Intensively Monitored Watersheds

for Salmon Restoration


Millions of dollars have been dedicated to the restoration of freshwater habitat since the listing of many populations of salmon in the Pacific Northwest in the 1990s. Little is known about the effectiveness of these efforts. The most effective means of determining the contribution of restoration projects to salmon recovery is to put into action experimental, watershed-scale evaluations. Several organizations in the Pacific Northwest have begun to establish such projects. This document describes a series of intensively monitored watersheds (IMWs) being established in Washington for the purpose of better understanding how salmon and trout respond to current approaches to restore habitat

This research is funded by the Salmon Recovery Funding Board (SRFB) with in-kind support from the Washington State Department of Ecology and Department of Fish and Wildlife, Weyerhaeuser, Lower Elwha Tribe, Environmental Protection Agency, and NOAA-Fisheries. Tasks for FY 2004 included:

  1. Developing study plans for the three coho, steelhead, cutthroat IMW complexes.
  2. Identifying opportunities for effectiveness monitoring of chinook salmon restoration activities.
  3. Ranking other basins for their potential as IMW basins.
  4. Continuing to develop a landscape classification scheme that will allow extrapolation of research results from the IMW basins to other unmonitored basins.

Annual progress reports will update the study design and describe the monitoring and restoration activities in each basin complex. Data analysis and interpretation will be published separately. Data collected by this project will be posted on this site as they become available. Inquiries should be directed to:

Dr. William Ehinger
Department of Ecology
360 407 6416

General Concept

The basic premise of the IMW project is that the complex relationships controlling salmon response to habitat conditions can best be understood by concentrating monitoring and research efforts at a few locations. The type of data required to evaluate the response of fish populations to management actions that affect habitat quality or quantity are difficult and expensive to collect. Focusing efforts on a relatively few locations enables enough data on physical and biological attributes of a system to be collected to develop a comprehensive understanding of the factors affecting salmon production in freshwater.

IWM is an efficient method of achieving the level of sampling intensity necessary to determine the response of salmon to a set of management actions. Evaluating biological responses is complicated, requiring an understanding of how various management actions interact to affect habitat conditions and how system biology responds to these habitat changes. The response of the fish is dependent on the relative availability of the habitat types it requires, which change through the period of freshwater rearing (Table 1), and the manner in which these habitat types are influenced by application of a management action. Further complicating the issue is the fact that the relative importance of each habitat type in determining fish survival changes from year-to-year due to variations in weather and flow, the abundance of fish spawning within the watershed and other factors. For example, smolt production can be dictated by spawning habitat availability and quality during years when flood flows occur during incubation and greatly decrease egg survival. However, during years of more benign flow conditions during egg incubation, population performance may be more influenced by the availability of food during spring and summer or adequate winter habitat. Untangling the various factors that determine performance of salmon and how these factors respond to land use actions or restoration efforts can only be accomplished with an intensive monitoring approach.

Initially, this project focused on coho, steelhead and cutthroat for several reasons:

  1. These species spend more time in freshwater (1-3 years) than most other species of anadromous salmonids. Thus, they should be more responsive to changes in the quality and quantity of freshwater habitat than species which only reside in streams and rivers for a short period of time (e.g. ocean-type chinook, chum, pink).
  2. In order to cause a change in smolt production, a fairly substantial change in freshwater habitat conditions across a watershed will need to occur. The relatively small size of the watersheds within which coho, steelhead and cutthroat complete their freshwater rearing will make it much more practicable to alter enough freshwater habitat to affect smolt production. Response to the treatments also should be detectable in a shorter period of time than would be the case with species requiring much larger watersheds to support their freshwater rearing requirements.
  3. Many of the restoration projects and land use regulations that have been implemented in the region have been based on the habitat requirements of coho salmon. Therefore, this species should be the most likely to respond to many of the restoration activities that are being funded.
  4. Because these species complete freshwater rearing in a smallwatershed, fish responses to management actions can be assessed using a before-after/control impact design. Use of this type of design should make the responses by the fish easier to detect. Such a design would not be possible with species requiring a much more extensive area to complete rearing.

