Bull trout (Salvelinus confluentus) COSEWIC assessment and status report 2012: chapter 9

Population Sizes and Trends

Sampling Effort and Methods

Visual counts of redds have been the primary stock assessment tool for adult Bull Trout populations (Dunham et al. 2001; USFWS 2008). This is one of the least expensive and uninvasive adult population assessment methods. The characteristic form and bright, clean appearance of redds, as well as the low water conditions generally present during the early fall mean that they can be a reliable indicator of spawner abundance. Tight correlation of redd counts with independent estimates of population size have verified their usefulness (Dunham et al.2001; Al-Chokhachy et al.2005), although a number of caveats about their reliability and repeatability apply to their use.

Errors (omissions and false identifications) must be reasonably low if redd counts are to accurately indicate the status of a population, and provide an index of population trends. High levels of inter-observer variability can be a significant source of error in redd count accuracy and precision (Dunham et al. 2001) and will confound the ability to detect trends in streams over limited time scales (Rieman and Myers 1997). These discrepancies can be greatly reduced, however, when detailed criteria for redd identification and experienced observers are used (Muhlfeld et al. 2006; Decker and Hagen 2008).

In addition to inconsistent methodology, variations in detection rate among streams and at different times also contribute to inconsistencies. While a near complete count of redds may be possible under certain environmental conditions, weather and stream type may result in underestimates (Decker and Hagen 2008). For example, high flows may delay redd counts and lead to underestimates, as redds become more difficult to identify with the passing of time after spawning (Decker and Hagen 2008). Also, in areas of limited gravel or high redd abundance, or where spawning site selection is highly specific, superimposition of redds upon one another can occur (Baxter and McPhail 1996). Redd counting in these instances can only be based on a subjective evaluation (Decker and Hagen 2008).

Caution must be applied when estimating the number of adults from the number of redds counted. The Bull Trout’s propensity for alternate-year spawning or resting periods between consecutive spawning events (Pollard and Down 2001) means that only the number of spawning adults can be estimated. In addition, the expansion factor from redds to number of spawners can vary among populations as a result of single females constructing more than one redd (Leggett 1980) and sneak fertilizations from inconspicuous satellite and sneaker males in some populations (Kitano et al. 1994; Baxter 1997; McPhail 2007). Some males may fertilize more than one redd while some redds may be fertilized by more than one male (Fraley and Shepard 1989). A review of three calibration studies that used independent estimates of population size (two in British Columbia and one in Idaho) found the average number of Bull Trout spawners/redd to be 2.2 (Decker and Hagen 2008), while two other BC rivers yielded expansion factors of 1.5 and 3 (Pollard and Down 2001). This range is confirmed by Al-Chokhachy et al. (2005), whose review of five studies in the Columbia River basin suggested an average expansion factor of 2.7 (range of 1.2 - 4.3).

Alternative methods for estimating the spawning population include trapping migratory populations, electro-fishing, snorkel surveys, aerial surveys and, more recently, resistivity counters (Hagen and Decker 2011). All of these methodologies are more labour intensive than redd counting, and each comes with their own potential drawbacks. For example, trap avoidance may bias trapping estimates. Electro-fishing gear is size selective and its capture efficiency diminishes in Bull Trout’s preferred habitat of flowing waters with low conductivity and high cover (Bonneau et al. 1995; Peterson et al. 2004). Day and night snorkel counts can compensate for diel shifts in Bull Trout habitat use but counting errors depend partly upon water clarity and habitat type (Thurow and Schill 1996; Dunham et al. 2001; Thurow et al. 2006). The deployment of resistivity counters is a costly procedure and their reliability has yet to be evaluated for Bull Trout (Decker and Hagen 2008). Quantitative estimates of adult densities in lakes are rare, but hydroacoustic surveys have been used (McPhail and Baxter 1996). The diel differences in habitat use by adfluvial populations needs to be taken into consideration when selecting sampling locations and techniques.

Electro-fishing and snorkeling surveys are most commonly used to estimate juvenile Bull Trout densities. As with all surveying techniques, their potential drawbacks make them vulnerable to bias. The diel and seasonal shifts in habitat use by juvenile Bull Trout, in particular, will affect the density of fish in sampling locations and the effectiveness of these techniques (Jakober et al. 2000). Juvenile Bull Trout’s preference for cover during the day makes it difficult to assess their populations from daytime surveys (Jakober et al. 2000). Size selection and variable capture efficiency suggest electro-fishing based estimates are more biased than those from night time snorkeling surveys (Decker and Hagen 2005).

