experimental removal of lake trout in swan lake, mt:...

Experimental Removal of Lake Trout in Swan Lake, MT: 2013 Annual Report

Prepared for the Swan Valley Bull Trout Working Group

By:

Leo Rosenthal, Fisheries Biologist

Montana Fish, Wildlife & Parks

And

Wade Fredenberg, Fisheries Biologist U.S. Fish and Wildlife Service

April, 2014

2

Background The Swan Valley has historically been home to a stable bull trout population. However, in 1998 anglers began to occasionally catch adult sized (20-30 inch) lake trout from Swan Lake and the Swan River. This caused alarm because lake trout are not native and are notorious for rapidly expanding and dominating fish communities in lakes with Mysis shrimp, particularly at the expense of bull trout and kokanee salmon (Martinez et al. 2009). In 2003, the level of concern was compounded when biologists gillnetted juvenile lake trout from Swan Lake during standard low-intensity sampling efforts, indicating that wild reproduction was occurring. Since 2003, lake trout catch by anglers as well as during Montana Fish, Wildlife and Parks (FWP) biological sampling continued to increase, indicating that the population was expanding. Research efforts from 2006-2008 focused on lake trout population demographics, and exploring potential techniques to reduce lake trout numbers while minimizing bull trout bycatch. Based on case histories from nearby waters, long-term management alternatives for this increasing lake trout population are necessary in order to maintain the popular bull trout and kokanee fisheries. In 2009 FWP released an EA for a three-year experimental removal of lake trout in Swan Lake. This removal experiment was a feasibility study to determine the effectiveness of using targeted gillnetting as a technique to reduce the number of lake trout, and thus minimize threats kokanee and bull trout. From 2009-2011 over 20,000 lake trout were removed from Swan Lake. Modeled total annual mortality rates for lake trout vulnerable to the nets used in the project were higher than literature suggests are sustainable. While much has been learned with regard to our ability to affect lake trout cohort strength from one year to the next, the overall effect this level of removal has on the lake trout population and other fish populations remains unknown. Therefore, in May 2012 FWP released another EA for a five-year extension of the lake trout removal project to further evaluate the effectiveness of the current lake trout suppression effort. Measurable goals and specific success criteria outlined in the original 2009 EA are being used to evaluate the feasibility and effectiveness of alternatives to control expansion of the lake trout population. Based on the results of this assessment and other relevant considerations, FWP, with recommendations from the Swan Valley Bull Trout Working Group (SVBTWG), will consider whether these actions are appropriate or if other changes are warranted in fisheries management of Swan Lake and the lake trout population. Previous annual reports can be found at: http://montanatu.org/resources/swan-valley-bull-trout/ Methods The five-year extension of the lake trout suppression project closely follows the efforts conducted from 2009-2011. This replication of effort allows researchers to maximize lake trout removal in Swan Lake, while providing consistent data for year-to-year comparisons and analysis to evaluate the efficacy of gill net removal of lake trout. Consistent with the work from 2009-2011, the project continues to be comprised of two

3

distinct netting events. The first event (Juvenile Netting) is aimed at removing juvenile and subadult lake trout throughout the two deep (>60’) basins of Swan Lake. This removal is carried out using small-mesh (1.5 – 2.75 inch stretch) gill nets, set by professional fisheries contractors over a three-week period in late August. This netting is conducted during a time in which most adult bull trout are upstream in the Swan River drainage in preparation for fall spawning and also occurs during the period in which Swan Lake is thermally stratified. Only habitat below the thermocline (>60’) is sampled, in order to reduce incidental bycatch of bull trout and other fish species which occupy shallower depths. The second netting event (Spawner Netting) is directed at removal of adult lake trout during spawning and thus is targeted to directly affect further recruitment. This portion of the project is carried out largely by SVBTWG members and takes place during the month of October. Large-mesh gill nets (4.5 – 5 inch stretch) are set during the nighttime and early morning hours, along spawning areas identified by telemetry work conducted from 2007-2009 (Cox 2010). Results

Juvenile Netting Netting for juvenile lake trout has been contracted with Hickey Brothers Fisheries of Baileys Harbor, Wisconsin since 2009. Each year the boat is cleaned and disinfected following a Hazard Analysis and Critical Control Point Plan (HACCP) to minimize the risk of spreading aquatic invasive species. The boat is inspected annually by FWP personnel prior to entering Swan Lake to ensure proper cleaning procedures have been followed. Netting took place from August 11-30, 2013 which was similar to work conducted in 2012, but a week earlier than 2009-2011. This change in timing simplified logistics for the contractor and still provided lake conditions similar to previous years. Additionally, shifting the Juvenile Netting a week earlier ensured that post-spawning bull trout would not have returned to Swan Lake during netting operations. The amount of netting effort and the locations of nets set for juvenile lake trout has been consistent since 2009. The contract with the Hickey Brothers provides 30 lifts, with one lift being described as an event in which nets are set and retrieved. While the number of lifts has not changed since 2009, the number of net panels set has varied slightly since the beginning, as more panels of small mesh net were set to increase the catch of juvenile lake trout (Table 1). Although the number of net panels has varied slightly since inception of the project, the locations of the nets have remained constant (Figure 1).

4

Table 1: Dates and numbers of nets set for Juvenile Netting 2009-2013.

Year Netting Dates # Lifts # 900’ Net Panels 2009 Aug 24-Sept 11 30 248 2010 Aug 23-Sept 10 30 311 2011 Aug 22-Sept 9 30 399 2012 Aug 13-Aug 31 30 382 2013 Aug 11-Aug 30 30 347

Figure 1: Netting locations and numbers of net panels for Juvenile Netting in Swan Lake 2009-2013. A total of 7,051 lake trout from 5”-37” were removed during the 2013 Juvenile Netting period (Figure 2). This represented a slight decrease from 2012. All fish less than 22” in length were cleaned, packed on ice, and sent to local area food banks for distribution. Fish greater than 22” were not retained for food bank distribution because of human consumption guidelines related to mercury content. Those fish were either given to local wildlife rehabilitation centers or were used by FWP wildlife personnel for trapping lure. The length frequency distribution of lake trout caught during the Juvenile Netting period continues to be skewed heavily toward smaller fish, as a result of targeting their location and fishing smaller mesh nets as the primary method (Figure 3). The majority of the juvenile lake trout catch is comprised of age-3 and age-4 lake trout (Cox 2010). Soak times of each panel of net and the depth of nets fished have remained relatively constant throughout the five years of the project. The depth was maximized and duration of these net sets was minimized in an effort to reduce bycatch and associated mortality of non-target species. Bycatch of other fish species during the Juvenile Netting period can be found in Table 2.

5

Figure 3: Relative length frequency of lake trout caught during Juvenile Netting in Swan Lake 2009-2013.

Figure 2: Total number of lake trout removed during Juvenile Netting 2009-2013.

6

Table 2: Bycatch of non-target fish species captured during Juvenile and (Spawner) netting events 2009-2013. Most fish were released alive.

