Introduction

The recent major accomplishments of this project are described in this section.  These accomplishments are organized under headings consistent with the project objectives described in this proposal, which include: 1) improve stream habitats; 2) track trends in westslope cutthroat trout demographics and population structure; 3) evaluate effectiveness of habitat restoration; 4) address impacts from non-native introduced fishes; and 5) increase coordination and participation among stakeholders.  To avoid redundancy in presentation, several key initiatives to improve stream habitats are described under the heading “Evaluate Effectiveness of Habitat Restoration” to emphasize the role of this type of monitoring in informing the approach and direction of management actions over different spatial scales and timeframes (i.e. adaptive management).  The representative deliverables for this recent work are referenced by contract and work element where appropriate.

Objective 1: Improve Stream Habitats

Prioritizing Restoration

The approach to restoration in the target tributaries has evolved over time to more successfully translate watershed analyses, resource inventories and assessments and monitoring results into the management actions needed to achieve project goals (Contract 42560/WE C114; Contract 33533/WE D115; Contract 37842/WE D115).  The recent project history reflects a shift from opportunistic implementation of restoration projects to a more systematic approach for prioritizing management actions consistent with the refugia approach described by Reeves et al (1995) and Frissell and Bayles (1996) and a multispecies, analytical approach (Beechie and Bolton 1999).  The former approach is rooted in the idea of protecting the best first and expanding restoration outward from areas of relatively intact habitats and populations.  The approach acknowledges studies that indicate that restoration efforts focused near sources of colonists result in more-rapid species recovery (Huxel and Hastings 1999) and that recovery time increases with distance from colonization sources (Gore and Milner 1990).  The later approach has been implemented as more detailed knowledge of factors limiting recovery have been developed and focuses on suites of landscape processes considered necessary to conserve multiple species.

A fundamental goal of the Coeur d’Alene Tribe Fisheries Program is to identify restoration and enhancement needs and opportunities in areas that have the greatest potential to improve habitat and translate into positive biological responses to recover depressed native cutthroat trout populations.  Within this context we are interested in answering the question, “What are the highest priority restoration actions at the watershed scale and at finer spatial scales?” To help structure the process of identifying and prioritizing restoration actions we utilized a four-step process that connects watershed analyses and status and trend monitoring to prioritization through 1) setting clear goals and objectives for restoration activities, 2) selecting a prioritization scheme that is consistent with the goal, 3) using watershed assessments to identify restoration actions, and 4) prioritizing the list of actions (Beechie et al. 2008).  These steps fit within the broader restoration process that we have used in developing other programmatic plans, such as the Fisheries Program Management Plan (Lillengreen et al. 1999) and Research Monitoring and Evaluation Plan (Vitale et al. 2002), which includes restoration planning, implementation, and evaluating the success of restoration actions (Figure 1).

Figure 1.  Diagram of the restoration process and the steps used for identifying and prioritizing restoration actions that are nested within this broader process.

Restoration Goals and Objectives

Ranked aquatic resource goals for the Coeur d’Alene Subbasin were developed as part of the NPPC Subbasin Planning process.  The highest priority goal included protection and restoration of harvestable surpluses of naturally reproducing adfluvial adult fish from Lake, Benewah, Evans and Alder creeks and other populations well-distributed in tributaries throughout the basin (Intermountain Province Subbasin Plan 2004).  The restoration goal that is corollary to this subbasin goal is to “support recovery of resident and migratory westslope cutthroat trout through restoration of landscape processes that form and sustain riverine habitat diversity, while managing the riparian/aquatic interface for both wildlife and limited domestic uses that do not conflict with protection of water quality, public health and the fisheries resource”.  This goal is stated in the context of landscape and aquatic processes that drive habitat degradation and species declines, as well as socioeconomic considerations, so as to be both realistic and explainable.

We developed specific process-based objectives and criteria for describing impairment to watershed process functions that would be useful in identifying the restoration actions needed to achieve the above goal and in prioritizing those actions (Table 1).  The watershed processes that were considered included sediment, flood hydrology, riparian and channel processes, water quality and biological productivity.  For each of these processes, criteria were developed that described the degree of impairment relative to the watershed or sub-watershed scale.  It was difficult to find suitable criteria in the peer-reviewed literature for all of these functions.  Where existing criteria were not available, we developed definitions based on results from our long-term status and trend monitoring (e.g., for biological productivity) and based on the range of measured values identified during watershed assessments (e.g., for sediment, flood hydrology, riparian and channel).  Ratings of high and moderate indicated a degree of process impairment warranting restoration action.

Table 1.  Restoration objectives for watershed process functions and definitions for process impairment criteria; ratings of high (H); moderate (M), and low (L) indicate the degree to which each impaired process alters riverine habitat conditions.

Prioritization Approach

There are a variety of available prioritization approaches, and selecting an approach that matches restoration goals and assessment capabilities is helpful in linking restoration goals, watershed assessments, and prioritization into a coherent strategy for river restoration (Beechie et al. 2008).  We selected a prioritization approach consistent with the logic approach (decision support system) described by Lewis et al. (1996), SRSRC (2004) and Cipollini et al (2005).  This approach utilizes an array of semi-quantitative tools for prioritizing restoration actions, including, information developed from watershed assessments that describe causes of impairment, biological benefits associated with classes of restoration actions, as well as estimated costs.  The fundamental objective is to assemble and weigh information considered important to setting priorities (Cipollini et al. 2005).  The approach is considerably more flexible than others that were reviewed and allowed for incorporating values important to the Coeur d’Alene Tribe and local landowners.

A decision support system score sheet was developed to obtain a relative “score” for each planned project.  Criteria were drafted to reflect the values embodied in the goal statement as well as the constraints of implementing projects within the target watersheds.  The criteria include consideration of species that benefit from restoration, the degree to which restoration actions address causal processes, uncertainty associated with project actions and habitat/biological responses, and how the project accommodates local socioeconomic goals.  The criteria are scored on either a discrete or continuous scale, as well as being weighted, then summed to a total score.  Total scores are useful in differentiating projects (Table 2).

In the initial use of this approach, we selected just two of the focal watersheds, Benewah and Lake creeks, to develop a list of projects and conduct the preliminary prioritization of restoration efforts.  These watersheds were chosen because they both have resident and adfluvial westslope cutthroat trout and relatively more restoration actions have been implemented compared with Alder and Evans creeks.  In proceeding with this approach we recognize the importance of watershed scale restoration as well as the value in maintaining a treatment/control approach to monitoring action effectiveness.  Within these two watersheds, 12 subbasins were delineated so that priorities could be viewed and implemented at multiple scales.  These subbasins encompass the distribution of cutthroat within the watersheds and contain the critical habitats for spawning and early life stage rearing.  The Lake Creek watershed includes three subbasins: Bozard, Upper Lake, and WF Lake; while the Benewah Creek watershed includes nine subbasins: Bull, Coon, Hodgson, Schoolhouse, SF Benewah, WF Benewah, Whitetail, Whittrock, and Windfall.

Identifying Restoration Actions

Watershed assessments and long-term monitoring data collected as part of this BPA funded project, provided most of the information needed to identify and prioritize restoration actions (Table 3).  The most recent assessments included 1) large wood recruitment inventory and analysis which examined existing in-stream wood loads, stream conditions and the wood recruitment capacity of riparian forests associated with more than 74 km of streams (Miller et al 2008); and 2) inventory and analysis of road conditions and fish passage associated with 540 km of forest roads and more than 400 stream crossings (Duck Creek Associates 2009).  These assessments provided the critical understanding of natural potentials as they relate to sediment, flood hydrology and riparian and channel function, and the degree to which restoration efforts can move habitats toward a re-expression of natural habitat capacity and quality (Poff and Ward 1990; Ebersole and Liss 1997; Frissell et al. 1997; Pess et al. 2003).

Table 2.  Example of a project score sheet to facilitate prioritization of a list of restoration actions.

Table 3.  Summary of assessments and inventories used to identify the condition of watershed processes and function.

In order to translate the watershed assessment results into a list of necessary restoration actions, we first prepared a summary to clearly identify which processes or functions were most impaired and most responsible for habitat degradation (Table 4).  The summary identifies the degree of impairment for each of the subbasins in the Lake and Benewah creek watersheds consistent with the definitions for process impairment that were developed and described above.  The summary of impairments was then translated into a list of restoration needs, which includes types of restoration actions, their locations, and approximate levels of effort needed to address each of the impaired processes (Table 5).

Prioritizing Restoration Actions

Prioritization of restoration actions is an important part of the overall exercise to ensure that limited restoration funds can be focused on actions that will have the greatest impact and locations that will receive the greatest benefit. To this end, the delineated subbasins were further ranked by relative restoration priority according to the overall level of impairment, proximity to restored habitats and the potential for increasing fish production (Table 5).  A weighted impairment value was calculated for each subbasin, wherein a moderate impairment rating was scored as 1 point and a high rating was scored as 2 points and the scores were summed.  Subbasins with the highest impairment values were considered higher priorities for restoration.  Where impairment values for subbasins within the same watershed were equal, the rankings were modified to favor priority for subbasins in closer proximity (connectivity) to restored habitats or with greater potential for increasing fish production.  This potential was indicated by the “current productivity distance”, defined as the difference in mean maximum/minimum cutthroat trout densities within the subbasin.

Table 4.  Summary of process impairments identified by watershed assessment in subbasins within the Lake Creek and Benewah Creek watersheds. Subbasins lacking assessment data are indicated by ND.

Table 5.  Summary of restoration needs and relative restoration priority by subbasin within the Lake Creek and Benewah Creek watersheds. Proximity to restored mainstem habitat is indicated as near (N) or far (F), where applicable.

A list of spatially explicit projects was developed to meet the stated process objectives for each of the highest priority subbasins (6 in Benewah Creek and 3 in Lake Creek).  A total of 105 projects were identified and prioritized in these subbasins.  Only two percent of the ownership in these project areas is Tribal, while 49 percent is owned by three private companies and an additional 39 percent is owned by 18 individual landowners (Figure 2).  Therefore, sharing information generated through this project and coordinating planning, implementation and monitoring, with the goal of increasing participation with affected landowners, becomes a critical component of meeting our goals and objectives for recovery.  Cumulatively, these projects affect 41.1 km of stream and riparian habitat (29.7 km in Benewah Creek, 11.4 km in Lake Creek), with fish passage projects expected to result in a significant short- and long-term response. The list of projects will be used over the next several years to negotiate landowner agreements for implementation, and serves as the core of on-the-ground work that is identified in this project proposal.

