In 1939, construction of Grand Coulee Dam blocked anadromous fish from 1,835 linear stream kilometers of spawning and rearing habitat in the upper Columbia River, causing the extinction of upper Columbia River Chinook (O. tshwaytscha), coho (O. kisutch), sockeye (O. nerka), and steelhead (O. mykiss). Collectively, these fish accounted for approximately 1.1 million adult migrants to the upper Columbia River annually (Scholz et al. 1985). At the time, biologists and engineers believed it was not possible to move fish the 113 vertical meters necessary to pass them over the dam, hence, Grand Coulee Dam was constructed without a fish ladder or other means of conveying fish above it (Fish and Hanavan 1948).

Figure 1.  Map of Lake Roosevelt.  Spokane Tribal Hatchery, Sherman Creek Hatchery, and the net pens are also identified.

In addition, the completion of Grand Coulee Dam altered the natural ecology of the area by inundating approximately 243 river km of the upper Columbia River, creating Franklin D. Roosevelt reservoir (Lake Roosevelt; Figure 1). Inundation of the upper Columbia River and subsequent hydro-operations changed the river from a lotic to a lentic-type habitat, leading to resident fish extinctions of redband, cutthroat, and bull trout species (Bryant and Parkhurst 1950; Earnest et al. 1966). The loss of wild indigenous anadromous and resident fishes combined with major environmental changes resulted in a substantial ecosystem imbalance within the upper Columbia River. The once salmonid-based ecosystem became a predominantly cyprinid, centrarchid and catostomid-based system, with remnant populations of redband trout and residualized land-locked sockeye salmon, or kokanee (Gangmark and Fulton 1949).

In the 1960’s, fishery surveys indicated an abundant population of kokanee salmon in Lake Roosevelt, with anglers harvesting large numbers from the forebay. Snyder (1967) found kokanee comprised 99% of gill net and purse seine collections from the forebay, which suggested favorable ecological conditions existed for kokanee in Lake Roosevelt at that time. The relative abundance of rainbow trout was approximately only 1% throughout the reservoir.

By the mid to late 1970’s, the walleye (Sander vitreus) fishery had gained popularity, leading to implementation of harvest regulations (Nielson 1975). During the same time period, kokanee abundance was declining, prompting a review of factors limiting their production. Stober et al. (1977) concluded that the recent decline in kokanee abundance was attributable to the impact of annual drawdown regimes on reproductive success. Also during this time, Grand Coulee Dam was retrofitted with a third powerhouse, which became operational in 1975 and doubled the turbine hydro-capacity of the dam. Stober et al. (1977) evaluated the effect of the third powerhouse on the kokanee population and concluded the dam penstocks were located in a position that could cause significant entrainment of kokanee. Rainbow trout abundance remained approximately 1% of the gill net surveys.

By the early 1980’s walleye had become the predominant fishery in Lake Roosevelt with a relative abundance of 30% in a survey completed by the U.S. Fish and Wildlife Service and were reported as the most dominant and important sport fish in the reservoir with over 90% of the total fish examined in the creel (Harper et al. 1980; Beckman et al. 1985). In the 1990's, Cichosz et al. (1999) reported walleye abundances ranging from 15-48%, and more recent studies from 2000-2007 have shown walleye continuing to be one of the most dominant species collected with annual relative abundances reported from 21-32% (Lee et al. 2003, 2006; Scofield et al. 2004, 2007; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Pavlik-Kunkel et al. 2008; Lee et al. 2010). Meanwhile, the kokanee population continued to decline. Beckman et al. (1985) examined factors limiting kokanee salmon and rainbow trout production in the reservoir. They determined the two primary factors that limited the kokanee and rainbow fisheries were; 1) tributary spawning habitat within the system was insufficient to maintain a viable stock, and 2) lake operations and resultant water level fluctuations made shoreline spawning futile for kokanee salmon and rainbow trout. Additionally, Jagielo (1984) examined zooplankton availability as a limiting factor for kokanee, and determined that zooplankton abundance was sufficient to support a much larger kokanee population than existed in Lake Roosevelt. Recent surveys conducted by the Lake Roosevelt Fisheries Evaluation Program support the Jagielo (1984) conclusion, indicating ample zooplankton abundance to support kokanee populations (Cichosz et al 1999; Fields et al. 2005).

Based on the history of fisheries in the upper Columbia River/Lake Roosevelt, co-managers for Lake Roosevelt (Washington Department of Fish and Wildlife (WDFW), Spokane Tribe of Indians (STI) and Colville Confederated Tribes (CCT)) began examining alternatives to produce a viable, and in the future, a self-sustaining fishery in Lake Roosevelt that would meet the Tribes’ subsistence needs, as well as create a recreational fishery. Scholz et al. (1986) formulated a restoration plan for the Lake Roosevelt fishery that centered on spawning habitat enhancement and artificial production of kokanee salmon. It was hoped that utilizing artificial production would reduce potential conflicts with reservoir operations for power production and providing flow for downriver anadromous salmonids. The management plan also included the following elements: 1) enhance natural spawning populations of rainbow trout by improving spawning and rearing habitat; 2) maintain existing walleye stocks through natural production and harvest management and; 3) conduct a monitoring program designed to evaluate the efficacy of the above measures. This plan was submitted to the Northwest Power Planning Council (NWPPC) and, in 1987, the NWPPC published its Fish and Wildlife Program (FWP), which directed BPA to fund the construction and operation of two kokanee/rainbow trout hatcheries, a rainbow trout habitat improvement project, and a monitoring and evaluation project on Lake Roosevelt. The result was the creation of the Lake Roosevelt Monitoring Program (now referred to as Lake Roosevelt Fisheries Evaluation Program), the Spokane Tribal Hatchery, Sherman Creek Hatchery, and the Rainbow Trout Habitat Improvement Project. The LRFEP is a cooperative effort amongst the Lake Roosevelt co-managers (STOI, CCT, WDFW) and Eastern Washington University.

The Lake Roosevelt Monitoring Program began operations in 1988 to establish baseline data on the ecosystem and to monitor the performance of hatchery salmonids released into the reservoir. The program combined with the Lake Roosevelt Data Collection Program, and became the Lake Roosevelt Fisheries Evaluation Program (hereinafter referred to as the LRFEP) in 1996. The LRFEP now functions to perform the activities of both projects as well as the objectives specifically identified for the LRFEP. Additional tasks include determining the effects of the supplementation program on native resident fish and on the ecology of the reservoir, and determining the effects of hydro-operations on the artificial production program, the native fishery, and Lake Roosevelt ecology as a whole.

History of Artificial Production in Lake Roosevelt

The Lake Roosevelt/Banks Lake artificial production program is a supplementation program developed as partial mitigation for the loss of anadromous and resident fishes above Grand Coulee and Chief Joseph dams. Rainbow trout production began in 1986 at the Spokane Trout Hatchery (WDFW facility), which provided rainbow trout fingerlings to the volunteer net pen program operated by the Lake Roosevelt Development Association (Cichosz et al. 1997a). The number of rainbow trout provided by the Spokane Trout Hatchery began at 50,000 rainbow and increased to 276,500 rainbow trout by 1990 (Spotts et al. 2002). Kokanee salmon production began at Ford Trout Hatchery (WDFW facility), which provided approximately 850,000 fry to Lake Roosevelt annually from 1988 through 1990 (Cichosz et al. 1997a). The purpose of these early releases was to begin fisheries enhancement in Lake Roosevelt prior to construction of the Spokane Tribal and Sherman Creek hatcheries, the two hatcheries that were dedicated solely for production to Lake Roosevelt. The Spokane Tribal and Sherman Creek Hatcheries began releasing fish in 1991 and 1992, respectively. Initial hatchery releases were comprised of approximately 1.7 million kokanee fry and 400,000 Spokane stock (coastal origin) yearling rainbow trout annually through 1994. Additionally, approximately 100,000 kokanee salmon yearlings were released from 1992-1994 to allow the LRFEP to assess yearling versus fry survival (Cichosz et al. 1997a). Beginning in 1995, based on results of these assessments, the LRFEP recommended shifting production goals to approximately one million kokanee yearlings and 500,000 rainbow trout yearlings. In 1995, the hatcheries released approximately 440,000 yearling kokanee salmon, 500,000 kokanee salmon fry, and 500,000 rainbow trout yearlings before shifting to nearly 100% yearling kokanee production in 1996. Annual assessments indicated the rainbow trout fishery increased substantially, while the kokanee salmon supplementation program has had more limited success (Underwood and Shields 1996b, McLellan et al. 1999, Spotts et al. 2002, Lee et al. 2003, 2006; Scofield et al. 2004, 2007; McLellan et al. 2004a; Fields et al, 2005; Pavlik-Kunkel et al. 2005; Pavlik-Kunkel et al. 2008; Lee et al. 2010; Miller et al. 2011).

Concern regarding genetic integrity and introgression with native species/stocks have lead co-managers to begin assessing alternative kokanee and rainbow trout stocks to be utilized by the artificial production program in Lake Roosevelt. Originally, a coastal, non-native rainbow trout was stocked into Lake Roosevelt, and in an attempt to release a more native species, redband rainbow trout (from the Kettle River tributary) were released simultaneously to determine which stock performed best in the reservoir (recruitment to the creel). In addition to the redband testing, a pairwise comparison of performance of diploids vs. triploids determined that triploid rainbow trout performed similarly to diploids. Following the final tagging studies completed in 2009, managers determined that triploid Spokane stock (coastal origin) outperformed redband rainbow trout. Based on lower performance, and increasing concern by managers regarding the genetic impacts of releasing a stock comprised of mixed Kettle River tributaries would have on stocks native to other Lake Roosevelt tributaries, the stocking of redbands ceased in 2009. Additionally, the artificial production program increased triploid Spokane stock (coastal origin) rainbow trout production from 50% to 100% beginning in 2006. Kokanee stocks are also under review by co-managers. Fisheries managers have prioritized Lake Roosevelt mixed stock, comprised of Meadow Creek kokanee (up-river stock) and Lake Whatcom stock (the coastal, non-endemic stock historically used for hatchery purposes in Lake Roosevelt), as the first choice for releases into the reservoir, followed by Meadow Creek, and then Lake Whatcom stock. All hatchery raised kokanee salmon yearlings released into Lake Roosevelt have been adipose clipped since 1998, and all kokanee fry released into Lake Roosevelt since 2004 have been otolith marked (either thermally or chemically). All hatchery rainbow trout fingerlings released into Lake Roosevelt are 100% adipose clipped beginning with 2006 releases (Peone 2007). Efficacy testing of adipose clip marking has ranged from 94.5-100% with a four year average since 2007 of 98.6% marked (LRDA, Plant Reports, 2007-2011: WDFW unpublished data).

In 2000, the LRFEP began monitoring early maturation and sex-ratios in the hatcheries and found that a large percentage of kokanee salmon that were being released as yearlings at 17 months were sexually mature as 2 year olds and exhibited skewed sex ratios (4 males to 1 female; McLellan et al. 2004a). The cost of raising kokanee salmon yearlings that were part of the fishery for only 4 months led co-managers to begin assessing alternatives that exhibited better cost-effectiveness. Beginning in 2004, the production program began releasing 400,000 yearling kokanee salmon, approximately 3 million kokanee salmon fry, 500,000 rainbow trout yearlings and 60,000 redband rainbow trout yearlings annually. The shift back to a kokanee salmon fry production program was a management decision based on eight years of yearling assessment findings, and the limited success and associated cost of the kokanee program during that time. Adaptive management actions during those assessments included moving kokanee away from predator traps, releasing them in upper reservoir tributaries, and increasing the overall number of fry released. Currently, production goals for kokanee fry releases into Lake Roosevelt are limited by kokanee egg availability, increasing the need to meet the program goal of creating a self-sustaining kokanee fishery in Lake Roosevelt. This has become more critical as the Lake Whatcom stock kokanee are slated to lose their pathogen-free status in the future, which will disallow them from being utilized for stocking in Washington State (Whatcom County Water Resources 2003; Peone 2007).  Additionally, Meadow Creek egg availability has been erratic. The Kootenay Lake fertilization program has increased their reliability, but the cost of eggs is expected to increase in the future and ultimately decrease the number of eggs available to the program. Management direction is to saturate designated release locations with kokanee fry. To meet this end, managers are developing a locally adapted Lake Roosevelt mixed stock of kokanee that can be collected at trapping locations on the reservoir such as Hawk Creek. Initial data indicate the strategy has been successful. Return rates observed for the Lake Roosevelt mixed stock are the highest recorded for hatchery kokanee in the reservoir at 4.5% in 2004, 7.1% in 2009, and 7.8% in 2010 (McLellan et al. 2010, Scholz et al. in progress), and sex ratios for this stock have been closer to expected ratios at 3 females to 5 males (McLellan et al. 2010). A total of 5,538 kokanee returned to Hawk Creek in 2010 (Scholz et al. in progress), from which a total of 75,691 kokanee eggs were collected and transported to the Spokane Tribal Hatchery for rearing and release in 2012. Scholz et al. (in progress) estimated that approximately 515,611 kokanee eggs could have been collected from adults returning to Hawk Creek if a full complement of eggs had been taken from all females. High mortality rates from the stress of handling, transporting, and holding adults until they were ripe limited the number of eggs that remained viable. While refining collection and trapping techniques at Hawk Creek will be necessary to reach management goals, clearly the recommended strategy is achieving the desired outcome.

A brief history of fish enhancements for Lake Roosevelt is covered in the other proposal sections. Additionally, the annual and supplemental reports submitted for the program, and referenced at the end of this report, document the entire history of fish enhancement work completed by this program and the associated artificial production program.