Chinook Salmon

Chinook salmon require a substantially larger watershed to complete their freshwater rearing than coho, steelhead and cutthroat. The larger area required by this species makes it very difficult to use a treatment-reference comparison at the level of an entire watershed, as we are doing for the other species. However, because many stocks of chinook are listed under the ESA and they are an important commercial and sport species, there is considerable interest in understanding how this species responds to the application of various restoration measures. We have expanded the IMW effort to include watersheds that support large populations of the chinook using two approaches:

  1. We are in the process of integrating two existing projects assessing the response of chinook salmon to restoration measures into the IMW effort. One of these projects is a BPA-funded, watershed-scale evaluation of restoration efforts in the Wenatchee River. The Wenatchee River watershed includes both listed stream-type chinook and steelhead. We also have begun discussions with researchers conducting an ongoing monitoring effort on the Skagit River to evaluate the response of ocean-type chinook to restoration measures. Bringing the Skagit monitoring effort into the IMW project would cover the two, primary life history types of chinook in the region. Integration of these projects into the IMW effort will help ensure that data are collected in a compatible manner and information is shared.
  2. In addition, the IMW project is identifying watersheds where investigations on ocean-type chinook would be feasible. The screening process we used for identifying appropriate watersheds is described later in this document.

Objectives for Coho, Steelhead, Cutthroat Watershed Complexes

These three IMW watershed complexes vary in physical characteristics, land use patterns, climate and relative abundance of the focal species (Figure 1; Table 2). The variation in conditions will enhance our ability to extend our results to other watersheds. The variation in conditions also provides an opportunity to address a wider range of factors contributing to habitat degradation than would be the case if all watersheds were similar.

Figure 1. Locations of the three coho, steelhead, cutthroat IMW basin complexes, Straits Juan de Fuca (SJF), Hook Canal, and Lower Columbia, and two chinook salmon IMW's, Skagit Estuary and Wenatchee basin.

state map shoing IMW basins

Straits of Juan de Fuca

The watersheds in this complex (West Twin Creek, East Twin Creek, and Deep Creek) have been logged since early in the 20th century. As a result, much of the wood that historically created pools and regulated the movement of sediment and organic matter in these watersheds had been depleted. Wood loss contributed to channel incision at some sites, isolating the floodplain and reducing access to off-channel habitats. The primary treatment for this watershed complex will be the addition of wood to a large proportion of the channel accessible to anadromous fishes in Deep Creek and West Twin Creek. In addition, off-channel habitats will be developed at several locations. No treatments will be applied in East Twin Creek during the period of our evaluation.

Of all the watershed complexes, this location offers the best opportunity for maintaining the integrity of control and treatment watersheds. The watersheds are almost completely owned by USFS and one private forestry company. We have the full cooperation of both organizations. Relatively little timber harvest or road construction will occur in these watersheds over the next decade. Therefore, interpreting any responses of the fish to the restoration treatments at the watershed scale will not be complicated by other activities that might affect habitat condition.

Hood Canal

Land use in the four watersheds in this complex range from urban and residential in Little Anderson Creek to almost entirely forestry in Stavis Creek. We plan to implement restoration treatments in all the watersheds except Stavis Creek. The types of treatments applied will vary by watershed depending on the factors perceived to be limiting fish production. In Little Anderson Creek, lack of wood and off-channel habitat has been identified as likely factors constraining fish production. We are currently planning several restoration projects that will address these concerns. Seabeck Creek displays evidence of channel incision in some locations and significant amounts of sediment deposition in other channel segments. The incision in this watershed may actually be contributing to low summer flows by reducing groundwater storage. We are currently conducting a hydrologic assessment of this watershed to determine the potential for increasing summer flow by reducing incision in key reaches. Big Beef Creek has a small impoundment that impacts water temperature downstream and provides habitat for various warm water fishes that may prey on coho and steelhead smolts. As the factors most likely to be limiting fish production become evident, appropriate restoration actions will be applied and the fish response compared with Stavis Creek, where no restoration applications will be applied.

The watersheds in this complex offer us the best opportunity to evaluate the impact of urban and residential development on our ability to increase salmon production with restoration efforts. These watersheds also offer the advantage of being quite small making it possible to treat a significant proportion of the channel network relatively easily.

Lower Columbia

The available data for the watersheds in the Lower Columbia complex are not as complete as for the other complexes. Land use in the three watersheds is dominated by commercial forestry. Of the three complexes, these watersheds provide the best opportunity to assess the effect of commercial forest management on aquatic habitat and fish. Lack of large wood in the channels, reduction in off-channel habitat and alterations in sediment delivery and transport are likely to be factors that have influenced habitat conditions in these watersheds. Because the currently available habitat data is relatively incomplete, we have not yet determined which of the three watersheds would be most appropriate as a reference site nor have we begun to identify potential restoration projects. Activities during 2004 focused on obtaining the data necessary to complete the experimental design for this complex.

One potential positive aspect of these watersheds is the presence of populations of chinook and chum salmon, possibly providing us with an opportunity to evaluate the response of these species to restoration actions. However, there is some speculation that the chinook populations are supported primarily by hatchery strays, confounding any responses that might be generated by alterations in habitat condition. The extent to which chum salmon spawn in these watersheds is not clear and further assessment of their population would be required to determine the feasibility monitoring the response of this species.

Objectives for Chinook Salmon Studies

Wenatchee River

The Upper Columbia Regional Technical Team and Upper Columbia Salmon Recovery Board have designed a monitoring effort for the Upper Columbia Basin. The plan described here addresses the following basic questions:

  1. What are the current habitat conditions and abundance, distribution, life-stage survival, and age-composition of fish in the Upper Columbia Basin (status monitoring)?/programs/eap/images/beach_header.gif
  2. How do these factors change over time (trend monitoring)?
  3. What effects do tributary habitat actions have on fish populations and habitat conditions (effectiveness monitoring)?

The plan is designed to address these questions and at the same time eliminate duplication of work, reduce costs, and increase monitoring efficiency by coordinating current monitoring efforts conducted by the U.S. Forest Service, U.S. Fish & Wildlife Service, Washington Departments of Fish and Wildlife, and Ecology, Chelan County, and Chelan County Public Utility District. The coordination is overseen by NOAA Fisheries. The Wenatchee River is represented on the IMW Scientific Oversight Committee to ensure close cooperation and information flow among the various IMW efforts. However, no SRFB funding has been requested for the Wenatchee River.

Skagit River

In 1994 the Skagit River System Cooperative (SRSC) initiated field studies to understand wild Skagit Chinook fish habitat relationships for population recovery purposes. The studies were developed in the context of a lifecycle model framework that includes discrete life stages and habitats for multiple juvenile life history types of ocean-type Chinook salmon. Field studies include:

  1. Identification of juvenile life history types.

  2. Inventories of current and historic habitat conditions.
  3. Fish use patterns for freshwater, estuarine delta, and Skagit Bay near shore life stages.

Results after a decade of study show:

  1. A strong negative relationship of peak flow during incubation with egg-fry survival.
  2. A large historical loss of delta estuarine habitat and a high percentage of wild juvenile Chinook positioned to utilize this habitat for extended rearing.
  3. Evidence for density dependence in the delta and possibly freshwater habitat areas.
  4. Density-dependent movement by individual migrants.
  5. Strong seasonal preferences in nearshore habitat utilization.

The results of the field studies lead independently to a solid biological rationale for for specific recovery actions that would benefit specific juvenile life history types. However, it is critical to understand how chinook salmon populations respond to recovery actions, to be extrapolate these results to other estuaries within Puget Sound and elsewhere that have been lost to agriculture and urbanization. We will work with the SRSC to develop an integrated monitoring plan that will complement the existing monitoring and focus the effects of population response to estuary restoration.

Experimental Design

Multiple experimental designs will be utilized in the IMW watersheds, depending upon the question being addressed and the scale at which the assessment must be conducted. Regression analysis has been used to investigate the relationship between stream flow and smolt production and this approach will also be used in the IMW investigations. Many of the objectives of the IMW project can be addressed by a before-after/control-impact (BACI) experimental design. This type of design enhances the ability to differentiate treatment responses from responses due to variations in weather or other factors not directly affected by the treatments.

At least one watershed will serve as a reference site where no experimental treatments are implemented during the study. A calibration period prior to applying treatments is required to determine how the reference and treatment watershed compare in the key response variables prior to any habitat manipulation. The length of time required to develop this baseline will vary among watersheds. The calibration period for sites with existing information on spawner abundance and smolt output would be much shorter than for watersheds where these data have not been collected.

