fish assemblages; fish responses to habitat and habitat responses to
landscape/watershed processes; and all four habitat attributes of interest (physical structure,
stream flow, temperature, and nutrients). Through intensive field sampling, we will characterize
spatial and seasonal patterns in habitat quality within each watershed and relate these patterns to
spatial and seasonal patterns in fish assemblages and the abundance, movement, survival, and
growth of juvenile coho salmon. Biologically important habitat attributes will then be related to
landscape/watershed processes and historical records of human land use. In 2002 and 2003, we
propose to concentrate our efforts in a single watershed, West Fork of the Smith River, with
14
-------
limited sampling in Winchester Creek watershed (Figure 3.1). Both are ODFW life-cycle
watersheds (see Section 3.3). We hope eventually to expand the project, contingent on funding,
to encompass 3-4 watersheds. Watershed selection criteria include (1) existing monitoring and
ongoing studies (in particular, smolt trap monitoring), (2) diversity of habitat types and patterns
within the watershed (lowland as well as upland habitats), and (3) covering the range of
watershed characteristics important in coastal Oregon (in particular, geology and land use).
Broader scale analyses of among-watershed patterns, described in Section 5, will rely
principally on existing data or data being collected by others, with limited supplemental data
collected in this project. Multicollinearity among potential causal factors and substantial
unexplained "noise" are likely to make it difficult to distinguish individual causal relationships
based solely on large-scale correlative analyses. Such analyses, however, still provide an
important check on the consistency of mechanisms observed in the integrated watershed study
across the larger region. They can also help generate hypotheses for testing in the integrated
watershed study. Broad-scale analyses to relate landscape features to regional patterns of salmon
habitat quality and juvenile salmon abundance are already underway for coastal Oregon as part
of the Coastal Landscape Analysis and Modeling Study (CLAMS; see Section 3.3). We will
coordinate with CLAMS to compare their results with data we collect in the integrated watershed
study. Broad-scale analyses that we will conduct include: (1) variations in nutrient
concentrations as an additional factor influencing regional patterns in juvenile salmon
abundance, by adding measurements of nutrients to ODFW coast-wide juvenile salmon surveys
and expanding CLAMS analyses to include nutrients; (2) associations between fish assemblage
structure, in-stream habitat, and landscape features using data collected as part of EPA's
Environmental Monitoring and Assessment Program (EMAP) and related studies; (3) empirical
models to predict stream flow, derived from existing flow records with limited supplemental data
collection; and (4) comparisons of freshwater coho survival among ODFW life-cycle watersheds
and across years, with supplemental habitat and watershed characterization in this project. All
these analyses will be conducted collaboratively with the scientists responsible for original data
collection, as described in Section 3.3.
As part of previous EPA research (Compton et al., In review), field studies began in 2000
to assess watershed processes controlling stream nutrient concentrations in the Salmon River
watershed. We propose an additional 1-2 years of sampling in this watershed to complete these
15
-------
studies within this project. Data from the Salmon River will then be used in conjunction with the
integrated watershed study and broad-scale analyses to identify the watershed characteristics that
control nutrient loadings and the relative importance of watershed- and marine-derived nutrient
sources. This supplemental nutrient sampling in the Salmon River watershed is described in
Section 6. Some fish sampling also will be conducted in the Salmon River watershed, but
unfortunately coho populations in this watershed are heavily hatchery-influenced, making it a
poor choice for a comprehensive, integrated watershed study of wild salmon.
Data from the integrated watershed study and broad-scale analyses will be used to
calibrate simulation models of fish response, described in Section 7. The fish simulation models,
in turn, will help interpret and extrapolate results from the field studies to other watersheds and
over longer time frames. Simulation modeling is useful particularly for exploring the role of
long-term dynamics and interactions between life history strategies and spatial-temporal habitat
patterns, issues that cannot be easily addressed with field data alone because of the long time
frame of response and complexity of interacting factors. We will develop separate simulation
models for coho salmon populations and native fish assemblages. Ultimately, both models will
be run using a common set of scenarios to assess, for example, how management strategies
targeted to restore coho salmon are likely to affect the overall fish assemblage. Initially,
modeling will deal only with fish responses to in-stream habitat patterns. We hope eventually to
combine these fish-habitat models with model components dealing with landscape-habitat
relationships, developed by others or in later years of the project, to address Research Question
3.
3.3 PARTNERSHIPS WITH OTHER AGENCIES
A major consideration in designing this research project was to complement other related
research in coastal Oregon. As a new organization entering the salmon research arena, we
recognize that it is essential to complement ongoing research and, to the extent possible, build
collaborative research partnerships. Five other organizations have major research efforts in
coastal Oregon: (1) ODFW, (2) CLAMS (cooperative venture of the USDA Forest Service,
Oregon State University and the Oregon Department of Foresty), (3) EPA's Environmental
Monitoring and Assessment Program (EMAP), conducted cooperatively with the Oregon
16
-------
Department of Environmental Quality (DEQ), (4) NMFS, and (5) U.S. Geological Survey
(USGS). As a research team, our established strengths are in biogeochemistry, hydrology, and
landscape-scale analysis and modeling, and we are building our fisheries expertise (Table 3.1;
Appendix 1). By working closely with investigators involved in other research programs, we can
best apply our expertise to policy-relevant questions about salmon and aquatic integrity in
coastal streams. During January 2002, we met with ODFW, CLAMS, NMFS, and USGS
researchers to discuss common research goals, potential collaboration, research needs, and
approaches to address those needs. The integrated watershed study design proposed in Section 4
is a direct outgrowth of that meeting.
ODFW is charged with evaluating the status and trends in Oregon salmon stocks. They
collect data for coho, chinook, and chum salmon, and steelhead, although the most
comprehensive monitoring in coastal Oregon is for coho salmon. Numbers of spawning coho
adults are estimated for about 120 randomly selected stream segments per year in each of the five
Gene Conservation Areas (GSA) along the coast (Jacobs et al., 2001); juvenile abundance at 50
sites/year/GSA (Rodgers, 2000; 2001; Figure 3.1), and stream habitat conditions at 45
sites/year/GSA (Moore et al., 1997). Monitoring of both coho adult returns and smolt
escapement at eight life-cycle watersheds provide aggregate estimates of freshwater and marine
survival (Solazzi et al., 2001; Figure 3.1). Other activities include long-term records of coho
spawner abundance at 46 hand-picked sites monitored since about 1950 and comprehensive
basin-wide surveys of habitat conditions in selected basins, including all life-cycle watersheds.
These studies provide a rich database on salmon and salmon habitat in coastal Oregon. We
designed our project to work with ODFW, to supplement their basic monitoring and help
interpret observed patterns and trends.
CLAMS' goal is to evaluate the ecological, economic, and social consequences of forest
policies and practices across multiple ownerships in coastal Oregon (Spies et al., in press). Major
components include comprehensive GIS-based characterizations of vegetation (Cohen et al.,
2001; Ohmann and Gregory, in press) and physical conditions for the entire Coast Range;
modeling to simulate vegetation changes 100-200 years into the future in response to different
forest management strategies (Bettinger et al., In review); and modeling the likely effects of
these landscape changes on salmon and salmon habitat (Burnett, 2001; Burnett et al.,
unpublished). Aquatic response models are being developed based primarily on statistical
-------
associations between landscape features and ODFW data on stream habitat and salmon status
(spawner and/or juvenile abundance). Eventually, the statistical models will be applied to project
future changes in salmon habitat and status in all streams throughout the Coast Range in
response to simulated changes in vegetation. We can take advantage of CLAMS' landscape
characterizations and, at the same time, hopefully provide information that helps evaluate and
extend their statistical models of aquatic response.
EMAP's goal is to quantify the ecological condition of U.S. resources (Messer et al.,
1991). As part of this effort, 57 randomly selected, wadeable stream reaches in coastal Oregon
were surveyed to assess the composition of fish and benthic invertebrate communities, water
quality, and physical habitat condition during summer 1994 and 1995 (Merger and Hayslip,
2000). Using comparable survey techniques, this initial sample has been supplemented by
continued monitoring of stream biointegrity by DEQ, enhancements of the sampling grid within
selected watersheds (Rose, 2000), and additional coastal sites included in statewide and western
U.S. EMAP assessments. In total, EMAP-type data are now available for over 150 stream
reaches in coastal Oregon (Figure 3.1). We propose to use these data to develop models offish
assemblage responses to habitat and landscape change that would complement the salmon-
focused modeling being conducted by CLAMS (see Sections 5 and 7).
NMFS and USGS are conducting a diversity of projects in coastal Oregon. Key
opportunities for collaboration include: (1) models of coho salmon population dynamics
developed jointly by ODFW and NMFS (Nickelson and Lawson, 1998; see Section 7); (2)
juvenile coho survival and growth in restored estuarine wetlands in Salmon River Estuary
(Daniel Bottom, NMFS, personal communication; see Section 6); (3) fish and benthic
invertebrate responses to boulder weir restoration projects, including field sampling proposed in
the West Fork Smith River watershed (Phil Roni, NMFS, personal communication); and (4)
population dynamics and movement of non-anadromous cutthroat trout (Robert Gresswell,
USGS, personal communication).
In future years, we will continue to build collaborative relationships with agencies and
organizations involved in salmon and fish assemblage research and management in Oregon
coastal streams. We are especially interested in expanding the research project via joint, multi-
agency research proposals in future years.
18
-------
4.0 INTEGRATED WATERSHED STUDY
The goal of the integrated watershed study is to contribute to understanding the processes
linking native stream fish diversity, productivity and persistence to the watershed environment.
Although relationships of salmonids to stream habitats have been relatively well-studied,
particularly for coho salmon (e.g., Chapman, 1962; Peterson, 1982; Hartman and Brown, 1987;
Holtby, 1988; Holtby et al., 1989; Reeves et al., 1989; Swales and Levings, 1989; Sandercock,
1991; Nickelson et al., 1992; Nielsen, 1992; Quinn and Peterson, 1996; Cederholm et al., 1997;
Bilby et al., 1998; Nickelson and Lawson, 1998; Solazzi et al., 2000; Bell et al., 2001), linking
responses of salmonids and other native fishes to watershed-scale habitat attributes has been
challenging. For example, studies examining productivities of distinct rearing habitats for stream
salmonids must often assume that habitats are fully accessible to fish regardless of their spatial
location relative to habitats for preceding life history stages (Beechie et al., 1994). Where
individuals have high mobility or where spawning and rearing habitats are in close proximity,
spatial relationships may be relatively unimportant in influencing productivity. But where
habitats for distinct life history stages are spatially and/or temporally disjunct, or where fish
mobility is limited, spatial habitat relationships can be a strong factor influencing watershed-
scale population dynamics (Kocik and Ferreri, 1998). In addition, the importance of spatial and
temporal heterogeneity of watershed physio-chemical features to fish productivity is well
recognized but has been difficult to quantify at watershed or large scales (Poff and Ward, 1990;
Reeves et al., 1995). Increased understanding of the population dynamics of mobile stream fish
species in heterogeneous landscapes will require spatially and temporally extensive evaluations
of stream fish distributions and fitness in relation to the habitat template. The integrated
watershed study component of this research is an attempt to characterize spatial and temporal
patterns offish distribution and fitness within entire coastal third to fifth-order drainage
networks. The specific objective of the integrated watershed study is to examine interactions
between fish life history strategies and spatial/temporal patterns of habitat quality within a
watershed, relating seasonal patterns of distribution, growth, and survival of stream fishes to
watershed/stream habitat characteristics, including the availability of watershed-derived and
marine-derived nutrients, variation in stream flow, and stream temperatures.
19
-------
The major questions of this research, outlined previously in Section 3, are listed below
(Questions 1- 3). In addition, specific questions that will be addressed within the integrated
watershed study are included under major research questions 1 and 2 below. Specific sub-
questions for research question 3 will be developed in years 1-3 from the integrated watershed
study and modeling efforts.
1. How do coho salmon and other fish use the stream network during their freshwater
lifecycle? How important is the interplay among fish distributions, movement
patterns, and spatial - temporal patterns of habitat quality in sustaining coho
populations and native fish assemblages?
la. Do juvenile salmon move between habitat types seasonally in a manner that
increases their growth and survival potential?
Ib. Is the quantity of suitable overwinter habitat predominantly important to coho
salmon productivity, independent of the spatial distribution, distance, and
connectivity among habitats used during different seasons?
Ic. Do reach-level habitat features adequately explain patterns of native fish
richness and species composition, or are larger-scale network properties also
important?
Id. How do potential banners to dispersal (e.g., culverts, high-gradient stream
reaches, thermally-unsuitable reaches) influence seasonal distributional
patterns of native fishes?
2. What roles do nutrients, temperature, and flow play, relative to physical habitat, in
determining coho salmon freshwater survival and growth? How do these factors
influence fish assemblage structure?
2a. Do juvenile coho grow faster or survive better in stream reaches with higher
nutrient inputs? What is the relative contribution of watershed-derived vs.
marine-derived nutrients to coho salmon freshwater productivity?
2b. Are physical refugia (e.g., off-channel habitats in winter, deep pools in
summer) more critical to coho survival in streams with flashier hydrographs?
2c. Are stream temperatures in coastal streams sufficiently elevated, over a large
enough area, to be an important factor limiting coho salmon productivity?
2d. How do stream flow, temperature, and nutrient status interact with physical
habitat conditions to influence salmon productivity?
20
-------
2e. Which habitat aspects (temperature, stream flow, nutrients, physical habitat)
are most strongly associated with fish assemblage structure within coastal
streams?
3. How does human land use interact with natural processes at watershed to landscape
scales to affect the long-term sustainability of coho salmon populations and native
fish assemblages in Oregon coastal streams?
Efforts within the integrated watershed study will focus on Questions 1 and 2 during the
first two years of this study. We will emphasize collecting field data relevant to seasonal
distributional patterns of coho salmon and other stream fish species within individual
watersheds. We will also derive habitat-specific growth and survival estimates for coho salmon
particularly as they relate to the spatial proximity of summer rearing habitats and high quality
winter habitats. A preliminary hypothesis is that overwinter survival of coho salmon is a function
of spatial proximity of summer and winter rearing habitats, such that survival rates would be
expected to be higher for coho salmon moving lesser distances to access winter habitats.
Overwinter survival also is expected to be positively related to size (or condition factor) and
growth, which may in turn reflect both summer and winter rearing conditions (Quinn and
Peterson, 1996; Nickelson and Lawson, 1998). This research will examine quantitative
relationships between spatial habitat relationships and fish distribution, survival and growth. By
building upon existing knowledge and expanding investigations of fish-habitat relationships to
watershed scales, this research will contribute to understanding the relative historical importance,
and potential future capacity, of diverse salmon life histories and the suite of habitats essential
for the expression of those life histories within entire watersheds. The research also will examine
whether native fish assemblages respond to the same types of in-stream habitat attributes and
dynamics as salmon.
4.1 APPROACH
The emphasis during the first two years of this study will be directed toward
comprehensively characterizing habitat quality (physical structure, flow, temperature, nutrients)
for stream fishes within small coastal watersheds, including the inter-dependent and interactive
21
-------
effects of these factors as they relate to the distribution, growth and survival of stream fishes. In
2002 and 2003, we propose to concentrate our efforts in the West Fork of the Smith River, with
limited sampling in Winchester Creek watershed (Figure 3.1b). We selected the West Fork Smith
River and Winchester Creek as watersheds for the initial phase of this study because they
provide the best opportunity to collaborate with existing life history research being conducted by
the ODFW. These watersheds are used by ODFW to monitor trends in freshwater and marine
survival of coho salmon in the Oregon Coast Evolutionary Significant Unit (ESU; Solazzi et al.,
2001). Each watershed is equipped with a weir to capture returning adult coho salmon and
steelhead, and a smolt trap to capture emigrating smolts. These trapping efforts provide annual
estimates of both coho salmon spawning escapement and smolt production for the entire
watershed, and will provide a context for our more detailed life history investigations within
each watershed described below. In addition, operation of smolt traps near the mouth of each
watershed will allow the recapture offish tagged for growth and survival estimates that would
not be possible without the cooperation and assistance of ODFW.
4.2 SITE CHARACTERISTICS
The West Fork Smith River and Winchester Creek are located within the coastal
mountain ranges and coastal valleys of the Coast Range Ecoregion of Oregon (Omernik, 1987).
They share characteristics of climate and geology but differ significantly with regard to
topography, watershed area, stream habitat configuration, vegetation cover, fish assemblage
composition, and abundance and distribution of native salmonids (Table 4.1). The West Fork
Smith River is a 4-5th order watershed with relatively high topographic relief (elevation range 60
to 850 m) and drainage area of approximately 68 km2. Coho salmon escapement to West Fork
Smith River was estimated to be 1517 wild adults plus a small number of hatchery strays in
2001-2002. Spawners were distributed throughout accessible reaches of the mainstem and
tributaries, although spawner densities were highly variable among surveyed reaches (Figure
4.1). When last surveyed in 1999, channel substrates in the mainstem West Fork Smith River
were composed of 38% exposed bedrock (ODFW unpublished Aquatic Habitat Inventory Data).
The mainstem West Fork Smith River is apparently unable to retain gravels due to a lack of
channel structure and/or high transport capacity. These conditions are believed to reflect a legacy
22
-------
Table 4.1 Characteristics of candidate integrated watersheds.
Stream length
(mainstem, km)
Watershed area (km2 )
Stream order
Land ownership
(owner, %)
Geology
Elevation (m)
Winchester Creek
9
25
3
Coos County forest 76%
Private Non-Industrial
Forest 23%
Tyee sandstone
0-100
W. Fork Smith River
25
68
4-5
BLM 63 %
USFS 4 %
Private Industrial
Forest 30 %
Tyee sandstone
60-850
Land Use History
Logging, livestock grazing
Logging, splash-damming
Vegetation Cover (%)
Water
Open forest
Broadleaf
Mixed forest
Small-medium conifer
Large- very large conifer
Open nonforest
Woodlands
Coho smolts
1998
1999
2000
2001
2002
Smolt trap efficiency
Coho adults
1999-00
2000-01
2001-02
0
18
46
6
18
3
1
6
not sampled
2247 (after Feb 1)
3535 (after Feb 1)
5074 (after Feb 1)
600+ (as of May 1)
30%
(plus estuary seining)
40
5
302
0
8
21
26
30
14
0
1
22412
10866
14855
20091
7000+ (as of May
30%
(late winter and spring
264
538
1517
1)
only)
Adult trap efficiency
Near 100%
8-20%, with additional
recoveries on spawning surveys
23
-------
of historic splash-damming, large wood removal, and subsequent channel downcutting (Pat
Olmstead, Coos Bay BLM, personal communication). As a result, habitat conditions in the
mainstem West Fork Smith River are sub-optimal for coho salmon spawning and rearing
(ODFW HabRank model output; Andy Talabere, Corvallis ODFW, personal communication).
Tributaries to the West Fork Smith River contain higher loadings of large wood, more complex
pool habitats, and overall higher rankings of coho salmon habitat quality than mainstem reaches
(ODFW unpublished Aquatic Habitat Inventory Data).
hi contrast to the West Fork Smith River, Winchester Creek is a 3rd order watershed with
low topographic relief (elevation range 0 to 100 m) and a drainage area of approximately 11 km2.