Hatchery Kokanee

The primary objective of the LRFEP during the first four years (1987-1990) of operation was to collect data on Lake Roosevelt before hatchery supplementation began. This provided baseline data co-managers could use to assess the effectiveness of the Sherman Creek and Spokane Tribal Hatchery’s kokanee salmon program. Initial annual relative abundance and creel survey estimates indicated kokanee abundance levels were less than 2.8% from 1989 to 1992, and annual kokanee harvest rates were between 11,906 and 17,756 (95% CI ± 3,597 and ± 1,382 respectively; Peone et al. 1990, Griffith and Scholz 1991; Griffith et al. 1995; Thatcher et al., 1996). Rainbow trout abundance (3-7%) was slightly higher than kokanee, while walleye abundance (10-19%) was considerably greater (Peone et al. 1990; Griffith and Scholz 1991; Griffith et al. 1995, Thatcher et al. 1996). Comparisons of annual harvest rates of rainbow trout [65,515 to 140,609 (95% CI ± 25,373 and ± 11,369 respectively)] and walleye [80,626 to 118,863 (95% CI ± 33,513 and ± 6,710 respectively)] indicated kokanee harvest was considerably less than either rainbow trout or walleye (Peone et al. 1990, Griffith et al. 1995, Thatcher et al. 1996). Kokanee growth rates were back-calculated and found to exhibit greater growth than observed in 19 other kokanee producing lakes located in the Columbia River Basin, suggesting that favorable growth controlling factors occurred in the reservoir (Peone et al. 1990, Griffith and Scholz 1991).

The Spokane Tribal and Sherman Creek Hatcheries began fish production of Lake Whatcom stock kokanee in 1991 and 1992, respectively, with the intent of raising 10 million kokanee fry (NWPPC 1987) with a goal of 1% escapement. However, due to egg accumulation constraints, the maximum number of kokanee fry released was closer to 1 million. After release, the LRFEP began assessment of hatchery performance and effects in the reservoir. Early kokanee studies conducted by the program were focused on two questions; 1) to determine the critical period for imprinting; and 2) to determine the best release strategies for maximizing kokanee recruitment to the fishery and returns to egg collection locations.

In order to assess performance of hatchery kokanee in the fishery, test groups of 1.8 million fry and 100,000 post smolt (16 months) kokanee salmon were released from the Spokane Tribal Hatchery with coded wire tags between 1992 and 1994. A total of 923,167 kokanee were released with coded wire tags in the three year period (Tilson et al. 1995). Of these, 67% were released as fry and 33% were released as residualized smolts (Tilson et al. 1995). A total of 431 coded wire tagged fish were recaptured between 1992 and 1994, of which 99% had been released as residualized smolts (4 fry and 427 post-smolt recoveries; Tilson et al. 1995). Entrainment was identified as the most likely reason for low recruitment of adult kokanee to the fishery based on the findings from the physiological and behavioral studies the LRFEP also conducted during this time. Scholz et al. (1993) and Tilson et al. (1994, 1995) found that kokanee undergo partial smoltification during their first and, to a lesser extent, second year of life, increasing their susceptibility to entrainment. When fish were held past smoltification, they had a greater likelihood of remaining in the reservoir (Scholz et al. 1993, Tilson et al. 1994, 1995). Based on these findings, co-managers recommended releasing the kokanee as yearlings to minimize hatchery kokanee entrainment. In 1995, the hatcheries began shifting production to kokanee yearling releases (50% released as yearlings, 50% released as fry) based on recommendations of the LRFEP.

In 1995, the creel survey conducted by the program estimated a harvest of 32,353 kokanee salmon, nearly double the harvest observed in 1994 (16,567 kokanee; Underwood et al, 1996b). Not all hatchery kokanee were adipose fin clipped at this time, therefore contributions from the hatchery program were estimated between 7% and 56% (Underwood et al, 1996b). Additionally, kokanee relative abundance increased to 22% in boat electrofishing surveys in 1995, a sizeable increase from previous years (<1-4% from 1989-1994; Cichosz et al. 1997a, McLellan et al. 1999; Miller et al. 2011). These data, combined with the partial smoltification study results and success of yearling releases compared to fry releases, led co-managers to believe that yearling kokanee releases would result in a strong kokanee salmon fishery and reiterated the recommendation that the hatcheries fully transition from fry releases to primarily yearling releases. In 1996, yearling kokanee comprised 85% of the kokanee hatchery releases.

The switch to yearling releases (1.0 million) in 1995-96 necessitated that the hatcheries begin utilizing net pens to decrease densities in the hatcheries to meet the goal of raising 100% yearling kokanee. A limited number of net pens were placed throughout the reservoir where access was already in place through the rainbow trout net pen program and in locations that were close to egg collection sites or had the potential to allow relatively easy collection of returning adults. The program monitored hatchery kokanee recruitment to the fishery, and examined release location and strategies for net pen and hatchery released kokanee from 1996-2000.

The early adult return studies also examined the effectiveness of utilizing synthetic chemicals to increase returns to collection locations. From 1992-1998, fish were imprinted to morpholine or phenethyl alcohol, and returns were assessed. The necessity of using chemicals was questioned when a portion of the 1993 and 1995 brood years not exposed to the synthetic chemicals returned to their release locations. Based on this, in 1998 Eastern Washington University examined the efficacy of imprinting kokanee salmon with morpholine to improve homing to collection locations and found no significant difference in return rates between the morpholine and non-morpholine exposed kokanee (McLellan et al. 2001). These results led co-managers to eliminate the synthetic chemical imprinting process in an effort to streamline artificial production/monitoring efforts and costs in 1999 (McLellan et al. 2001).

In 1998 and 1999, McLellan and Scholz (2001) found that fish released directly from Sherman Creek Hatchery exhibited greater returns than fish released from Kettle Falls net pens. Additionally, McLellan et al. (2001) found that lower reservoir net pens experienced the highest entrainment compared with other release groups, leading to a recommendation that kokanee not be released from net pens located in the lower reservoir. Early work indicated that returns were not increasing with yearling releases and numbers were still not sufficient to develop a self-sustaining fishery (McLellan et al. 2001). Overall return rates for all kokanee collected in 1998 and 1999 were 0.51% and 0.25%, which were similar to overall return rates of 1997 (0.55%), 1996 (0.20%) and 1995 (0.30%; McLellan et al. 2001). Of the fish collected at egg collection sites in 1998 and 1999, 90% were age 2 kokanee, which are not optimum to development of a self-sustaining fishery (McLellan et al. 2001).

Low adult kokanee return data was supported by harvest and abundance estimates for these years, which indicated limited recruitment of hatchery kokanee to the fishery. Kokanee relative abundance via electrofishing decreased in 1996 (4%) and 1997 (1%) compared with 1995 abundance numbers (22%; Cichosz et al. 1997a and 1999; Fields et al. 2005). Additionally, annual kokanee harvest in 1996 (1,265) and 1997 (588) was the lowest observed in 10 years of creel (Fields et al. 2005).

We know from early entrainment studies that reservoir operations can impact the fishery, but effects are not fully understood (Underwood et al. 1996, Underwood and Shields 1996, Cichosz et al. 1997, Cichosz et al. 1999). Reservoir operations during 1994 and 1995 were the most stable since the LRFEP began operating, characterized by minimal drawdown and high water retention time (Cichosz et al. 1999). As a result, the yearling kokanee hatchery releases of 1994 performed well in the 1995 fishery, partially as a result of yearling release strategies, but partly because of the favorable reservoir conditions. In contrast, the 1996 and 1997 hydro-operations were the most severe in the history of the LRFEP. In 1997, the lake was drawn down to minimum operating pool [1208 ft above mean sea level (AMSL)] with water retention times the lowest in the 50 year record (minimum 8 days; Cichosz et al. 1999, Fields et al. 2005).

The more moderate reservoir conditions that characterized later years (1998-2000) allowed kokanee to rebound from 1997 lows (Fields et al. 2005). High water retention times tended to decrease entrainment of fish, allowing greater numbers to remain in the reservoir, which was observed in the creel. High kokanee catch and harvest observed in the 1998 creel were believed to be kokanee that were entrained during the high flows of 1997 from British Columbia, Canada and were retained in Lake Roosevelt. This was consistent with a hypothesis that a portion of the wild kokanee found in the reservoir entrained through Canadian dams and remained in Lake Roosevelt as a wild component of the kokanee population. See Wild Kokanee portion below for current information on wild kokanee immigration.

Kokanee harvest increased in 2000 and 2001, but decreased in 2002 (Fields et al. 2005). This may have been due to suspect data collection and possible creel design flaws. To avoid future biases, and as suggested by the Independent Scientific Review Panel, the creel was redesigned by a biostatistician and initiated with data collected in 2004. In 2000, the creel indicated that catch (n=42,705) and harvest (n=29,145) estimates were the highest in the 10-year record for kokanee (Fields et al. 2005). Overall, the kokanee 10-year average estimated just 13,598 fish were harvested annually, with the majority being wild (unmarked) origin (67% wild; McLellan and Scholz 2001, 2002b; Fields et al. 2005). While the available hatchery kokanee fishery had been primarily facilitated by the yearling release strategies, early maturation limited their availability to the fishery. Yearling fish were released in the spring and matured the following fall, leaving only months until the hatchery fishery peaked during the first summer after release. This left the winter and early spring kokanee fisheries to be comprised primarily of wild fish. Despite low catch rates, kokanee were reported as the third most targeted fish, after rainbow trout and walleye in the 2002 creel (Fields et al. 2005). The contribution to the kokanee fishery by hatchery fish continued to be below co-manager goals. Despite modifications in the Sherman Creek and Spokane Tribal Hatcheries and an expanded net pen program to rear 1 million yearling kokanee, hatchery kokanee were not reaching management objectives.

Results from annual assessments of Whatcom stock post-smolt kokanee and adult collections from 1997-2004 found that less than 1.0 % of the releases returned to release sites and were comprised primarily of age 2 (21 month) precocial males (McLellan et al. 2001; McLellan and Scholz 2001; McLellan and Scholz 2002b; McLellan and Scholz 2003; McLellan et al. 2004a). The runs were typical of a precocial run, where a proportion of the males became sexually mature at a smaller size a year before the large run composed of relatively equal numbers of larger males and females (Foote et al. 1991). However, the larger run of age-3 kokanee never materialized in Lake Roosevelt (McLellan et al. 2004a). While small numbers of precocial and adult fish were collected during surveys or returned to collection sites, very few fish were observed in the recreational and subsistence fishery (Pavlik-Kunkel et al. 2008).

As a result, co-managers considered a native stock of kokanee that could potentially be locally adapted to conditions found in Lake Roosevelt. Lake Whatcom origin kokanee (a western Washington state coastal stock used since the inception of the hatcheries) were tested against a Kootenay Lake, British Columbia kokanee stock (Meadow Creek). Using the Meadow Creek kokanee stock addressed two factors: 1) rising concern over the genetic integrity of the kokanee populations in Lake Roosevelt and downstream of Grand Coulee and Chief Joseph dams and 2) provide a local stock that could potentially perform better in the reservoir. Paired release studies of Lake Whatcom and Meadow Creek kokanee were conducted from 2000-2004.

McLellan and Scholz (2001, 2002b and 2003) found age-2 Meadow Creek kokanee returned in significantly higher numbers compared to age-2 Lake Whatcom stock kokanee in both 2000 and 2001, indicating Meadow Creek kokanee may perform better in the reservoir than Lake Whatcom stock kokanee. However, no significant difference in return rates occurred for age-3 Meadow Creek and Lake Whatcom kokanee stocks (McLellan and Scholz 2003), and the number of age-3 kokanee of both stocks collected was extremely small [2000 (10), 2001 (45) and 2002 (88)] compared with age-2 kokanee collections [2000 (2,604), 2001 (1,858) and 2002 (527)]. While Meadow Creek kokanee outperformed Lake Whatcom kokanee as age-2 fish, an age-3 run never materialized for either stock. Because Meadow Creek kokanee returned in higher numbers, the managers decided to utilize Meadow Creek stock as the preferred stock for artificial production in Lake Roosevelt.

At this same time, adult kokanee returns continued to be plagued with early maturation and skewed sex ratios. Three-year old kokanee had been identified by the Lake Roosevelt co-managers as the primary goal due to the short period of time two-year old kokanee were available to the fishery (approximately 4 months when raised to yearlings). However, it was unclear whether Meadow Creek kokanee would perform to meet this objective. Lewis (1970) found that Meadow Creek kokanee tended to mature at ages one and two, while Lake Whatcom kokanee matured at ages two and three. Grant Gale (Ministry of Fisheries; personal communication in McLellan and Scholz 2001) stated that Meadow Creek kokanee were known to mature at age 4, when they reached 200 mm in total length. Ultimately, the rapid growth observed for kokanee in Lake Roosevelt may have contributed to the lack of three-year old returns. Age-2 hatchery kokanee in Lake Roosevelt average 289 mm in length after only three months in the reservoir (McLellan and Scholz 2002b), but tended to be sexually mature (McLellan et al. 2004a).

In 2002, investigations into hatchery rearing practices, rather than stock selection, were initiated to discover potential factors influencing sex ratio and early maturation. Higher frequencies of early maturation were observed when fry were raised at the Spokane Tribal hatchery and exposed to higher than optimal water temperatures (McLellan et al. 2007) in combination with a protein and lipid rich diet (STI unpublished data). Early maturation of kokanee prior to release was reduced by rearing fry at the Ford Trout Hatchery under cooler water temperatures (WDFW unpublished data). Precocity was also lower when fry were reared in spring water (a cooler water source than previously used) at the Spokane Tribal Hatchery (McLellan et al. 2007). However, the reduction in early maturation under these altered rearing conditions did not translate into increased age-3 harvest or returns (McLellan et al. 2007). Additionally, skewed sex ratios were observed primarily in Meadow Creek stock kokanee (McLellan et al. 2007). The skewed sex ratios were not as apparent when fish were reared under the lower water temperatures of the Ford Trout Hatchery or the spring water treatment at the Spokane Tribal Hatchery (WDFW unpublished data; McLellan et al. 2007). Kokanee are density-dependent and mature earlier if rearing conditions are favorable, therefore, it seemed likely that as long as the program maintained a yearling release strategy, the hatchery kokanee would mature at age-2 due to excessive growth experience during hatchery rearing. Because of hatchery water sources and infrastructure limitations, steps to reduce precocity could not be implemented.