Figure 2. Number and distribution of planned restoration projects by ownership and project type.

Instream CWD Loading and Recruitment Processes

We completed a survey of 74.1 km of 2^nd order tributaries and adjacent riparian habitats in 2008 (Contract 33533/WE D115) with the goal of identifying conditions with respect to coarse wood recruitment (Miller et al. 2008).  The objectives of this study were to:

1. Measure existing instream wood loads and stream conditions to better understand fisheries production potential as it relates to riparian stand condition and wood recruitment potential;
2. Measure the wood recruitment capacity of riparian forests using growth, yield and recruitment models to examine the potential effects of management prescriptions on stream habitats; and
3. Draft a planning document to identify priorities for conservation and restoration based on the study results.

Riparian habitats in the study area were classified into relatively homogenous polygons according to variables describing channel confinement, channel slope, vegetation type and composition, and lidar-derived canopy height and canopy cover.  Field data was then collected for a representative sample constituting 26.6% of stream habitats in the study area to describe riparian vegetation, instream wood characteristics and channel attributes.  Non-surveyed riparian stands were populated with tree lists derived from field-sampled stands for each structural class.  We utilized several computer models to analyze the effects of three different management scenarios representing a range of current and potential future riparian conditions over a 150 year planning window.  To model growth, yield and mortality we used the Forest Vegetation Simulator (FVS) model, developed and maintained by the USDA Forest Service.  Data from 113 sampled stands containing 4,259 individual tree list records were used to drive simulations for silvicultural prescriptions associated with each management scenario at a 5-year time interval.  For the wood recruitment modeling effort, we chose a spatially explicit model for in-stream wood recruitment and accumulation that adapted existing riparian wood recruitment algorithms (e.g., Sobota et al. 2006, Van Sickle and Gregory 1990) into a GIS framework, so that model outputs could be incorporated into other GIS-based models to 1) track in-channel wood abundance based on integration over time of wood inputs and wood depletion by decay and fluvial transport (e.g., Benda et al., 2003); and 2) relate wood-piece characteristics and channel attributes to the likelihood and type of habitat function.

We found that wood volumes in Benewah and Lake creeks are significantly lower than the median values from comparable studies (Young et al. 2006; Fox and Bolton 2007) and the observed reach scale distributions of CWD are indicative of impaired recruitment processes.  In this study, we used spatially referenced survey data and spatially distributed models to characterize wood-loading potential and management effects at multiple scales.  We found differences in predicted tree mortality between stand types, which translate to stand-specific recruitment rates.  We found differences in habitat-forming potential with piece size and channel attributes, which translate to differences in the effects of wood between reaches.  The site-specific aspects of these channel systems suggest that effects of riparian management will also vary reach to reach in terms of both the contribution of CWD to the channel (Figure 3) and in the frequency of wood-formed habitats (Figure 4).  Additional detailed results from this assessment are described in the objectives section of this proposal under the headings Technical Background and Emerging Limiting Factors.

Figure 3. Comparison of modeled wood volume by channel segment under the No management and Idaho Forest Practices Act (IFPA) scenarios.  Scatter in these plots indicates differences in the proportional increase in wood abundance between segments and differences in the pattern of points between graphs indicate differences between management alternatives.  Points falling below the one-to-one line indicate a loss of modeled in-stream wood over time (from decay of current wood) and points above the line indicate an increase in modeled wood over time, from mortality of riparian stands.  The graph on the right compares modeled wood abundance at 150 years under the no-management and IFPA scenarios.  The greater the distance from the one-to-one line, the greater is the reduction of modeled wood abundance associated with thinning of the riparian stands.

Figure 4. Model predictions of basin average wood-formed pools for three management alternatives at 150 years (Miller et al. 2008).

We suggest that this limiting factor can be addressed through a combination of management actions, including additions of CWD, adoption of alternative management practices for some riparian areas, and conservation of areas with well-functioning riparian wood recruitment processes.  Furthermore, the availability of both fish density and habitat data in our study area provide avenues for use of forest growth and recruitment model results for conservation and restoration planning following the general recommendations of Beechie and Bolton (1999) and Roni et al. (2002).  From the perspective of habitat quality, management and restoration activities have two primary goals: 1) to develop and maintain well functioning riparian-channel interactions that promote a diverse channel environment with high-quality habitat; and 2) to improve conditions where current habitat quality is low.  The appropriate activities to achieve these goals vary with channel and riparian zone attributes and current condition.  We identified three primary attributes, each divided into nominal rankings, to aid in determination of appropriate actions and in setting of priority levels, as follows:

1. Fish abundance, defined as the average surveyed number of fish (cutthroat trout) per unit channel length.  Rankings are:

• High abundance: > 10 fish/100m
• Moderate: 5 – 10 fish/100m
• Low = < 5 fish/100m

2. Current in-stream wood abundance, in terms of wood volume per unit channel length.  Rankings are:

• High abundance: > 9 m³/100m
• Low: < 9 m³/100m

3. Sensitivity to riparian management (thinning from below) in terms of the modeled difference for in-stream wood abundance (in terms of number of pieces per unit channel length) in 150 years with no thinning (no-management scenario) and with riparian thinning consistent with Idaho-Forest-Practices rules.  Rankings are:

• High sensitivity: > 30 pieces/100m difference
• Moderate: 30 – 10 pieces/100m difference
• Low: < 10 Pieces/100m difference

Each attribute is associated with specific types of management and restoration actions and unique combinations of these three attributes and their rankings provide 18 categories (Table 6).  Management and restoration actions appropriate for the rankings in each category are then identified, with four resulting management categories: conservation, riparian management, addition of large woody debris with riparian management, and no action.  Priority levels are based on the potential to maintain high and moderate fish abundances in light of sensitivity of future wood loading to riparian management.  We have added two additional variables to further condition management guidelines and prioritization:

1. Channel distance to the nearest reach with high fish abundance. Channel reaches with low fish abundance, but within 500 meters of reaches with high fish abundance, receive higher priority rankings than those further than 500 m. This conditional change is based on the potential for fish from high-abundance reaches to utilize these nearby reaches as habitat conditions improve.
2. Modeled wood abundance under the no-management scenario at 150 years. Reaches for which modeling indicates low potential for future wood recruitment even under no management, identified by low wood abundance (< 9 m³/100m) at 150 years, default to the high priority rating.

Table 6.  Categorical summary of large wood recruitment results based on unique combinations of nominal attribute rankings with recommendations for management actions and priority.

Applying this strategy to the target watersheds provides detailed maps of management options and conservation/restoration priorities that can be addressed systematically at either the watershed or sub-watershed scale and be coordinated with implementation of other management actions.  The overall effect of this recent study is to strengthen the restoration approaches we have adopted by applying detailed knowledge of factors limiting recovery and focusing management activities on suites of landscape processes considered necessary to conserve multiple species.

Forest Road Condition and Fish Passage Assessment

We completed an inventory and assessment of 539 km of forest roads and 407 stream crossings (Contract 37842/WE D115) in portions of four target watersheds in 2009 (Middel et al., 2009).  The streams targeted for this survey included the most important habitats for spawning and rearing of westslope cutthroat trout in the respective watersheds.  Roads in the survey area are variously managed by tribal, state, county, and private landowners.  These roads provide access to a wide variety of activities important to Tribal members including access to traditional hunting and gathering areas, recreation, cultural/spiritual practices, timber harvesting, and access to private property.  The objectives of the study were to: 1) evaluate sediment contributions from roads that are within proximity to critical areas for spawning and rearing habitat; 2) identify any complete or partial barriers that may affect the ability of native westslope cutthroat trout to access key spawning and rearing habitats; and 3) draft a planning document to identify priorities for restoration based on the study results.  The methods and key results are summarized below, with additional detailed results described in the objectives section of this proposal under the headings Technical Background and Emerging Limiting Factors.

In this study, we applied both the State of Oregon and the State of Washington methodologies to help determine the current and near-term condition of roads in the four study watersheds.  The State of Oregon’s “Rapid Watershed Risk and Current Condition Survey” (Mills et al. 2007), was used to determine the current and near-term condition of roads by rating road segment attributes and features with attention priority codes.  These codes range from 1 to 5, with 1 indicating a need for immediate road work and 5 indicating that the road is in good working condition.  Different environmental indicators were examined including location of critical areas, locations of hydrologically connected roads, stream crossing washout potential, condition of stream crossing structures, and condition of the road prism.  Hydrologic connectivity occurs when water intercepted by the road prism flows from a road directly to a stream or waterway.  The second model used for surveying roads was the Washington Road Surface Erosion Model (WARSEM) (Dube et al. 2004).  This model predicts relative long term soil erosion and sediment delivery from roads.  In this case, road segments that contributed sediment to stream were those that drained to an active stream channel, drained to a gully that drained to an active stream channel, or road segments that drained to areas that were located less than 200 feet from a stream.  Geology, topography, precipitation, land use, and road characteristics are important variables used to estimate sediment delivery in WARSEM and these were derived using local values.