Treated and untreated sites can be paired at a multiple spatial scales within the IWM design, the scale dependent on the question being addressed. In fact, reference sites for some reach-level projects could be within the basin designated for treatment. These reference sites would consist of portions of the basin comparable in initial condition to the location where a restoration action is applied but where no habitat manipulation would occur during the period of evaluation. Questions that can be addressed at this finer scale include life-history specific biological responses or physical habitat responses to management actions. For evaluations of effects at the scale of the entire basin, a comparison with a nearby basin that is not undergoing treatment is required. Therefore, the IMW approach does require sufficient management discipline to ensure that reference sites remain untreated through the duration of the study. This does not imply that any management activities in the reference watershed will compromise the integrity of the study. The validity of the study design will be maintained provided that the management activities not directly related to the restoration actions being evaluated are comparable at the reference and treated locations. For example, the effectiveness of restoration actions can be evaluated in watersheds being actively managed for wood production provided that the type and intensity of forest management activities in the treated and reference watersheds are comparable.

Variables Measured

The specific parameters measured in each watershed will vary depending on the questions being addressed and the types of treatments being applied. However, a basic set of data will be collected at all of the watersheds (Table 3). These common measures are of attributes that reflect the impact restoration actions are having on a watershed scale or of factors that provide context for interpretation of response to treatments. These measures include measures of water quantity and quality, habitat characteristics and characteristics of the fish populations.

Extension of Results to Other Watersheds


Because only a few watersheds can be included in the IMW project, extension of the results to other watersheds cannot be accomplished by the traditional method of increasing the sample size (number of watersheds monitored) until a sufficient level of statistical certainty is achieved. We will determine the applicability of our results by classifying watersheds across Washington State based on similarity of physical and biological characteristics in relation to the ten watersheds included in the IMW project. Watersheds which have biophysical characteristics and patterns of human activities comparable to IMW sites will be locations where IMW results can be extended with the greatest degree of certainty.

Identifying Other Potential IMW Opportunities

In order to measure population-level response to restoration activities, consideration was given to all watersheds that met the following criteria:

  1. Watersheds containing populations of naturally-produced anadromous salmonids, and
  2. Watersheds draining directly to saltwater, or
  3. Watersheds draining directly into Lake Washington/Sammamish, or
  4. Watersheds draining directly into the Columbia/Snake Rivers.

This list was first partitioned into Statewide Salmon Recovery Regions and then into large (>500 km2), medium (125 to 500 km2), and small (<125km2) watersheds. Basins were then ranked according to criteria that included the number of listed salmonids species, data record, and the feasibility of collecting accurate smolt production data. Results are included in the progress report.


Table 1. Habitat requirements of coho salmon during freshwater rearing.

The changing requirements of the fish stress the need to develop monitoring designs that evaluate responses at a spatial scale large enough to encompass the full range of habitat types required by the fish to complete freshwater rearing.

Life History Stage Habitat
Spawning and egg incubation Gravel bedded riffles and pool tail outs in proximity of cover suitable for adult spawners (e.g., deep pools, undercut banks, debris jams)
Early fry rearing Low velocity areas with cover in close proximity to food source. Typically associated with shallow, channel margin habitat with cover from wood and overhanging vegetation
Summer rearing Pool habitat with cover in close proximity to food source. Typically found in low gradient channels with a pool/riffle morphology
Winter rearing Low velocity areas with cover. Often associated with
off-channel habitat on floodplains including low gradient tributaries, secondary channels and ponds

Table 2. Characteristics of the three coho, steelhead, cutthroat IMW complexes in western Washington.

  Straits of Juan De Fuca Hood Canal Lower Columbia
watersheds West Twin
East Twin
Stavis Little Anderson Seabeck Big Beef Germany Abernathy Mill
focal species coho steelhead cutthroat coho cutthroat steelhead coho steelhead cutthroat (chum, chinook)
land use forestry private, state, and federal urban, rural, residential, forestry private and state forestry - private and state
total area 111 km2 75 km2 206 km2
geology mixed sedimentary and metamorphic glacial till flow basalt w/ interbedded sandstone
precipitation 190 cm 105 cm 160 cm

Table 3. Variables measured in all coho, steelhead, and cutthroat watersheds.

Variable Frequency Collection period
Flow Continuous Begin June 2004
Climate Continuous Begin Summer 2004
Water temperature Continuous Begin Summer 2004
Water chemistry Monthly Begin October 2004
Hankin & Reeves survey Annual July-August
Probabilistic sampling Annual (Hood Canal and SJF only) July-August
Smolt production Annual March-June
Juvenile abundance Annual July-August
Spawners Annual (varies by species)


* Also see Environmental Information Management (EIM) System.