Habitat and Fish Assemblage Associations and Current Status of Northern Leatherside Chub Lepidomeda Copei in Western Wyoming

Western North American Naturalist

Volume 76 Number 4 Article 4

12-31-2016

Habitat and fish assemblage associations and current status of northern leatherside chub Lepidomeda copei in western Wyoming

Luke D. Schultz Wyoming Game and Fish Department, Pinedale, WY, [email protected]

Peter A. Cavalli Wyoming Game and Fish Department, Pinedale, WY, [email protected]

Hilda Sexauer Wyoming Game and Fish Department, Pinedale, WY, [email protected]

David Zafft Wyoming Game and Fish Department, Laramie, WY, [email protected]

Follow this and additional works at: https://scholarsarchive.byu.edu/wnan

Recommended Citation Schultz, Luke D.; Cavalli, Peter A.; Sexauer, Hilda; and Zafft, David (2016) "Habitat and fish assemblage associations and current status of northern leatherside chub Lepidomeda copei in western Wyoming," Western North American Naturalist: Vol. 76 : No. 4 , Article 4. Available at: https://scholarsarchive.byu.edu/wnan/vol76/iss4/4

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Western North American Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Western North American Naturalist 76(4), © 2016, pp. 427–440

HABITAT AND FISH ASSEMBLAGE ASSOCIATIONS AND CURRENT STATUS OF NORTHERN LEATHERSIDE CHUB LEPIDOMEDA COPEI IN WESTERN WYOMING

Luke D. Schultz1,3, Peter A. Cavalli1, Hilda Sexauer1, and David Zafft2

ABSTRACT.—Human activities have extensively altered native fish assemblages and their habitats in the western United States. Conservation and restoration for long-term persistence of these fishes requires knowledge of their distributional patterns and life history requirements. Northern leatherside chub Lepidomeda copei (hereafter northern leatherside) is a cyprinid native to the Snake and Bear River Basins of Wyoming, Idaho, Nevada, and Utah, and it is believed to have declined in distribution relative to historical records. To address information gaps in the species’ ecology and assess its status in the state, the objectives of this study were first to document the distribution (2010–2011) of northern leatherside in Wyoming and then to examine habitat factors related to the entire fish assemblage and to evaluate specific habitat associations of northern leatherside in the Bear River Basin, Wyoming. In the Bear River and Upper Snake River Basins, we documented the distribution of northern leatherside and compared it to the previously known distribution. Across the Bear River Basin, we used habitat measurements to assess abiotic features related to the distribution and abundance of northern leatherside. Northern leatherside was found across the Bear River Basin and was present in 2 streams each in the Upper Snake River and Green River Basins in Wyoming. Populations in Wyoming appear to represent the core of northern leatherside range, and our work provided a finer-scale delineation of the species’ occurrence. Northern leatherside was collected from a variety of habitats, but multivariate analyses and occurrence modeling indicated it was associated with increased channel depth and depth variability, and positively associated with other native fishes (including mountain sucker Catostomus platyrhynchus, redside shiner Richardsonius balteatus, and speckled dace Rhinichthys osculus). These findings on the distribution and ecology of northern leatherside provide important new information to assist successful management and conservation efforts within Wyoming and across the species’ range.

RESUMEN.—Las actividades humanas han alterado significativamente las comunidades de peces nativos de agua dulce y sus hábitats en el oeste de Estados Unidos. La conservación y la restauración enfocadas en la persistencia a largo plazo de estos peces requieren del conocimiento de los patrones de distribución y de los requerimientos de las historias de vida. Lepidomeda copei es una especie nativa de los ciprínidos en las cuencas de los ríos Snake y Bear, en Wyoming, Idaho, Nevada y Utah, que se cree ha disminuido en relación con su distribución histórica. Para llenar huecos de información en su ecología y evaluar su estado de conservación, los objetivos de este estudio fueron, en primer lugar, documentar su distribución (2010–2011) en Wyoming y, posteriormente, estudiar los factores del hábitat relacionados con la comunidad de peces y asociaciones específicas del hábitat de L. copei en la cuenca del río Bear, en Wyoming. En las cuencas del río Bear y las cuencas del Upper River Snake, documentamos la distribución de L. copei y la compara- mos con la distribución previamente conocida. Al otro lado de la cuenca del río Bear, utilizamos medidas del hábitat para evaluar los factores abióticos relacionados con la distribución y la abundancia de la especie. Encontramos L. copei a lo largo de la cuenca del río Bear, y estaba presente en dos corrientes de las cuencas del Upper Snake y el río Green, en Wyoming. Las poblaciones de Wyoming parecen representar el núcleo de la distribución de esta especie, y nuestro trabajo proporcionó una escala más fina de la presencia de esta especie. Colectamos L. copei en hábitats muy variables, pero el análisis multivariado y el modelo de ocurrencia indicó una asociación con el aumento de la profundidad del canal y la variabilidad de la profundidad, y una asociación positiva con otros peces nativos (incluyendo Catostomus platyrhynchus, Richardsonius balteatus y Rhinichthys osculus). Estos resultados sobre la distribución y la ecología de L. copei proporcionan nueva e importante información que ayudará en el desarrollo de un manejo exitoso y en los esfuer- zos de conservación en Wyoming y a lo largo de su distribución.

Due to the cumulative effects of overex- patterns, information on basic biology, an ploitation, habitat degradation, and interactions understanding of community and autecology, with introduced species, numerous western and an assessment of potential limiting factors native fishes are currently imperiled (Jelks et al. for the species (or assemblage) of interest 2008). Effective conservation efforts for these (Minckley and Deacon 1991). Though address- fishes require knowledge of distributional ing these information gaps has improved the

1Wyoming Game and Fish Department, Pinedale, WY 82941. 2Wyoming Game and Fish Department, 528 S. Adams St., Laramie, WY 82070. 3Present address: Forest and Rangelands Ecosystem Science Center, U.S. Geological Survey, Corvallis, OR 97331. E-mail: [email protected]

