Center for Streamside Studies

joint with

Center for Urban Water Resources Management

Annual Review
abstracts

Sex, lies, and steelhead: A genetic assessment of reproductive
success in hatchery and wild steelhead on Forks Creek, Washington

Jennifer McLean
jenm34@u.washington.edu

Hatcheries have artificially propagated salmon in the Pacific Northwest for more than 100 years, however, only recently have their effects on wild salmon become a concern. When hatchery fish and wild fish spawn in sympatry, there are four possible outcomes: the two stocks may remain distinct, the two stocks may homogenize and become one new “hybrid” stock, the hatchery stock may not do well in the wild and only survive through propagation in the hatchery, or the hatchery stock may thrive and replace the endemic stock. Hatcheries are sometimes blamed for native salmon declines, however, it is currently unclear what success hatchery salmon have had spawning in the wild, or what effects they have had on the reproductive success of wild salmon. A unique situation in which to investigate this question exists on Forks Creek in southwest Washington. The Forks Creek hatchery has been operating since 1895, but it was not until 1994 that steelhead trout, Oncorhynchus mykiss, were propagated and released. During the first two years in which hatchery steelhead returned, surplus hatchery fish were allowed upstream to spawn in the river where the wild fish, which return five months later, also spawn. Molecular markers such as microsatellite loci, highly variable repetitive nuclear DNA regions, can be used to evaluate genetic relationships among individuals. Genotyping individuals at many loci results in unique patterns or “DNA fingerprints” for each individual. These fingerprints allow assessment of kinship patterns, such as parent-offspring or sibling-sibling. Information gained from this kinship analysis will be used to evaluate the reproductive success of hatchery steelhead, as well as their effect on wild steelhead in Forks Creek.

Implications of flow regulation for spawning, incubating and emerging Cedar River steelhead: The development of an adaptive water management tool

Karl Burton
Karl.Burton@ci.seattle.wa.us
(UW graduate student)

Adequate instream flow is important for the protection and restoration of freshwater fish habitat in regulated river systems. Because of their varied and complex life history patterns, salmonid fishes are particularly vulnerable to managed flow regimes that alter the natural hydrographic pattern of rivers. Potential impacts from instream flow regulation vary depending on the frequency, intensity, duration and timing of flow events. The temporal and spatial distributions for spawning, incubation, rearing and outmigration vary among species and populations of Pacific salmon. Therefore, the relative impacts from a regulated flow regime will also vary among sympatric salmonid populations. For example, winter steelhead, the anadromous version of rainbow trout (Oncorhynchus mykiss), spawn in the spring and, consequently, are especially vulnerable to redd dewatering and fry stranding because their early life history stages occur during the descending limb of the annual hydrograph. As river inputs from snowmelt and rain events decrease with the onset of summer, instream flow levels decline and the potential for redd dewatering and associated egg and alevin mortality is increased. The timing for natural flow reductions is concurrent with an increase in water demand. Increased water withdrawal in response to increased water demand can further the potential for redd dewatering.

Recent declines in the Cedar River winter steelhead population prompted fisheries managers to question whether the minimum instream flow regime was providing adequate protection for incubating steelhead. These concerns resulted in a cooperative effort between Seattle Public Utilities and Washington Department of Fish and Wildlife to investigate the effects of instream flow regulation on steelhead spawn site selection and the associated susceptibility to dewatering for steelhead redds. A long term study was designed to determine: a) the timing and durations for Cedar River steelhead spawning, incubation and emergence, b) the relationships between flow levels at the time of spawning, redd location and subsequent vulnerability to redd dewatering, and c) the steelhead incubation protection performance of the present minimum instream flow regime provided by the City of Seattle’s Cedar River Habitat Conservation Program. An adaptive water management regime will be developed using the results from these investigations.

Factors limiting salmonid distribution in managed and
unmanaged drainages of the Puget Sound Cascades

Joshua J. Latterell and Robert J. Naiman
latterel@u.washington.edu

Fish distribution is a key criterion in the prescription of forest harvest practices in Washington. We investigated variability in physical structure of stream reaches immediately upstream and downstream of the upper distribution limit of fish in 58 headwater streams draining logged (managed) and unlogged (unmanaged) drainages of the Puget Sound Cascades. The uppermost fish were either coastal cutthroat, rainbow, or brook trout in all but two drainages. Salmonid distribution appears be predominantly restricted by dispersal barriers or the availability of pool habitat. Distribution limits were associated with geomorphic dispersal barriers (falls, cascades, steep bedrock chutes) more frequently than hydrologic (shallow stream, narrow channel, intermittent flow), organic (debris jams), anthropogenic (road culverts), or non-physical factors. Channel gradient was greater and pool abundance lower in reaches immediately upstream of the distribution limit than adjacent downstream reaches, in both managed and unmanaged drainages. In managed drainages, wetted channel width was more narrow upstream of the distribution limit than downstream. In unmanaged drainages, reaches upstream of the distribution limit had larger substrate and fewer large woody debris (LWD)-formed pools than downstream reaches. Comparisons of analogous reaches in managed and unmanaged drainages suggests that the physical structure of stream reaches immediately downstream of the uppermost salmonid is quite similar. Salmonids occupied reaches of similar width, gradient, pool attributes, LWD abundance, and substrate in both types of drainages. The mean increase in channel gradient at the distribution limit was smaller in managed drainages. This may, in part, reflect differences in drainage shape among managed and unmanaged drainages. Current work involves the use of logistic regression and discriminant analysis to further investigate the influence of climatic, geomorphic, and anthropogenic factors on salmonid distribution. We are also working on genetic analyses of salmonids in our study streams to examine the role of stocking on fish distribution and to describe the genetic population structure of salmonids in isolated headwater streams under different management regimes.

Effect of marine-derived nutrients on aquatic macroinvertebrate
production in a salmon spawning stream

Jon Honea
jhonea@u.washington.edu

Many hypothesize that the macroinvertebrate community of salmon spawning streams is an important conduit of marine-derived nutrients (MDN) from adult to juvenile salmonids. To test this hypothesis, I am analyzing macroinvertebrate production as a response to MDN released from spawning salmon. The size-frequency method will be used to estimate macroinvertebrate production upstream and downstream of a natural barrier to upstream migration on Kennedy Creek in the South Puget Sound. These production estimates will provide a measure of the total energy the macroinvertebrates make available to higher consumers, including juvenile salmon. I will also evaluate the relative importance of the different pathways of MDN to aquatic macroinvertebrates by monitoring changes in production in each of the major trophic categories. A comparison of changes in survivorship, condition factor, biomass, and length in the resident cutthroat common to both stream sections will determine if these salmonids are more successful in the presence MDN, as has been shown elsewhere. Stable isotope analysis of the relative concentrations of ¹²C to ¹³C and ¹⁴N to ¹⁵N in salmon and macroinvertebrate tissue will indicate whether increases in macroinvertebrate production can be attributed to MDN from spawning salmon and whether macroinvertebrate production is supporting the production of salmon.

Responses of fishes and salamanders to instream restoration
efforts in western Oregon and Washington

Philip Roni
phil.roni@noaa.gov
(recent UW graduate student)

Thirty streams in western Oregon and Washington were sampled during summer and winter to determine the responses of juvenile salmonids, juvenile lamprey (Entosphenus tridentatus and Lampetra spp.), sculpin (Cottus spp.) and giant salamanders (Dicamptodon spp.) to artificial large woody debris (LWD) placement. Total pool area, pool number, LWD loading, and LWD forming pools were significantly greater in treatment (LWD placement) than paired reference reaches. Juvenile coho salmon (Oncorhynchus kisutch) densities were 1.8 and 3.2 times higher in treated reaches compared to reference reaches during summer and winter, respectively. Densities of age 1+ cutthroat trout (O. clarki) and steelhead (O. mykiss) did not differ between treatment and reference reaches during summer but were 1.7 times higher in treatment reaches during winter. No significant difference was detected between densities in treatment and reference reaches for trout fry (age 0+ cutthroat and steelhead), giant salamanders, sculpin or larval lamprey. However, lamprey and coho response to LWD placement was positively correlated with LWD forming pools. Mean lengths of all species were similar in treatment and reference reaches. However, coho length was negatively correlated with densities and coho in treatment reaches were generally smaller than those in reference reaches. These results indicate that the largest fish response to restoration occurs for those species that prefer pools such as coho and lamprey and at those sites with the largest increase in LWD and pool area. However, the results of my study in no way negate the need to focus on restoring natural processes that create and maintain habitat rather than relying on in-stream manipulations.