Abundance

In order to review current abundance, trends and conservation status of current Bull Trout populations, core area assessments have been conducted in Alberta (Rodtka 2009; Appendix 1, Figure 11) and British Columbia (Hagen and Decker 2011; Appendix 2). These assessments of core areas, which are analogous to meta-populations, used modifications of the methodology employed by the US Fish and Wildlife Service for their Bull Trout core area analysis in the USA (Fredenberg et al. 2005). Briefly, a combination of empirical data and expert opinion were used to determine the most likely population abundance using approaches developed by Master et al. (2003). This approach was adopted by the US Fish & Wildlife Service and applied to Bull Trout across core areas throughout their US range (Fredenberg et al. (2005) and Alberta (Rodtka (2009) and BC (Hagen and Decker 2011) followed the same protocol.


Figure 11. Spatial distributions of Bull Trout core areas in Alberta and their conservation ranking

Map of the spatial distributions of Bull Trout core areas in Alberta (see long description below).

Assessment was performed by the Fish and Wildlife Division of Alberta Sustainable Resource Development and is based upon a modification of the Natural Heritage Network ranking methodology using NatureServe Conservation Status Assessment Criteria. Extirpated core areas are not shown. Figure prepared by Velma Hudson (Alberta Conservation Association) and sourced from Rodtka (2009).

Description of Figure 11

Map of the spatial distributions of Bull Trout core areas in Alberta and their conservation ranking: high risk; at risk; potential risk; and unranked. The Bull Trout’s historical range is indicated.

Several factors were considered when defining the 51 Bull Trout core areas that have been identified in Alberta (Appendix 1, Figure 11), including: historical distribution; abundance of adult fish; barriers to movement, and; the probability of permanently losing (or likelihood of natural re-establishment if extirpated) a population (Girard pers. comm. 2010). A comprehensive assessment that estimated the abundance of adults within each of the 51 Albertan Bull Trout core areas was based on available data (e.g., population estimates, reconnaissance inventories, fish trapping results, redd surveys; Girard pers. comm.2010). After data compilation, density estimates were extrapolated from specific habitats to the area of occupancy within a core area (assuming same habitat quality). Some core areas had no data, in which case abundance was estimated as the median from range categories (Fredenberg 2005; Girard pers. comm. 2010).

The delineation of the 115 Bull Trout core areas that have been identified to span 26 Ecological Drainage Units (EDUs) as defined for British Columbia (Appendix 2) and described in full by Hagen and Decker (2011). Briefly, core areas were established using the following guidelines: they contain or have the potential to contain multiple, interconnected local populations; be typically 100-250 km along their longest dimension unless further restricted by migration barriers (or if they can be estimated more reliably from telemetry/genetic studies); provide all critical habitat elements, and; be distributed within the known range of the species in the Province. Expert opinion was then canvassed from a range of biologists to estimate distribution, abundance of mature individuals, trends in abundance, and threats within each of these putative core areas. It is important to note that most of these core areas are considered provisional as availability of data describing Bull Trout distribution, population structure, movement and barriers varies significantly among areas (Hagen and Decker 2011).

The genetic population structure of the vast majority of these core areas has not been defined in either Alberta or British Columbia. Given that genetic differentiation has been detected among Bull Trout populations at the fine scale e.g., within watersheds over distances as small as a few kilometers (Spruell et al. 1999; Taylor et al. 2001; Costello et al. 2003; Taylor and Costello 2006), it is possible, or even likely in some instances, that the number of genetically distinct Bull Trout populations exceeds the core areas identified so far within each DU. A summary of our current knowledge for each of the Canadian DUs is outlined below.

DU1 [Genetic Lineage 1: Southcoast BCpopulations]

Bull Trout populations from this DU are restricted to British Columbia: only three of the 26 EDUs occupied by Bull Trout within this province are known to contain Genetic Lineage 1 Bull Trout (Appendix 2). And, of the 115 provisional Bull Trout core areas in this province, just five have been identified with a reasonable level of certainty within this DU (Appendix 2). Some short-term monitoring data exists for three populations (Skagit River, Phelix Creek and Cheakamus River) that represent three of these provisional core areas; in all cases, well over 100 spawners have been counted in most years of monitoring (Table 1). In fact, the most recent snorkel count in the Skagit River (2010) estimated over 1500 adults were present. Based on this information, expert opinion estimates that several thousand spawners (1,000-2,500 or more) may be present in this DU (Hagen and Decker 2011).