Fish Species 2009 2010 2011 2012 2013

bull trout 238 (26) 212 (87) 237 (104) 334 (103) 168 (135)

kokanee 205 (23) 414 (110) 159 (46) 521 (114) 388 (300)

mountain whitefish 107 (0) 28 (5) 31 (2) 67 (0) 104 (2)

pygmy whitefish 139 (0) 63 (0) 9 (0) 79 (0) 27 (0)

longnose sucker 86 (50) 49 (306) 65 (145) 17 (207) 7 (157)

northern pikeminnow 27 (36) 14 (136) 31 (131) 2 (68) 1 (132)

largescale sucker 0 (58) 0 (109) 0 (111) 0 (54) 0 (96)

rainbow trout 6 (3) 5 (10) 7 (11) 0 (11) 1 (11)

northern pike 0 (2) 0 (0) 0 (7) 0 (2) 1 (7)

Spawner Netting Removal of the adult component of the lake trout population, in efforts to directly affect further recruitment of lake trout cohorts, continues to be an important aspect of the project. Netting for spawning lake trout in 2013 was conducted from October 7-25, with nets being set and lifted twice daily, Monday-Friday. While netting did not occur twice every day (Friday afternoons, Saturdays, Sundays, and Monday mornings were not fished), the schedule and subsequent effort was similar to previous years of the project. Adult lake trout catch in 2013 was 216 fish, nearly identical to the 215 captured in 2012. After realizing in 2009 that FWP-owned boats and equipment was inadequate to set and retrieve large amounts of gill net for covering the entire spawning area, the contract with the Hickey Brothers was modified to incorporate the use of their boat during Spawner Netting. This effort has been replicated since 2010. Since 2010, the number of adult lake trout captured during Spawner Netting has decreased by 47% (Figure 4). Relative length frequency of lake trout captured during Spawner Netting has also shifted to smaller individuals, suggesting that efforts are effectively removing the larger, older individuals from the population (Figure 5). Bycatch of fish species other than lake trout during Spawner Netting was similar to past years (Table 2).

7

Figure 4: Total number of lake trout removed during Spawner Netting in Swan Lake 2009-2013.

Figure 5: Relative length frequency of lake trout captured during Spawner Netting 2009-2013.

8

Bycatch, Bull Trout, and Kokanee Bull trout bycatch continues to be closely monitored as this project has progressed. Since the inception of the project, reducing the amount of bull trout bycatch and minimizing the mortality rates of inadvertently caught bull trout has been a priority. While this project continues to be first and foremost a research experiment, with data being collected in a consistent manner annually, slight changes are being made to reduce bycatch mortality for bull trout. An example of this was the elimination of 3.5” and 4.0”-mesh gill nets during the Spawner Netting period in 2013. Analysis in 2012 revealed that 4.5” and 5.0”-mesh nets accounted for the majority of adult lake trout catch and resulted in the least amount of bull trout bycatch. This change in mesh size still allows for comparable data and should be considered a benefit of the adaptive management approach the working group is taking on this project. A total of 303 bull trout were inadvertently captured during netting activities in 2013. The Juvenile Netting period resulted in 168 bull trout and the Spawner Netting period added another 135. Despite using every effort to reduce mortality of accidentally netted bull trout, roughly 40% of these fish do not survive. Research led by the USFWS (see Appendix) suggests that bycatch mortality is likely higher, as some released bull trout may swim off but perish later. Therefore, a calculated mortality rate of 53.6% is being applied throughout this project. Estimated bycatch motality of bull trout in 2013 was 162 fish. Bull trout bycatch from 2009-2013 has averaged 329 fish/year, resulting in an estimated annual mortality of 176 fish/year. Bull trout bycatch was identified as a considerable concern during the 2012 EA process for a five-year extension of this project. During that time, a commitment was made to further evaluate the cumulative impacts bycatch could have on the bull trout population, and determine if there is a threshold in which changes in the netting project are warranted. The Appendix attached at the end of this annual report contains analysis by the USFWS on the potential cumulative effects of bull trout bycatch since the beginning of this study. While an exact number was not identified as a point in which netting operations would come to a halt, the analysis does suggest that in its current configuration, the level of bycatch mortality (170-180 fish/year) should be sustainable. This analysis is consistent with the fact that until 2012 anglers were able to harvest one bull trout daily from Swan Lake, with an estimated annual harvest of 176 fish. The USFWS analysis provides a “worst case scenario” of what the potential maximum effect of the bull trout bycatch could be. While the Appendix suggests that bycatch can negatively affect bull trout redd numbers, the subsequent benefits to bull trout from the removal of lake trout remains unknown. These potential benefits are the primary reason this experimental removal was initiated and will be the focus for evaluating the success of the project in future years.

9

Introduced kokanee salmon are another important fish species for Swan Lake. Kokanee provide one of the more popular angling opportunities in Swan Lake for both ice and open-water fishermen, and also represent an important food resource for adult bull trout. Case histories from surrounding area lakes have demonstrated that the combination of Mysis, kokanee, bull trout, and lake trout typically results in decreased abundance of bull trout and elimination of kokanee. Therefore, kokanee represent another indicator of potential project success, as increases in abundance may suggest a reduction in lake trout density. Kokanee abundance in Swan Lake is monitored annually through redd counts along an index reach of Swan Lake. Similar to adult bull trout trends, kokanee spawner abundance had declined from 2005-2011 (Figure 7). Redd count surveys in 2011 revealed the lowest abundance throughout the period of record. Surveys conducted in 2012 and 2013 both produced higher redd numbers and the increase over the past two years is encouraging. This trend will continue to be closely monitored to determine if the project is effective in relieving predation from lake trout. In addition to the slight increase in kokanee redd numbers, the total number of kokanee caught as bycatch in both netting periods has also increased in both 2012 and 2013. While bycatch is not typically viewed as favorable, this increase in catch could be representative of increased abundance of kokanee in Swan Lake. As mentioned previously, netting locations and effort have remained relatively constant throughout the project, so the kokanee bycatch data can potentially be viewed as a standardized index. Length frequency analysis of kokanee inadvertently captured during the 2013 netting reveals no missing age classes in Swan Lake (Figure 8). Kokanee in the bycatch ranged from 6-20 inches (150-500 mm). Kokanee smaller than 6” were likely not captured as a result of the mesh sizes used for the netting. The observation of numerous small kokanee should be viewed as positive, as these fish are likely to be most affected by an increasing lake trout population.

10

Evaluation Criteria This lake trout removal project in Swan Lake was initiated to evaluate the efficacy of using gill nets to control the expansion of the lake trout population and benefit bull trout and kokanee. Criteria to evaluate this project were outlined in the original 2009 EA, and continue to be monitored throughout the study. An in-depth review of these criteria with

Figure 7: Kokanee redd count data from Swan Lake 1987-2013.

Figure 8: Length frequency of kokanee captured during both Juvenile and Spawner Netting in 2013.