Channel substrates are dominated by sand and silt (70-95%), with very limited occurrence of
gravel (ODFW unpublished Aquatic Habitat Inventory Data), hi 2001-2002, approximately 290
wild adult and 12 hatchery coho salmon returned to spawn in Winchester Creek (Bruce Miller,
ODFW personal communication). Distribution of coho spawning within Winchester Creek is
spatially restricted by the availability of suitable spawning gravels to approximately 2 km of the
upper West Fork Winchester Creek (Figure 4.1; Bruce Miller, ODFW personal communication).
This spatially-restricted pattern of spawning makes Winchester Creek of particular interest for
examining patterns of coho fry dispersal from known natal reaches and subsequent influences on
watershed smolt productivity (e.g., Kocik and Ferreri, 1998). Additionally, we hypothesize that
the distribution of marine-derived nutrients reflects the restricted distribution of spawners in this
watershed. Spatial variation in coho salmon growth has been documented among headwater,
beaver pond, and tidally-influenced portions of this stream network (Miller and Sadro, 2000), but
the influence of within-watershed variation in nutrient status or physical habitat quality on these
difference in realized growth are presently unknown. The diversity of potential rearing habitats
and an ability to trap migrant coho at the head of tide in Winchester Creek year-round presents
an excellent opportunity to examine movements, distribution, and growth of juvenile coho
salmon within an entire freshwater watershed.
Together, these two watersheds provide contrasting patterns of spawning and
overwintering habitat configuration that are well suited to our questions of interest. Although
these two watersheds provide a very limited sample size they represent a starting point for
identifying patterns of habitat configuration and fish distribution and survival that we will
24
-------
subsequently explore across a wider array of coastal Oregon watersheds (see Future Direction,
below).
Coho Spawning Survey
2001 -2002
Winchester Creek
»„•«,„ ">< ,, "'«
Coho Spawning Survey
20C1 - 2002
Figure 4.1 Distribution and relative abundance of coho spawners 2001-2002 in the
Winchester Creek and West Fork Smith River watersheds. Counts
based upon sighting of live fish and carcasses, thus do not account
for loss or movement offish between survev intervals.
In addition to coho salmon, fish species potentially present within coast range watersheds
include pacific lamprey (Lampetra tridentatd), brook lamprey (Lampetra richardsoni}, river
lamprey (Lampetra ayresi), fall chinook salmon, winter steelhead, sea-run and resident
cutthroat trout (Oncorphynchus clarki), reticulate sculpin (Cottus perplexus), coastrange sculpin
(Cottus aletiticiis), riffle sculpin ('Cottus gulosus), prickly sculpin (Coitus asper), torrent
sculpin (Cottus rhoihem), speckled dace (Rhinichthys osculus), longnose dace (Rhinichthys
cataractae), redside shiner (Richardsonius balteatus), largescale sucker (Catostomus
mact'ocheilus), Umpqua pikeminnow (Ptychocheihis oregonetisis), and three-spined stickleback
-------
(Gasterosteus aculeatus). This list likely is incomplete. Information on the distribution,
abundance and habitat associations of many of the non-salmonid species in coast range streams
is limited (Zaroban et al., 1999).
4.3 FIELD STUDY DESIGN
The objectives of the integrated watershed study are to 1) examine associations between
fish distributions, movement patterns, and spatial/temporal patterns of habitat quality within
watersheds, and 2) identify and characterize relationships between watershed/landscape
characteristics and stream habitat quality (Table 4.2). To address objective 1, we will quantify
distributional shifts and habitat-specific growth and survival of individual juvenile coho salmon
prior to emigration from the study watersheds. We will relate seasonal patterns of distribution,
growth, and survival of coho salmon, as well as overall fish assemblage structure, to
watershed/stream habitat characteristics, including the availability of watershed-derived and
marine-derived nutrients, variation in stream flow, and stream temperatures. To address objective
2, we will characterize the spatial and temporal patterns of critical habitat attributes for stream
fishes in relation to riparian and landscape characteristics. We will integrate results from both
objectives into spatially-explicit fish assemblage and coho population models (Section 7).
4.3.1 Seasonal Distributional Surveys
Coho abundance and fish assemblage structure, defined by species presence, will be
estimated during summer base-flow conditions using snorkel counts from a systematic sample of
pools (e.g., every 5Ih pool) throughout the entire watershed. Systematic pool surveys are being
increasingly used in coastal Oregon watersheds to assess distributional trends in juvenile coho
salmon summer abundance (Garono and Brophy, 2001). While we assume that coho salmon
distribution within the network will be accurately assessed using a calibrated pool-only snorkel,
we will test this assumption by conducting multiple-pass depletion electrofishing surveys in a
subset of stream reaches comprised of multiple habitat unit types (Hankin and Reeves, 1988).
Estimation of seasonal bias in snorkel ing estimates due to changes in fish behavior and
concealment (Roni and Fayram, 2000) may require mark-recapture population estimates in
26
-------
Table 4.2 Objectives and tasks for the integrated watershed study.
Objective
Tasks
Methods
Objective 1. Years 1-5
Examine associations among
fish distributions, movement
patterns, and spatial - temporal
patterns of habitat quality
within watersheds
Quantify distributional shifts of coho
salmon August through March
Characterize freshwater fish
assemblages
Seasonal (summer, fall, winter) snorkel
surveys of coho salmon distribution
Seasonal (summer, winter)
electrofishing surveys of freshwater
fish assemblage
Quantify growth of individual coho
salmon August through March
Characterize freshwater habitat
quality
PIT tag and externally mark (late
summer) and recapture (fall through
early spring) individual coho salmon
Seasonal (summer, winter) surveys of
physical habitat availability and quality
(physical structure, water chemistry);
Continuous monitoring of srreamflow
and temperature
Objective 2. Years 3-5
Identify and characterize
relationships between
watershed/landscape
characteristics and stream
habitat quality, particularly
with regard to the historical
and present-day distribution of
critical habitats for native
fishes
Characterize landscape distribution
of critical habitat attributes (e.g., high
quality overwintering habitat.
thermal or flow refugia) in relation to
riparian and landscape characteristics
Characterize historical (or potential)
distribution of critical habitats
Characterize spatial patterns and
concentrations of critical nutrients in
relation to salmonid spawner
distribution/abundance and
watershed vegetation patterns
Surveys and mapping of critical habitat
attributes in GIS overlays for spatial
analysis
Model habitat-landscape relationships
to simulate historic versus present day
critical habitat distributions
Water chemistry monitoring and
mapping, spatial analysis of salmonid
distributions, vegetation, and nutrient
status (see Section 6)
pools. Mark-recapture calibration efforts may also be needed for pools having high structural
complexity (Hankin and Reeves, 1988; Rodgers et al., 1992). Reliability of pool-only snorkel
estimates of species presence and derived assemblage estimates will also be assessed during
these multiple-pass depletion electrofishing surveys. This will allow us to account for negative
27
-------
bias inherent in fish assemblage structure estimates due to variation in habitat preferences and
detectability between species (Zaroban et al., 1999).
4.3.2 Coho Salmon Growth and Survival
We will focus on growth and survival of coho salmon during the initial two years of the
project, but will also tag and capture lesser numbers of juvenile steelhead trout and cutthroat
trout to complement research being conducted by the US Forest Service Pacific Northwest
Research Laboratory. Growth and survival of juvenile coho salmon, cutthroat trout, and
steelhead will be quantified using a mark and recapture/resighting approach. We will tag
salmonid individuals from a set of sites within the watershed in late summer using internally-
implanted passive integrated transponder (PIT) tags (e.g., Prentice et al., 1990). In addition,
coho salmon will be marked with externally visible dye markings (e.g., Thedinga and Johnson,
1995). PIT tags carry unique codes that can be read with an external reader, and can allow
individual identification and quantification of specific growth rates of recaptured fish. External
marks will be used to identify batches of PIT-tagged coho salmon captured from the same initial
reach, and marks visible to snorkelers will be used to identify locations of PIT-tagged fish during
snorkeling surveys. This information will be used to direct recapture efforts toward those
individuals, thus enhancing recapture probabilities for fish that have moved from original
tagging locations. We will select stream reaches for PIT tag studies to control, to the extent
possible, for variation in salmonid density (estimated from snorkel surveys) and summer and
winter habitat quality as estimated by existing ODFW Aquatic Habitat Inventory data. We will
re-capture and record length and weight of individually tagged coho salmon, cutthroat trout and
steelhead during fall, winter and spring surveys using electrofishing, seining, and minnow-
trapping. The fish assemblage will also be characterized at these sites. To increase likelihood of
recapturing tagged coho salmon individuals, recapture efforts will be focused on locations where
externally marked coho salmon were observed during watershed-wide snorkel surveys, hi
addition, we will use portable PIT tag readers to interrogate for PIT tags without capturing and
handling fish. Stationary PIT tag antenna arrays located at the mouths of Gold Creek, Beaver
Creek, Moore Creek, Crane Creek and additional locations as available will provide date and
time data on individual PIT tagged fish detected moving past these detectors. Stationary readers
28
-------
will allow us quantify patterns offish residency/movement within individual tributaries. Smolt
traps located 1.6 km above the mouth of the West Fork Smith River (and at the mouth of
Winchester Creek) will allow capture and PIT tag interrogation of recaptured emigrants leaving
the watersheds.
Given sufficient recapture rates, these data will allow us to compare growth rates,
condition factor, and survival rates of groups of juvenile salmonids tagged and recovered from
different habitats within the watersheds. Recapture and relocation data will allow minimum
estimates of between-habitat movement within the watershed and allow comparisons of growth
and condition factors between "movers" and "stayers" (e.g., Kahler et al., 2001). We will also be
able to compare retention rates of salmon and trout among stream reaches of varying habitat
complexity and presumed winter habitat suitability (Bell et al., 2001) and subsequent differences
in individual condition and growth. By starting tagging studies in August, we will capture
influences of summer habitat conditions on salmonid distribution and condition factor, and will
incorporate summer distributional patterns and rearing densities into the layout of the reaches
selected for PIT tag growth and survival studies.
PIT-tagged fish surviving ocean rearing may be detected upon return to the watershed as
they move past stationary PIT tag antenna arrays or are trapped at adult trapping facilities.
Recovery rates of these fish are expected to be very low, and will be insufficient to quantify
effects of smolt size on ocean survival, but will provide individual case-studies of adult survival
that we will be able to relate to juvenile rearing, movement and growth patterns that could be
used to parameterize life-history models (Section 7).
Influence of PIT tagging on coho salmon growth and survival will be assessed by
conducting laboratory trials using treatment (tagged) and control (untagged) coho salmon held at
a constant 20°C temperature. Survival and growth differences between tagged and control fish
will be compared over a four-week period.
4.3.3 Physical and Chemical Habitat Characterization
Environmental characterization of the intensive watersheds will consist of monitoring
temporal and spatial trends in key environmental factors believed to be important to native
fishes, including water chemistry, temperature, discharge, and channel morphology. Physical and
29
-------
chemical characteristics will be measured throughout each watershed, specifically at locations
where fish assemblages are characterized and where juvenile coho salmon are captured for PIT
tagging and marking.
Stream nutrients will be sampled monthly at a subset of the seasonal fish distribution
survey sites. Sampling will follow EMAP protocols (Herlihy, 1998) and samples will be
analyzed according to the "QA Plan for the Willamette Research Station" (Erway et al., 2001).
To help understand watershed process controls on nutrient concentrations in streams, we will
intensively sample at least one major hydrologic event at one site using continuous water
samplers (ISCO Instruments) within the West Fork Smith River. Water samples will be analyzed
for nutrients that are expected to play a major role in aquatic productivity, including ammonium,
nitrate, phosphate, total dissolved N, dissolved organic carbon and total organic carbon and
nitrogen. The following chemical constituents will also be determined, for use in constructing
ionic charge balances, to serve as indicators for geologic parent material, and to characterize the
stream environment: dissolved oxygen, dissolved inorganic carbon, total suspended solids, silica,
chloride, sulfate, calcium, magnesium, sodium, potassium, acid neutralizing capacity, pH and
specific conductivity.
Because paniculate nutrient inputs from terrestrial sources (allochthonous inputs) play
important food web roles in headwater streams, we will sample this component of the stream
nutrient environment also. We will place litterfall traps near or over the streams at a subset of the
fish sampling locations (5 replicates per station) to represent inputs at the watershed scale. The
litter will be collected monthly, sorted into two litter types (coarse and fine), and ground and
analyzed for carbon and nitrogen by Carlo-Erba CN analyzer.
Macroinvertebrate biomass and/or functional group composition will be assessed within
fish sampling sites to provide an index of food availability. Benthic samples and aerial pan traps
will be used to capture potential invertebrate prey for juvenile salmonids. Invertebrate
assemblages will be compared to gut contents sampled seasonally from juvenile salmonids.
These prey availability data will be used as a covariate in analyses of juvenile salmonid growth
and movement, and will also be related to stream nutrients.
Our initial water temperature monitoring will focus primarily on the West Fork Smith
River with more limited sampling in Winchester Creek. We will install Onset TidBit data loggers
(accuracy =H-/- 0.3°C) at a spacing of approximately 1-km throughout the stream networks to
30
-------
monitor temperature variability through an annual cycle. The quasi-continuous (e.g., half-hourly)
stream temperature records from these instruments will allow us to capture the diurnal and
seasonal patterns and variability of water temperature cycles as well as to detect maximum daily
temperatures during the critical summer season. Because of the much wider ranges in air
temperature compared to water temperature during a 24-hour period, placement of temperature
data loggers throughout the stream network will also alert us to the spatial and temporal patterns
of stream dewatering (i.e., portions of the stream network that go dry and result in the loss of
coho summer rearing habitat). Based on knowledge gained during this initial stage of
temperature monitoring, we will investigate finer-scale patterns of stream temperature variability
in West Fork Smith River and Winchester Creek, as well as additional integrated watersheds.
These investigations will allow us to better identify and understand temperature refugia
associated with logjams, beaver dams, off-channel habitats, tributaries, and locations
downstream from hyporheic flows. This information will complement the results from the
juvenile fish distribution snorkeling surveys.
Both West Fork Smith River and Winchester Creek have existing discharge gauging
stations installed and maintained by the State of Oregon located near the mouths of both
watersheds from which we will obtain continuous watershed discharge. In addition, we will
periodically (e.g., monthly or bi-weekly frequency) measure instantaneous discharge on the
major tributaries within each watershed starting in the summer of 2002. During the second year
of the project, we will install pressure transducers and Campbell data loggers to continuously
monitor tributary stage. By developing regression equations between instantaneous discharge
and stage we will be able to estimate daily tributary discharge and water yield for hourly to
yearly periods. Stream discharge monitoring will expand to additional gaged integrated
watersheds in the second through fifth years of the project. For other watersheds, e.g., ODFW
salmon juvenile survey sites, and on a regional scale, we will estimate discharge for un-gaged
streams by developing and verifying a regression model that uses precipitation and similar gaged
watershed data. This effort is described in Section 5.
During high flow period of November 2003 through March 2004 we will measure
turbidity of stream water at the mainstem and tributary study reaches on a monthly basis. This
sampling effort may be increased during subsequent years in we find a large degree of variability
in turbidity within the stream network.
31
-------
In addition to reach-specific habitat data collected as described herein, channel and
riparian habitat inventory data are available from the ODFW Aquatic Habitat Inventory project
for the West Fork Smith River and Winchester Creek watersheds (Moore et al., 1997). Reach,
sub-watershed and watershed scale summaries of physical habitat attributes will be extracted
from these databases for use in analyses offish distributional and movement patterns.
4.3.4 Analyses
Data will consist of stream network-scale seasonal distributional patterns of coho salmon
and associated fish assemblages, individual growth estimates from recaptured tagged coho
salmon, estimates of distances moved by recaptured individuals, survival and growth estimates
for groups of coho salmon tagged from different reaches of origin, and survival and growth
estimates from groups of coho salmon recaptured from different locations throughout the
watershed. We will relate network-scale fish distributional patterns to network-scale habitat
features using graphic, correlative, and regression approaches that explicitly incorporate spatial
structure of responses such as fish occurrence or density and explanatory variables such as
stream channel gradient, wood volume, pool depth, or temperature (Legendre, 1993). We will
also relate multivariate indices offish assemblage structure to physical and chemical habitat
factors using ordination approaches including non-metric multidimensional scaling (NMDS; e.g.,
Brazner and Beals, 1997). Growth, survival and movement data from individual fish will be used
to classify individuals into groups offish sharing similar performance characteristics using
multivariate clustering. Alternatively, groups may also be able to be defined based upon tagging
reach of origin, size or condition factor at time of tagging, or extent of movement (e.g., "mover"
or "stayer") depending upon specific questions of interest. From these data, we will be able to
identify habitat factors associated with differences in growth, condition, and survival among
coho salmon groups using multiple group comparisons. As described in the methods above, PIT
tagging reach selection will be designed to partition "treatments" into a priori classes of habitat
quality and spatial location. However, reaches selected for PIT tag studies will not be true
replicates due to our inability to control the broad suite of factors known to influence juvenile
salmonid growth and survival. By using physical and chemical habitat characteristics as potential
covariates in group comparisons, we will attempt to account for additional confounding factors
32
-------
influencing variation in coho salmon growth and survival differences between and among study
reaches in the analyses.
4.4 FUTURE DIRECTIONS
Results from the integrated watersheds in years 1 and 2 will provide preliminary data on
the spatial and temporal patterns of multiple, interacting aspects of habitat quality and associated
fish responses within coastal stream networks (Objective 1, Table 4.2). We will particularly
focus on growth and survival of coho salmon, and native fish assemblage structure. Preliminary
results will be used to design field experiments necessary to test specific hypotheses in years 3
through 5. Potential applications include controlled manipulations of nutrient inputs, or
derivation of more precise estimates offish dispersal or season-specific growth and survival
rates. A sub-set of watershed and fish fitness indicators specifically related to spatial
arrangement of important rearing and spawning habitats will also be developed from these initial
data that will be examined at additional sites (1-3 watersheds) in the Oregon coast range. These
additional integrated watershed surveys will be designed to capture sources of variation
hypothesized to be important factors influencing the interplay of coho salmon life histories and
watershed characteristics. Characterization of watershed-stream habitat linkages will also be
pursued in years 3 through 5 (Objective 2, Table 4.2), focusing on aspects of habitat quality of
demonstrated importance to coho salmon and native fish assemblages as determined in the initial
years of the integrated watershed study.
Results from the integrated watershed study (e.g., observed seasonal distributional
patterns of coho salmon and other native fish species) will be linked to spatially-explicit salmon
and native fish population modeling efforts (Section 7). We will derive inferences about spatial
habitat relationships from observed distributions, and predict outcomes of changes to habitat
configuration and quality. Ultimately, these efforts could help identify coast-wide restoration
potential of re-connecting life history patterns and habitat for coho salmon and other native
fishes.
33
-------
5.0 BROAD-SCALE ANALYSES
The integrated watershed study described in Section 4 involves a small number of
watersheds sampled relatively intensively to assess within-watershed patterns and processes. We
use the term broad-scale analyses to refer to analyses involving larger numbers of watersheds,
typically with less intensive data per watershed, where the emphasis is on among-watershed
comparisons. For the most part, our broad-scale analyses will rely on existing data sets and data
being collected by others, with very limited new data collection as part of this project. Thus, the
types of analyses we can conduct are highly dependent on the types of data available.
