Resource use by threatened, native Bull Trout (Salvelinus confluentus) in the presence of two non‐native salmonids

Abstract

Introduction of non‐native species contributes to the global decline of freshwater fishes by disrupting the diet and resources of native species through competition. Despite the potential for non‐native trout (Family: Salmonidae) to compete with native trout, non‐native trout species have been widely introduced worldwide to support recreational fishing. In this study, we examined resource use by native Bull Trout (Salvelinus confluentus), a Threatened species, and non‐native Brook Trout (Salvelinus fontinalis) and Brown Trout (Salmo trutta) in headwater streams in Alberta, Canada. Our objective was to investigate potential for competition and its implications for Bull Trout populations. To assess changes in diet, we used stable isotope analysis. Brook Trout and Brown Trout had higher trophic positions (3.49 ± 0.02 and 3.53 ± 0.02) and greater proportions of terrestrial resources in their diets (0.53 ± 0.02, 0.54 ± 0.03) than Bull Trout (TP = 3.38 ± 0.02, proportion terrestrial = 0.36 ± 0.02), suggesting the non‐native species may outcompete Bull Trout for preferred resources and larger prey. Moreover, a relatively wide isotopic niche of Bull Trout led to a lack of niche partitioning and substantial overlap in niche utilization between sympatric Brook Trout and Bull Trout, supporting the potential for future niche displacement, particularly in scenarios where resources are scarce. These findings underscore the importance of monitoring and managing the invasion of non‐native salmonids in streams where native salmonid populations exist, and the challenges in understanding multi‐species interactions in dynamic habitats. It highlights the need for effective fisheries management and conservation efforts in regions where native freshwater fishes are facing population declines.

1. INTRODUCTION

Aquatic ecosystems face numerous stressors, and one of the primary threats to freshwater biodiversity is the introduction of non‐native species (Dudgeon et al., 2006; Reid et al., 2019). Introduction of non‐native species has caused native species to decline in terms of their range and abundance (Hill & Lodge, 1999); by introducing harmful pathogens (Youngson et al., 1993); modifying habitat (Parkos III et al., 2003; Ruppert et al., 2017); and can ultimately lead to extirpations (Lowe et al., 2000). These alterations not only reduce biodiversity but also have significant impacts on trophic structure and dynamics, which are essential for ecosystem functioning (Eby et al., 2006; Elmqvist et al., 2003; McMeans et al., 2015; Moore & Hunt, 1988; Sakai et al., 2001).

Understanding effects of non‐native species on native species is complicated by food web and habitat dynamics and the particular species involved. A good example of this phenomenon is the interaction between trout species (family: Salmonidae). Trout are among the most widely introduced species globally, mainly to support recreational fishing (Crawford & Muir, 2008). Despite this popularity, understanding potential impacts of introducing non‐native trout on native species is filled with uncertainty. For example, Brook Trout (Salvelinus fontinalis) and Brown Trout (Salmo trutta), which are native to North America and Europe, respectively, have displaced each other in their respective native ranges (Fausch & White, 1981; Korsu et al., 2007; Waters, 1983). Both species have been introduced globally, impacting native salmonids in regions such as Japan (Hasegawa, 2017; Kitano, 2004; Takami et al., 2002), New Zealand (Jones & Closs, 2017; Townsend, 1996) and other locations (e.g., McDowall et al., 2001; Pascual et al., 2007).

In Alberta, Canada, Bull Trout (Salvelinus confluentus) populations have undergone dramatic decline, highlighting growing conservation concern. For example, Bull Trout have declined by more than 80% of their historical distribution, leading to their listing as a Threatened species (COSEWIC, 2012). Whereas multiple factors contribute to their decline, including increases in stream temperature (Kovach et al., 2019), habitat loss (COSEWIC, 2012) and reduced connectivity (Maitland, Anderson, & Poesch 2016; Maitland, Poesch, et al., 2016), competition with Brook Trout and Brown Trout is considered a primary driver of decline (COSEWIC, 2012; Rodtka et al., 2009; Sawatzky, 2016). Studies have identified Brook Trout as a threat through hybridization (Kanda et al., 2002), interspecific dietary aggression (Gunckel et al., 2002; McMahon et al., 2007; Warnock & Rasmussen, 2013b) and displacement (Paul & Post, 2001; Rieman et al., 2006; Warnock & Rasmussen, 2013a).