In addition to early maturation investigations, co-managers began assessing other potential limiting factors affecting kokanee. In the mid 1990’s it was evident that data gaps existed for the limnetic oriented fish (kokanee, rainbow trout, burbot (Lota lota), and lake whitefish (Coregonus clupeaformis)) because the majority of the monitoring efforts had focused on the littoral zone (Cichosz et al. 1997a, 1999; Baldwin et al. 2005). Black et al. (2003) performed a carbon and nitrogen stable isotope analysis of the aquatic food web and noted that the majority of fishes in Lake Roosevelt occupied a fairly high trophic level and primarily utilized pelagic carbon. Results indicated that the aquatic food web in Lake Roosevelt is pelagically driven. Thus in 1998, the LRFEP began a limnetic sampling regime for Lake Roosevelt that was conducted by WDFW. Hydroacoustic surveys were used to collect data necessary to the bioenergetics model that was used to determine limiting factors for kokanee in Lake Roosevelt. The hydroacoustic surveys and model permutations focused on determining effects of predation, entrainment, food limitations, and abiotic influences (temperature and dissolved oxygen; Baldwin et al. 2005).

Reservoir temperatures were found to be in the preferred range for most limnetic fishes in June and October, but were too warm for kokanee in August in the upper 50 m, of the water column (17-23ºC; Baldwin et al. 2005). Surveys conducted in August, found that the majority of kokanee collected in the lower reservoir [from Grand Coulee (rkm 960) Dam to Whitestone Rock (rkm 999] were of wild origin and found below 50 m where water temperatures were less than 16^oC (Baldwin et al. 2005), suggesting kokanee may be migrating 30-50 meters on a diel cycle to feed in the photic zone during the day. Attempts were made to evaluate daytime and dawn distribution of kokanee in certain transects, but results were inconclusive (Baldwin et al. 2005). Kokanee captured in the middle section [from Lincoln(rkm 1019) to Hunters (rkm 1065)] of the reservoir were also captured in the lower half of the water column, again where stratification provided refuge from warm temperatures. In the upper, shallower portion of the reservoir [from Gifford (rkm 1085) to Kettle Falls (rkm 1128)], temperatures were nearly isothermal and kokanee were forced to occupy the warmer water. These data indicated a possible reduction in optimal cool water habitat in the late summer months forcing kokanee to congregate below 50 m depth (Baldwin et al. 2005).

The hydroacoustic survey and bioenergetics modeling also examined the potential impacts of predators on the artificial production program at multiple spatial and temporal scales (Baldwin and Polacek 2002; Baldwin et al. 2003). Baldwin et al. (2003) collected abundance and diet data on walleye (Sander vitreus) to determine predation effects on hatchery kokanee salmon and rainbow trout in Lake Roosevelt. The Wisconsin Bioenergetics model was used to generate daily consumption estimates for each age class of walleye, ages three through seven, over a 31-41 day modeling period (Baldwin et al. 2003). Walleye consumed 9.4% of hatchery kokanee salmon and 7.3% of hatchery rainbow trout released into the reservoir within 41 days of release (Baldwin et al. 2003). These numbers are generally considered to be an underestimate of total predation on hatchery kokanee and rainbow trout because walleye less than 300 mm and other predacious fish species were not included in the estimates (Baldwin et al. 2003). When this approach was applied to the whole reservoir on an annual basis, it was determined that walleye could account for 31-39% of the kokanee mortality and 6-12% of the rainbow trout mortality (Baldwin and Polacek 2002). Based on these findings, predation clearly impacts the artificial production program; however, other limiting factors were also contributing to the lack of recruitment to the fishery and hatchery facilities.

Co-managers began developing alternative strategies that were designed to release kokanee in areas where predator impacts would potentially be lessened in an attempt to improve returns to egg collection sites. In 2002, yearling kokanee that were directly released at the Fort Spokane and Gifford boat launches exhibited the highest return rates compared with other, more traditional release locations and methodologies (McLellan et al. 2004b). The Fort Spokane and Gifford locations were characterized by large, deep pelagic areas that offered a refuge from predators, and were not located in areas known to possess the highest walleye densities (McLellan et al. 2004b). Comparisons of recovery rates utilizing more traditional release locations and strategies such as the net pens (0.03%) and the annual Little Falls Dam release (0.02%), indicated that recovery rates for Fort Spokane and Gifford were much higher (0.45% and 0.52%, respectively; McLellan et al. 2004b).

The success of releasing kokanee directly into areas with predator refugia, prompted the move to release kokanee away from predators rather than at holding sites where predators were most likely located. Initially, yearling kokanee were released into tributaries in the upper reservoir that would allow easy collection of returning adults. Meyers Falls was selected as a release location because it had the potential to mimic conditions wild fish immigrating from Canada experience and possessed a natural barrier that blocked upstream migration where adult kokanee could be collected. In 2001 and 2002, approximately 25,000 kokanee were released at Meyers Falls in the plunge pool (McLellan et al. 2004b, 2006). A relatively high return rate of age two kokanee (0.40%) was observed for this release group, suggesting this release strategy may have future potential (McLellan et al. 2004b). However, low returns were observed the following year (0.03%; McLellan et al. 2006). The reason for reduced returns was not clear, but the limited number of kokanee released at the location and potentially confounding effects by early maturation, skewed sex ratios, predation and reservoir operations may have played a key role in low returns that year.

Approximately 250,000 to 475,000 (initial release was 25,000) post-smolt kokanee released annually at Fort Spokane serendipitously created a put-and-take fishery in the middle reservoir. Every year in August since 2000, anglers competed in the Two Rivers Casino Trout Derby, which provided the opportunity to collect catch data from a large number of anglers specifically targeting species that are supplemented by the artificial production program. The success of the Fort Spokane release was indicated by recoveries of marked fish captured during the Derby. From 2001-2008, the catch of hatchery supplemented kokanee varied from 44% (n = 24) to 93% (n = 338) (McLellan et al. 2004b; McLellan et al. 2006; McLellan et al. 2007; McLellan et al. 2008; McLellan et al. 2009; Miller et al. 2011), and the contribution to the total catch (rainbow trout and kokanee salmon) ranged from 4.4 % to 27.2 %. In 2009, the Derby date was moved to September from August, and hatchery kokanee comprised 86% of the catch (n = 63), which was consistent with past years. All hatchery kokanee were Lake Whatcom stock released from Fort Spokane. No mixed stock kokanee were captured, which is logical because the mixed stock fish were early run, and were already peaking on the spawning run during the September Derby. The success of the Fort Spokane release strategy led co-managers to establish it as an annual put-and-take release strategy until other release strategies proved more effective. Based on past failures and successes in the kokanee artificial production program, the current management direction is to permanently maintain the portion of the kokanee program that is perceived as a success. The success of this open water release was also evident both in adults returning to Hawk Creek (> 1% returns in most years, which met escapement goals) and fish observed in the creel (Pavlik-Kunkel, et al. 2008; McLellan et al. 2011; Miller et al. 2011). Kokanee released at Fort Spokane tend to return to Hawk Creek in abundant numbers, recruit to the summer fishery and have higher survival rates compared to fish released at other locations (McLellan et al. 2009).

In order to maintain the put-and-take fishery, managers recommend 250,000 Lake Whatcom yearling kokanee be released annually. From 2009-2010, an additional 16,000 Lake Roosevelt (mixed) stock yearling kokanee were stocked for the management of the put-and-take fishery (Peone, 2010, 2011). In both 2006 and 2008, escapement to Hawk Creek was low, which coincided with deep drawdown events (McLellan et al. 2011). In 2009, yearling return rates for the Fort Spokane release of both the Lake Whatcom (1.7%) and mixed stocks (7.1%) were encouraging and releases were recommended to continue (McLellan et al. 2011). Favorable conditions likely provided increased escapement.  In 2009, drawdown was shallow and the mixed stock was released at a larger size (2-5 fish/pound), which may have been a factor in their normal sex ratio (McLellan et al 2010). Future studies will use this model (i.e., smaller release amount and larger fish size) to try to increase escapement. Fish were released post-drawdown to prevent entrainment loss and to assure that zooplankton biomass was sufficient to support the released fish.

In 2003, the LRFEP embarked on a new strategy to increase hatchery kokanee performance in Lake Roosevelt. The yearling program had maximized the capacity at the Spokane Tribal Hatchery because of the larger sized kokanee, but space was available in the hatchery if the fish were released as fry. The Lake Roosevelt managers decided to take the opportunity to re-examine fry plants in Lake Roosevelt due to the availability of large numbers of Meadow Creek stock eggs. Strategies used during previous fry plants could possibly explain the poor performance of fry observed in the early 1990’s. Lake Whatcom stock kokanee was used for initial fry experiments because it was the only stock available at the time. Studies conducted through this program demonstrated Meadow Creek kokanee perform better than Lake Whatcom kokanee in Lake Roosevelt. Managers hypothesized that utilization of the coastal Lake Whatcom stock, rather than a locally adapted stock, could be one of the factors that limited the success of initial fry attempts. Additionally, kokanee fry were released below Little Falls Dam (on the Spokane River) and from Sherman Creek (in the upper reservoir). Fry stocked at both areas were immediately exposed to the fast flows found in Lake Roosevelt during the spring runoff, which potentially increased entrainment out of the reservoir. Also, the Little Falls Dam area of the Spokane River had been characterized as the primary walleye spawning area in Lake Roosevelt (McLellan et al. 2002a). Likewise, the Sherman Creek area of Lake Roosevelt had been documented as the primary walleye summer home range (McLellan et al. 2002a). Walleye predation on post-smolt kokanee after release from Sherman Creek was documented by Baldwin et al. (2003). The high predator densities found to exist at the previous fry release locations could also have limited kokanee fry survival under those release strategies. The new fry release strategies would take into consideration past fry planting limitations (entrainment and predation) and take advantage of techniques that had improved the yearling portion of the hatchery kokanee program (use of an upper Columbia River native stock and spatial removal from predators). It was hoped that saturation planting of fry at a site removed from walleye predators and far enough from Grand Coulee Dam would reduce entrainment and increase the success of fry plants in Lake Roosevelt.

In an attempt to overcome some of the obstacles of releasing post-smolt fish and their limited contributions to the recreational and subsistence fisheries, the managers again adjusted release locations and timings, and size at release. Two tributaries in the northern section of the reservoir were selected for Meadow Creek stock fry plants (Big Sheep Creek and Onion Creek) and one tributary in the middle reservoir (Hawk Creek) was selected to receive Lake Whatcom stock kokanee. From 2004-2006, the artificial production program released approximately 400,000 post-smolts annually at various reservoir locations, and 3 million kokanee fry annually into tributaries in the northern reservoir. Assessments following these release strategies, indicated that releasing kokanee post-smolts in open water following the start of refill led to greater adult kokanee returns to collection sites, but had limited success in recruiting kokanee to the fishery (McLellan and Scholz 2002b, 2003; Pavlik-Kunkel, et al. 2008). Fry release strategies aimed at mimicking the natural recruitment from upriver sources were unsuccessful, with low returns to stocking tributaries and low recruitment to the recreational and subsistence fishery (McLellan et al. 2008, 2009). No kokanee fry returned to either Big Sheep or Onion creeks in the northern reservoir or to Hawk Creek in the middle reservoir the first year of collection (2005) (McLellan et al. 2007). During the second sampling year, escapement was minimal. Only 10 kokanee (5 hatchery origin) were collected in Big Sheep Creek, and only 41 kokanee (24 hatchery origin) were collected in Onion Creek (McLellan et al. 2008). The third year of collection also had limited returns for both Big Sheep Creek (n = 2; 1 hatchery origin) and Onion Creek (n = 35; 32 hatchery origin), while Hawk Creek had the largest adult returns (n = 4,702) (McLellan et al. 2009). Of these 4,094 (87%) were Meadow Creek hatchery kokanee, 601 (13%) were Lake Whatcom stock hatchery kokanee and 7 were unmarked (McLellan et al. 2009). No fry or yearling plants occurred in Onion or Big Sheep creeks in 2008; however, monitoring continued in the fall to capture returning adults (McLellan et al. 2009). Due to low escapement of both fry and yearling kokanee in Big Sheep and Onion creeks, the studies were terminated. In the spring of 2008, in an attempt to get kokanee released as fry to return to Hawk Creek, fry were released above the falls (a natural barrier) of Hawk Creek.  Future analysis will determine if this study is successful.

Many factors influence escapement in Lake Roosevelt, including predation, stock, release location, timing, and size.  The LRPEP developed protocols to increase kokanee escapement in the reservoir (McLellan et al 2010).  These include stocking the Meadow Creek stock; however, availability is unpredictable. The Lake Roosevelt mixed stock (Meadow Creek and wild, unmarked kokanee) is often variable and depends on annual adult returns.  Lake Whatcom is more predictable, yet they are not a localized stock. Fort Spokane has proven to be a successful release location, probably do to the open water release, which provides refuge from predators.  Fish released from Fort Spokane return in generous numbers and have helped to provide an egg source to create a mixed stock within Lake Roosevelt. Fish are released after the reservoir reaches 1260 msl, depending on fish health, extreme weather, and varying hydro-operations, usually near the end of May.  The Lake Roosevelt mixed stock was consistently released between 1.5-2.5 fish/pound, while other stocks were released between 6-21 fish/pound.  Due to the success of the mixed stocks in most years, release size has changed to 5-7 fish/pound.  Co-managers agreed to meet this goal, even if less fish are stocked (McLellan et al. 2011).

The Lake Roosevelt managers required all hatchery origin fish stocked in Lake Roosevelt to be marked. All yearlings were given an adipose fin clip and an additional fin clip was given to experimental fish. All fry were originally marked by heating/cooling water to thermally mark otoliths, and in 2009, strontium chloride was used to mark kokanee fry otoliths. Concern regarding additional handling of hatchery fish required with chemical marking of otoliths led managers to re-assess marking techniques, and return to thermally marking otoliths, albeit with a more structured approach. In 2009, managers agreed that converting to a thermal marking program for kokanee fry released into the reservoir was a priority. LRFEP project proponents, with assistance from Bonneville Power Administration, coordinated a within-project funding modification to procure a chiller unit to be installed at the WDFW Spokane Fish Hatchery. Kokanee fry released into the reservoir are 100% thermal otolith marked using the facility.