One reason we decided to place less emphasis on broad-scale analyses, compared to the
integrated watershed studies, is because of the regional focus of CLAMS. As discussed in
Section 3.3, CLAMS is a multi-investigator, multi-year project that will predict future changes in
vegetation, stream habitat, and salmon freshwater production potential for the entire Oregon
coastal region. As part of that effort, CLAMS scientists have developed comprehensive regional
characterizations (e.g., GIS coverages of vegetation, elevation, soils, geology, stream networks)
which we will use to assess, among other things, the representativeness of the watersheds
selected for the integrated watershed study. They are also examining statistical correlations
between landscape features, ODFW stream habitat data, and ODFW measures of juvenile coho
salmon abundance (Burnett 2001, Burnett et al., unpublished). Broad-scale analyses we conduct
will add value to those already underway in CLAMS by focusing on other fish species and
habitat parameters, such as nutrients, not included in CLAMS analyses. The primary purpose of
our analyses is not regional prediction. Instead, we will use correlations among fish, habitat, and
landscape features to infer likely causal mechanisms and the relative importance of causal
factors, to help address Research Questions 2 and 3 (Section 3.1). We propose to analyze three
primary data sets: (1) EMAP and related surveys using EMAP protocols, (2) ODFW surveys of
juvenile coho abundance, (3) ODFW Life-Cycle watersheds. In addition, we will use long-term
stream flow records from gaged sites, to develop models for estimating stream flow for years
and watersheds without flow records.
34
-------
5.1 EMAP AND RELATED DATA
Over 150 stream reaches within the project study area have been sampled using EMAP
protocols and probability survey design (see Figure 3.1; Merger and Hayslip, 2000; Rose, 2000).
For each, we have data on fish, benthic invertebrates, physical habitat, and water quality
collected during summer low flow 1994-2001, with generally only one visit per stream. Fish
were surveyed using 1-pass electroshocking over reach lengths equal to 40 times the stream
width, providing indicators offish species presence/absence, richness, relative abundance, and an
aggregate index of biotic integrity (IBI; Hughes et al., 1998). The data are not adequate to
estimate juvenile coho abundance. Measurement errors include (a) differential susceptibility of
fish species to electroshocking, (b) influence of stream depth, complexity and other features on
the effectiveness of electroshocking, and (c) among-year and among-season variations in fish
communities (Herlihy et al., 1997; Bayley and Peterson, 2001). Nevertheless, these data still
provide a valuable snapshot of regional patterns in fish assemblages in coastal streams. We will
use these data to address the following sub questions under Research Questions 2 and 3.
2e. What habitat aspects (physical habitat, stream flow, nutrients, temperature) are most
strongly associated with fish assemblage structure in coastal streams?
3a. What natural and anthropogenic landscape features are most strongly associated with fish
assemblage structure in Oregon coastal streams?
We will examine several different indicators offish assemblage structure: the
presence/absence of individual fish species, native fish richness, fish IBI, and major fish
community types (identified using clustering methods; Anderberg, 1973; Legendre and
Legendre, 1998). We will analyze the association between each indicator offish assemblage
structure and (a) in-stream habitat data and (b) in-stream habitat and landscape data combined.
EMAP measures of in-stream habitat include a wide range of reach-scale physical habitat
variables, such as stream gradient, pool depth and frequency, volume of large woody debris, and
substrate type (Kaufrnann et al., 1999). In addition. EMAP collects water quality samples
(including nutrients) and makes instantaneous measurements of stream flow and temperature.
Using the stream flow models described in Section 5.4, we will also estimate for each reach
additional hydrologic parameters, including an index of hydrologic flashiness (see section 5.4 for
35
-------
definition), peak flow, and low flow. Landscape variables will include reach-level EMAP
measurements (e.g., canopy cover and riparian condition), reach-level metrics derived from CIS
layers (e.g., valley-floor width index), and upstream riparian-buffer and whole-watershed
characterizations derived from GIS layers (such as vegetation composition and landslide
potential). Because of the large number of potential explanatory variables, and high degree of
collinearity among many of these variables, we will approach the analysis in three stages:
1. Limit candidate variables to those identified as important for juvenile coho abundance in
CLAMS analyses. Do native fish assemblages respond to the same types of in-stream and
landscape variables as coho salmon?
2. Develop and test specific hypotheses about important habitat and landscape factors
derived from literature reviews and existing knowledge about fish habitat requirements.
3. Subsequently, conduct analyses with all candidate explanatory variables (both individually
and using principal components analysis to aggregate multiple variables, Legendre and
Legendre, 1998). Are there any surprises, that is, variables more highly correlated with
fish assemblage structure than those identified in stages 1 and 2? If so, are there logical
mechanisms that could explain the high correlation? These results could lead to additional
hypotheses for testing in the integrated watershed studies.
The specific statistical technique will vary depending on the characteristics of the data, but will
include multiple linear regression, logistic regression, classification and regression tree analysis,
and multivariate regression trees (Breiman et al., 1984; Harrell, 2001; De'ath, 2002). We
conducted similar analyses using EMAP data for the Willamette River Basin, Oregon (Van
Sickle et al., In review). We can also draw upon the experience of EMAP researchers examining
fish-habitat-landscape relationships (e.g., Herlihy et al., 1997; Rose, 2000) as well as a wealth of
literature describing similar analyses for other regions (e.g., Matthews, 1985; Maret et al., 1997;
Angermeier and Winston, 1998; Porter et al. 2000; Olden et al., 2002).
5.2 ODFW JUVENILE COHO SALMON SURVEYS
As part of their salmon monitoring program, ODFW researchers estimate juvenile coho
abundance from snorkel surveys in 50 stream reaches per year selected using a rotating panel
design in each of the five Gene Conservation Areas along the Oregon coast (Rodgers, 2000;
36
-------
2001). Between 1998 and 2001, 441 reaches were surveyed within our study area (see Figure
3.1). Most of these sites also have physical habitat data measured in the same year using ODFW
protocols (Moore et al., 1997). As noted earlier, these data will play a major role in CLAMS'
analyses of relationships between stream habitat and landscape features, and coho abundance and
habitat-landscape features. We propose to expand these analyses in two ways, by adding: (1)
improved estimates of stream How and hydro logic characteristics, using the models described in
Section 5.4 and (2) measurements of stream nutrients, as additional candidate predictors of coho
abundance.
Stream reaches surveyed previously by ODFW for juvenile coho abundance will be
compared with available EMAP and REMAP data for correspondence (spatial and temporal) and
will be evaluated for explanatory power (Section 5.1). Based on the results of these analyses of
prior data, we will perform supplementary sampling of water quality (especially nutrients) at
selected juvenile coho study sites beginning in 2003. Information on which site selection will be
based will include existing data on juvenile abundance, water chemistry and site characterization
(index of habitat quality) available from the ongoing CLAMS Project. Approximately 50
juvenile coho survey sites will be sampled for water quality twice a year - once during late
summer low-flow just prior to fall rains (September) and once at late-winter/early-spring
baseflow (March). We have found from our prior two-years of high frequency sampling of the
Salmon River on the Oregon Coast that these are the two most reliable and informative index
times for sampling Coast Range stream chemistry.
Water samples will be analyzed for nutrients that we expect to play a major role in stream
productivity, including ammonium, nitrate, phosphate, total dissolved nitrogen, and dissolved
organic carbon. In addition, the following chemical constituents will be analyzed as indicators of
geologic parent material and for use in constructing ionic charge balances for quality control:
dissolved inorganic carbon, total suspended solids, silica, chloride, sulfate, calcium, magnesium,
sodium, potassium, acid neutralizing capacity, pH, and specific conductivity. Sampling and
analysis will follow existing EPA protocols (Lazorcheck et al., 1998; Erway et al., 2001).
We will regress juvenile coho abundance as a function of: (a) indicators of stream
hydrology (stream flashiness, peak flow, low flow), and (b) measures of stream nitrogen and
phosphorus, both with and without accounting for the physical habitat and landscape variables
37
-------
identified as important determinants of coho abundance in CLAMS analyses. These analyses
will address the following sub questions under Research Question 2:
2d. Are juvenile coho salmon less abundant in streams with high peak flows and flashier
hydrographs?
2e. Are juvenile coho salmon more abundant in streams with higher nutrient concentrations?
The nutrient data collected at these sites will also be used to evaluate empirical models relating
stream nutrients to watershed characteristics, as described in Section 6.2.2.
5.3 ODFW LIFE-CYCLE WATERSHEDS
ODFW monitors the numbers of both adult coho returning and smolt outmigrating to
estimate overall freshwater survival and marine survival for eight Life-Cycle watersheds (Solazzi
et al., 2001). These watersheds are distributed across coastal Oregon (Figure 3.1), although
watershed selection was tightly constrained by the logistical requirements of the monitoring
protocols. Data are available since 1995, although data collection began in different years in
different watersheds. Together with ODFW and CLAMS, we will evaluate the primary habitat
and landscape variables responsible for both among-watershed and among-year variations in
coho freshwater survival. ODFW has collected physical habitat data for all eight watersheds, and
CLAMS is providing extensive landscape characterizations. Similar to the juvenile coho analyses
in Section 5.2, our contributions will be primarily (1) improved estimates of stream flow and
hydrologic characteristics from the models described in Section 5.4 and (2) measurements of
stream nutrients. Synoptic samples of water chemistry will be collected seasonally at the base of
each Life-Cycle watershed, and analyzed as described in Section 5.2. Using these data, in
combination with the physical habitat and landscape data from ODFW and CLAMS, we can
address the following sub questions of Research Question 2:
2f. For any given life-cycle watershed, is coho freshwater survival lower in years with higher
peak flows or lower base flows?
2g. In years with relatively high winter flows, is coho freshwater survival lower in
watersheds with flashier hydrographs, and is there an interactive effect between stream
38
-------
flow and the availability of high quality overwinter habitat (refugia from high flows in
off-channel habitats and deep pools)?
2h Do watersheds with higher coho freshwater survival have higher nutrient concentrations
(with and without accounting for other habitat variables likely to affect coho survival)?
In addition, the detailed understanding of factors that affect coho survival and growth obtained in
the integrated watershed study (Section 4) should inform the search for mechanisms explaining
among-watershed and among-year variations in freshwater survival across all eight Life-Cycle
watersheds. Thus, based on results from the integrated watershed study, we expect to target
additional habitat and landscape parameters to measure in each life-cycle watershed in years 3-5
of the project.
5.4 STREAM FLOW
As discussed in Section 2, stream flow and its interactions with physical structure of
streams are important factors in the survival and productivity of fish in Oregon coastal streams.
Stream gaging is ongoing at the West Fork Smith River and Winchester Creek integrated
watershed study sites (Section 4.0), but unfortunately, most of the EMAP sites (Section 5.1),
ODFW juvenile survey sites (Section 5.2), and ODFW Life-Cycle watershed sites (Section 5.3)
do not have current or historic stream flow data.
To allow the inclusion of stream flow as a variable in our broad-scale analyses, we will
employ an empirical modeling approach to estimate stream flow for watersheds and time periods
for which hydrologic data are not available. Our goal is to be able to estimate hydrologic
parameters such as fiashiness (i.e., hydrologic response = stormflow/total flow), annual peak
daily stream flow, maximum annual 3-day or 7-day duration stream flow, annual minimum daily
stream flow, minimum annual 3-day or 7-day duration stream flow, and total annual water yield.
Estimation of these parameters will require that we be able to make reasonable estimates of the
annual hydrographs for coastal streams.
Modeling stream flow of coastal Oregon streams with mechanistic models is challenging.
The terrain is mountainous with highly variable precipitation among and within stream
drainages. There are a limited number of long-term precipitation sites, and they tend to be
located near low-elevation towns and cities (Oregon Climate Service website:
39
-------
http://www.ocs.orst.edu). Consequently, precipitation data for mid and high elevations of the
Oregon Coast Range are quite sparse. Most traditional streamflow modeling approaches are
dependent on precipitation measurements at the same time scale of the stream flow being
modeling (Haan et al., 1982). In addition, detailed watershed information often is required.
A fairly extensive collection of stream flow data exist for streams within our coastal
Oregon study area and nearby drainages (Oregon Department of Water Resources website,
http://www.wrd.state.or.us/surface_water; Figure 5.1). Unfortunately, relatively few streams that
currently are gaged also have long-term discharge records (Figure 5.2). The Nehalem River,
Wilson River, Siletz River, Alsea River, South Fork Coquille River, and Smith River (northern
California) are currently gaged and have at least 50 years of streamflow data. Stream gaging on
many streams with long-term historic stream flow data records has been discontinued, and many
streams were gaged only for relatively short periods of time (1 to 10 years).
Preliminary analyses of selected stream flow records from Oregon coastal streams and
rivers indicate that a relatively simple modeling approach may allow estimation of daily stream
flow in ungaged streams using a combined watershed classification and regression modeling
approach. Figure 5.3 shows the daily stream flow (normalized by drainage area) for a series of
Oregon coastal rivers during December 1965 - June 1966. Not surprisingly, a clear pattern exists
of high flow during the winter months followed by declining flows in the late spring. A more
unexpected observation is that a similar pattern of storm flow events occur in the all of the rivers,
even though the rivers are located in the north, the central and the south Oregon coast.
Furthermore, the overall stream flow patterns in the three rivers with predominately igneous
bedrock watersheds, Wilson, Trask and Siletz Rivers, are very similar. The three rivers with
primarily sedimentary bedrock watersheds, Alsea, Smith (Oregon) and South Fork Coquille
Rivers, have lower stream flow during recessional periods than do the igneous watershed
streams, and differences in stream flow among the rivers with sedimentary watersheds are
greater than the differences among the streams with igneous watersheds. The stream flow
patterns evident in Figure 5.3 suggest stream flow for streams with similar bedrock geology
could be estimated using a simple regression model from a nearby stream, with similar bedrock
geology.
40
-------
Figure 5.1. Stream gaging stations in coastal Oregon with historic streamflow data.
im Csustal
Nanfaem California HUCS
Figure 5.2. Current stream gaging stations in coastal Oregon,
41
-------
To test this idea, we used Alsea River stream flow data to estimate Smith River (Oregon)
stream flow during October 1, 1965 through September 30, 1966. First we divided the annual
hydrograph into a high flow (mid-November through April) and a low-flow, recessional period
(May through September). We developed two simple linear regression models (Neter and
Wasserman, 1974) with Smith River stream flow as the dependent variables and Alsea River
stream flow as the independent variables during each of the two flow periods (r^ 0.84 for high
flow period, r2 = 0.99 during low-flow, recessional period). We then used these models to
estimate Smith River stream flow during the entire water year (October 1, 1965 - September 30,
1966). Stream flow during October - mid November was estimated with the low-flow,
recessional model. Results, shown in Figure 5.4, are encouraging because the estimated annual
hydrograph for the Smith River was quite similar to the measured annual hydrograph. There are
some obvious differences in the peak discharge for several storms, but nevertheless, the overall
results are encouraging.
0.1 -
CD
O5
TO
"I
0.01 -
0.001
Alseu Rlvui
Smith River
S.F Coqutlle Rlv
Dec
Jun
Jul
Figure 5.3. Streamflow (discharge/watershed area) for six coastal Oregon rivers,
December 1,1965-June 30, 1966.
42
-------
Smith River
500
400 -
•5f
IE 300 -
0)
O)
CO
•£ 200 -
100-
Oct
-•— Measured Q
• o- Estimated Q
Dec
Feb
Apr
Date
Jun
Aug
Figure 5.4 Estimated and measured streamflow of Smith River during October 1,1965
September 30,1966.
During the next two years, we will expand on the results from our preliminary analyses
to implement a modeling effort that will allow the estimation of stream flow (discharge) in
ungaged basins in our coastal Oregon study area. The major steps in the modeling effort are as
follows.
1. Compile available discharge data (minimum of one complete year of data) for streams and
rivers in the study area. Data are available at U.S. Geological Survey
(http://wvvw.waterdata.usgs/or) and Oregon Department of Water Resources websites
http://www.wrd.state.or.us/surface_water). We will analyze data by water year (October 1
- September 30). We will express annual hydrographs on a unit area basis (e.g., Figure
5.3).
2. For each stream with long-term discharge records (>50 years), rank water years according
to total water yield, maximum annual flow, and minimum low-flow.
3. Classify streams, for which discharge data are available, into similar groups based on
watershed bedrock geology, watershed morphology and topography, annual precipitation
patterns, and geographic position along the coast. We may include land use and other
43
-------
classes of watershed information in the analyses. We will use PRISM annual average
precipitation maps (Daly et al., 1997) to determine annual average precipitation patterns.
We then will use analysis-of-variance procedures (Neter and Wasserman, 1974) to test for
differences in annual hydrograph characteristics among groups and to assess the validity
of groups. (Note: To the extent possible, we will use data from the same period of record
to accomplish this task. Ideally, we would like to have data from at least one wet year, a
moderate year, and a dry year. This is unlikely for all streams because of the gaps in the
discharge records for many streams. Consequently, we most likely will need to include
data from different water years for some streams. The ranking exercise in step two
provides the basis for selecting similar water years for our analyses.)
4. For each group of watersheds, calculate average discharge per unit watershed area for
each water year for which the majority of watersheds within the group have discharge
records.
5. Develop empirical models using regression techniques (Neter and Wasserman, 1974) to
estimate average watershed group discharge (step 4) based on discharge from a coastal
Oregon stream that is currently gaged and has a long-term flow data. If possible, we will
use two thirds of the available data for model building and one third of the data will be
used for model testing. We will quantify the degree to which the empirical models can
estimate the total water yield, peak flows and minimum flows of the average annual
hydrographs representing the watershed groups.
6. Use empirical models developed in step 5 to estimate stream flows at ODFW Life Cycle
watersheds and juvenile coho survey locations and EMAP locations to support activities
described in sections 5.1 - 5.3.
5.5 FUTURE DIRECTIONS
In general, we intend to complete the work described in Section 5 during the first two
years of the study. We do, however, expect to identify important habitat and landscape
parameters in the integrated watershed study (Section 4), and will plan to measure them in the
ODFW Life-Cycle watersheds during years 3-5 of the project. Furthermore, we may need to
conduct additional study-wide sampling or analyses to allow application of our study results to
the Oregon coastal area as a whole.
44
-------
6.0 ROLE OF NUTRIENTS IN SALMON HABITAT
6.1 BACKGROUND
Many freshwater environmental factors such as temperature, light and stream flow have
an important influence on habitat for salmon and other native fishes. Nutrient availability is also
a major factor regulating aquatic ecosystem structure and productivity. Nitrogen and phosphorus
additions have been shown to stimulate algal production in many streams (Elwood et al., 1981;
Peterson et al., 1983; Triska et al., 1983). Increased abundance of stream algae leads to increased
abundance of stream herbivores (Gregory, 1983) and this increased productivity is carried
through the aquatic community to higher trophic levels.
hi recent years, researchers have recognized the existence of cultural oligotrophication,
where humans have reduced the natural processes of nutrient delivery to aquatic ecosystems, in
some areas of the Pacific Northwest (Stockner et al.. 2000). This cultural oligotrophication is in
strong contrast to the more commonly recognized cultural eutrophication in areas where human-
accelerated nutrient loading has stimulated the productivity of many aquatic ecosystems. One
such example of oligotrophication has been the decline in runs of anadromous fish (Gresh et al.,
2000; Stockner et al., 2000), and one possible consequence is the loss of nutrient resources for
the next generation of salmonid juveniles (Larkin and Slaney, 1997; Cederholm et al., 1999;
Stockner et al., 2000). Decaying salmon carcasses increased stream nutrient concentrations and
stream periphyton biomass in small spawning streams of Lake Superior (Schuldt and Hershey,
1995). Decaying salmon carcasses increased stream biofilm ash-free dry mass as well as
macroinvertebrate densities in streams of southeastern Alaska (Wipfli et al., 1998), when
compared to stream reaches where carcasses were excluded. Direct additions of inorganic
nutrients to a small coastal stream in British Columbia increased periphyton biomass as well as
fry size of steelhead trout and coho salmon (Johnston et al., 1990). Such bottom-up effects have
resulted from nutrient additions to an Alaska tundra stream also, with detectable effects
throughout the food web up to, and including, Arctic grayling (Thymallus arcticus) (Peterson et
al., 1983).