11

regard to the 2009-2011 efforts can be found in the 3-year Summary Report (Rosenthal et al. 2012). Netting mortality during Juvenile Netting continues to be evaluated on an annual basis. Total annual mortality rates in excess of 50% have been shown to cause population declines in some fisheries (Healey 1978). Conservative estimates of exploitation (mortality) of age-3 and age-4 lake trout have exceeded 50% in most years since 2009. Estimated lake trout mortality during Juvenile netting in 2013 was slightly below 50%. These modeled estimates are most accurate for the two aforementioned age classes of fish, as they are the most vulnerable to the nets being deployed and the locations being sampled. While the estimated mortality rates are encouraging, model validation is likely required. The overall decline in catch throughout the netting period is used to calculate the estimated mortality rates. Catch has consistently declined over the 3-week period each year of the project. Work in upcoming years should test whether this decline is real or a result of fish behavior related to netting activities. Lake trout length frequency during Spawner Netting continues to be skewed toward smaller fish. This shift in length suggests that netting has been effective in removing many of the older, larger individuals, as most fish captured have recently reached sexual maturity. Initial netting in 2009 resulted in the capture of many fish greater than 700 mm (27.5”), with the mode of the distribution centered on 750 mm (29.5”). However in 2013 the mode of the fish caught on the spawning grounds was 575 mm (22.6”), indicating a decrease in the length of spawning fish. In addition to the decrease in length of lake trout captured on the spawning areas, catch per unit effort of lake trout has also decreased over the past four years, suggesting that fewer lake trout are showing up on the spawning areas (Figure 10).

12

Figure 10: Lake trout catch per unit effort during Spawner Netting 2010-2013. Juvenile Netting in 2013 provided researchers the first glimpse of how past efforts have affected the overall lake trout population. Spawner Netting for adult lake trout began in 2009. Undoubtedly some adult lake trout spawned that fall and their offspring would then be age-0 fish during Juvenile Netting in 2010, age-1 in 2011, and age-2 in 2012. Because of fish behavior and the nets being used for this project, lake trout do not fully recruit to the nets until they are age-3. Therefore, netting in 2013 represented the first year of sampling lake trout that had been affected by past Spawner Netting. While the overall number of fish removed during Juvenile Netting in 2013 was considerably lower than in 2012, more years of data are necessary to determine if this trend is real. The upcoming years will determine whether past efforts have been sufficient to significantly reduce the lake trout population. 2014 Plans Netting in 2014 will follow the same schedule as was completed from 2009-2013. Juvenile Netting will be conducted the last three weeks in August. This schedule is identical to that of 2012 and 2013, starting a week earlier than 2009-2011. Lake conditions are similar throughout the month, and the shift of one week ensures that no post-spawn bull trout will be returning to Swan Lake during netting operations. Spawner Netting will follow the same schedule as was done from 2009-2013. As mentioned previously, removal of the spawning adult lake trout is an important component of this project, as it directly affects further recruitment to the lake trout

13

population. While the decrease in catch per unit effort of lake trout on the spawning grounds is encouraging, researchers must be able to determine if this trend is a result of fewer fish or displacement of fish and/or the presence of alternate spawning areas. Therefore in 2014 an additional 30 adult lake trout will be implanted with sonic transmitters and tracked throughout the spawning period. This valuable information will be collected to determine the efficacy of Spawner Netting activities and whether changes to the current operations are necessary. In addition to work being conducted in Swan Lake, research will continue to examine the lake trout populations in Lindbergh Lake and Holland Lake. Lake trout first appeared in FWP sampling in Lindbergh Lake in 2009 and were found in Holland Lake in 2012. Both of these headwater lakes also contain their own populations of bull trout and the implications of these newly established lake trout remain unknown. Biologists are concerned with this expansion of the lake trout population, as bull trout numbers in these two lakes are limited. A sonic telemetry project was initiated by the United States Geological Survey (USGS) in 2013. During the summer of 2013, 9 adult lake trout in Lindbergh Lake were implanted with transmitters. Concurrently, three adult lake trout were given transmitters in Holland Lake. These fish were tracked during the fall of 2013 to begin to identify potential spawning areas. This project will continue in 2014, with an additional 12 transmitters being put into fish. .This information will better inform biologists of the potential for lake trout reproduction in these two lakes and if the potential exists for future suppression. Sonic telemetry was used effectively to identify the spawning area in Swan Lake and this has been the focus of netting operations from 2009-2012 (Cox 2010). Monitoring of the other aquatic organisms will also continue in the Swan Lake system. Annual Mysis sampling will occur in early June, bull trout juvenile estimates will occur in August, bull trout redd counts will happen in October, and kokanee redd counts will be completed in early December. Additionally, spring gill net monitoring will be conducted in Swan Lake, Lindbergh Lake, and Holland Lake to look for trends of all fish species.

14

References Cox, B.S. 2010. Assessment of an invasive lake trout population in Swan Lake, Montana.

Master’s thesis. Montana State University, Bozeman. Healey, M.C. 1978. The dynamics of exploited lake trout populations and implications

for management. Journal of Wildlife Management 42:307-328. Martinez, P. J., P. E. Bigelow, M. A. Deleray, W. A. Fredenberg, B. S. Hansen, N. J.

Horner, S. K. Lehr, R. W. Schneidervin, S. A. Tolentino, and A. E. Viola. 2009. Western lake trout woes. Fisheries 34(9):424-442.

Rosenthal, L., W. Fredenberg, J. Syslo, and C. Guy. 2012. Experimental Removal of

Lake Trout in Swan Lake, MT: 3-year Summary Report. Prepared for the Swan Valley Bull Trout Working Group. Kalispell, Montana.

15

APPENDIX

2007-‐2013 Bull Trout Bycatch Analysis

Wade Fredenberg, U.S. Fish and Wildlife Service

Bull Trout Background and Swan Lake Population Dynamics

Because bull trout are listed as a threatened species under the U.S. Endangered Species Act (USFWS 1998), the U.S. Fish and Wildlife Service has the lead role in monitoring, evaluating, and coordinating efforts to minimize bycatch of bull trout in Swan Lake associated with the lake trout suppression effort. Swan Lake is considered to be an important bull trout “core area”, one of the more robust of over 100 bull trout core areas in the U.S. range of the species. Important tributaries to Swan Lake (e.g., Lion, Woodward, Elk, etc.) are considered to host “local populations”, each supporting spawning and rearing (SR) habitat for genetically identifiable and diverse stocks of bull trout. The mainstem Swan River and Swan Lake are considered foraging, migrating, and overwintering (FMO) habitat, where the separate local populations mingle to grow, mature, and develop prior to completing their life cycle. The bull trout population of the Swan is primarily a migratory adfluvial population, meaning that individuals require both SR (tributary) and FMO (lake and river) habitat to complete their complex life cycle.

An extensive data base on Swan Lake bull trout populations has been collected, extending back to the early 1980’s. A combination of research efforts and ongoing monitoring has resulted in a well-‐documented history of bull trout numbers, spawning locations, life history attributes, and other demographic information. We are using this information to our advantage, both in predicting impacts of lake trout expansion as well as to examine unintended consequences of bycatch from lake trout suppression efforts, angling, and other mortality sources.

In order to effectively suppress lake trout and not exert excessive mortality on bull trout or other important species (such as rainbow or cutthroat trout and kokanee) the lake trout suppression effort has been carefully designed and constantly adjusted in both its timing and method. We conducted a preliminary benefit/risk analysis prior to implementing the 2007-‐2011 suppression efforts (Fredenberg and Rumsey 2007). At that time, we estimated an adult bull trout population of around 5,000 fish in the Swan Lake core area. We knew from extensive surveys that bull trout spawned annually at detectable levels in at least 10 Swan Lake tributaries. Figure 1 (see also Table 2) shows the trend in bull trout redd counts in those ten combined tributaries, since the

16

beginning of comprehensive basinwide surveys in 1995. Counts from the comprehensive basinwide surveys do not monitor every single bull trout redd in the system, but are known (based on several more extensive basinwide searches) to encompass the vast majority of spawning reaches (estimated 90% or more) and these redd counts are thought to provide highly accurate tracking of bull trout spawner trends.

Figure 1. Swan Lake basinwide bull trout redd counts 1995-‐2013.

The 1995 to present trend in basinwide redd counts covers a period of relatively high bull trout population levels, based upon what is known about the latter 20th century historical condition for this core area. These high levels reflect largely favorable conditions of an abundant forage base (primarily Mysis relicta and kokanee) in Swan Lake, stable or improving habitat trends in the spawning tributaries, and restricted angling mortality. Angler harvest of bull trout declined (despite stable or increasing bull trout populations) from an estimated 482 bull trout in 1995 (Rumsey and Werner 1997) to an estimated 176 bull trout in 2009 (MFWP 2010, unpublished), as fishing regulations were gradually tightened; all legal bull trout harvest was eliminated beginning in 2012.

Additional data from comprehensive bull trout studies in the Swan basin include work by Leathe and Enk (1985) and Leathe et al. (1985 a, 1985 b) which described physical and fishery attributes of the important bull trout spawning tributaries. Those

0 100 200 300 400 500 600 700 800 900

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

17

reports determined that bull trout spawning populations in Swan tributaries were comprised, on average, of 38% Age 5, 36% Age 6, 20% Age 7, 6% Age 8, and <1% Age 9 and older bull trout. Those same studies contained extensive analysis of age and growth information for bull trout.

In 2007, we (Fredenberg and Rumsey 2007) estimated, based largely on previous netting experience in Swan Lake, that a lake trout gillnet suppression effort conducted during midsummer (late August-‐ early September) when most adult bull trout were up the river on spawning migrations would catch between 225-‐900 bull trout annually, with an expected mortality rate of netted fish between 10-‐25% and a maximum estimated bull trout mortality of 315 fish. We determined that this rate of bycatch mortality should be sustainable given demographic parameters, thereby allowing the project to proceed. We concluded that this rate of loss would likely not cause an irreversible decline in a bull trout population that was also being harvested by angling and was estimated at over 5,000 adults.

Bull Trout Bycatch During Lake Trout Gillnetting Suppression Activities

During 2007 and 2008, early experimental gillnetting efforts occurred. In 2009-‐2011 a more comprehensive three-‐year experimental lake trout suppression effort was conducted. That 3-‐year effort was evaluated and, subsequently, five more years of lake trout suppression were approved starting in 2012. Procedures, overall catch, and effort are summarized elsewhere. The purpose of this analysis is to determine potential effects of bycatch from ongoing lake trout suppression, specific to the impact on threatened bull trout. During the seven years of netting to date (2007-‐2013), we have captured a total of over 2,200 bull trout (n = 2,237) in combined lake trout suppression efforts. Annual catch totals are shown in Table 1.

Table 1. Bull trout bycatch in lake trout suppression efforts, by year, with ratio.

Netting Year Total Bull Trout Bycatch Lake Trout : Bull Trout Ratio 2007 378 5.6 2008 240 15.8 2009 238 22.9 2010 299 33.5 2011 342 15.9 2012 437 24.3 2013 303 23.8 TOTAL 2,237 20.0

18

The cumulative size structure of bycaught bull trout that were measured (n = 2,047) in 2007-‐2013 (Figure 2), depicts a relatively normal distribution with diminishing numbers above the peak size range of 340-‐359 mm (Figure 2). This distribution continues to reflect a generally strong population of age 3 and 4, mostly juvenile bull trout in the samples (~ 320-‐340 mm and smaller), with declining numbers of age 5 and older fish, many of which would be mature adults. During typical August-‐September netting windows, an unknown, but presumably high proportion of the mature population would have left the lake for the annual spawning migration and those fish were typically absent from the lake population and unavailable for capture in our gear.

Figure 2. Length frequency of bull trout bycatch (includes 2,047 fish that were measured) from gillnets, 2007-‐2013.

There was a pronounced shift to smaller bull trout in the bycatch sample, beginning in 2009 (Figure 3), when netting effort became more focused on juvenile lake trout in deeper water. The size range of fish captured directly reflects changing net dimensions (especially mesh size) and to a lesser degree, changing effort, but does not necessarily correlate to the size and age structure of the bull trout population. More than 60% of all bull trout captured (including both types of netting for juvenile lake trout and spawners) in 2009-‐2013 were under 400 mm total length (Figure 3) and nearly all those fish are considered immature juveniles and subadults based solely on their size.

0

20

40

60

80

100

120

140

160

180

200

220

240

260

280

300

320

340

360

380

400

420

440

460

480

500

520

540

560

580

600

620

640

660

680

700+

Length Categories (20 mm)

Num

ber M

easured

19

While net standardization was considered important for assessing program success, that consideration was sometimes overshadowed by adjustments made in the mesh sizes we fished in order to minimize bull trout bycatch. In recent years, we have increasingly focused on use of smaller gillnet mesh sizes for juvenile lake trout suppression in August (1.5” – 2.75” stretch) backed by increasingly larger mesh sizes (4.5” – 5” stretch) during lake trout spawner netting in October. The separate targeting of small fish in August juvenile netting and large fish in October spawner netting accounts for the increasingly bimodal distribution of bull trout lengths exhibited in the 2011-‐2013 netting results (Figure 3), since nets that would target medium sized fish (3”-‐ 4” stretch) are no longer routinely used.

20

Figure 3. Length frequency (by 20 mm size group by year) of bull trout captured in annual gillnet bycatch associated with lake trout suppression netting in Swan Lake during 2007-‐2013.

In comparing the bull trout samples captured in the gillnetting efforts, those captured in spawner netting conducted in October and early November, after adult bull trout had returned to the lake, contained an expectedly higher proportion of adult sized fish. Roughly two-‐thirds of the bull trout captured during spawner netting in 2009-‐2013 were mature size (i.e., presumably Age 6 and older and greater than ~425 mm) specimens. Up

0

10

20

30

40

50

60

% o

f Ann

ual D

istr

ibut

ion

Length Categories (20 mm)

2013 2012 2011 2010 2009 2008 2007

....Age 3.../......Age 4......../........Age 5....../........Age 6......../.........Age 7........./........Age 8........./...9+

21

to half of those fish appeared, based on external characteristics, to be post-‐spawning adults that had recently returned to the lake from upriver spawning activity; but visual identification of postspawn adults was likely conservative.