Table 1. Summary of 31 adult Bull Trout abundance datasets compiled from 22 core areas from 12 of 26 Bull Trout Ecological Drainage Units (EDUs) identified in British Columbia. Trend data (simple regression analysis) available for 23 of these datasets covering 11 Bull Trout EDUs (datasets with more than five years of data collected using a consistent methodology). Modified from Hagen and Decker (2011). Accessible version of Table 1
Core area Stream or lake No. years data Estimated abundance Short-term trend
DU1[Genetic Lineage 1: Southcoast BCpopulations]
EDU Lower Fraser
Lillooet Phelix 5 27-185 no
EDU Puget Sound
Skagit Skagit 5 159-1650 positive (P=0.03)
EDU South Coastal
Squamish Cheakamus 13 75-316 multiple trends
DU2 [Genetic Lineage 2: Western Arctic populations]
EDU Upper Peace
Finlay Reach Davis 9 37-85 no
Parsnip Reach Misinchinka 5 35-58 no
  Scott 2 58-106 Unknown
Peace Reach Point 5 5-39 no
Thutade Thutade Lake 16 122-288 positive (P=0.01)
EDU Lower Peace
Halfway-Peace Chowade 6 55-864 positive (P=0.01)
  Needham 3 52-103 Unknown
  Cypress 3 18-120 Unknown
Lower Murray Wolverine 3 25-67 Unknown
DU5 [Genetic Lineage 2: Pacific populations]
EDU Columbia-Arrow
Pend d'Oreille Salmo 12 38-109 no
ALR all Arrow Lakes 23 0.02-0.13
CPUE (fish/hr)
no
ALR southern Arrow tribs 2 198-260 Unknown
ALR northern Arrow tribs 2 586-755 Unknown
EDU Lower Kootenay
Kootenay Lake Irishman 8 13-32 no
  Duncan 9 202-725 no
  Kaslo 5 716-1219 no
  Crawford 3 336-486 Unknown
  Kootenay Lake 34 0.02-0.15
CPUE (fish/hr)
multiple trends
EDU Upper Kootenay
Elk Line 19 28-184 positive (P=0.001)
Upper Kootenay R Skookumchuck 14 64-189 no
  White 10 93-193 no
Koocanusa Wigwam 17 105-2298 multiple trends
EDU Upper Skeena
Upper Sustut Sustut 19  3-70 negative (P=0.04)
Mid-Skeena Kitwanga 7 31-495 no
Lower Sustut/ Skeena Damshilgwet 11 22-302 positive (P=0.01)
EDU Upper Fraser
Upper Fraser Goat 5 55-163 no
EDU Middle Fraser
Chilko Long Valley 2 433-693 Unknown
EDU Thompson
Upper Shuswap Sugar Lake 4 0.01-0.26
CPUE (fish/hr)
*positive(P=0.02)

* only four years of data, but included in trend analysis as data spanned a 20-year time period.

DU2 [Genetic Lineage 2: Western Arctic populations]

Bull Trout populations from the vast area of the Mackenzie River drainage basin are found in two Canadian provinces and 2 Territories that harbor this species. Of the 51 Bull Trout core areas that have been identified in Alberta, 15 fall within this DU (in the Athabasca and Peace-Smoky River basins; Appendix 1, Figure 11). Approximately 23,000 adult Bull Trout are estimated to inhabit Alberta’s lakes and streams in this DU within the Athabasca and Peace-Smoky River basins (Appendix 1). The mean population size for these 15 core areas is 1545 but population sizes vary widely (standard deviation 1960); extant populations range in size from 25 adults for the Peace River to 7450 for the Kakwa River, both from the northern Peace-Smoky River basin (Appendix 1).

Of the 115 provisional Bull Trout core areas estimated in British Columbia, 30 occur within this DU across four EDUs (Appendix 2). There is, however, significant uncertainty regarding the number of core areas within at least one of these EDUs (Upper Liard; Hagen and Decker 2011). Some monitoring data exists for nine populations that represent six provisional Bull Trout core areas from the Upper and Lower Peace EDUs (Table 1). The Thutade Lake inlets Bull Trout population has been subject to the longest period of monitoring, 16 years. The most recent survey in 2009 recorded 235 redds (Table 1, Hagen and Decker 2011). Although redd counts have varied across time and systems among the eight remaining populations from five other provisional Bull Trout core areas that have received short-term monitoring, fluctuations have been restricted to at or below 100 redds annually with one exception; the Chowade River, where estimates in 2010 exceeded 800 redds (Table 1, Hagen and Decker 2011). There is no information on abundance from either the Upper or Lower Liard EDUs (Hagen and Decker 2011). Expert opinion could only consider 5 of the 30 provisional core areas, with a total estimate of 5,000-10,000 adults. The abundance for 25 provisional core areas remains unknown (Hagen and Decker 2011).