Previous studies investigating competitive dietary interactions between Bull Trout and Brook Trout and Brown Trout have yielded diverse results, suggesting potential utility of a conceptual framework for co‐occurring salmonids that considers two features of the diet, trophic position (TP) (inferred from stable nitrogen isotopes, δ¹⁵N) and % terrestrial feeding (inferred from stable carbon isotopes, δ¹³C) (Figure 1). Preferred food sources for drift‐feeding salmonids are high‐quality, large‐bodied terrestrial invertebrates (Baxter et al., 2007; Marcarelli et al., 2011) that, although they lack long‐chain omega‐3 polyunsaturated fatty acids (Hixson et al., 2015), are energy dense and tend to peak in availability during the summer growing season for trout species (Nakano & Murakami, 2001). As such, we would expect superior competitors to have higher % terrestrial feeding and a higher TP. Laboratory studies have shown that sympatric Brook Trout exhibit increased interspecific antagonism and outcompete Bull Trout, particularly in warmer temperatures, resulting in a significant growth advantage (McMahon et al., 2007; Rodtka & Volpe, 2007). In‐stream experiments have further observed competitive interactions between Brook Trout and Bull Trout, without any apparent shifts in their respective niches (Gunckel et al., 2002; Nakano et al., 1998). In regard to Brown Trout, which have been introduced to western Canada and occupy a wide range of Bull Trout habitat (MacCrimmon & Campbell, 1969), limited research has been conducted on their feeding interactions with Bull Trout. Nevertheless, experimental studies conducted in Japan show that Brown Trout exhibit competitive superiority over two native salmonid species, White‐spotted Charr (Salvelinus leucomaenis) and Masu Salmon (Oncorhynchus masou) (Hasegawa et al., 2004; Kitano, 2004).

FIGURE 1.

Conceptual framework of predicted effects of non‐native species (red circles) on native species (blue circles) isotopic niches in streams. Solid blue lines enclose the niche area for the native species, solid red lines enclose the niche area for the non‐native species, and dashed lines enclose the combined niche area for two species together. Following introduction of the non‐native species, in scenario (a) (competition), there is no change in the native species' niche area which is occupied fully or partly by the non‐native species, resulting in a similar combined niche area. In scenario (b) (unfilled niche occupancy), native species have a small niche that is not made smaller by non‐native species because the latter feeds on different resources, leading to a larger combined niche area. In scenario (c) (partial displacement), native species' niche area is reduced because non‐native species occupy a portion of the native species' niche. Combined niche area remains the same because the isotopic niche is partitioned. In scenario (d) (full displacement), native species niche area is not reduced but there is a shift in its resource use because the former niche is occupied by non‐native species. Combined niche area increases.

The objective of this study was to investigate potential dietary interactions between Bull Trout, Brook Trout and Brown Trout by examining trophic structure and niche overlap in streams across five different invasion scenarios. These scenarios included (a) streams dominated by Bull Trout, (b) streams dominated by Brook Trout, (c) streams dominated by Brown Trout, (d) streams with a mix of Bull Trout and Brook Trout and (e) streams with a mix of Bull Trout, Brook Trout and Brown Trout. The specific objectives were to

1. determine if there are isotopic niche shifts of Bull Trout in response to varying levels of invasion by Brook Trout and Brown Trout,

2. assess the degree of isotopic niche overlap between Bull Trout, Brook Trout and Brown Trout and

3. infer potential for competition, unfilled niche occupancy, niche shifts and/or resource partitioning between Bull Trout, Brook Trout and Brown Trout.

In this context, competition refers to a high degree of niche overlap between native and non‐native species without an increase in total niche area or change in resource use (Figure 1a). Unfilled niche occupancy occurs when a non‐native species uses resources that were previously unexploited by native species, without affecting the native species' niche (Figure 1b). Displacement is characterized by a native species shifting its resource use from when it is allopatric to when it is sympatric, accompanied by niche partitioning with the non‐native species, which can involve a reduction in isotopic niche size (Figure 1c) or an isotopic shift (Figure 1d).

Based on previous studies examining feeding interactions and behaviours between Bull Trout and Brook Trout (Gunckel et al., 2002; McMahon et al., 2007; Rodtka & Volpe, 2007; Warnock & Rasmussen, 2013b), our hypothesis is that neither Bull Trout nor Brook Trout will undergo a niche shift when they coexist in the same habitat (sympatry), resulting in a high degree of niche overlap and subsequent competition. Conversely, with Brown Trout present, we hypothesize that Bull Trout and Brown Trout will undergo niche shifts to mitigate competitive interactions, leading to niche partitioning (Hasegawa et al., 2004; Kitano, 2004). As a result, we predict a lesser degree of niche overlap between sympatric Brown Trout and Bull Trout than their hypothetical overlap while in allopatry.

2. METHODS

The study took place in the Rocky Mountain foothills of Alberta, Canada (Figure 2), where Bull Trout are designated as Threatened (Fisheries and Oceans Canada, 2020). Stream sites were chosen based on historical electrofishing records to encompass a range of conditions, including sites containing only Bull Trout, sites with Bull Trout and either Brook Trout or Brown Trout and sites with all three (Alberta Environment and Parks, 2022). There were no sites with only a mixture of Brook Trout and Brown Trout or Bull Trout and Brown Trout due to the distributions of these species in this region. In July and August 2020, individuals were captured using backpack electrofishing at 48 sites located within 11 creeks: Yara, Wildhorse, Burnt Timber, Wilson, Pineneedle, Rocky, Cutoff, Elk, Trout, Rough and Fallentimber. The aim was to capture 10–30 individuals of each target species at each creek, as this sample size is statistically robust for stable isotope analysis (Jackson et al., 2011). Electrofishing was conducted in an upstream direction using a Smith‐Root LR24 backpack unit.

FIGURE 2.

Map of the study area. Individual sites noted by white circles within the Red Deer River basin (blue) and North Saskatchewan River basin (green).