The Lake Roosevelt kokanee artificial production program had been plagued with walleye predation, early maturation, skewed sex ratios and entrainment, which likely contributed synergistically to reduced hatchery kokanee numbers in the reservoir. The goals of developing a fishery that could be utilized for subsistence and recreational purposes as well as be self-sustaining had not been reached, despite extensive monitoring and adaptive management based on study results. Therefore, alternatives were considered that would mimic the wild kokanee population. Based on low recruitment of hatchery produced kokanee to the creel and to egg collection facilities, an investigation into the contribution of kokanee immigrating into Lake Roosevelt from upriver sources was warranted. Additionally, because wild kokanee had been the primary contributors to the reservoir-wide creel efforts (Fields et al. 2005), and little was known about their life history, co-managers believed a better understanding of life history strategies was warranted. Consequently, life history strategies would be mimicked at the hatchery level (i.e., timing, size, and location of release [Polacek 2008]), and this information would also be used in bioenergetics modeling to assist co-managers to better understand kokanee survival.

Wild Kokanee

Kokanee and sockeye salmon existed in the Upper Columbia River before the construction of Grand Coulee Dam blocked anadromous fish migrations. The wild kokanee population status has been monitored through limnetic fish surveys for multiple years (2000-2008), indicating that the population has varied little in abundance with a mean estimated population size based on acoustic and gill net surveys of 20,845 or 0.63 fish/ha (Baldwin et al. 2006). Kokanee densities in the reservoir were low but they were still observed in the creel. In 2008, the kokanee harvest was 0.16 fish/angler and relative abundance was 8% (Miller et al. 2011). From 2000-2008, kokanee relative abundance has ranged from 3-25% (mean 8.6%).

Determining what drives the wild kokanee population in Lake Roosevelt is a critical link for managers, as wild origin kokanee have been an essential part of the Lake Roosevelt fishery for decades. By understanding the timing and size of kokanee immigration from British Columbia, Canada (Kootenay and Arrow Lakes) into Lake Roosevelt, hatcheries would be able to mimic wild kokanee behaviors and adapt release strategies to ultimately increase kokanee recruitment and escapement to egg collection facilities (Polacek 2008). From October 2006 to September 2008, screw traps were placed in Lake Roosevelt in an attempt to collect wild origin kokanee salmon (Polacek 2008). Overall, the traps collected 34 kokanee (mean TL 218mm); however, most of these fish were of hatchery origin from a recent plant from upstream Onion Creek. Only two wild kokanee were collected, and it was suggested this was due to multiple factors, including low kokanee densities, kokanee followed different currents, kokanee location within the water column was deeper than the trap fished, and/or low numbers of kokanee were entrained that year (Polacek 2008). It was estimated that nearly 51,000 kokanee passed the trapping site, of which 51% were from the Onion Creek release, assuming that kokanee may have remained for extended periods south of the Little Dalles and that walleye predation may have occurred at some point prior to reaching the trapping site (Polacek 2008). Unfortunately, multiple mechanical failures, log damage, turbulent water, and trap placement made sampling difficult (Polacek 2008), and the expected results could not be obtained. The author recommended that future immigration studies continue, and to use the current study as a pilot project if screw traps were to be used again (Polacek 2008).

An early genetic study conducted by the Chief Joseph Kokanee Enhancement Project, in cooperation with the Lake Roosevelt Fisheries Evaluation Program, originally supported the contention that wild kokanee immigration was occurring in Lake Roosevelt. The study indicated that the majority of unmarked kokanee in Lake Roosevelt were not from wild produced hatchery origin fish, but comprised of Sanpoil River and upper Columbia River stocks, indicating immigration from Canada (LeCaire et al. 2000; Loxterman and Young, 2003), potentially due to entrainment from Arrow Lakes. At the time, the study results may have been deceiving, despite their assignment to those populations, because the numbers of wild kokanee (adipose intact kokanee of unknown origin will be referred to as wild kokanee) collected may not have been enough to accept this conclusion (Kassler et al 2010). These fish may have assigned to the Sanpoil River because there were no wild reaches in the baseline for that analysis or that other sources of kokanee (Meadow Creek or Arrow Lakes) were genetically similar to the Sanpoil River and were contributing to the wild population of kokanee in Lake Roosevelt (Kassler et al. 2010). However, a more recent genetics characterization study indicated that Lake Roosevelt wild kokanee may be a genetically distinct population with some similarities to the Sanpoil River kokanee and Nespelem River kokanee (Kassler et al. 2010). While showing genetic similarities to each other, the Lake Roosevelt wild kokanee population showed no similarities to all other analyzed collections (Kassler et al. 2010). Lake Whatcom and Meadow Creek kokanee shared only minimal genetic ancestry with the wild kokanee, and the authors suggested that the majority of genetic ancestry of the Lake Roosevelt wild kokanee was from a naturally reproducing population within the Lake Roosevelt system (Kassler et al. 2010). Due to the genetic differences between the wild Lake Roosevelt kokanee and the upper Columbia stocks (below Keenlyside Dam and Arrow Lakes), the authors suggest that kokanee do not migrate downstream into Lake Roosevelt (Kassler et al. 2010). Additional genetic collection and further analysis needs to be conducted before there is a final conclusion on stock origin and the presence of an in-reservoir reproductive population. In Lake Roosevelt, tributary escapement data indicates that few kokanee spawn in streams. In the Sanpoil River, escapement ranges between 2 and 5 wild kokanee each year (Colville Tribes, unpublished data). Due to these low numbers, researchers have hypothesized that deep-water lakeshore spawning could explain the sustained survival of Lake Roosevelt’s wild kokanee population, and recommend studies to examine deep-water spawning hypotheses.

In 2008, Eastern Washington University initiated a four-year kokanee acoustic tagging study to determine several strategies used by kokanee. These included monitoring how kokanee seasonally moved throughout the reservoir, what depth was occupied thermally and dielly, how temperature affected kokanee movement, how hydro-operations affected kokanee, and identification of fall spawning migration patterns and spawning grounds (McLellan and Scholz 2010). During the first two years of the study, most kokanee were found to use the lower third of the reservoir, and only two migrated above the Spokane River confluence. Kokanee frequently moved within the study area, contrary to the belief that they remained in localized areas. Nearly half of the tagged fish utilized the Sanpoil River in winter and spring possibly due to increased water temperatures and/or food. Kokanee were found to have distinct spring and summer diel migrations. Eighty-three percent (n = 15) exhibited spring diel movement (early morning/late evening diel movements near the surface and deeper depths during the day), and once water temperatures rose, 100% exhibited summer diel movements (daytime diel movements to the surface with deeper depths at night). Kokanee must vertically migrate from 50-100m to avoid increased temperatures, find food during the day near the surface, and seek refuge at night at increased depths. This suggests that kokanee had substantial physiological stress due to the high metabolic demands of migrating ≥50m to feed and then digest. These extreme migrations are caused by a mostly isothermal water column with higher temperatures (≥20°C) than kokanee prefer (12.2°C), forcing them to cooler water. Baldwin and Woller (2006b) did not determine food to be a limiting factor for littoral kokanee; however, accessibility to food may be a limiting factor to these populations. The effects of hydro-operations on kokanee recruitment have not been assessed; however, the effects on rainbow trout have (McLellan et al. 2008). Drawdown depth, water retention time, and release time all impact rainbow trout recruitment to the fishery. Entrainment over Grand Coulee Dam has consistently been a limiting factor for fish in the reservoir, and it has been difficult to quantify kokanee entrainment. Acoustic data showed two fish entrained over the dam and two with similar movement patterns. Drawdown was not a factor in kokanee movements, but refill caused an increased downstream movement, which could have been the increased push of water or the following of zooplankton being flushed downstream. No large spawning locations have ever been discovered in Lake Roosevelt, and only a few (10-20) wild kokanee have been found annually in tributaries throughout the reservoir (McLellan and Scholz 2010). Genetics testing discovered that most wild kokanee within Lake Roosevelt were genetically distinct from upriver Canadian stocks, suggesting that Lake Roosevelt has its own spawning population (Kassler et al 2010). Only a small percentage of kokanee were found to be genetically similar to the upriver stocks (Kassler et al 2010), which are the few that migrate into Lake Roosevelt from Canada. To date, annual sampling has yet to locate a large, wild spawning population. The only exception is one tagged fish that was detected below Brilliant Dam, in British Columbia. It swam 219km from Grand Coulee Dam in only 5 days and stayed below the dam for 30 days. It eventually moved back to Lake Roosevelt (Kettle Falls), and detection ceased soon after. This fish may have been returning to its original spawning grounds, trying to find another spawning ground, or was a subject of uncertain mortality. The acoustic tagging study will continue through 2011, at which time a complete data set will fill in the life history data gaps to help better manage the wild kokanee population in Lake Roosevelt.

Current Program

Hatchery Kokanee 

The current Lake Roosevelt Fisheries Evaluation Program management goals for hatchery Lake Roosevelt kokanee include: 1) develop and maintain a fishery; 2) develop a Lake Roosevelt fishery that could be used for the artificial production program; and 3) restore a natural run of kokanee to Lake Roosevelt (Pavlik-Kunkel et al. 2008). Managers have tried several release strategies with limited success, yet managers are determined to reach the current goals with continued strategies.

A current, ongoing strategy (beginning in 2008) to compare post-smolt kokanee to kokanee fry releases was to saturate portions of the reservoir by releasing larger numbers of both post-smolts and fry into the reservoir. The first part of the strategy was to conduct an assessment of hatchery kokanee fry versus post-smolt release success. Approximately 475,000 post-smolts were released at Fort Spokane and approximately 2 million fry were released into the lentic portion of the reservoir (near Seven Bays). All releases of fish were delayed until after refill began to avoid entrainment. Additionally, releasing fish in open water areas may have reduced overlap with predators in the reservoir. This fry release strategy is different than past efforts because of the timing, release location, and release method. Unlike the current proposal, the previous fry releases were early spring direct releases into the upper reservoir, which likely resulted in substantial losses due to predation, lack of suitable forage, and entrainment.

The second part of the strategy, proposed by the Colville Tribe, was to supplement the Sanpoil River with kokanee. The proposal includes releasing up to 800,000 kokanee fry into the Sanpoil River (Gold Creek), with the intent that those fish migrate to the reservoir, remain throughout their development, and return as adults to the Sanpoil River. The assumptions for this release strategy are that fish will be less susceptible to entrainment out of Lake Roosevelt following their migration from the Sanpoil River, and that they will distribute to, and reside in, the portion of reservoir that offers the best habitat and survival opportunities for kokanee.

Past attempts of collecting eggs from adult returns to Sherman Creek to reestablish a self sustaining kokanee population were unfeasible with returns at relatively low levels (<5,000). However, the number of adult returns to Hawk Creek had shown potential for a local egg source.  It was decided that these fish were able to survive and adapt to local reservoir conditions and continue to return in fair numbers suggesting better future survival than fish that did not return (McLellan et al. 2011). Due to temporal separation in spawn timing and physical appearance versus Whatcom stock, it is possible to separate and spawn only localized Meadow Creek and “Lake Roosevelt” stocks (aka: mixed stock). Therefore in 2002, kokanee egg collection from select spawning sites (Spokane River and Hawk Creek) began. Assessment of the Lake Roosevelt stock returns was positive with higher return rates (2.0-4.5%) than the historical average (0.50%) (McLellan et al. 2006). Also, annual egg collections contributed between 10-30 thousand additional kokanee yearlings for Lake Roosevelt (McLellan et al. 2011). In spring 2008, kokanee fry were released above Hawk Creek falls. Adults from this experiment were

expected to return fall 2010. In 2009 and 2010, (2,790 and 12,420, respectively) Lake Roosevelt Stock yearlings were released at Fort Spokane. In 2009, an adult weir trap was placed at Hawk Creek just below the plunge pool to collect returning kokanee to use as a potential internal egg source (McLellan et al. 2011). In 2009, a record 8,411 kokanee were captured at Hawk Creek during the fall spawning run in 2009. The majority were age-2 Lake Whatcom fish from the Fort Spokane release; however, the greatest return rate was by the Lake Roosevelt stock kokanee released at Fort Spokane (7.1% return). This was the highest return rate observed for any release group of kokanee since hatchery kokanee monitoring began in 1989 in Lake Roosevelt (McLellan et al. 2011). The sex ratio for Lake Whatcom returning adults was highly skewed towards males (1 female:34 males), while the sex ratios for the mixed stock were within the normal range (1 female:7 males). Kokanee have been released from the Fort Spokane boat launch periodically, beginning in 2004, with relatively consistent, positive escapement data. In years without deep drawdowns, return rates have enabled managers to develop a Lake Roosevelt hatchery mixed stock.

Adaptive Management Strategies: Concern regarding the health of the wild kokanee salmon population in Lake Roosevelt increased due to the potentially high numbers of wild kokanee harvested from Lake Roosevelt. As an interim measure to protect native kokanee salmon, the co-managers cooperatively agreed that WDFW should implement a protective regulation for kokanee, excluding all kokanee with an adipose fin from the harvest beginning in 1998. Beginning with the 1998 release year, all hatchery origin kokanee salmon released as yearlings have been adipose fin clipped.

During the 2000 winter supplementary creel conducted by the LRFEP, a ratio of 50 wild kokanee were harvested to only 1 hatchery kokanee, and annual creel surveys indicated that the majority of kokanee harvested were wild fish (Fields et al. 2005). This generated concern regarding incidental-hooking mortality of wild kokanee and led to a protective regulation (only two kokanee per angler per day, regardless of whether their adipose fins were clipped or not). The co-managers felt the new regulation had a higher likelihood of conserving a greater number of wild kokanee than the previous regulation. In 2010, in an attempt to encourage anglers to harvest hatchery kokanee, yet still conserve the wild fish population, the regulation was changed and currently allows for a daily bag limit of 6 kokanee (only 2 wild, unmarked with adipose fin intact), and no minimum size limit.