Studies using stable isotopes of carbon and nitrogen have provided direct experimental
evidence of the incorporation of nutrients from decaying salmon or from eggs of spawning
45
-------
salmon into resident fish or into the fry and juveniles of the next generation of anadromous
salmonids (Kline et al., 1990; Bilby et al., 1998; Bilby et al., 2001). The incorporation of such
marine-derived nutrients (MDNs) into surrounding riparian and upland animal and plant
communities also has been demonstrated, where they may play an important role in the dynamics
of those forests (Wilson et al., 1998).
With the recognition of the loss of nutrient inputs to streams as a consequence of steeply
declining salmon returns, fisheries researchers and managers have started a discussion of
supplementing these streams with additions of externally derived salmon carcasses or of
fertilization with inorganic nutrients (Cederholm et al., 2000). Against this background of
research and restoration studies, however, it must be recognized that within a region, streams can
have widely varying nutrient concentrations and even limiting nutrients, dependent upon the
watershed bedrock composition (affecting phosphorus concentrations) or forest community
composition (e.g., red alder dramatically increasing stream nitrogen concentrations). Streams of
the Oregon Coast range are quite variable, and may well be subject to higher "natural" loading of
key nutrients (Wigington et al., 1998) controlling or influencing stream productivity than streams
for which other nutrient or carcass loading studies have been done. Thus, previous experiments
or observations in other regions within the Pacific Northwest may not translate directly to
Oregon coastal stream ecosystems. The role of stream nutrient concentrations on the viability of
salmonid populations has yet to be studied directly in the Oregon Coast range. Key questions
remain on both the sources of nutrients for salmon juveniles (e.g., returning salmon adults versus
watershed sources) and their relative importance. The sources of nutrients and the role that these
different sources play are important questions with regards to maintenance and restoration of
salmonid populations that could have important implications for fish harvest as well as land
management.
6.2 APPROACH
As presented in Section 3, we will address the following over-arching questions. "What
roles do nutrients, temperature, and flow play, relative to physical habitat, in determining coho
salmon freshwater survival and growth? How do these factors influence fish assemblage
structure?" Addressing these questions will require a close examination of the links among
landscape characteristics, stream nutrients and fish populations. We propose two approaches:
46
-------
• examining the relationships between stream nutrients and fish
• identifying the links between landscape characteristics and stream nutrients.
We present an overview of the questions we plan to address, combined with the approaches and
datasets, in Table 6.1.
Since January 2000, we have been examining the factors influencing stream chemistry in
40 small watersheds within the Salmon River basin of the Oregon Coast Range. Our prior work
and that of Wigington et al. (1998) has shown that stream nitrogen and phosphorus across this
forested region are quite variable, and are strongly related to land use/cover, geology and sea salt
inputs (Compton et al., In review; Church et al., In prep.). The Oregon Coast Range landscape
has a variety of bedrock types and nitrogen supply, resulting in tremendous variability in
streamwater nutrient concentrations. These soils are very rich in organic matter and have
substantial N leaching losses due to atmospheric inputs from nitrogen fixation by red alder
(Binkley et al., 1992). In addition, variations in bedrock types influence stream phosphorus and
cation concentrations. Nitrogen and phosphorus availability are potentially important controls on
stream productivity in the Pacific Northwest. Variations in stream nutrients could strongly affect
aquatic productivity of all trophic levels across small watersheds in the Coast Range. Therefore,
we propose to build upon our prior work examining the controls on stream nutrients by
developing an improved understanding that leads to approaches for extrapolating across larger
areas of the Coast Range. In addition, as described in Section 5 (Broad-Scale Analyses), we will
link this variation in stream nutrients to success of salmonid populations, by examining new and
existing data sets that combine stream chemistry and juvenile and adult fish populations.
Specifically, we will address the questions outlined in Table 6.1.
6.2.1 Question 1. What is the relative importance of watershed-derived versus marine-derived
nutrients to fish nutrition?
We will examine the comparative influence of watershed-derived versus marine-derived
nutrients in two ways. First, we will compare patterns and numbers of adult salmon returns with
natural patterns of nutrient concentrations. This will be done at one or more Integrated
Watershed Study sites (see Section 4), where we have the combination of adult surveys plus
measurements of nutrient concentrations in the following stream components:
47
-------
• returning adults
• juvenile fish
• stream periphyton
• stream macroinvertebrates
• coarse and fine particulate organic matter in litterfall and stream water
Table 6.1 Questions, goals and methods for the nutrient studies.
Question
Goal
Methods
What is the relative
importance of watershed-
derived versus marine-
derived nutrients to fish
nutrition?
Examine the role of MDNs in stream
reaches with varying levels of adult
returns
Determine the importance of MDNs
in stream food webs across sites of
varying N status
Use natural abundance stable
isotope approach in Integrated
Watershed Studies sampling of
natural gradients in fish returns
Carcass planting experiments,
sampling of foodweb components
and using stable isotope data
What are the major
processes and landscape
factors, both natural and
anthropogenic, which
control spatial patterns
and concentrations of
nutrients?
Testing the relationships between
watershed characteristics (focusing
on land use/cover and geology) with
stream nutrient concentrations and
fluxes.
Compare the relative magnitude of
watershed nutrient losses with the
nutrient inputs associated with
returning anadromous fish
Determine the role of sea salt inputs
and soil exchange processes in
coastal stream nutrient
concentrations and watershed
budgets.
Testing of empirical models in other
Coast Range Watersheds, using data
from Juvenile Watersheds and
Integrated Watershed Study
Biogeochemical mass balance
approach using stream chemistry,
discharge and estimated peak
counts for the Salmon River
Watershed from ODFW.
Soil column and iysimeter chloride
addition experiments
In addition, we plan to conduct experimental carcass manipulations, in order to examine
the effects of carcasses on stream N fate and juvenile salmon. This work will be conducted in
streams with varying N status, building on our previous understanding of the variations in stream
chemistry across the Salmon River basin. Because returning adult salmon have a carbon and
nitrogen isotopic signature that is naturally enriched in I5N and 13C when compared to dissolved
nitrogen forms, this marine N and C signal can be detected in freshwater organisms (Bilby et al.,
48
-------
1996). In the fall of 2003, we will begin studies to examine the fate of MDNs in streams of
differing inherent nitrogen and phosphorus status. This work will involve gathering salmon
carcasses from the Salmon River hatchery, transporting them to selected streams of varying
nutrient status, tagging these carcasses, placing them in the stream and monitoring the fate of
these carcasses for 1, 3, 7, 14, 30 and 60 days after placement. This work will be done in
conjunction with ongoing work by ODFW.
We will measure the carbon, nitrogen and phosphorus concentrations and stable isotopic
composition of the foodweb components identified above in this section with regards to stable
isotopes of nitrogen to determine the relative contribution of sources to juvenile salmonids. In
addition, we will collect insects on carcasses, juvenile salmon, and algae on nearby rocks that
may take up salmon-derived nitrogen. Sampling will be conducted for isotope analysis of
functional foodwebs (e.g., periphyton, epixylon, shredders, and scrapers). We will place litterfall
collectors immediately adjacent to the stream to characterize the inputs and composition of
terrestrially derived organic matter to the stream foodwebs. These biological samples will be
freeze dried, ground and analyzed for C and N by automated CN analyzer (Carlo-Erba
Instruments), P by modified Kjeldahl digestion and 515N and 813C by isotope ratio mass
spectrometry (Finnigan Delta Plus). These analyses will allow us to follow the flow of I5N from
returning adults through to the next generation of juveniles, providing key information on the
relative contributions of nutrients from returning adults (+12 %o 6I5N) versus terrestrially derived
watershed sources (expected values, -5 to +5 %o 8I5N) to juvenile salmonids, a question of
significant recent concern in the Pacific Northwest (see Section 6.1). Although there are no
stable P isotopes to use in an equivalent manner, we will follow P concentrations in biological
samples, because P could play an important role in these food webs. Future efforts will build
upon this exploratory work.
6.2.2 Question 2. What are the major processes and landscape factors, both natural and
anthropogenic, which control spatial patterns and concentrations of nutrients?
Three factors play primary roles in nutrient delivery to streams of the Oregon Coast
Range: a) watershed characteristics (including land use/cover and geology), b) atmospheric
deposition (wet + dry) of sea salts originating from the ocean, and c) MDNs in the form of
returning anadromous fish. Once we have generated an empirical understanding of the
49
-------
relationship among these landscape factors and stream chemistry, we will test this approach
across broader spatial scales within the Oregon Coast Range. Our specific goals are to:
2A. Examine the relationships between watershed characteristics (focusing on land use/cover
and geology) with stream nutrient concentrations and fluxes.
2B. Compare the relative magnitude of watershed nutrient losses with the nutrient inputs
associated with returning anadromous fish using a biogeochemical mass balance
approach.
2C. Determine the effects of sea salt inputs and soil exchange processes on coastal stream
nutrient concentrations and watershed budgets.
Goal 2A. Examine the relationships among watershed characteristics (focusing on land
use/cover and geology) with stream nutrient concentrations and fluxes.
We will compare stream chemistry with spatial data on land use/cover and geology. For
the past 3 years, we have been sampling stream chemistry in 40 small watersheds within the
Salmon River Basin within the Coast Range of Oregon. The Salmon River basin encompasses a
variety of forest land uses, bedrock types, salmon returns, and distances from the ocean.
Therefore we are able to examine the relative importance of land use/cover, geology, sea salt
deposition and marine derived nutrients in the processes and factors controlling nutrient delivery
to streams.
The Salmon basin comprises approximately 220 km2. The Coast Range soils are
generally old and highly weathered, and developed from Miocene and Eocene age sedimentary
and volcanic rocks (USDA, 1997). The soils in this watershed can be very rich in organic matter
and experience substantial N leaching losses where red alder is present (Binkley et al., 1992).
The watersheds are largely forested, dominated by sitka spruce (Picea sitchensis) and western
hemlock (Tsuga heterophylld) near the coast, shifting to western hemlock-Douglas-fir
(Pseudotsuga menziesii) forests toward the higher end of the watershed. Ownership in the lower
watershed is a mixture of private non-industrial and federal (Siuslaw National Forest) lands,
whereas ownership in the upper watershed is largely private industrial forest. Land cover data
were taken from the CLAMS project (Ohmann and Gregory, In press). Although this region
experiences low rates of atmospheric N deposition, soil N contents are very high, primarily due
to the widespread presence of red alder. Indeed, red alder stands are a dominant component of
the landscape, averaging 25% of watershed area for our study streams.
50
-------
Our studies to date in the Salmon River Basin indicate that there are several important
controls on nutrient concentrations in tributaries of that system. Inputs of N from pure and mixed
alder stands appear to play a major role in controlling N losses from these small watersheds
(Figure 6.1). In addition, heavy inputs of chloride from atmospheric deposition of sea salts near
the ocean may displace nitrate from the soil, resulting in a spatial pattern of higher nitrate
leaching close to the coast (Church et al., In prep), mimicking patterns of chloride (Wigington et
al.. 1998; Church et al.. In prep). Our data show that inputs from point sources or septic systems
appear to be of minimal importance in the basin, relative to the background levels from forested
watersheds. There is little grazing or other agriculture in the basin so we have not been able to
judge their relative importance. Bedrock geology appears to exert strong controls on stream acid
neutralizing capacity, base cation concentrations and sulfate concentrations, and also appears to
be important in controlling stream phosphate concentrations. Most of the
^-v
' i.
'«
ja
•^
y,
js
c
V
sr<
o
£
Z
40 -i —
30 -
20 -
10 -
0 -
Nloss = 24.0x2 + 8.6x+ 1.4
R2 = 0.76 ,
. f'
^ y,»*
+•''
r**
+ *** !--'"
^^* ^^ ^^
A-'*' ^T ^
,,..-%V
0%
20% 40% 60% 80%
Broadleaf plus mixed forest cover (weighted)
100%
Figure 6.1 Watershed N export as a function of broadleaf plus mixed (conifer-broadleaf)
cover, weighted by the slope coefficient for both cover
types. Data are for for 27 streams in the Salmon River basin in
2000. (Compton et al., in press).
streams have inorganic N:P ratios that suggest that phosphorus could limit in-stream
autochthonous production. Larned (In prep.) found very long uptake lengths for both inorganic N
51
-------
and P in two low-order stream reaches in the Salmon River basin. His findings indicate that
neither N nor P were strongly retained by in-stream processes, and suggest that light availability
may be the principal control on in-stream production in this system.
We will test the empirical relationships that we develop among stream nutrients and
watershed characteristics by examining other Coast Range watersheds. The sites where we will
conduct this research include the Integrated Watershed Study sites (Section 4) and the ODFW
juvenile watershed study sites. Annual discharges will be estimated using either available
gauging information for the individual streams or from inference using regional
precipitation/runoff relationships (Section 5, Broad-Scale Analyses). We will use available
CLAMS information as well as available bedrock geology and National Resources Conversation
Service (NRCS) county soil maps to characterize the watersheds. On a seasonal basis, these
streams will be sampled synoptically, and analyzed for streamwater chemistry. Sampling will
follow EMAP protocols (Lazorcheck et al., 1998). All water chemistry analyses will be
conducted according to the "QA Plan for the Willamette Research Station" (Erway et al., 2001).
Samples will be analyzed for nutrients that are expected to play a major role in aquatic
productivity, including ammonium, nitrate, phosphate, total dissolved N, and dissolved organic
carbon. In addition, the following chemical constituents will be used in constructing ionic charge
balances and to serve as indicators for geologic parent material: dissolved inorganic carbon, total
suspended solids, silica, chloride, sulfate, calcium, magnesium, sodium, potassium, acid
neutralizing capacity, pH and specific conductivity. We will compare patterns from these
analyses with our more in-depth analyses on controls in the Salmon River basin to develop
understanding of controls across the Coast Range.
Goal 2B. Compare the relative magnitude of watershed nutrient losses with the nutrient inputs
associated with returning anadromous fish using a biogeochemical mass balance approach.
To investigate this question we will perform mass-balance calculations for C, N and P
for the Salmon River, for which we have detailed information on stream chemistry and flow
throughout tributaries of the system. The data we will use to compute watershed output budgets
for N and P will be discharge and stream chemistry. Discharge for the mainstem of the Salmon
River is measured by the State of Oregon just above the tidal influence. Discharge is measured
at 15-minute intervals at this site. We have measured instantaneous discharge at roughly 28
52
-------
tributary sites throughout the Salmon River basin coincident with our chemical sampling. The
frequency of these chemical sampling and discharge measurements was monthly in 2000 and
every six weeks in 2001. Current and future sampling frequency will be seasonal. From
relationships among sites (and especially with the mainstem gage site) we will be able to infer
patterns and quantities of discharge at the tributaries that we will combine with stream chemistry
measurements to compute output budgets. Preliminary estimates of nutrient losses ranged from 2
to 24 kg N ha"1 yr'1 from 27 of the sub watersheds within the Salmon River basin.
To estimate return of MDNs, we will use 1986-2001 ODFW hatchery return data for the
Salmon River basin. These data will include peak counts for the main stem, and spot counts and
an estimate offish length, sex and species for several other tributaries within the Salmon River
basin. By using literature relationships between fish length and weight, and the known
percentages of C, N and P in returning fish, we can compute an annual return estimate of those
nutrients for comparison to watershed-derived output budgets.
Goal 2C. Determine the role of sea salt inputs and soil exchange processes in coastal stream
nutrient concentrations and watershed budgets.
To investigate this question we will conduct soil and lysimeter studies in the Salmon
River basin. This work will focus on clarifying the role of the upland soil nitrogen stocks and
availability on stream nitrogen dynamics. In order to better define the relative role of exchange
mechanisms vs. direct leaching from red alder stands in controlling stream chemistry, we plan to
conduct soil sampling, column leaching and lysimeter experiments to examine closely the
exchange processes.
Basin-Scale Soil Sampling and Analysis
We will use the CLAMS land cover and NRCS soils data layers for the Salmon basin to
select 20 sites for soil sampling. Based on all available GIS information, we will choose these
sites to represent the watersheds sampled in the basin. Initially, the GIS-based soil map will be
used in combination with aerial photos and GIS-based vegetation cover to select 20 sites that
represent the range of soils observed throughout the Salmon basin. These 20 sites will be
sampled (O horizon, 0-20 cm, 20-50 cm and 50-100 cm depth increments) using a quantitative
corer. We will sample three replicate locations within a site. The samples will be dried and
53
-------
analyzed for C and N by automated combustion techniques (Carlo-Erba Instruments) and 15N
isotope ratios by isotope ratio mass spectrometry (Finnigan Delta Plus), both housed in our
laboratory facilities. A subset of these soil samples will be used fresh in the soil column leaching
experiments (see below). We will use the isotope ratios as an integrator of net ecosystem N
inputs and losses (e.g., Robinson, 2001).
Soil Column Leaching
Soils from four of the sites described above will be used to examine the exchange
characteristics of nitrate and chloride in these soils. These exchange columns will be leached
with three levels of chloride (representing a range of atmospheric inputs) and 2 levels of nitrate
(representing a range of soil solution concentrations). After the initial 2-month period, a subset of
the columns will be leached for longer periods. Samples will be provided at approximately
monthly intervals in groups of 48, beginning in May 2002.
Lysimeter Studies
In fall 2001, approximately 30 lysimeters were installed in an old-growth Sitka spruce stand
within Cascade Head Experimental Forest (Salmon River Basin). The lysimeter solutions will be
collected twice per season until fall 2002, when we will begin chloride additions. Several levels
of chloride will be added to the soil surface as NaCl or CaCh in solution once per month, and the
lysimeters will be sampled biweekly for several months. Control plots will receive water only.
Collections from the 30 lysimeters installed for the chloride-nitrate exchange studies will
continue through FY 2003.
6.3 FUTURE DIRECTIONS
Results from all aspects of the work described above for the first two years, as well as
from interconnections with the Integrated Watershed Analyses, Broad-Scale Analyses and Fish
Modeling will guide the future directions of this research.
Further work addressing Question 1 could include additional carcass planting
manipulations and experiments, particularly within study sites of the Integrated Watershed
Studies where a strong knowledge base generated in years 1 and 2 would aid in productive
experimental design and interpretation. We may take advantage of natural variations in salmonid
54
-------
returns across the Integrated Watersheds, in order to examine the effects on stream chemistry and
foods webs.
Further exploration of the relationships among watershed characteristics and stream
nutrient concentrations and fluxes (Goal 2A) might well include incorporation of knowledge
gained within the Integrated Watershed Studies 1o develop additional ways to extrapolate
observed relationships across the Coast Range to other watersheds. Mass-balance comparisons of
the relative magnitude of watershed nutrient losses versus nutrient returns via anadromous fish
(Goal 2B) will be expanded to include data from other basins. Per the approach of Van Sickle et
al. (1997), evaluation will be made of the effect of sampling frequency on estimation of episodic
chemistry and thus budget calculations.