The results of the two different methodologies were combined to guide planning efforts for restoration.  Priority road segments were identified for inspection and potential repair based on a positive hydrologic connectivity classification and an attention priority rating of 1 or 2.  There were 4.7 miles of roads that fit this category.  High sediment delivery road segments were also assigned high priority.  A range of best management practices were identified to address these road condition priorities, including installation of cross drains, culvert replacement, reducing road gradient, increasing vegetation on cutslopes, and improving surface conditions.  Reducing hydrologic connectivity of roads to streams is an especially important objective of forest road management and one that depends on engineering strategies which effectively divert road surface runoff to the forest floor where it can be filtered before entering waterways.  We identified all such opportunities for “disconnecting” existing road segments from the drainage network.  Furthermore, the list of BMP’s we identified was evaluated as to their potential effectiveness in reducing sediment delivery as a means of further prioritizing future restoration actions.  For example, WARSEM predicts a greater than 60% reduction in sediment loading to streams when resurfacing from pit run to gravel and keeping all other attributes the same.  This type of resurfacing has no effect on the cutslope sediment delivery.  Another scenario we explored involved upgrading from a native surface to a gravel surface.  In these instances, WARSEM predicts an overall reduction from 209 mean tons/annually to 45 mean tons/annually.  When we applied a BMP of providing 90% - 100% cover to cut slopes, either by applying a hydro-mulch or straw on all cut slopes with less than 90% cover, WARSEM predicted a 36% reduction in sediment loading to creeks.  Road gradient has a noticeable effect on sediment loading to creeks and we considered the effect of reducing road gradient on segments that were greater than 10%.  WARSEM predicted a 60% reduction in sediment loading to creeks when road gradient was reduced from > 10% to 5 – 10%.  The compilation of data derived from this evaluation exercise was particularly useful in identifying process-based objectives for sediment reduction.  Additionally, it provides important information for monitoring the application of BMP’s such that estimates of sediment delivery can be tracked and modeled to allow for before and after comparisons of project effectiveness.

Physical stream data and culvert information was collected at each of 407 sites to assess fish passage.  A fish passage screen developed by the US Forest Service Northern Region (Hendrickson 2008) was used to do a preliminary assessment of fish passage for adult and juvenile cutthroat trout at each stream crossing that was determined to be fish bearing.  Fish bearing streams were defined as having less than 20% gradient and more than 3 feet bankfull width (WDFW 2000).  This methodology was chosen primarily because it was developed by targeting resident-adult and resident-juvenile westslope cutthroat passage at existing road-stream crossings.  Where the Northern Region Fish Passage Screen proved inconclusive, we utilized methodology from the National Inventory and Assessment Procedure for identifying Barriers to Aquatic Organisms Passage at Road-Stream Crossings (NAIP) developed by Clarkin et al. (2005).  The data gathered for these inconclusive sites, 70 crossings in total, were input into the computer software FishXing (Beta V.3) to model fish passage.  For modeling purposes, fish lengths of 23 cm for adults and 10 cm for juveniles were used.  Low flow and high flow discharge was estimated using USGS regional regression equations developed for Idaho.  A system was developed to rank each potential fish bearing crossing by watershed.  A combination of barrier status and extent of habitat upstream was used to rank the identified barriers as low, moderate, or high priority for replacement.  Fish barriers were ranked first according to the results of FishXing.  If the percent of passable flows for adults was less than 50%, the crossing was given a higher priority.  For those crossings not modeled in Fish Xing, a crossing that was identified as being both an adult and juvenile barrier was given higher priority than a crossing identified just as a juvenile barrier.  Next, the amount of stream length upstream of the barrier that was less than 20% grade was determined.  If this length was more than 1500 meters, a crossing was given a higher priority.  Twenty four out of 121 identified barriers were given a high priority ranking (Figure 5).  Development of this specific list of priorities will help to identify key landowner contacts to develop on-the-ground projects and to facilitate project implementation in areas where barrier removals should illicit the greatest biological responses.

Figure 5.  Percentage of low, moderate, and high priority stream crossings per watershed based on total number of fish barriers identified during the 2008 fish passage assessment.

Objective 2: Track Status And Trends In Westslope Cutthroat Trout Demographics And Population Structure

Adfluvial Spawner Abundance

It is imperative that we have the capability to reliably track temporal changes in adfluvial spawners given that one of the primary objectives of our recovery efforts is to augment the number of returning adult cutthroat to our adfluvial watersheds.  As such, sampling techniques for capturing and tagging fish have been improved since 2004, and as a result, we have been able to obtain more accurate and precise spawner abundance estimates.  Beginning in 2005 in Lake Creek and in 2007 in Benewah Creek, we replaced the fixed-weir trap design with a floating weir design to intercept upriver migrating spawners (Tobin 1994; Stewart 2002; Photo 1).  Owing to its flexible nature, the floating weir can more effectively accommodate high discharge and debris loading than the former design, and was able to be fished at a greater range of flows in our watersheds throughout migratory periods.  However, because brief periods of high spring discharge were still observed to depress panels below the water surface permitting upriver migrants to pass and giving rise to variable capture efficiency across years (e.g., Lake Creek, 2005-2008; Table 7), additional changes were respectively made to floating weirs in Lake and Benewah creeks in 2009 and 2010 to further improve performance.  Floating weirs were modified so that the panels could be manually elevated or lowered with a hoist to maintain their position above the water surface at a much greater range of discharge levels than before.  Modifications were also made to the fixed-weir traps used to intercept post-spawn outmigrating adults in Lake and Benewah creeks in 2005 and 2007, respectively.  Removable pop-out panels were incorporated into the trap design to alleviate hydrostatic pressure on the trap during high flow events (Photo 2).  Although trap performance is sacrificed during brief high flow events when pop-outs are removed, structural trap damage and consequent lost trapping opportunities during rebuilding efforts are prevented.  Trapping modifications likely explain the greater numbers of adult cutthroat captured in Lake Creek in more recent years and the consistency in trapping efforts since 2009, and, also explain, in part, the recent increase in adults captured in Benewah Creek since 2007 (Table 7).

Because of the increased number of adults captured in our traps, most notably as ascending adults at our floating weirs, we were able to obtain a sufficient sample size to initiate a mark-recapture program to estimate spawner abundance.  Beginning in 2009, we began to opercle punch every ascending adult intercepted at our floating weirs, and from the recapture of a considerable percentage of these marked fish as post-spawn outmigrants in our fixed weir traps, have been able to obtain rather precise abundance estimates in recent years (Table 7).  Currently, our migrant traps are installed at rkm 6.0 on the mainstem of Lake Creek and at river kilometer (rkm) 14.5 on the mainstem of Benewah Creek.  In both watersheds, traps have been installed downriver of principal spawning tributaries and of most of the recently implemented and projected habitat restoration projects.  In Benewah Creek, however, several perennial spawning/rearing tributaries exist below our trap location, and as a result spawner abundance estimates obtained at the current trap location do not reflect the production potential of the entire watershed.  As a result, we are currently in the process of installing a floating weir at the mouth of Benewah Creek that has the capability to trap both ascending and descending spawners.  This trap is expected to be operable in the spring of 2012 and our marking protocol will be employed to obtain a spawner abundance estimate for the entire watershed.  The ability to obtain rather precise estimates of annual adult abundance should permit us to reliably assess the status of adfluvial spawners in our watersheds and track trends in this high-level indicator over time.

Photo 1.  Floating weir used to capture ascending spawners at river kilometer 6.0 in Lake Creek.  Photo does not depict the hoist that modified the trap design in 2009.

Table 7.  Abundance estimates and number of fish captured for adfluvial cutthroat trout spawners and outmigrating juveniles in Lake and Benewah watersheds, 2004-2010.  Captured adults were intercepted as ascending spawners and descending post-spawn outmigrants. 

Photo 2.  Fixed weir used to intercept post-spawn adults and outmigrating juveniles at river kilometer 6.0 in Lake Creek.  Close-up of pop-out panel insert is depicted in the inset photo in upper-right corner.

Adfluvial Juvenile Outmigrant Abundance

Our status and trend monitoring program also tracks adfluvial juvenile production in Lake and Benewah creek watersheds.  In combination with our adult spawner estimates, juvenile outmigrant abundance estimates and associated age structure information will permit the derivation of outmigrant per spawner ratios, a watershed-wide indicator that would allow tracking of trajectories in juvenile production in addition to aiding in the assessment of in-stream population response to our restoration actions (Bradford et al. 2005).  Outmigrating juveniles are captured with the same fixed weir traps that are used to capture post-spawn adults.  The greater numbers of juveniles captured in the Lake Creek trap since 2005 and in the Benewah Creek trap since 2007, compared to earlier years in both watersheds, likely reflect our improvements to the outmigrant fixed-weir trap design that were described above (Table 7).

Because of the increased success in capturing juveniles, we began a PIT-tagging program in 2006 in Lake Creek and in 2008 in Benewah Creek to annually generate outmigrant abundance estimates.  Throughout the outmigration period, subsamples of captured juveniles (e.g., 40-80) were periodically tagged and released upstream of the trap to estimate capture efficiencies (Carlson et al. 1998).  Release trials were typically conducted every 4-6 d, with a subsample of fish from each release trial held overnight before release to evaluate tag retention and survival rates (since inception of the release trial protocol, we have not had a tag shed nor a mortality).  Generally, under amenable flow conditions, our capture efficiencies have exceeded 90% which give rise to rather precise outmigrant estimates (e.g., flow years of 2009 and 2010 in Lake Creek, Table 7).  However, under years with extended high flow periods (e.g., 2008, Table 7) capture efficiencies can be extremely low which decrease our confidence in the overall abundance estimates.  Furthermore, notwithstanding the inability to capture fish during periods of suspended trap operation during high flow events, we consistently are unable to effectively deploy traps early enough to sample the early part of the outmigration as evidenced by the large numbers of juveniles typically captured immediately upon trap installation.  Consequently, outmigrant abundance estimates are undoubtedly biased low in most years.  Given these concerns, we are considering an alternative mark-recapture protocol to address these biases that will be discussed more fully in our proposed deliverables.

Juvenile-to-Spawner Return Rates

The PIT-tagging program was also initiated to address the uncertainties that may be limiting survival rates of adfluvial cutthroat trout during lake residence.  Accordingly, full duplex PIT-tag pass-through antenna arrays have been installed in close proximity (5-10 m) and downriver of floating weirs in the mainstem of both watersheds to interrogate tagged fish and supplement information obtained from adults captured in our traps.  Given that the fixed PIT-tag antennas span the entire channel width in Lake Creek and interrogate the wetted channel in Benewah creek under most flows, the likelihood of detection would be great as upriver spawners linger in the vicinity of the detection field as they attempt to negotiate the trap.  Further, tagged juveniles released as test batches upriver of the antenna array in Lake Creek demonstrate detection rates of 95-100% (Firehammer et al. 2010).