Flood-related channel changes in a small forested watershed

Bryan Berkompas
bberko@u.washington.edu

Channel morphology can be dramatically altered during a flood. This study will look at the influences of discharge, reach scale conditions, and local conditions on channel morphology during a flood event. I will test three hypotheses. First, channel change will increase with increasing discharge. Second, channel change will be greatest in response reaches as defined by Montgomery & Buffington or in sensitive reaches as defined by Rosgen. Third, channel morphology at each site is controlled by local conditions. In February 1996 and spring 1997 large flood events occurred in Mica Creek in northern Idaho. Cross-sectional surveys in the summers before and after the events recorded the changes in the channel. Pebble counts and thalweg surveys were also completed. In addition, each cross section was rated for stability following the reach scale methods of Rosgen and Montgomery & Buffington and on a cross-sectional scale. The cross-sectional method evaluates large-scale roughness, vegetation, slope and network geometry at each cross section. The Distributed Hydrology Soil Vegetation Model (DHSVM) was used to calculate discharge at each cross section. I will compare discharge, reach scale conditions, and local conditions to the changes in the cross sections resulting from the 1996 flood event using regression analysis. Simple models will be based on these relationships. I will then test the relationships using the 1997 flood event.

Observations on the morphology of headwater channels

Mike Liquori
Mike.Liquori@ipaper.com
(UW graduate student)

As stream protection strategies shift higher into channel networks, it becomes important that we better understand how ecosystem processes influence channel behavior in headwater streams. While some patterns in headwater streams are similar to those found in higher-order channels, many significant differences exist that may shed light on the dependency on riparian communities and other ecosystem processes to create and maintain important habitats. I’ve examined data collected on 980+ stream segments, representing over 150 miles of headwater streams in the western cascades near Mt. Rainier. Trends include upstream fining and downstream increases in reach-average shear stress, which are opposite observations from fully alluvial (e.g. higher-order) channels. There are marked disequilibria in local gradient for some stepped headwater channels that may be linked to limited stream power, high sediment supply, or limits in step-forming structures. LWD load has little influence on channel type (at least between stepped and cascade channels). There is also no relationship between LWD size and residual pool depth. Small woody debris, either in jams or as single pieces, often provides step features, as can large clasts (cobbles/boulders), bedrock, or hydraulic scour. LWD frequency is also significantly lower than found in higher-order channels, likely due to fewer active LWD recruitment processes. These results suggest that greater focus on headwater channels may elucidate management objectives in headwater channels that may be significantly different than higher-order channels.

Influence of riparian alder on aquatic communities in
headwater streams: Olympic peninsula, WA

C.J. Volk, P.M. Kiffney, R.L. Edmonds
cvolk@u.washington.edu

In the Pacific Northwest, stands of red alder (Alnus rubra) dominate riparian corridors of many 2nd -growth forest streams. Red alder leaf litter is plentiful and rich in nitrogen and may provide an important source of carbon and nutrients to oligotrophic streams. Coniferous species, however, provide a more stable supply of pool-forming woody debris, which is important for juvenile Pacific salmon. As a result, current management practices include the removal of riparian red alder and replanting with coniferous species. To determine the effects of red alder nutrient inputs on aquatic communities, surface water chemistry, leaf litter inputs, periphyton and grazer communities, benthic invertebrates, and amphibian populations were measured on four red alder dominated streams and three coniferous old growth streams on the Olympic peninsula. During the fall, surface water nitrogen concentrations and leaf litter inputs were higher in red alder dominated streams than coniferous old growth streams. The nitrogen content of red alder litter was approximately 3 times greater than coniferous litter. Higher periphyton biomass and insect grazer abundance were also observed in red alder dominated streams. These data suggest that inputs from red alder forests may provide important nutrient and food resources for aquatic communities in Pacific Northwest streams.

Analysis of historic stream shade in eastern Washington

Gardner Johnston
gardnerj@u.washington.edu

Eastern Washington riparian forests have experienced substantial change over the past century due to fire suppression, timber harvest, road building and other land management activities. These actions have affected stream temperature, tree species composition, fuel loading, and insect and disease susceptibility in complex ways. Current management prescriptions, specifically stream shade requirements, need to consider historic condition and past management in order to accomplish forest health and species conservation goals. This study has three objectives: 1) characterize historic range of riparian canopy cover based on geomorphic and climatic variables, 2) compare current riparian canopy cover against the historic range, and 3) investigate the effect of fire suppression, level of timber harvest, and road building on riparian canopy cover. Air photos from the 1930’s and 40’s are being used to measure historic and current canopy cover for each reach. Stream reaches will be stratified by geomorphic and climatic variables using existing GIS coverages. Government land surveys from before 1900 will be used to obtain tree species and tree density information. This analysis covers selected areas of northeast Washington where data are available, including portions of the Kettle, Sanpoil, Colville, Spokane, and Upper Columbia river basins. The results of this study will provide insight into how the level of riparian cover has changed since the early 1900s. Additionally, the historic ranges of canopy cover will be useful for riparian management planning.

Instream wood: How much is enough?

Martin Fox, Loveday Conquest, Susan Bolton, and Pete Bisson
mjfox@u.washington.edu

Instream wood is recognized as an important feature linked to channel processes that benefit salmonids. Stream channel assessments associate the size, distribution, and abundance of woody debris to salmon habitat quality. Existing wood targets used by resource managers in the Pacific Northwest to evaluate, manage, and initiate stream restoration do not account for potential differences in wood distribution in different types of streams. To address this issue, we are developing improved estimates for in-stream wood characteristics that incorporate the variability found in natural basins with different climatic, geomorphological, and hydrological influences.

Instream wood surveys were conducted in 150 stream segments within National Parks and Wilderness areas in Washington State in order to characterize the spatial distribution, orientation, size, quantity, and other characteristics typical of wood found in streams draining unmanaged forests. Surveyed sites represent a wide array of geomorphic channel types where the process of wood input and distribution has most likely evolved under a natural rate of disturbance (with the possible exception of fire suppression). Channel types were distinguished by gradient, confinement, basin size, bed-form (alluvial or bedrock), origin (glacial, snow/rain, and rain-dominated), and eco-region (based on elevation and fire succession). All wood pieces with at least a 10 cm mid-point diameter and 2 m in length were counted, measured, and assessed for channel position, distribution relative to other in-channel wood, and species.

Preliminary analysis suggests that existing wood targets used in management may not be applicable to all channel types and regions. For example, in forums such as state Watershed Analysis, a stream is considered to have adequate quantities of wood if there is an average of 2 or more pieces (with qualifying dimensions) per unit channel width (Washington Forest Practices Board, 1997). However, in confined streams draining small basins (<4 km²) of eastern Washington, we found a mean of less than one piece of wood per unit channel width. In contrast, unconfined streams draining large basins (>100 km ²) in western Washington were found to have a mean of over 100 pieces per channel width. We found the variation of wood quantity and volume in streams is attributed to several factors. The data suggest that wood quantity and volume increases with 1) basin size, 2) adjacent riparian basal area, and 3) lack of confinement. The distribution of size classes of wood is also influenced by these factors. The data also imply that logjam size and frequency increases as the basin size becomes larger and less confined; however, the proportion of large “key members” to smaller “racked members” in jams decreases. Wood was evenly distributed throughout most of the surveyed streams only if jams were not present; however, when jams were found, they accounted for the majority of the wood within the segment.

Application of remotely-sensed data to regional analysis and
assessment of stream temperature in the Pacific Northwest

Nir Naveh, Stephen Burges, Alan Gillespie, Derek Booth, Chris Konrad, and Rebecca Handcock
navehn@u.washington.edu

The principal goals of this work (begun in April 2000) are to develop efficient methods for regional assessments of stream temperature and to demonstrate how the methods can be applied to assess effects of land use on stream temperature. We will evaluate the utility of remotely sensed thermal infrared (TIR) and visible images of streams and stream corridors for increasing the data coverage for regional stream temperature analysis and assessment. We have selected water temperature to illustrate and explore methods for regional water quality assessments because water temperature is biologically important; it is affected by anthropogenic activities; and surface (skin) temperature can be measured from remote instruments that detect TIR signals.