Fish were identified, weighed and measured upon capture. To obtain tissue samples for stable isotope analysis, a fin clip was collected from the upper lobe of the caudal fin of fish >65 mm using a sterilized scalpel and forceps. Fin clips were stored in dry 1‐mL microcentrifuge tubes. Non‐lethal caudal fin clips result in minimal disturbance to species and provide no significant differences in δ¹⁵N and δ¹³C isotope measurements than using a more invasive tissue plug (Sanderson et al., 2009). As a minimum of 0.4 mg of dried tissue is recommended for analysis, obtaining a non‐lethal fin clip large enough for stable isotope analysis was not feasible for fish <65 mm, and therefore, no samples were taken from Bull Trout or Brown Trout <65 mm (Sanderson et al., 2009). Brook Trout were the only fish we were allowed to retain under our approved provincial Fish Research Licences (20‐2208 & 21‐2207), and federal Species at Risk Act Permits (20‐PCAA‐00026 & 21‐PCAA‐00031). We euthanized Brook Trout individuals with a 200 ppm buffered solution of Tricaine methane sulphonate (MS222) and then ensured death using single‐blow blunt force trauma. Methods were approved by the University of Alberta Research Ethics Board (AUP00003132_AME2). Euthanized fish were promptly frozen on‐site and stored at −20°C until tissues were later dissected. All fish that could not be retained were released at least 25 m downstream of the start of the next electrofishing transect to avoid recapture.

At the start of each site, we set up two pan traps on each stream bank to collect terrestrial invertebrates. After completing electrofishing surveys, we emptied pan traps onto a fine mesh and transferred terrestrial invertebrates to 2.5‐mL centrifuge tubes. We collected aquatic invertebrates using a 3‐min kick sample from a pool, riffle and run at each site, for a maximum of three kick samples per site. All samples collected were placed on ice at site, held in portable freezers in the field and stored at −20°C.

2.1. Laboratory methods—Stable isotope analysis

Stable isotope analysis is increasingly used to examine effects of non‐native species on trophic structure and dynamics of native species. Stable isotope analysis involves monitoring ratios of naturally occurring isotopes in organism tissues to trace their pathways within a food web (Fry, 2006). Ratios of ¹³C/¹²C (δ¹³C) and ¹⁵N/¹⁴N (δ¹⁵N) are commonly measured as they are associated with specific organic matter sources (e.g., aquatic vs. terrestrial) and TPs, enabling inferences about dietary changes (Jardine et al., 2012; Vander Zanden et al., 1997). When combined with probabilistic models, δ¹³C and δ¹⁵N can assess trophic niches of species and degrees of overlap between species in a system (e.g., Cucherousset et al., 2020). Overall, stable isotope results can provide insights into potential competition, niche occupancy and/or resource partitioning between native and non‐native species (Cucherousset et al., 2007, 2020; Gerig et al., 2018; Jensen et al., 2017) (Figure 1).

For each kick sample of benthic invertebrates, we identified up to 200 individuals down to family, or to the next highest taxonomic rank where family could not be identified, according to Clifford (1991). In instances where families were repeated in pool, riffle and run samples, we consolidated samples into a single group per family per site. Aquatic invertebrates were further categorized into five functional feeding groups: collectors, filterers, scrapers, shredders and predators, based on the classification system described by Merritt and Cummins (1996). Terrestrial invertebrates were identified to order, following Borror and White (1970). As our focus was on lower trophic level organisms, which served as a baseline for calculating percentages of dietary terrestrial organic matter and TP (as described below), we did not analyse predatory invertebrates for stable isotopes.

Fin clips and invertebrate samples designated for stable isotope analysis were freeze‐dried in microcentrifuge tubes for a minimum of 24 h. This was conducted in a LABCONCO® FreeZone 1 Liter Benchtop Freeze Dry System at a temperature of −55°C and a pressure of 0.015 Barr. Following freeze‐drying, tissue was homogenized to a fine powder using dissecting scissors and a stainless‐steel rod in individual microcentrifuge tubes. Approximately 0.4 mg of each sample was weighed into 4 mm × 6 mm tin capsules using a UMX2 Ultra‐microbalance (Mettler Toledo). All samples were analyzed using a Thermo Delta V Advantage Isotope Ratio Mass Spectrometer and reported in standard delta notation (δ¹³C vs. PDB and δ¹⁵N vs. Air). To ensure precision, every twelfth sample was repeated for an internal standard. Mean absolute difference between replicates for δ¹³C and δ¹⁵N was 0.22 and 0.29, respectively.

2.2. Data processing

We categorized streams into five levels of invasion, taking into account the number of individuals captured during electrofishing surveys (Table S1). Categorization was as follows: (1) Bull Trout dominant streams, where Bull Trout constituted over 90% of catch; (2) Brook Trout dominant streams, where Brook Trout made up over 90% of catch; (3) Brown Trout dominant streams, where Brown Trout made up over 90% of catch; (4) Bull Trout/Brook Trout mixed streams, where both species each account for 10% to 90% of catch; (5) Bull Trout/Brook Trout/Brown Trout mixed streams, where each species accounts for 10% to 90% of catch.