Additional work may also be pursued relative to the roles of sea salt inputs and soil
exchange processes in determining nutrient leaching rates from coastal watershed soils (Goal
2C). Collaborative studies may develop with Swedish researchers in the emerging research area
of chlorine biogeochemistry as a fruitful avenue of research to understand the interactions
between sea salt inputs and nitrogen dynamics in forests of the Oregon Coast Range.
55
-------
7.0 FISH MODELING
7.1 RATIONALE FOR MODELING APPROACH
One of the major components of our research approach is the use of models to simulate
fish response to environmental variability and human disturbance and management. Two models
will be included: an adaptation of an existing simulation model (Nickelson and Lawson, 1998)
that examines the dynamics of coho salmon, and construction of a new model focusing on native
fish assemblages. Specific goals of the salmon and assemblage modeling efforts are discussed in
Sections 7.3 and 7.4, respectively. Fish simulation models are being included in the study for
three main reasons. The first reason is that, from a scientific perspective, simulation models can
serve as an efficient tool to conduct exploratory evaluations of mechanisms hypothesized to
affect fish survival and variability; for example, effects of climate on marine survival, changes in
in-stream habitat quality on abundance, and decreased dispersal, due to barriers, on abundance
and survivability. Given the cost and difficulty of conducting field studies, modeling might be
the only feasible way to address certain important questions. As an example, little is known
about the spatial distribution of coho straying. Past efforts have modeled straying as an
equiprobability function (Nickelson and Lawson, 1998). However, a number of other
mechanisms are also possible. Testing these different mechanisms through field studies would
be difficult, since a large number offish would have to be tracked and their natal spawning
grounds would need to be known. Alternatively, a model could be used to compare these
different mechanisms. The first issue would be to determine if population dynamics vary
significantly by straying mechanism. In other words, is the mechanism of straying a significant
source of variation, compared with other sources? If so, then the analysis could lead to testable
hypotheses - e.g., differences in expected fish distributions - that would provide insight into
which mechanism was operating. Simulation models can also be useful in differentiating
between factors that are correlated at a landscape scale, and which therefore cannot be separated
by regression analyses. This is possible if the factors have distinct mechanisms that are
incorporated into the model.
The second reason for including modeling elements in our study is that models have the
ability to address a wide-range of management issues by allowing various scenarios to be
56
-------
evaluated. For example, models can help address complex questions such as: How do land use
decisions affect yield and sustainability? What is the effect of hatchery fish on wild salmon
stocks? How important are fish carcasses and other nutrient additions as a supplement to marine-
derived nutrients? How does managing for wild salmon impact other native populations? Are
changes in fish stocks due to human actions or natural variations? Models are especially useful
for exploring and comparing the effects of various management options before the actions are
taken. Ultimately it is our intent that the type of questions that can be addressed through our
modeling effort include how natural factors operating at various scales interact with human
actions to affect fish abundances, distribution, and long-term sustainability. The need for such
models has been recognized by various groups. For example, the Committee on Environment
and Natural Resources (2000) stated that "modeling and decision support tools are required to
incorporate land use change relative to habitat on this extensive spatial scale, and must
incorporate temporal changes".
Finally, our decision to include modeling as a part of our study represents a significant
value-added contribution that our group can make to other agencies already involved in salmon
research. A coho salmon model (Nickelson and Lawson, 1998) is currently in use by ODFW and
NMFS for Oregon coastal coho. However, this model is not spatially explicit, has a simplified
representation of straying, and considers winter habitat as the only important type of freshwater
habitat. As a result, a number of critical issues cannot be addressed through this model. Also this
single species model does not address other native populations. Modifying the basic Nickelson-
Lawson model so it is spatially explicit and contains different habitat types and separately
developing an assemblage model are significant contributions that will benefit the wider inter-
agency effort by broadening the scope of our modeling capabilities.
7.2 STREAM NETWORK
We ultimately wish to be able to run both the coho and assemblage models using a
common set of scenarios. This would allow us to assess, for example, how a particular
management decision aimed specifically at coho affected both target and non-target species. To
accomplish this, we will jointly develop software to represent stream networks. This will serve as
a common spatial driver for both modeling efforts (Figure 7.1).
57
-------
7.2.1 Representation of the Stream Network
The spatial driver will have the capability both to generate hypothetical stream networks
with specified characteristics and to import real stream networks from geographic information
Network Structure
Coho
Salmon
Assemblages
.Biological Response/
Figure 7.1 Schematic representation on interaction between stream network and fish
models. HC, is an array representing the habitat characteristics of stream
segment /. Upland effects on HQmay be included in later years.
system (GIS) representations. Tools will be developed for computing graph theoretical properties
of networks that are relevant to this simulation approach.
The stream network will be structured as a directed acyclic graph of nodes and links
(Foulds, 1992), where each node has two or more, but usually only two, upstream links entering,
and one or more, but usually one, downstream link leaving. The nodes will not carry any
58
-------
information relevant to the problem; they will only serve as conceptual connections. Each link
will consist of one or more stream segments, which will represent the environmental aspects of
the network. The segments belonging to a link will be structured as a directed path across the
link. The segments will represent the subdivision of the network into units that are relatively
homogeneous in properties presumed to control fish distributions. These properties will often be
geomorphological attributes of the landscape. For example, in the Oregon Coast Range, we will
use the CLAMS definitions of stream reaches determined by topographical gradient and valley
morphology.
For initial development of the network algorithms, a wide range of hypothetical stream
topologies can be generated with recursive functions using a simple parameterization to obtain
desired branching properties and sizes. The "Q model" (Costa-Cabral and Surges, 1997) is one
such method that incorporates a single parameter that is the ratio of the probability of internal
branching to the sum of the probabilities of internal plus external branching. In the early stages
of development of the model, networks will be generated stochastically according to these
probabilities.
7.2.2 Environmental Properties of Segments
The properties of stream segments will be focused on instream habitat attributes that are
important to fish. Examples include stream order, gradient, valley and stream channel type, flow,
summer maximum temperature, amount of spawning habitat, frequency of pools, mean pool
depth, abundance and type of cover, and physical barriers to fish upstream movement. The
model will be designed to allow varying numbers and types of properties to be considered in a
specific application, so that the habitat attributes incorporated can be tailored to represent those
expected to be most important in controlling fish distributions in a particular system or in
response to a particular type of disturbance.
Properties of stream segments will be expressed as spatial and temporal probability
distributions of either continuous or discrete quantities, depending on the property. Examples of
continuous properties are flow, temperature, gradient, and frequency of pools. Examples of
discrete properties are valley type and cover type. Some properties, such as stream temperature,
for example, may also have their distributions shifted over time to represent natural variability,
59
-------
natural disturbances, and human impacts. Through these changes the effects of different
management activities may be examined.
For Oregon coastal streams, properties for some segments will be available from the
extensive habitat surveys done by ODFW (Moore et al., 1997;
http://osu.orst.edu/Dept/ODFW/freshwater/inventory/index.htm). For segments not included in
these habitat surveys, or for properties that are included but not available from surveys, other
methods of estimating these properties will be used. For example, physical properties of
segments such as gradient and geomorphology may be estimated from maps, air photos, or
digital elevation models.
In the longer term development of coho and assemblage analyses, there is an intention to
include functional connections to upland properties that are relevant to population and
assemblage composition (Figure 7.1). For example, functional predictions of temperature and
large wood inputs may be desirable and feasible as the models evolve, or the models may be
linked to other models that predict in-stream habitat from watershed and landscape features.
7.3 COHO SALMON MODEL
The goal of the coho salmon modeling effort is to develop a flexible, spatially explicit
simulation model that can be used to study and evaluate issues related to habitat use, response to
natural and anthropogenic disturbance, and long-term sustainability of coho salmon, with
particular emphasis on the importance offish movement through the stream network. Given this
capacity, the model will also be useful in exploring and comparing the utility of various human
management options on coho abundance and sustainability. The broad objectives of this work are
to examine: how coho salmon use the stream network during their freshwater life cycle, and how
important the interplay between life history diversity and spatial - temporal patterns in habitat
quality is in sustaining coho salmon populations (Research Question 1); and how human land use
interacts with natural processes at watershed to landscape scales to affect the long-term
sustainability of coho salmon populations in Oregon coastal streams (Research Question 3).
Specifically, this research will focus on three areas:
(1) Examine the long-term response of coho salmon to natural disturbance cycles, both
marine and freshwater, focusing in particular on local and regional extinction risks.
60
-------
(2) Consider the importance of various types of movement - straying of adult spawners
and among- and within-season habitat movement - to population dynamics and long-term
sustainability. This will include how network configuration and habitat distribution affect
the ability of salmon to disperse and recolonize local extinctions.
(3) Investigate the effect of human disturbances at various scales on population dynamics
and long-term sustainability. In particular, this will focus on the direct effects of human
disturbance, through harvesting and habitat loss, as well as indirect effects on the ability
of salmon to respond to natural disturbance regimes (i.e., through recolonization).
Barriers to movement (e.g., culverts) will be included as a disturbance.
7.3.1 Approach
The basic approach for salmon simulation modeling will be to modify a stochastic life-
cycle model developed by Nickelson and Lawson (1998) that has been used to assess the risk of
extinction for coho in the Oregon coast (Figure 7.2). The model incorporates basic life cycle
information, including survival rates for different life history stages, density dependence, and
genetic effects of small spawner populations. The Nickelson-Lawson model also includes a
number of sources of stochastic variability, e.g., marine cycles and climate-induced changes in
freshwater habitat quality. This model allows recolonization of reaches that have experienced
local extinctions through the use of a straying function. However, the Nickelson-Lawson model
has several significant limitations. First, it represents straying in a simplistic fashion: strays from
a given reach are equiprobably distributed to all other reaches. A number of alternative
mechanisms are possible and probably more realistic. For example, fish might have a higher
likelihood of straying near their natal spawning ground, which is equivalent to assuming that
their fidelity occurs at a scale larger than the stream reach. This would result in a two-tailed
Gaussian distribution around the natal stream reach. Or if fish began to run out of energy as they
swam to their natal reach, straying might occur as a one-tailed Gaussian. A simple model for
Atlantic salmon (Salmo salar) demonstrated that the way in which dispersal to spawning habitat
is represented can affect both parr production and extirpation probability (Kocik and Ferreri,
1998). Alternative representations of coho straying may have similar results. However, the
original Nickelson-Lawson model could not include such alternative straying mechanisms
because the model does not include a spatial representation of the stream network. Therefore the
spatial relationship between different stream reaches - and distances between them - are not
61
-------
known. The model has been modified so that it can incorporate such spatial information, thereby
allowing alternative straying mechanisms to be implemented (Peter Lawson, National Marine
Fisheries Service, pers. comm.). This modified version of the model works with input of specific
GIS maps; however, it does not allow for generation of random networks (see "Stream Network"
discussion above). Also, simulations with this new model are proceeding extremely slowly
because of new research priorities.
Smolts
(OvtmtntMl
1 Swvlvtf ^
Summer
Parr
o
Eggs
Spawners
Stream Reach /
Adults
Ocean
Figure 7.2 Schematic representation of habitat-based life cycle model developed for coho
salmon (Nickelson and Lawson, 1998)
Secondly, the Nickelson-Lawson model assumes that overwintering habitat is limiting,
and therefore it is the only habitat type included. While overwintering habitat may be in smallest
supply, it is possible that existing overwintering habitat may not be used to capacity if it is not
accessible from summer rearing habitat. More generally, a seasonal habitat type can only be used
if the different seasonal habitat types are distributed in such a way that movement between them
62
-------
is not limiting. Kocik and Ferreri (1998) provide an example of this, showing that Atlantic
salmon have access to more nursery habitat when spawning habitat is broken up into two smaller
regions rather than one larger area. The issue of movement between seasonal habitat must not be
viewed statically, since these properties can dynamically change in response to natural and
anthropogenic disturbance. The spatial juxtaposition of different seasonal habitats and movement
between them cannot be considered by the original or modified Nickelson-Lawson models. A
more subtle nuance related to this characteristic of the model is that habitat fidelity is assigned to
overwintering habitat, rather than spawning habitat. This makes the implicit assumption that all
individuals hatching from the same spawning habitat use overwintering habitat in the same
reach, which is a questionable and unnecessary assumption.
Finally, the Nickelson-Lawson model limits smolt abundance by a reach-specific,
maximum smolt capacity. This means that fish are eliminated if their abundance exceeds the
capacity of a reach, even if nearby reaches contain under-utilized habitat. Exclusion of this kind
of movement introduces an artifact that produces lower abundance than would be expected if
supplemental habitat could be exploited. Including this kind of between habitat movement
should allow for more realistic estimates of abundance. More importantly, this kind of movement
could represent another mechanism that could allow for recolonization following local
extinctions. Consider the following scenario: A stream reach with high quality habitat is
unoccupied because of a local extinction. This allows juveniles from another area to move into
that habitat if their local habitat was limiting. If imprinting happens after this movement occurs,
then these individuals would return to this new stream reach as adult spawners. However, if coho
fidelity is strictly based on imprinting to the spawning habitat where they hatched from, the
adults would return to the original spawning grounds and the supplemental habitat would remain
uncolonized. These two mechanisms could be compared if the model allowed movement
between the same seasonal habitat types.
All three of the limitations discussed above concern different types of coho movement.
Addressing these limitations will require that the basic model be changed in such a way that
more realistic movement - straying of adult spawners and among- and within-season habitat
movement - can be included. This will require a spatially explicit model that includes all
seasonal habitat types. A number of spatially explicit models have been developed for salmonids
(e.g., Bartholow et al., 1993; Jager et al., 1997; Railsback and Harvey. In press). However, we
63
-------
believe it is better to modify the Nickelson-Lawson model, so that it is spatially explicit, rather
than use one of these models for two main reasons: First, none of these other models have been
applied to coho salmon in the coastal Oregon ESU, as is the case for Nickelson-Lawson. We
believe that using a model parameterized to the target population of interest far outweighs the
advantages of using a pre-existing, spatially explicit model developed for a different population.
Secondly, both Nickelson and Lawson work in the Corvallis area, making it much easier to
collaborate with them on this project. As an example of this collaboration, they have both been
consulted with in developing EPA's study plan.
In practical terms, modifying the complex code of the Nickelson-Lawson model would
be more difficult than developing a new model from scratch. Thus the basic biological structure
of the Nickelson-Lawson model will be incorporated into a new model that includes the
following modifications:
• A spatially explicit framework that allows for various types of movement (adult straying,
between season habitat, and within season habitat);
• Various mechanisms for spatially explicit straying of adult spawners, including inter-
basin straying;
• Inclusion of spawning and rearing habitat in addition to overwintering habitat, and
movement between these different seasonal habitats; and
• Removal of the maximum smolt capacity limit and inclusion of the ability to move into
underutilized habitat.
For the sake of simplicity, survival rates will be standardized by dividing by an indexed
survival rate (either a maximum observed value or values from some benchmark year). We can
then easily link changes in these indexed survival rates to stochastic variations in marine and
freshwater conditions - both natural and anthropogenic. The model will calculate the relative
proportion of capacity that is occupied in addition to abundance; i.e., abundance divided by the
capacity of that reach, where capacity is a dynamic value that annually varies with freshwater
conditions. This is useful for two reasons. First, the abundance to capacity ratio can be used to
trigger movement to search out supplemental habitat. This would occur when the potential
abundance exceeds capacity, i.e., when the ratio is greater than one. Secondly, an observation of
low coho abundance is ambiguous from a management perspective, because it is not known if
64
-------
this occurred due to demographic reasons (poor recruitment or freshwater survival) or due to
habitat constraints. The proportion of capacity that is occupied is therefore less ambiguous, and
tells us how much of the habitat potential is being utilized. This ratio will be modeled as a
function of the ratio for the previous generation, the inherent habitat capacity, and marine and
freshwater variability.
7.3.2 Model Evaluations and Experiments
Sensitivity Studies
The behavior of the model will first be explored through a series of sensitivity analyses.
This will make use of Monte-Carlo simulations and randomly generated stream networks to gain
an understanding of the model behavior and the various processes. The following permutations
will be included:
! Case 1: Base case. This is meant to represent the Nickelson-Lawson model
(overwintering habitat only) without straying. In this case all movement between
freshwater habitat and the marine environment occurs with complete fidelity. Survival
rates for and movement between other seasonal habitat types will be set to one to
represent conditions where overwintering habitat is limiting. Only natural sources of
variability are included.
! Case 2: Base plus straying. Add various straying mechanisms, including equiprobability
(equivalent to the Nickelson-Lawson model), two-tailed distance weighted (probability of
straying decreases both upstream and downstream with distance from natal spawning
ground), and one-tailed distance weighted (only straying downstream of natal spawning
ground).
! Case 3: Base case with among-season habitat movement survival rates. This will allow
us to investigate the effect of seasonal habitat distribution on coho dynamics. In this case
survival rates for and movement between other seasonal habitat types will vary between
zero and one, so that overwintering habitat is not automatically limiting (survival across a
habitat unit type will vary by season).
! Case 4: Base case with within-season habitat movement. Allows fish to seek out
underutilized habitat in adjacent reaches if the abundance to capacity ratio exceeds one.
Survival stays constant across a habitat unit type within season. Will include two
scenarios: fidelity remains with old stream reach and fidelity switches to new stream
reach.
65
-------
! Case 5: Base case with all movement. Combines straying (Case 2), among-season habitat
(Case 3), and within-season habitat (Case 4) movements.
Effects of Anthropogenic Stress
Repeat Cases 1-4 and add various sources of anthropogenic stress (random and spatially
autocorrelated, both acute and chronic). Barriers to movement (e.g., culverts) will be included as
a form of anthropogenic stress.
Lowland/Estuarine Habitat
There has been significant scientific focus on the effects of upstream habitat loss on coho,
and a great deal of effort has gone into restoring this habitat. However, there has been relatively
little research examining the effect of lowland and estuarine habitat loss on these populations,
even though this could represent a significant impact (Tschaplinski, 1988; Miller and Sadro,
2000; Cornwell et al., 2001). We will conduct an initial set of exploratory simulations to examine
and compare the importance of this habitat relative to upstream habitat.
7.3.3 Future Directions
Integrated Watersheds
The model will be calibrated using data characterizing the network and habitat of the
West Fork Smith River and possibly other integrated watersheds (Section 4). Information
obtained from PIT tagging will be used to assign movement values between various seasonal
habitat types. Straying data obtained on the Smith River from ODFW tagging studies may also
be used. Model results will be compared with actual coho data. Conduct simulations to assess
long-term sustainability of current (i.e., impacted) environment. Compare with pre-impact
conditions and various management scenarios (e.g., removal of culverts or habitat restoration).
Lowland/Estuarine Habitat
Contrast model results for a watershed having relatively intact lowland/estuarine habitat,
e.g., Winchester Creek, with those from a watershed having more impacted lowland/estuarine
66
-------
habitat (e.g., West Fork Smith River), using data from the integrated watershed studies (Section
4). This effort will be contingent on the results of the exploratory analyses described above.
Modifications
The model will be modified to include temperature, discharge, and/or nutrient effects,
contingent upon the findings of other project elements (Sections 4 - 6).
Combined Coho/'Assemblage Modeling Effort
Coho and assemblage model will be run using common scenarios, e.g., for same
calibrated stream network using the same management scenarios. Assess the combined effect of
various management scenarios on both coho and native populations.
