Study design to assess response of steelhead to the removal of Hemlock Dam from Trout Creek of the Wind River Subbasin, Washington

Goal:  The purpose of the Trout Creek environmental impact assessment (EIA) study design is to evaluate the response of adult and juvenile steelhead to the removal of the dam.  These EIA designs are generally variations of the Before-After-Control-Impact (BACI) design (Meissner 2000).  In fisheries research, the BACI design has been applied or recommended to evaluate the responses of fish to nuclear power plant construction and operation (Smith et al. 1993), to the response of juvenile salmonids to habitat restoration (Korman and Higgins 1997, Solazzi et al 2000, Paulsen and Fisher 2003, Bradford et al. 2005, Roni et al. 2005), and to the habitat use of juvenile salmonids of different origin (Hill et al. 2006).

Background: The U.S. Forest Service (USFS) proposed to remove Hemlock Dam in Trout Creek a tributary of the Wind River, as an action to aid in the recovery of Wind River steelhead (USFS 2005).  The dam was removed in 2009 and the Bonneville Power Administration (BPA) funded the majority of dam removal costs.  The removal of Hemlock Dam is anticipated to have an overall positive benefit for ESA-listed steelhead (Barber and Perkins 1999, USFS 2005), but the short-term and life-stage specific response may not be positive because of the disturbance associated with removal of the dam.  Other aspects may prevent realization of the expected long-term benefits, such as access and use by hatchery steelhead strays and non-native fishes (e.g., spring Chinook salmon), and global warming.  These factors suggest that an impact assessment is the appropriate study design for tracking the response (Meissner 2000).

Location: The Wind River subbasin is located in the Columbia Gorge Province and this river enters the Columbia River approximately 15 miles upstream of Bonneville Dam.  


Objective 1.  Determine change in migration timing, adult age (including repeat spawner rate), and adult number (including summer versus winter run) of Trout Creek steelhead before and after removal of Hemlock Dam.
Background: The primary purpose in the removal of Hemlock Dam from an ESA perspective is to improve adult fish passage (USFS 2005).  The adult fishway did not meet state fish passage criteria and likely impeded adult steelhead passage (Orsborne 1987).  Passage deficiencies may have lead to the passage of fish only during times that fish passage criteria were met or for large fish that may have been better  at jumping ladder weirs or swiming through high velocity orifices.  The before data from the Trout Creek trapping project are the basis for hypothesis testing on abundance, timing, and age structure.

In addition, adult abundance and diversity have been identified as Viable Salmonid Population criteria (McElhaney et al. 2000), and these parameters are required to assess if protection for Wind River steelhead is required under the ESA.  Therefore, determination of adult abundance and diversity as measured by migration timing and age structure should be afforded the highest priority for dam removal effectiveness monitoring.

Historical data for the Impact population, Trout Creek was a census of adult steelhead based on Hemlock Dam counts.  Trap records begin in 1992 and continue through dam removal in 2009. In addition to count by date, biological data collected from the trap include: sex, length, and scales, which can provide age structure.  The control population is the Wind River steelhead population estimate minus the population estimate above Hemlock Dam.  WDFW has estimated the annual Wind River summer steelhead population based on mark-recapture-resight methods since 2000 (Rawding and Cochran 2007).  For this project, WDFW is proposing to incorporate different mark-recapture methodologies from Rawding and Cochran (2007) into a single estimate using a sequential Bayesian (Gazey and Staley 1986) or similar approach, and include uncertainty of tag loss and delayed trap mortality.  In addition, 1988 to 1999 Wind River snorkel survey estimates will be used to estimate Wind River population abundance based on run timing and snorkel efficiency from 2000 to 2010 population estimates.    


Objective 2.  Determine change in the number of steelhead a) smolts b) parr, c) parr-to-smolt ratio, and d) age at smoltification based on smolt trapping at TC, UW, and PC before and after removal of Hemlock Dam.

Background:  The most extensive data available for assessing the removal of Hemlock Dam are from juvenile trapping efforts (Rawding and Cochran 2007).  In general, population estimates are available from 1998 to 2010.  The most testable hypothesis for dam removal is to assess the change in smolt abundance.  However, the construction and operation of Hemlock Dam may have changed steelhead life history patterns.  For example, Wind River steelhead have developed many juvenile life history patterns but the most common one is a transient pattern (Rawding and Cochran 2005).  In this pattern, adults spawn in a tributary or upper maintsem reach.  Following a year of rearing, parr migrate into the mainstem Wind River to complete freshwater rearing.  They emigrate as smolts at age two or three from the mainstem Wind River.  Therefore, we propose to test for changes in parr abundance, smolt abundance, parr-to-smolt ratio, and age at smoltification.

Pre-dam removal smolt abundance data are available from Rawding and Cochran (2007), Rawding and Cochran (2001), and Rawding et al. (1999).  WDFW has not incorporated all uncertainty from mark-recapture methods, such as tag loss, delayed mortality, missed trap days, and other factors into population estimates.  WDFW plans to incorporate this uncertainty into final smolt population estimates.  WDFW proposes to provide estimates of parr and parr-to-smolt ratios from previous juvenile trapping data.  Age data for approximately 400 smolts per year are available from the juvenile age database.


Objective 3. Determine change in survival of steelhead smolts from Trout Creek before and after removal of Hemlock Dam.

Background: One of the potential effects of Hemlock Dam was the indirect or direct morality of emigrating steelhead smolts (USFS 2005).  Direct or indirect mortality can occur through predation and prolonged exposure to warm water associated with the delay of smolts when the dam boards impound the reservoir and obscure emigration routes.  Smolt mortality may also occur as plunge over the dam and into substrate and large wood below the dam.  WDFW has PIT tagged thousands of smolts from the TC, UW, PC, and LW from 2003 onward, and these fish are captured at the Lower Wind trap and detected at BON and the estuary.

Objective 4.  Determine if there is a change in smolt-to-adult return and freshwater productivity from Trout Creek before and after removal of Hemlock Dam.

Background: Smolt-to-adult return of Trout Creek steelhead based on adults that return to Trout Creek was influenced by many factors related to Hemlock Dam operations.  Primary factors that could reduce smolt-to-adult survival are smolt mortality (objective 5) and adult passage at this facility, including lack of attraction flow, insufficient ladder flow, and trap rejection.  If these factors are significant, then the smolt-to-adult survival of Trout Creek steelhead should be lower compared to the survival from PIT tagging at other sites.  The analysis will incorporate past data from WDFW/USGS PIT tagging, USFS/WDFW Trout Creek adult trap operations, and USGS PIT-tag interrogation in the Trout Creek ladder.  By providing a PIT tag interrogation system at the Upper Wind River, two years of control pre-data were collected on upper Wind River smolt-to-adult return.

Korman and Higgins (1997) used simulations to assess the utility of monitoring spawners and subsequent adult recruits to assess response to fish habitat improvements.  They concluded this type of spawner-recruit relationship was too variable to be useful in detecting fish response to habitat alteration.  Bradford et al. (2005) expanded on this analysis to examine the spawner to smolt relationship, termed freshwater productivity, to assess fish response to habitat alteration trough simulations.  Based on their assumptions for coho salmon response, variability, productivity, and capacity, and autocorrelation, they found that monitoring freshwater productivity was more precise for detecting a fish response to habitat alteration compared to monitoring adults or smolts.   Based on objectives (1) and (3), spawner-to-smolt data can be estimated from brood year 1993 to the present.