Managers continue to attempt to address predation effects in Lake Roosevelt through proposed harvest regulations on walleye and smallmouth bass (Micropterus dolomieu). Beginning in 2007, anglers were allowed a daily harvest of eight walleye with no more than one over 22 inches with no minimum size limit. This was intended to reduce the number of small walleye and balance the population (2012- 2013 Sport Fishing Rule Change Proposals 2011) and at the same time allow for a more aggressive harvest than in previous years, while still protecting larger fish. A current proposal is in place to increase the daily limit of walleye in Lake Roosevelt from 8 to 16 fish per day with only 1 over 22 inches. The proposal is intended to improve the walleye fishery, increase walleye condition, and limit walleye predation on salmonid species (WDFW 2011). WDFW estimated that between 75,000 and 100,000 walleye need to be harvested annually, and in the past few years (2007-2009), only about half (40,000) were harvested (WDFW 2011). In order to reach this goal, the amount of walleye harvested needs to increase to more than double.

Also, smallmouth bass were placed in a category entitled Statewide Freshwater Species Rules so that daily harvest limits were uniform across the entire state of Washington. The daily bag limit increased to ten fish, with only one allowed over 14 inches. This would promote aggressive harvest of the most abundant sized bass, while protecting the harvest of trophy fish. A new proposal is also being put forth to allow anglers a two pole endorsement in the lower Spokane River (a mid-reservoir tributary of Lake Roosevelt; from the SR 25 bridge to 400 feet below Little Falls Dam). The kokanee and walleye regulations listed above will also be allowed on the specific section of the Spokane River. The two-pole endorsement will allow anglers increased recreational opportunity to harvest mitigated kokanee as well as increased opportunity to harvest the proposed increased daily limits of walleye.

The LRFEP developed a set of protocols to maximize escapement (current goal, 1%). These include: 1) continue to stock Meadow Creek kokanee, from British Columbia, Canada. The availability of this stock is unpredictable, and some years Lake Roosevelt receives no eggs, others up to 4 million. The Lake Whatcom stock is more predictable, and the program usually receives 1.8 million eggs a year. The preferred stock is Lake Roosevelt mixed stock, comprised of F₁ generation Meadow Creek kokanee captured at spawning sites, primarily through trapping at Hawk Creek. However, this stock is also variable and depends on the number of adults returning each year. Therefore, stocking strategies are developed annually based on the number of eggs obtained each fall. 2) Continue Fort Spokane releases to increase post-stocking survival. The mechanisms for this are unclear, but are likely the result of reduced predation because of the immediate access to deep limnetic waters at the confluence. These fish return to Hawk Creek and attempt to spawn there. This has become the primary egg collection location.  3) Continue to release kokanee after reservoir refill begins, typically near the end of May. The protocol cannot always be met, due to various issues including fish health, extreme weather, and variable reservoir operations. 4) Continue to release the Lake Roosevelt mixed stock at a larger size (1.5 – 2.5 fish/ lb). Other stocks were released at smaller sizes (6-21 fish/lb), and therefore, a new recommendation was developed to stock fish at 5-7 fish/lb. Co-managers agreed to meet this goal with future releases, even if less fish were released (McLellan et al 2010).

A new study was initiated in 2010 to evaluate movements of hatchery kokanee from the Fort Spokane release. Since the put-and-take fishery is only four months long, it was prudent to put the fish where anglers targeted kokanee, which is primarily in the lower reservoir. Due to entrainment, it was not ideal to put fish close to Grand Coulee Dam; however, anglers target walleye in the upper sections and kokanee in the lower sections (McLellan et al. 2011). The objective was to determine if hatchery kokanee prefer to stay in the local vicinity of release for the four months prior to spawning, or if they spread throughout the entire lower reservoir system, thereby staying in prominent kokanee angling areas.

Hatchery kokanee management in Lake Roosevelt has always been complex because of the many inconsistencies every year. It is difficult to know on an annual basis if the program will receive eggs, how many it will receive, where they will come from, how to mark them, where and how many to stock, and how variable the reservoir ecosystem will be? Unlike the rainbow trout program, which has consistent egg allotments, the kokanee program needs constant evaluation and adjustments (McLellan et al 2010). The LRFEP strives to assist the Lake Roosevelt Management Team (LRMT) to reach the goals for Lake Roosevelt, which are stated in the Guiding Document for Lake Roosevelt (LRMT 2009) to be: 1) conservation, enhancement and restoration of native species; 2) provide and maintain subsistence fishing opportunities for Native American Tribes; and 3) provide and maintain sport fisheries that are economically productive for Lake Roosevelt and surrounding communities. Kokanee have been and will continue to be a vital part of these goals.

Despite the issues challenging both the kokanee artificial production program and the wild kokanee population, co-managers will continue to examine remaining alternatives towards the development of a viable kokanee fishery, determine how wild kokanee survive and recruit to the fishery, and strive towards developing a self-sustaining fishery that meets subsistence and recreational fishing needs based on requests for the return of a viable subsistence salmon fishery by local tribes. Salmon are part of the cultural heritage of Northwest Tribes, including the Spokane and Colville Confederated Tribes. The loss of salmon to the tribal people is deeply felt on cultural and social levels, and the tribes continue to support development of a kokanee salmon fishery as long as alternatives exist that may provide a successful fishery and until such time as anadromous salmon return to the upper Columbia River. Co-managers plan to continue to develop, implement, and monitor alternative strategies to address limiting factors and kokanee success in Lake Roosevelt.

Rainbow Trout/Redband Trout:

Rainbow Trout

Production and release of coastal rainbow trout into Lake Roosevelt began in 1986 as partial mitigation for the loss of anadromous fishes in Lake Roosevelt due to the construction of Grand Coulee Dam (Spotts et al. 2002; Peone 2007). The Lake Roosevelt Net Pen Program (operated by the Lake Roosevelt Development Association [LRDA], a nonprofit, volunteer organization) acquired 50,000 fish for release from the WDFW Spokane Fish Hatchery (Spotts et al. 2002; Peone 2007). The LRDA operated approximately 30 net pens at Hunters, Seven Bays, Two Rivers, Keller Ferry, Lincoln, Hall Creek and Kettle Falls. By 1990, releases had increased to 276,500 (Spotts et al. 2002). Baseline data on rainbow trout populations were gathered by the LRFEP prior to initiation of the mitigation program and release of fish from the Spokane Tribal and Sherman Creek hatcheries in 1991 and 1992, respectively. Relative abundance of rainbow trout during 1989-1992 ranged from 3-7% (Peone et al. 1990; Griffith and Scholz 1991; Griffith et al. 1995; Thatcher et al. 1996). Annual harvest estimates for rainbow trout during the same period ranged from 65,515 (95% CI ± 25,373) to 140,609 (95% CI ± 11,369) (Peone et al. 1990; Griffith and Scholz 1991; Griffith et al. 1995; Thatcher et al. 1996). For the first few years after establishment of the rainbow trout mitigation program, the Spokane Tribal and Sherman Creek hatcheries released 400,000 Spokane stock (coastal origin) yearling rainbow trout. From 1995-2006, hatchery production of rainbow trout increased to 500,000 yearlings annually (Spotts et al. 2002; McLellan et al. 2004b; Fields et al. 2005), and from 2007-2010, production increased to 750,000 yearlings annually (Peone 2007, 2008, 2009, 2010, 2011).

Historically, stocking strategies for rainbow trout have included hatchery rearing from egg to fingerlings, then transfer to net pens to grow to yearlings, at which time they were released into the reservoir. The WDFW Spokane Trout Hatchery provides 1.1 million (annual egg take goal) triploid Spokane stock eggs to the Spokane Tribal Hatchery in the winter (Dec-Jan) for culturing and fry to juvenile production (Jan-Oct). Approximately 300,000 juveniles are transferred to the Sherman Creek Hatchery in July to alleviate rearing capacity issues at the Spokane Tribal Hatchery. Assuming survival rates of 80 percent egg to fry and 90 percent fry to juvenile, approximately 750,000 triploid Spokane stock rainbow trout are collectively produced for transfer to Lake Roosevelt net pen rearing operations in the fall (Oct-Nov). All fish are marked with adipose fin clips prior to transfer to the net pens. The Lake Roosevelt Rainbow Trout Net Pen Project coordinates volunteers to feed the trout, maintain the net pens and release the fish after reservoir drawdown and when zooplankton biomass is adequate for forage (May-June).

The rainbow trout mitigation program has proven to be an extremely successful hatchery production program (Cichosz et al. 1999; Spotts et al. 2002), and rainbow trout are consistently the most targeted and harvested species in Lake Roosevelt (McLellan et al. 1999; Lee et al. 2003, 2006; Scofield et al. 2004, 2007; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Pavlik-Kunkel et al. 2008; Lee et al. 2010; Miller et al. 2011). Rainbow trout abundance from electrofishing surveys were fairly stable from 1993-1997 (5-9%) and showed a slight increase in 1998 (12%) (Peone et al. 1990; Griffith and Scholz 1991; Griffith et al. 1995; Thatcher et al. 1996; Cichosz et al. 1999; McLellan et al. 1999; Spotts et al. 2002). Rainbow trout annual harvest estimates for the same period ranged from 76,782 (95% CI ± 3,672; 1996) to 398,943 (95% CI ± 16,669; 1993) with the exception of a low annual harvest in 1997 of only 5,336 (95% CI±182), which was characterized as a high water year (deep drawdown) (Underwood and Shields 1996a, 1996b; Cichosz et al. 1999). The reservoir was drawn down to the minimum operating pool (1208 AMSL) with water retention times the lowest in the 50-year record (minimum 8 days; Cichosz et al. 1999, Fields et al. 2005). Reduced harvest estimates most likely occurred because of a combination of high entrainment due to decreased water retention time, reduced angling pressure due to drawdown, and lessened creel inquiry during periods of low fishing pressure (Cichosz et al. 1999). From 1999-2002, electrofishing surveys showed a marked increase in rainbow trout relative abundance averaging 21% (McLellan et al. 1999; Miller et al. 2011). From 2003-2008, relative abundance was variable, ranging from 14-26% (mean 17%) (Miller et al. 2011). Rainbow trout annual harvest estimates from 1999-2004 ranged from 92,090 (95% CI. ± 6,265; 2001) to 281,370 (95% CI. ± 16,729) (McLellan et al. 1999; Lee et al. 2003; Scofield et al. 2004; Fields et al. 2005; Pavlik-Kunkel et al. 2005). Due to the change in creel protocol in 2004, the estimated number of fish harvested decreased in subsequent years, but it is believed this is a more accurate representation of angler harvest (Miller et al. 2011). From 2005-2008, annual rainbow trout angler harvest rates ranged from 12,445 (95% CI. ± 11,581-13,592) fish harvested for boat anglers and 4,713 (95% CI. ±2,950-6,963) fish harvested for shore anglers to 19,105 (95% CI. ± 18,020- 20,080) fish harvested for boat anglers to 7,081 (95% CI. ± 5,173-9,306) fish harvested for shore anglers (Lee et al. 2006; Scofield et al. 2007; Pavlik-Kunkel et al. 2008; Lee et al. 2010; Miller et al. 2011). Goals for annual harvest for hatchery rainbow trout are 50,000 to 150,000 Spokane stock fish in the recreational and subsistence fishery. Harvest is adjustable based on a logistic regression model that estimates harvest percents based on water year, reservoir retention times and reservoir elevation when fish are released (McLellan et al. 2008).

The number of anglers visiting Lake Roosevelt has increased over the last several years and the rainbow trout production program has remained substantially reliable (Miller et al. 2011). The amount of anglers targeting rainbow trout was consistently high (40.7%-74.6%; mean 52.9%) over the past 10 years (1999-2008). When interviewed, most anglers targeted rainbow trout more often than any other species. The economic value of the Lake Roosevelt fishery has ranged from $1.3-20.7 million from 1990-2008 and has averaged $5.1 million over the last 10-year period (1999-2008; (McLellan et al. 1999; Lee et al. 2003, 2006; Scofield et al. 2004, 2007; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Pavlik- Kunkel et al. 2008; Lee et al. 2010; Miller et al. 2011). The hatchery rainbow trout program has been one of the most successful components, with a 10-year average of 116,278 fish harvested annually (McLellan et al. 1999; Lee et al. 2003, 2006; Scofield et al. 2004, 2007; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Pavlik-Kunkel et al. 2008; Lee et al. 2010; Miller et al. 2011).

Over the last 10 years (1999-2008), growth and condition of both hatchery and wild (non-marked) rainbow trout was evaluated. Based on condition factors (K_TL), both hatchery and wild rainbow trout perform well in the reservoir. K_TL’s are consistently above 1.0, averaging 1.15 for hatchery and 1.06 for wild rainbow trout. High K_TL’s are indicative of a high quality food source for zooplanktivorous fish like rainbow trout (Fields et al. 2005; Miller et al. 2011). Studies have consistently shown that food is not limiting to rainbow trout in Lake Roosevelt and that the population is below carrying capacity (Fields et al. 2005). Even though studies show high diet overlap and high K_TL’s, indicating that little competition exists between hatchery and wild rainbow trout (Fields et al. 2005), impacts from hatchery triploid rainbow trout on the food supply are unknown (Lee, unpublished). The difference in condition between hatchery and wild rainbow trout could be the result of a smaller sample size of wild fish or possibly a difference in age composition, sex, maturity, gonadal condition, morphometry or physiology, sampling time or location, etc. (Fields et al. 2005).

