Abstract
The Wood River Basin in central Idaho has been isolated from the surrounding Snake River Basin by Malad Gorge Falls for at least 50,000 years, and recent genetic analyses suggest that Redband Trout Oncorhynchus mykiss gairdneri in the basin represent a previously undescribed lineage. To assess their contemporary status, we electrofished 22 study reaches in 2021–2022 previously occupied by Redband Trout when originally surveyed in 2003. Our objective was to assess changes in the occurrence and density of Redband Trout, Brook Trout Salvelinus fontinalis, and Brown Trout Salmo trutta. In 2021–2022, Redband Trout were absent in five of the 22 originally occupied reaches with Brook Trout as the only salmonid species present at all extirpated reaches. Brook Trout have now been extirpated from two reaches previously occupied with Redband Trout as the only salmonid species present at both reaches. Brown Trout colonized one new reach since 2003 and are now present at all study reaches (n = 3) which exceed 10 m average stream wetted width. Average fish density increased from 2003 to 2021–2022 across all study reaches for total trout (all species combined) and at nearly all study reaches for each individual trout species. Changes in Redband Trout density were unrelated to changes in nonnative trout density. On average, Redband Trout composition (of all trout at each reach) across all study reaches decreased by 13% whereas Brook Trout and Brown Trout increased by 8% and 4%, respectively, from 2003 to 2021–2022. Nevertheless, Brook Trout are increasingly present in headwater streams, whereas Brown Trout occupy larger, lower elevation rivers, potentially threatening the long-term conservation of Redband Trout in the Wood River Basin. Although additional data is warranted to more thoroughly understand the distribution of trout, these preliminary results suggest that management activities may be necessary to control the further spread of nonnative salmonids in the basin.
Introduction
The native range of interior Redband Trout Oncorhynchus mykiss gairdneri encompasses a wide variety of habitats and climatic regimes ranging from high alpine to low desert elevations across six states in the United States of America and two Canadian provinces in western North America (Behnke 2002). Historically (circa 1800), Redband Trout occupied over 60,000 km of stream habitat across their native range in the United States, but invasive species, habitat alterations and fragmentation, and climate change have reduced their contemporary distribution by more than one-half (Muhlfeld et al. 2015). Despite extirpation and population declines in many populations from historical abundance levels, none currently have elevated status or protection (Penaluna et al. 2016), and strongholds remain (e.g., Dambacher et al. 2009; Meyer et al. 2014).
The Wood River Basin in central Idaho lies at the eastern edge of the native range of Redband Trout. This basin has been hydrologically isolated from the surrounding Snake River Basin by Malad Gorge Falls, an 18 m waterfall, for at least the last 50,000 years (Lamb et al. 2014). This isolation has resulted in unique fishes occupying lotic habitats in the basin, including the endemic Wood River Sculpin Cottus leiopomus (Simpson and Wallace 1982) and genetically divergent populations of Bridgelip Sucker Catostomus columbianus (Smith 1966) and Mountain Whitefish Prosopium williamsoni (Miller 2006). A recent genetic investigation by Campbell et al. (2022) indicated that Redband Trout in the Wood River Basin may also represent a divergent lineage that has not been previously described. While stocking of fertile hatchery Rainbow Trout (of various origins) in the basin ceased decades ago (Kozfkay et al. 2006), such stocking occurred for nearly a century and, surprisingly, there is limited introgression in the basin between native Redband Trout and nonnative coastal Rainbow Trout O. mykiss irideus (Campbell et al. 2022). Consequently, determining the status of this diverged form of O. mykiss is an important conservation need.