We conducted a visual assessment of the data to identify any potential carbonate contamination by examining scatter plots of %C and %N. Additionally, we mathematically checked fish tissue for lipid variation by analysing the C:N ratios, ensuring that values were below 4.0 (Logan et al., 2008). It is important to note that chitin, a common protein in invertebrates, has a higher C:N ratio than typical proteins (Logan et al., 2008). Consequently, the commonly assumed C:N ratio of 3 for lipid free muscle does not apply well to invertebrates, and most published mathematical lipid corrections are not effective on invertebrate tissues (Kiljunen et al., 2006; Logan et al., 2008). Therefore, we chose not to correct invertebrate tissue δ¹³C values for lipids.

2.3. Baseline corrections

To standardize consumer isotope values and account for variation in δ¹⁵N and δ¹³C across waterbodies, it is necessary to establish a baseline reference (Newsome et al., 2007; Syväranta et al., 2013). Since our streams did not have long‐lived primary consumers like snails and clams, we selected scrapers (Trichoptera, Ephemeroptera) as the aquatic baseline due to their strict algal diet (Finlay, 2001) and their low δ¹⁵N values (Anderson & Cabana, 2007). Trichoptera families included Glossosomatidae and Helicopsychidae, and Ephemeroptera families were Baetidae, Ephemerellidae and Heptageniidae. For the terrestrial baseline, we used pooled primary consumer terrestrial invertebrates (Lepidoptera, Thysanoptera, Psocoptera, Homoptera) as their δ¹³C values resemble those of terrestrial vegetation (Finlay, 2001). Due to low invertebrate availability within streams, a single mean value for the aquatic source and a single mean value for the terrestrial source were used for baseline corrections rather than site‐specific means. Subsequently, we standardized δ¹³C values to represent the proportion of carbon derived from terrestrial sources (1−α) using Post's (2002) Equation (1):

$$1-\mathrm{\alpha }=\left(\left({\mathrm{\delta }}^{13}{\mathrm{C}}_{\text{consumer}}-\mathrm{TEF}\right)-{\mathrm{\delta }}^{13}{\mathrm{C}}_{\text{baseline2}}\right)/\left({\mathrm{\delta }}^{13}{\mathrm{C}}_{\text{baseline1}}-{\mathrm{\delta }}^{13}{\mathrm{C}}_{\text{baseline2}}\right),$$ (1)

where α represents the proportion of carbon derived from aquatic sources, δ¹³C_consumer represents δ¹³C of individual fish, δ¹³C_baseline2 represents average δ¹³C of scrapers across all waterbodies and δ¹³C_baseline1 represents average δ¹³C of terrestrial invertebrates across all waterbodies. We set the trophic enrichment factor (TEF) of δ¹³C to 2.3‰ as an average reported for trout species in previous studies (Beltrán et al., 2009; McCutchan et al., 2003; Pinnegar & Polunin, 1999). After correction, some α values lie outside of 0.0–1.0, and we recognize that values outside this range owe to uncertainty in source and TEF values. α values <0.0 were set to 0.0 and α values >1.0 were set to 1.0 to constrain estimates of species means. Constraining α values between 0 and 1 did not affect comparisons among species. δ¹⁵N values were standardized to represent TP using Post's (2002) Equation (2):

$$\mathrm{TP}=\left(\left({\mathrm{\delta }}^{15}{\mathrm{N}}_{\text{consumer}}-\left[{\mathrm{\delta }}^{15}{\mathrm{N}}_{\text{baseline1}}\times \mathrm{\alpha }+{\mathrm{\delta }}^{15}{\mathrm{N}}_{\text{baseline2}}\times \left(1-\mathrm{\alpha }\right)\right]\right)/\left(\text{∆n}\right)\right)+\mathrm{\lambda },$$ (2)

where δ¹⁵N_consumer represents the δ¹⁵N of individual fish, δ¹⁵N_baseline1 represents average δ¹⁵N of scrapers across all waterbodies and δ¹⁵N_baseline2 represents average δ¹⁵N of terrestrial invertebrates across all waterbodies. We set the TEF of δ¹⁵N, enrichment in ¹⁵N per trophic level, to 3.4‰ as per Post (2002). α represents the proportion of carbon derived from aquatic sources (see Equation 1 above) and λ is the TP of the baseline organism (2 for primary consumers).

2.4. TP and proportion of terrestrial carbon

To assess impacts of non‐native species on resource use (α) and TP of Bull Trout, we aggregated streams based on their level of invasion and calculated the mean TP for each species within each invasion level. Using a two‐way analysis of variance (ANOVA), we examined whether there were significant differences in TP between species within invasion levels and between each species across invasion levels. In the global model, we included species and invasion level as predictors and then selected the model with the lowest Akaike Information Criterion (AICc). We visually assessed the best model for normality and homoscedasticity, and if necessary, we log‐transformed the data to meet assumptions. Subsequently, we conducted a Tukey honestly significant difference (HSD) post hoc test to identify significant differences (p value <0.05) between invasion levels and species. We repeated the same process to evaluate effects of invasion level and species on the proportion of carbon derived from terrestrial sources (1−α) and to determine if fork length, known elsewhere to influence (1−α) and TP (Cucherousset et al., 2007; Sánchez‐Hernández & Cobo, 2016), significantly varied between species and invasion levels.

