Biotic and abiotic factors and their interaction influence the seasonal movement patterns of a dispersal limited benthic species Authors: Marie F Veillard, Douglas A Watkinson, Mark S Poesch Published: 2026 ABSTRACT AbstractDispersal is an important mechanism linked with population viability. Increases in species-specific dispersal allow for improved connectivity between habitat patches and populations. Here, we seek to understand the role of both biotic and abiotic factors, and their interactions, in influencing the movement of the recently identified and federally threatened Rocky Mountain sculpin (Cottus sp.). We conducted a mark-recapture study in a 400 m reach of Lee Creek in Alberta, Canada, using passive integrated transponder and visible implant elastomer tags across approximately 4 months. Boosted regression tree models were used to assess the movement of (1) all recaptured individuals (global model) and (2) only mobile individuals (movement only model) in response to abiotic and biotic factors. Biotic factors, such as congeners at the destination (8.7%), congeners at the origin (8.0%), and competitors at the origin (7.2%) were the most important variables for predicting movement in the global model. Alternatively, cobble (18.7%) followed by biotic factors including congeners at the origin (9.6%) and competitors at the origin (9.1%) were the most important variables selected in the movement only model. Biotic and abiotic factors showed strong interactions, providing a clear example of the importance of competition in the understanding of movement. Although the vast majority of restoration activities for endangered species are aimed at abiotic (i.e. habitat-related) factors, this study shows how these may be limited without considering biotic interactions, such as the role of inter- and intraspecific competition.Supplementary InformationThe online version contains supplementary material available at 10.1007/s10641-026-01841-9.Keywords: Freshwater fishes, Biodiversity, Inter- and intraspecific competition, Optimal foraging, Mark-recapture, Rocky Mountain sculpin (Cottus sp.) Introduction Dispersal is a mechanism underlying many ecological processes. Understanding movement patterns is important for species conservation, including factors that can influence dispersal (Radinger and Wolter 2014). While some fish undertake long-distance migrations, many stream fish populations are largely sedentary (Gerking 1959) with a subset of mobile individuals (Funk 1957; Rodriguez 2002). The sedentary portion of the population is often associated with concepts such as home range (Gerking 1959; Radinger and Wolter 2014) and territoriality (Gerking 1953, 1959), resulting in density-dependent movement (Jenkins 1969). On the other hand, the mobile portion of the population has been suggested to contribute to a host of ecological processes such as range shifts (Hynes 1970; Radinger and Wolter 2014), meta-population dynamics (Hanski 1998), colonization or recolonization of suitable habitat (Knaepkens et al. 2005), and gene flow (Wright 1978). The ability to respond to environmental changes may then be influenced by mobile individuals in the population, inter- and intra-species interactions, and changes to suitable habitat, indicating the importance of both biotic and abiotic factors to movement (Gilliam and Fraser 2001). Movement is known to be influenced by a suite of biotic and abiotic factors as individuals attempt to alleviate stress from competition (Petty and Grossman 2004) and predation (Owens and Bergstedt 1994), while optimizing preferred habitat (Tabor et al. 2007). Despite the influence of abiotic and biotic factors on stream fish movement, few studies have quantified their importance (Turchin 1998). Abiotic factors have been more readily incorporated into fish movement studies (Albanese et al. 2004; Breen et al. 2009; Petty and Grossman 2004), whereas biotic factors such as predation and competition are less frequently used in movement studies despite their importance in structuring fish communities (Jackson et al. 2001). In a movement study conducted by Gilliam and Fraser (2001), they found an interaction between biotic pressures and abiotic physical streambed structure that provided cover to promote movement of the killifish (Anablepsoides hartii (Boulenger, 1890)) in the presence of its predator trahira (Hoplias malabaricus (Bloch, 1794); Gilliam and Fraser 2001). However, the role of competitive interactions is less clear for structuring stream fish communities (Jackson et al. 2001). In a study of competitive interactions between the benthic mottled sculpin (Cottus bairdii Girard, 1850) and fantailed darter (Etheostoma flabellare Rafinesque, 1819), Resetarits (1997) found that juveniles from each species under heterospecific competitive pressure resulted in reduced growth and decreased relative condition, respectively. Additionally, Jenkins (1969) found brown trout (Salmo