The ecological integrity of many rivers and streams in Washington State are threatened by elevated temperature. The National Water Quality Assessment Program (NAWQA) has selected summer stream temperature as a focal point for regional water quality assessment in the Puget Sound Basin. Regional-scale assessments are needed for: 1) monitoring water temperature because it is a spatially distributed condition; and 2) analyzing water temperature because it is influenced by spatially distributed conditions including anthropogenic activities near to and distant from stream channels.

Regional temperature assessments, however, are limited by sparse sampling in both space and time, given the area (or length of stream) of concern. In the Puget Sound ecoregion for example, the State of Washington relied on periodic data collected at 76 stations to assess water quality conditions of 12,721 km of streams and rivers (i.e., one station for 167 km of stream). An objective of our investigation is to develop and evaluate methods for extending spatial coverage of regional stream temperature assessments.

Three applications of remotely-sensed, thermal and visible, images to regional stream temperature assessments will be considered:

1. Locating ground stations in a temperature monitoring network. The objective is to evaluate whether remote imagery can be used to identify stream reaches that have strong temperature gradients. This information will be used to determine the length of stream that can be represented by a monitoring station, to evaluate whether temperature-monitoring stations are representative of streams in the basin, and to identify reaches in a stream network that may not require monitoring because temperature is likely to be uniform and cool.

2. Remote measurement of stream temperature. There are three objectives for this application: a. To develop empirical relationships between surface (top 100 mm) and kinetic (moving and mixed) temperature in relatively shallow water (<1 m); b. To identify the information (data quality) lost when using remote platforms (i.e., aircraft and satellite) to determine temperature; and c. To characterize the types of stream that are amenable to remote temperature monitoring. If stream temperatures can be estimated from images with known and acceptable levels of confidence, then regional temperature assessments will be less sensitive to the uncertainty associated with sampling temperature at a relatively small number of ground stations.

3. Remote collection of local, spatially distributed data for stream temperature analysis. While remote measurement of stream temperature may not be feasible for smaller streams, temperature in these streams may be strongly influenced by near-stream ground temperatures. The objective is to estimate ground (and shallow ground water) temperatures using remote imagery and incorporate this information in a stream temperature model. This application will improve representation and analysis of stream temperature dynamics.'

Evaluation and rehabilitation of urban streams in the Pacific Northwest

Derek B. Booth, Chris P. Konrad, Sarah A. Morley, Marit G. Larson,
James R. Karr, Sally Schauman, and Stephen J. Burges
dbooth@u.washington.edu

Urban streams are usually degraded from their natural, pre-urban state. Our goal in this three-year project has been to document the consequences of urban development on urban streams and to use that knowledge to evaluate the likely effectiveness of specific rehabilitation strategies. A multiplicity of conditions is required for “best” stream conditions, and the loss of any one generally produces degraded conditions. Yet a few specific consequences of human activity are ubiquitous in most urban and urbanizing watersheds; in particular, degraded hydrologic conditions are virtually universal. Other stream conditions with physical, biological and/or aesthetic consequences (notably water quality, woody debris loading, channel-bank scour, and riparian-corridor conditions) are also commonly, but not consistently, impacted. “Traditional” hydrologic parameters, such as flood recurrences and flow durations, incompletely characterize urban influences; hydrologic changes that affect sediment transport and biological integrity can occur without commensurate change in these parameters. Urban development, however, is not the sole determinant of hydrologic conditions, insofar as watershed geology, climate and weather, and channel-network hydraulics also influence flow patterns. Thus no single assessment, particularly “impervious area percentage in the watershed,” can adequately predict stream conditions, although this measure can be a useful indicator variable for a variety of environmental stresses. Similarly, any single spatial scale for assessing the magnitude of urbanization will not capture, nor accurately predict, the full range of biological and physical impacts that urban development may have on streams. We anticipate that rehabilitation success is most likely in those watersheds with relatively low levels of development, and that display paradoxically poor biological and/or physical conditions. We base this conclusion on numerous empirical examples where low watershed development and good in-stream conditions coexist. Rehabilitation is least likely to produce improvements in highly developed watersheds, because the inverse state (high levels of development with very good biological and/or physical conditions) is simply not observed. Also, in highly developed watersheds streamside residents more frequently maintain, but more typically diminish, healthy stream conditions. As yet, we can offer no examples of effective educational or stewardship programs to improve this trend.

Spatial effects of urbanization on Puget Sound Lowland streams

Maeve McBride
mcbridem@u.washington.edu

Strong and complex links tie the condition of a stream to the state of its contributing watershed. An urbanized landscape causes significant changes to the hydrologic regime of a watershed. With development, once forested land in the Puget Sound lowlands has been replaced with buildings, roads, and lawns. These impervious surfaces and other anthropogenic changes promote a “flashy” hydrologic system. The increased peak stream flows and lowered base flows ultimately result in stream degradation. The amount of impervious surface in a watershed is often used as a measure of urbanization, yet this measure may be too simplified. Stream health is frequently rated by the benthic index of biological integrity (B-IBI), assessments of a stream’s physical features, and/or records of stream flow. Prior research has shown that the amount of impervious surface in a watershed can only roughly predict these physical or biological measures. Other methods are required to explain the variability of the conditions in our urban stream systems.

We are starting to characterize watersheds using several spatial metrics including land cover distribution, connectivity of urban land, and other GIS-generated metrics. This analysis of landscape patterns is part of a larger, joint project (Center for Urban Water Resources Management (CUWRM), College of Architecture and Urban Planning, College of Forest Resources), which is looking at the relationship between urban patterns and ecological conditions. To date, I have completed a rapid assessment of physical features along the mainstem channels of four Puget Sound lowland watersheds. The collected field data will be compared with the spatial metrics. Preliminary results suggest that streams may degrade locally in response to nearby urban land, but may have the capability to recover downstream through reaches with intact, forested buffers, active floodplains, and/or riparian wetlands.

A comparative analysis: Watershed issues and science
in Massachusetts and Washington

Mindy Roberts
mindyr@u.washington.edu

Despite a distance of nearly 3,000 miles, the watersheds of Washington face many of the same challenges as the watersheds of New England. Eastern Massachusetts and the Puget Sound Lowlands both exhibit the impacts of historic and ongoing urban development. Maine watersheds are the subject of timber operations and fisheries discussions, spurred by the recent listing of Atlantic salmon as endangered. A brief introduction to New England issues should provide some lessons for the state of Washington from a different region of the country with many similar natural-resource issues, but with a much longer history of European settlement and urban development.

The Northeast and Northwest were both glaciated, resulting in a mix of glacial till, bedrock, and stratified drift deposits, and both have similar annual average precipitation. Headwater elevations are significantly higher in western Washington: this combined with climate and weather patterns, contribute to significantly higher discharge per unit area in western Washington compared to eastern Massachusetts.

Settlement patterns were similar, with coastal areas receiving the earliest settlers, followed by inland water corridors. Urban development in the greater Boston area began in the 1600s, and nearby forests were cleared for agriculture and to provide building materials. The first dam on the Charles River was constructed in 1634, sparking a confrontation between colonists and Native Americans living upstream who relied on anadromous fish runs for sustenance. Early colonial law required that all dams provide for fish passage, buth this was not strictly followed. In fact, freshwater fish populations declined precipitously, due to construction of over 20 dams on the 80-mile Charles River alone as well as significant industrial and domestic pollution. The recovery of the Charles River suggests the level of investment needed to mitigate the impacts of urban development. Earliest efforts involved the construction of large centralized wastewater treatment plants, followed by significant decreases in pollutant loads spurred by the NPDES permit requirements. More recent attention has focused on mitigating combined sewer overflows and improving stormwater through regulatory requirements. The extensive and diligent efforts have resulted in much improved water quality; however, contaminated sediments remain an issue.