2.5. Standard ellipses area and niche overlap

To assess isotopic niche widths of individual species at each invasion level, we employed standard ellipse area (SEA_B), a Bayesian estimate of isotopic niche width that is suitable for small sample sizes (n <30) (Jackson et al., 2011). We used the R package SIBER (Jackson et al., 2011). Since proxies for δ¹⁵N and δ¹³C, namely, TP and 1−α, allowed for pooling of streams and comparison among invasion levels, we used them in SEA_B calculations. The model was run for 10,000 iterations with two chains, a burn in of 1000 and a thinning rate of 10. To determine if niche width varied between species within invasion levels and for each species across invasion levels, we compared SEA_B 95% posterior credible intervals (Olsson et al., 2009; Syväranta et al., 2013). If the 95% posterior credible intervals did not overlap, they were considered significantly different from one another (Jackson et al., 2011).

To quantify the degree of overlap between species within invasion levels, we employed a Bayesian approach using the R package nicheROVER (Lysy et al., 2021). Niche area, representing a 95% probability region, was calculated for each species at each invasion level. Mean overlap then was calculated in a directional manner and expressed as a percentage, indicating the proportion of species of A within species B's niche and vice versa (Lysy et al., 2021; Swanson et al., 2015). By comparing percent overlap between different levels of invasion, we aimed to identify evidence of competition, displacement or resource partitioning (Figure 1).

3. RESULTS

A total of 610 fin clips were collected from 48 sites within 11 streams (Figure 2). Among these were 222 Brook Trout, 224 Bull Trout and 164 Brown Trout (Table S2). Catch per unit effort (CPUE) varied between streams (Table S1) with average CPUE rates across invasion levels as follows. In Bull Trout dominant streams, mean CPUE was 28 ± 6 fish/h. In Brook Trout dominant streams, CPUE was 41 ± 20 fish/h. In the Brown Trout dominant stream, CPUE was 118 fish/h. In streams with Bull Trout and Brook Trout in sympatry, CPUE was 20 ± 1 fish/h. In streams with all three species in sympatry, CPUE was 41 ± 17 fish/h (Figure S1).

Environmental conditions varied between streams (Table S1) with average conditions across invasion levels as follows. In Bull Trout dominant streams, mean water temperature was 9.48 ± 0.02°C, mean wetted width was 8.40 ± 2.65 m, mean depth was 26.37 cm and mean elevation was 1528 ± 145 m. In Brook Trout dominant streams, water temperature was 9.70 ± 2.22°C, wetted width was 3.99 ± 1.25 m, depth was 18.25 ± 4.79 cm and elevation was 1465 ± 42 m. In the Brown Trout dominant stream, water temperature was 13.28°C, wetted width was 15.66 m, mean depth was 27.97 cm and the mean elevation was 1260 m. In Bull Trout and Brook Trout sympatric streams, water temperature was 9.58 ± 0.76°C, wetted width was 5.96 ± 2.93 m, depth was 22.29 ± 1.72 cm and elevation was 1553 ± 86 m. In streams where all three species were present, water temperature was 10.05 ± 0.34°C, wetted width was 6.02 ± 1.50 m, depth was 31.62 ± 9.01 cm and elevation was 1365 ± 132 m.

Prior to baseline correction, mean stable isotope values for Brook Trout ranged from −33.2 to −28.6‰ (δ¹³C) and 6.9 to 8.2‰ (δ¹⁵N). For Bull Trout, values ranged from −33.5 to −29.2‰ (δ¹³C) and 6.4 to 8.2‰ (δ¹⁵N), and for Brown Trout, it was −31.9 to −25.9‰ (δ13C) and 7.1 to 9.7‰ (δ¹⁵N), as summarized in Table S2. No effects of carbonate contamination were detected and, except for one fish, C:N ratios of fish tissues were below 4.0; therefore, lipid correction was not required. Baseline isotopic values for aquatic invertebrates ranged from −38.7 to −33.8‰ (δ¹³C) and 1.4 to 4.4‰ (δ¹⁵N), whereas for terrestrial invertebrates, values ranged from −31.8 to −26.6‰ (δ¹³C) and 0.4 to 4.4‰ (δ¹⁵N).

3.1. TP, terrestrial carbon and fork length

Among the three species, Bull Trout exhibited the lowest TP with an average value of 3.38 ± 0.02 relative to Brook Trout (3.45 ± 0.02) and Brown Trout (3.75 ± 0.03). When comparing between invasion levels, Bull Trout TP remained unchanged in the presence of non‐native species. Correspondingly, Bull Trout fork length did not significantly differ across invasion levels (Figure 3c). In contrast, Brook Trout TP significantly increased from allopatry to both sympatric invasion levels (Figure 3a), but changes were modest (7% increase in two species systems and 3% increase in three species systems). Body size played a minor role in these changes, where sympatric Brook Trout fork length was only significantly greater than that of allopatric Brook Trout when in sympatry with all species (Figure 3c). Interestingly, Brook Trout TP was significantly higher when in sympatry with Bull Trout compared to when they coexisted with both Bull Trout and Brown Trout, even though Brook Trout were not significantly longer in the former scenario (Figure 3a). There was also a significant difference in TP between Brook Trout and Bull Trout while in sympatry despite no difference in body size, but this difference was weak with a small effect size (Brook Trout were only 5% higher in TP). Brown Trout TP was significantly higher when in allopatry relative to sympatry (by 12%), and correspondingly, fork length was also significantly greater. Within invasion levels, allopatric Brown Trout had a significantly higher TP than allopatric Brook Trout and Bull Trout (by 23% in each case) and a significantly higher fork length than both species. When all three species coexisted, both sympatric Brook Trout and Brown Trout had significantly greater fork length than sympatric Bull Trout; however, only sympatric Brown Trout had a significantly higher TP than sympatric Bull Trout (9% increase, Figure 3a).