trutta Linnaeus, 1758) and rainbow trout (Oncorhynchus mykiss (Walbaum, 1792)) to exhibit movements in relation to inter- and intraspecific agonistic acts resulting in chasing and displacement of individuals. Freshwater sculpins (genus Cottus) are a useful group for investigating movement in relation to both biotic interactions and abiotic environmental factors. They are widely used as sentinel species in environmental monitoring (Environment Canada 2012; Gray et al. 2018) because effective sentinels should reflect local conditions where they reside and provide realistic estimates of exposure to site-specific effluents and contaminants (Gray et al. 2018). Accordingly, sculpins have been applied in monitoring programs for contaminants such as heavy metals (Environment Canada 2012), including mercury (French et al. 2024) and selenium (Miller et al. 2015), as well as persistent organic pollutants such as PCBs (Kuzyk et al. 2005). Understanding sculpin movement is therefore important for interpreting sentinel responses and for defining the spatial scale over which individuals integrate local conditions. Sculpins are generally more sedentary than many other stream fishes (Radinger and Wolter 2014). Schwalb et al. (2011) reported that freshwater sculpins moved an average of ~ 23 m over a mean study duration of 13 months, indicating strong site fidelity. Limited movement can increase vulnerability to disturbances (McCormick et al. 1994), including waterway alterations and movement barriers (Maitland et al. 2016), and may also intensify the importance of local biotic processes such as intra- and interspecific competition because individuals cannot readily disperse to access new resources (Anderson 1985; Greenberg 1991). Freshwater sculpins typically occur in cool to cold habitats (Adams and Schmetterling 2007) and often require clean, unembedded cobble substrates to complete key life-history stages (Bond 1964; Finger 1982; Haro and Brusven 1994). Taken together, these characteristics suggest that studies of sculpin movement should explicitly integrate both biotic context (e.g. densities of congeners, competitors, and predators) and abiotic habitat constraints (e.g. temperature, substrate condition, and connectivity). Several freshwater sculpin species have been identified in recent years (Lemoine et al. 2014; Neely et al. 2007), including the Rocky Mountain sculpin (Cottus sp.) (David A Neely 2003 unpublished). Rocky Mountain sculpin straddles the Rocky Mountains on the west coast of North America and is listed as Threatened under Canada’s Species at Risk Act (Fisheries and Oceans Canada 2012). Local changes to environmental conditions, such as flow, are hypothesized to negatively affect Rocky Mountain sculpin species, as found in laboratory and field studies (Rudolfsen et al. 2019; Veillard 2016; Veillard et al. 2017). It is critical we understand the role of both abiotic and biotic interactions in determining species-specific movement to predict how these changes may impact broader dispersal-mediated population and community processes. Using Rocky Mountain sculpin as a model organism, we predict the following: (1) as a benthic and burrowing species, the movement ability of the species over the course of the summer and fall seasons (July–October) will not be influenced by biotic and abiotic factors, and (2) movement may be influenced by the number of congeners (competition). Understanding the factors associated with Rocky Mountain sculpin movement is paramount for determining the sustainability of this range-limited species and the broader response to conservation efforts by dispersal-limited species. Methods Study area The movement patterns of Rocky Mountain sculpin were observed in Lee Creek, a tributary to the St. Mary River in Alberta, Canada (Fig. 1a, b). Lee Creek is located within the Oldman River sub-watershed in southern Alberta (Fig. 1c). Flow is predominantly driven by snowmelt and precipitation in this system, with lows of 0.2 m3s−1 (November) and highs of 7.5 m3s−1 (June; Alberta Environment and Parks 2014). Catch per unit effort (CPUE) is high within the Canadian range of Rocky Mountain sculpin, including Lee Creek (16.62 fish min−1; COSEWIC 2005). This high density of fish, together with a wide range of habitat types found along Lee Creek, makes this area an ideal system to study the movement of Rocky Mountain sculpin and determine biotic and abiotic factors influencing the movement of the species. Map of the study area where (a) indicates the location of the Oldman River watershed (polygon) and study site (star) within Alberta, Canada; (b) indicates the location of the study site (star) within Oldman River watershed (polygon); and (c) illustrates the study site design over a 400-m stretch of Lee Creek, part of the Oldman River watershed We conducted a mark-recapture field study within a 400 m stretch of Lee Creek, consisting of an inner 200 m core tagging area with 100-m buffer zones extending upstream and downstream of the core in the summer and fall as outlined by Veillard (2016) and Veillard et al. (2017). Cross-sectional transects were established every 10 m throughout the site, totalling 41 transects (Fig. 1d). The site was chosen to include two main riffle sections connected by runs and pools which characterized a range of available habitat. Otherwise, natural, historic stream channelization occurred in the upper 150 m of our study area, creating a straight, uniform stretch. Fish tagging and movement Fish abundance and abiotic habitat variables were measured across six sampling events, which commenced every 17 ± 1.45 days from July through to October 2013. To track movement, passive integrated transponder (PIT) and visible implant elastomer (VIE) tags were deployed (Tables 1 and 2). Long-lasting remote PIT tagging systems are often used in studies of small-bodied fish (Breen et al. 2009; Bruyndoncx 2002; Roussel et al. 2000) as an alternative to traditional mark-recapture techniques, such as VIE, that require physical recapture (e.g. electrofishing; Neufeld et al.2015). Table 1A summary of field data collected from abiotic, biotic, and temporal variables included in boosted regression tree movement modelsVariablesDescriptionMean (Range)ResponsedistDistance (m) between transects for each movement event8.42 (0–170)PredictorsAbiotic factorsOriginDestinationdepth.odepth.dAverage transect depth (cm)23.9 (13.4–57.1)24.5 (13.4–61.4)sd.depth.osd.depth.dStandard deviation of depth (cm)9.5 (2.4–22.1)9.5 (2.8–22.1)vel.ovel.dAverage transect velocity (cm/s)31.39 (4.5–81.5)27/94 (4.5–81.5)sd.vel.ods.vel.dStandard deviation of velocity (cm/s)17.59 (4.8–52.2)15.55 (4.6–41.2)fines.ofines.dPresence of fine substrate (< 2 mm) at 6 points across each transect0.56 (0–3)0.67 (0–3)peb.opeb.dPresence of pebble substrate (2–32 mm) at 6 points across each transect1.35 (0–4)1.40 (0–4)gra.ogra.dPresence of gravel substrate (32–64 mm) at 6 points across each transect2.11 (0–4)1.90 (0–4)cob.ocob.dPresence of cobble substrate (64–256 mm) at 6 points across each transect4.53 (1–6)4.60 (0–6)bou.obou.dPresence of boulder substrate (256–1000 mm) at 6 points across each transect0.66 (0–3)0.68 (0–3)bed.obed.dPresence of bedrock (> 1000 m) at 6 points across each transect0.65 (0–5)0.65 (0–5)Biotic factorsrmsc.ormsc.dCount of congeners at each transect14.75 (2–30)14.35 (1–30)comp.ocomp.dCount of competitors at each transect10.50 (0–58)16.24 (0–60)pred.opred.dCount of predators at each transect0.07 (0–1)0.023 (0–1)Temporal factormonthMonth of recapture event (factor with levels: July, August, or September)-- A summary of field data collected from abiotic, biotic, and temporal variables included in boosted regression tree movement models Table 2Population estimates per time period for Rocky Mountain sculpin (Cottus sp.). Population estimates were derived using the Jolly-Seber open population model where M^i is the estimated number of marked fish in the population just prior to the ith sample; mi is the number of marked fish in the ith sample; ni is the total fish captured in the ith sample; Ri is the number of marked fish returned to the population after the ith sample, and ri is the number of the Ri fish released in the ith sample, not captured in the ith sample, but captured subsequently; and zi is the number of fish captured before the ith sample, not captured in the ith sample, but captured subsequently and (Ni 1000 mm) using a modified Wentworth Scale (Wentworth 1922). Abundance of each substrate type was ranked from 0 to 6 based on its presence or absence in the 6-point measurements; a substrate class present in all points of a cross-section was given a rank of 6 (Table 1). Mean depth and velocity were calculated from the 6-point measurements taken at each transect; standard deviation was calculated to represent the variability in depth and velocity across each transect (Table 1). Field data were summarized into July (period 1–2), August (period 3–4), and September (period 5–6) months, and a movement was defined as two or more detection events for an individual that occurred within the month (without VIE tags included). This was done to categorize monthly temporal variation for use in the models. Within each sampling period, the last known location for each individual was selected for analysis. Analysis We assessed all movements (total, n = 363; PIT, n = 309; VIE, n = 54) using a kernel density smoothing technique for movement events between sampling periods using the density function in the R package stats (bw = 2.068, adjust = 2, kernel = gaussian; R Core Team 2015). Kernel density estimation allows ecologists to make conclusions about populations based on an estimate of the probability density of a sample (Bolker 2008). Further, to test for differences in upstream and downstream movement distances, we used a two-sample t-test assuming equal variances. To determine what factors influenced movement, we used boosted regression tree (BRT) models fitted to minimize a Gaussian loss function (Elith et al. 2008; Hijmans et al. 2012). Distance between unique transect combinations was analyzed in response to abiotic, biotic, and temporal variables at the origin and destination transects (Table 1). Models were assessed for the core area (Fig. 1d; 200 m), where data on biotic factors were collected during electrofishing (total, n = 329; PIT, n = 275; VIE, n = 54). BRT models were fitted using the gbm.step function in the dismo library in R as described by Elith et al. (2008). Due to a small sample size, tenfold cross-validation was utilized to fit the models rather than splitting the dataset into training and testing data frames for prediction (Elith et al. 2008). We constructed two BRT models: (a) a global model with all individuals and (b) a movement-only model with only individuals that moved, to describe factors affecting movement when zero-movement events were included and excluded, respectively. Final parameters for the global (lr = 0.05) and movement-only (lr = 0.05) models were chosen by systematically altering the learning rate to reduce predictive deviance using > 1000 trees. A bag fraction of 0.5 and tree complexity of 3 were used following recommendations for small sample sizes by Elith et al. (2008). Finally, both models were simplified following Elith et al. (2008) to remove redundant variables, thereby increasing parsimony. Results Rocky Mountain sculpin dominated catch followed by longnose dace, plains sucker, lake chub, burbot, longnose sucker, white sucker (Catostomus commersonii (Lacepède, 1803)), and trout-perch. Population size of Rocky Mountain sculpin ranged from 142 (± 38) in late July to 580 (± 284) in late August (Table 2). Average recapture rate was higher for PIT-tagged individuals (81.6%) than those marked with VIE tags (12.6%). Despite high variability in abiotic and biotic parameters throughout the study period (Fig. S1), there was low movement of sculpins throughout the study period (Fig. 2). Of the 363 detections, 244 records (67.2%) indicated no movement at a 10-m scale. The vast majority of detections (89.3%) were within 30 m of the tagging transect (Fig. 2); the remaining 10.7% of detections were between 30 and 240 m with a slight increase at 60 m (Fig. 2). The longest movement event from one sample period to the next was 240 m over 13 days between August 16 and August 29. There was no significant difference in upstream or downstream movement (t-test, df = 118, p = 0.79) or by body size, e.g. VIE of only small individuals versus PIT tags (large individuals; t-test, df = 882, p = 0.88). Kernel density smoothing estimates for movement distances (m) of Rocky Mountain sculpin summarized as all detections and broken down by each sampling period (dates shown in Table 2). Dotted vertical line indicates average movement distance per period Biotic and abiotic components influencing movement Global model The global model emphasized biotic factors as important predictors of movement including congeners at the destination (8.7% contribution), congeners at the origin (8% contribution), and competitors at the origin (7.2% contribution) as the three most important variables (Fig. 3a; Deviance Explained = 53.0%, Table S2). Movement was best explained by a high abundance of congeners (weighted mean = 16) and competitors (weighted mean = 22) at the origin and a low abundance of congeners at the destination (weighted mean = 6; Fig. S2; Table S2). Variable importance for: (a) the global and (b) movement only models. Numbers at the end of bars represent percent contribution to the model. Bar colour indicates if the variable was included in the final model (dark grey) or not (light grey). Variable names are shown in Table 1 The top abiotic factors included variables related to depth and velocity, contributing a total importance of 24.6% and 20.6%, respectively (Fig. 3a). Modal peak in mean depth shifted from between 15 and 25 cm at the origin to 30–40 cm at the destination (Fig. S2, Table S2). Standard deviation of depth at the destination showed a bi-modal shape peaking at approximately 5 cm and 20 cm (Fig. S2). Standard deviation of velocity at the destination most strongly influenced movement at low values of variation (weighted mean = 10.6; Fig. S2, Table S1). Conversely, standard deviation of velocity at origin transects had a bi-modal shape, peaking at variation less than 10 cm s−1 and between 20 and 40 cm s−1 (Fig. S2). The strongest interaction occurred when the standard deviation of velocity was approximately 10 cm s−1 at the destination, and congener abundance was greater than 25 individuals at the origin (Fig. 4a, b). The top four interactions all included congeners or the standard deviation of velocity at either the origin or destination transects (Fig. 4a, b). Top panel: Global model interaction effects calculated by holding all other variables at their means (Elith and Leathwick 2008). a Interaction rank; dark grey bars indicate the top 4 interaction effects illustrated in (b); light grey bars are the remaining interactions. Bottom panel: Movement only model for all interaction effects. c Interaction rank; dark grey bars indicate the top 4 interaction effects illustrated in (d); light grey bars are the remaining interactions. Variable names are shown in Table 1 Movement only model The most important biotic factors in the movement only model (Deviance Explained = 47.03%, Table S1) included congeners at the origin (9.6% contribution), competitors at the origin (9.1% contribution), and competitors at the destination (5.6% contribution; Fig. 4b). Partial plots and weighted means indicated a shift from origin transects with > 20 congeners or competitors to destinations with < 10 individuals for either biotic group (Table S2), demonstrating a strong biotic effect influencing movement. The variable importance plot (Fig. 3b) showed cobble at the destination to have a strong modal effect (Fig. S2) on movement (contributing 18.7%) towards a weighted mean of 6.2 (Table S1) signifying the importance of destinations with a high presence of cobble. Additional abiotic factors of importance included velocity and depth variables, with overall importance of 17.8% and 13.1%, respectively (Fig. 3b). Standard deviation of velocity at the origin showed a similar bi-modal shape as the global model, with peaks at low (< 10 cm s−1) and high (20 cm s−1) variation (Fig. S2). Likewise, mean velocity at the origin showed two peaks at less than 20 cm s−1 and at approximately 40 cm s−1 (Fig. S2). Standard deviation of depth at the origin peaked between 7 and 10 cm, while standard deviation of depth at the destination had a weak bimodal shape peaking at approximately 5 cm and 17 cm (Fig. S2). Substrate variables were more important in the movement only model (40.2% contribution) than in the global model (26.9% contribution). Three of the four most important interaction terms included the substrate variables cobble and boulder at either the origin or destination (Fig. 4c), showing a preference for cobble and movement away from boulder (Fig. 4d). The second most important interaction occurred between the standard deviation of depth and congeners at the origin (Fig. 4a, b). For this interaction, Rocky Mountain sculpin movement was best explained when the standard deviation of depth was approximately 7–10 cm and the abundance of congeners at the origin was greater than 25 individuals (Fig. 4b). Discussion The role of biotic components influencing movement Inter- and intraspecific competition were among the most important factors explaining Rocky Mountain sculpin movement. Our study found movement to be clearly influenced by a shift from high to low abundances of Rocky Mountain sculpin, providing evidence of a density-dependent response. Previous work suggests that high density of congeners can elicit intraspecific interactions between sculpins through both competitive interactions for shelter space (Grossman et al. 2006; Petty and Grossman 2004) and predatory responses from larger sculpins (Downhower and Brown 1979; Freeman and Stouder 1989). Similarly, interspecific competition declined from high to low abundances of small-bodied benthic fishes. Benthic species also compete for the allocation of limited resources such as shelter space (Van Kessel et al. 2011) and food availability (Greenberg 1991). The importance of biotic interactions for Rocky Mountain sculpin movement may make the species vulnerable to non-native species introductions and shifts in community composition. Sculpins have shown adverse responses to introduced competitor species with drastic declines in abundance (Lauer et al. 2004) due, in part, to competition for shelter space (Van Kessel et al. 2011) and nesting sites (Janssen and Jude 2001). Introduction of non-native predators has also resulted in sharp declines in cottid abundance as well as a shift in habitat use away from deeper areas towards riffle habitat too shallow for predators (White and Harvey 2001). In these instances, abundant suitable habitat may buffer the impact of biotic changes up to a certain capacity. Our study suggests sculpins in this system, with high recorded levels of CPUE, may already be experiencing competitive pressure. During the study, food availability in Lee Creek appeared abundant (M. Veillard, personal observation), emphasizing that suitable habitat space may be the limited resource driving inter- and intraspecific competition. The lack of a predator effect was surprising, given the seminal works of Power et al. (1985), Matthews et al. (1994), and Gilliam and Fraser (2001) who all show predator-mediated movement in small-bodied fishes. Perhaps the reason for this was that we had a single predator in the system, burbot, which occurred in relatively low abundances. Further, our