Little is known about the number or size of Western Arctic Bull Trout populations in the Northwest Territories, where recent surveying is only now establishing the northern range of this species distribution. Two recent surveys (electro-shocking, angling and set lines) of 29 streams in the southern (Deh Cho) and central (Sahtu) Northwest Territories found Bull Trout represent 1% (Mochnacz and Reist 2007) and 4% (Mochnacz et al. 2009) of the total catch, respectively. This is in line with the general observation that Bull Trout typically comprise less than 5% of the total catch from broad faunal surveys (reviewed in McPhail and Baxter 1996). Given that productivity generally decreases with increasing latitude due to colder temperature and shorter growing seasons, the initial indication of small but wide ranging populations is likely an accurate reflection of Bull Trout populations in their northern reaches.

The single population estimate that is available for Bull Trout from the Northwest Territories comes from Funeral Creek, a suspected resident headwater population (Department of Fisheries and Oceans pers. comm.2010). Here, four randomly selected reaches (~200 m) were surveyed using electro-shocking (Mochnacz et al. 2006). Maximum-likelihood population size estimates at 95% confidence intervals showed adult ranges (N= 17 [95% CI 16-18] and 21 [95% CI 18-23]) to be similar to those for juveniles (17[95% CI 16-18] and 23 [95% CI 20-28]; Mochnacz et al. 2006). This suggests that both groups have small populations compared to other fish species, typical of Bull Trout populations throughout their range.

In southeast Yukon, Bull Trout are known to occur in numerous drainages and lakes of the Liard River (Can-nic-a-nick Environmental Sciences 2004) and are likely widespread in this drainage basin (Miller pers. comm. 2010). These northern populations are likely to be small in size, although there is currently no information available on the number or size of these Western Arctic Bull Trout populations.

DU3 [Genetic Lineage 2: Yukon River Watershed populations]

Bull Trout populations from the Yukon River watershed are believed to be found in both Yukon and British Columbia, although there is very little information on their distribution (see ‘Distribution’ section).

DU4 [Genetic Lineage 2: Saskatchewan-Nelson Rivers populations]

Bull Trout populations from this DU are restricted to Alberta. Of the 51 Bull Trout core areas that have been identified in this province, 36 fall within this DU (Appendix 1, Figure 11). Approximately 10 000 adult Bull Trout are estimated to inhabit Alberta’s lakes and streams in this DU within the Oldman, Bow, Red Deer and North Saskatchewan River basins (Appendix 1). Population sizes for these 36 core areas in southern Alberta tend to be smaller than those from the more northerly Western Arctic populations in Alberta; the mean population size for these 36 core areas is 300. Once again, however, these population sizes vary widely (standard deviation 368); extant populations range in size from 10 adults for the Middle Bow River from the southern Bow River basin, to 1275 in the Brazeau River (Appendix 1).

DU5 [Genetic Lineage 2: Pacific populations]

Bull Trout populations from this DU are widespread throughout British Columbia; they occur in the majority (n = 78) of the 115 provisional Bull Trout core areas identified here, and are spread across 17 EDUs (Appendix 2). Although a number of short and longer-term monitoring initiatives for Bull Trout have been undertaken within this DU, the majority (n = 13) of the 19 abundance datasets available occur within the Columbia drainage (Table 1). These 13 datasets within the Columbia drainage fall across seven provisional Bull Trout core areas in three EDUs. Each of these EDUshas at least one population whose abundance estimates remain below 200 (Table 1). Nevertheless, they all also harbor populations whose abundance estimates consistently exceed this (Table 1). Little information about abundance exists for north coastal watersheds, the Thompson River, or the mid- and upper Fraser River within this DU (Table 1). Expert opinion estimated to be much more than 39,000 adults in this DU, but this was limited to 25 provisional core areas where some abundance information was available (Hagen and Decker 2011).

Effective Population Size

Based on a generalized, age-structured simulation model that incorporated a range of life histories and other conditions characteristic of Bull Trout populations, the effective population size for Bull Trout has been estimated to be approximately 0.5 to 1.0 times the average number of adults spawning annually in a population (Rieman and Allendorf 2001). Achieving the recommendation that Bull Trout populations should include an average of at least 1,000 adults spawning each year for long-term management goals (Rieman and Allendorf 2001) will be challenging given that many Bull Trout populations tend to be smaller (data herein; Rieman and McIntyre 1993). Conserving interconnected populations as groups may be a strategy that can meet this suggested minimum, while simultaneously providing for the full expression of life history variation and the natural processes of dispersal and gene flow (Rieman and Allendorf 2001).