Annual rainbow trout tagging studies began in 1988 and have allowed managers to gather information on fishery recruitment, the timeliest release periods and locations, entrainment rates, movement and growth, stock performance, triploid versus diploid performance, and to acquire long-term data on tag returns during a variety of water years (high, low, average). Results from studies conducted from 1989-1994 show optimal release times of rainbow trout to avoid entrainment through Grand Coulee Dam. These studies indicated that fish released in late spring (May-June) rather than early spring (March-April) exhibit reduced entrainment (Peone et al. 1990; Griffith and Scholz 1991; Griffith and McDowell 1996; Voeller 1996; Cichosz et al. 1999). Griffith and McDowell (1996) found that water retention times less than 40 days during spring increased the number of tagged fish collected downstream of Grand Coulee Dam (24-50% entrainment), while water retention times greater than 40 days during spring exhibited rainbow trout entrainment estimates at less than 2%. Cichosz et al. (1999) suggested that water retention times, as well as annual reservoir drawdown and refill, could affect entrainment rates based on the timing of when fish are released into Lake Roosevelt. Cichosz et al. (1999) also discovered that entrainment may be reduced (i.e., increased recruitment) by approximately 12-20% with water retention times of 30-40 days, respectively and concurred with the release time of late May or early June when the reservoir reached 384 m above mean sea level. Griffith and McDowell (1996) and Griffith et al. (1995) linked water retention time and smoltification to increased entrainment of rainbow trout. Smoltification may increase the likelihood of rainbow trout entraining during the spring, while high entrainment rates were not associated with low water retention times in the summer, based on low numbers of tags collected below Grand Coulee Dam during summer (6%) when water retention times were low (34 days; Griffith et al. 1995; Griffith and McDowell 1996). Based on these findings, they also recommended the release of net pen rainbow trout in late May or early June annually.

From 1988-2008, tagging studies also focused on coastal rainbow trout entrainment through Grand Coulee Dam, as it is thought to be a major cause for losses of mitigated fish. In the last 10 years (1999-2008), entrainment has ranged from 0.41-23.06%, averaging 5.61%. Although a lack of concentrated effort below the dam has limited the ability to quantify entrainment, voluntary angler tag returns have verified that it is a problem (Miller et al. 2011). The percentage of tags collected below Grand Coulee Dam ranged between 0.41-32.3% of the 21-year record (1988-2008), excluding 1997 (Miller et al. 2011). Of the tagged fish released in 1997 (classified as a deep drawdown year), 96.7% were collected below Grand Coulee Dam most likely due to decreased water retention time, timing of release, water elevation at time of release, reduced angling pressure due to drawdown, and lessened creel inquiry during periods of low fishing pressure (Cichosz et al. 1999). Some studies have attempted to quantify entrainment through the use of hydroacoustics; however, it was difficult to determine species and/or origin (hatchery or wild) (Baldwin and Polacek 2002; LeCaire 2000). Authors of a recent study developed a model to predict how hydropower operations affected the coastal rainbow trout fishery in Lake Roosevelt based on the probability of annual tag returns (McLellan et al. 2008). They found that release location, water year (especially deep drawdown years), release elevation, and mean water retention time at 4 weeks post-release were the most significant variables to impact the fishery. The authors discovered that anglers were most likely to return tags 1.86 times after a shallow drawdown than after a deep drawdown, indicating that more fish may have been entrained, were lost to mortalities (i.e., predation, gas bubble trauma), or both, due to unstable water conditions after their release date. A benefit of this study is the reduction in cost to evaluate the success of the coastal rainbow trout fishery. By evaluating hydro-operations on the reservoir, the cost was much less than trying to quantify entrainment using more laborious methods. This study is significant, as it states that hydro-operations at Grand Coulee Dam negatively impact the coastal rainbow trout fishery in Lake Roosevelt.

Tagging studies were also used to examine movement and growth of rainbow trout within the reservoir. Studies have shown that rainbow trout tend to move downstream more frequently than upstream (Underwood and Shields 1996a, 1996b; Cichosz et al. 1997a; Miller et al. 2011). The downstream migration may have been due to a higher abundance of food items available in the lower sections of the reservoir or possibly to a smoltification process. Shields and Underwood (1996a, 1996b) also confirmed that growth of rainbow trout is greater in the lower than in the upper reservoir, which may be indicative of more suitable habitat available. Growth in the lower and mid reaches of the reservoir were attributed to higher zooplankton levels found in the lacustrine and transitional portions of the reservoir compared with the more riverine upper reservoir (Underwood and Shields 1996a, 1996b; Cichosz et al. 1997a; Fields et al. 2005, Miller et al. 2011).

From 1999-2008, tagging studies were used to determine stock performance of locally adapted stocks and triploided fish to address upriver, downriver and in-reservoir concerns regarding entrainment susceptibility, localized adaptations and genetic introgression. Studies employed release strategies using paired release groups of rainbow trout. Local (Kettle River tributaries) diploid redband rainbow trout (aka Phalon Lake) were paired with diploid coastal Spokane stock rainbow trout, triploid coastal Spokane stock rainbow trout, and triploid Troutlodge stock (through Columbia River Fish Farm). Items considered in stock performance were genetic introgression with remaining rainbow/redband stocks of Lake Roosevelt, impacts to downstream rainbow and steelhead (i.e., genetic, competition), in-lake feeding behavior, and successful recruitment to the fishery. The redband stock was the preferred stock, as it was believed to be more closely related to wild indigenous stocks of the upper Columbia River and would be utilized as a replacement for the non-indigenous Spokane stock in sensitive areas of the reservoir. Due to the limited availability of redband trout, the LRFEP also examined triploid Spokane stock as an alternative to address rising interest in protecting native stocks of rainbow and redband trout that remain in the reservoir, adjacent tributaries, and downriver from Grand Coulee and Chief Joseph dams.

The results of the comparative studies met with mixed results. The triploid Troutlodge stock rainbow trout performed well for only half of the years it was released into the reservoir (Miller et al. 2011), and was only utilized from 2000-2003 while the WDFW hatcheries began triploiding Spokane stock for release into Lake Roosevelt. The triploid Troutlodge stock was only used under extreme circumstances to backfill when Spokane stock rainbow trout were unavailable, if hydro-operations were particularly harsh due to entrainment, or if higher mortality occurred in the net pens. Data on redband trout performance compared to diploid and triploid Spokane stock rainbow trout also indicated varied success. Redband trout released in October of 2000 performed better than diploid and triploid rainbow trout; however, redband trout released in September 1999-2007 did not perform as well as diploid and triploid Spokane stock rainbow trout (Lee et al. 2006; Miller et al. 2011). Based on these findings, co-managers recommended increasing the number of redband trout released into Lake Roosevelt from 60,000 to 100,000 (based on availability). In 2007, following the assessment of performance of the October released redband trout, the LRFEP finalized tagging of Phalon Lake redband rainbow trout.

Concerns whether triploid Spokane stock would perform as well as the diploid Spokane stock in the fishery prompted paired release studies examining the performance of diploid and triploid Spokane stock rainbow trout in Lake Roosevelt. From 2000-2005, WDFW triploidy tests proved to be highly successful with 86-98% of fish successfully triploided (Morrison 2006).  Results from 2003-2006 tag recovery studies indicated that triploid Spokane stock performed similarly to diploid Spokane stock (Miller et al. 2011). In 2006, based on the positive return results and on a recommendation by the Independent Scientific Review Panel (ISRP), the LRFEP began triploiding all hatchery raised rainbow trout released into Lake Roosevelt, an increase from 50% diploid/50% triploid. In 2007, the annual production of 100% triploid trout increased from 500,000 to 750,000, as it was hypothesized that adjusting the percentage of triploid Spokane stock rainbow trout would not adversely affect the success of the rainbow trout artificial production program. Costs from increasing the rainbow trout program were fairly limited as the program would utilize kokanee net pens, as kokanee rearing in net pens did not meet expectations in Lake Roosevelt. There is insufficient data to determine whether any negative interactions exist between wild and hatchery rainbow trout. Suggestions of diet overlap, competition for mates, female gamete waste by wild rainbow trout, have been made and need further research in order to determine if any negative behaviors/interactions exist between wild and hatchery rainbow trout (Lee, unpublished). Triploidy rates from 2010 ranged from 98.2-99.5%, reiterating that the process continues to be successful.

At the same time all hatchery rainbow trout were being released as triploids, 100% of all hatchery rainbow trout released into Lake Roosevelt were being adipose fin clipped. The decision to mark and triploid all fish was two-fold; to limit genetic interference with spawning native rainbow and redband trout within reservoir tributaries and the upper Columbia River above Lake Roosevelt, as well as downriver steelhead populations; and to allow managers to easily identify between artificially produced and wild origin rainbow trout. The ability to readily identify hatchery fish allows fisheries managers to develop research/studies to define and more clearly understand hatchery versus wild rainbow trout interactions, temporal and spatial distribution of hatchery fish in the reservoir and tributaries, and proportional contribution of hatchery versus wild rainbow trout in the harvest. Marking and triploidy levels are assessed annually to determine the efficacy of each measure.

In the future, studies to minimize impacts of hatchery rainbow trout on wild redband trout will consist of PIT tagging 10% of the hatchery rainbow trout to determine if they co-exist on spawning grounds within redband natal streams. Future studies will also include PIT tagging 10% of the hatchery rainbow trout to estimate entrainment through Grand Coulee Dam. PIT tagged fish will be identified at downstream collection facilities along the Columbia River. Data will provide information about the impacts of hydro-operations on hatchery released rainbow trout, with the possibility of quantifying the number of fish lost to entrainment (Lee, unpublished).

The rainbow trout supplementation program has been exceptionally successful in Lake Roosevelt and the volunteer-based Lake Roosevelt Net Pen Program (BPA Project No.1995-009-00) is an excellent example of a low cost, effective program. Through annual reservoir-wide creel surveys and sampling, the LRFEP has been able to fulfill the objective of assessing management actions and performance of the Lake Roosevelt artificial production program. Managers have been able to determine hatchery rainbow trout health and growth, total harvest and catch rates, annual recruitment to the fishery, as well as angler preferences and the long-term economic contribution of the Lake Roosevelt fishery to the region. In the future, the co-managers for Lake Roosevelt plan to make the most of this program to maintain the fishery under increasing demands, as sustaining the thriving rainbow trout fishery remains of critical importance to both the co-managers and stakeholders of Lake Roosevelt.

Kettle River Redband Rainbow Trout

A population of native redband rainbow trout was known to exist within the Kettle River tributary, a tributary of the upper Columbia River near Kettle Falls, WA. The Kettle River was originally open for year-round angling; however, reports from anglers and data from a Washington Department of Fish and Wildlife (WDFW) boat electrofishing survey conducted in the late 1980’s (WDFW, unpublished data in McLellan and Vail 2005) suggested the recreational fishing opportunities were less than desired. A study in 1992 confirmed low densities of rainbow trout in both the upper and lower Kettle River (Jackson and Lovgren 1992). New regulations were then passed in order to protect trout and to allow them to spawn at least once (McLellan and Vail 2005). Due to the low abundance of redband rainbow trout in the Kettle River, a supplementation program was implemented to increase the local population. To increase fishing opportunities and genetic integrity, redband rainbow trout broodstock from Kettle River tributaries were transferred to Phalon Lake, a closed water, for rearing (Sherman Creek Hatchery memo 4/12/2000; McLellan and Vail 2005). From 1992-1994, the broodstock offspring were reared at the WDFW Colville Hatchery and released as fall fry in the upper Kettle River (WDFW, unpublished data in McLellan and Vail 2005). The fish were adipose fin clipped to identify them as hatchery plants. Due to fin regeneration, fin clips were difficult to identify, so fish were held until spring and released as yearlings for better identification purposes. From 1996-1998, 26,000 yearlings were released into the upper river. Through 1999, initial broodstock takes were approximately 100-150 fish ≥ 100 mm TL, which were relocated into Phalon Lake annually. Inconsistent numbers of broodstock consequently led to variable egg takes, and hence, annual protocol was changed. In the early 2000's, a minimum of 200 redband rainbow trout ≥ 100 mm TL were collected from a wider variety of local Kettle River tributaries in order to create a more "mixed stock" to be reared in Phalon Lake (Broodstock Replacement Plan, WDFW memo). From 2000-2003, releases took place in the lower Kettle River, with the exception of 2002 due to broodstock offspring losses in 2001. Stocking of the Kettle River ceased in 2003; however early sampling by WDFW in the upper river indicated an initial survival of at least one year after release with 40% of the sample being hatchery fish (WDFW, unpublished data in McLellan and Vail 2005).

Lake Roosevelt Redband Rainbow Trout

Redband rainbow trout within the upper Columbia River are a subspecies of rainbow trout that were once anadromous prior to the construction of Grand Coulee Dam in 1939 (Lee, unpublished). In 1998, the LRFEP, in an attempt to increase angler harvest of rainbow trout in the upper reaches of Lake Roosevelt and to introduce a more native stock to the reservoir, opted to raise and release redband rainbow trout in replacement of the traditionally stocked non-native coastal rainbow trout. Originally, non-native coastal rainbow trout were stocked into Lake Roosevelt to provide both angling opportunities and partial mitigation for hydro-operations. Over time, co-managers were concerned about negative interactions with native redband rainbow trout and moved to stocking a percentage of redbands to test the feasibility of stocking a more native species.

Redband rainbow trout were collected locally from Kettle River tributaries (an upper Columbia River tributary) as yearlings and released into Phalon Lake, a closed water, for broodstock rearing. Juveniles from the Phalon Lake broodstock were then reared at Colville Fish Hatchery and transferred to Sherman Creek Hatchery until release into net pens as yearlings in the fall. Final release of net penned fish into Lake Roosevelt occurred the following spring, while some fish were held until fall and released directly into the Colville River at Meyer's Falls, into Sheep Creek (an upper river tributary of Lake Roosevelt), or directly into Lake Roosevelt. Others were held over until the subsequent year as part of an ongoing evaluation of release strategies comparing local redband rainbow trout to diploid and triploid coastal rainbow trout in the net pen program (Combs 2011).

To test the success of the redbands, a tagging study was developed in 1999 that compared redband rainbow trout with both diploid and triploid coastal rainbow trout. From 2000-2006, various rearing and release methods were tried and evaluated. Biologists ultimately determined that holding fish for 1½ years was optimal to achieve a large enough size (similar to coastal rainbow trout) for survival. However, returns were typically 2-3 times less than diploid and/or triploid coastal rainbow trout returns (ranging from <0.01-0.78 for redbands, 0.94-2.58 for Spokane diploid stock, and 0.30-1.39 for Spokane triploid stock for Kettle Falls releases). In 2007, returns were closer than in previous years (0.73 redbands, 0.93 triploids), but in 2008, redband rainbow trout were recruited in slightly higher numbers than coastal triploids (0.55 redbands; 0.30 triploids); however, lower numbers of triploids were stocked in comparison to previous years. In brief, it was determined that redband rainbow trout released in the spring, rather than in fall, had comparable results to triploid rainbow trout and should continue to be released at that time. In 2009, despite the possible increase in recruitment, and due to historically limited availability of redband rainbow trout to the program, redband rainbow trout stocking was discontinued so that only triploid rainbow trout would be stocked into Lake Roosevelt.