The Edwards Dam on the Kennebec River in Augusta, Maine, was removed July 1, 1999, to open up 17 miles of habitat for numerous species, including Atlantic salmon and sturgeon. The Federal Energy Regulatory Commission (FERC) denied an application by Edwards Manufacturing Company to relicense the dam, after considering three options: (1) approve the application, (2) approve the application under the condition that a $10 million fish passage system be installed, or (3) remove the dam. FERC’s analysis concluded that installation of the fish passage system, which would not benefit all species of interest, would be more expensive than retiring and removing the dam, and that continued operation would result in unacceptable environmental damage. As of summer 2000, water quality has improved and mayfly and stonefly populations have increased dramatically. American shad, alewives, Atlantic salmon and sturgeon have returned. Gravel bars began reappearing in November 1999. Subsequent monitoring should provide insight to the short- and long-term management of impounded sediments.

The older development and experience of New England could benefit Washington as it addresses urban impact mitigation and dam removal.

Crafting collaboration: An implementation analysis of
Washington State’s Watershed Planning Act

Jacqueline Klug
jsklug@u.washington.edu

Collaborative resource management is becoming one of the preferred methods for making decisions on how to manage natural resources. There are now hundreds of collaborative management partnerships and watershed councils in existence across the West. While a majority of these partnerships are “bottom up” grassroots initiatives, increasingly the partnerships are being structured and initiated through legislation and agency directives.

A substantial literature on collaborative management exists, with most work focusing on the successfulness of the partnerships on reducing conflict in the region. While these studies provide useful insight on the factors influencing the collaborative process, the question of what it means to institutionalize collaborative decision-making processes into policy remains unexplored. This study investigates this question by examining collaborative management through the lens of policy implementation theory. We focus on Washington State’s Watershed Planning Act of 1998 (ESHB 2514) and how the process of collaboration is structured by the Act. We draw upon both “top-down” implementation theories and “bottom up” implementation theories to explore implementation of collaboration as outlined by the Act and how the implementation process is altered by the actors involved in the process.

Four watershed planning units were studied in our work: two planning units which formed as result of the Act and two units which existed prior to the Act. Interviews were conducted with participants from the relevant institutions involved in each planning unit. The results of this study will provide new insight on how collaborative processes work within a formal policy structure as well as the unique challenges of implementing collaborative management policies.

Factors influencing amphibian colonization of
stormwater control ponds in King County

Elissa Ostergaard
eoster@u.washington.edu

Stormwater ponds are built to protect human health and safety as well as aquatic resources, primarily wetlands and streams. Stormwater ponds are generally designed to detain stormwater and improve water quality, and may have unintended impacts on wildlife. A two-year study of 53 stormwater ponds in King County is being undertaken to determine the extent of use by amphibian species, whether mortality occurred prior to larvae metamorphosis due to pond drying, and if landscape and in-pond factors correlate with amphibian use. Pond age ranges from 1 to approximately 16 years. Preliminary results indicate that three species of caudates (Northwestern salamander - Ambystoma gracile, long-toed salamander - A. macrodactylum, rough-skinned newt -Taricha granulosa) and three species of anurans (red-legged frog - Rana aurora, bullfrog - R. catesbeiana, Pacific treefrog - Hyla regilla) breed in stormwater ponds. All ponds with standing water greater than 15 cm deep in March-April supported breeding amphibians. All species can colonize within two years of pond construction, and Pacific treefrogs colonized ponds in less than six months. Bullfrogs do not appear to exclude native amphibian species, and are occasionally present in temporarily flooded ponds. Northwestern salamander gilled adults were observed in approximately 25% of ponds. Pacific treefrogs occurred most frequently, and rough-skinned newts were the least common of the six species observed. The Oregon spotted frog (Rana pretiosa) and Western toad (Bufo boreas) were not observed. Species richness appears to be positively correlated with the presence of forest within 100 meters of stormwater ponds. Fewer than five percent of ponds were observed to dry before larvae metamorphosis. Future work will examine the extent of egg stranding and correlations between water level fluctuations and species occurrence.

Monitoring biological condition in urbanizing headwater streams

Heidi Wachter
hwachter@u.washington.edu

Instream benthic macroinvertebrates are commonly used indicators in biological monitoring and have been shown to respond to human-induced changes in a watershed. However, the scoring criteria and/or sampling protocols, as developed for larger Puget Sound lowland streams, may not be applicable to measure biological condition of headwater streams. Two urban planned developments (UPDs) being constructed in a rural headwater region in northern King County, under a single development plan with the articulated management goal of “no downstream impacts,” provide an opportunity to test this concern. Incorporated into the development permit is a comprehensive monitoring plan, designed to assess whether avoiding downstream impacts is even possible. The monitoring plan includes chemical and physical assessment, although these monitoring elements alone cannot indicate the response from aquatic biota. Therefore, benthic macroinvertebrate sampling is also being implemented to monitor the biological condition of the UPD headwater streams.

Benthic invertebrate sampling was initiated in fall 1997. A regional index, the Puget Sound lowland Benthic-Index of Biological Integrity (B-IBI), was used for analysis of the fall samples using the B-IBI sampling protocols. However, fall B-IBI scores were much lower (indicating poor biological condition) than expected for the low-level of urbanization in the UPD watersheds. Furthermore, low or non-existent late summer flows in the UPD streams could not guarantee annual fall sampling of benthic invertebrates. Thus, in 1999, sampling protocols were altered to accommodate the seasonal nature of these streams by sampling in late spring (early- to mid-June), after major spring storms and prior to significant loss of surface flow. This has solved the problem of inadequate flow for sampling but not the likelihood that community structure and assemblages respond differently to urbanization in headwater streams. Therefore, benthic invertebrate sampling was conducted in spring 2000 in seventeen headwater streams throughout the Puget Sound lowlands in watersheds with low, moderate, and high levels of urbanization. Stream reaches sampled met similar criteria for channel size, percent grade, riparian cover, substrate composition, and local geology. To facilitate results that are indicative of impacts to biological condition from basin land use, sampling locations were selected to minimize disturbance from local conditions (e.g. high-use roadways, or no riparian canopy).

Analysis of the spring 2000 samples involves examining of a range of biological attributes to identify community characteristics that are sensitive and respond to increased levels of urbanization within a basin. Preliminary results indicate B-IBI metrics developed as indicators of disturbance for perennial streams may be applicable for spring sampling in headwater streams.

Distribution patterns of hyporheic fauna in a floodplain riparian terrace

Holly Coe and Rick T. Edwards
hollycoe@fish.washington.edu

The hyporheic zone is an area of saturated sediments located beneath and adjacent to streams and rivers and has recently been recognized as an important part of the hydrologic continuum, effectively linking terrestrial, surface and subsurface components of rivers. Studies investigating hyporheic invertebrates have focused on communities within the active channel, either beneath the wetted channel or channel gravel bars. The floodplain hyporheic zone has received little attention in the literature but may be important in rivers with active floodplains. The objectives of this study are to describe hyporheic invertebrate community structure, beyond the active channel, in the floodplain hyporheic zone and to relate spatial and temporal patterns in community structure to hydrology and biological and physical parameters. Wells in a floodplain riparian terrace on the Queets River were sampled during summer low-flow 1999 and winter and spring 2000. Preliminary identification of the invertebrates indicates the presence of at least 18 major taxa including archiannelids, amphipods, chironomids, copepods, hydrachnids and tardigrades. Cyclopoid copepods and archiannelids dominate in summer, winter and spring with highest densities occurring along the perimeter of the terrace. There are no differences in total invertebrate density between seasons suggesting that the floodplain hyporheic zone may be a stable environment for total system invertebrate production. However, high variability between wells suggests that there may be seasonal differences in individual taxa densities. Hyporheic invertebrate communities can provide unique insights into the linkage among terrestrial, surface and subsurface components of rivers and may ultimately play an important role as indicators of water quality and ecosystem health.