collection methods targeted small-bodied benthic species, and therefore, the community composition and abundance may not accurately represent the fish community of non-target fishes (Mahon 1980). The role of abiotic components influencing movement The three most important abiotic variables in our study included cobble, velocity, and depth. Movement in the movement-only model was largely explained by high amounts of cobble at destination transects. While cobble is known to be an important substratum for Rocky Mountain sculpin (Bailey 1952; COSEWIC 2005), these substrates must also maintain a low degree of embeddedness for microhabitat use by sculpins (Bateman and Li 2001; Haro and Brusven 1994). Riffle habitats maintain lower siltation rates and embeddedness (Berkman and Rabeni 1987) attributable to a high ratio of velocity to depth (Jowett 1993). The specific relationship between these three abiotic variables maintains interstitial spaces utilized by sculpins for shelter and spawning, lending to the importance of these factors observed in our study. Under competitive pressure, the importance of cobble could be an indication of movement towards riffle margins still containing cobble but where competition may be lower due to slightly slower (Emery et al. 2003) and deeper (Jowett 1993) water, often resulting in lower benthic invertebrate abundances (Hynes 1970). Depth variables suggested a shift from preferred riffle transects at the origin (M. Veillard, personal observation) towards both shallower riffles and deeper waters on the periphery of riffle habitat when cross-validated with field data. Rocky Mountain sculpin in our study appear to be utilizing a wide range of depths similar to other cottid species experiencing congener competition (Finger 1982; Freeman and Stouder 1989) and interspecific benthic fish competition (Greenberg 1991). The range of depths at the destination inhabited by Rocky Mountain sculpin in our study (13.4–61.4 cm; Table 1) was slightly higher than previously reported depth ranges (5–42 cm) for the species in southern Alberta (COSEWIC 2005), potentially due to density-induced movement towards sub-optimal habitat. Consistently, velocity parameters illustrated a density-based movement from preferred riffle habitat represented by the high variation peak (20 cm s−2 < sd.vel < 40 cm s−2) towards slightly slower riffle margins (standard deviation of velocity ~ 10 cm s−2). Rocky Mountain sculpin utilized an assortment of habitats under competitive pressure, demonstrating the adaptive ability of the species to a range of velocities. Bailey (1952) noted Rocky Mountain sculpin to tolerate a wide spectrum of velocities, only avoiding slow-moving reaches where fine particles were deposited on the substrate. In field trials, Facey and Grossman (1992) observed Mottled Sculpin strongly preferring habitats with lower than available river velocities while simultaneously utilizing cover to reduce additional energetic costs of velocity. Similarly, Rocky Mountain sculpin under competitive pressure in our study area displayed analogous behaviour to the Mottled Sculpin studied by Facey and Grossman (1992). Rocky Mountain sculpin exhibited an additional behaviour in our study where individuals occupying pool and run habitats at the origin (standard deviation of velocity < 10 cm s−2) moved towards areas with higher variation in velocity, likely representing the margins of riffles. This behaviour may simply represent movement towards more suitable habitat or could be comparable to findings by White and Harvey (2001) as a response to predator presence in deeper pool areas. Interactions between abiotic and biotic factors As a hierarchical tree-based method, boosted regression trees models successively split data into increasingly more homogenous groups, inherently assessing interaction terms without having to manually choose important interactions (Elith et al. 2008). The interaction between standard deviation of velocity and congeners was the highest ranked interaction for the global model. For the movement only model, the interaction between standard deviation of depth and congeners was the second highest ranking interaction. These results provide further evidence that movement towards slightly slower velocities at riffle margins and away from depths associated with optimal riffle habitat is linked to congener abundance. In a study on the movement of Mottled Sculpin, Petty and Grossman (2004) concluded that the interaction between stream flow and abundance of conspecifics influenced juvenile movement. Similarly, we found a relationship in both our models between intraspecific competition and velocity and depth variables, which together are a function of stream flow (Hynes 1970). Additionally, substrate variables were important for interactions in the movement only model, indicating the importance of optimizing physical substrate selection by movement towards cobble and