Following guidelines based on the Northwest Power and Conservation Council to work towards restoring native resident fish species, the co-managers are currently working towards maintaining a native redband trout population in the upper sections of the reservoir. This has become more critical as researchers in British Columbia, Canada expressed increasing concern regarding non-native, coastal rainbow trout invading their native trout streams in the upper Columbia River. Redband rainbow trout still naturally occur and reproduce in the upper Columbia River, and through DNA testing, it was determined that some stocks of redband trout were pure, native stocks (Taylor 2002; Powell and Faler, 2002, Small and Dean 2006, 2007, Small et al. 2007). The current goal is “to conserve, enhance and restore (native fishes and) redband trout populations in Lake Roosevelt and its tributaries, and where appropriate, provide opportunities for subsistence harvest by Native American Tribes and recreational harvest” (Lee, unpublished). To meet this goal, several objectives are in place, which include 1) management of stocks of wild redband trout in Lake Roosevelt and the upper Columbia River to allow for conservation, enhancement, and restoration, 2) minimize impacts of the Lake Roosevelt hatchery rainbow trout program on wild redband trout, and 3) assess entrainment of hatchery triploid rainbow trout and wild redband rainbow trout from Lake Roosevelt to determine fish losses and the effect on native fisheries downstream of Grand Coulee Dam. This work is crucial as identification of wild redband rainbow trout stock structure, life history, interactions with hatchery rainbow trout, habitat use, and migration patterns have not been completed within Lake Roosevelt/upper Columbia River and its tributaries (Lee, unpublished). The most current data exists only within the Sanpoil River drainage (Sears 2008). Future studies will include estimates of recruitment, age at recruitment, recreational harvest, interval and instantaneous exploitation rates, growth parameters, age at maturity, spawner escapement, mortality, and spatial and temporal characteristics of harvest data for select populations of redband trout in Lake Roosevelt and the Upper Columbia River. A comprehensive stock assessment will provide managers with the tools to describe population dynamics and the impacts of various harvest alternatives for redband trout, allowing informed management decisions that help meet co-manager goals to conserve, enhance and restore redband trout populations in Lake Roosevelt and its tributaries, while providing for subsistence and recreational harvest where appropriate.

Walleye

Management objectives for walleye are to maintain a harvestable population while minimizing impacts on kokanee, rainbow trout, white sturgeon (Acipenser transmontanus), and other native fish. An initial study was conducted in 1998 to estimate the walleye population in Lake Roosevelt (McLellan et al. 1998, McLellan et al. 1999, McLellan et al. 2002). The number of walleye was estimated to be approximately 122,109, but the authors acknowledged the limitations of the estimate. Statistically strong population estimates on the reservoir are difficult due to the large area that must be sampled and from which the fish must be recaptured (McLellan et al. 2002). Despite the limitation, the estimate was utilized in the Wisconsin bioenergetics model, as it was the best estimate available. The LRFEP does not plan to complete a more recent reservoir-wide population estimate unless it is identified as a need for future bioenergetics modeling.

To determine movement and growth of walleye, tagging studies took place from 1997-2000, where over 10,000 walleye were tagged and released (McLellan et al. 2002). Tagging information was also used to determine abundance, age, condition, mortality, spawning frequency, and post-spawning migrations. It was determined in the early 1980’s and from 1997-1999, that the primary walleye spawning grounds were located in a 10 km free-flowing stretch of the upper Spokane River arm 40 river kilometers (rkm) from the Lake Roosevelt confluence below Little Falls Dam (Beckman et al. 1985; McLellan et al. 2002). Spawning occurs from early April to early May, with peak activity occurring the last two weeks of April (Beckman et al. 1985; McLellan et al. 2002). Post-spawn, some walleye remain in the Spokane River arm, while others disperse throughout the reservoir, traveling upstream as far north as Canada, or downstream (Hall et al. 1985; McLellan et al. 2002), with some passing through Grand Coulee Dam (Beckman et al. 1985). Tagging studies have shown evidence that most walleye in Lake Roosevelt establish summer home ranges and move limited distances (20-25 rkm) post-spawn (Beckman et al. 1985; McLellan et al. 2002). Tagging efforts were ceased in 2001 due to financial and time constraints, along with priorities of other research objectives, so a decision was made to revisit walleye tagging in Lake Roosevelt at a later date, or when the need arose.

From 2000-2007, walleye relative abundance has been fairly consistent, ranging from 21-32% (Lee et al. 2003, 2006; Scofield et al. 2004, 2007; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Pavlik-Kunkel et al. 2008; Lee et al. 2010). In 2008, walleye relative abundance dropped considerably to 2%; however, the survey design was altered to reduce sampling bias to catch young-of-year species that were virtually nonexistent in previous surveys (Miller et al. 2011). This in turn changed the analysis and results and should be examined in the future to reduce biases (i.e., young-of-year should be analyzed separately from adults). CPUE of walleye from 2000-2008 ranged from 3.12-9.39 fish/hour (mean 6.92 fish/hour) (Miller et al. 2011). In the Sanpoil River from 2009-2010, walleye relative abundance was 74.2% and CPUE was 21.9 fish/hour (Stroud et al. 2011a, in press, 2011b, in press). Walleye are one of the most abundant piscivores and highly effective opportunistic feeders, so their impact on prey species is likely one of the greatest (Carlander 1997; McLellan et al. 1999; Lee et al. 2003, 2006; Scofield et al. 2004, 2007; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Pavlik-Kunkel et al. 2008; Lee et al. 2010).

It has been established that walleye impact the artificial production program based on predation estimates of salmonids (31- 39% annual kokanee loss; 6-12% annual rainbow trout loss) (Baldwin and Polacek 2002; Baldwin et al. 2003). Also, since the establishment of walleye, the relative abundance of cyprinid fishes has steadily decreased from 83% in 1949; 72% in 1966; 22% in 1973; 23% from 1980-1983; 4.7-7.1% from 1988-1991 (Thatcher et al. 1996). From 2000-2008, native cyprinids were only <1-3% of the total relative abundance. Additionally, surveys conducted by the Lake Roosevelt Sturgeon Recovery Program (LRSRP) and associated work in the upper Columbia River above the international border has indicated that the white sturgeon (Acipenser transmontanus) population is declining as well. Recent assessments have shown that wild white sturgeon juveniles were only rarely encountered in the reservoir (UCWSRI 2002). The LRSRP recognizes that many variables may be limiting the recruitment of sturgeon past the larval stage, including contaminants (i.e., slag), predation (possibly from walleye), food availability, low turbidity, and water discharge rates (Howell and McLellan, unpublished data). Working in conjunction with the LRSRP, co-managers will consider future studies designed to determine predation effects on these sensitive species.

In 2009-2010, the Colville Confederated Tribe (CCT) and Eastern Washington University (EWU) led a study to investigate predation by non-native predators (walleye and smallmouth bass) on stocked kokanee salmon and wild rainbow trout to determine a possible reason for recruitment failures on the Sanpoil River, a lower reservoir tributary of Lake Roosevelt (Stroud et al. 2011a, in press, 2011b, in press). In an attempt to enhance survival and increase returns of adfluvial rainbow trout (which had steadily declined over the years) and stocked kokanee salmon, the CCT restored spawning and rearing habitat in the Sanpoil River (Sears 2008). However, returns continued to decrease, suggesting that the reduction may be due to predation by walleye and smallmouth bass on juvenile salmonids migrating out of the Sanpoil River. Walleye and smallmouth bass populations were estimated in the Sanpoil River to determine consumption and predation of salmonids within that system. In 2009, the walleye population <100 mm TL was estimated to be 25,068 (95% CI 13,793–46,509) (Stroud et al. 2011a, in press, 2011b, in press), and the walleye superpopulation estimate for fish >170 mm TL was 14,952 (95% CI 3,147 - 26,757) from 24 March to 14 July, 2010 (Stroud et al. 2011a, in press, 2011b, in press). Based on population, consumption, and predation estimates, they determined that walleye consumed 22.8% (95% CI= 4.8–40.8%) of the rainbow trout yearlings, 32.8% (6.9–58.7%) of the rainbow trout 2- and 3–year olds, 42.5% (8.9–76.1%) of the kokanee fry and 32.9% (6.9–58.9%) of the kokanee yearlings (Stroud et al. 2011a, in press, 2011b in press). Due to the high level of predation, the CCT provided options to reduce walleye and smallmouth bass predation in the Sanpoil River, including (1) trapping salmonids for relocation away from predators or swamping predators all at once; (2) physically reducing the total number of predators within the system; and (3) releasing larger kokanee yearlings, rather than kokanee fry (Stroud et al. 2011a, in press, 2011b, in press). The CCT is currently undergoing a process of predator removal of walleye and smallmouth bass in the Sanpoil River.

Diet analysis has consistently shown the preference for salmonids and some native species in walleye diets. For littorally caught walleye over the last 16 years (1993-2008), salmonids and cottids were consistently found as the primary fish prey items based on the relative importance index (Ria). Salmonids were found to be the principal prey species in six of those years [(1996, 1999, 2000, 2001, 2004, 2006) (Ria ranging from 14-28)] ( Cichosz et al. 1997; McLellan et al. 1999; Lee et al. 2003; Scofield et al. 2004; Lee et al. 2006; Pavlik-Kunkel et al. 2008) with cottids as the primary prey species in six alternate years [(1993, 1998, 2002, 2003, 2005, 2008) (Ria ranging from 9-48)] (Underwood and Shields 1996; Spotts et al. 2002; Fields et al. 2005; Pavlik-Kunkel et al. 2005; Scofield et al. 2007; Miller et al. 2011). Percidae was the top prey species in 3 years [(1994, 1995, 2007); (Ria ranging from 11-17)], while cyprinids were the top prey species one year (1997; Ria=14) (McLellan et al.2003; Miller et al. 2011). WDFW limnetic surveys from 2000-2008 show cottids and salmonids as the most consumed fish prey items in August (29-83%, 0-15%, respectively) and October (20-50%, 2-17%, respectively) (Baldwin et al. 2005, 2006; Baldwin and Woller 2006a, 2006b, 2006c, 2007; Polacek and Woller 2010; Polacek and Woller 2011).

Over a ten year period (1999-2008), walleye growth and condition has remained stable in Lake Roosevelt, with condition factors averaging 0.83 (ranging from 0.78 in 2008 to 0.89 in 2002) (McLellan et al. 1999; Miller et al. 2011). Lower condition in 2008 may have been due to changes in sampling protocol, gear bias, reduced sample size from previous years, or lack of suitable forage available in the reservoir at the time of sampling. FWIN sampling (see below) has showed similar results with condition factors ranging between 0.81-0.86 (7-year avg.: 0.84) (Fields et al. 2005; Pavlik-Kunkel et al. 2005; Lee et al. 2006, 2010; Scofield et al. 2007; Pavlik-Kunkel et al. 2008; Miller et al. 2011).

To further investigations of walleye, beginning in 2002, WDFW adopted a new statewide protocol for sampling walleye populations, known as Fall Walleye Index Netting (FWIN; Morgan 2000). The protocol was borrowed from researchers from Ontario, Canada, who developed the methods to help manage walleye, their most important sport fishery. This was not a new objective for the LRFEP, but an improved methodology to try to assess the status of walleye in Lake Roosevelt due to increasing concerns as to whether walleye, the apex predator in the reservoir (Cichosz et al. 1999; Baldwin et al. 2003), is in balance with the ecosystem. From 2002-2008, the results from FWIN surveys indicated that walleye were the most abundant predator ranging from 25-41% of the species collected (Fields et al. 2005; Pavlik-Kunkel et al. 2005; Lee et al. 2006; Scofield et al. 2007; Pavlik-Kunkel; et al. 2008; Lee et al. 2010; Miller et al. 2011); however, the condition of walleye is low when compared to other aquatic systems in Washington (WDFW 2008). FWIN data has allowed for more aggressive management of walleye by focusing on making changes to fishing regulations to increase harvest of smaller fish, while protecting larger fish to allow development of a trophy fishery. The LRFEP believes the annual FWIN surveys are successful, and in order to continue to better manage the walleye population, FWIN has become a necessary, ongoing, annual project. An original recommended minimum of five years of data collection was established to determine trend analysis and to evaluate the regulation change; however, a more in depth analysis of the FWIN data will be completed in the future. Once the data has been analyzed, the recommendation will be revisited to determine if additional years may be necessary or consider alternating assessments on a two year basis.

Managers continue to attempt to address predation effects on kokanee, rainbow trout, white sturgeon, and other native fish in Lake Roosevelt through proposed harvest regulations on walleye (and smallmouth bass). Since the revised creel survey was initiated in 2004, 40-64% of angler caught walleye have been harvested, making walleye the most harvested species in the last 4 out of 5 years (2004-2008) (Lee et al. 2006; Scofield et al. 2007; Pavlik-Kunkel et al. 2008; Lee et al. 2010; Miller et al. 2011). WDFW estimated that between 75,000 and 100,000 walleye need to be harvested annually, and in the past few years (2007-2009), only about half (40,000) were harvested (WDFW 2011). In order to reach this goal, the amount of walleye harvested needs to increase to more than double.

The LRFEP co-managers worked with WDFW to establish new regulations for walleye on Lake Roosevelt. Beginning in 2007, anglers were allowed a daily harvest of eight walleye with no more than one over 22 inches with no minimum size limit. This was intended to reduce the number of small walleye and balance the population (WDFW 2011) and at the same time allow for a more aggressive harvest than in previous years, while still protecting larger fish. A current proposal is in place to increase the daily limit of walleye in Lake Roosevelt from 8 to 16 fish per day with only 1 over 22 inches. The proposal is intended to improve the walleye fishery, increase walleye condition, and limit walleye predation on salmonid species (WDFW 2011). Another new proposal is also being put forth to allow anglers a two pole endorsement in the lower Spokane River (a mid-reservoir tributary of Lake Roosevelt; from the SR 25 bridge to 400 feet below Little Falls Dam). The walleye regulations listed above will also be allowed on the specific section of the Spokane River. The two-pole endorsement will allow anglers increased recreational opportunity to harvest mitigated kokanee as well as increased opportunity to harvest the proposed increased daily limits of walleye.