Bull Trout PIT Tags and Returns

In 2010, we also began inserting PIT tags into the dorsal sinus, anterior to the dorsal fin, of all bull trout released, in order to directly verify that survival was occurring and to further evaluate the accuracy of the survival classification scheme. Between 130 and 261 unique PIT-‐tagged bull trout were released annually, 734 in total (Table 2). During sampling, we assess mortality of gillnetted and released bull trout by assigning each fish to one of four condition categories. Class “0” = Mort (dead or moribund); Class “1” = Poor (not orienting, possible bleeding, respiration shallow); Class “2” = Fair (tired, but orienting and respiring, considered “likely to survive”); Class “3” = Good (vigorous, struggles to escape, swims away upon release). For each of these mortality classes we have subjectively ascribed a rough estimate of survival probability. For Class “0” survival probability is nil. Class “1” survival probability is estimated at 10-‐50% (average 30%). Class “2” survival probability is estimated at 51-‐90% (average 70%). Class “3” survival probability is estimated at 91-‐100% (average 95%). During the progression of the netting surveys we have further attempted to enhance survival capacity of netted fish, particularly those in survival classes 1 and 2, by holding them for up to 30 minutes in an oxygenated, chilled, recirculation tank known as a Fraser Tank. We have documented instances where fish, upon capture, appear moribund due to suffocation but are later observed to revive in the Fraser Tank to be released as condition Class 1 or even Class 2.

Eighty-‐four of the PIT-‐tagged fish we released were condition Class 1 (11.4%), another 226 were condition Class 2(30.8%), leaving 424 that were condition Class 3 (57.8%). Results are preliminary, but in 2010-‐2013 a total of 55 marked bull trout were recaptured, 3 of which had finclips but no PIT tag (apparent tag loss approximating 5.5%). Of the 52 PIT-‐tagged fish that were recaptured, 21 were at large 3 weeks or less, indicating they were recaught during the same juvenile or spawner netting session when they were tagged. One of those fish (5%) was originally designated Class 1 upon release, 10 each (47.5%) were originally designated as Class 2 or Class 3 upon release. Length at first capture ranged from 238-‐622 mm and mean length at first capture was 378 mm. Seven of those fish (33%) were caught/recaught during spawner netting and 14 other fish (67%) during the higher effort associated with juvenile netting events. Most mesh sizes of gillnet were represented in this sample, from 1.5” to 5” stretch. Both captures and recaptures were distributed across both North and South basins of Swan Lake.

22

Table 2. Number of unique PIT-‐tagged bull trout released, by year and condition class.

Year Class 1 Class 2 Class 3 Total 2010 22 33 75 130 2011 21 48 95 164 2012 24 88 149 261 2013 17 57 105 179 TOTAL 84 226 424 734

There were three PIT-‐tagged bull trout caught originally in juvenile netting, then recaptured in spawner netting events 49-‐63 days later during the same year. All three were, almost by necessity, larger fish (452-‐631 mm in length), since that’s what the spawner netting mostly captures. One of these was condition Class 2, and the other two were condition Class 3 when originally captured.

The most interesting group of recaptures were 28 PIT-‐tagged fish that were at large 316 days or longer, from one netting year to the following year or longer. Nine of those fish (32%) were originally captured during spawner netting events and 19 (68%) were originally captured during juvenile suppression netting. Of the recapture events, 13 (46%) occurred during spawner netting and 15 (54%) during juvenile suppression netting, with again most mesh sizes represented. Of the 28 longterm recaptures, 2 fish (7%) were originally classified as survival Class 1, another 5 fish (18%) were survival Class 2, and the remaining 21 fish (75%) were survival Class 3 when originally handled. These results would seem to corroborate our hypothesis that subjective evaluation of health status when released is a reasonable indicator of long-‐term survival probability. Comparison to short-‐term recapture results (above) may also indicate that some delayed mortality of Class 2 fish occurs, as fish released as Class 2 were much more prevalent in short-‐term returns than in long-‐term returns. Initial lengths of long-‐term recaptures were again representative of the spectrum, ranging from 212-‐520 mm, though no fish longer than 520 mm when originally tagged was recaught a year or more later. This may reflect higher natural or angling mortality or possibly netting mortality, but as previously noted the sample is biased, because larger adults are mostly vulnerable only during spawner netting.

Recaptures of fish at large 316 days or longer ranged from 310-‐586 mm total length at time of recap. All fish at large 316 days or longer showed positive growth increments. Fifteen of the 28 longterm recaptures grew 100 mm or more. Seventeen fish that were at large roughly one year (316-‐423 days) grew an average of 78.6 mm (range 5-‐139 mm). Ten fish that were at large roughly two years (661-‐787 days) grew an average of 186.3 mm (range 66-‐212 mm). One exceptional specimen grew 277 mm,

23

doubling in length from 277 mm when tagged to 554 mm after being at large 1157 days, an average increment of over 7 mm per month over 3+ years.

There were four PIT-‐tagged bull trout that survived multiple recapture/release events. One bull trout originally PIT-‐tagged as a 250 mm fish on 10/21/11 and released as a condition Class 3, was recaught on 8/13/12 at 310 mm and released as a Class 2, then caught again on 8/12/13 at 373 mm and succumbed. Another, originally PIT-‐tagged as a 433 mm fish on 8/22/11 and released as a Class 2, was recaught more than two years later, on 10/16/13 at 564 mm and released as a Class 1, then caught again the following day during spawner netting and released again as a Class 1. A third fish, originally PIT-‐tagged as a 358 mm fish on 8/25/10 and released as a Class 3, was recaught more than two years later, on 10/18/12 during spawner netting at 570 mm, and released as a Class 2, only to be caught again and succumbing the following day. The ultimate survivor, however, was a fish originally caught on 8/24/10 at 277 mm on the north end of the lake, released as a Class 3 and then caught and released again 8 days later as a Class 3 nearby. This fish was caught again two years later at 448 mm on 8/22/12, and again released as a condition Class 3 on the north end of Swan Lake. Finally, this fish was caught a 4th time, during a spawner netting event on 10/24/13 on the south end of the lake at a length of 554 mm, and it survived again to be released as a condition Class 2. It is presumably still at large, after being gillnetted and released four times.

In summary, from 2010-‐2013 a total of 734 uniquely tagged bull trout were released; 52 recapture events of these PIT-‐tagged fish occurred, with 5 multiple recaptures involving 4 different fish. Of the 47 first-‐time recapture events, 3 involved fish originally marked as condition Class 1, another 15 involved fish originally marked as condition Class 2, and the other 29 involved fish originally marked as condition Class 3. This means Class 1 fish at large were recaptured 3.6% of the time (3/84), Class 2 fish were recaptured 6.6% of the time (15/226), and Class 3 fish were recaptured 6.8% of the time (29/424).