As of 2010, 4545 and 757 juveniles have been tagged in Lake and Benewah creeks, respectively.  Interrogation data from these tagged fish suggest that a low percentage of outmigrating juveniles return to spawn as adults.  Of the 2272 juveniles that have been tagged from 2005 to 2007 in Lake Creek (cohorts for which enough years have elapsed to evaluate return rates for juveniles with variable lake residence times), only 1.7% have been uniquely detected.  Recent interrogation data from tagged Benewah creek outmigrating cohorts reveal similar results.  Although empirical estimates of in-lake survival rates for adfluvial cutthroat trout are scarce, several studies have provided values with which comparisons may be drawn.  Annual survival rates of 49% were estimated in Lake Koocanusa for cutthroat trout from reservoir entry as juveniles to first time spawning two years later which equates to around a 25% return rate (Huston et al. 1984).  Gresswell et al. (1994) estimated a 16-25% return rate for adfluvial juvenile Yellowstone cutthroat trout emigrating from Arnica Creek in the Yellowstone Lake system in the early 1950’s.  Compared with these studies, our juvenile-to-spawner return estimates are substantially lower.  We also began PIT-tagging ascending spawners (tag placement in pelvic girdle) in 2009 that were captured in our floating weirs to evaluate post-spawn survival rates and return frequency of adults.  Interrogations from the 2010 and 2011 seasons indicate that approximately 45% of the adults tagged in 2009 have already returned as repeat spawners.  Apparently, our data suggest that juvenile survival, but not post-spawn adult survival, is limiting in-lake adfluvial production in our watersheds.  Demographic modeling analyses for cutthroat trout have also found in-lake juvenile survival to be a key vital rate in determining overall population growth (Stapp and Hayward 2002).

Stream Densities of Cutthroat Trout

Our monitoring program has also conducted population surveys at established index sites distributed across tributary and mainstem reaches to evaluate cutthroat trout abundance trends at a much finer spatial scale than that attainable using our migrant trap data.  Trend trajectories permit an examination of whether conditions appear to be improving or declining at local tributary, watershed, and regional scales.  Trend monitoring also permits a description of temporal changes in spatial distributions to assess expansion rates of cutthroat trout populations to examine whether newly created suitable habitat is undergoing colonization.  In addition to our adfluvial watersheds (Benewah and Lake creeks), these surveys are also conducted in Evans creek which supports a prevailing resident cutthroat trout population and serves as a reference watershed for our monitoring program within the basin given its more suitable rearing habitat for cutthroat trout.  Our monitoring protocol entails electrofishing index sites during baseflow periods, and has employed multipass-depletion sampling procedures to obtain linear density estimates (fish / 100m).

Surveys conducted since 2004 generally reflect a spatial pattern of cutthroat trout distribution within each watershed that is consistent across years.  In Lake and Benewah creeks, cutthroat trout were primarily found in tributary rather than mainstem reaches, though distributions within tributaries varied between the two watersheds.  Whereas in Benewah Creek densities of cutthroat trout were typically similar across sites within each of the surveyed tributaries, cutthroat in Lake Creek were most often captured at uppermost tributary index sites, with estimated densities at least 4-5 times greater in upper than in lower reaches (Figure 6).  In comparison to our adfluvial watersheds, cutthroat trout in Evans Creek were spatially distributed across the watershed and typically observed at comparable densities in both mainstem and tributary index reaches during annual surveys.  These distributional data have allowed us to identify reaches in both Benewah and Lake creeks where suboptimal rearing conditions may be present and has guided prioritization efforts for implemented and propective habitat improvements in mainstem and tributary reaches.

Index site abundance data collected from 2003 to 2009 also revealed the presence of temporal trends in age one and older cutthroat trout in our monitored watersheds, though the abundance trajectories varied among the systems surveyed.  Synchronous trends in cutthroat abundance among reaches were detected in both the Evans Creek watershed and the upper Benewah Creek watershed as supported by a repeated measures analysis.  In Evans Creek, though there was evidence of a linear increase in abundance over time, a greater portion of the annual variability in abundance was explained by a cyclical trend (Table 8).  Generally, densities were found to decrease from 2003 to 2005, exhibit an increase from 2005 to 2007, and then decrease over the next two years, so that in half of the reaches densities in 2009 were not appreciably different from those in 2003 (Figure 7).  In comparison, cutthroat trout in upper Benewah Creek (i.e., upriver of the 9-mile bridge) displayed a more pronounced linear increase in tributary-wide densities over time (Table 8).  Densities from 2007 to 2009 were approximately twice that estimated from 2003 to 2006 in each of the five monitored tributaries (Figure 8).  Similar concurrent trends, however, were not apparent in tributary and mainstem reaches downriver of 9-mile bridge in the Benewah watershed (Table 8).  These results suggest that processes that influenced cutthroat trout demographics in tributaries in the upper Benewah watershed were operating similarly and contributing to an overall increase in juvenile abundance, whereas those that influenced abundance in the lower reaches may be operating independently from one another.  In contrast to Evans and Benewah creeks, a watershed-scale trend in cutthroat trout abundance in Lake Creek was not detectable over this time period, and reach trajectories did not exhibit similar patterns of change over time (Table 8).  For example, whereas densities in the Bozard tributary generally increased from 2003 to 2007, densities in the lower and upper reaches in the West Fork Lake tributary were respectively decreasing and high variable over the same time period (Figure 6).  This suggests that processes regulating juvenile cutthroat abundance may have been operating independently from one another in sampled reaches in the Lake Creek watershed.

Figure 6.  Depletion removal estimates (fish/100 m) of age 1 and older cutthroat trout for four reaches in the Bozard and West Fork drainages in the Lake Creek watershed, 2003-2009.

The positive trajectories in cutthroat trout abundance observed in both Evans Creek and tributaries of the upper Benewah watershed may in part be explained by a recent, regionally favorable environment where basin-wide stream conditions were conducive to spawning success and increased survival rates of early life stages.  Though the lack of a detectable similar watershed-scale trend in Lake Creek is not consistent with such an explanation, tributary densities in Lake Creek may be approaching carrying capacity with not much potential for further increase.  Concordant population abundances, indicative of regional climatic influence, have commonly been reported for small salmonid stream in other regional networks (Platts and Nelson 1988; Gowan and Fausch 1996).  On the other hand, the observed trend in cutthroat trout abundance in upper Benewah may have also been a collective response to the large-scale habitat restoration and the aggressive brook trout suppression program that have proceeded in upper Benewah reaches since 2004, given that the abundance trajectory in Benewah Creek exhibited a more linear profile than that in Evans Creek, a watershed that has received minimal management intervention in recent years.  As additional years of data are collected, further comparisons among watersheds will allow us to better evaluate whether population responses are the result of our remedial actions.

Table 8.  Summary of repeated measures analysis for series of standardized depletion estimates and first pass catch to detect trends in age one and older cutthroat trout over the years 2003-2009.

Figure 7.  Depletion removal estimates (fish/100 m) of age 1 and older cutthroat trout for six reaches in the Evans Creek watershed, 2003-2009.

Figure 8.  Depletion removal estimates (fish/100 m) of age 1 and older cutthroat trout for five tributary reaches in the upper Benewah Creek watershed, 2003-2009.

Though multipass-depletion estimates are useful for examining site-specific temporal trends, the large degree of within-year variability among site estimates that has commonly been documented in our surveys does not permit a reliable examination of abundance when estimates are expanded to larger spatial scales (Firehammer et al. 2009).  Because of this shortcoming and the desire to increase sampling efficiency, an index of site abundance would be preferable to an absolute estimate, providing that the index tracks true abundance over time.  As such, a study was conducted in 2009 at 23 of our sites, which varied in salmonid abundance, to examine the predictive abilities of a single-pass index.  Age 1+ fish were marked and released at each site during an initial pass (marked fish ranged from 2 to 78), and the following day the site was re-sampled using our multipass-depletion removal protocol.  Depletion estimates for marked fish were generated and compared to the actual number of marked fish released, and the precision of the relationship between marked fish captured during the first removal pass and known marked fish released at each site was examined.

A strong linear relationship was detected between numbers of marked trout recaptured in the first depletion pass and those released the day before across sampled sites (r = 0.95).  Others have also found single-pass indices to perform well in predicting abundances for salmonid populations in small-streams (Jones and Stockwell 1995; Kruse et al. 1998; Bateman et al. 2005).  More importantly, we found that depletion estimates underestimated the true abundance, primarily as a result of an overestimation of capture probability during subsequent passes, which was consistent with other studies that documented biases associated with depletion-removal estimates for salmonids in small stream systems (Peterson et al. 2004; Rosenberger and Dunham 2005).  Results from our repeated measures trend analyses also indicated that first pass catch data provided similar interpretations of watershed-wide abundance trends in cutthroat trout as did our removal-depletion estimates (Table 8).  These results lend support to using first-pass catch rather than multi-pass depletion estimates to examine long-term trends in our watersheds, and consequently, since 2010 we only conduct single passes.   Moreover, because of the reduced effort associated with single-pass efforts, we intend to expand our summer surveys across a greater percentage of our watersheds to better understand salmonid demographics.

Objective 3: Evaluate Effectiveness of Habitat Restoration

Barrier Removal and In-Channel Wood Additions

Barrier removals have been implemented in several locations recently as a priority for reconnection of isolated habitats and blocked tributaries, which has been demonstrated to provide a quick biological response, is likely to last for many decades, and has a high likelihood of success (Pess el al. 2005; Roni et al. 2008).  The first of these projects, completed in 2004, was a culvert replacement at the confluence of Windfall Creek and Benewah Creek (RM 11.5) to restore fish passage for adfluvial westslope cutthroat trout to 4,344 meters of high quality spawning and rearing habitat (Contract 10885/no WE reference).  Hydraulic analysis indicated the old 122 cm diameter pipe was a barrier for all flowrates during the migration period due to excessive leap height and velocity.  The old pipe was replaced with a 221cm x 160cm pipe-arch placed below grade and lined with natural stream substrate.  Tailwater control at the outlet of the culvert was created by constructing a series of riffles in a 200 m reach of Benewah Creek downstream of the outlet.  This had the effect of reducing channel entrenchment and increasing rearing habitat capacity.  Approximately 28 m³ of large wood was placed in the treatment reach to increase roughness in overbank areas and provide instream habitat complexity.  Figure 9 shows the difference in longitudinal profile and wood frequency before and after restoration activities were completed.  The LWD volume was increased from 0.057m³/100 m to 5.59m³/100 m.  Mean residual pool depth increased from 0.41 m to 0.78 m.  Bank height ratio, the ratio of total bank height to bank full height, was reduced by 54% and estimated stream bank erosion rates and sediment yield were reduced by 47% and 69%, respectively.