Nonnative Brook Trout Salvelinus fontinalis and Brown Trout Salmo trutta were introduced throughout western North America more than a century ago and have established self-sustaining populations in many river basins in the west (MacCrimmon and Campbell 1969; Dunham et al. 2002; Penaluna et al. 2016), including the Wood River Basin of Idaho. The naturalization of both Brook Trout and Brown Trout outside their native range has often led to the demise of native salmonid populations (reviewed in Dunham et al. 2002). For example, Brook Trout compete with native salmonids for food and space (Kanda et al. 2002) and have repeatedly been found to negatively impact Pacific salmon (Levin et al. 2002), Bull Trout Salvelinus confluentus (Rieman et al. 2006), and multiple species of native inland Cutthroat Trout O. clarkii (Dunham et al. 2002; Benjamin and Baxter 2010; McGrath and Lewis 2011), often exhibiting greater density and biomass over their native counterparts. Studies have also demonstrated that Brook Trout can negatively affect Redband Trout (e.g., Miller et al. 2013), although they are less commonly considered a direct threat to Redband Trout conservation. Similarly, Brown Trout negatively impact numerous species of native salmonids, largely because of competition and predation (reviewed in Rinne and Calamusso 2007 and Budy and Gaeta 2017; also see Hasegawa 2020). For both Brook Trout and Brown Trout, negative impacts often manifest in longitudinal occupancy patterns in river networks where the nonnative species dominate the lower portions of a drainage, displacing native salmonids to upstream habitat (e.g., Shepard 2004; McHugh and Budy 2006; Zeigler et al. 2019). However, reverse patterns have also been observed, such as when Brook Trout reside upstream of O. mykiss populations (Larson and Moore 1985; Weigel and Sorensen 2001).
Prior research conducted in the Wood River Basin suggested that in 2003, Redband Trout occupied 19% of the stream network in the basin (at a 1:100,000 hydrologic scale) and comprised about two-thirds of the total abundance of salmonids at that time, with Brook Trout and Brown Trout comprising the remainder (Meyer et al. 2014). Our primary objective was to compare the contemporary occurrence and density of Redband Trout in the Wood River Basin to the prior research conducted in 2003 by Meyer et al. (2014). Our secondary objective was to assess whether nonnative Brook Trout and Brown Trout experienced changes in occurrence and density in the basin over the same time period.
Study Area
The Wood River Basin, located in central Idaho, has a drainage area of 7,778 km2 and consists of three subbasins: the Big Wood River, the Little Wood River, and Camas Creek. The Malad River forms at the confluence of the Big Wood and Little Wood rivers and flows downstream to the Snake River (Figure 1). Geologic processes of glaciation and episodic volcanic activity likely contributed to the isolation of fish populations in the basin through the formation of the Malad Gorge Falls, eliminating upstream fish passage from the Snake River (Lamb et al. 2014). Two large irrigation storage reservoirs and multiple irrigation diversions on the main stem rivers and tributaries further limit connectivity and modify stream discharge, which is driven by alpine snowmelt and peaks between April and June. Stream elevations are highest at the mountainous headwaters (over 3,000 m above sea level) and lowest (930 m) at the confluence with the Snake River.
Map outlining the species composition at each study reach in 2003 (left) and 2021–2022 (right) in the Wood River Basin (where RBT = Redband Trout Oncorhynchus mykiss gairdneri, BKT = Brook Trout Salvelinus fontinalis, and BNT = Brown Trout Salmo trutta). Study reach numbers correspond with reach numbers in Table 1. Main stem rivers, reservoirs, and main Idaho cities are also outlined.
Map outlining the species composition at each study reach in 2003 (left) and 2021–2022 (right) in the Wood River Basin (where RBT = Redband Trout Oncorhynchus mykiss gairdneri, BKT = Brook Trout Salvelinus fontinalis, and BNT = Brown Trout Salmo trutta). Study reach numbers correspond with reach numbers in Table 1. Main stem rivers, reservoirs, and main Idaho cities are also outlined.
Methods
In 2003, a list of spatially balanced randomly selected study reaches were generated with the help of the Environmental Protection Agency’s Environmental Monitoring and Assessment Program. This technique maps two-dimensional space (in our study, a 1:100,000 scale hydrography layer) into one-dimensional space with defined, ordered spatial addresses and uses restricted randomization to randomly order the spaces. Systematic sampling of the randomly ordered spaces results in a spatially balanced sample of study reaches (Stevens and Olsen 2004). For further details on study reach selection, see Meyer et al. (2014).
In 2003, Meyer et al. (2014) captured Redband Trout at 24 of the 114 study reaches sampled in the Wood River Basin. Most (72%) of the unoccupied reaches were dry during sampling, either naturally or because of irrigation diversions. In 2021–2022, we resampled all reaches where Redband Trout were formerly present, except for two reaches on private property where we could not obtain access (22 study reaches total). All fish and habitat sampling methods were consistent for both sampling periods (2003 and 2021–2022) unless otherwise noted, with electrofishing sampling methods only differing between sampling periods at the two Little Wood River study reaches (Table 1).