427 428 WESTERN NORTH AMERICAN NATURALIST [Volume 76 success of management actions for native to those of southern leatherside (Belk and sport fishes (e.g., salmonids of the genera Johnson 2007). Southern leatherside is usu - Oncorhynchus and Salvelinus; Dunham et al. ally restricted to lotic habitats below 2200 m 2008), conservation of native nongame fish has in elevation, is strongly associated with pools been hampered by a limited understanding of and other low-velocity areas (Wilson and their life histories and a general perception of Belk 2001), and is thought to have broad tol - these fishes as being undesirable (Cooke et al. erance for extreme environmental conditions 2005). Conservation and restoration of native (Belk and Johnson 2007). At the reach scale, fishes and ecosystems for long-term persis - southern leatherside is negatively associated tence can benefit from addressing these issues with brown trout Salmo trutta (Wilson and (Brouder and Scheurer 2007). Belk 2001). In the presence of brown trout, Northern leatherside chub Lepidomeda southern leatherside tends to occupy side- copei (hereafter northern leatherside) is a small channel refuge habitats, when they are avail - cyprinid native to the Bear River Basin in able (Walser et al. 1999, Olsen and Belk Idaho, Utah, and Wyoming and the Snake 2005). Loss of natural channel forms due to River Basin of Idaho, Nevada, and Wyoming habitat manipulation (e.g., channelization and (see UDWR 2009 for historical distribution dam/diversion construction) and degradation information). Although the species was origi - (e.g., riparian habitat impairment due to nally described to include populations of overgrazing) has been identified as poten - leatherside chub from southern Utah, recent tially contributing to decreased productivity molecular, morphological, and ecological evi - and fragmentation of northern leatherside dence supports designation of northern leather - populations (WGFD 2010). side and southern leatherside chub Lep - Conservation actions for northern leather - idomeda aliciae as separate species (Johnson side have recently been coordinated among et al. 2004). Across the range of both species, multiple state, federal, and nongovernmental leatherside chub have declined in distribution agencies (UDWR 2009). The conservation relative to historical records, have apparently agreement and recovery plan for northern been extirpated from numerous locations leatherside coordinated the conservation across their range, and are highly fragmented efforts of several state and federal agencies in areas of persistence (Wilson and Belk 2001, and targeted areas of management action, Belk and Johnson 2007). Jelks et al. (2008) including habitat protection and enhancement, included both species on their list of imperiled restoration of natural hydrologic conditions, freshwater fishes of North America and and control of nonnative species. The plan reported northern leatherside as endangered, also addressed efforts to restore northern while others have listed northern leatherside leatherside (via translocations) in areas where as critically imperiled (NatureServe 2011). In it has been extirpated (e.g., UDWR 2010). To Wyoming, northern leatherside is classified inform management decisions relative to the as a species of greatest conservation need conservation agreement, and to improve (WGFD 2010). understanding of the basic ecology of north - Historically, the distribution of northern ern leatherside and other native fishes in leatherside in Wyoming was described only Wyoming, our study contained 3 objectives. coarsely, and recent studies have documented First, we looked to document the current numerous previously unknown occurrences (i.e., 2010–2011) distribution of northern (Baxter and Stone 1995, Sigler and Sigler leatherside in the Bear and Upper Snake 1996, Zafft et al. 2009). Currently, some native River Basins of Wyoming. We then examined populations within Wyoming are thought to be habitat associations of stream fishes in the located in the Bear River Basin and a small Bear River Basin with multivariate analyses, portion of the Upper Snake River Basin, and and finally we evaluated biotic and abiotic recent genetic and geologic evidence suggest factors specifically associated with the distri - that 2 small northern leatherside populations bution of northern leatherside (i.e., 2011). in the Upper Green River Basin are also Distribution and abundance information for native (Blakney et al. 2014). The habitat and northern leatherside in Wyoming can be biotic associations of northern leatherside coupled with other data to identify and pri - (Wesner and Belk 2012) appear to be similar oritize areas of conservation interest (Williams 2016] NORTHERN LEATHERSIDE STATUS AND HABITAT 429

Fig. 1. Historical distribution (a) and current distribution (b) of northern leatherside chub Lepidomeda copei in southwestern Wyoming. Historical distribution is all known collections from Wyoming Game and Fish records between 1951 and 2009, adapted from Zafft et al. (2009). Current distribution of northern leatherside is from extensive sampling in 2010–2011. Closed circles are locations where northern leatherside was collected, open circles are locations that were sampled without collecting northern leatherside. Fish distributions were described for major drainage basins (i.e., Bear River, Green River, Snake River) and are depicted in panel a. et al. 2011), identify potential source popula - current distribution within its native waters in tions for translocation (Schultz and Cavalli Wyoming, excluding the Green River Basin. 2012), and identify habitat protection and Prior to the work of Blakney et al. (2014), restoration actions (UDWR 2009, 2010). Fur - northern leatherside from the Green River thermore, an analysis of factors associated Basin were generally believed to be intro - with the distribution of native fishes across duced (Zafft et al. 2009), so we did not target multiple spatial scales might allow managers these areas during our sampling. Effort was to identify factors critical for the persistence initially prioritized in areas where northern of multiple species and the scale at which leatherside historically occurred (Zafft et al. management actions may be effective (e.g., 2009; Fig. 1a). After we completed thorough Quist et al. 2005). inventories in areas of past occurrences, our additional sampling efforts targeted areas METHODS that historically had not been sampled. We performed extensive sampling in the Distribution Surveys Upper Snake River and Bear River drainages, We sampled stream reaches throughout but sampling approaches differed in these 2 the known and probable distribution of north - drainages. In the Upper Snake River Basin, ern leatherside (2010–2011) to document the most of the habitats where northern leatherside 430 WESTERN NORTH AMERICAN NATURALIST [Volume 76 were historically collected were lateral flood - accommodate these limitations. Generally, plain habitats of the Snake River—frequently block nets were placed at the upstream and side channels, beaver pond complexes, and downstream end of sample reaches; natural isolated backwaters—and were usually associ - barriers (e.g., diversion or beaver Castor ated with some type of cover (i.e., woody canadensis dams) were occasionally used as structure or aquatic vegetation; Zafft et al. upstream or downstream barriers. We primar - 2009). The Snake River itself is a large non - ily used single-pass backpack electrofishing wadeable river (mean width 50–100 m), and it to assess assemblage composition and species was impractical to sample small-bodied richness within sample reaches by moving in fishes from it (Curry et al. 2009). Therefore, an upstream direction with an electrofisher. multiple-person crews opportunistically sam - Output (100–350 V, 20–80 Hz) was adjusted pled these habitats using backpack electro - to a level that allowed capture of individual fish ing in >10-km river segments to delineate fish with dip nets. Larger stream systems (i.e., northern leatherside distribution where these mean width > 6 m) were sampled with longer small habitat patches occurred. We started reaches (i.e., 400 m) using a raft-mounted in the 2 locations where northern leatherside electrofishing unit (Curry et al. 2009). If condi - we re previously collected and then moved tions precluded effective use of electrofishing upstream and downstream to assess distribu - gear (e.g., high turbidity, extremely high or tion in similar habitats. In contrast, most of low conductivity, equipment malfunction), bag the Bear River drainage contained wadeable sein ing was used to characterize fish assem - streams and small stock ponds within the blages in study reaches. stream network. Therefore, we used a system - In addition to standardized reaches, lentic atic approach where upstream sampling along and lotic habitats (e.g., stock ponds and stream networks ceased when fish assem - reaches consisting of mainly isolated pools) blages were dominated by coldwater species, that appeared to be suitable for fish were because northern leatherside were thought to opportunistically sampled with electrofishing occur primarily downstream of habitats dom - gear or seins to supplement distributional inated by trout and sculpin (Wilson and Belk information. Northern leatherside is known to 2001). We sought to allocate sample points occur in these habitats and may persist longer evenly throughout the study area by select - than other species in these extreme environ - ing sites broadly, starting in proximity to ments (Belk and Johnson 2007). Although we previous collection/sampling locations. As attempted to sample all habitats encountered, sampling progressed through both field sea - permission from landowners and accessibility sons, we added supplemental locations to determined which could be included. increase overall sample sizes and improve All fish captured were identified to species the spatial resolution of sampling. Although and counted following completion of each our sampling did not focus on habitats in the pass. The exception was sculpins (i.e., mottled Green River Basin, we did note collections sculpin Cottus bairdii and Paiute sculpin C. by other Wyoming Game and Fish personnel beldingii ) which are difficult to differentiate in 2010–2011 in the description of northern in the field and were therefore not separated leatherside distribution. to species. The first 30 northern leatherside In the Bear River drainage, we typically and salmonids were measured (total length), quantified fish assemblages in 200-m sample and the length range for other fishes (e.g., reaches (following Barbour et al. 1999). redside shiner Richardsonius balteatus ) was Although we didn’t specifically quantify cap - noted by measuring the longest and shortest ture probability, this sampling approach individuals of each species captured. should allow an accurate assessment of pres - Habitat Associations of Stream Fishes ence/absence of fish within the reach (Patton et al. 2000, Rabeni et al. 2009). In some To evaluate habitat features related to the instances, stream access and other logistical distribution of fishes in the Bear River Basin, constraints resulted in shorter reaches. For we quantified stream habitat variables within example, if a stream reach was positioned wadeable sample reaches during field sam - between a road crossing and a property pling in 2011 (see Schultz and Cavalli 2012 boundary, we adjusted the reach length to for details). At each sample reach, we placed 2016] NORTHERN LEATHERSIDE STATUS AND HABITAT 431