Fluctuations and Trends

In recent decades, Bull Trout populations have experienced declines in abundance across their range but particularly in southern and eastern parts in the USA (Rieman et al. 1997; USFWS 1999, 2008) and Alberta (Rodtka 2009). For the most part, this range reduction is comprised of localized extinctions, although it is known to have become extinct in two systems in the USA (McCloud, California; Willamette, Oregon; McPhail and Baxter 1996). The status of Bull Trout populations appears to show a general north to south trend, with decreasing abundance towards its southern margins (Haas and McPhail 1991; McPhail 2007). This trend is likely due, at least in part, to the more pristine and suitable environments in northerly regions (Haas and McPhail 1991).

In addition to this general trend of declining abundance, there is evidence to suggest that the full range of life histories is also being lost from populations. There is particular concern that migratory Bull Trout may be especially susceptible to declines in larger, highly fecund, individuals (Nelson et al. 2002; Post et al. 2003). For example, large-bodied fluvial or adfluvial Bull Trout were common in southwestern Alberta prior to 1950, but many extant populations are now comprised of small-bodied residents that only occupy a fraction of their former range (Fitch 1997). It has also been noted that adfluvial Bull Trout populations in the upper Columbia Basin frequently include individuals that are larger and older than those found in more southerly populations, suggesting these more northerly populations experience less exploitation as well as lower growth rates (Hagen 2008).

Although this general pattern of decline in abundance is clear, two factors make it difficult to assess its extent in populations. Firstly, broad natural fluctuations in abundance (Paul et al. 2000) make it difficult to assess population trends over short periods of time. This natural variation, combined with the limitations of surveying methods (Rieman and Myers 1997; Dunham et al. 2001; Al-Chokhachy et al. 2009), means that considerable resource and temporal commitments are required to detect moderate changes in abundance of Bull Trout. Evidence from long-term studies conducted in the USAsuggests that more than a decade of Bull Trout monitoring may be required to detect a large population decline statistically (Rieman and Myers 1997; Al-Chokhachy et al. 2009). Given this sensitive species’ tendency towards naturally low population sizes, it may not be possible to prove significant trends for many monitored Bull Trout populations before they drop below a critically low level (Rieman and Myers 1997).

Despite these hurdles, monitoring is often proposed as a mechanism to assess trends in abundance in order to recognize and mitigate land management effects. Standardized quantitative information gathered from a number of Bull Trout populations over a period of decades will be necessary for a thorough evaluation of the trends and status of Bull Trout in each DU. However, few long-term monitoring efforts on the abundance of Bull Trout exist in Canada, and the limited long-term quantitative data that is available is often confounded by non-standardized sampling techniques. Much of our current knowledge of population trends actually relies on qualitative expert opinion (Rodtka 2009; Girard pers. comm. 2010; Hagen and Decker 2011).

Nevertheless, more monitoring and baseline assessments are now being established. For example, 31 abundance datasets have been compiled for Bull Trout populations in British Columbia (which have two or more years of information collected using a consistent methodology; Table 1). Of these datasets, 15 have at least seven years of data (which is a reasonable approximation of one generation for Bull Trout in British Columbia; Westover and Conroy 1997), seven have 14 or more years of data (i.e. two generations), and two have 21 or more years of data (i.e. three generations). Coupled with core area assessments (Rodtka 2009; Hagen and Decker 2011), abundance assessments such as these should provide useful data to assess trends in population size for future Bull Trout status assessments. Even so, substantial gaps remain. For example, the 31 abundance datasets compiled for Bull Trout populations in British Columbia are found in just 22 of the 115 provisional Bull Trout core areas, spanning only 12 of the 26 Bull Trout EDUsidentified in this province (Table 1).

A summary of our knowledge about trends in Bull Trout populations for each of the Canadian DUs is outlined below.

DU1 [Genetic Lineage 1: Southcoast BCpopulations]

Short-term trend data are available for three local populations in this DU (Skagit River, Phelix Creek and Cheakamus River, Table 1). They are from three of five identified provisional Bull Trout core areas, and represent each of the three EDUs within this DU. A positive trend was observed for the Skagit River with most recent counts (2010) almost six times greater than earliest counts (1998); this trend largely reflects population recovery following the implementation of more restrictive fishing regulations (Hagen and Decker 2011). A similar response was observed in the Cheakamus dataset until 2006, a year after a caustic soda spill into the river. Since this time, adult numbers steadily declined but most recent counts (2010-2011) indicate the population is increasing (Hagen and Decker 2011). A similar response to altered angling regulations was not observed in Phelix Creek (where no trend was observed), and further regulatory restrictions may still be required (Jesson pers. comm. 2011).