Debris flows and stream benthic communities: A natural experiment
at West Twin Creek, Olympic National Park, Washington

Peter Kiffney, Carol Volk, Robert Edmonds, and Georgia Murray
Peter.Kiffney@noaa.gov
(Affiliate professor, Center for Streamside Studies)

We have been monitoring inputs of riparian vegetation; suspended organic matter; periphyton; stream invertebrates and vertebrates in West Twin Creek and Lindner Creek since May 1999. The objectives of this research are (1) to couple the dynamics of biological communities with watershed chemistry; (2) to compare energy flow and food web dynamics in streams dominated by old-growth conifers (West Twin) vs. streams dominated by riparian red alder (Lindner Creek); and (3) to understand the role of stream amphibians in headwater food webs. This talk will present data on the first two objectives and describe studies planned for the third. In addition, we will present data on the ecological effects of an unusual natural disturbance in West Twin Creek. In late December 1999, a rain-on-snow event occurred in the Hoh River drainage. In response to this storm event, a large debris flow occurred in West Twin Creek where massive amounts of wood and sediment traveled almost 2 km, and obliterated our research site. Before this event, stream periphyton biomass was higher and insect grazers were more abundant in Lindner compared to West Twin Creek. Following this event, periphyton is higher and there are more insects in West Twin compared to Lindner. This is one of the first studies that will document the temporal dynamics of water chemistry and biological communities before and after a large-scale natural disturbance in an old-growth watershed. Capturing such events without long-term monitoring is virtually impossible.

Influence of Pacific Salmon on nutrient availability and
vegetation in riparian areas in western Washington

Robert E. Bilby, Eric W. Beach, Brian R. Fransen, Jason K. Walter, and Peter A. Bisson
bob.bilby@weyerhaeuser.com
(Affiliate professor, Center for Streamside Studies)

Pacific salmon can transport large quantities of nutrients from the Pacific Ocean to streams and rivers where they spawn and die. Thus, these fish may influence the ecological characteristics of riparian areas bordering streams where they spawn. We compared riparian plant communities along streams in western Washington (Griffin Creek and Kennedy Creek) above and below barriers to spawning salmon. Salmonberry (Rubus spectabilis) foliage at the channel edge and at 20 m, 50 m, and 100 m upslope from the stream was analyzed for nitrogen stable isotope ratio (d¹⁵N, an indicator of salmon-derived nitrogen) and total nitrogen (N) and phosphorus (P) content. In addition, we compared cover, plant density and species richness of shrub and understory vegetation between sites with and without salmon. d¹⁵N declined with distance from the channel at sites with and without salmon at both locations. d¹⁵N values in salmonberry leaves at sites with salmon were higher than at corresponding distances from the channel at sites without salmon at Kennedy Creek but not Griffin Creek. Salmonberry foliage possessed significantly higher levels of total N and P adjacent to salmon spawning reaches of both streams. Shrub cover was greater at sites with salmon at Griffin Creek and shrub species diversity was higher at sites with salmon at Kennedy Creek. Understory plants at Griffin Creek sites with salmon exhibited greater cover, species diversity and plant density than sites without salmon. Understory plant density and species diversity was higher at sites with salmon at Kennedy Creek. Results indicate that Pacific salmon carcasses contribute to nutrient availability in riparian areas and influence the cover, density and diversity of riparian shrub and understory vegetation.

p o s t e r s

Sediment from stream-crossing restoration practices in western Washington

Tim Brown
tmbrown@u.washington.edu

Forest transportation systems can significantly affect streams and the habitats they provide including those with to anadromous fish. Documented examples of adverse changes associated with forest roads include changes in water quality and quantity, increases in fish passage barriers and increases in access to previously remote and pristine areas once capable of supporting habitat-forming processes. A greater emphasis on protecting fish and wildlife habitat and providing clean water has shifted policy from developing transportation systems to managing current systems in a more environmentally and financially responsible manner. These changes combined with budget shortfalls have forced many managers to resort to road removal and closure alternatives rather than spend the money required to maintain the roads in their inventory.

With a sharp increase in the use of road closure and removal techniques, often referred to as road decommissioning, it is important to understand the extent of their effects, at least in comparison with the “do nothing” alternative. Transportation managers currently make decisions under the premise that decommissioning is “better” than doing nothing. However, it is important to validate this position and determine just how much better and if practices can be improved. Thereby, we avoid the mistake of doing something simply for the sake of doing something, and actually address the problem.

This is a pilot study examining contemporary road decommissioning techniques in western Washington public forests. This study will focus strictly on what occurs at stream crossings, and will be done in cooperation with another study examining sediment inputs from decommissioned running surfaces. Oftentimes, stream-crossing decommissioning is actually a restoration of the crossing to some semblance of previous conditions (e.g., culvert removal, reestablishment of slope and channel width, etc.), and is referred to as stream crossing restoration. Moreover, diversion of flow around a site while work is in progress is strongly encouraged, but for a number of reasons including practicality is not always practiced. The primary objective of this study is to examine and document these techniques and attempt to quantify the amount of sediment introduced during active restoration in perennial or live streams.

Currently water quality samples (6 x 250 mL grab samples) are being taken directly above and below a site and before, during and after active restoration to determine the total suspended solids resulting from the site and work being done. Because many of the roads being decommissioned are along ephemeral streams rather than perennial (or live) streams a problem arises. To deal with this problem, standard surveying equipment (a rotary laser level and total station) are being used to establish cross-sections or in a few cases digital elevation models that can be used to monitor erosional patterns and estimate volumes of transported material from year-to-year. Therefore a distinction between the sedimentation from both types of streams can be made. Installation of continuous monitoring devices for suspended solids and/or turbidity is a possibility for the future if the resources become available. Currently work is being done on decommissioning in the Beckler and White River watersheds of the Snoqualmie-Baker National Forest, and additional study sites are being planned for the Cedar River watershed. Any information on recent or upcoming decommissioning on public or private lands interested in being included in this study would be greatly appreciated.

Mapping the historic river landscape of the Puget Sound basin

Brian D. Collins, Amir J. Sheikh, and David R. Montgomery
bcollins@u.washington.edu

We are using archival and modern sources with a GIS to map historic rivers, wetlands, and vegetation in the river valleys of Puget Lowland. Archival sources from the mid-to-late 19th century include plat maps and field notes of the General Land Office surveys and charts by the US Coast & Geodetic Survey. Modern sources, such as detailed topography from LIDAR imagery are useful for refining historic data. We hope to publish the results in an historic river resource atlas and to make data available online for interactive use with a map server. Potential uses for the data include estimating historic salmonid production, creating an ecological context for forest restoration, and providing a reference condition for river and floodplain restoration.

Forest development, log jams and restoration of
floodplain rivers in the Puget Lowland

Brian D. Collins and David R. Montgomery
bcollins@u.washington.edu

In Puget Lowland rivers historically, woody debris jams created anastomosing channels, a dynamic channel-floodplain connection, and deep pools. In the late 1800s, debris was removed from rivers, rivers isolated from floodplains, and floodplain forests cut down, thereby limiting debris recruitment. An 11-km-long reach of the Nisqually River is a unique site in the Puget Lowland to study natural river processes because it has natural banks and a mature floodplain forest. The dynamic between floodplain forests, wood debris recruitment, and debris jams in the Nisqually shows that reestablishing floodplain forests and the fluvial processes that create stable log jams is integral to restoring Puget Lowland rivers. A critical question for contemporary restoration planning is how long it may take to reestablish debris recruitment and in-channel habitat. One necessary condition is for floodplain forests to include trees that provide wood debris large enough to function as key pieces in stable jams. While the frequency of large trees in the Nisqually valley-bottom is comparable to that documented in 1873 land surveys, many of the formerly-dominant Thuja plicata (western redcedar) were cut down in the late 1800s, and now hardwoods, including Populus trichocarpa (black cottonwood) and Acer macrophyllum (bigleaf maple), are also abundant. Pseudotsuga menziesii (Douglas fir) and the fast-growing Populus trichocarpa commonly form key pieces, suggesting that reforested floodplains can develop naturally-recruited stable debris jams within 40 to 80 years, much faster than generally assumed. A second necessary condition for restoring Puget Sound rivers is a dynamic river-floodplain connection to recruit enough debris to create jams. We propose a planning framework, based on these two conditions, within which to develop programs to restore self-sustaining, dynamic river morphology and habitat in Puget Lowland rivers.