Currently, the Colville Confederated Tribe's regulations match WDFW daily catch limits (8 fish for 2011 and plan to change to 16 in 2012 with WDFW revision), and no size limit for walleye or smallmouth bass within the Sanpoil River during the fishing season from April 15 to December 31. Furthermore, all captured walleye and bass must be kept and not released under a tribal permit. Stroud et al. (2011a, in press, 2011b, in press) suggested that state regulations be changed to match tribal regulations in order to remove a higher number of predators. They also recommended a revision to the current regulation: to reopen the Spokane River arm for walleye fishing year-round, as it is currently closed to angling during walleye spawning. They conclude that this would be the best option to restore native fisheries within the region (Stroud et al. 2011a, in press, 2011b, in press). In order to create a more balanced ecosystem, where native fish prosper, non-native predators should be kept at lower densities (Stroud et al. 2011a, in press, 2011b, in press).

Smallmouth Bass

Smallmouth bass have been identified as one of the four main sport fish species in Lake Roosevelt (along with rainbow trout, kokanee salmon, and walleye), and over the last three decades have become one of the top, non-native piscivores in the reservoir (Lee et al. 2010; Stroud et al. 2011a, in press, 2011b, in press).  From the mid-1940’s to about 1980, the relative abundance of smallmouth bass in Lake Roosevelt was minimal (0.00-<.05%), yet from the mid-1980’s a dramatic increase began to occur, and by the current decade, relative abundance had increased (Stroud et al. 2011a, in press, 2011b, in press) to a high of 36% (Miller et al. 2011).  Sanderson et al. (2009) stated that the percentage of smallmouth bass had increased in many other Pacific Northwest systems, greatly expanding their range, and unfortunately the impacts of their establishment may not become apparent for decades, which appeared to be what happened in Lake Roosevelt.

Annual surveys have shown a dramatic change in smallmouth bass relative abundance and harvest making them one of the most abundant species in the reservoir (Scofield et al. 2007), and over the last decade have been ranked among the top four species collected in both electrofishing and gill net surveys (Lee et al. 2010).  During 1989-1990, relative abundance was low (1 and 3 %, respectfully), and only 1-2% of anglers targeted smallmouth bass.  Estimated harvest was also low with less than 1% of the total fish harvested (1,538±578) in 1989.  However, in 1990, 83% of the smallmouth bass caught were harvested by anglers, and by 1991, smallmouth bass relative abundance increased considerably to 15%, the third highest species collected that year.  From 1992-1995, relative abundance had increased from previous years, but remained fairly stable, ranging from 7-10%, and in 1995, due to their consistent appearance in annual surveys, co-managers determined that monitoring smallmouth bass impact on kokanee predation was a priority (Underwood and Sheilds 1996).  Relative abundance from 1996-1999 was similar to previous years (6-9%); however, harvest estimates dropped much lower than in past years (<1%-13%).  Cichosz et al. (1997) noted that harvest rates had decreased greatly (83% in 1990, 53% in 1993, 11% in 1994, 13% in 1995 to 0.7% in 1996) and suggested that consistent catch rates and decreased harvest might be due to the angler’s opinions that the smallmouth bass fishery should become a “catch and release” fishery.  They also insinuated that smallmouth bass may be viewed as incidental catch when targeting other species because the percentage of anglers targeting smallmouth bass was low in earlier years (1% in 1989; 2% in 1990; <1% in 1992).  In 1997, 90% of the smallmouth bass caught were harvested, which may imply that more anglers may have specifically targeted smallmouth bass in 1997 relative to previous years (Cichosz et al. 1999).  However, 1997 was characterized as a high water/deep drawdown year, and this may have affected the lower number of fish caught, so that anglers chose to harvest rather than release fish.  Relative abundance from 2000-2003 ranged between 7% and 14%, and while harvest rates were 12-16%, only about 2% of anglers were targeting smallmouth bass.  Catch and harvest may have been underestimated at this time due to a low number of angler interviews conducted in the lower reservoir, as not all anglers with completed trips were interviewed (Fields et al. 2005).  Smallmouth bass were highly targeted in the lower reservoir, and the low number of interviews was most likely not due to a lack of anglers (Fields et al. 2005).  New creel protocol began in 2004, which corrected this issue in order to better estimate the number of fish harvested in all sections of Lake Roosevelt.  From 2002-2004, angler target percentages rose slightly, (4-7%) and continued to rise in subsequent years (11-19% from 2005-2008), while from 2000-2005, harvest estimates leveled out between 12-18%.  Total annual angler pressure rose at this time, and was most likely due to anglers targeting smallmouth bass (and rainbow trout) in the lower reservoir (Lee et al. 2006).  In 2004-2005, relative abundance increased dramatically to 23% and 27%, respectively.  Angler success may have been related to increased fish density in the lake resulting from favorable hydro-operations (i.e., above average water retention time and minimal spring drawdown) in recent years (Lee et al. 2006).  Relative abundance decreased in 2006 (14%) and 2007 (12%), which may have been due to changes in release strategies of hatchery origin salmonids (McLellan et al. 1999, Fields et al. 2005).  In 2008, relative abundance rose considerably to 36%, which was most likely due to the change in sampling design to include beach seining, as a majority of smallmouth bass captured were young-of-year (Miller et al. 2011).  This drastically changed the perspective of the fish assemblage, and future analyses would benefit from the separation of adults and juveniles.  From 2000-2008, CPUE for smallmouth bass ranged from 2.02-12.90 fish/hour (mean 5.04 fish/hour) (Miller et al. 2011). Centrarchidae (mostly smallmouth bass) have also contributed considerably to the relative abundance of species during FWIN sampling, ranging from 10-23 % from 2002-2008 (Miller et al. 2011).

Smallmouth bass have been extremely successful in Lake Roosevelt, and it is evident when analyzing the annual condition factor (K_TL) of the surveyed population.  K_TL’s greater than 1.00 are associated with greater robustness of the population by determining how fish add weight in relation to their length and are associated with the well-being, age, sex, and maturity of each fish (Williams 2000).  From 1997-2008, annual K_TL’s ranged from 1.34 to 1.85 (mean 1.46), and during annual Fall Walleye Index Netting surveys (2003-2008), K_TL’s ranged from 1.40-1.52 (mean 1.46).  A slight decrease occurred in K_TL’s in 2003 (1.35) and 2004 (1.36), which was thought to be associated with changes in kokanee release strategies that were implemented in 2003.  Historically, kokanee were released into high predator density areas (Little Falls Dam and Sherman Creek), and then as strategies changed, were released into deep, pelagic areas of the reservoir (Fort Spokane, Gifford, and/or Seven Bays), which provided refuge from predators (McLellan et al. 1999, Fields et al. 2005).  In 2004, managers additionally began stocking kokanee salmon as fry in upper reservoir tributaries also in an attempt to reduce predation.  These changes to the kokanee salmon stocking regime may have limited smallmouth bass (and walleye) predation on a food source they were both dependent upon, and may have resulted in the reduced K_TL (Scofield et al. 2007).  From 2005-2008, an increase in both smallmouth and walleye K_TL occurred, which happened to coincide with a strong year class (2005) of yellow perch (Perca flavescens) and an increase in the relative importance (Ria) of yellow perch in the diets of both species.  Yellow perch relative abundance also increased three fold from 2003 to 2004.  This may have been an instance where the top predators changed the composition of the fish assemblage (Scofield et al. 2007).

Competition between smallmouth bass and walleye has been documented repeatedly within Lake Roosevelt.  Inter- and intra-specific competition for food resources between and among smallmouth bass and walleye could negatively affect the abundance and quality of these piscivores as forage becomes limited (Miller et al. 2011).  There is also potential for competition with other species as smallmouth bass alter feeding habits as they grow, going from insects to zooplankton to crayfish and fish (Olsen and Young, 2003, Carey et al. 2011).  Chinook salmon and juvenile smallmouth bass in the Willamette River were found to have diet similarities, which might eventually lead to limited resources creating a competitive environment (Carey et al. 2011).  In Lake Roosevelt, diet overlap with smallmouth bass has been documented in other species, including rainbow trout (hatchery and wild), eastern brook trout (Salvelinus fontinalis), lake whitefish (Coregonus clupeaformis), northern pikeminnow (Ptychocheilus oregonensis), yellow perch, and cottidae species (McLellan et al. 1999; Pavlik-Kunkel et al. 2005; Lee et al. 2006).

Predation is the most quantifiable impact of nonindigenous fishes on native species (Sanderson et al. 2009).  Predation on salmonids by smallmouth bass in the Pacific Northwest (Columbia and Snake River systems) was quantified by Sanderson et al. (2009), reiterating the fact that nonindigenous species are detrimental to native trout within these main watersheds.  Smallmouth bass are aggressive, opportunistic feeders, have a large prey base available, and will consume most any prey smaller than the size of their gape (Sanderson et al. 2009); however, fish make up a large proportion of the diets in Lake Roosevelt (Ria ranged from 24.81-51.2 from 2000-2008 [Miller et al. 2011]).  Because of the high incidence of fish in their diets, beginning in 2002, co-managers recommended that predation by smallmouth, as well as walleye and burbot, be assessed as a limiting factor to salmonids produced by the artificial production program.  Smallmouth bass prey mostly on cottids and in some years on perch, yet for years salmonids have consistently been found in their diets.  Between 2000 and 2006, the relative importance of salmonids ranged from 0.6-6.1, yet in 2007-2008, salmonids were not observed in smallmouth bass diets.  Changes in release strategies of hatchery origin salmonids have potentially reduced piscivory by smallmouth bass causing a shift towards other forage species, although recent diet analyses have not focused on location and time of release impacts (Miller et al. 2011).  In 2008, besides the lack of salmonids in their diets, a reduction in cottids and an increase in catastomids and percids occurred.  Substantial predation is likely still occurring near release locations in other parts of the reservoir and is possibly limiting survival and performance of hatchery origin salmonids (Miller et al. 2011).  However, little is known about the extent of smallmouth bass predation on hatchery salmonids shortly after release into Lake Roosevelt.  Future studies should determine the extent of smallmouth bass predation on salmonids with respect to location and timing of release, as well as the impact on native fishes within the reservoir.

A recent study (2009-2010) by the Colville Confederated Tribe and Eastern Washington University determined that a high percentage of salmonids were being consumed by both smallmouth bass and walleye, non-native predators, in the Sanpoil River (a tributary of Lake Roosevelt).  High abundance [walleye (74.2%); smallmouth bass (84.4%)] and predation by walleye and smallmouth bass was believed to be the cause of recruitment failures within this system (Stroud et al. 2011a, in press, 2011b, in press), despite restored spawning and rearing habitat in the Sanpoil River (Sears 2008). Smallmouth bass populations were estimated in the Sanpoil River to determine consumption and predation of salmonids within that system. In 2009, the smallmouth bass population over 100 mm TL was estimated to be 37,634 (33,429 – 42,449), and the superpopulation estimate for smallmouth bass >170 mm TL in 2010 was 32,780 (95% CI 25,669 – 39,892) (Stroud et al. 2011a, in press, 2011b, in press).

Based on population estimates and consumption based on age-specific bioenergetic modeling, the CCT and EWU determined that in almost a four month period (March-July 2010), smallmouth bass consumed an estimated 61% (95% CI= 48–75%) of the kokanee fry, 5% (4–6%) of the kokanee yearlings, 17% (14–21%) of the rainbow trout yearlings, and 2% (1–2%) of the rainbow trout 2- and 3-year olds migrating out of the Sanpoil River into Lake Roosevelt (Stroud et al. 2011a, in press, 2011b, in press). Both walleye (see Walleye section above) and smallmouth bass together consumed 103% (95% CI = 57–150%) of the total kokanee fry release (612,610 of 589,580 released), 38% (10–65%) of the kokanee yearling release (3,836 of 10,080 released), 40% (18 – 62%) of the rainbow yearling population estimate (5,874 of 14,578 estimated), and 34% (8–60%) of the rainbow trout 2- and 3-year old population estimate (8,153 of 23,738 estimated) (Stroud et al. 2011a, in press, 2011b, in press). Fritts and Pearsons (2004) found that smallmouth bass can consume ≥35% of outmigrating juvenile salmon.  To combat these crucial predation issues, authors suggested three main ways to effectively increase kokanee and rainbow trout survival within the Sanpoil River.  They included: (1) trapping salmonids to either relocate them away from predators or swamp the predators all at once with large numbers of fish released; (2) predator removal by (a) offering anglers a monetary reward, (b) hiring members of the Colville Confederated Tribe to catch and harvest predators, or (c) manually removing predators using electrofishing, gill netting and/or angling; and (3) release larger-sized kokanee into the river, as they showed higher survival than fry, plus they successfully returned to the Sanpoil River in the fall [n = 6.2 % of stocked kokanee yearlings returned to spawn (fall 2010)].

The CCT is currently undergoing a process of predator removal of walleye and smallmouth bass in the Sanpoil River.  Harvest regulations have been made more aggressive so that CCT tribal permits have no size or catch limit and all catches of smallmouth bass (and walleye) must be kept. Regulations on Lake Roosevelt (non-tribal waters) are not as aggressive. Currently, anglers must follow the statewide freshwater species rule for smallmouth bass, which states that there is no minimum size, only one over 14” retained, with a daily limit of 10.  This rule was designed to promote aggressive harvest of the most abundant sized bass, while protecting the harvest of trophy fish; however, this may not be substantial enough to control the smallmouth bass population in Lake Roosevelt. Bag limits may need to be increased, catch/harvest limits may need to be removed, or mandatory eradication policies may have to be introduced (Carey et al. 2011).