Figure 9.  Comparison of channel bed form and large wood frequency before and after restoration for a site in Benewah Creek.

An additional fish passage project was conducted on the North Fork Alder Creek where Tribal staff identified a stream crossing as a complete barrier in 2005 (Contract 10885/WE E175).  The existing stream crossing, which consisted of logs and fill, was replaced with a 11.5 m x 4.4 m x 3 m Horizontal Ellipse CMP culvert.  Construction for the project was completed in June -July 2006.  This project restored connectivity with the upper North Fork Alder Creek watershed and opened up 2,469 m of high quality rearing and spawning habitats upstream of the new culvert.  The new culvert allows for passage of all size classes of westslope cutthroat trout.  The crossing is more stable and less susceptible to erosion.  Flood flows can pass through the culvert instead of being blocked and forced over the road surface, effectively eliminating the road as a sediment source to the stream channel.

Several recent projects have been implemented to add coarse wood to discrete stream reaches where deficiencies in instream wood have been identified and where recruitment processes have been impaired by past management.  Large woody debris structures were added to a site in Evans Creek in October 2005 with the objectives of increasing pool habitats, accumulating spawning substrate and providing enhanced cover opportunities for fish (Contract 10885/WE H29).  This project involved placing 4 MBF of natural wood and 16 ELWD^TM (Type 20 N) structures along 152 m of Evans Creek.  Approximately 44 pieces of natural wood were placed on the site, these consisted of pulp logs that came in a variety of sizes as large as 10 m long and 0.6 m in diameter.  The ELWd^TM structures were formed from eight smaller diameter logs to form structures that were approximately 63-68 cm in diameter and 6 m long.  Habitat data was collected for the site from before and after wood additions.  The density of large wood increased by 256% from 4.11 m³/100 m in 2005 to 14.63 m³/100 m in 2006.  Much of the 2005 wood density was due to a single large rootwad that had a volume of 3.97 m³.  This rootwad moved off the site in 2006.  It was found that localized scour of the bed surface occurred in conjunction with placement of the ELWd^TM structures during the first year following implementation.  However, any changes in pool metrics attributed to restoration were transient and we did not observe significant changes in either pool depth or frequency relative to a control site when comparing pre-treatment and post-treatment data.  Though trout densities were very similar in lower Evans creek from 1996 to 2002, extremely different trends were observed from 2003 to 2009 between the treated reach and nearby control reaches located upstream and downstream of the project (Figure 10).  Prior to restoration, from 2003 to 2005, the mean density of cutthroat trout at the treated site was 12.8 fish/100 m.  After restoration in 2005, mean density over the years 2006 to 2009 increased dramatically to 93.2 fish/100 m.  Conversely, over the same time periods, mean densities decreased from 48.3 to 32.3 fish/100 m at a downriver control site, and from 30.8 to 21.1 fish/100 m at an upriver control site.  Evidently, the ELWd^TM structures provided important habitat at the treated site and pool depth, as intended to be created with the addition of the ELWd^TM structures, may not be as important in creating suitable rearing habitat in Evans Creek as in our other systems.  One explanation is that water temperatures in the Evans Creek watershed are not as limiting as temperatures found in Benewah Creek.  For example, water temperatures at Evans site 3 exceeded 17° C less than 2% of the time compared to approximately 50% of the time in Benewah mainstem sites near 9-mile bridge in 2006 (Vitale et al. 2008).  Thus, the thermal refugia that have been observed in deep restored habitats in the upper Benewah watershed may not be as critical for cutthroat trout in Evans Creek.

In 2007, a instream restoration project was completed on Whitetail Creek, a tributary to Benewah Creek (Contract 33533/WE I29).  The objectives of the project were to to increase habitat complexity, increase channel stability, reduce bank erosion, and increase the frequency of overbank flooding by adding large wood to the stream channel.  A total of 305 m of stream channel was treated.  Approximately 20 MBF of wood was used to create single and multiple log structures for this project.  Portions of the logs were buried below the predicted depth of scour to act as anchors for the structures.  Other logs were placed along and across the stream in different configurations to form bank protection structures and dams.  Twenty structures were built.  Eroding stream banks were reshaped in areas where structures were placed in order to form new bankfull benches.  Prior to construction wood volumes for the project reach was 1.54 m³/ 100 m.  One year post construction saw a wood loading of 11 m³/ 100 m.

Figure 10.  Density estimates and associated 95% confidence intervals for cutthroat trout at restored site 3 and unrestored sites 2 and 4 in lower Evans Creek, 1996-2009.   In-channel habitat enhancement occurred at site 3 in fall 2005 after fish sampling occurred (depicted by arrow).  The accompanying photo shows juvenile cutthroat trout using ELWd™ structure as cover.

Developing Adaptive Approaches to Restoring Watershed Processes

Since 2005, much of the restoration effort for this project has been directed at recovering a suite of watershed processes based on increasing the connectivity of the channel and floodplain and restoring native plant communities for a 5.1 km reach of upper mainstem habitats in the Benewah Creek watershed, with the downstream end of the reach located 14.3 km from Coeur d’Alene Lake.  This reach was prioritized, in part, based on the measured deviation from historic and reference conditions for several habitat, water quality and landscape elements (e.g., wetland loss, channel stability, riparian conversion), and its proximity to intact habitats and source populations in the upper watershed (Vitale et al. 2002).  The implementation of this larger reach scale project is consistent with the multispecies analytical approach described by Beechie et al. (2008) in which landscape functions at the scale of species and ecosystems are prioritized for restoration based on the degree of impairment to processes and/or rarity of habitat types.

In this reach, the historical engagement of flood flows on the valley floor was most likely in response to both (i) blockage effects of large wood pieces falling into the channel and aggregating smaller wood, and (ii) beaver dams, with local gravel and fine sediment accumulations upstream.  Following removal of the valley forest, beaver trapping, and 70+ years of cattle grazing, the effective flood level control provided by flow obstructions and the associated upstream gravel accumulations was likely removed, resulting in less frequent and shorter duration inundation of the valley floor during snowmelt and runoff events compared with pre-settlement times.  In addition, impacts to the former riparian forest, which provided root cohesion, resulted in reduced stability of stream banks and a general unraveling via widespread lateral bank failures and channel avulsions throughout the course of Benewah Creek.  Hydraulic analysis of representative channel cross-sections through this degraded reach show the overall level of incision/enlargement is approximately equivalent to the capacity of a 5-year return interval peak flow event, with some areas exhibiting channel capacity that approaches the 10-year peak flow.  Implementation has proceeded with the objective of recovering a more stable channel and floodplain geometry by restoring natural fluvial geomorphic processes and riparian vegetation that allow for channel migration without incision.  During the first phase of implementation (2005-2008), the design took the approach of filling the stream channel to historical elevations and utilizing historical alignments where possible (Contract 10885/no WE reference).  In areas that laterally expanded following entrenchment, new banks and floodplain were created.  Excavation to enhance channel planform was generally not employed, except in areas where deposited sediments occluded abandoned channel segments that were reactivated.  The designed planform creates channel grade and profiles within the range of what is believed to be historical conditions, based on topographic and field analysis.  A second phase of implementation (2009-2012) takes an approach involving minimal in-channel work and a reliance on beavers and beaver assist structures to aggrade the incised stream channel over time (Contract 37842/WE E175).  Discrete segments of relict channels were reactivated where these opportunities existed, but more importantly, in-channel structures were constructed to facilitate overbank flooding in key areas to support establishment and growth of native plant communities across the valley bottom and to help reinforce a more stable complex of natural beaver dams.

A total of 2,523 m of channel was constructed between 2005-2008, and changes in response variables are summarized in Table 9 using a combination of before/after and treatment/control comparisons from monitored sites (Contract 10885/WE G30).  Restoration activities have increased channel length by 506 m, resulting in an overall 25% increase in sinuosity from 1.28 to 1.68.  Slope deceased by 58% from 0.0048 pre-construction to 0.002.  Mean residual pool depth increased significantly (p <0.001) from 0.57 m pre-construction to 1.18 m and mean low-flow thalweg depth also increased significantly from 0.38 m pre-construction to 0.52 m (p<0.0167) (Figure 11).  Pool volume increased by 41% from 155 m³/100 m to 218 m³/100 m.  Instream large wood volume increased 143% from 3.15 m³/100 m pre-construction to 7.67 m³/100 m.  For some constructed channel segments, the increase in wood frequency and volume was even greater.  Changes in stream bank erosion rates were estimated for treated and untreated control sites using the BANCS model (Rosgen 2006), which combines quantitative measures of stream bank characteristics with derived values of near-bank sheer stress to generate estimates of average annual erosion rates.  The control site was characterized by unstable stream banks with accelerated erosion rates and increased sediment yield to the channel; 30% of stream banks showed active erosion with erosion rates of 0.7 metric tons/year/m.  Restoration efforts have significantly improved stream bank conditions to reduce erosion potential (Photo 3).  Significant response variables include the bank height ratio, which was reduced by 50%, and the rooting character (e.g., root density and depth) of stream bank vegetation.  We estimate that erosion rates have been reduced by 73% with a reduction in total sediment yield of greater than 1,294.6 metric tons/yr for the 2,523 m of channel that has been treated to date.  Active bank erosion was evident at 10% of stream banks 2 years post-restoration, a reduction of 65% compared with the untreated control.  A mix of 23 native plant species, including more than 78,000 herbaceous plants and 35,000 woody trees and shrubs have been planted in 7.2 hectares of valley bottom floodplain influenced by the project (Contract 10885/WE J47, WE K47).  In these areas, the extent of wetland habitats has been increased by 48% and wetland function has been improved over a broad range of indices.  Improvements in functional capacity are most likely attributable to the increase in hydrologic interaction between the restored Benewah Creek and it’s floodplain, with the greatest improvements evident in maintaining detrital biomass, dissipation of energy, sediment and nutrient retention/removal, and dynamic and long term surface water storage.  Mean depth to groundwater during summer base flow, as measured at shallow groundwater wells (N=10) in 2008, indicate much higher water levels in restored reaches (mean = 0.62 m) compared with untreated controls (mean = 1.32 m).  Together these changes reflect a significant improvement in the quantity and quantity of instream habitats available to native fishes as well as an improvement in the geomorphic characteristics often correlated with increased biological productivity and associated with ecological restoration.