Historic changes in the distribution and functions of
large woody debris in Puget Lowland rivers

Brian D. Collins, David R. Montgomery, and Andrew D. Haas
bcollins@u.washington.edu

We examined changes in wood debris loading and functions in Puget Lowland rivers resulting from the last ~150 years of land use by comparing field data from an 11-km-long protected reach of the Nisqually River with field data from the Snohomish and Stillaguamish rivers, and archival data from several Puget Lowland rivers. The field and archival data indicate that historic wood debris load has been reduced by one to two orders of magnitude. This is primarily due to a decline in wood debris jams, which are generally now absent because of a lack of very large debris that can function as key pieces, and low rates of debris recruitment because of levees. The historic change in wood debris abundance and size fundamentally changed the morphology and habitat of lowland rivers at all spatial and temporal scales. Based on our field studies, rivers had substantially more and deeper pools historically. Archival data and field studies indicate debris jams were integral to creating and maintaining a dynamic, anastomosing river pattern with numerous floodplain channels and abundant edge habitat, and that wood debris routed floodwaters and suspended sediment onto floodplains and deltas. Targets for restoring rivers and habitat must take these historic changes into account.

Monitoring the response in headwater streams to increasing urbanization

Karen Comings
kcomings@u.washington.edu

Studies of the effects of urban development typically must rely on comparisons between different watersheds at various levels of urbanization. The construction of a 5-km² development in western Washington provides a rare opportunity to study catchment response to urbanization as changes in land use occurs within a particular watershed. Prior to the start of construction in 1998, this site typified “undisturbed” conditions of the late-20th century in the Pacific Northwest: covered with mature second-growth forest, and many large wetlands that serve as headwaters for over half a dozen streams. The quality of the ecosystem was very high, supporting diverse populations of fish, amphibians, and aquatic insects, with downstream reaches that support wild salmonids. A central condition of development approval was that the construction activities and the development itself have minimal effects on the local aquatic ecology. In response, the project has an extensive drainage plan with state-of-the-practice best management practices (BMP’s) for mitigating hydrologic changes and a fully funded 15-year monitoring plan. This project is an opportunity to evaluate how effective our best efforts are at protecting aquatic ecosystems from the impacts of urbanization. Careful monitoring is essential, yet raises many challenges such as data resolution and repeatability; differentiating natural variability from disturbance effects; and coping with inconsistencies in collection methods between baseline data, construction data, and post-construction data. This project provides the opportunity to evaluate the effectiveness of current BMP’s on a basin-wide scale, and improve methods of ecological monitoring.

Using remote underwater video cameras to study salmon spawning behavior

Manu Esteve
manu19b@u.washington.edu

A remote underwater video system is being used to monitor and study salmon spawning behavior. The following behaviors are quantified: females cutting the redd, males quivering, attacks, probings, false spawnings, and actual spawnings. The objectives are to match the possible similarities and differences in spawning behavior of various salmonid species (Salmo and Oncorhynchus) with their evolutionary phylogenetic tree. This project will test the hypothesis that more closely related species will share the most common behaviors. The adaptive significance of the different behaviours will be discussed. A general conclusion about the sexual and natural selection processes driving salmonid reproduction will be drawn.

Investigating the effects of buffer width on organic matter input in the riparian zone

Johnny Grady, Jr.
grady96@u.washington.edu

One of the primary purposes of buffering headwater channels is to protect fish-bearing waters downstream from the adverse effects of timber harvesting. However, questions about what constitutes adequate buffer widths persist. State and Federal agencies have developed different policies regarding timber harvest practices in riparian zones. The policies require differentbuffer widths along streams and allow different activities within these zones.

To determine an optimum buffer width, one that protects the integrity of the riparian ecosystem while allowing profitable timber harvest, researchers have examined riparian processes as a function of distance from stream channels.

Few studies have examined the effects of buffer width on organic matter input in riparian zones. This relationship is important because a buffer strip of insufficient width may not provide adequate food resources for the stream ecosystem. Over the years, researchers have investigated the nutrient dynamics of streams by focusing on the terrestrial inputs. Several studies have demonstrated that litterfall from the forest explains the predominately hetertrophic bases for the stream food web.

From September to November 1999 and June to November 2000, coarse organic matter input into the stream riparian zone were measured at four tributaries to the Green River, Washington. All sites were on north-facing slopes, two of which had recently been clearcut (Summer 1999) leaving buffer strips of varying widths between the clearcuts and the streams. Ten litterfall traps were positioned at each site along the stream channel outside of the bankfull channel. Organic material from traps were collected every two weeks and evaluated in the lab to quantify litterfall associated with riparian buffers.

The objective of this project is to study the effects that the managed riparian zones of Sweeney Creek and Humphrey Creek have on organic matter dynamics. These study sites are compared to old-growth riparian forest habitats of Wolf Creek and Gold Creek.

Response of stream periphyton and insects to a gradient of riparian buffer widths

Peter Kiffney, John S. Richardson, and Jennifer P. Bull
Peter.Kiffney@noaa.gov
(Affiliate professor, Center for Streamside Studies)

At the University of British Columbia’s Malcolm Knapp Research Forest, Maple Ridge, BC, Canada, an ecosystem-level riparian management experiment was initiated in 1995. This study is taking a multi-disciplinary approach to examining the structural and functional responses of stream and terrestrial ecosystems to riparian buffer width. After 2-5 years of baseline data collection, 10 stream reaches were logged from April 1998 to February 1999 and left with either no buffer (n=4), a 10-m (n=3) or a 30-m (n=3) buffer. In addition, three streams serve as unmanipulated controls. In this research, we used unglazed ceramic tiles to measure the response of stream periphyton and invertebrate grazers to riparian manipulation. We collected 12 to 24 months of pre-cut data and approximately 17 to 24 months of post-cut data. Light levels were higher in all treatments compared to the controls. In response to changes in light regimes, and potentially other abiotic variables (temperature and nutrients) we have observed significant increases in periphyton biomass in the clear cut and 10-m buffer treatments compared to controls. Invertebrate grazers, primarily Chironomidae, have also increased in response to riparian buffer width with the largest increase in the clearcut and 10-m treatments. In some cases, however, there have been significantly more chironomids in 30-m treatments compared to controls. Although these data suggest that periphyton and primary consumers were higher in reaches with no or a narrow buffer (10-m), there has been a concomitant increase in the amount of fine sediment embedded in the periphyton matrix in these treatments. This increase in fine sediment may affect grazing by some invertebrates, such as Trichoptera.

Spatial patterns of bed material entrainment in gravel-bed streams

Chris Konrad
cpkonrad@u.washington.edu

Floods are a frequent agent of sediment transport and ecological disturbance in gravel-bed streams. The spatial extent of bed material entrainment during a flood determines, in part, the mass of sediment transported and the severity of the disturbance. Field experiments at seven gravel bars in three streams in the Puget Lowland, Washington, were conducted to document spatial patterns of bed material entrainment during floods. The fraction of a bar's surface entrained during a flood, or partial entrainment, was related to the peak dimensionless shear stress of the flood (i.e., total boundary shear stress divided by the median of the particle size distribution of the bed surface). The locations of bed material entrainment, however, were widely distributed over the bars and not predictable from flood to flood. Bed material entrainment can be represented as a spatially stochastic process, but the probability of entrainment during a flood is only approximately uniform along a gravel bar. Sections of a gravel bar with convergent flow and regions near the center of the channel are less stable than sections with divergent flow and regions along stream banks.

Factors influencing the distribution of conifers and alder
in riparian zones in western Washington

Kerri N. Mikkelsen and Robert L. Edmonds
knm5c@u.washington.edu

Riparian forests bordering second and third-order streams in old-growth forests were originally dominated by conifers in lowland western Washington. After extensive cutting during the early to mid-1900s, these riparian systems are now dominated by red alder (Alnus rubra). There is considerable interest in converting alder dominated riparian systems to more “natural” conifer stands. The methodology for conversion, however, is still being developed. This study in Browns and LeBar Creeks, tributaries of the Skokomish River Basin in the Olympic National Forest, was designed to determine what factors promote the establishment, growth, and survival of conifer species in riparian forests. Specific objectives were to: (1) Determine if soil parameters differ under red alder and conifer patches in riparian forests; (2) define the role of soil, coarse woody debris, seed source, and landform in conifer establishment in riparian forests; and (3) monitor stream chemistry, particularly in relation to placement of salmon carcasses. This study measured tree size, age, and spatial distribution; soil pH, organic matter, moisture, N, P, texture, microbial biomass, and concentrations of stream N. Results indicate that conifer survival is greater on terraces, while alder prefer floodplain soils. Coarse woody debris, landform, and proximity to seed source are important for conifer establishment. Organic matter is low in both alder and conifer soils. Soil pH is not different under alder and conifer. Alder adds significant amounts of nitrogen to streams during the fall. Salmon carcasses may influence stream nitrate. Management strategies will be developed from study results.

Incorporating prior knowledge in environmental sampling:
Ranked set sampling and other double sampling schemes

Nicolle A. Mode, Dr. Loveday L. Conquest, Dr. David A. Marker
conquest@u.washington.edu

Scientists design environmental studies to optimize precision and allow for generalization of results, while keeping costs of associated field and laboratory work at a reasonable level. Ranked set sampling (RSS) is a two-phase sampling procedure that increases precision and reduces costs by using “rough but cheap” qualitative or quantitative information to obtain a more representative sample before the more accurate (and more expensive) response is obtained. We have investigated under what conditions RSS becomes a cost-effective sampling method for ecological and environmental field studies. The first example, a small field study on the Taylor River, shows no difference between RSS and ratio estimation, although other issues of quality control (e.g. where does a habitat begin and end) were very much in evidence. The second example uses simulation on habitat measurements from Oregon streams to investigate the effect of different degrees of information, by comparing RSS to simple random sampling; to an RSS procedure with “cutpoints” taken from having general information on a distribution; and to ratio estimation, which assumes a linear relationship between the cheap and expensive measurement. We conclude that RSS is a viable alternative to other sampling approaches; it is a way to incorporate human judgment and prior information when costs or other factors (e.g., destribution of sampling material) act to keep sample sizes small.

Spring and fall movements of juvenile salmonids

Jamal Moss and Susan Bolton
sbolton@u.washington.edu

Salmon are an important part of Pacific Northwest culture and economy, but have disappeared from almost half of their historical breeding grounds in Washington, Oregon, Idaho, and California. Despite large investments in fish ladders, hatcheries, and regulations protecting riparian zones, there are continuing reductions in salmon abundance. Historical watershed reconstruction has shown that key habitat features for coho have been lost, and currently numerous efforts are underway to restore off-channel habitat throughout the state with the hope of restoring stocks.

Culvert repair or redesign for better fish passage has the potential to open up good habitat. In the State of Washington culverts are designed to meet average barrel velocity requirements that can range from three to six feet per second depending on culvert length and the species of fish for which passage is designed. Presently, passage regulations are designed for returning adult salmon and not juveniles. However, recent studies have documented the importance of upstream passage for salmonid juveniles and described the swimming capabilities of salmon fry. Smaller fish have slower swimming speeds, which causes them to expend more energy relative to their size during burst and prolonged speeds. To improve design guidelines, better information on when fry and juvenile salmon move is needed.

This pilot project had the following objectives:

1. Assess upstream movement of swim-up coho (Onchorhynchus kisutch).

2. Assess Spring juvenile coho movement in streams with and without culverts.

3. Analyze existing tributary trap data for fall movement out of rivers and into tributaries by examining trends in movement timing and environmental conditions.

4. Provide recommendations for future studies investigating these questions.

At the Remote Site Incubator ( RSI) site in Andrews Creek, 17 fry were caught in minnow traps as much as 174 m upstream from the RSI. None was caught downstream. At the Snow Creek RSI site, 11 fry were caught as far as 210 m upstream from the RSI and 21 were caught as far as 560 m downstream. Upstream distances of captured fry tended to increase over time.

The distribution of juvenile coho between 30 and 50mm fork length appears to be more influenced by water velocity rather than habitat type. In early April, 286 of the 410 fish captured in Griffin Creek inhabited regions with water velocities between 0 and 0.1 feet/second and 371 of the 410 juveniles inhabited areas with velocities less than 0.4 feet/second. Juvenile coho were most commonly found in pools and along the bank, but were also associated with large woody debris (LWD). Habitat did not appear to be as important in fish distribution as velocity. Juveniles were often found milling about in large aggregations (20 or more fish) suggesting that feeding territories had not yet been established.

Recapture rates from mark and recapture studies designed to track juvenile coho movement through culverts were low (2/36 fish in Stossel Creek and 2/60 fish in Griffin Creek). Repeating the mark and recapture study with larger juvenile cutthroat yielded acceptable recapture rates (48%). Tagging of fish < 50mm can be done successfully with Bismarck Brown dye and adipose fin clipping. For meaningful movement data, sites where 300-400 fry can be tagged are necessary. Fry traps upstream and downstream of the study site would help determine the number of fish that move out of the study reach entirely. Working in streams less than 4 meters wide is also better for juvenile work.

Water temperature, flow, date, and moon phase were compared to juvenile trap data from tributaries on the Olympic Peninsula and the North Cascades. Preliminary analysis of fall trap data indicate that movement of juveniles out of mainstem rivers into tributaries occurs over a ten week period from early October to mid-December (Julian weeks 40-50). At some sites, movement is concentrated into 2 or 3 weeks and at other sites it is extended over the entire period. The greatest amount of movement occurred between new moon and half moon phases and with flow increases of 2000 cfs.

Nutrient processing in an Alaskan salmon stream —
fish, alder, and the hyporheic zone

Tom O’Keefe and Rick Edwards
okeefe@u.washington.edu

Evidence for the importance of marine-derived nutrient (MDN) inputs to freshwater ecosystems and the adjoining riparian zone is rapidly accumulating, but the mechanisms by which MDN inputs are stored and transferred within the stream corridor are poorly understood. Annual inputs of MDN increase stream productivity and may result in enhanced growth rates and altered community composition in riparian forests. Since floods and mammals can remove carcasses and decomposition products from streams, storage of nutrients in the hyporheic zone can be an important mechanism by which salmon-derived nutrients are retained within stream ecosystems and subsequently made available to primary producers. Uptake of enriched hyporheic waters by riparian root systems represents a potentially important pathway for nutrient transfer to riparian vegetation. Within Lynx Creek, a sockeye spawning stream in SW Alaska, surface water moves into extensive areas of floodplain hyporheic zone along spawning reaches and flows downstream before returning to the channel. Nutrients entering the hyporheic zone are retained and transformed as they move along flowpaths. The relative importance of this transfer in any given year varies with the size of the salmon run and the capacity for animals to remove carcasses from the stream. At our study sites in SW Alaska, distribution of alder within watersheds has a large impact on the N dynamics within individual streams. While MDN are clearly important for stream ecosystems in general, their role must be interpreted within the context of landscape features of the watershed that control other nutrient transfers.

Land-use and suspended sediment in Puget Lowland streams

James Packman and Susan Bolton
jpackman@u.washington.edu

The presence of suspended sediment and other solids in the aquatic environment is sensitive to upstream land-use and largely determined by the source and pathway of sediment input. Suspended sediment can have a wide range of deleterious effects on salmonids depending on the concentration and composition of particles. This study examines suspended sediment loads (measured from instantaneous discharge and total suspended solids) and grain sizes from storm-event samples during water year 2000.

Eight low-medium gradient salmonid streams in King and Snohomish counties were studied that have a single primary land-use: 3 urban streams, 2 agricultural streams and 3 forest streams. Results indicate identifiable differences in suspended sediment loads among land-uses, with urban streams having the highest sediment loads (up to 2,700 kg/s of instantaneous sediment flow) as well as the greatest variability in sediment concentration. Highest instantaneous sediment loads recorded in forest and agricultural streams were 47 kg/s and 39 kg/s, respectively. D50 grain size for suspended sediments are consistently in the 0.001-0.01 mm size range across all land-uses. While land-use explains some of the variation in suspended sediment loads recorded, other factors also show a significant influence, including normalized storm precipitation depth, basin size, stream gradient, season and time of sampling on hydrograph.

Heavy metal transfer from mines in eastern
Washington to organisms in the Methow River

Dan Peplow
dpeplow@u.washington.edu

Historic hardrock mining that occurred near Twisp, in Okanogan County, Washington, has resulted in surface and shallow groundwater contamination by heavy metals, which poses a risk to endangered salmonids in the Methow River. Water-borne metals at concentrations that exceed State and Federal criteria levels are the only ones regulated. Sediments in the Methow River, however, are causing adverse effects on aquatic biota even where waters meet water-quality guidelines. The elements Zn, Cu, and Pb appear to be present as cations adsorbed to coatings on the surface of mineral particles. The accumulation of particulate metals following ingestion with food results in the formation of metal granules in cell mitochondria by the caddis fly Ecclisomyia sp. Although metals are regulated in the environment based on the concentration of metals that are dissolved in water, this study shows that food is important in the transfer of metals in aquatic food webs.