Study reach characteristics, fish survey type (where D = multiple-pass depletion and MR = mark-recapture), and trout density (fish ≥ 100 mm total length) for Redband Trout Oncorhynchus mykiss gairdneri (RBT), Brook Trout Salvelinus fontinalis (BKT), and Brown Trout Salmo trutta (BNT) in the Wood River Basin, Idaho surveyed in 2003 and 2021–2022.

Fish sampling occurred at baseflow conditions from July to October to minimize differences in fish capture efficiency and seasonal shifts in habitat use. We captured fish using electrofishing gear, generally using settings of 50–60 Hz, 10–25% duty cycle, and 200–500 volts. We anesthetized all captured trout, identified them to species, measured them for total length to the nearest millimeter, and released them back into the stream (Data S1, Supplemental Material). The few triploid hatchery Rainbow Trout we encountered were stocked at catchable size and thus were readily identifiable based on size and fin condition. We released these fish and did not consider them further. Once fish data collection was complete, we measured reach length (in the thalweg) using a tape measure or a rangefinder accurate to ± 1 m. We measured stream wetted width (m) across multiple transects at each study reach and averaged these measurements to estimate mean wetted width for the reach (Data S2; Supplemental Material). We determined elevation (m) at each study reach from U.S. Geological Survey (USGS) 1:24,000-scale topographic maps using GPS-acquired geospatial coordinates obtained at the lower end of each study reach.
At study reaches less than ∼15 m wide, we formed crews of two to seven people to perform multiple-pass depletions (two to four passes) using at least one but up to three backpack electrofishing units, depending on stream size (Table 1). We used block nets or natural velocity breaks to minimize fish movement into and out of the study reach during depletion surveys. Reach length averaged 103 m in 2003 and 101 m in 2021–2022 across all study reaches (Table 1, Data S3, Supplemental Material). We used the Zippin removal estimator (Zippin 1958) to estimate the abundance of trout ≥ 100 mm total length and did not consider fish < 100 mm because of low capture efficiency. We captured too few fish to estimate trout abundance for additional size classes. If we captured all trout on the first pass, we considered that catch to be the estimated trout abundance. At two reaches, we conducted only one electrofishing pass in 2021–2022, for which we estimated abundance using the linear relationship between the first pass and the resulting abundance estimates from all other multi-pass depletion study reaches (cf. Kruse et al. 1998).
At wider, deeper reaches where multi-pass depletion electrofishing was not feasible, we conducted mark-recapture abundance estimates using a barge-mounted electrofishing unit with crews of seven to 10 people (Table 1). We marked all trout with a caudal fin clip during a single marking run, with marked and unmarked trout captured a few days later during a single recapture run. We assumed that there was no movement of trout into or out of the study reach between runs, and reaches were much longer than depletion reaches to help minimize movement between the mark and recapture runs (mean reach lengths of 879 m in 2003 and 912 m in 2021–2022; Table 1). We estimated the abundance for trout ≥ 100 mm total length using the Lincoln-Petersen mark-recapture model as modified by Chapman (1951) and did not consider fish < 100 mm because of low capture efficiency. We estimated abundance separately for the smallest size classes possible (generally 25–50 mm) while meeting the criteria that (1) the number of fish marked in the marking run multiplied by the catch in the recapture run was at least four times the estimated population size and (2) at least three recaptures occurred per size class; meeting these criteria creates biased estimates by less than 2% (Robson and Regier 1964). We summed estimates for each size class for an estimate of total trout abundance.
We pooled all trout captured for an overall estimate of trout abundance in the study reach (e.g., Isaak and Hubert 2004; Carrier et al. 2009) for both depletion and mark-recapture electrofishing surveys. We estimated abundance for each species based on the proportion of catch comprised by each species (Meyer and High 2011). We converted fish abundance to density (fish/100 m2) by dividing the abundance estimate for each study reach by the product of the mean wetted width and length obtained from that study reach multiplied by 100. We used simple linear regression to assess whether changes in Redband Trout density from 2003 to 2021–2022 were related to changes in nonnative trout density (Brook Trout and Brown Trout combined). We used R version 4.2.2 (R Core Team 2023) for this analysis. We calculated changes in fish density as the difference between the 2021–2022 and 2003 fish densities.
Results
Of the 22 study reaches in the Wood River Basin where Redband Trout were present in 2003, they were still present at 17 (77%) reaches in 2021–2022 (Table 1). From 2003 to 2021–2022, Redband Trout were apparently extirpated from five study reaches, all of which were small streams (< 4.1 m mean wetted width), and Brook Trout are now the only salmonid present (Table 1). In comparison, Brook Trout were present at 20 of the 22 reaches in 2003 (91%) and were still present at 18 reaches (82%) in 2021–2022 and remained absent from the two reaches they did not occupy in 2003. Redband Trout are the only salmonid to occupy the two reaches where Brook Trout were apparently extirpated, both of which were small (3.1 and 5.0 m wide in 2021–2022). Brown Trout were present at two reaches in 2003 and three reaches in 2021–2022, having colonized one new reach. In 2021–2022, Brown Trout occupied all three stream reaches in the study that exceeded 10 m average wetted width (Table 1). From 2003 to 2021–2022, Redband Trout composition (of all trout at the study reach) across all study reaches decreased on average by 13%, whereas Brook Trout and Brown Trout increased on average by 8% and 4%, respectively (Figure 1).
Average total trout density across all reaches increased from 5.42 trout/100 m2 in 2003 to 11.46 trout/100 m2 in 2021–2022 (Table 1) and increased individually for both Redband Trout and nonnative trout at nearly every reach. In 2003, average density was 3.38 fish/100 m2 for Redband Trout, 2.01 fish/100 m2 for Brook Trout, and 0.02 fish/100 m2 for Brown Trout. In 2021–2022, average density increased to 6.15 fish/100 m2 for Redband Trout, 4.67 fish/100 m2 for Brook Trout, and 0.65 fish/100 m2 for Brown Trout. Changes in Redband Trout density were not related to changes in nonnative trout density (F = 0.14; r2 = −0.043; P = 0.709).
Discussion
From a conservation perspective, Brook Trout likely pose a threat to the long-term persistence of Redband Trout in the Wood River Basin, as evidenced by the fact that during the last two decades Brook Trout apparently displaced Redband Trout at one-quarter of the reaches where they were formerly sympatric. However, although fewer in number, there were also some formerly sympatric reaches where Redband Trout persisted and Brook Trout surprisingly did not, indicating that displacement of Redband Trout by Brook Trout is not a foregone conclusion. In western North America, Brook Trout have repeatedly displaced both Cutthroat Trout and Bull Trout (reviewed in Dunham et al. 2002), often resulting in little sympatry between Brook Trout and either of these two native salmonids (e.g., Meyer et al. 2022; Voss et al. 2023). In contrast, our results suggest that Redband Trout can exhibit some level of biotic resistance to Brook Trout invasion, and prior research has demonstrated similar resistance by Redband Trout to Brook Trout displacement (e.g., Benjamin et al. 2007; Miller et al. 2013). In Oregon desert streams, the presence of Brook Trout diminished Redband Trout size structure and abundance compared to stream reaches where Brook Trout were absent, but overall trout density was higher at sympatric sites, suggesting that some resource partitioning was occurring between the two species (Miller et al. 2013). Relative to the vast amount of literature relating Brook Trout interactions with other native salmonids in western North America, assessments of Brook Trout interactions with Redband Trout are surprisingly rare. Further research is clearly needed to more fully understand the environmental conditions where Redband Trout are most vulnerable to displacement by Brook Trout.
Brown Trout may also pose a long-term conservation threat to Redband Trout in the basin (Budy and Gaeta 2017), although this threat is likely limited to main stem, larger river segments. Brown Trout can successfully occupy small, high-elevation Rocky Mountain streams (Young 1999; Al‐Chokhachy and Sepulveda 2019). However, environmental conditions often limit Brown Trout recruitment in high elevation streams (Wood and Budy 2009) since Brown Trout prefer warmer, lower elevation main stem rivers (de la Hoz Franco and Budy 2005). Intermittent stocking of Brown Trout has taken place at three locations in the Little Wood River subbasin from 1970 to 2023 to provide angling opportunities to the public, whereas stocking in the Big Wood River subbasin has only occurred once, in 1971 (IDFG 2024). As with Brook Trout, there is also limited research on interactions between native Redband Trout and nonnative Brown Trout, though Brown Trout have displaced Rainbow Trout in some mountain streams (Gatz et al. 1987). Considering the collective threat likely posed by Brook Trout in headwater streams and Brown Trout in larger rivers, an invasive species management approach―potentially utilizing rotenone treatments (reviewed in Ling 2003; McClay 2005), manual electrofishing suppressions (Shepard et al. 2002), and perhaps even MYY stocking (Schill et al. 2017; Vincent et al. 2022)–may be needed in some areas to control their spread in the Wood River Basin. Of the study reaches with mean width < 5 m, one-half experienced extirpation of either Redband Trout or Brook Trout, whereas none of the study reaches with mean width > 5 m experienced any species extirpations. As such, we recommend initially focusing Redband Trout restoration actions or Brook Trout eradication efforts in the smallest streams in the basin.