7.4 SIMULATION OF FISH ASSEMBLAGES
This part of the research will explicitly simulate a set of mobile fish species occupying a
stream network, where sub-populations of a species may be colonizing new segments of the
network if their attributes match the environmental attributes of the segment, possibly replacing
existing species if their match is better, possibly suffering local extinction, and possibly
migrating to neighboring segments (Figure 7.3). The objects of analysis will be sub-populations,
however, rather than individual organisms. Sub-populations will be defined as all individuals of
a species occupying a segment of the network. The objective is to predict the assemblage offish
species expected to occur in each stream segment and pattern of fish species presence/absence
within the stream network as a function of the environmental attributes of the network.
No simulation model currently exists for predicting changes in entire fish assemblages in
large stream networks. Thus, the first several years of effort will be exploratory, to determine
whether a model can be developed that reasonably mimics fish assemblage responses based on
model parameters and inputs that are realistic to obtain. We will begin with a simple,
hypothetical network, a small pool of species, and relatively simple model structure. If
promising, the work will evolve into the use of Oregon coastal networks and species pools.
While initial model applications will occur in the Oregon Coast Range, the model will be
designed for broader use. The ultimate objective is to develop a model that can be applied to real
67
-------
stream networks to address questions about the potential long-term impacts of human
disturbance at watershed to landscape scales.
7.4.1 Research Questions
Applications of the proposed model will help address two research questions: How do
native fish use the stream network, and how important is the interplay between life history
diversity and spatial - temporal patterns in habitat quality in sustaining native fish assemblages
Assemblages
Competition
Predation
Species Replacement'
Local Extinction
CD TS
O .25
0> |Q
5- (D
Species
b
,t!
s
pecies A
DeciesC
Temperature
Figure 7.3 Schematic representation of assemblage model which will predict
presence/absence of native fish sub-populations by stream reach.
68
-------
(Research Question 1)? And how does human land use interact with natural processes at
watershed to landscape scales to affect the long-term sustainability of native fish assemblages in
Oregon coastal streams (Research Question 3)? Examples of more specific questions the model
will be designed to address are:
1. What is the space/time pattern of effects on fish assemblages of habitat disturbance?
a. What are the effects of network wide changes such as increased temperature
(due to climate change, for example)?
b. What are the effects of local changes (due to logging, toxics, or other habitat
alteration, for example)?
c. How do changes with different spatial extent interact with different time
durations of changes?
2. What are the effects of insertion or removal of partial or complete barriers to fish
passage?
3. What are the effects of introduction of new (exotic, for example) species to the
species pool?
Such applications of the model will occur only in years 3-5 of the project. In years 1-2, efforts
will focus on model development and evaluation, for example:
1. How does the simulation model predict the distribution of species compared to
empirical data and models?
2. Which parameters in the simulation model appear to have the greatest effect on
performance?
7.4.2 Representation of Fish Species and Sub-populations
For any network being modeled, there will be a pool of species that are presumed to be
available to occupy the network. A regional pool of species for basins appears to be a reasonable
starting point for investigating patterns over a basin (Matthews, 1998). Species pools will be
developed from empirical data with possible additions from expert opinion. For Oregon coastal
streams, for example, we will estimate the composition of the pools from EMAP/DEQ data, from
this project's snorkel surveys, and from other sources.
The unit of a species that occupies a segment will be called a sub-population. Initially,
all sub-populations of a species will have the same attributes. This specification could be relaxed
69
-------
as the modeling evolves to simulate local variations or shifts in habitat use resulting from inter-
species interactions (e.g., White and Harvey, 2001).
Habitat Suitability
Species requirements will be conceptualized as habitat suitability functions that operate
on the habitat attributes of stream segments (e.g., gradient, mean pool depth, temperature; see
earlier section on stream network). Each species will have a unique function that determines its
suitability on a scale of 0-1 for a segment. Random draws from the probability distributions of
segment properties will be the inputs to the suitability functions.
There is a rich literature on approaches for assessing habitat suitability for fish, which we
will build upon in developing the habitat suitability functions for the model. We anticipate
pursuing three alternative approaches. The first will be strictly statistical. Based on available
EMAP and related surveys offish assemblages and habitat in Oregon coastal streams (Merger
and Hayslip, 2000; Rose, 2000), logistic regression will be used to generate models of the
probability offish species presence as a function of instream habitat attributes (e.g., Porter et al.,
2000). The presence or absence of a species in a given stream reach is determined by more than
just habitat suitability, however. Thus, statistical models have significant drawbacks as
surrogates for habitat suitability functions, but are still worth examining as an initial starting
point.
The second approach will follow procedures developed by the U.S. Fish and Wildlife
Service and others for defining habitat suitability indices (e.g., Terrell, 1984; Stoneman et al.,
1996; Terrell and Carpenter, 1997). Based on the combination of field survey data, available
literature, and expert judgment, functional relationships (metrics) will be defined for each habitat
attribute and then the multiple metrics combined into an overall habitat suitability score. While
this approach is highly flexible and takes full advantage of all available information, it is also
highly uncertain and the uncertainty is difficult to quantify.
Thus, in later years of the project (years 3-5), we will also consider a third approach,
using Bayesian techniques to combine multiple sources of information, including both expert
judgment and empirical data, to produce functional relationships with quantified uncertainty
(Reckhow, in press). We will initiate the third approach, however, only if initial exploratory
70
-------
modeling, based on the first and second approaches, suggests that the additional effort is
warranted.
Species Interactions
Competitive and predator-prey interactions among species may also be important in
determining fish assemblage composition and species distributions. Both of these factors will be
represented with a species co-occurrence matrix. Values in the matrix, on a scale of 0 to 1, will
indicate the degree to which the occurrence of one species restricts the co-occurrence of a second
species. Initially, species pairs will be simply ranked as having either zero, small, moderate, or
high degree of negative interaction (e.g., matrix values 0.0, 0.2, 0.5, and 0.7, respectively) based
on the available literature and expert judgment.
Statistical analyses of EMAP and other available survey data will then be used to refine
or replace these relative rankings. For example, for species with similar habitat needs (based on
the available literature), the co-occurrence matrix could be defined based simply on the
frequency of species co-occurrence in habitat surveys. Alternatively, the occurrence of one
species could be included as an additional predictor variable, together with habitat attributes, in
logistic models predicting the probability of presence of a second species. However, habitat and
biological interactions are likely to be highly confounded and difficult to tease apart based solely
on statistical correlations. As for habitat suitability functions, we will initially explore the utility
of expert/literature-based and statistical analyses for defining co-occurrence. If warranted,
Bayesian techniques to quantitatively combine these different sources of information could be
pursued in later years.
Movement Property
Each species will have an intrinsic movement property that will be represented as a
probability distribution of the number of segments that sub-populations are likely to be able to
move in one time step. Species will be grouped into movement classes based on body size and
form. Realistic movement properties for each class will be defined based on available studies of
fish movement in the literature. In later phases of model development, we will also add physical
barriers to fish movement to network properties. The same or separate classification will be used
71
-------
to rank species with regard to their probability of moving upstream over low, moderate, or high
barriers.
Sub-population extinction
Species can go extinct from a given stream reach. This local extinction may be caused by
poor habitat or species interactions, as discussed in the next section. However, independent of
any biological interactions or environment effects, species may vary in their likelihood of local
extinction, i.e., their intrinsic extinction rate. For example, species with higher intrinsic growth
rates, r, may have lower intrinsic extinction rates. Initial model applications will assume a
common intrinsic extinction rate for all species. Later applications may vary this rate for
different types of species, e.g., species with "r" vs. "K" life history strategies.
Future model enhancements (years 3-5) could include life cycle stages, such as spawning,
juvenile, and adult, matched to seasons of the year in which the stages occur. Habitat
requirements could differ from stage to stage.
7.4.3 Simulation of Ecological Processes
Time Units and Initial Conditions
The time step for the simulations will initially be one year. In future development, sub-
year or seasonal time steps may be introduced to capture important life cycle stages.
Initial conditions may be of two types. A current conditions initialization would stock
species in the network according to how they are believed to be distributed. A tabula rasa
initialization would allow species to colonize the network from the entrance.
Migration
Each species will attempt to migrate to neighboring segments at each time step. A
random draw will be made from the probability distribution of movement for the species and a
sub-population of the species will be placed in potential colonization status for the subsequent
time step in the neighboring segment. The movement probability distributions will be either
symmetric or asymmetric to allow either equal or non-equal probabilities, respectively, of
upstream and downstream migration. Physical barriers to movement may constrain upstream
72
-------
movement, with the probability offish passage over the barrier dependent on the height of the
barrier and mobility class of the species.
Colonization
A sub-population that has arrived at a segment can remain for at least one time step if its
probability of occupancy is greater than a system-wide threshold, initially set to 0.5. The
probability of occupancy for a species' sub-population will be calculated as its habitat suitability,
minus the sum of the co-occurrence probabilities of all other species currently occupying the
segment, minus the species saturation for the segment. Species saturation will be calculated as
the ratio of the current number of species to the maximum predicted number of species for the
segment. Maximum richness will be computed from segment properties, predominately
indicators of stream size, at each time step, and calibrated with empirical data (surveys offish
assemblages in coastal streams) or determined by expert opinion. Initially, maximum richness
will be determined by stream order and set to one higher than the maximum observed species
richness in any reach of that order within the area being modeled.
Replacement
Species arriving at a stream segment also have the potential to replace species already
resident in the segment from the previous time step. The probabilities of occupancy will be
determined for each eligible species (resident plus all species migrating in) with the same
formula as in colonization with one modification. The probability of occupancy for resident
species will be increased slightly (small multiplier), with the degree of increase a function of the
number of years of prior residency. For species with probabilities of occupancy below the
system-wide threshold (initially 0.5), the species with the lowest value will be deleted.
Probabilities of occupancy will then be recalculated, and the species with the lowest value (of
those below the threshold) will again be deleted. This process will be repeated until all remaining
species have probabilities of occupancy above the system-wide threshold. All these species will
remain in the segment unless the number of species is greater than the estimated maximum
richness. In this case, species will be eliminated starting with those whose probabilities of
occupancy are lowest. Ties will be broken randomly.
73
-------
Local Extinction
At any time step, a species will die out in a segment if its intrinsic extinction rate divided
by its habitat suitability is greater than a system-wide extinction threshold.
7.4.4 Model Evaluations
Model evaluation approaches will include sensitivity studies and field data comparisons.
Sensitivity Studies
There are many parameters proposed for this model. It will be important to understand
the behavior of these parameters and how sensitive possible outcomes are to them. Drechsler
(1998) has a useful discussion on studying the sensitivities of complex models, with particular
attention to population biology models and models with non-linear dynamics. The model
proposed in this research will certainly exhibit non-linear behavior due to the interactions of
species with the topology of the network. Drechsler's "sensitivity analysis of sensitivity
analyses" may help to focus the tuning and initial experimentation of the model in order to learn
how it is behaving.
Another approach to model assessment is that of Reynolds and Ford (1999). Their
"Pareto Optimal Model Assessment Cycle" uses genetic algorithms and related techniques to
explore the parameter space of a model in the attempt to identify parameterizations that satisfy
an optimal number of assessment criteria. Assessment criteria may be empirical data,
theoretically derived criteria, or criteria determined in some other way. As a method for
systematically investigating proposed model parameters against multiple measures of
performance, these ideas have appeal for the work proposed here.
Comparisons to Field Data
Model outputs will be compared to available field data to evaluate whether the model is
producing reasonable results. Such comparisons require measures offish assemblage
composition at multiple locations within a stream network together with measures or estimates of
habitat attributes throughout the network. Data of this type are available for the Tillamook and
Kilchis watersheds in coastal Oregon (Rose, 2000) and will be collected in the integrated
74
-------
watershed component of this project in West Fork Smith, Winchester, and other watersheds
selected in future years (see Section 4). Examples of the types of data-model comparisons
planned are (1) does the longitudinal pattern of species richness predicted by the model mimic
that observed, (2) are the predicted and observed species co-occurrence matrices significantly
correlated, and (3) for individual species, are the correlations between fish presence/absence and
habitat attributes similar in the predicted and observed data sets.
We are also interested in comparing the simulation model to simpler reach-scale
regression models. As described in Section 5, regression models will be developed using EMAP
and other available surveys offish assemblages in Oregon coastal streams, to predict individual
species presence/absence, species richness, and major assemblage groups as a function of in-
stream and landscape features. We will then apply both regression and simulation models within
the same watersheds (Tillamook, Kilchis, West Fork Smith, Winchester) to compare model
predictions to observed fish assemblage patterns.
7.4.5 Future Directions
If model development efforts in years 1-2 are promising, the model will be applied to
evaluate fish assemblage responses to human alterations and several model extensions will be
explored in years 3-5.
Effects of Human Alterations
Two major human effects that will be simulated in future developments are species
introductions and habitat loss or alteration. Species introductions can be simulated by adding a
species to the species pool and introducing a population either at the entrance to the network or
to some random link. Exotics may often be at the generalist end of the habitat utilization
spectrum, but not always. Experiments will look at whether exotic effects on assemblages occur
and what the effects are. Data on fish assemblage responses to nonindigenous Sacramento
pikeminnow in coastal streams of northern California (Harvey et al., 2002) could be used to
evaluate the reasonableness of model outputs.
Alteration of environmental conditions can take many forms. A degradation of habitat
can be simulated by narrowing and/or shifting the probability distributions for properties of
75
-------
segments. Change in climatic conditions can be simulated by shifting temperature distributions
upward. Habitat changes can be changed over the entire network, or only in portions to simulate
more local effects.
Extensions to the Model
Two ways of extending the model, if it proves successful, would be adding
representations of lakes and ocean connections to the stream network. Lakes and oceans would
be conceptualized, initially, as segments in the network with (1) special habitat properties, and
(2) constricted connections to neighboring segments. Prior to such modifications, only the
freshwater component of the life cycle of anadromous species will be modeled. By adding
oceans, movements between freshwater and ocean environments can be more explicitly
addressed, as well as potential straying of subpopulations between basins connected only via the
ocean or estuaries.
76
-------
8.0 Project Integration
As we have described in sections 3-7, we are proposing a range of research activities
including watershed-based field studies, broad-scale empirical analyses, and modeling to address
our project goal. The purpose of this section is to show how these research components
collectively address our project goal and research questions during the five year duration of the
project. In review, the overall goal of this project is:
To quantify the influence of human and natural disturbances at landscape and watershed
scales on salmon populations and native fish assemblages in Oregon coastal streams.
To accomplish this goal, we have defined three research questions to guide our research.
1. How do coho salmon and other fish use the stream network during their freshwater
lifecycle? How important is the interplay among fish distributions, movement patterns,
and spatial - temporal patterns of habitat quality in sustaining coho populations and
native fish assemblages?
2. What roles do nutrients, temperature, and flow play, relative to physical habitat, in
determining coho salmon freshwater survival and growth? How do these factors
influence fish assemblage structure?
3. How does human land use interact with natural processes at watershed to landscape
scales to affect the long-term sustain ability of coho salmon populations and native fish
assemblages in Oregon coastal streams?
Simplified versions of the project goal and research questions are show in Figure 8.1.
Project Goal and Research
Quantify Influence of Disturbance
Q1. Use of stream network habitat?
Q2. Roles of environmental factors?
Q3. Land use - natural processes
Figure 8.1 Diagrammatic representation of project goal and research questions.
77
-------
Each of the research activities described in this plan contribute directly to accomplishing
our research goal and to addressing our research questions. Figure 8.2 shows how the research
described in sections 4-7 contributes to each of the research questions.
Contributions of Project Components to Research Questions
Q1. Use of stream network habitat?
Integrated Watershed Study (Section 4)
Fish Modeling (Section 7)
Q2. Roles of environmental factors?
Integrated Watershed Study (Section 4)
Broad-Scale Analyses (Section 5)
Role of Nutrients in Salmon Habitat (Section 6)
Fish Modeling (Section 7)
Q3. Land use - natural processes interactions/influence?
Broad Scale Analyses (Section 5)
Role of Nutrients in Salmon Habitat (Section 6)
Research to be developed in years 3 - 5
Figure 8.2 Summary of contributions of research components to project research
questions.
This research project is part of a larger nationwide NHEERL aquatic stressors research
program designed to provide the scientific basis for assessing the role of essential habitat in
maintaining healthy populations of fish, shellfish, and wildlife and the ecosystems on which they
depend (NHEERL, 2002). As part of this overall effort, a number of project deliverables (annual
performance measures - APMs) have been established (Figure 8.3). This project will supply the
Pacific Northwest portion of APM FY07, as well as the entirety of the other 3 APMs.
78
-------
Aquatic Stressors
Framework and Implementation Plan for Effects Research
APM FY03 Prototype watershed-stream network modeling
approach for Pacific salmon
APM FY04 Report characterizing the relationship between
habitat in stream networks and salmon and native
fish for coastal Oregon watersheds
APM FY05 Develop indices of watershed integrity based on
land use/land cover and relationships to fish
APM FY07 Regional models of landscape influence of
salmon/native fish in the Pacific Northwest and
native fish in Great Lake coastal wetlands
Figure 8.3 Project deliverables (annual performance measures - APMs) under the
NHEERL Aquatic Stressor Research effort.
Figure 8.4 illustrates the timing and development of our research project. This plan
describes research that will be conducted over a five year period. It provides the greatest detail
on work that will be completed during the first two years of the project. After two years, we will
evaluate our research results and refine work to be conducted during years 3 to 5 to address our
research questions. During years 3 -5 we expect there will be an increased emphasis on
79
-------
establishing and quantifying linkages between watershed attributes, explicitly including
nutrients, and important stream habitat features identified during the first two years of the
project. We also foresee that years 3 to 5 will be an important time for the incorporation of
results from our field and broad-scale analyses into salmon and fish assemblage models and the
development of models that can be used for assessments on a coastal Oregon scale. Figure 8.5
explains the source of information for each of the project APMs.
Project Critical Path
2003
2004
2005
2006
2007
Integrated Watershed Study
Broadscale Analyses
Fish model development
Nutrients
t
ARM FY03
m
tu
c
0)
l-»
0
^
Tl
a
**•
r
< APM FY04 APM FY05 ARM FY07
III 1
In
Integrated Watershed Study
Watershed - Nutrient Influences on Habitat
Incorporation of Field Results into Models
Formulation of Regional Models
Figure 8.4 Timing of major research activities annual performance measures.
80
-------
Contributions of Project Components to APMs
ARM FY03
Fish Modeling (Section 7)
APM FY04
Integrated Watershed Study (Section 4)
Broad-Scale Analyses (Section 5)
Role of Nutrients in Salmon Habitat (Section 6)
APM FY05
Integrated Watershed Study (Section 4)
Broad-Scale Analyses (Section 5)
Role of Nutrients in Salmon Habitat (Section 6)
Fish Modeling (Section 7)
APM FY07
Integrated Watershed Study (Section 4)
Broad-Scale Analyses (Section 5)
Role of Nutrients in Salmon Habitat (Section 6)
Fish Modeling (Section 7)
Figure 8.5 Contribution of project components to annual performance measures (APMs
See Figure 8.3).
81
-------
9.0 NOTICE
The information in this document has been funded wholly by the U.S. Environmental Protection
Agency. It has been subjected to review by the National Health and Environmental Effects
Research Laboratory's Western Ecology Division and approved for publication. Approval does
not signify that the contents reflect the views of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
82
-------
10.0 LITERATURE CITED
Anderberg, M. R., 1973. Cluster Analysis for Applications. Academic Press, London, United
Kingdom.
Angermeier, P. L. and M. R. Winston, 1998. Local vs. Regional Influences on Local Diversity in
Stream Fish Communities in Virgina. Ecology 79:911-927.