Assessment of Bull Trout Bycatch by Cohort

We made a previous effort to assess the potential impact to the bull trout population of selective losses to bycatch in 2007-‐2011 (Rosenthal 2012). As the redd counts declined in 2008-‐2011 (Table 3, see also Fig. 1), there was appropriate concern that the netting could be a causative or contributing factor in the decline. In order to best evaluate this concern, we first estimated the ages of bull trout captured in the

24

netting by back-‐correlating their sizes with the previous information on age and growth (Leathe and Enk 1985). We determined that roughly 10% of the total gillnetted sample was comprised of age 3 fish (~181-‐248 mm); 24% were age 4 (~249-‐337 mm); 25% were age 5 (~338-‐427 mm); 22% were age 6 (~428-‐518 mm); 12% were age 7 (~519-‐611 mm); 5% were age 8 (~612-‐694 mm); and 2% were age 9 or older (> ~695 mm).

There are multiple assumptions involved in this analysis. By assuming bull trout captured were of known age, we were then able to assign individuals back to the spawning cohort from which they originated and essentially reconstruct mortality estimates by age cohort (Table 3). We assumed the netting affects six age classes of bull trout, ages 3-‐8; since age 1 and 2 fish are generally in the tributaries and/or below sizes of recruitment to our gear, and age 9+ fish are an insignificant portion of the spawning population.

Table 3. Bull trout spawning cohorts, basinwide redd counts, and calculated maximum mortality by cohort due to lake trout suppression netting during 2007-‐2013.

Spawning Cohort

# Redds Basinwide

Years During Which Netting Affects Cohort

Netting Bycatch for Cohort

(# bull trout Age 3-‐8)

Calculated Mortality = 53.6% of bycatch (# of bull trout)

1996 748 1997 833 1998 829 2007 24 13 1999 682 2007-‐2008 76 41 2000 719 2007-‐2009 137 73 2001 687 2007-‐2010 279 150 2002 540 2007-‐2011 173 93 2003 592 2007-‐2012 151 81 2004 597 2008-‐2013 284 152 2005 588 2009-‐2014 243 130 2006 656 2010-‐2015 231 124 2007 762 2011-‐2016 366 196 2008 598 2012-‐2017 150 80 2009 481 2013-‐2018 20 11 2010 378 2014-‐2019 0 0 2011 312 2015-‐2020 0 0 2012 405 2016-‐2021 0 0 2013 335 2017-‐2021 0 0

Cumulatively, the 2007-‐2011 bull trout database included 1,477 fish for which survival class was recorded (20 fish in the total sample lacked a survival code). About 39.7% (n = 586) were direct mortalities, 8.5% (n = 126) were Class “1”, 13.9% (n = 205)

25

were Class “2”, and the remaining 37.9% (n = 560) were Class “3”. By assigning the predicted average survival to each class, we estimated total gillnet mortality was 586 (Class “0”) + 88 (70% of Class “1”) + 62 (30% of Class “2”) + 56 (5% of Class “3”) = 792. Thus, our calculated estimate of total bull trout gillnet mortality was 792/1,477 = 53.6% of the fish handled. We did not recalculate this ratio to include 2012-‐2013 sampling, but saw no reason to expect significant variance. In Table 3, we assigned 53.6% mortality to each cohort in order to better estimate the impact of mortality on the bull trout population.

Applying the four-‐stage condition survival scale to the bull trout bycatch resulted in a calculated total mortality by year associated with juvenile netting of 101-‐189 fish per year in 2007-‐2013 (Table 4). We began spawner netting in 2009. Calculated mortality for spawner netting, using the same protocol, was 18-‐53 fish per year in 2009-‐2013. In cumulative, juvenile netting from 2007-‐2013 resulted in a calculated mortality of 937 fish and spawner netting from 2009-‐2013 added 214, for a total calculated mortality of 1,151 bull trout in 2007-‐2013 (an average of 164 per year).

Table 4. Bull trout bycatch associated with 2007-‐2013 lake trout suppression efforts, by year, with estimated total mortality after factoring in estimated delayed mortality.

Netting Year Total Bull Trout Bycatch

Estimated Suppression Netting Mortality

Estimated Spawner Netting Mortality

Estimated Total Mortality

2007 378 189 189 2008 240 131 131 2009 238 109 18 127 2010 299 121 48 169 2011 342 133 49 182 2012 437 153 53 206 2013 303 101 46 147 TOTAL 2,237 937 214 1,151

Projected Impact of Gillnetting Bycatch on the Bull Trout Population

The impact of gillnetting bycatch loss on the potential basinwide bull trout redd production for the Swan Lake core area is a function of the number of fish removed, their age (which relates to the number of potential spawning years they could contribute), size (related to fecundity), and sex ratio (undetermined, but assumed equal). The impact is also partially dependent upon whether or not spawning occurs

26

annually, whether or not compensatory mortality is occurring, as well as mortality rates and other factors. In the simplest model, we assume all spawners make an equal contribution (standard age, size, and sex ratio) and that there are 3.2 adults per redd, an oft-‐used conversion factor applied in the bull trout literature that is based on earlier studies in Flathead Lake (Fraley and Shepard 1989) and Lake Pend Oreille (Downs et al. 2006). Some historical information (largely pre-‐Mysis and applicable to bull trout populations sympatric with kokanee) indicated that adfluvial bull trout in Flathead Lake and similarly productive headwater systems such as Swan Lake routinely trended toward alternate year spawning. More recent studies (e.g. Johnston and Post 2009, studying Lake Kananaskis, Alberta; Downs et al. 2006, studying lake Pend Oreille), and our current Swan Lake observations indicate that very few adult-‐sized bull trout are present to be captured during the immediate pre-‐spawning period (late August – early September) when juvenile lake trout netting is occurring. This suggests that most adult Swan Lake bull trout at this time are either annual spawners or otherwise avoiding capture. For purposes of our simple model, we assume that no compensatory mortality is occurring and that once bull trout reach age 3 the annual mortality is nil. We assume that every bull trout that reaches the age of 5 spawns annually at ages 5, 6, 7, and 8; an assumption that would maximize the potential lifetime productivity of individual fish. This results in a worst-‐case scenario in order to predict the maximum possible impact of gillnetting bycatch.

Applying these assumptions, we calculated the projected maximum annual loss of spawners and subsequent potential impact to redd counts due to gillnetting bycatch removal. In 2007, for example, when the project began, we estimated gillnet mortality of 189 bull trout. Since most spawners were likely out of the lake by the time netting began (or avoiding capture), the impact on the redd count in 2007 (n = 762 basinwide) would have been nil. To estimate the impact of 2007 gillnet bycatch on the bull trout redd count in 2008, we applied the average age class distribution (10% age 3; 24% age 4; 25% age 5; 22% age 6; 12% age 7; 7% age 8+) to the calculated bull trout mortality from 2007 (n = 189) and estimated that about 19 of the mortalities were age 3, ~45 were age 4, ~ 47 were age 5, ~42 were age 6, ~23 were age 7, and finally about 13 age 8+ bull trout were removed (Table 5). We further assumed those fish age 4 and older that we removed in 2007 would have all survived to spawn (age 5 and older) in 2008; with exception of those age 8 and older, which we eliminated from the spawner population due to senescence. In this fashion, 2007 gillnetting was estimated to remove 45+47+42+23 = 157 adult spawners from the 2008 population (Table 6). Converting this number to the observed number of redds, using the standard 3.2 adults/redd formula: 157/3.2 = 49. Thus, our estimate was that the 2007 gillnetting potentially removed a total of up to 49 redds from the 2008 basinwide count.