FIGURE 3.

Trophic position (TP) (a), proportion of terrestrial‐based carbon in the diet (b), and fork length of fish (c) for individual species within invasion levels. First tile includes Brook Trout dominant, Bull Trout dominant and Brown Trout dominant streams. Second tile includes streams of the Brook Trout/Bull Trout invasion level, and third tile includes streams of the Brook Trout/Bull Trout/Brown Trout invasion level. Significant differences (p value <0.05) between species indicated by *. Significant differences within a species denoted by differing letters.

Bull Trout exhibited the lowest proportion of carbon derived from terrestrial sources overall at 0.36 ± 0.02, while Brook Trout (0.48 ± 0.02) and Brown Trout (0.55 ± 0.02) had roughly half their carbon from terrestrial sources. Comparing between invasion levels, Bull Trout exhibited a moderate but insignificant decrease in the proportion of carbon derived from terrestrial sources, whereas both Brook Trout and Brown Trout exhibited moderate but insignificant increases (Figure 3b). Correspondingly, the diets of sympatric Brook Trout and Brown Trout were significantly higher in terrestrial‐derived carbon than that of sympatric Bull Trout, by 65% and 43% for Brook Trout and 58% for Brown Trout (Table S3).

3.2. Standard ellipses area and niche overlap

There was a moderate, but insignificant, decrease in the niche width of Bull Trout across invasion levels (Figure 4) and both Brook Trout and Brown Trout had significant increases in niche width from allopatry to sympatry with all three species. When comparing total niche width of all species within each invasion level, there were no significant changes in total niche width. Stream by stream differences in niche expansion/contraction matched all of the theoretical possibilities (Figure 1). That is, Bull Trout niche sizes in allopatry were both large (Burnt Timber) and small (Rocky) (Figure S2; Table S3). When in sympatry, Bull Trout niches were also both large (Cutoff, Elk, Trout, Wildhorse) and small (Rough, Wilson) (Figure S2). At locations where Bull Trout niches were smallest in sympatry (Rough, Wilson), TPs were lowest and terrestrial derived carbon was least (Figure S2), best matching the patterns in scenarios c and d of Figure 1.

FIGURE 4.

Mean niche width (SEA_B) and their 95% credible intervals of Brook Trout (BKTR), Bull Trout (BLTR), Brown Trout (BNTR) and all species together (All) by invasion level (Number of Species).

In terms of niche overlap, there was a high median probability that Brook Trout (94% probability) and Brown Trout (79% probability) would be found within the overall dietary niche of Bull Trout (Table 1). Similarly, there was a high median probability that Bull Trout would be found within the dietary niche of Brook Trout (85% probability) and Brown Trout (83% probability). When comparing niche overlap between invasion levels, the likelihood of Brook Trout in the niche of Bull Trout, and vice versa, did not change significantly (Table 1). However, there was a significantly lower likelihood of allopatric Bull Trout and Brook Trout occupying the dietary niche of allopatric Brown Trout (Table 1).

TABLE 1.

Mean isotopic niche overlap between Brook Trout (BKTR), Bull Trout (BLTR) and Brown Trout (BNTR) with 95% confidence intervals.

		BKTR	BLTR	BNTR
Allopatry	BKTR	‐	98.74 (95.4, 100)	15.49 (3.4, 39.51)
	BLTR	66.38 (53.9, 78.6)	‐	23.31 (11.1, 40.33)
	BNTR	11.33 (3.3, 25.31)	65.45 (35.5, 92.2)	‐
BLR/BKTR	BKTR	‐	94.27 (90.3, 97.4)	‐
	BLTR	84.68 (78.9, 90.2)	‐	‐
	BNTR	‐	‐	‐
BLTR/BKTR/BNTR	BKTR	‐	88.24 (80, 94.9)	90.43 (83.3, 96.2)
	BLTR	86.52 (78.1, 93.4)	‐	83.26 (73.1, 91.1)
	BNTR	83.75 (73.7, 91.5)	82.59 (72.8, 90.2)	‐
Total overlap	BKTR	‐	94.25 (90.3, 97.3)	94.19 (89.1, 97.6)
	BLTR	85.01 (78.5, 90.71)	‐	83.09 (75.9, 90.2)
	BNTR	72.22 (62.6, 80.51)	78.79 (70.5, 86.1)	‐

Note: Means represent the probability that a species from the column two will be found in the isotopic niche of a species in the top row. Percent overlap is separated by allopatric streams, Bull Trout/Brook Trout streams (BLTR/BKTR), Bull Trout/Brook Trout/Brown Trout streams (BLTR/BKTR/BNTR) and overlap across all streams in the study area (Total overlap).

4. DISCUSSION

We found native Bull Trout were generally smaller and fed at lower TPs and less on terrestrial resources compared with non‐native Brook Trout and Brown Trout. Though Bull Trout did not exhibit significant niche contractions or shifts when in sympatry with the non‐native species, suggesting the non‐native species were partly occupying previously unoccupied niches, there remained considerable overlap in niches, which suggests the potential ongoing competition for resources. Global introduction and stocking of Brook Trout and Brown Trout have had significant impacts on native fish populations, leading to increased invasion pressure (Kitano, 2004; Lowe et al., 2000; Townsend, 1996; Waters, 1983). Presence of these fishes can ultimately lead to replacement of native species (Eby et al., 2006).

When native trout coexist with non‐native trout following invasion, resource availability can become limited, leading to shifts in resource use by different species (Eby et al., 2006; Kitano, 2004; Nakano et al., 1998; Nakano et al., 1999; Taniguchi et al., 2002). While there was a slight downward shift in SEA_B for Bull Trout in sympatry, overall there was no significant decrease in TP or 1−α when the non‐native species were present. As such, our findings do not support significant displacement either by niche reduction or niche shifts along these two niche axes (Figure 1). Importantly, the use of multiple streams with different combinations of species allowed us to avoid drawing conclusions that may have been made if only one allopatric and one sympatric stream were compared (Figure S2). Bull Trout did show the lowest consumption of terrestrial‐based carbon sources, especially compared with Brown Trout, which also consumes more terrestrial‐based carbon when in sympatry with Brook Trout (Cucherousset et al., 2007). Brown Trout also exhibit ontogenetic shifts, increasing consumption of terrestrial invertebrates in larger size classes (Sánchez‐Hernández & Cobo, 2016). Aggressive interference of Brown Trout with Brook Trout has been suggested as a driving factor behind this increase in terrestrial‐based carbon consumption (Fausch & White, 1981; Nyman, 1970). Our results support this notion and demonstrate that Bull Trout access fewer terrestrial resources, especially when in sympatry with the two non‐native species.

Changing foraging tactics can enable species coexistence (Holbrook & Schmitt, 1989; Nakano et al., 1999); however, any minor shifts in resource use by Bull Trout did not lead to a significant change in niche overlap between allopatric and sympatric conditions. As a result, Bull Trout may be adapted to partly rely on less desirable resources (low‐TP aquatic insects) that allow them to coexist with Brook Trout and Brown Trout. This reliance on alternative resources may hinder Bull Trout growth and abundance, particularly when terrestrial invertebrates are scarce or usurped, as seen when Dolly Varden (Salvelinus malma) are sympatric with non‐native Rainbow Trout (Oncorhynchus mykiss) (Baxter et al., 2007). Brook Trout, known for their fast maturation and prolific reproduction (Dunham et al., 2002; Raleigh, 1982), have proven to be strong competitors in streams they have invaded, often occurring in higher densities compared to species they replace (Benjamin & Baxter, 2010). With size playing a role in determining interspecific dominance (Nakano & Furukawa‐Tanaka, 1994), Bull Trout are at a disadvantage as they grow slower and mature later (Dunham et al., 2008; Meyer et al., 2006). Laboratory studies have further demonstrated competitive and growth advantages of Brook Trout over Bull Trout, particularly in warmer water temperatures (McMahon et al., 2007).

Our findings indicate a significant increase in TP for Brook Trout from allopatry to sympatry with no change in TP for Bull Trout, but both species had significantly lower TP than Brown Trout. Body size may be a contributing factor to a higher TP when all three species are present since both Brook Trout and Brown Trout were larger than Bull Trout, but size does not explain the weak but significant differences in TP between Bull Trout and Brook Trout when in sympatry with only each other. While the TP difference in this latter scenario was small and unlikely to be biologically meaningful (one fifth of one trophic level difference), the species differed strongly in their use of terrestrial resources when in sympatry. Thus, our findings suggest that Brook Trout are likely to maintain their competitive advantage over Bull Trout even at similar body sizes. Piscivory may contribute to increasing TP, especially in large Brown Trout; however, piscivory in riverine populations of Brown Trout is rare (Sánchez‐Hernández, 2020). Instead, the higher TPs in Brook Trout and Brown Trout may be due to selective targeting of larger‐sized invertebrates (Bechara et al., 1993). If non‐native trout usurp all larger invertebrates, Bull Trout will need to exert more energy to obtain an equal amount of caloric intake, ultimately decreasing their fitness (Stephens & Krebs, 1986).

Lack of significant change in total niche width when considering all species together, along with the lack of significant change in individual species' niche widths and a slight increase in percent overlap, does not provide evidence of niche partitioning between sympatric Bull Trout and Brook Trout. Our findings align with previous experimental and observational studies that also found no evidence of a niche shift when examining feeding and behavioural interactions between Brook Trout and Bull Trout (Gunckel et al., 2002; Nakano et al., 1998). Absence of niche partitioning and high degree of niche overlap (over 85%) between sympatric Brook Trout and Bull Trout indicates potential for future displacement, particularly when resources are limited. Potential for displacement is likely heightened by contrasting feeding behaviours of the two species. Bull Trout typically employ a sit‐and‐wait foraging strategy (Bonneau & Scarnecchia, 1998; Nakano et al., 1998), whereas sympatric Brook Trout are more inclined to engage in competitive interference for resources (Gunckel et al., 2002; McMahon et al., 2007; Rodtka & Volpe, 2007; Warnock & Rasmussen, 2013b). Consequently, Bull Trout are susceptible to being displaced from terrestrial invertebrates and large prey when resources are scarce, and abundance of Brook Trout is high.

While previous studies have identified terrestrial invertebrates as an important resource for trout (Baxter et al., 2007; Nakano et al., 1999), availability of terrestrial prey is limited during the winter months (Nakano & Murakami, 2001) especially in high‐elevation streams, so the extent to which they are incorporated into trout diets may depend instead on availability of aquatic invertebrates. It is anticipated that high‐elevation, cold‐water streams will serve as refuges for salmonids amidst a changing climate (Dunham et al., 2003; Eby et al., 2014; Poesch et al., 2016), and our Bull Trout dominated sites tended to be cooler, narrower and at higher elevations than those dominated by Brown Trout. While these streams often have reduced benthic secondary production (Huryn & Wallace, 2000), some aquatic prey taxa have larger average body sizes in such colder environments (McCarty et al., 2022) and aquatic invertebrates do contain key omega‐3 fatty acids (Hixson et al., 2015), which increases their quality and may allow Bull Trout to persist while feeding exclusively on these resources, as they did in Rough and Wilson (Figure S2). Both Brook Trout and Brown Trout are able to thrive in low productivity cold‐water streams (MacCrimmon & Campbell, 1969; Smialek et al., 2021) but may do so by continuing to exploit terrestrial resources. Coupled with the competitive interference posed by Brook Trout, Bull Trout are likely to be excluded from accessing terrestrial invertebrates when benthic resources are limited. Assessing the availability of both terrestrial and aquatic resources in combination with isotopic assessments of dietary overlap would be a fruitful area for further investigation in this study system. Given our findings, it is critical for monitoring and management efforts to prioritize limiting further invasion of Brook Trout and Brown Trout into streams that already harbour healthy populations of Bull Trout, especially in cases where resources are already constrained due to environmental conditions.

AUTHOR CONTRIBUTIONS

Jacqueline Pallard, Mark S. Poesch: Conceptualization. Jacqueline Pallard, Tim Jardine, Mark S. Poesch: Developing methods. Jacqueline Pallard: Data analysis; Preparation of figures and tables. Jacqueline Pallard: Conducting the research. Jacqueline Pallard, Tim Jardine, Mark S. Poesch: Data interpretation and writing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Table S1. Environmental and electrofishing summary statistics by stream with the number sites sampled per stream (Sites). Time spent electrofishing (hrs); the number of BLTR (BLTR SIA), BKTR (BKTR SIA) and BNTR (BNTR SIA) fin clips used; and the total number of BLTR (BLTR Catch), BKTR (BKTR Catch) and BNTR (BNTR Catch) caught are totalled by stream. Water temperature (°C), wetted width (m), depth (cm), elevation (m) and CPUE fish/h are averaged across each stream.

Table S2. Sample sizes for stable isotope analysis for Brook Trout (BKTR), Bull Trout (BLTR), Brown Trout (BNTR), aquatic invertebrates (Collectors, Filterers, Scrapers and Shredders) and terrestrial invertebrates (Terrestrial, primary consumers only). Streams are separated by invasion levels of Bull Trout dominant (BLTR), Brown Trout dominant (BNTR), Brook Trout dominant (BKTR), Bull Trout/Brook Trout mix (BLTR/BKTR) and Bull Trout/Brook Trout/Brown Trout mix (BLTR/BKTR/BNTR).

Table S3. Means and standard deviation of raw δ¹³C and δ¹⁵N values for Brook Trout (BKTR), Bull Trout (BLTR), Brown Trout (BNTR), aquatic invertebrates (Collectors, Filterers, Scrapers and Shredders) and terrestrial invertebrates (Terrestrial, primary consumers only). Streams are separated by invasion levels of Bull Trout dominant (BLTR), Brown Trout dominant (BNTR), Brook Trout dominant (BKTR), Bull Trout/Brook Trout mix (BLTR/BKTR) and Bull Trout/Brook Trout/Brown Trout mix (BLTR/BKTR/BNTR).

Figure S1. Histogram of the fork length (cm) of fish caught by electrofishing by species across the study area.

Figure S2. Trophic position (TP, based on δ¹⁵N) versus proportion of terrestrial carbon (based on δ¹³C) of Bull Trout (BLTR), Brook Trout (BKTR) and Brown Trout (BNTR) across the 11 study streams, with Bayesian ellipses that capture 95% of the data.

DATA AVAILABILITY STATEMENT

Data supporting this research are available by contacting the administrator of the Alberta Fish and Wildlife Information System from the Government of Alberta (email: aep.fwdatamgmt@gov.ab.ca). Instructions on how to access the data are available online at: https://www.alberta.ca/access-fwmis-data.aspx including the requisite data sharing agreement.