11 evenly spaced transects (1 each at the total number of thalweg depth measurements upstream and downstream ends with 9 in to calculate the percent pool metric. Thalweg between) across the stream channel. At each depth CV was a metric of depth variability in transect we measured wetted stream width. each reach. Water depth and dominant substrate (follow - MULTIVARIATE ANALYSIS OF FISH ASSEM - ing Platts et al. 1983) were measured at 5 BLAGES .–– We used canonical correspondence evenly spaced points across each transect. At analysis (CCA, vegan package in R version each transect, the amounts of aquatic and over - 2.3-4; Oksanen et al. 2016) to analyze the hanging vegetation, woody structure coverage, effect of multiple habitat factors on the catch and bank undercutting were visually esti - rates of different fishes. This technique is a mated within 5 m upstream and downstream direct gradient analysis that uses empirically of the transect. Categories were the following: derived species and environmental data to 0 = absent, 1 = sparse (0%–25%), 2 = medium create multivariate gradients, which maxi - (26%–50%), 3 = heavy (51%–75%), and 4 = mize dispersion among species and among very heavy (76%–100%). To delineate indi - environmental (habitat) variables in ordination vidual channel units and describe substrate space and which identify important variables conditions, we measured thalweg depth and in community composition (ter Braak 1995, substrate at 100 evenly spaced points longi - ter Braak and Verdonschot 1995). Ordination tudinally along each sample reach. Elevation biplots were constructed to help visualize at the midpoint of each sample reach was dispersion of species along eigenvectors obtained from a handheld Global Positioning weighted by environmental variables (ter System (GPS) unit. Some of these habitat fea - Braak and Verdonschot 1995). tures vary seasonally (e.g., vegetation), so it is Fish assemblage, sample reach, and habitat important to note that our inferences were variables were included in the CCA, and initial limited to the factors that were present on the variable screening and data processing were day of sampling. used to maximize interpretability of analyses. For environmental variables, we included Analyses multiple reach-scale habitat variables and one HABITAT DATA SUMMARIES .–– For each sample segment-scale variable (i.e., elevation). From reach, we summarized habitat variables for these potential variables, Pearson correlations species-habitat association analyses. Reach- were used to identify redundancy among scale habitat metrics included maximum and variables. For any pair of redundant variables mean depth, mean substrate size, mean wetted (r > 0.60), we chose one as an explanatory stream width, mean wetted width-to-depth variable based on which was more biologically ratio, aquatic vegetation, bank undercutting, encompassing; the corresponding factor was overhanging vegetation, and woody structure. eliminated from further analyses. For example, Mean substrate size was calculated as the maximum depth and mean depth were corre - mean particle size of the dominant substrate lated ( r = 0.86), so maximum depth was omit - in each quadrat in each sample reach. North - ted in further analyses because mean depth ern leatherside is positively associated with encompassed multiple measurements. To better increased pool habitat and depth variability approximate normality, the mean reach scores (Wesner and Belk 2012, Dauwalter et al. for overhanging vegetation data were logarith - 2014), so we computed 2 additional metrics mically transformed (log 10 + 1); Shapiro–Wilk from our longitudinal depth profile: percent tests indicated that the transformed data bet - pool habitat and thalweg depth coefficient of ter approximated normality. Similar screening variation (CV). We defined pool habitat as procedures were used with species data for where the thalweg profile had a residual inclusion in the CCA. Because we only com - depth measurement greater than one-tenth pleted one electrofishing pass at most sample the mean channel width. We defined residual reaches, we standardized the fish-species matrix depth as the difference in depth between for the CCA to the density of fish (fish per m2 each thalweg profile depth measurement and stream area sampled) of each species cap - the shallowest depth measurement along the tured in the first pass of each sampling reach. profile (Lisle 1987). The number of all points Single-pass and multiple-pass catch-per-unit defined as pool habitat was divided by the effort were strongly correlated in most areas, 432 WESTERN NORTH AMERICAN NATURALIST [Volume 76 and single pass samples provide a reliable variable in reaches where they occurred ( n = approximation of both density and species 35). We examined habitat factors associated diversity (Jones and Stockwell 1995, Bertrand with density, and species that tended to co- et al. 2006, Reid et al. 2009). To reduce the occur with northern leatherside. confounding influence of rare species, species Prior to implementing statistical models, were only included in the CCA if present in we performed additional variable screening. >5% of all the sample reaches. Of the 108 sample reaches where we mea - Initially, all candidate variables were sured habitat in 2011, we omitted 14 fishless included in the CCA, but variables were reaches from multivariate analyses and species removed from the analysis post hoc if they distribution modeling. These reaches were all did not explain variation along major axes in very small (<1 m mean width) and may have an easily interpretable way (ter Braak and been fishless and ephemeral naturally or due Verdonshot 1995). In addition to our initial to water diversion impacts. Similarly, northern variable screening, we eliminated variables leatherside were not collected in any of the 26 with variance inflation factors (VIFs) >20 sample reaches in the Thomas Fork drainage. from the CCA (ter Braak and Verdonschot We assumed that northern leatherside was not 1995). The final CCA from this iterative pro - available in the potential species pool within cess was used for interpretation of species- this drainage, so these reaches were also omit - habitat associations. ted from habitat association analyses. After NORTHERN LEATHERSIDE HABITAT AND this screening process, our total sample size BIOTIC ASSOCIATIONS .–– To examine factors for multivariate and species distributional related to the distribution of northern models was 68 reaches. leatherside in Wyoming, we used logistic regression to model occurrence as a function RESULTS of reach-scale habitat variables and fish Northern Leatherside Distribution assemblage variables (Meyer et al. 2010, Schultz et al. 2015). We used habitat and We collected northern leatherside from biotic factors that the CCA identified as asso - several portions of the Bear River and Upper ciated with northern leatherside to develop a Snake River Basins in western Wyoming set of predictive variables to explain the dis - (Fig. 1b); the limited sampling in the Green tribution of northern leatherside. To identify River Basin also yielded northern leatherside explanatory variables from these candidates, from 2 locations (details in Schultz and Cav - we used a stepwise backward-elimination alli 2012). We captured northern leatherside procedure to remove variables with little from a variety of habitats across all areas sam - explanatory power (R version 2.3-4; R Core pled, but we did not collect it from standing Development Team 2011). At each step, we waters larger than 1 ha. Northern leatherside used Akaike’s information criterion (AIC) to occurred in densities ranging from 0.0015 to evaluate the relative support for each model. 0.197 fish ⋅ m−2. Numerous length classes We followed the recommendations of Arnold were represented in most populations, and (2010) and computed 85% confidence inter - length ranged from 25 to 137 mm. The highest vals for parameter estimates. If 85% confi - densities of northern leatherside were col - dence intervals included zero, parameters lected in Sulphur Creek (up to 0.197 fish ⋅ were interpreted to be uninformative. We m−2) and Dry Fork (up to 0.114 fish ⋅ m−2). assessed logistic regression model performance Extensive sampling (26 sample reaches) in based on a receiver operating characteristic the Thomas Fork drainage failed to detect (ROC) curve, where ROC values 0.5–0.7 northern leatherside. indicated low, 0.7–0.9 indicated medium, Northern leatherside was found in rela - and 0.9–1.0 indicated high model accuracy tively small patches within this broad distri - (Manel et al. 2001). bution in the Twin Creek, Upper Bear River, To assess factors potentially related to and Smiths Fork drainages. Populations in northern leatherside abundance, we used sim - each of these drainages were located over 30 ple Pearson correlations to identify relation - river kilometers from populations in adjacent ships between northern leatherside abundance drainages. Some populations appeared to (density) and each measured environmental be relatively isolated within drainages. For 2016] NORTHERN LEATHERSIDE STATUS AND HABITAT 433

TABLE 1. Weighted correlation matrix of stream habitat variables and species values for stream fish assemblages in the Bear River Basin, southwestern Wyoming, 2011. Eigenvalues equal the dispersion of species scores on the axis and indicate the importance of the axis. Habitat factors in bold are strongly associated with the corresponding canonical axes. For interpretive purposes, species values that load strongly on a given axis are associated with habitat factors that also load strongly on that axis. For example, Bonneville cutthroat trout loaded strongly on axis 1 and were positively associated with depth coefficient of variation and overhanging vegetation and negatively associated with mean stream width and depth.

______C__a_n_o_n_ic_a_l_ a_x_i_s ______Variable 1 2 3 Habitat factors Depth coefficient of variation 0.66 0.43 0.24 Mean stream width −0.73 0.20 −0.24 Mean depth −0.81 0.33 −0.17 % Pool habitat 0.43 0.46 0.08 Aquatic vegetation −0.42 −0.36 0.67 Woody structure 0.33 −0.31 0.16 Overhanging vegetation 0.49 −0.34 −0.22 Species Brown trout Salmo trutta −0.13 −2.26 −0.32 Brook trout Salvelinus fontinalis −1.06 −1.04 0.02 Bonneville cutthroat Oncorhynchus clarkii utah 1.13 −0.54 −0.87 Northern leatherside chub Lepidomeda copei 0.58 0.26 0.33 Sculpins Cottus spp. −0.09 −0.19 0.15 Mountain sucker Catostomus platyrhynchus 0.61 0.21 −0.02 Speckled dace Rhynichthys osculus 0.10 0.09 −0.15 Longnose dace Rhinichthys cataractae −0.67 −0.32 0.32 Redside shiner Richardsonius balteatus −0.36 0.20 0.04 Mountain whitefish Prosopium williamsoni −1.58 1.34 −2.24 Fathead minnow Pimephales promelas −1.50 0.29 −0.53 Utah sucker Catostomus ardens −0.95 0.73 −0.81 Eigenvalue (% variation explained) 0.192 (39.5) 0.163 (33.4) 0.074(15.3) example, northern leatherside populations in at 35 (frequency of occurrence 32%). In these Muddy Creek and Dry Fork (both in the collections, speckled dace Rhinichthys osculus Smiths Fork drainage) were separated by was the most frequently occurring species more than 20 river kilometers. (56%), followed by sculpins (52%), Bon - The majority of northern leatherside popu - neville cutthroat trout Oncorhynchus clarki lations were in the Bear River Basin. In addi - utah (51%), and mountain sucker Catostomus tion to sampling in the Snake River and its platyrhynchus (46%). tributaries downstream of Jackson Lake Dam, MULTIVARIATE ANALYSIS OF FISH ASSEM - we also collected northern leatherside from BLAGES .— Variable screening resulted in the portions of Pacific Creek and Triangle X Creek retention of 7 explanatory variables in our (a new location for the species). final CCA: thalweg depth coefficient of varia - tion, mean depth, mean stream width, percent Habitat and Biotic Associations pool habitat, amount of aquatic vegetation, Fish assemblage and stream habitat sam - presence of woody structure, and amount of pling in 2011 represented typical habitats of overhanging vegetation. Variables included in foothills streams in the Bear River Basin of the final CCA were not redundant; no southwestern Wyoming. Mean stream width retained variables had VIFs >5. The first 2 ranged from 0.5 to 14.7 m and elevation canonical axes of our CCA explained 39.5% ranged from 1971 to 2507 m. As is typical in and 33.4% of variation in species abundance Mountain West streams, fish assemblages among sample reaches, respectively (Table 1). transitioned from cyprinid-catostomid assem - The first axis was correlated with the depth blages in downstream reaches to coldwater coefficient of variation, mean depth, and mean trout-sculpin assemblages in higher gradient stream width (Fig. 2). The second axis was and colder reaches. Of the 108 reaches we correlated with percent pool habitat and sampled, northern leatherside was collected amount of aquatic vegetation. The third CCA 434 WESTERN NORTH AMERICAN NATURALIST [Volume 76

Fig. 2. Canonical correspondence analysis ordination of the distribution of fishes in the Bear River Basin, southwest - ern Wyoming. Habitat variables are represented as vectors, and species are represented by 3-letter species codes: BNT = brown trout, BKT = brook trout, BRC = Bear River (Bonneville) cutthroat, LND = longnose dace, MSC = sculpins, SPD = speckled dace, FHM = fathead minnow, RSS = redside shiner, MTS = mountian sucker, LSC = northern leatherside chub, UTS = Utah sucker, MWF = mountian whitefish. Habitat vectors project toward higher values for each variable. Species centroids can be interpreted relative to these vectors. Centroids parallel with habitat vectors indicate strong relationships with that habitat factor; centroids perpendicular to habitat vectors are unrelated to those habitat factors. Longer habitat vectors indicate more influence in explaining variation in assemblage structure relative to shorter vectors. For example, Utah sucker and mountain whitefish are positively associated with deeper/wider reaches with less overhanging vegetation but are neutrally associated with aquatic vegetation and pool habitat.

TABLE 2. Parameter estimates ( bi), standard error, and western Wyoming. The centroids of sculpins 85% confidence intervals for a logistic regression model and speckled dace were positioned close to the that predicts occurrence of northern leatherside chub in the Bear River Basin of southwestern Wyoming, 2011. origin, indicating no strong association with any Thalweg depth CV is the coefficient of variation for habitat factor. Bonneville cutthroat trout was depths measured longitudinally along stream reaches. associated with overhanging vegetation and 85% Confidence shallower/narrower stream channels. Mountain Variable bi SE limits sucker and northern leatherside were both associated with increased thalweg depth coef - (Intercept) −7.18 2.006 −10.31, 0.012 Mean substrate −0.01 0.0038 −0.016, 0.99 ficient of variation and percent pool habitat. size (mm) Other species were less common and displayed Thalweg depth CV 11.38 3.220 7.04, 16.38 variable habitat associations (Fig. 2). Mean depth (cm) 0.127 0.0367 0.08, 1.20 NORTHERN LEATHERSIDE HABITAT AND BIOTIC ASSOCIATIONS .— Using the backward- elimination procedure, we selected a model axis was correlated with aquatic vegetation, containing 3 predictor variables associated but only explained 15.3% of variation in with the occurrence of northern leatherside: assemblage structure. Within the CCA, thal - mean substrate size, thalweg depth coeffi - weg depth coefficient of variation, mean cient of variation, and mean depth (Table 2). depth, and stream width showed the strongest However, confidence intervals (85%) for mean gradients, while aquatic vegetation, percent substrate size included zero, suggesting this pool habitat, and woody structure had rela - parameter was uninformative. Northern leather - tively weak gradients. side tended to occur at sites with greater Centroids of species distributions elucidated mean depth, higher depth variability, and finer habitat associations of stream fishes in south - substrates (Table 2). The ROC value for the 2016] NORTHERN LEATHERSIDE STATUS AND HABITAT 435

TABLE 3. Pearson correlation coefficients ( r) for correlations between northern leatherside chub abundance and (1) habitat features and (2) abundance of other fishes in the Bear River Basin, 2011. “Width” is mean stream width and “Elev” is the elevation of the sample reach. Species codes include the following: MSC = sculpins, MTS = mountain sucker, RSS = redside shiner, SPD = speckled dace, BRC = Bear River (Bonneville) cutthroat trout, BNT = brown trout, LND = longnose dace.

______H__a_b_i_ta_t_ f_e_a_tu_r_e_s______Aq_veg CVdepth Mean depth Overhang %Pool Width Wood Elev r 0.12 0.49 −0.32 −0.05 0.11 −0.22 0.08 0.41

______F_i_s_h_ a_s_s_e_m_b_l_a_g_e_ v_a_r_ia_b_l_e_s ______MSC MTS RSS SPD BRC BNT LND r 0.03 0.34 0.38 0.38 0.08 0.05 −0.01 selected model was 0.84, suggesting accept - were concentrated in small areas within this able model performance. broad distribution (see Schultz and Cavalli At reaches where northern leatherside 2012 for details). The distribution of northern occurred, its abundance was related to depth leatherside in Wyoming represents more than variability ( r = 0.415) and elevation ( r = half of the known rangewide distribution of 0.229). Northern leatherside abundance also the species (USFWS 2011), which presents showed a positive association with abundance unique conservation challenges and opportu - of 3 native nongame fishes: mountain sucker nities for Wyoming. (r = 0.34), redside shiner ( r = 0.38), and The current distribution of northern leather - speckled dace ( r = 0.38; Table 3). side in Wyoming relative to its ostensible historical distribution (Zafft et al. 2009) DISCUSSION appears to be relatively intact when compared to other imperiled fishes of the Intermountain This study documented the distribution West. In contrast, native bluehead sucker and habitat associations of northern leather - Catostomus discobolus, flannelmouth sucker side and other stream fishes in southwestern C. latipinnis, and roundtail chub Gila robusta Wyoming. Although the coarse distribution occur in less than half of their historical range of northern leatherside in this study differed (Bezzerides and Bestgen 2002), and Colorado little from Zafft et al. (2009), our work pro - River cutthroat trout Oncorhynchus clarkii vides a finer-scale delineation of the species’ pleuriticus occupies approximately 14% of its distribution and provides measures of relative historical range (Hirsch et al. 2006). Northern abundance in the state. The size of the distri - leatherside does not appear to have been bution did not appear to have diminished from extirpated from any of its confirmed historical that of the previously known distribution, and distribution in Wyoming (Zafft et al. 2009). we documented additional populations (Fig. Because of the species’ intact distribution, we 1). This information is directly applicable to can infer that Wyoming populations of north - conservation and management efforts directed ern leatherside (1) are likely fairly resilient, towards northern leatherside in Wyoming (2) represent core populations of the species, (Schultz and Cavalli 2012). Furthermore, our or (3) are exposed to minimal threats in habitat analyses can address information gaps Wyoming. For these reasons, Wyoming popu - and inform management efforts across the lations of northern leatherside could play a range of northern leatherside. These habitat critical role in conservation of the species analyses are also applicable to management across its range. In addition to providing fish of co-occurring nongame fishes across the for translocations and reintroductions, man - expanse of the Intermountain West, species agement work in Wyoming may provide valu - for which little information exists. able information that is broadly applicable (e.g., UDWR 2010). For instance, transloca - Northern Leatherside Distribution tions within Wyoming might be used to Northern leatherside was widely distributed develop predictive tools for assessing suitable across western Wyoming. Most individuals reintroduction sites in other areas. Wyoming 436 WESTERN NORTH AMERICAN NATURALIST [Volume 76 populations may also be studied to examine very similar to those described in other assess - the life history of the species (e.g., growth ments. Sculpins and speckled dace were posi - rates, movement, swimming performance, tioned near the origin of our ordination biplot. response to habitat perturbation/restoration) Because these small-bodied fishes are riffle to supplement existing studies on some of dwellers (Baxter and Stone 1995), suitable these questions (Billman et al. 2008a, 2008b). habitat is likely available in small patches The broadly spread but locally abundant within most 200-m reaches. Redside shiner distribution of northern leatherside is charac - was associated with deeper and wider stream teristic of imperiled fishes (Haak et al. 2010). reaches, possibly because it is more produc - Connectivity of these population patches tive in warmer habitats (Reeves et al. 1987). might present issues for the long-term persis - Bonneville cutthroat trout was associated with tence of northern leatherside. Hilderbrand narrow, shallower stream reaches with wood and Kershner (2000) found that inland cut - cover, similar to results from other studies throat trout required at least 8 km of stream (e.g., White and Rahel 2008). However, the for high-density population persistence, and habitat associations for Bonneville cutthroat more for lower-density populations. As a are likely not truly reflective of its preferences small-bodied fish, northern leatherside likely across the Bear River Basin because only a requires less stream distance than larger narrow portion of the entire habitat it uses salmonids (Pyron 1999). Because most popula - was sampled. Utah sucker Catostomus ardens tions contained multiple size (and presumably and mountain whitefish Prosopium williamsoni age) classes, northern leatherside in these were associated with larger stream reaches, as patches currently appear to have all necessary would be expected of these larger-bodied habitat elements to complete their life history. insectivores (Moyle and Herbold 1987). Over evolutionary time, however, isolation Northern leatherside responded to habitat and lack of connectivity between adjacent conditions in small streams in patterns gener - populations might present long-term genetic ally similar to those observed for southern diversity or population viability issues due to leatherside. Occurrence of northern leather - stochastic demographic processes (e.g., Perkin side was explained by depth variability, mean et al. 2015). Furthermore, seasonal use pat - depth, and (weakly) by substrate variability. terns (Schultz 2014) suggest that northern Southern leatherside is typically found in leatherside may be more mobile than origi - low-velocity habitats with fine substrate nally thought. Maintaining and restoring con - (Walser et al. 1999). It is vulnerable to habitat nectivity between population patches will changes induced by channelization because increase the likelihood of viability of northern the loss of lateral habitats forces it to occupy leatherside in Wyoming, with concomitant microhabitats containing predators (Olsen benefits to other associated native fishes. and Belk 2005, Belk and Johnson 2007). Although northern leatherside was occasion - Habitat Associations ally collected from riffle habitats, its occur - Multivariate and logistic modeling showed rence was associated with increased mean congruent results, allowing strong inferences depth—habitats that contain slower water about important habitat components for velocities. In larger, complex streams, north - northern leatherside and other native fishes ern leatherside used off-channel habitats of southwestern Wyoming. Northern leather - more frequently. Similar to the Upper Snake side occurred in a variety of habitats, but River populations from this study, southern showed a strong association with highly vari - leatherside in Diamond Fork Creek, Utah, was able stream depths and several other native found nearly exclusively in lateral habitats nongame fishes (Schultz and Cavalli 2012). (Walser et al. 1999). These findings can help prioritize conserva - Northern leatherside showed habitat asso - tion areas for northern leatherside, as well as ciations similar to those of mountain sucker help identify suitable areas for translocation and other native fishes. These species were (UDWR 2009, 2010). generally associated with higher depth vari - Our multivariate analyses identified habitat ability and pool habitat, and were found in components structuring fish assemblages in relatively narrower stream reaches. In Utah, the Bear River Basin of Wyoming that were mountain sucker has been suggested as a 2016] NORTHERN LEATHERSIDE STATUS AND HABITAT 437 suitable indicator of leatherside chub occur - within a stream reach was also supportive of rence (M. Grover, Utah Division of Wildlife greater species diversity, as has been shown Resources, personal communication), and elsewhere (Schlosser 1982). recent evidence suggests that both species Our results indicate that the distribution respond similarly to introduced salmonids and abundance of northern leatherside were (Wilson and Belk 2001, Schultz and Bertrand related to habitat conditions in the Bear River 2012, Schultz et al. 2015). Similarly, Quist et Basin, and that other native fishes also shared al. (2004) found that northern leatherside these associations. In these small foothills occurred with catostomids in warmer low- streams, watershed processes that maintain elevation stream segments with abundant and enhance natural channel form (Rosgen deep-pool habitat in the Salt River drainage of 1994) can be promoted to enhance habitats Idaho and Wyoming. Northern leatherside for these native fishes. Management actions has also been linked with higher fish species to achieve these conditions include maintain - diversity in other portions of the Great Basin ing water quantity and quality in stream (Wesner and Belk 2012). Though it is uncer - channels for natural hydrologic function, pro - tain whether this is driven by habitat condi - tecting and restoring native riparian vegeta - tions or factors related to the associated fish tion, and improving grazing management assemblage, these studies suggest that conser - (Kauffman et al. 1997). Riparian grazing vation measures directed at any of these fishes exclosures, a common management tool in should be mutually beneficial to the entire western Wyoming, effectively remedy histori - native nongame assemblage. cally degraded stream conditions (Ranganath In reaches where it occurred, abundance et al. 2009, Batchelor et al. 2015). Focused of northern leatherside varied as a function of grazing management is another tool available depth variability and elevation; higher eleva - to rangeland managers to meet objectives of tion and more variable depths supported riparian habitat recovery without construction higher abundances. Our study contrasts with of grazing exclosures (Swanson et al. 2015). the findings of Wilson and Belk (2001), who Work on northern leatherside in the Goose were unable to detect any reach-scale factors Creek Basin, Idaho, also suggests that resto- related to abundance of southern leatherside ra tion of beavers can also be an important but did find an elevational limit to its distri - natural source of habitat complexity in stream bution. They attributed the lack of strong ecosystems (Dauwalter et al. 2014). While relationships to the broad range of environ - many of these actions are currently being mental conditions that southern leatherside implemented, this study provides additional evolved in and suggested that the elevational support for these ongoing activities. pattern observed was reflective of longitudi - nal thermal patterns of stream systems. In our ACKNOWLEDGMENTS study, reaches in the Upper Bear River drainage were at higher elevations than most This study was funded by a state wildlife other reaches sampled and also exhibited grant to the Wyoming Game and Fish Depart - some of the highest population densities we ment. Discussions with D. Miller, P. Dey, and C. observed. Because our upstream sampling Amadio improved study design, and guidance ceased when assemblages transitioned to from J. Luginbill greatly facilitated sampling. trout without northern leatherside presence, Fieldwork was completed by N. Thompson, an elevation gradient likely does constrain K. Buer, J. Luginbill, W. Gordon, M. Horwitz, northern leatherside distribution, but we J. Stratton, P. Baigas, J. Peterson, B. Hines, T. were unable to detect it. We also hypothe - Stephens, M. Devine, J. Blakney, S. Walker, C. sized that northern leatherside would be Girard, T. Neebling, C. Matthews, J. Bailey, S. associated with naturally functioning stream Green, G. Edwards, P. Mathias, K. Gelwicks, channels, which create a diversity of habitats B. Compton, and W. Stacy. Access logistics for multiple life history stages (Wesner and were facilitated by N. Hymas. Access to pri - Belk 2012, Dauwalter et al. 2014). The positive vate lands was provided by numerous landown - rela tionship between depth variability and ers throughout western Wyoming. We thank northern leatherside abundance supported J. Dunham, S. Beldin, and M. Fitzpatrick at this hypothesis. Higher habitat diversity USGS FRESC for administrative assistance to 438 WESTERN NORTH AMERICAN NATURALIST [Volume 76 complete this manuscript. Finally, C. Amadio, proceedings. Symposium 53, American Fisheries B. Bradshaw, J. Burckhardt, B. Compton, M. Society, Bethesda, MD. Fitzpatrick, T. Stephens, and 5 anonymous COOKE , S.J., C.M. B UNT , S.J. H AMILTON , C.A. J ENNINGS , M.P. P EARSON , M.S. C OOPERMAN , AND D.F. M ARKLE . reviewers provided input on prior drafts of 2005. Threats, conservation strategies, and prognosis the manuscript. Any use of trade, product, or for suckers (Catostomidae) in North America: firm names is for descriptive purposes only insights from regional case studies of a diverse fam - and does not imply endorsement by the U.S. ily of non-game fishes. Biological Conservation 121:317–331. Government. CURRY , R.A., R.M. H UGHES , M.E. M CMASTER , AND D.J. ZAFFT . 2009. Coldwater fish in rivers. Pages 139–158 LITERATURE CITED in S.A. Bonar, W.A. Hubert, and D.W. Willis, editors, Standard methods for sampling North American ARNOLD , T.W. 2010. Uninformative parameters and model freshwater fishes. American Fisheries Society, selection using Akaike’s information criterion. Jour - Bethesda, MD. nal of Wildlife Management 74:1175–1178. DAUWALTER , D.C., S.J. W ENGER , AND P. G ARDNER . 2014. BARBOUR , M.T., J. G ERRITSEN , B.D. S NYDER , AND J.B. The role of complexity in habitat use and selection STRIBLING . 1999. Rapid bioassessment protocols for by stream fishes in a Snake River Basin tributary. use in streams and wadeable rivers: periphyton, Transactions of the American Fisheries Society benthic macroinvertebrates and fish. 2nd edition. 143:1177–1187. U.S. Environmental Protection Agency 841-B-99- DUNHAM , J., C. B AXTER , K. F AUSCH , W. F REDENBERG , S. 002, Washington, DC. KITANO , I. K OIZUMI , K. M ORITA , T. N AKAMURA , B. BATCHELOR , J.L., W.J. R IPPLE , T.M. W ILSON , AND L.E. RIEMAN , AND K. S AVVAITOVA , ET AL . 2008. Evolution, PAINTER . 2015. Restoration of riparian areas follow - ecology, and conservation of Dolly Varden, white ing the removal of cattle in the northwestern Great spotted char, and bull trout. Fisheries 33:537–550. Basin. Environmental Management 55:930–942. HAAK , A.L., J.E. W ILLIAMS , H.M. N EVILLE , D.C. D AUWAL - BAXTER , G.T., AND M.D. S TONE . 1995. Fishes of Wyoming. TER , AND W.T. C OLYER . 2010. Conserving peripheral Wyoming Game and Fish Department, Cheyenne, populations: the values and risks of life on the edge. WY. Fisheries 35:530–549. BELK , M.C., AND J.B. J OHNSON . 2007. Biological status HILDERBRAND , R.H., AND J.L. K ERSHNER . 2000. Conserv - of leatherside chub: a framework for conservation ing inland cutthroat trout in small streams: how of western freshwater fishes. Pages 67–76 in M.J. much stream is enough? North American Journal of Brouder and J.A. Scheurer, editors, Status, distri- Fisheries Management 20:513–520. bution, and conservation of native freshwater fishes HIRSCH , C.L., S.E. A LBEKE , AND T.P. N ESLER . 2006. of western North America: a symposium proceed - Range-wide status of Colorado River cutthroat trout ings. American Fisheries Society, Symposium 53, (Oncorhynchus clarkii pleuriticus ): 2005. Colorado Bethesda, MD. Division of Wildlife, Denver, CO. BERTRAND , K.N., K.B. G IDO , AND C.S. G UY . 2006. An evalua - JELKS , H.L., S.J. W ALSH , N.M. B URKHEAD , S. C ONTRERAS - tion of single-pass versus multiple-pass backpack BALDERAS , E. D IAZ -P ARDO , D.A. H ENDRICKSON , J. electrofishing to estimate trends in species abun - LYONS , N.E. M ANDRAK , F. M CCORMICK , AND J.S. dance and richness in prairie streams. Transactions NELSON , ET AL . 2008. Conservation status of imper - of the Kansas Academy of Science 109:131–138. iled North American freshwater and diadromous BEZZERIDES , N., AND K.R. B ESTGEN . 2002. Status review fishes. Fisheries 33:372–407. of roundtail chub Gila robusta , flannelmouth sucker JOHNSON , J.B., T.E. D OWLING , AND M.C. B ELK . 2004. Catostomus latipinnis , and bluehead sucker Catosto - Neglected taxonomy of rare desert fishes: congruent mus discobolus in the Colorado River basin. Contri - evidence for two species of leatherside chub. Sys - bution 118, Larval Fish Laboratory, Department of tematic Biology 53:841–855. Fishery and Wildlife Biology, Colorado State Univer - JONES , M.L., AND J.D. S TOCKWELL . 1995. A rapid assess - sity, Fort Collins, CO. ment procedure for the enumeration of salmonine BILLMAN , E.J., E.J. W AGNER , AND R.E. A RNDT . 2008a. populations in streams. North American Journal of Reproductive ecology and spawning substrate pref - Fisheries Management 15:551–562. erence of the northern leatherside chub. North KAUFFMAN , J.B., R.L. B ESCHTA , N. O TTING , AND D. L YT - American Journal of Aquaculture 70:273–280. JEN . 1997. An ecological prospective of riparian and BILLMAN , E.J., E.J.W AGNER , R.E. A RNDT , AND E. stream restoration in the Western United States. VAN DYKE . 2008b. Optimal temperatures for growth Fisheries 22:12–24. and upper thermal tolerance of juvenile northern LISLE , T.E. 1987. Using “residual depths” to monitor pool leatherside chub. Western North American Naturalist depths independently of discharge. USDA Forest 68:463–474. Service. General Technical Report RN-394, Pacific BLAKNEY , J.R., J.L. L OXTERMAN , AND E.R. K EELEY . 2014. Southwest Forest and Range Experimental Station, Range-wide comparisons of northern leatherside Berkeley, CA. chub populations reveal historical and contemporary MANEL , S., H.C. W ILLIAMS , AND S.J. O MEROD . 2001. Evalu - patterns of genetic variation. Conservation Genetics ating presence-absence models in ecology: the need 15:757–770. to account for prevalence. Journal of Applied Ecol - BROUDER , M.J., AND J.A. S CHEURER , EDITORS . 2007. Sta - ogy 38:921–931. tus, distribution, and conservation of native freshwa - MEYER , K.A., J.A. L AMANSKY JR., AND D.J. S CHILL . 2010. ter fishes of western North America: a symposium Biotic and abiotic factors related to redband trout 2016] NORTHERN LEATHERSIDE STATUS AND HABITAT 439

occurrence and abundance in desert and montane R C ORE DEVELOPMENT TEAM . 2011. R: a language and streams. Western North American Naturalist 70:77–91. environment for statistical computing. R Foundation MINCKLEY , W.L., AND J.E. D EACON ,EDITORS . 1991. The for Statistical Computing, Vienna, Austria. http:// battle against extinction: native fish management in www.R-project.org/ the American West. University of Arizona Press, REEVES , G.H., F.H. E VEREST , AND J.D. H ALL . 1987. Inter - Tucson, AZ. actions between the redside shiner ( Richardsonius MOYLE , P.B., AND B. H ERBOLD . 1987. Life-history patterns balteatus ) and the steelhead trout ( Salmo gairdneri ) and community structure in stream fishes of western in Western Oregon: the influence of water tempera - North America: comparisons with eastern North ture. Canadian Journal of Fisheries and Aquatic America and Europe. Pages 25–32 in W.J. Matthews Sciences 44:1603–1613. and D.C. Heins, editors, Community and evolu - REID , S.M., G. Y UNKER , AND N.E. J ONES . 2009. Evalua - tionary ecology of North American stream fishes. tion of single-pass electric fishing for stream fish University of Oklahoma Press, Norman, OK. community monitoring. Fisheries Management and NATURE SERVE . 2011. NatureServe Explorer: an online Ecology 16:1–9. encyclopedia of life [web application]. Version 7.1. ROSGEN , D.L. 1994. A classification of natural rivers. NatureServe, Arlington, VA; [accessed 19 April Catena 22:169–199. 2011]. http://www.natureserve.org/explorer SCHLOSSER , I.J. 1982. Fish community structure and func - OKSANEN , J., F.G. B LANCHET , R. K INDT , P. L EGENDRE , tion along two habitat gradients in a headwater P.R. M INCHIN , R.B. O’H ARA , G.L. S IMPSON , P. S OLY - stream. Ecological Monographs 52:395–414. MOS , M.H.H. S TEVENS , AND H. W AGNER . 2016. SCHULTZ , L.D. 2014. Use of seasonally available habitat by Vegan: community ecology package [software]. R northern leatherside in Wyoming. Southwestern package version 2.3-4. http://CRAN.R-project.org/ Naturalist 59:118–120. package=vegan SCHULTZ , L.D., AND K.N. B ERTRAND . 2012. Long-term OLSEN , D.G., AND M.C. B ELK . 2005. Relationship of diur - trends and outlook for mountain sucker in the Black nal habitat use of native stream fishes of the eastern Hills of South Dakota. American Midland Naturalist Great Basin to the presence of introduced sal- 167:96–110. monids. Western North American Naturalist 65: SCHULTZ , L.D., K.N. B ERTRAND , AND B.D.S. G RAEB . 2015. 501–506. Factors from multiple scales influence the distribu - PATTON , T.M., W.A. H UBERT , F.J. R AHEL , AND K.G. G EROW . tion and abundance of an imperiled fish—mountain 2000. Effort needed to estimate species richness in sucker in the Black Hills of South Dakota, USA. small streams on the Great Plains in Wyoming. Environmental Biology of Fishes 99:3 –14. North American Journal of Fisheries Management SCHULTZ , L.D., AND P. C AVALLI . 2012. Distribution and 20:394–398. habitat associations of northern leatherside ( Lep - PERKIN , J.S., K.B. G IDO , A.R. C OOPER , T.F. T URNER , M.J. idomeda copei ) in Wyoming. Wyoming Game and OSBOURNE , E.R. J OHNSON , AND K.M. M AYES . 2015. Fish Department Administrative Report, Cheyenne, Fragmentation and dewatering transform Great WY. Plains stream fish communities. Ecological Mono - SIGLER , W.F., AND J.W. S IGLER . 1996. Fishes of Utah: a graphs 85:73–92. natural history. University of Utah Press, Salt Lake PLATTS , W.F., W.F. M EGAHAN , AND G.W. M INSHALL . 1983. City, UT. Methods for evaluating stream, riparian, and biotic SWANSON , S.R., S. W YMAN , AND C. E VANS . 2015. Practical conditions. General Technical Report INT-138, grazing management to meet riparian objectives. USDA Forest Service, Intermountain Forest and Journal of Rangeland Applications 2:1–28. Range Experimental Station, Ogden, UT. TER BRAAK , C.J.F. 1995. Ordination. Pages 91–173 in PYRON , M . 1999. Relationships between geographical size R.H.G. Jongman, C.J.F. ter Braak, and O.F.R. van range, body size, local abundance, and habitat Tongeren, editors, Data analysis in community and breadth in North American suckers and sunfishes. landscape ecology. Cambridge University Press, Journal of Biogeography 26:549–558. United Kingdom. QUIST , M.C., W.A. HUBERT , AND D.J. I SAAK . 2004. Fish TER BRAAK , C.J.F., AND P.F.M. V ERDONSCHOT . 1995 . assemblage structure and relationships with envi - Canonical correspondence analysis and related ronmental conditions in a Rocky Mountain water - multivariate methods in aquatic ecology. Aquatic shed. Canadian Journal of Zoology 82:1554–1565. Sciences 57:255–289. QUIST , M.C., F.J. R AHEL , AND W.A. H UBERT . 2005. Hierar - [USFWS] U NITED STATES FISH AND WILDLIFE SERVICE . chical faunal filters: an approach to assessing effects 2011. 12-month finding for a petition to list the of habitat and nonnative species on native fishes. northern leatherside chub as endangered or threat - Ecology of Freshwater Fish 14:24–39. ened. Federal Register 76:63444–63478. RABENI , C.R., J. L YONS , N. M ERCADO -S ILVA , AND J.T. [UDWR] U TAH DIVISION OF WILDLIFE RESOURCES . PETERSON . 2009. Warmwater fish in wadeable 2009. Rangewide conservation agreement and streams. Pages 43–58 in S.A. Bonar, W.A. Hubert, strategy for northern leatherside chub ( Lep - and D.W. Willis, editors, Standard methods for sam - idomeda copei ). Publication number 09-11, Salt pling North American freshwater fishes. American Lake City, UT. Fisheries Society, Bethesda, MD. [UDWR] U TAH DIVISION OF WILDLIFE RESOURCES . 2010. RANGANATH , S.C., W.C. H ESSION , AND T.M. W YNN . 2009. Northern leatherside chub ( Lepidomeda copei ) Livestock exclusion influences on riparian vegeta - monitoring summary, statewide 2010. Publication tion, channel morphology, and benthic macroinver - number 11-20, Salt Lake City, UT. tebrate assemblages. Journal of Soil and Water WALSER , C.A., M.C. B ELK , AND D.K. S HIOZAWA . 1999. Conservation 64:33–42. Habitat use of leatherside chub ( Gila copei ) in the 440 WESTERN NORTH AMERICAN NATURALIST [Volume 76

presence of predatory brown trout ( Salmo trutta ). WILSON , K.W., AND M.C. B ELK . 2001. Habitat character - Great Basin Naturalist 59:272–277. istics of leatherside chub ( Gila copei ) at two spatial WESNER , J.S., AND M.C. B ELK . 2012. Habitat relationships scales. Western North American Naturalist 61: among biodiversity indicators and co-occurring 36–42. species in a freshwater fish community. Animal [WGFD] W YOMING GAME AND FISH DEPARTMENT . 2010. Conservation 15:445–456. Wyoming state wildlife action plan—2010. Wyoming WHITE , S.M., AND F.J. R AHEL . 2008. Complementation of Game and Fish Department, Cheyenne, WY habitats for Bonneville cutthroat trout in watersheds ZAFFT , D.J., C. A MADIO , P. C AVALLI , H. S EXAUER , R. G IP - influenced by beavers, livestock and drought. Trans - SON , AND D.D. M ILLER . 2009. Northern leatherside actions of the American Fisheries Society 137: distribution in Wyoming. Wyoming Game and Fish 881–894. Department Administrative Report, Cheyenne, WY. WILLIAMS , J.E., R.N. W ILLIAMS , R.F. T HUROW , L. ELWELL , D.P. P HILLIP , F.A. H ARRIS , J.L. K ERSHNER , Received 8 January 2015 P.J. M ARTINEZ , D. M ILLER , AND G.H. R EEVES , ET AL . Accepted 19 July 2016 2011. Native fish conservation areas: a vision for large-scale conservation of native fish communities. Fisheries 36:267–277.