Photo 3. Comparison of representative stream bank and channel conditions in restored (A) and untreated (B) mainstem reaches of Benewah Creek.

Figure 11. Box plots comparing residual pool depth and low-flow thalweg depth pre-construction in 2005 and 1-year post-construction in 2006.  The horizontal lines within the boxes are median values, the upper and lower edges of boxes the central 50% of the distribution, and the whiskers the highest and lowest values, including “outliers” (asterisks).  Median values are significantly different (p < 0.0167).

Table 9. Summary of change for selected response variables following restoration in Benewah Creek.

More recently, our monitoring of natural beaver dams in the upper 3 km reach of the Benewah Creek mainstem, beginning in fall 2008, greatly influenced our approach during the second phase of restoration (Contract 47583/WE W157).  Our surveys indicated that only 26% of the dams appeared to be built with or upon stable materials such as large wood; most of the dams were built with small alder pieces (Vitale and Firehammer 2011).  Because of the small size of dam-building materials, in combination with the entrenched nature of the reach and the lack of other large wood to attenuate current velocities, many of the dams were either compromised or destroyed during ice breakup and peak flows.  Notably, of those dams that tended to persist, most were located in areas with a proximate relatively intact riparian forest and the presence of large wood in the channel (Figure 12).  Though dams tend to be rebuilt during late summer and early fall periods, their lack of stability precludes the consistent availability of deep pool habitat for cutthroat trout during both critical overwintering and warm, mid-summer rearing periods.  Furthermore, the lack of intact structures during spring runoff does not provide the mechanism that will consistently promote the engagement of spring flows with floodplain habitats and facilitate restoration of a riparian forest.

Figure 12. Mean change in dam height with associated 95% confidence intervals from the fall 2009 to the spring 2010 survey (left panel), and the mean height (± one standard deviation) of dams surveyed during the spring of 2010 (right panel).  Circle, squares, and triangles represent three discrete reaches, where reach 2 (squares) is laterally bounded by relatively intact riparian forest and exhibits relatively higher large woody debris loading.

We developed our natural analog approach for beaver assist structures based in part on the surveyed characteristics of natural dams, using an arrangement of large logs to create backwater effects and flow constriction to allow for more frequent and extensive floodplain connection during annual floods (DeVries et al. 2011 In Press).  Structures were sited at strategically selected locations to facilitate connection of key floodplain swales during high flows.  This approach was conceived as an adaptive application, where additional structures may be identified for construction based on future monitoring of the present structures.  Because we do not know the best approach a priori, we conceived two alternative types of engineered flow choke structures with different means for upstream water surface elevation control.  Both designs involved constructing log “walls” with the appropriate hydraulic constriction to back up water to the floodplain level at approximately the target bankfull flow (Photo 4).  The structures extend downwards sufficiently to account for predicted scour depths, and extend into the floodplain sufficiently to prevent the stream from accomplishing an end-run around the structure.  The sill elevation of the structures was designed to emulate general low flow control elevations formed by numerous beaver dams present in the reach, where the median depths of dams were 0.35 m at the riffle crest and 0.98 m below the floodplain.  We also conceived a non-engineered approach involving placing 2-4 large logs in the channel to provide a key building block to aid beavers in dam construction.  This was based on observations that the most persistent, existing dams throughout the Benewah Creek stream corridor are built with mountain alder keyed into remnant in-channel large wood.  MacCracken and Lebovitz (2005) found that this technique can work under specific circumstances, when the channel is unconfined with a wide floodplain, there are no logjams nearby, and there are deep pools and banks suitable for dens nearby.  The objective of this approach is to assist the beaver in constructing new dams by supplying them with foundational dam material.  Additionally, fresh black cottonwood, Aspen saplings and conifers have been placed along the stream banks and in the channel to encourage beavers to finish the dam construction.  It has been demonstrated that beaver will use saplings placed along stream banks for dam construction (Muller-Schwarze and Sun 2003).

Photo 4.  Two types of flow choke structures were built, one with a combined orifice and weir (a), and the other with a simple contracted weir (b).  Both structures back water upstream during high flow and are strong enough to withstand ice jams.

Between 2009 and 2011, we built 9 flow choke structures within the active channel - 4 additional structures were built in side-channel habitats and in channels that will be activated in the future - and we used a simpler approach to reinforce natural beaver dams and/or provide beaver assist structure in 7 additional locations (Contract 42560/WE D30).  These structures influence 797 m of stream habitats at low flow through backwater effects, and provide increased connectivity to the valley bottom floodplain for about 2.1 km (70%) of the project reach.  Importantly, from a riparian floodplain restoration perspective, we have documented overbank flows across the valley bottom at discharges equal to about the 1.5 year return interval flood in reaches associated with more stable natural dam complexes and in the vicinity of our structures (Photo 5).  Other reaches require much higher discharge for overbank flow.  Recovery of beaver-generated floodplain wetlands and their wet meadow, scrub-shrub and forested plant communities depends upon restoring lost hydraulic linkages between the channel and its floodplain (Westbrook et al. 2006).  Characteristic riparian floodplain vegetation depends on annual flood-pulses and a locally high water table.  We note that the choke structures were not designed to increase summer pool habitat directly over current conditions.  In contrast, natural beaver dam building significantly increases the amount and depths of pool habitat in Benewah Creek.  We have observed, however, that most dam building activity appears to occur later in the summer, which leads us to suggest that thermal refugia and habitat cover typically associated with deeper pools and increased trout abundance (e.g., Ebersole et al. 2003; Firehammer et al. 2010) is currently less available during the warmest mid-summer periods.  The loss of unstable dams during high flows and ice breakup may also contribute to poor overwinter survival in mainstem habitats because both juvenile and adult cutthroat trout are known to use deep pools as winter refuge habitat in small stream systems (Jakober et al. 1998; Brown and Mackay 1995; Harper and Farag 2004; Lindstrom and Hubert 2004).  In instances where flow choke structures have replaced less stable natural dams, the effect has been to increase residual pool depth by an average of 0.33 m and increase pool area by 141%.  In addition to the above mentioned structures, approximately 48 cubic meters of wood (80 20-33 ft. long logs), primarily aspen and conifer logs, has been added to the stream channel and near bank region within a 700 meter reach to aid beavers in dam construction and increase wood loading to approximate a target wood loading of 6-9 m³/100 m for mainstem and tributary habitats in the watershed.  Collectively, the approaches to channel wood additions that have been implemented most recently allow for more frequent and extensive floodplain connection during annual floods, increase deep pool habitats associated with improved summer and winter rearing, and is a natural analog alternative to large scale riffle construction that maintains connectivity with cooler groundwater during summer months.

Photo 5. Over bank flows in discrete channel segments of upper Benewah Creek are initiated at the 1.5 year return interval flood (168 cfs) in areas that are associated with engineered structures that have been constructed in the channel, as well as with stable, persistent beaver dams, large wood aggregations and intact riparian forest.

In combination with the in-channel work described above, restoration of riparian plant communities has proceeded in floodplain habitats throughout the reach, and especially in areas where project work is resulting in more frequent overbank flows (Contract 42560/WE E47, F47).  A primary strategy being utilized for floodplain restoration is the utilization of black cottonwood’s unique life history characteristics to rapidly “flip” or change the current degraded riparian ecosystem into a diverse self-sustaining riparian forest.  Although black cottonwood’s regenerative strategy (seedling establishment on bare alluvial substrates and branch fragment vegetative propagules) likely resulted in it historically playing a non-dominant role in the riparian forest, its life history characteristics make it ideal for rapidly establishing a complex riparian forest.  The planting restoration design calls for establishing a matrix of floodplain cottonwood and other woody shrubs, interplanted with understory cedar, Engelmann spruce and pine species.  Cottonwood will establish a closed canopy within about 5 years and act as nursery cover for establishing understory conifers.  Cottonwood break-up will occur at about 60-90 years, relinquishing understory conifers to a dominant canopy position.  This technique has been used successfully with cottonwood and western red cedar in trials in British Columbia (Peterson et al. 1996).  From 2009-2011 a total of 59,053 herbaceous plants and 20,953 woody plants representing 26 species were planted to treat 9.45 hectares of floodplain habitats and 2.85 km of stream bank.  In much of these areas, invasive reed canarygrass (Phalaris arundiacea) that had become established was mechanically removed from planting areas prior to treatment.  The establishment of an interior forest micro-climate following canopy closure will provide significant enhancement of fish and wildlife habitat throughout the Benewah Creek valley riparian ecosystem.  Specifically, the new riparian forest will provide for maintenance of habitat interspersion and connectivity, reflecting the capacity of a wetland to permit aquatic organisms to enter and leave the wetland via permanent or ephemeral surface channels, overbank flow, or unconfined hyporheic grave aquifers, and access of terrestrial or aerial organisms to contiguous areas of food and cover (Jankovsky-Jones 1999).  The forest will support enhanced fish habitat through stream shading, allochthonous input of fine, coarse and organic carbon to the aquatic ecosystem, and input of large wood to the stream channel.

Another major stream and wetland restoration project was recently completed on the WF Lake Creek between 2009-2011 (Contract 37842/WE F175).  The project treated 762 m of WF Lake Creek and 305 m of an unnamed tributary.  The project reach was an incised section of stream channel that had been straightened and ditched prior to 1937.  The stream reach had severe bank erosion, lack of large woody debris, low riparian plant diversity, and a confined floodplain (Photo 6).  Flood flows were concentrated in the channel, located 4-6 ft. below the valley bottom, and rarely able to access the historic floodplain.  The restoration approach involved filling 610 m of the existing incised channel and diverting flows into a newly constructed 922 m long channel that is well connected with the valley bottom to allow dissipation of flood flows over a broad, vegetated floodplain.  New stream habitat was constructed over the channel subgrade using imported gravels and logs to create streambed and streambanks (Photo 6).  Rock was placed in the channel in combination with large wood to form riffles and pools.  The seasonal stream was also filled to repair the degradation that had occurred as headcuts from the incised mainstem migrated headward over a period of several decades.  Goals for this project included: 1) create wetland habitats and increase hydraulic connections with the valley bottom; 2) reduce bank erosion; 3) provide a long-term source of large woody debris for natural recruitment; and 4) provide measurable increase in abundance and distribution of westslope cutthroat trout.