away from boulder. The strong interactions between biotic and abiotic variables further emphasize our need to include both components in movement studies. Restricted movement of Rocky Mountain sculpin Historical studies have been conducted on Rocky Mountain sculpin in Montana, under the previous species’ names Cottus bairdi punctulatus (Bailey 1952) and Cottus bairdii (McCleave 1964) show the majority of individuals travel less than 46 to 47 m, respectively. Our study confirms the restricted movement of Rocky Mountain sculpin using updated tagging and tracking methods since fin clips (McCleave 1964) and jaw tags (Bailey 1952; McCleave 1964). While some individuals ventured longer distances (up to 240 m between recapture events), the majority of our detections (89.3%) were within 30 m of the original transect lending to the paradigm that fish populations may be comprised of an assortment of stationary and mobile individuals rather than a homogenous population of fish (Rodriguez 2002). This leptokurtic dispersal pattern has been observed in many freshwater fish families, including Cottidae, Fundulidae, Percidae, Salmonidae, Esocidae, Cyprinidae, Centrarchidae, Catostomidae, and Ictaluridae, as reviewed by Radinger and Wolter (2014). While these families exhibit leptokurtic dispersal, the stationary component does not always indicate zero movement, rather the most common distance moved by individuals (Radinger and Wolter 2014). For example, Cottidae and Fundulidae are the most dispersal-limited families with the stationary component moving up to a median of 10 m; comparatively, the stationary component of ictalurids, the most mobile family reviewed, moved up to 1000 m (Radinger and Wolter 2014). Limited movement demonstrated by fish in our study indicates that local river reaches of suitable habitat may be heavily relied upon by the majority of the population to complete their life history. Implications for conservation Sedentary movement observed in cottids suggests limited dispersal, which can reduce connectivity between suitable habitat patches and gene flow. Changes in Rocky Mountain sculpin morphology (Rudolfsen et al. 2018) and behaviour (Tierney et al. 2025) appear to be plastic responses to environmental variation, supporting local adaptation. However, limited movement may restrict gene flow and promote population homozygosity, with potential genetic consequences over time (Ruppert et al. 2017). These genetic consequences and isolation for freshwater fish populations can potentially imperil dispersal-limited species, such as freshwater sculpin, more so than those with greater dispersal abilities (Maitland et al. 2016). Conservation efforts have primarily focused on abiotic factors, including habitat restoration, with little emphasis on biotic interactions. Our study shows that this strategy may be limited without considering the importance of biotic factors in species conservation. The inclusion of these factors into future movement studies can help managers define important parameters of fine scale movement and focus conservation efforts accordingly. Supplementary Information Below is the link to the electronic supplementary material. DOCX (311 KB) Acknowledgements This project was made possible with funding from Fisheries and Oceans Canada, Species at Risk (SARCEP) funds, and NSERC Discovery Grant to Mark Poesch. Invaluable local area knowledge was provided by Terry Clayton and Emeric Janssens with Alberta Environment and Parks. The authors would like to thank Kelly Mulligan, Caitlin Good, Christine Lacho, Kenton Neufeld, and Troy Adams for their tireless contributions to field collections. Land access was graciously permitted by LaRae and Murray McKinnon, and Alberta Parks accommodated our field stay by providing access to a private camping site in the area. Nathan Medinski provided valuable comments on early manuscript drafts. All procedures were conducted with Canadian Council of Animal Care Approval (AUP00000759) under an approved Fish Research License (13-2408 FR) and federal Species at Risk Permit (SARA C&A 13-007). Author contribution MFV completed all field research and the initial draft of the manuscript (including analyses), DW and MP edited and provided advice on statistical analyses, and MP provided funding and supervision. Funding This project was made possible with funding from Fisheries and Oceans Canada, Species at Risk (SARCEP) funds, and NSERC Discovery Grant to Mark Poesch. Data availability Data is available upon reasonable request to the corresponding author. Declarations Ethics approval All procedures were conducted with Canadian Council of Animal Care Approval (AUP00000759) under an approved Fish Research License (13–2408 FR) and federal Species at Risk Permit (SARA C&A 13-007). Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials DOCX (311 KB) Data Availability Statement Data is available upon reasonable request to the corresponding author.