27

Similar calculations were performed annually for years 2007-‐2013 (Table 6). Using this model, removal of fish that are Age 4 has the greatest long-‐term impact, even though they do not spawn until Age 5, as we assumed noncompensatory mortality (and unrealistically low 0% annual mortality), meaning loss of those fish would continue to affect the redd count baseline for four successive years.

Table 5. Calculated age-‐class structured bycatch mortality of bull trout for each year of lake trout suppression netting to date.

Year Netted

Calculated Mortality (juv+spawn) =

Mortality Age Class 3 (0.10)





Mortality Age Class 8+ (0.07)

2007 189 19 45 47 42 23 13 2008 131 13 31 33 29 16 9 2009 109+18 = 127 13 30 32 28 15 9 2010 121+48 = 169 17 41 42 37 20 12 2011 133+49 = 182 18 44 45 40 22 13 2012 153+53 = 206 21 49 52 45 25 14 2013 101+46 = 147 15 35 37 32 18 10

Table 6. Calculated cumulative bycatch mortality of bull trout in relation to spawning year, for combined lake trout netting to date.

Spawn Year

Bycatch Mortality Loss (5 years prior)




Bycatch Mortality Loss (1 year prior)

Cumulative Bycatch Mortality Loss by Cohort

2008 45+47+42+23 157 2009 19+45+47+42 31+33+29+16 262 2010 19+45+47 13+31+33+29 30+32+28+15 322 2011 19+45 13+31+33 13+30+32+28 41+42+37+20 384 2012 19 13+31 13+30+32 17+41+42+37 44+45+40+22 426 2013 13 13+30 17+41+42 18+44+45+40 49+52+45+25 474 2014 13 17+41 18+44+45 21+49+52+45 35+37+32+18 467 2015 17 18+44 21+49+52 15+35+37+32 320 2016 18 21+49 15+35+37 157 2017 21 15+35 71 2018 15 15

In summary, we theoretically observe an accumulating trend in the loss of potential spawners which, as discussed, would have a maximum effect to bull trout after

28

five consecutive years of netting (removing age classes 4-‐8). This is logically an overestimate because some level of compensatory mortality may be occurring and annual mortality certainly occurs, though levels are not known. Conversely, some bull trout may spawn past age 8, but those numbers are believed to be low. Nonetheless, these bycatch mortality totals were designed to estimate the maximum potential effect of the bycatch on the bull trout population. We observe from the last column (in bold) from Table 6 that the maximum loss of bull trout spawners due to netting begun in 2007 did not accumulate until spawning year 2012, when an estimated 426 spawners were potentially removed from the bull trout population. Similarly, in 2013 we lost an estimated maximum of 474 spawners and for 2014 the estimated loss will accrue to 467 spawners. For those three years where we estimate complete bycatch loss totals, estimated reduction of spawners was up to 426-‐474 annually, averaging 456. Using the existing conversion factor of 3.2 adults per redd, a reduction of 456 spawners would net a maximum reduction of about 142 redds annually.

Conclusions

Basinwide bull trout redd counts for the Swan Lake core area steadily declined from 2007 through 2011 (Figure 1), but appeared relatively stable in 2011-‐2013. Since netting began in 2007, a recent peak of 762 redds was counted in 2007; with a more recent range of roughly half that (312-‐405 redds in 2010 through 2013). Based on the preceding analysis, it appears the gillnetting bycatch likely contributed to the 2007-‐2011 decline, but is at most partially responsible. The maximum potential reduction in redds attributable to bycatch mortality can be roughly estimated by dividing the cumulative bycatch mortality (last column in Table 6) by 3.2 (# adults/redd). Thus, the maximum amount that the gillnet bycatch potentially reduced redd counts was by approximately 49 in 2008, 82 in 2009, 101 in 2010, 120 in 2011, 133 in 2012, and 148 in 2013. The highest (2013) estimate is unlikely to be exceeded in the future under the current netting protocols. Further, as discussed, these are absolute maximum estimates since we have chosen to discount annual and compensatory mortality and assume annual spawning for every fish. If the lake trout gillnetting program continues at existing levels, then bycatch of bull trout can be expected to contribute to the loss of, at most, 130-‐150 bull trout redds per year. If the gillnetting program were to be discontinued, reduced in intensity or duration, or otherwise improved by reducing the level of bycatch or increasing survival of bycaught bull trout, then the ongoing impacts to the redd counts would still be felt through at least 2018, though at increasingly reduced levels for the next 5 years.

29

It is important to note, however, that if redd counts decline further, effects of netting bycatch could become more consequential to the bull trout population. Loss of 150 redds from a peak of 762 (as in 2007) would represent only 16% off the top, but loss of 150 redds from a count of 312 (as in 2011) represents nearly a one-‐third (32%) decline in potential. Another way to measure the potential impact of the loss of redds due to netting relative to potential bull recruitment could be through the ongoing bull trout juvenile abundance surveys conducted annually by MFWP in four primary spawning streams. Multi-‐pass population estimates of age 1 and older bull trout have been conducted more or less annually in the same sections of Elk, Goat, Lion, and Squeezer Creeks since 1987 (MFWP 2013, unpublished). Results are expressed in terms of fish/100m2. A composite index number is obtained by averaging the point estimate from individual streams, typically for all four streams but at minimum two (in some years not all streams were surveyed). For 2001 through 2013 the data set is continuous. Because juvenile abundance (mostly age 1 and 2) is primarily a measure of recruitment from redds counted 2 and 3 years prior to the juvenile survey, we use a 2-‐year time lag between redd counts and juvenile numbers to make a direct line-‐by-‐line comparison in the table (Table 7).

The composite juvenile abundance index ranged from 2.26 to 4.64 bull trout / 100m2 in (MFWP 2013, unpublished) in 2003-‐2013, corresponding to juveniles produced during redd count years 2001-‐2011, and averaged 3.41, very near the longterm (23-‐year) average of 3.45 compiled since 1985. The average composite juvenile abundance in the first seven of those years, immediately prior to any effects from the initiation of lake trout suppression (juveniles sampled in years 2003-‐2009, representing redds counted in 2001-‐2007), was equivalent to the longterm average at 3.45 bull trout / 100m2. Redd counts in 2001-‐2007 ranged from 402-‐521, averaging 458. In the succeeding four years, when bull trout bycatch associated with lake trout suppression efforts could have had an impact on redd counts (2008-‐2011) redd counts ranged from 210 to 395 and averaged 310; about two-‐thirds of the level posted in the preceding seven years. Yet, the composite estimate of juvenile abundance measured in 2010-‐2013 (correlated to those 2008-‐2011 redd counts) ranged between 2.26 and 4.39, averaging 3.34 juvenile bull trout / 100m2. The observation that a one-‐third decline in redd counts resulted in only an insignificant decline (0.11 juvenile bull trout / 100m2) in juvenile abundance provides some assurance that juvenile recruitment is still robust, despite lower redd numbers. In fact, at recent redd count levels there appears to be, at most, a weak positive relationship between redd counts and juvenile recruitment two years later (Figure 4).