Photo 6.  WF Lake Creek before (2008) and after (2011) restoration treatments were applied to address severe channel incision and bank erosion stemming from historical channelization of the stream.

A comparison of response variables for pre/post construction is summarized in Table 10 for the site.  Restoration activities have increased channel length by 312 m, resulting in an overall 51% increase in sinuosity from 1.13 to 1.71.  Slope decreased by 19% from 0.0047 pre-construction to 0.0038.  Restoration efforts have significantly improved stream bank conditions to reduce erosion potential.  Bank height ratio (the ratio of total bank height to bankfull height) was reduced by 75% from 4.37 to 1.  A mix of 23 native plant species, including more than 20,199 herbaceous plants and 11,566 woody trees and shrubs have been planted in 3.64 hectares of newly created floodplain, increasing the plant diversity on site (Contract 42560/WE H30, I47).  The extent of wetland habitats has been increased by 302%.  Together these changes reflect a significant improvement in stream and riparian processes that should translate into improved quantity and quantity of in-stream habitats available to native fishes.

Table 10. Summary of change for selected response variables following restoration in WF Lake Creek.

Thermal Response to Restoration in Benewah Creek

Temperature monitoring in mainstem reaches of the upper Benewah watershed have revealed the creation of thermal refugia that were the result of our large-scale channel restoration activities.  Sequences of temperatures, collected in riffles and in the deepest part of adjacent upstream pools, have been measured continuously along mainstem reaches during mid-summer periods before and after their restoration.  Before restoration, most pools in upper Benewah mainstem reaches were less than 1.0 m deep and generally were not more than 0.5°C cooler than their associated downstream riffles (Figure 13).  However, in the reach of the mainstem that underwent channel reconstruction in 2005 and 2006, post-reconstruction pools were typically greater than 1 m, and stream temperatures measured along the bottom of these pools were frequently between 2 and 5°C cooler than ambient riffle temperatures.  Pools in the 2007 restoration reach were also generally deeper after than before channel construction, with temperature differences ranging between 2.5 and 6.5°C during the post-restoration survey.  In summary, cool-water refugia were more prevalent in reaches after than before large-scale channel restoration, apparently created by the concomitant deepening and lengthening of pool habitats during the process of streambed elevation within the designated riffles.  The creation of these refugia should increase the availability of suitable rearing habitat for WCT in mainstem habitats of the Benewah watershed (Torgersen et al. 1999; Ebersole et al. 2001, 2003).

Figure 13.  The relationship between temperature difference and residual pool depths for surveys conducted above 9-mile bridge in the upper mainstem of Benewah Creek along reaches that were restored in 2005 and 2006 (top panel) and in 2007 (lower panel).  Temperature difference was calculated as the temperature measured along the pool bottom minus the temperature measured in the associated downstream riffle.  Years in which surveys were conducted are identified in parentheses.

Biological Response to Mainstem Habitat Restoration in Benewah Creek

Despite the mosaic of thermal refugia and the complex habitat (e.g., deep pools and LWD additions) that have been created in reaches of the upper Benewah mainstem that had been restored from 2005 to 2008, we have yet to see direct evidence of a significant response by cutthroat trout.  Various explanations have been proffered for the apparent lack of utilization of these restored habitats, which have been described in detail in previous annual reports (Firehammer et al. 2009, 2010).  Briefly, these include, but are not limited to, the following: (1) a sufficient degree of isolation between core rearing tributaries and restored mainstem habitats, mediated by distance or other physic-chemical barriers (e.g., temperature), that inhibit dispersal (Bond and Lake 2003; Pretty et al. 2003); (2) insufficient tributary densities to induce density-dependent emigration into these habitats (Johnson et al. 2005; Shrank and Rahel 2006); (3) a lag in positive fish response because of the repeated, acute artificial disturbances imposed by channel reconstruction on ecological and hydrological stream properties over the four years of restoration; and (4) the persistence of limiting factors in reaches adjacent to those restored (Moerke and Lamberti 2003; Cowx and Van Zyll de Jong 2004).  We realize that because we are not only amending local deficiencies in habitat complexity but also addressing impaired processes that operate at larger spatial scales, the re-establishment of natural processes will occur gradually, and as such, detection of positive responses by cutthroat trout may require a longer timeframe.  As we progressively address contiguous reaches in the upper Benewah mainstem with Phase 2 implementation, we expect to continue to increase the extent of favorable rearing habitats that are conducive for cutthroat trout colonization and growth.

Another explanation for the absence of cutthroat trout in restored habitats, which has not been adequately tested, is our inability to capture fish using our current sampling techniques.  Given the thermal refugia that have been observed at the bottom of deep pools in restored reaches, cutthroat trout, if present, would most likely be using these micro-habitats.  However, restored pools are frequently over 4 ft deep, and not only is visibility poor but both wading and netting prove challenging at these depths.  Furthermore, because of the low conductivities in our watersheds, the electrical fields generated by our backpack electrofishing equipment are exceptionally small and consequently may not elicit electrotaxis in fish lying along the bottom.  Currently, we are experimenting with other gear types (e.g., fyke nets) to evaluate whether cutthroat trout are using deep restored pools as summer rearing habitats.

Though a direct numerical response to restoration has not been observed in mainstem reaches, the significant increase in cutthroat trout densities in tributary habitats demonstrated by our trend analysis may suggest an indirect response to restoration.  Deepened mainstem reaches may have provided suitable overwintering habitat that was available only in a limited capacity before restoration.  Both juvenile and adult cutthroat trout have been found to prefer deep pools as winter refuge habitat in small stream systems (Jakober et al. 1998; Brown and Mackay 1995; Harper and Farag 2004; Lindstrom and Hubert 2004).  In addition, cutthroat trout have been found to respond positively to improvements to winter refuge habitat.  Solazzi et al. (2000) found cutthroat trout abundance to increase, presumably owing to higher overwinter survival rates, following the creation of winter habitat for salmonids in coastal Oregon streams.  In addition, Roni and Quinn (2001) found higher densities of cutthroat trout at sites with experimental large woody debris additions than at control sites, but only during winter and not summer sampling.  Evaluating the winter distribution of cutthroat trout in upper Benewah mainstem habitats may reveal benefits of our channel construction activities that were not realized from summer surveys.  In order to perform such an evaluation, cutthroat trout captured in tributaries during summer and fall electrofishing surveys will be PIT-tagged and their movements monitored throughout the fall and winter using strategic placement of antenna arrays in mainstem habitats.  A more detailed description of this methodology is provided in the proposed deliverables section.

Objective 4: Address Impacts from Non-native Introduced Fishes

Non-native interactions in stream environments

A brook trout control program was initiated in 2004 to suppress the numbers of brook trout found in main-channel and tributary habitats in the upper portion of the Benewah watershed.  However, unlike other brook trout removal projects that have focused on eradication and subsequent preventative recolonization measures, such as passage barriers (Shepard et al. 2003), our approach was tempered by the desire to maintain connectivity with the lake to promote the migratory life-history variant of our cutthroat trout population and its concomitant high productivity potential.  We felt that the benefits of unimpeded access and the expression of the cutthroat adfluvial life-history greatly outweighed the benefits of brook trout eradication in isolated tributaries (Peterson et al. 2008a).  Further, eradication and subsequent barrier installation have not always proven entirely successful (Thompson and Rahel 1998), and, within our watershed, would require large-scale chemical treatments and an extensive trapping and hauling program to supply migratory adult cutthroat trout to the various isolated spawning tributaries.  Our control strategy entailed annually removing fish before fall spawning periods by conducting single-pass electrofishing efforts through contiguous mainstem reaches upriver of 9-mile bridge and in tributaries that supported relatively high densities of brook trout.  Numerical responses in brook trout to our efforts were examined at a 2.0 km main-stem index reach that has been consistently targeted in each year, and at our tributary index sites in the upper watershed.

From 2004 to 2011, approximately 8000 brook trout have been removed from the upper reaches of the Benewah watershed.  Over the first three years of the suppression effort, numbers of fish removed increased from 601 to 2405, in large part due to the progressive targeting of additional mainstem river kilometers rather than just tributary reaches (Figure 14).  Not only were higher densities of fish found in mainstem reaches (most notably from the 12-mile bridge to the confluence of the upper Benewah forks) than in tributaries, but fish were also generally greater in length (Vitale et al. 2008, 2009).  The increase in the percent of mature adults removed from 26% in 2004 to 55% in 2006 also reflects the spatial re-distribution of our suppression efforts.  Since 2006, numbers of brook trout removed from the upper watershed have steadily declined (Figure 14).  Although the decline may partly be explained by the large numbers of fish removed during the early years, the low numbers of fish removed since 2009 were largely due to reduced efforts applied to mainstem reaches.  Notwithstanding the reduced effort in recent years, both numbers of brook trout and CPUE (fish removed / seconds of shock effort) have steadily declined in the 2.0 km mainstem index upriver of 12-mile bridge (Figure 14).  A significant overall reduction in brook trout densities across upper Benewah tributaries, however, has not been able to be detected.  The lack of a measurable reduction across tributaries is likely explained by the differences in trends observed among the monitored tributaries.  Whereas densities generally remained at low levels in Whitetail Creek and the South Fork since 2004, they declined substantially in the West Fork, but displayed increasing and variable trends in Schoolhouse and Windfall creeks (Figure 15).

The differences in trends observed across tributaries may be attributed to one or more of several factors including the proximity to colonizing sources, changes in reach accessibility, and varying degrees of effort applied in previous removal activities.  First, the location of Schoolhouse and Windfall creeks in the upper part of the watershed may in part explain the positive trends observed in both tributaries.  The mouths of both creeks are located along the mainstem reach where densities estimated during removal efforts have consistently been found to be the highest, thus increasing the probability for mobile individuals to colonize these tributary reaches.  Others have noted the importance of both proximity and connectivity to source localities in determining probabilities of brook trout establishment (Benjamin et al. 2007).  Brook trout expansion into Windfall Creek, however, was likely inhibited until 2004 when culvert replacement virtually eliminated this barrier.  Thus, local sub-populations, colonized by the more mobile individuals, may not have yet had the opportunity to become firmly established in Windfall Creek (Peterson and Fausch 2003), which may partly explain the variable densities observed in this tributary over the last four years.  As a result of the recent re-connectedness of Windfall Creek with the mainstem, this tributary should continue to be monitored in the future to assess rates of brook trout expansion into this newly accessible habitat.  Additionally, the differences observed among tributaries may have been due to the focus of removal efforts during the first couple of years.  Initially, before it was discovered that many of the larger adults were residing in upper mainstem habitats, efforts were concentrated in tributaries, most notably the South and West Forks.  Given that the most marked decrease in abundance was demonstrated in the West Fork, the unequal distribution of past sampling efforts may partly explain the results from our survey data.

Figure 14.  Number of brook trout removed from the upper Benewah watershed and from an index mainstem reach in upper Benewah in addition to the CPUE (fish removed / shock sec) of brook trout in the index reach, 2004-2011.

Figure 15. Depletion-removal estimates (fish/100 m) of age one and older brook trout for five tributary reaches in the upper Benewah Creek watershed, 2003-2009.

Though a tributary-wide appreciable decline was not detected in upper Benewah, a comparison with the neighboring Alder Creek watershed suggest that our suppression program has been effective at regulating numbers of brook trout at a manageable level.  Overall trends in the upper Benewah watershed have yet to display trajectories that would project densities similar to those observed in Alder Creek (Figure 16).  In most reaches, present densities are more than five times lower in the upper Benewah than in upper Alder.  Watershed comparisons also may provide insight into the productive potential for brook trout in the Benewah watershed.  Even before commencement of the suppression program, densities of brook trout in Alder Creek have been consistently higher than those documented in Benewah Creek.  In addition, whereas distributions of cutthroat and brook trout are almost entirely disjunct in Alder Creek, suggesting probable displacement by the latter (Dunham et al. 2002), distributions of both species overlap in Benewah Creek.  Differences between these two watersheds could be explained by an invasion process that is still in its incipient stage in Benewah, though given the proximity of these watersheds to each other, expansions should have proceeded at similar rates if colonizing migrants arrived from common downriver sources (Peterson and Fausch 2003).  As another possible explanation, the productive adfluvial life-history strategy that is prevalent in the Benewah but not the Alder watershed may confer an advantage to cutthroat trout in the former that permits a greater biotic resistance to invasion (Griffith 1988).  Differences in apparent vulnerabilities of proximate systems have also been reported by others that have examined brook trout invasions in the west (Adams et al. 2002; Dunham et al. 2002; Shepard 2004; Benjamin et al. 2007).

Figure 16. Depletion-removal estimates (fish/100 m) of age one and older brook trout for five reaches in the upper Alder Creek watershed, 2003-2009.

Alternatively, habitat conditions that are more conducive to brook trout establishment may be more prevalent in Alder than in Benewah.  For example, the spatial distribution of brook trout and their habitat preferences have commonly been associated with low gradient reaches with deep, low velocity habitats (e.g., beaver ponds) that serve both as summer rearing and overwintering habitat (Chisholm et al. 1987; Cunjak 1996; Lindstrom and Hubert 2004; Benjamin et al 2007).  Recent habitat surveys conducted across our watersheds have indicated that pool habitat is approximately three times as great in Alder Creek than in Benewah Creek (Miller et al. 2008).  Additionally, the surveys found that 33% of the pool habitat documented in Alder Creek was formed by dams, whereas only 3% of the pool habitat in Benewah Creek was dammed.  Given our current restoration approach to encourage the stability of beaver dam complexes and augment associated pool habitat in the upper Benewah watershed, we may also be increasing suitable habitat for brook trout.  Continued monitoring of brook trout numbers in the upper watershed should inform whether such unintended responses are occurring.

Our suppression program also entailed monitoring changes in maturation metrics in brook trout to detect potential compensatory reproductive responses to our removal efforts owing to a release from conspecific competition.  Specifically, we were interested in whether residual brook trout expressed changes in the average size at maturation or in the fecundity-at-length relationship.  From 2004 to 2008, a subsample of fish was dissected to ascertain maturation status, gonad weight, and, in the case of females, fecundity.  Similar data were collected from a representative sample of sacrificed brook trout from Alder Creek to obtain comparative life-history data for control purposes.  Thus far, our removal program apparently has not induced compensatory responses in the brook trout population.  Female brook trout were not more likely to mature at a given length in 2008 than in 2004 (logistic regression, odds-ratio = 1.403, p = 0.511).  Though we did detect an increase in fecundity from 2004 to 2008 for brook trout from the upper Benewah watershed, fecundities for Alder Creek females were also comparably higher in 2008 than 2004, suggesting similar mechanisms may have been operating in both watersheds (ANCOVA; Watershed, p = 0.684; Year, p < 0.001).  Though continued monitoring would better inform the potential for long-term compensatory responses, it appears that the maintenance of low brook trout densities in the upper Benewah watershed through periodic removals should not increase individual reproductive investment (e.g., increased fecundity at a given length) nor induce an earlier maturation schedule (inferred from length at maturation) that would shorten generation times.  More importantly, our results illustrate the advantage of using a control watershed when evaluating the effectiveness or potential undesired impacts of a non-native removal program.

Figure 17.  Cumulative distributions of total length (mm) of brook trout removed from the 2.0 km mainstem reach upstream of 12-mile bridge in the Benewah watershed in 2005 and 2009-2011.

Non-native Interactions in Lake Coeur d'Alene

As a result of the PIT-tag information, a cooperative study is currently being conducted through the Fisheries and Wildlife Resources Department at the University of Idaho to evaluate the impact of two non-native piscivores, northern pike and smallmouth bass, on cutthroat trout survival in Lake Coeur d’Alene.  Cutthroat trout have been found to be a major dietary item for northern pike in earlier studies conducted on Lake Coeur d’Alene (Rich 1992), and smallmouth bass, a documented salmonid predator, have apparently increased in numbers in the last 10 years (Maiolie et al. 2010).  The study will incorporate two field seasons (2012 and 2013) in which both Windy Bay and the southern end of the lake, into which Lake and Benewah creeks respectively enter, will be intensively sampled during spring periods where there exists a high potential for spatial and temporal overlap of migratory cutthroat trout and both predators.  Demographic (e.g., age structure, growth, seasonal abundance) and dietary data will be collected from both predators during these repeated sampling efforts and incorporated into bioenergetic models to estimate the consumption of cutthroat trout by both species.  Information gained from this study will support the development of alternative actions that may be considered for implementation to manage the fish assemblage in Lake Coeur d’Alene, or re-direct questions and uncertainties research to other potential limiting processes in the lake (Figure 18).

Figure 18.  Diagram depicting alternative management actions that could be considered for implementation contingent on the findings of the study assessing the consumptive impact of northern pike and smallmouth bass on cutthroat trout during lake residence.  The weight of the arrows linking study outcomes to proposed alternatives denotes the relative projected feasibility of alternative in achieving success.  Two-way arrows linking proposed alternative to effectiveness monitoring denotes that monitoring will inform the implementation of alternatives under an adaptive management context.

Objective 5: Increase Coordination and Participation Among Stakeholders

Outreach and education activities are regularly completed in conjunction with restoration and monitoring efforts to effectively connect many thousands of people that are affected directly and indirectly to project work each year (Contract 37842/WE AA99).  There are two education/outreach objectives:  1) improve awareness of Fisheries Program activities within the Reservation community to solicit support and participation in projects; and 2) provide for educational opportunities that raise awareness for natural resources issues in the local schools and communities of northern Idaho (Table 11).

Outreach activities have focused on informing the public about different projects and topics important to project work.  The Fisheries Program publishes a newsletter, The Watershed Wrap, which is distributed to reservation landowners and local agencies throughout the year.  This newsletter contains a variety of articles describing work being completed by the Fisheries and Wildlife programs, as well as other natural resource management efforts conducted by the Coeur d’Alene Tribe.  Watershed and inter-agency work group meetings are also held to inform stakeholders about different projects and provide a forum to receive public input and solicit participation in future projects.  Other outreach events that are held to engage the public include hosting a periodic speaker series and annual fishing derbies.  A variety of educational activities have been completed to engage local students.  The program holds two major educational events each year that collectively have more than 500 participants:  Water Awareness Week and Water Potato Day Celebration.  Local students participate in learning stations that focus on fish, water quality, plants, soil, wildlife, macroinvertebrates, and Tribal culture.  In addition to these events, the Fisheries Program hosts summer internships for local high school students and supports undergraduate/graduate research opportunities.  Currently the program is employing two seasonal technicians as they pursue their college education.

Involvement with the public is instrumental to the success of restoration work being completed by the Fisheries Program.  The Tribe currently owns only 7.9% of the land area within the four target watersheds.  To gain access to areas not managed by the Tribe, the program works with individual landowners and local government agencies on a regular basis. Landowner agreements or memorandums of agreements are developed between the Tribe and various landowners in order for restoration work to be completed.  In order to educate the public about current and future work, landowners are contacted through letters, workgroup meetings, and personal visits.  For two major watershed assessments completed in 2007 and 2008, landowner participation was required in order to access research locations to collect important data.  For the Wood Recruitment Study, letters were mailed to all landowners in the study area describing project objectives and soliciting participation.  Owners with controlling interest in more than 60% of the study area actively participated in the study.  Two landowner meetings were held; first to describe project methods and coordinate logistics to facilitate field work, and secondly to share project results and discuss next steps.  Similarly, for the road and fish passage study, letters describing the study were sent to 190 different landowners.  Seventy seven percent of the ownership in the study area agreed to participate, including six major timber companies and five government agencies along with many small private landowners.  Two landowner meetings were held to discuss study objectives, methods, and access issues.  By engaging the public in the planning process, we ensure the success of our work.

Table 11.  Summary of education and outreach activities.