Despite the apparent displacement of Redband Trout from several study reaches in the Wood River Basin, Redband Trout density increased substantially from 2003 to 2021–2022, as did the densities of nonnative Brook Trout and Brown Trout. Salmonid populations are notoriously variable in nature (Platts and Nelson 1988; House 1995), with stochastic and demographic population responses driving annual fluctuations (Cattanéo et al. 2003; Copeland and Meyer 2011). Therefore, we cannot rule out the possibility that the 2003 survey took place during a period of lower overall trout abundance, whereas the 2021–2022 surveys occurred during a more favorable period. Alternatively, the increased abundance we observed may be a reflection of improved habitat conditions. Indeed, there have been many habitat improvement projects implemented within the basin, including bank stabilization, removal of fish passage barriers, culvert replacements, fish screening of irrigation ditches, fish entrainment recoveries, floodplain reconnection, and deployment of beaver dam analogs (M.P. Peterson, personal observation). Although many river basins across the western United States and beyond widely implement such habitat improvement projects, these activities, which are generally beneficial (reviewed in Roni et al. 2008), are likely not the only factor currently explaining the substantial increases in trout abundance observed within the Wood River Basin.
It is admittedly difficult to draw firm conclusions on Redband Trout status and nonnative trout expansion in the Wood River Basin based on only two sets of surveys of only a few dozen study reaches across time. As a result of only resurveying reaches where Redband Trout occurred in 2003, our study design only allowed us to detect range contraction (not expansion), although we deem it unlikely that Redband Trout have significantly expanded into previously unoccupied habitat in the basin, considering the rarity of such expansions of western native salmonids in other investigations (e.g., Thurow et al. 1997; Meyer et al. 2006, 2022). Despite study limitations, we are concerned about the potential spread of nonnative salmonids (especially Brook Trout) in the basin. Consequently, we stress the need for continued periodic monitoring of these established reaches and other locations to thoroughly assess the status of this unique lineage of Redband Trout and to better monitor the potential expansion of nonnative salmonids in the basin. While Redband Trout appear to be at least somewhat resistant to Brook Trout displacement, a concerted effort to implement an invasive species management approach in the basin may be needed to control their spread.
Supplemental Material
Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.
Data S1. This datafile contains total length (mm) of all Redband Trout Oncorhynchus mykiss gairdneri (RBT), Brook Trout Salvelinus fontinalis (BKT), and Brown Trout Salmo trutta (BNT) captured at each study reach in the Wood River Basin, Idaho surveyed in 2003 and 2021–2022.
Available: https://doi.org/10.3996/JFWM-23-068.S1 (252 KB XLSX)
Data S2. This datafile contains the reach number, stream name, and location (latitude and longitude), of each study reach; and the stream wetted width (m) measured at each transect by study reach in the Wood River Basin, Idaho surveyed in 2003 and 2021–2022.
Available: https://doi.org/10.3996/JFWM-23-068.S2 (24.2 KB XLSX)
Data S3. This datafile contains the reach number, stream name, location (latitude and longitude), elevation (m), and length (m) of each study reach in the Wood River Basin, Idaho surveyed in 2003 and 2021–2022.
Available: https://doi.org/10.3996/JFWM-23-068.S3 (10.7 KB XLSX)
Acknowledgments
We owe a debt of gratitude to the numerous staff across the state who assisted with field surveys and data management in both 2003 and 2021–2022. In addition, we thank M. Corsi and J. Kozfkay, as well as two anonymous reviewers of this journal for providing valuable edits on this manuscript. Funding for this work was provided by anglers and boaters through their purchase of Idaho fishing licenses, tags, and permits and from federal excise taxes on fishing equipment and boat fuel through the Sport Fish Restoration Program.
Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.
References
Author notes
This Online Early paper will appear in its final typeset version in a future issue of the Journal of Fish and Wildlife Management. The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.