An expert opinion assessment found that trend varied between each of the five provisional core areas in this DU, ranging from increasing to stable, to decreasing and unknown (Appendix 2). In summary, no consistent trend is apparent from either the limited quantitative data or expert opinion assessment for this DU; status appears to vary by major watershed according to local pressures and threats.

DU2 [Genetic Lineage 2: Western Arctic populations]

Anecdotal information and limited historical records suggest that there has been a large decline in the abundance (as well as distribution) of Bull Trout in all river systems in Alberta where it has been found since the early 1900s, including in the drainages of the Peace and Athabasca rivers that fall within this DU (Rodtka 2009, Figure 6). As is the case in the USA (Rieman et al. 1997), most Albertan self-sustaining populations are now restricted to less accessible headwater areas (Rodtka 2009).

This historical pattern of decline is mirrored in today’s short-term trends in Alberta (Appendix 1). Although current population size estimates vary widely across the 15 Bull Trout core areas in this DU, their short-term trends are dominated by declines (N = 11, 73%; Appendix 1). Only three (20%) of them are considered to be stable and one to be increasing (6.7%; Appendix 1). These trends have been based on both quantitative data (multi-year abundance estimates) and qualitative, expert opinion using a modification of the Natural Heritage Network ranking methodology using NatureServe Conservation Status Assessment Criteria (Rodtka 2009; Girard pers. comm. 2010).

While monitoring and baseline assessments have now been established in all river systems in Alberta where Bull Trout is found, including the drainages of the Peace and Athabasca rivers within this DU, there is currently a lack of long-term trend data. Most monitoring efforts have been applied to populations in the southwest of Alberta, within DU4 [Genetic Lineage 2: Saskatchewan-Nelson populations](reviewed in Rodtka 2009). An exception to this is Eunice Creek in the Athabasca River drainage, whose monitoring reveals valuable insight into Bull Trout population dynamics under relatively unaltered conditions. Closed to angling since 1966 and protected from most development until 1985 (Hunt et al.1997), the abundance of Bull Trout fluctuated here by two orders of magnitudes over just fifteen years (Paul et al. 2000), reflecting the broad natural fluctuations that may occur in Bull Trout abundance.

 There is evidence that some less closely monitored adfluvial populations appear to be increasing as a result of conservative angling regulations in Alberta, including Pinto Lake within this DU (reviewed in Rodtka 2009). The trend for resident and fluvial populations in this province is less consistent. Within this DU, electro-fishing and angling surveys of Kakwa River have found no evidence of change in abundance in this system since the provincial wide, zero-harvest regulation implemented in 1995 and the total closure of angling on Lynx Creek, the key spawning stream for Bull Trout from the Kakwa River (reviewed in Rodtka 2009).

Adult trend data for Bull Trout populations in British Columbia exists for five populations in this DU (Table 1). They are from five of 30 identified provisional Bull Trout core areas. Four are found within the Upper Peace EDU, and one within the Lower Peace EDU but there is no trend data from either the Upper or Lower Liard EDUs (Hagen and Decker 2011). As in Alberta, there is a case of a previously exploited British Columbian Bull Trout population that has expanded once threats have been mitigated; the explosive increasing trend observed from six years of redd count data over 15 years for the Chowade River in the Lower Peace EDUis considered to represent the recovery of a depleted population following the implementation of more restrictive angling regulations in the 1990s (Hagen and Decker 2011). The more modest increase in spawner numbers observed during 16  years of monitoring of Bull Trout from the Thutade Lake watershed (assumed to represent a meta-population) from the Upper Peace EDU may simply reflect a range of normal variation, although compensation measures associated with the Kemess open-pit copper and gold mine (including fishway construction, spawning habitat creation and the removal of impassible beaver dams) have likely been beneficial (Bustard and Associates 2010). Angling regulation restrictions have likely had less impact on this remote watershed, which is subject to relatively little angling effort (Hagen and Decker 2011). The other three local populations within this EDU (Davis River, Misinchinka and Point), which are tributaries of Williston Reservoir, were more or less stable over time.

No data exists, however, for the large majority of provisional core areas regarding either abundance or distribution for Bull Trout. In summary, stable or increasing trends are evident from the limited quantitative data and expert opinion assessment although the status of the majority of provisional Bull Trout core areas in British Columbia remains unknown.