30

Table 7. Index redd counts and juvenile bull trout (age 1 and older) composite average abundance index (#/100m2) calculated two years later, for four combined index streams of the Swan Lake bull trout core area (Elk, Goat, Lion, and Squeezer creeks).

Spawning Year Index Redd Count Composite Juvenile Abundance Index Two Years Later (# age 1+ /100m2)

1985 109 6.72 1986 210 2.99 1987 290 2.59 1988 321 3.21 1989 371 3.11 1990 263 1991 366 1992 375 2.06 1993 432 2.61 1994 493 1995 508 3.36 1996 526 3.64 1997 586 2.83 1998 612 1999 501 4.38 2000 519 4.27 2001 502 3.91 2002 430 3.26 2003 425 4.64 2004 435 3.49 2005 402 2.30 2006 489 2.51 2007 521 4.01 2008 395 4.39 2009 366 2.26 2010 268 3.63 2011 210 3.07 2012 252 2013 201 MEAN 3.45

31

Figure 4. Association between composite juvenile bull trout abundance index (juvenile

bull trout / 100m2) in 2003-‐2013 and the redd counts two years prior that were responsible for producing those juvenile fish.

In future years, it will be important to continue to collect and track juvenile recruitment data as a potential early warning of any imminent population collapse. There could be a threshold effect below which bull trout juvenile numbers begin to decline. In the late 1980’s and early 1990’s, redd counts similar to the range seen in the recent past also produced composite juvenile abundance similar to the long-‐term average (3.45 bull trout / 100m2 ). Current redd counts do not appear to be approaching any threshold that limits juvenile recruitment.

Finally, in order to be proactive, MFWP closed the bull trout harvest fishery on Swan Lake beginning March 1, 2012. In 1983-‐1984 the estimated angler harvest of bull trout from Swan Lake was 738 fish and as recently as 1995, estimated angler harvest was 482 fish, well above numbers of bull trout currently removed by gillnet bycatch. The angler harvest has continued to decline since then, attributable to tighter regulations and perhaps changing angler ethics in regard to bull trout, with an estimate of only 176 bull trout harvested in the fishery 2009 (MFWP 2010, unpublished). It is logical to assume that the angler harvest in 2007-‐2011 was additive to the bycatch mortality from netting and so by closing the harvest fishery some additional buffer should be gained. The recent level of angler harvest (n =176) appears to have been roughly equivalent to the total number of bull trout removed (calculated bycatch mortality of 127-‐206 – see Table 3) with gillnets.

2

2.5

3

3.5

4

4.5

5

200 250 300 350 400 450 500 550 Compo

site Ju

venile A

bund

Inde

x

Index Redd Count

Index Redd 2001-‐2011 vs Juv Abundance 2003-‐2013

32

In conclusion, the bull trout bycatch associated with the lake trout gillnetting program, as it is currently configured (30 juvenile netting events followed by 20 spawner netting events), is likely to remain at around 300-‐400 fish per year (Table 1), with roughly half those fish killed. The predicted population level impact of that removal is a reduction of up to 150 bull trout redds per year from the systemwide total. Based on our current understanding of bull trout population dynamics in Swan Lake, that level of reduction should be sustainable and should not have long-‐term population level consequences to bull trout. The 2007-‐2011 redd count declines were steeper than what would have been predicted solely by gillnet bycatch. There is no way to feasibly determine the impact to bull trout from competition or predation by lake trout, but the removal of nearly 45,000 lake trout from Swan Lake by gillnetting (2007-‐2013) should benefit the bull trout population by reducing competition and predation risk and, importantly, should also help to maintain a viable kokanee population which is a vitally important forage species.

Literature Cited

Downs, C.C., D. Horan, E. Morgan-‐Harris, and R. Jakubowski. 2006. Spawning demographics and juvenile dispersal of an adfluvial bull trout population in Trestle Creek, Idaho. North American Journal of Fisheries Management 26:190-‐200.

Fraley, J.J. and B.B. Shepard. 1989. Life history, ecology, and population status of migratory bull trout (Salvelinus confluentus) in the Flathead Lake and River system, Montana. Northwest Science. 63:133-‐143.

Fredenberg, W. and S. Rumsey. 2007. Benefit and risk analysis. 2007 trap net and gill net survey – Swan Lake, Montana. U.S. Fish and Wildlife Service, Kalispell.

Johnston, F.D. and J.R. Post. 2009. Density-‐dependent life-‐history compensation of an iteroparous salmonid. Ecological Applications 19(2)449-‐467.

Leathe, S.A. and M.D. Enk. 1985. Cumulative effects of microhydro development on the fisheries of the Swan River drainage, Montana. I. Summary Report. Montana Department of Fish, Wildlife and Parks, Kalispell, Montana. Prepared for Bonneville Power Administration, Portland, Oregon. Project 82-‐19.

Leathe, S.A., S. Bartelt, and L.M. Morris. 1985a. Cumulative effects of microhydro development on the fisheries of the Swan River drainage, Montana. II. Technical

33

Information. Montana Department of Fish, Wildlife and Parks, Kalispell, Montana. Prepared for Bonneville Power Administration, Portland, Oregon. Project 82-‐19.

Leathe, S.A., S. Bartelt, and L.M. Morris. 1985b. Cumulative effects of microhydro development on the fisheries of the Swan River drainage, Montana. III. Fish and Habitat Inventory. Montana Department of Fish, Wildlife and Parks, Kalispell, Montana. Prepared for Bonneville Power Administration, Portland, Oregon. Project 82-‐19.

MFWP Unpublished. 2010. Results of 2009 Swan Lake creel survey. Montana Fish, Wildlife and Parks, Kalispell.

MFWP Unpublished. 2013. Compiled juvenile bull trout abundance estimates for Elk, Lion, Goat and Squeexzer Creeks, 1985-‐2013. Montana Fish, Wildlife and Parks, Kalispell.

Rosenthal, L. 2012. Draft Environmental Assessment for an extension of experimental lake trout removal efforts in Swan Lake, Montana. Montana Fish, Wildlife and Parks, Kalispell.

Rumsey, S. and T. Werner. 1997. Swan Lake, Montana angler creel survey -‐ 1995. Montana Fish, Wildlife and Parks. Helena.

U.S. Fish and Wildlife Service. 1998. Endangered and threatened wildlife and plants; Determination of threatened status for the Klamath River and Columbia River distinct population segments of bull trout. Final rule. Federal Register 63(111)31647-‐31674.

experimental removal of lake trout in swan lake, mt:...

Documents