Bartholow, J. M., J. L. Laake, C. B. Stalnaker, and S. C. Williamson, 1993. A Salmonid
Population Model with Emphasis on Habitat Limitations. Rivers 4:265-279.
Bauer, S.B. and S.C. Ralph, 1999. Aquatic Habitat Indicators and their Application to Water
Quality Objectives within the Clean Water Act. U.S. EPA Region 10, EAP 910-R-99-014.
Baxter, C., 2002. Fish Movement and Assemblage Dynamics in a Pacific Northwest Riverscape.
Ph.D. Dissertation, Oregon State University, Corvallis, OR. 188 pp.
Bayley, P. B. and J. T. Peterson, 2001. An Approach to Estimate Probability of Presence and
Richness of Fish Species. Transactions of the American Fisheries Society 130:620-633.
Beamish, R., D. Noakes, G. McFarlane, W. Pinnix, R. Sweeting and J. King, 2000. Trends in
Coho Marine Survival in Relation to the Regime Concept. Fisheries Oceanography 9:114-119.
Beechie, T. J., E. Beamer and L. Wasserman, 1994. Estimating Coho Salmon Rearing Habitat
and Smolt Production Losses in a Large River Basin, and Implications for Habitat Restoration.
North American Journal of Fisheries Management 14.
Bell, E., W. G. Duffy and T. D. Roelofs, 2001. Fidelity and Survival of Juvenile Coho Salmon in
Reponse to a Flood. Transactions of the American Fisheries Society 130: 450-458
Bettinger, P. J. Sessions, T. A. Spies, J. Brooks and A. Herstrom, In review. A Landscape
Simulation Model for Coastal Oregon Landscape Analysis and Modeling. Forest Science.
Bilby, R. E. and P. A. Bisson, 1992. Allochthonous Versus Autochthonous Organic Matter
Contributions to the Trophic Support of Fish Populations in Clear-Cut and Old-Growth Forested
Streams. Canadian Journal of Fisheries and Aquatic Sciences 49:540-551.
Bilby, R. and P. A. Bisson, 1998. Function and Distribution of Large Woody Debris. //;: River
Ecology and Management: Lessons from the Pacific Coastal Ecoregion, R. Naiman & R. Bilby,
(Editors). Springer-Verlag, New York, pp. 324-346.
Bilby, R. E., B. R. Fransen, P. A. Bisson and J. K. Walter, 1998. Response of Juvenile Coho
Salmon (Oncorhynchus kisutch) and Steelhead (Oncorhynchus mykiss) to the Addition of
Salmon Carcasses to Two Streams in Southwestern Washington, USA. Canadian. Journal of
Fisheries and Aquatic Sciences 55: 1909-1918.
83
-------
Bilby, R.E., B.R. Fransen, J.K. Walter, C.J. Cederholm and W.J. Scarlett, 2001. Preliminary
Evaluation of the Use of Nitrogen Stable Isotope Ratios to Establish Escapement Levels for
Pacific Salmon. Fisheries 26:6-14.
Binkley, D., P. Sollins, R. Bell, D. Sachs and D. Myrold, 1992. Biogeochemistry of Adjacent
Conifer and Aider-Conifer Stands. Ecology 73: 2022-2033.
Bisson, P. A., G. H. Reeves, R. E. Bilby and R, J. Naiman, 1997. Watershed Management and
Pacific Salmon: Desired Future Conditions. //;: Pacific Salmon and Their Ecosystems: Status and
Future Options, D.J. Stouder, P.A. Bisson & R.J. Naiman, (Editors). Chapman and Hall, New
York, pp. 447-474.
Brazner, J. C. and E. W. Beals, 1997. Patterns in Fish Assemblages from Coastal Wetland and
Beach Habitats in Green Bay, Lake Michigan: A Multivariate Analysis of Abiotic and Biotie
Forcing Factors. Canadian Journal of Fisheries and Aquatic Sciences 54: 1743-1761.
Breiman, L., J. H. Friedman, R. A. Olshen, and C. G. Stone, 1984. Classification and Regression
Trees. Wadsworth International Group, Belmont, CA.
Brown, J. H. and M. V. Lomoiino, 1998. Biogeography. Sinauer Associates, Inc., Sunderland,
Massachusetts, 691 pp.
Burke, J.L., K.K.. Jones, and J.M Dambacher, In prep. Hab-Rate: A Model for Rating Stream
Habitat Quality Based on a Review of Salmonid Life History Requirements. Oregon Department
of Fish and Wildlife, Corvallis Research Laboratory, Corvallis, OR.
Burnett, K. M., 2001. Relationships Among Juvenile Anadrornous Salmonids, Their Freshwater
Habitat, and Landscape Characteristics over Multiple Years and Spatial Scales in Elk River, OR.
Ph.D. Dissertation, Oregon State University, Corvallis, OR, 244 pp.
Burnett, K. M., G. H. Reeves, and S. Clarke, CLAMS Scientific Review: Aquatic Response
Models. U.S. Forest Service, Pacific Northwest Research Station, Unpublished Report.
Bustard, D. R. and D. W. Narver, 1975. Aspects of the Winter Ecology of Juvenile Coho Salmon
(Oncorhvndnts kisulch) and Steelhead Trout (Sahno gairdneri). Journal of the Fisheries
Research Board of Canada 32:667-680.
Cederholm, C. J., R. E. Bilby, P. A. Bisson, T. W. Bumstead, B. R. Fransen, W. J. Scarlett and J.
W. Ward, 1997. Response of Juvenile Coho Salmon and Sleelhead to Placement of Large
Woody Debris in a Coastal Washington Stream. North American Journal of Fisheries
Management 17: 947-963.
Cederholm, C.J., D.H. Johnson, R.E. Bilby, L.G. Dominguez, A.M. Garrett, W.H. Graeber, E.L.
Greda, M.D. Kunze, B.C. Mareot, J.F. Palmisano, R.W. Plotnikoff, W.G. Pearcy, C.A.
Simenstad, and P.C. Trotter, 2000. Pacific Salmon and Wildlife - Ecological Contexts,
Relationships, and Implications for Management. Special Edition Technical Report, Prepared for
84
-------
D.H. Johnson and T.A. O'Neil (Managing Directors), Wildlife-Habitat Relationships in Oregon
and Washington. Washington Department of Fish and Wildlife, Olympia, Washington.
Cederholm, C.J., M.D. Kunze, T. Murota and A. Sibatani, 1999. Pacific Salmon Carcasses:
Essential Contributions of Nutrients and Energy for Aquatic and Terrestrial Ecosystems.
Fisheries 24:6-15.
Chapman, D. W., 1962. Aggresive Behavior in Juvenile Coho Salmon as a Cause of Emigration.
Journal of the Fisheries Research Board of Canada 19:1047-1080.
Church, M.R., J.E. Complon and S.T. Larned, In preparation. Displacement of Nitrate from
Coastal Forest Watersheds by Atmospheric Inputs of Chloride.
Cohen, W. B., T. K. Maiersperger, T. A. Spies, and D. R. Getter, 2001. Modeling Forest Cover
Attributes as Continuous Variables in a Regional Context with Thematic Mapper Data.
International Journal of Remote Sensing 22:2279-2310.
Committee on Environment and Natural Resources (CENR), 2000. From the edge: Science to
support restoration of pacific salmon. Committee on Environment and Natural Resources,
National Science and Technology Council, Washington, DC.
Compton, J. E., M.R. Church, S.T. Lamed, and W.E. Hogsett, Nitrogen Losses from Forested
Watersheds in the Oregon Coast Range: The Role of N2-fixing Red Alder. Ecosystems (in
press).
Cooper, A. B. and M. Mangel, 1999. The Dangers of Ignoring Metapopulation Structure for the
Conservation of Salmonids. Fishery Bulletin 97:213-226.
Cornwall, T. J., D. L. Bottom, and K. K. Jones, 2001. Rearing of Juvenile Salmon in Recovering
Wetlands of the Salmon River Estuary. Oregon Dept. of Fish and Wildlife, Information Reports
2001-05, Portland.
Costa-Cabral, M. C. and S. J. Surges, 1997. Sensitivity of Channel Network Planform Laws and
the Question of Topologic Randomness. Water Resources Research 33:2179-2197.
Cunjak, R. A., 1996. Winter Habitat of Selected Stream Fishes and Potential Impacts from Land-
Use Activity. Canadian Journal of Fisheries and Aquatic Sciences 53 (Supplement l):267-282.
Daly, C., G. Taylor, and W. Gibson, 1997. The PRISM Approach to Mapping Precipitation and
Temperature, 10th Conf. on Applied Climatology, Reno, NV, Amer. Meteor. Soc.. pp. 10-12.
De'ath, G., 2002. Multivariate Regression Trees: A New Technique for Modeling Species-
Environment Relationships. Ecology 83:1105-1117.
Drechsler, M., 1998. Sensitivity Analysis of Complex Models. Biological Conservation 86:401-
412.
85
-------
Elwood, J.W., J.D. Newbold, A.F. Trimble andR.W. Stark, 1981. The Limiting Role of
Phosphorus in a Woodland Stream Ecosystem: Effects of P Enrichment on Leaf Decomposition
and Primary Producers. Ecology 62: 146-158.
Erway, M. M., K. Motter and K. Baxter, 2001. Quality Assurance Plan, Willamette Research
Station Analytical Laboratory. Revision 0. U.S. Environmental Protection Agency, National
Health and Environmental Effects Research Laboratory, Western Ecology Division, Corvallis,
OR.
Everest, F. H., 1973. Ecology and Management of Summer Steelhead in the Rogue River.
Oregon State Game Commission, 7. Portland, Oregon.
Fausch, K.D., 1984. Profitable Stream Positions for Salmonids: Relating Specific Growth Rate to
Net Energy Gain. Canadian Journal of Zoology 62:441-451.
Foulds, L. R., 1992. Graph Theory Applications. Springer-Verlag, New York, 385 pp.
Fresh, K. L., 1997. The Role of Competition and Predation in the Decline of Pacific Salmon and
Steelhead. In: Pacific Salmon and the Ecosystems: Status and Future Options, D.J. Stouder, P.A.
Bisson & R.J. Naiman, (Editors). Chapman and Hall, New York, pp. 245-275.
Frissell, C. A. and S. C. Ralph, 1998. Stream and Watershed Restoration. In: River Ecology and
Management: Lessons from the Pacific Coastal Ecoregion, R.J. Naiman & R.E. Bilby, (Editors).
Springer, New York, New York, pp. 599-624.
Garono, R. and L. Brophy, 2001. Midcoast Sixth Field Watershed Assessment - Final Report.
Prepared for the MidCoast Watershed Council, Newport, OR.
Gregory, S.V., 1983. Plant-herbivore interactions in stream systems. In: J.R. Barnes and G.W.
Minshall (Editors), Stream ecology: Application and testing of general ecological theory. New
York, NY: Plenum Press, pp. 157-189.
Gregory, S. V. and P. A. Bisson, 1997. Degradation and Loss of Anadromous Salmonid Habitat
in the Pacific Northwest. In: Pacific Salmon and the Ecosystems: Status and Future Options, D.J.
Stouder, P.A. Bisson & R.J. Naiman, (Editors). Chapman and Hall, New York, pp. 277-314.
Gresh, T., J. Lichatowich, and P. Schoonmaker, 2000. An Estimation of Historic and Current
Levels of Salmon Production in the Northeast Pacific Ecosystem: Evidence of a Nutrient Deficit
in the Freshwater Systems of the Pacific Northwest. Fisheries 25:15-21.
Groot, C. and L. Margolis, 1991. Pacific Salmon Life Histories. University of British Columbia
Press, Vancouver, B.C.
Haan, C.T., H.P. Johnson, and D. L. Brakensiek (editors), 1982. Hydrologic Modeling of Small
Watersheds. ASAE Monograph 5, American Society of Agricultural Engineers, St. Joseph,
Michigan.
86
-------
Hankin, D. G. and G. H. Reeves, 1988. Estimating Total Fish Abundance and Total Habitat Area
in Small Streams Based on Visual Estimation Methods. Canadian Journal of Fisheries and
Aquatic Sciences 45: 834-844.
Harrell, F. E., Jr., 2001. Regression Modeling Strategies. Springer-Verlag, New York, New
York, USA.
Hartman, G. F. and T. G. Brown, 1987. Use of Small, Temporary, Floodplain Tributaries by
Juvenile Salmonids in a West Coast Rain-Forest Drainage Basin, Carnation Creek, British
Columbia. Canadian Journal of Fisheries and Aquatic Sciences 44: 262-270.
Harvey, B. C., J. L. White, and R. J. Nakamoto, 2002. Habitat Relationships and Larval Drift of
Native and Nonindigenous Fishes in Neighboring Tributaries of a Coastal California River.
Transactions of the American Fisheries Society 131:159-170.
Healey, M. C. and A. Prince, 1995. Scales of Variation in Life History Tactics of Pacific Salmon
and the Conservation of Phenotype and Genotype. In: Evolution and the Aquatic Ecosystem:
Defining Unique Units in Population Conservation Ecosystem, J.L. Nielsen, (Editor). American
Fisheries Society, Bethesda, MD, pp. 176-184.
Herger, L. G. and G. Hayslip, 2000. Ecological Condition of Streams in the Coast Range
Ecosystem of Oregon and Washington. EPA 910-R-00-002, U.S. Environmental Protection
Agency, Region 10, Seattle, Washington.
Herlihy, A. T., 1998. Water Chemistry. In: Environmental Monitoring and Assessment
Programs-Surface Waters: Field Operations and Methods the Ecological Condition of Wadeable
Streams, J.M. Lazorchak, Klemm, D.J., and D.V. Peck, (Editor). EPA/620/R-94/004F. U.S.
Environmental Protection Agency, Washington, DC, pp. 57-66.
Herlihy, A. T., P. Kaufmann, L. Reynolds, J. Li, and G. Robison, 1997. Developing Indicators of
Ecological Condition in the Willamette Basin -- An Overview of the Oregon Prepilot Study for
EPA's EMAP Program. In: A. Laenen and D. Dunnette (Editors). River Quality: Dynamics and
Restoration. Lewis Publishers, CRC Press, Boca Raton, Florida, USA.
Holtby, L. B., 1988. Effects of Logging on Stream Temperatures in Carnation Creek, British
Columbia, and Associated Impacts on the Coho Salmon (Oncorhynchus kisutch ). Canadian
Journal of Fisheries and Aquatic Sciences 45: 502-515.
Holtby, L. B., B. C. Andersen, and R. K. Kadowaki, 1990. Importance of Smolt Size and Early
Ocean Growth to Interannual Variability in Marine Survival of Coho Salmon (Oncorhynchus
kisutch). Canadian Journal of Fisheries and Aquatic Sciences 47:2181-2194.
Holtby, L. B., T. E. MaMahon, and J. C. Scrivener, 1989. Stream Temperatures and Inter-Annual
Variability in the Emigration Timing of Coho Salmon (Oncorhynchus kisutch) Smolts and Fry
and Chum Salmon (O. keta ) Fry from Carnation Creek, British Columbia. Canadian Journal of
Fisheries and Aquatic Sciences 46: 1396-1405.
87
-------
Hughes, R. M, P. R. Kaufmann, A. T. Herlihy, T. M. Kincaid, L. Reynolds, and D. P. Larsen,
1998. A Process for Developing and Evaluating Indices of Fish Assemblage Integrity. Canadian
Journal of Fisheries and Aquatic Science 55:1618-1631.
Jacobs, S., J. Firman, and G. Susac, 2001. Status of Oregon Coastal Stocks of Anadromous
Salmonids, 1999-2000. Monitoring Program Report No. OPSW-ODFW-2001-3, Oregon
Department of Fish and Wildlife, Portland, OR.
Jager, H. I., H. E. Cardwell, M. J. Sale, M. S. Bevelhimer, C. C. Coutant, and W. Van Winkle,
1997. Modelling the Linkages Between Flow Management and Salmon Recruitment in Rivers.
Ecological Modelling 103:171-191.
Johnston, N.T., C.J. Perrin, P.A. Slaney, and B.R. Ward, 1990. Increased Juvenile Salmonid
Growth by Whole-River Fertilization. Canadian Journal of Fisheries and Aquatic Science 47:
862-872.
Kahler, T. H., P. Roni, and T. P. Quinn, 2001. Summer Movement and Growth of Juvenile
Anadromous Salmonids in Small Western Washington Streams. Canadian Journal of Fisheries
and Aquatic Sciences 58: 1947-56.
Kaufmann, P. R., P. Levine, E. G. Robison, C. Seeliger, and D. V. Peck, 1999. Quantifying
Physical Habitat in Wadeable Streams. EPA/620/R-99/003, U.S. Environmental Protection
Agency, Washington, DC.
Keeley, E. R. and J. W. A. Grant, 1995. Allometric and Environmental Correlates of Territory
Size in Juvenile Atlantic Salmon (Salmo salar). Canadian Journal of Fisheries and Aquatic
Sciences 52:186-196.
Kline, T.C.J., JJ. Goering, O.A. Mathisen, P.H. Poe, and P.L. Parker, 1990. Recycling of
Elements Transported Upstream by Runs of Pacific Salmon: I. 815N and 8I3C Evidence in Sashin
Creek, Southeastern Alaska. Canadian Journal of Fisheries and Aquatic Science 47: 136-144.
Kocik, J. F. and C. P. Ferreri, 1998. Juvenile Production Variation in Salmonids: Population
Dynamics, Habitat, and the Role of Spatial Relationships. Canadian Journal of Fisheries and
Aquatic Sciences 55(Suppl. 1):191-200.
Kostow, K., 1997. The Status of Salmon and Steelhead in Oregon. In: Pacific Salmon and the
Ecosystems: Status and Future Options, D.J. Stouder, P.A. Bisson & R.J. Naiman, (Editors).
Chapman and Hall, New York, pp. 145-178.
Kruzic, L. M., D. L. Scarnecchia, and B.B. Roper, 2001. Comparison of Midsummer Survival
and Growth of Age-0 Hatchery Coho Salmon Held in Pools and Riffles. Transactions of the
American Fisheries Society 130:147-154.
Labbe, T. R. and K. D. Fausch, 2000. Dynamics of Intermittent Stream Habitat Regulate
Persistence of a Threatened Fish at Multiple Scales. Ecological Applications 10:1774-1791.
88
-------
Larkin, G.A. and P.A. Slaney, 1997. Implications of Trends in Marine-Derived Nutrient Influx to
South Coastal British Columbia Salmonid Production. Fisheries 22: 16-24.
Lamed, S.T, In preparation. Nitrate and Phosphate Uptake Lengths and Short-term Retention in
Oregon Coast Range Stream Reaches with Contrasting Lithologies and Channel Substrates..
Lawson, P. W., 1993. Cycles in Ocean Productivity, Trends in Habitat Quality, and the
Restoration of Salmon Runs in Oregon. Fisheries 18:6-10.
Lazorcheck, J. M., D. J. Klemm, and D. V. Peck (Editors), 1998. Environmental Monitoring and
Assessment Program - Surface Waters: Field Operations and Methods for Measuring the
Ecological Condition of Wadeable Streams. EPA/620/R-94/004F. U.S. Environmental
Protection Agency, Washington, DC.
Legendre, P., 1993. Spatial Autocorrelation: Trouble or New Paradigm? Ecology 74: 1659-1673.
Legendre, P. and L. Legendre, 1998. Numerical Ecology. Second English Edition. Elesvier,
Amsterdam, The Netherlands.
Levins, R., 1970. Extinction. In: Some Mathematical Questions in Biology, M. Gerstenhaber,
(Editor). American Mathematical Society, Providence, RI, pp. 75-107.
Li, H. W.. C. B. Schreck, C. E. Bond, and E. Rexstad, 1987. Factors Influencing Changes in Fish
Assemblages of Pacific Northwest Streams. In: Community and Evolutionary Ecology of North
American Stream Fishes, W.J. Matthews & D.C. Heins, (Editors). University of Oklahoma Press,
Norman, OK, pp. 193-202.
Lichatowich, J. A., 1987. Habitat Alteration and Changes in Abundance of Coho Salmon
(Oncorhynchus kisntch) and Chinook Salmon (0. tshawytscha) in Oregon's Coastal Streams. In:
Proceedings of the National Workshop on Effects of Habitat Alteration on Salmonid Stocks,
C.D. Levings, L.B. Holtby & M.A. Henderson (Editors). Canadian Special Publication Fisheries
and Aquatic Sciences 105. pp. 92-99.
Maret, T. R., C. T. Robinson, and G. W. Minshall, 1997. Fish Assemblages and Environmental
Correlates in Least-Disturbed Streams of the Upper Snake River Basin. Transactions of the
American Fisheries Society 126:200-216.
Matthews, W. J., 1985. Distribution of Midwestern Fishes on Multivariate Environmental
Gradients, with Emphasis on Notropis lutrensis. American Midland Naturalist 113:331-343.
Matthews, W. J., 1998. Patterns in Freshwater Fish Ecology. Chapman & Hall, New York, 756
pp.
Matthews, W. J., 1999. Patterns in Freshwater Fish Ecology. Chapman & Hall, London.
McCullough, D. A., 1999. A Review and Synthesis of Effects of Alterations to the Water
Temperature Regime on Freshwater Life Stages of Salmonids, with Special Reference to
Chinook Salmon. U.S. E.P.A. Region 10, 910-R-010. Seattle, WA.
89
-------
McCullough, D. A., 1999. A Review and Synthesis of Effects of Alterations to the Water
Temperature Regime on Freshwater Life Stages of Salmonids, with Special Reference to
Chinook Salmon. U.S. E.P.A. Region 10, 910-R-010. Seattle, WA.
Mclntosh, B. A., J. R. Sedell, J. E. Smith, R. C. Wissmar, S. E. Clarke, G. H. Reeves, and L. A.
Brown, 1994. Historical Changes in Fish Habitat for Select River Basins of Eastern Oregon and
Washington. Northwest Science 68:36-53.
McMahon, T. E. and G. F. Hartman, 1989. Influence of Cover Complexity and Current Velocity
on Winter Habitat Use by Juvenile Coho Salmon (Oncorhynchus kisutch). Canadian Journal of
Fisheries and Aquatic Sciences 46:1551-1557.
Messer J. J., R. A. Linthurst, and W. S. Overton, 1991. An EPA Program for Monitoring
Ecological Status and Trends. Environmental Monitoring and Assessment
17(l):67-78.
Miller, B. A. and S. Sadro, 2000. Residence Time, Habitat Utilization and Growth of Juvenile
Coho Salmon (Oncorhynchus kisutch) in South Slough, Coos Bay, Oregon. Preliminary Report
of Findings. Available on-line at http://w\vw.southsloughestuary.com/.
Mobrand, L. E., J. A. Lichatowich, L. C. Lestelle, and T. S. Vogel, 1997. An Approach to
Describing Ecosystem Performance Through the Eyes of Salmon'. Canadian Journal of Fisheries
and Aquatic Sciences 54:2954-2973.
Moore, K., K. Jones, and J. Dambacher, 1997. Methods for Stream Habitat Surveys. Oregon
Department of Fish and Wildlife Information Report 97-4, Oregon Department of Fish and
Wildlife, Portland, OR.
National Health and Environmental Effects Research Laboratory (NHEERL), 2002. Aquatic
Stressors - Implementation Plan For Habitat Alteration Research (review draft). U.S.
Environmental Protection Agency, Office of Research and Development, NHEERL, Research
Triangle Park, North Carolina.
National Research Council (NRC), 1996. Upstream: Salmon and Society in the Pacific
Northwest. National Academy Press, Washington, DC.
Neter, J. and W. Wasserman. 1974. Applied Linear Statistical Models. Richard D. Irwin, Inc.,
Homewood, Illinois.
Nicholas, J. and M. O'Mealy, 2000. The Oregon Plan for Salmon and Watersheds - Annual
Report Summary 2000. The State of Oregon, Salem.
Nickelson, T. E. and P. W. Lawson, 1998. Population Viability of Coho Salmon, Oncorhynchus
kisutch, in Oregon Coastal Basins: Application of a Habitat-Based Life Cycle Model. Canadian
Journal of Fisheries and Aquatic Sciences 55:2383-2392.
90
-------
Nickelson, T. E., J. D. Rodgers, S. L. Johnson, and M. F. Solazzi, 1992. Seasonal Changes in
Habitat Use by Juvenile Coho Salmon (Oncorhynchus kisutch ) in Oregon Coastal Streams.
Canadian Journal of Fisheries and Aquatic Sciences 49: 783-789.
Nielsen, J. L., 1992. Microhabitat-Specific Foraging Behavior, Diet, and Growth of Juvenile
Coho Salmon. Transactions of the American Fisheries Society 121: 617-634.
Ohmann, J.L. and M.J. Gregory, In press. Predictive Mapping Of Forest Composition and
Structure with Direct Gradient Analysis and Nearest Neighbor Imputation in the Coastal
Province of Oregon, USA. Canadian Journal of Forest Resources
Olden, J. D., D. A. Jackson, and P. R. Peres-Neto, 2002. Predictive Models of Fish Species
Distributions: A Note on Proper Validation and Chance Prediction. Transactions of the American
Fisheries Society 131:329-336.
Omernik, J. M., 1987. Aquatic Ecoregions of the Conterminous United States - Map
Supplement. Annual of the Association of American Geographers 77: 118-125.
Peterson, B.J., L. Deegean, J. Helfrich, J.E. Bobbie, M. Hullar, N. Moller, T.E. Ford. A.
Hershey, A. Hiltner, G. Kipphut, M.A. Lock, D.M. Fiebig, V. McKinley, M.C. Miller, Peterson,
B.J., J.E. Hobbie, T.L. Corliss, and K. Kriet, 1983. A Continuous-Flow Periphyton Bioassay:
Tests of Nutrient Limitation in a Tundra Stream. Limnology and Oceanography 28: 583-591.
Peterson, N. P., 1982. Immigration of Juvenile Coho Salmon (Oncorhynchus kisutch} into
Riverine Ponds. Canadian Journal of Fisheries and Aquatic Sciences 39:1303-1307.
Poff, N. L. and J. V. Ward, 1990. Physical Habitat Template of Lotic Systems: Recovery in the
Context of Historical Patterns of Spatiotemporal Heterogeneity. Environmental Management 14:
629-645.
Poole, G. C. and C. H. Berman, 2001. An Ecological Perspective on in-Stream Temperature:
Natural Heat Dynamics and Mechanisms of Human-Caused Thermal Degradation.
Environmental Management 27:787-802.
Porter, M. S., J. Rosenfeld, and E. A. Parkinson, 2000. Predictive Models of Fish Species
Distribution in the Blackwater Drainage, British Columbia. North American Journal of Fisheries
Management 20:349-359.
Prentice, E. F., F. A. Flagg, C. S. McCutcheon, and D. Brastow, 1990. Feasibility of Using
Implantable Passive Integrated Transponder (Pit) Tags in Salmonids. In: Fish-marking
techniques, N.C. Parker, A.E. Giorgi. R.C. Heidinger, D.B. Jester, E.D. Prince & G.A. Winans
(Editors). American Fisheries Society Symposium 7. pp. 317-322.
Pulliam, H. R., 1988. Sources, Sinks, and Population Regulation. American Naturalist 132:652-
661.
91
-------
Quinn, T. P. and N. P. Peterson, 1996. The Influence of Habitat Complexity and Fish Size on
over-Winter Survival and Growth of Individually Marked Juvenile Coho Salmon (Oncorhynchus
kisutch) in Big Beef Creek, Washington. Canadian Journal of Fisheries and Aquatic Sciences
53:1555-1564.
Railsback, S. F. and B. C. Harvey, In press. Analysis of Habitat Selection Rules Using an
Individual-Based Model. Ecology.
Reckhow, K. H., 2003. Bayesian Approaches in Ecological Analysis and Modeling. In: The Role
of Models in Ecosystem Science, C. D. Canham, J. J. Cole, and W. K. Lauenroth (Editors).
Princeton University Press, Princeton, New Jersey (in press).
Reeves, G. H., L. E. Benda, K. M. Burnett, P. A. Bisson, and J. R. Sedell, 1995. A Disturbance-
Based Ecosystem Approach to Maintaining and Restoring Freshwater Habitats of Evolutionarily
Significant Units of Anadromous Salmonids in the Pacific Northwest. //;: Evolution and the
Aquatic Ecosystem: Defining Unique Units in Population Conservation Ecosystem, J.L. Nielsen,
(Editor). American Fisheries Society, Bethesda, MD, pp. 334-349.
Reeves, G. H., F. H. Everest, and T. E. Nickelson, 1989. Identification of Physical Habitats
Limiting the Production of Coho Salmo in Western Oregon and Washington. US Forest Service,
PSGTR-PNW-245.
Reeves, G. H., D. B. Hohler, B. E. Hansen, F. H. Everest, J. R. Sedell, T. L. Hickman, and D.
Shively, 1997. Fish Habitat Restoration in the Pacific Northwest: Fish Creek of Oregon. In:
Watershed Restoration: Principles and Practices, J.E. Williams, C.A. Wood & M.P. Dombeck,
(Editors). American Fisheries Society, Bethesday, MD, pp. 335-359.
Reimers, P. E., 1973. The Length of Residence of Juvenile Fall Chinook Salmon in Sixes River,
Oregon. Research Report of the Fish Commission of Oregon, Volume 4, Number 2. Portland,
OR.
Reimers, P. E., 1973. The Length of Residence of Juvenile Fall Chinook Salmon in Sixes River,
Oregon. Research Report of the Fish Commission of Oregon, Volume 4, Number 2. Portland,
OR.
Reynolds, J. H. and E. D. Ford, 1999. Multi-Criteria Assessment of Ecological Process Models.
Ecology 80:538-553.
Robinson, D., 2001. 6 15N as an Integrator of the Nitrogen Cycle. Trends in Ecology and
Evolution 16: 153-162.
Rodgers, J. D., 2000. Abundance of Juvenile Coho Salmon in Oregon Coastal Streams, 1998 and
1999. Monitoring Program Report No. OPSW-ODFW-2000-1, Oregon Department of Fish and
Wildlife, Portland, OR.
92
-------
Rodgers, J. D., 2001. Abundance Monitoring of Juvenile Salmonids in Oregon Coastal Streams.
Monitoring Program Report No. OPSW-ODFW-2001-1, Oregon Department of Fish and
Wildlife, Portland, OR.
Rodgers, J. D., M. F. Solazzi, S. L. Johnson and M. A. Buckman, 1992. Comparison of Three
Techniques to Estimate Juvenile Coho Salmon Populations in Small Streams. North American
Journal of Fisheries Management 12: 79-86.
Roni, P., T. J. Beechie, R. E. Bilby, F. E. Leonetti, M. M. Pollock, and G. P. Pess, 2002. A
Review of Stream Restoration Techniques and a Hierarchical Strategy for Prioritizing
Restoration in Pacific Northwest Watersheds. North American Journal of Fisheries Management
21:1-20.
Roni, P. and A. Fayram, 2000. Estimating Winter Salmonid Abundance in Small Western
Washington Streams: A Comparison of Three Techniques. North American Journal of Fisheries
Management 20: 683-692.
Rose, C. E., 2000. Environmental Variables as Predictors of Fish Assemblages in the Tillamook
Basin, Oregon. M.S. Thesis, Oregon State University, Corvallis, OR.
Sandercock, F. K., 1991. Life History of Coho Salmon (Oncorhynchus kisutch). In: Pacific
Salmon Life Histories, C. Groot & L. Margolis, (Editors). University of British Columbia Press,
Vancouver, B.C., pp. 397-445.
Schlosser, I. J. and P. L. Angermeier, 1995. Spatial Variation in Demographic Processes of Lotic
Fishes: Conceptual Models, Empirical Evidence, and Implications for Conservation. In:
Evolution and the Aquatic Ecosystem: Defining Unique Units in Population Conservation, J.L.
Nielsen, (Editor). American Fisheries Society, Bethesda, MD, pp. 392-401.
Schuldt, J.A. and A.E. Hershey, 1995. Effect of Salmon Carcass Decomposition on Lake
Superior Tributary Streams. J. N. Am. Benthological Soc. 14: 259-268.
Sedell, J. R., G. H. Reeves, F. R. Hauer, J. A. Stanford, and C. P. Hawkins, 1990. Role of
Refugia in Recovery from Disturbances: Modem Fragmented and Disconnected River Systems.
Environmental Management 14:711-724.
Shuter, B. J. and J. R. Post, 1990. Climate, Population Viability, and the Zoogeography of
Temperate Fishes. Transactions of the American Fisheries Society 19:314-336.
Sleeper, J. D., 1994. Seasonal Changes in Distribution and Abundance of Salmonids and Habitat
Availability in a Coastal Oregon Basin, Oregon State University, Corvallis, OR. 55 pp.
Solazzi, M. F., S. L. Johnson, B. Miller, and T. Dalton, 2001. Salmonid Life-Cycle Monitoring
Project 2000, Monitoring Program Report No. OPSW-ODFW-2001-2, Oregon Department of
Fish and Wildlife, Portland, OR.
93
-------
Solazzi, M. F., T. E. Nickelson, S. L. Johnson, and J. D. Rodgers, 2000. Effects of Increasing
Winter Rearing Habitat on Abundance of Salmonids in Two Coastal Oregon Streams. Canadian
Journal of Fisheries and Aquatic Sciences 57: 906-914.
Spies. T. A., G. H. Reeves, K. M. Burnett, W. C. McComb, K. N. Johnson, G. Grant, J. L.
Ohmann, and P. Bettinger. In press. Assessing the Ecological Consequences of Forest Policies in
a Multi-Ownership Province in Oregon. In: J. Liu and W. W. Taylor (Editors). Integrating
Landscape Ecology into Natural Resource Management. Cambridge University Press.
Stockner, J.G., E. Rydin, and P. Hyenstrand, 2000. Cultural Oligotrophication: Causes and
Consequences for Fisheries Resources. Fisheries 25: 7-14.
Stoneman, C. L., M. L. Jones, and L. Stanfield, 1996. Habitat Suitability Assessment Models for
Southern Ontario Streams: Model Development and Evaluation. SSC-FS 97-4/2345, Ontario
Fisheries and Habitat Management Branch, Burlington, Ontario.
Sullivan, K., D. J. Martin, R. D. Cardwell, J. E. Toll, and S. Duke, 2000. An Analysis of the
Effects of Temperature on Salmonids of the Pacific Northwest with Implications for Selecting
Temperature Criteria. Sustainable Ecosystems Institute, Portland, OR.
Swales, S. and C. D. Levings, 1989. Role of Off-Channel Ponds in the Life Cycle of Coho
Salmon (Oncorhynchus kisutch ) and Other Juvenile Salmonids in the Coldwater River, British
Columbia. Canadian Journal of Fisheries and Aquatic Sciences 46: 232-242.
Terrell, J. W., 1984. Proceedings of a Workshop on Fish Habitat Suitability Index Models. U.S.
Fish and Wildlife Service Biological Report 85(6), Washington, DC, 393 pp.
Terrell, J. W. and J. Carpenter, 1997. Selected Habitat Suitability Model Evaluations.
Information and Technology Report USGS/BRD/ITR-1997-0005, National Biological Service,
Fort Collins, Colorado.
Thedinga, J. F. and S. W. Johnson, 1995. Retention of Jet-Injected Marks on Juvenile Coho and
Sockeye Salmon. Transactions of the American Fisheries Society 124: 782-785.
Triska, F.J., V.C. Kennedy, R.J. Avanzino, and B.N. Reilly, 1983. Effect of Simulated Canopy
Cover on Regulation of Nitrate Uptake and Primary Production by Natural Periphyton
Assembleges. In: T.D. Fontaine III. & S.M. Bartell (Eds.), Dynamics of lotic systems. Ann
Arbor, MI: Ann Arbor Science, pp. 129-159.
Tschaplinski, P. J., 1988. The use of estuaries as rearing habitats by juvenile coho salmon. In:
Proceedings of the Workshop: Applying 15 Years of Carnation Creek Results, T. W.
Chamberlin (Editor). Carnation Creek Steering Committee, Pacific Biological Station, Nanaimo,
British Columbia, Canada.
94
-------
United States Department of Agriculture (USDA), 1997. Soil Survey of Lincoln County Area,
Oregon. United States Department of Agriculture, National Resources Conservation Service.
Van Sickle, J., J. Baker, A. Herlihy, P. Bayley, S. Gregory, P. Haggerty, L. Ashkenas, and J. Li.
In review. Projecting the Ecological Condition of Wadeable Streams, Under Alternative
Scenarios of Human Land and Water Use. Ecological Applications.
Van Sickle, J., P.J, Wigington, Jr., and M.R. Church, 1997. Estimation of Episodic Stream
Acidification Based on Monthly or Annual Sampling. Journal of American Water Resources
Association 33:359-366.
Warren, C. E., 1971. Biology and Water Pollution Control. W. B. Saunders, Philadelphia, PA,
434 pp.
Weavers, M. J., 1993. Life History and Evolutionary Adaptation of Pacific Salmon and Its
Application to Management. Ph.D. Dissertation, Oregon State University. Corvallis, OR.
Weitcamp, L. A., T. C. Wainwright, G. J. Bryant, G. B. Milner, D. J. Teel, R. G. Kope, and R. S.
Waples, 1995. Status Review of Coho Salmon from Washington, Oregon and California. NOAA
Technical Memorandum, NMFS-NWFSC-24. Seattle, WA.
White, J. L. and B. C. Harvey, 2001. Effects of an Introduced Piscivorous Fish on Native
Benthic Fishes in a Coastal River. Freshwater Biology 46:987-995.
Wigington, P.J., Jr., M.R. Church, T.C. Strickland. K.N. Eshleman, and J. Van Sickle, 1998.
Autumn Chemistry of Oregon Coast Range Streams. Journal of American Water Resources
Association 34:1035-1049.
Wilson, M.F., S.M. Gende, and B.H. Marston, 1998. Fishes and the Forest: Expanding
Perspectives on Fish-Wildlife Interactions. BioScience 48:455-462.
Wipfli, M.S., J. Hudson, and J. Caouette, 1998. Influence of Salmon Carcasses on Stream
Productivity: Response of Biofilm and Benthic Macroinvertebrates in Southeastern Alaska,
U.S.A. Canadian Journal of Fisheries and Aquatic Science 55:1503-1511.
Young, K. A., 1999. Managing the Decline of Pacific Salmon: Metapopulation Theory and
Artificial Recolonization as Ecological Mitigation. Canadian Journal of Fisheries and Aquatic
Sciences 56:1700-1706.
Zaroban, D. W., M. P. Mulvey, R. T. Maret, R. M. Hughes, and G. D. Merrit, 1999.
Classification of Species Attributes for Pacific Northwest Freshwater Fishes. Northwest Science
73:81-93.
95
------- |