There is no information available on the trend of Bull Trout populations in either the Northwest Territories, where recent surveying is only now establishing the northern range of this species distribution (Mochnacz et al. in review), or Yukon, where its distribution remains unclear. However, these northerly populations (which generally inhabit less productive habitat than their more southerly counterparts) are likely to be smaller, and hence more susceptible to perturbations, than those found further south. Indeed, Bull Trout are thought to be the most sensitive species in the upper Liard River basin (Can-nic-a-nick Environmental Sciences 2004).

DU3 [Genetic Lineage 2: Yukon River Watershed populations]

There is no information available on the trend of Bull Trout populations within this DU.

DU4 [Genetic Lineage 2: Saskatchewan-Nelson Rivers populations]

Historical patterns of decline in Bull Trout populations in Alberta have been most severe in southern areas of the province within this DU, with many areas in the South and North Saskatchewan River basins no longer supporting Bull Trout (Rodtka 2009). Brook Trout introductions in southwestern Alberta are thought to have contributed to this trend, where about 70% of native Bull Trout populations have been extirpated (Fitch 1997).

This historical pattern of decline is mirrored in today’s short-term trends in Alberta (Appendix 1). Although current population size estimates vary widely across the 36 identified Bull Trout core areas in this DU, the short-term trends of extant populations are dominated by declines (N = 19, 53%; Appendix 1). Fourteen are considered to be stable or increasing (48%; Appendix 1). This general pattern of decline is particularly pronounced in the south; in the South Saskatchewan River basin, both core areas in the Red Deer River basin are declining, 11 of the 15 in the Bow River basin are either extirpated or declining, as are five of the ten from the Oldman River basin.

Most of the monitoring efforts that have been established in Alberta have been applied to populations within this DU(reviewed in Rodtka 2009). The most comprehensive data set exists for the adfluvial population in Lower Kananaskis Lake, which has been monitored annually for adult numbers in a spawning creek from trapping data and redd counts over 12 years (Johnston et al. 2009). These data show a rapid recovery of a previously heavily exploited population since the introduction of strict angling regulations in 1992; a population low of fewer than 100 adults increased almost 28 fold by 2000 (Johnston et al. 2009). Other less closely monitored adfluvial populations also appear to be increasing as a result of conservative angling regulations (e.g., Jacques and Harrison Lakes) (reviewed in Rodtka 2009), although the trend for resident and fluvial populations in this DU is less consistent. There is no indication of change in some other rivers (e.g., Elbow and Highwood rivers, and Quirk Creek), although others do seem to be increasing (Clearwater and Sheep rivers). Lack of consistent methodology and long intervals between some assessments, however, inhibit robust interpretations (reviewed in Rodtka 2009).

DU5 [Genetic Lineage 2: Pacific populations]

Although this DU contains more provisional core areas for Bull Trout than any other DU in British Columbia (78 of 115 total), the number of adult trend datasets is very limited and mostly short-term; there are only 15 datasets representing just 12 provisional core areas from six of the 17 EDUs believed to harbor Bull Trout from this DU (Table 1). The majority (n = 10) of these datasets are from the Columbia drainage (Table 1).

The positive trend observed in Line Creek within this drainage reflects population recovery following the implementation of more restrictive fishing regulations (Hagen and Decker 2011). A strong positive trend from 1994-2006 in the Wigwam River from the Columbia drainage was followed by a decline (considered acceptable and within routine management zone for this healthy population). This reflects changes in regulations; a limited harvest has occurred in recent years (i.e. 2004-2010) following a decade of more restrictive fishing regulations (Hagen and Decker 2011). Some other systems within the Columbia drainage exhibited more or less stable trends (e.g., Salmo River watershed and Upper Kootenay River), while datasets from two other regulated systems (Kootenay Lake and Arrow Lake Reservoir) reflect more complex patterns, showing both positive and negative trends (Table 1). Fluctuations over a decade in Duncan Reservoir within the Kootenay Lake core area are thought to reflect nutrient levels and response to forage base (i.e. Kokanee; Hagen and Decker 2011). Trends in a long-term (nearly 50 yrs) catch-per-unit effort (CPUE) dataset from Kootenay Lake (assumed to represent a meta-population) also track nutrient loading patterns influenced historically by a fertilizer plant and its closure, and the damming of rivers, and more recently by an annual whole-lake fertilization program. However, CPUE survey design varied over time and may not be sensitive to decreases in abundance (Hagen and Decker 2011). Arrow Lakes Reservoir CPUE data, which is also assumed to represent a meta-population, is similarly varied and responsive to manipulations associated with nutrient additions, but is considered more or less stable over the past three decades (Hagen and Decker 2011).