Investigation of the effects of urbanization on
hyporheic flows in Puget Lowland streams

Cathy Reidy, Sandra Clinton, and Susan Bolton
creidy@u.washington.edu

With the listing of several salmon species as endangered or threatened under the Endangered Species Act, local and federal agencies are providing funding to restore salmon habitat in Puget Sound streams (Washington). Sucessful salmon recovery, however, can only occur with a full understanding of the baseline ecological condition of these systems. The role of shallow groundwater-surface water interaction zones, (hyporheic zones) has generally not been integrated into these efforts. Growing awareness of the importance of hyporheic flows in forested regions has caused many scientists to wonder if the same is true in urban areas. Data from a current study examining urban stream temperatures suggest that generic urbanization causes neither discernable increases in stream temperature nor systematically drier streams in the late summer. These findings lead us to suspect that the hyporheic zone in urban streams is playing an, as of yet, unexamined role in moderating urban stream ecology and hydrology. The primary goal of this study is to investigate how increasing urbanization influences subsurface-surface water interactions. We will use transient storage models to estimate the relative size of the hyporheic zone in 5 sites: 1) forested/hot, 2) forested/cold, 3) urban/hot, and 4) urban/cold. The fifth site is located on a forested urban stream that is slated for housing development. We will quantify the transient storage using injections of a conservative tracer (4 seasonally-distinct times during the year) to evaluate how changes in surface water discharge affect hyporheic exchanges. Data on stream temperature, nutrient fluxes and dissolved gas levels are also being collected. A comparison of results across the five different study sites should provide insight into the effects of urbanization on hyporheic zones.

Modeling forest road surface erosion using the
Water Erosion Prediction Project (WEPP) model

Hakjun Rhee
hrhee@u.washington.edu

The Water Erosion Prediction Project (WEPP) model was used to predict sediment from the roads in a timber sale on Boise National Forest, Idaho. An outsloped 4.4 km road was divided into road segments of different lengths, using three different level of detail (high, intermediate, and low), and the WEPP model was run for each road segment. For the high detail method, the road segments were identified by changes in road grade and azimuth angle, resulting in 145 segments with an average length of 31 m. Buffer slope was determined from 40 ft (12 m) contours along the flow path, with an average buffer length of 370 m. For the intermediate detail method, the criteria was less than 2 percentage points of road grade difference between, resulting 91 segments with an average length of 49 m. Buffer slope was measured where major slope breaks occurred along the flow path, with an average buffer length of 360 m. For the low detail method, the road segments were identified by road grade reversal, resulting in 42 segments with an average length of 110 m. Buffer slope was measured as an average overall slope between road and stream, with an average buffer length of 320 m. The results of the study indicate that road grade reversal is a sufficient criteria for acquiring input information for sediment prediction from roads. Modeling sediment delivery to stream, however, requires consideration of major slope breaks in buffer slope.

Fostering future conservationists: An interpretative walk along North Creek

Jenna Scholz
jscholz@u.washington.edu

Salmon conservation is an increasingly important issue within the Puget Sound region. We in western Washington have a unique opportunity to educate and energize young minds interested in environmental issues about the importance of these remarkable animals. The proximity of schools, homes, and businesses to existing yet potentially threatened salmon streams gives us the opportunity to inform our community and educate our children about both the life history of salmon and the impact of continued urbanization on their habitat.

A salmon education project for middle school students (grades 8-11) is underway at the North Creek School in Bothell, Washington. The goal of this project is to develop a 1/4 –mile interpretive trail along a portion of North Creek that flows through the school grounds. Signs relating to five salmon life history topics including species identification, water quality requirements, reproduction, habitat needs, and foraging will be placed along the trail.

The walk along North Creek will include both educational and interactive elements. Each sign will provide scientific facts and specific questions designed to stimulate teacher-student discussion. The signs will be located in areas where the students will have access to the stream and the opportunity to read a map, take water quality samples, observe fish, examine habitat (e.g. large woody debris or overhanging banks), or identify invertebrates. This hands-on experience will help students to understand how human activities affect salmon in an urbanizing environment. It will also highlight their personal role in ensuring the sustainability of wild salmon in the Pacific Northwest.

Bedload scour in bull trout spawning grounds: Influences of site
selection, channel morphology and watershed flow regime

Jeff Shellberg
jshellbe@u.washington.edu

The intent of this pilot study is to investigate the patterns of scour and fill along bull trout (Salvelinus confluentus) spawning grounds to determine if bedload movement could potentially influence embryo survival. Bull trout spawn in the fall, have shallower egg pocket depths than larger bodied salmon and have exceptionally long incubation periods. These facts led to the hypothesis that bull trout in the mountainous terrain of western Washington are affected by bedload scour induced by rain and rain-on-snow events. Three main research questions will be addressed to investigate the multi-scale factors influencing spawning gravel scour. First, what general factors influence spawning site selection (substrate, depth, velocity, cover, slope, hyporheic flow, general habitat type) and what are the patterns of scour and fill in these selected microhabitats? Second, how does the reach scale channel morphology (channel type, in-stream structure, sediment transport regime) affect scour and spawning habitat availability? Third, what influence does watershed hydrology (runoff timing, frequency, magnitude and duration) have on bedload scour during the incubation period? In order to address these questions, 80 sliding golf ball scour indicators were installed in three main watersheds in western Washington: the upper Cedar River, the North Fork Skykomish River and the South Fork Skokomish River. These watersheds all differ in their runoff patterns, channel morphologies, available habitat types and bull trout populations. The results of this study will shed light on the potential influence of fluvial processes on bull trout populations and species distribution as well as help resource managers identify restoration strategies for this threatened species.

Stream phosphorus transport

Sara Stanley
stanleys@u.washington.edu

In natural freshwater ecosystems phosphorus (P) availability often limits primary productivity. When P is present in excess, it may enhance the growth of algae, particularly nuisance species like cyanobacteria, and lead to eutrophication. Research suggests urbanized watersheds export more nutrients (especially phosphorus) than do low-density residential and undisturbed watersheds. The ability to model the movement of phosphorus through the environment would be a valuable tool for land use management, urban planning, and watershed restoration. The objective of this study is to determine to what extent phosphorus export varies with sub-watershed land use. Four highly resolved databases for three different stream types (highly urbanized-Thornton Creek, moderately urbanized-Swamp and North Creeks, and low-density residential-Issaquah Creek) streams are being developed. Preliminary results beginning from the middle of August indicate that North Creek has the highest phosphorus concentrations even though both Thornton and Swamp Creeks have a higher percentage of urbanized land. Land use type and stream concentrations of phosphorus was compared for several sub-basins in the Lake Washington-Lake Sammamish watershed. Results suggested there was no correlation for an urban landscape and stream P concentration, a negative correlation for a forested landscape, and a positive correlation within a 100m buffer distance from the stream for areas classified as natural.

Changes in developed land cover (1991-1998): Cedar River, Washington

Robert C. Wissmar, Douglas C. Pflugh, and Ray K. Timm
raytimm@u.washington.edu

The lower Cedar River basin of western Washington has experienced major land conversions along an urban-rural gradient. This basin lies within King County’s jurisdiction for managing population growth and providing open space for riparian habitats and other benefits. We addressed a major management question: what areas in the lower Cedar River basin are undergoing significant land cover changes that could impact riparian and aquatic habitats? The distribution of land cover changes was considered across physical, ecological, and political landscapes. Land cover within the basin was evaluated for change between 1991 and 1998. Significant increases in developed land cover in the more urbanized areas included the widening of a major highway on the river floodplains and conversion of isolated forest patches through infill development. More rural areas demonstrated increases in developed land cover which correspond with patterns of scattered low-density residential, clustered dense commercial and residential development near newly incorporated areas. Changes in land cover point to uncertainties for implementation of County programs for acquiring open spaces in riparian areas. Identification of changes in the extent of developed land cover classes, which infer changes in impervious cover, are designed to facilitate fine-resolution spatial hydrological modeling and future management.