There is little information available for north coastal watersheds, the Thompson River, or the mid- and upper Fraser River within this DU (Table 1). Within the Upper Skeena EDU of the north coast, there are three datasets from salmon counting fences; one (Sustut River) indicates a consistent negative trend, one (Damshilgwet Creek) a positive trend, and the other no trend. However, there is concern that these observations may be, at least in part, artifacts of a methodology (Hagen and Decker 2011).The positive trend observed from CPUE data in Sugar Lake in the Upper Shuswap drainage of the southern interior Thompson EDU (assumed to represent a meta-population) is considered to be a response to the implementation of more restrictive fishing regulations (Hagen and Decker 2011). Within the Upper Fraser EDU, population trend can only be evaluated over the relatively short-term (5 years) for one system, the Goat River, where abundance appears stable.

An expert opinion assessment found that trend appears to vary by major watershed according to local pressures and threats, ranging from increasing to stable, to decreasing and unknown (Appendix 2). In summary, no consistent trend is apparent from either the limited quantitative data or expert opinion assessment for this DU. Given the lack of quantitative data for most core areas within this DU, however, it would be inappropriate to consider existing trend data as representative of larger geographic areas. That said, the greatest concerns occur in the Flathead, Pend d’Oreille and Columbia (downstream of Arrow Lakes Reservoir) rivers where expert opinion considers both low abundance and declining trends to be of significant concern in all three core areas (Hagen and Decker 2011). In contrast, other core areas in the Columbia basin (upper Kootenay River and Kookanusa) are considered to be stable to increasing with large numbers of adults.

Rescue Effect

In theory, a diminished or extirpated population of Bull Trout could experience a rescue effect from neighbouring populations, be they within Canada or from the USA. The potential for such a rescue effect will, however, depend on several factors, including the amount of migration between populations, the viability of immigrants in their new environment and the status of neighbouring populations.

Genetic studies indicate low levels of gene flow between populations. Significant genetic differentiation among Bull Trout populations is common even within watersheds (Spruell et al. 1999; Taylor et al. 2001; Costello et al. 2003; Taylor and Costello 2006), although the degree of divergence is more pronounced at a more regional scale (Taylor et al. 2001; Costello et al. 2003; Whiteley et al. 2004; Taylor and Costello 2006). The Bull Trout’s typically strong site fidelity to spawning area and overwintering habitat revealed by radiotelemetry studies (Swanberg 1997a; Bahr and Shrimpton 2004) further suggests that migration between populations is low. This diminishes the likelihood of immigration providing a significant rescue effect for Bull Trout populations. Significant dispersal between watersheds seems particularly unlikely, although some evidence of straying at the local level (Swanberg 1997a; O’Brien 2001; Bahr and Shrimpton 2004) and at least one account of dispersal between watersheds (Brenkman and Corbett 2005) does suggest a potential role for dispersal from nearby sources in the repopulation of a declining or extirpated population.

While local adaptation of Bull Trout will reduce the viability of immigrants in new environments (Nosil et al. 2005) and diminish the possibility of rescue effects from neighbouring populations, phenotypic plasticity may counterbalance this to some extent. Divergence in quantitative traits will likely be most evident across different environments at larger scales, for example, among populations inhabiting the different DUs. However, local adaptation may exist even at the fine scale, given that microsatellite-based differentiation, which likely provides conservative estimates of adaptive divergence (Pfrender et al. 2000; Morgan et al. 2001), has commonly been detected among populations within localized areas (Spruell et al. 1999; Taylor et al. 2001; Costello et al. 2003; Taylor and Costello 2006).

Any rescue effects to be had will, therefore, most likely occur between close, adjacent populations that are connected by contiguous habitat suitable for Bull Trout migration. This could include several watersheds that have transboundary Bull Trout populations, such as the Flathead River, upper Kootenay River, Kootenay Lake, Salmo River, Skagit River and Chilliwack watershed. The direction of any transboundary rescue effect is most likely to be from Canadian populations to USA waters because Canadian Bull Trout are likely much more numerous in both the number of populations and abundance than their USA counterparts. The vast majority of Bull Trout’s range occurs in Canada (Rieman et al. 1997) and Canadian Bull Trout populations are generally considered to be more stable than Threatened (USFWS 2008) populations in the USA. The majority of Bull Trout populations in the northern USA range are considered to be depressed (Rieman et al.1997). With very few strong or protected populations near the US-Canada boundary (Rieman et al. 1997), it is very unlikely that a USA Bull Trout population could contribute to a rescue effect for a Canadian one.

Page details

Date modified: