by Jeff L. Arnold, Colleen R. Detjens, Brian D. Ertel, Michael E. Ruhl, & Todd M. Koel
The Madison and Gallatin rivers, two major headwaters of the Missouri River, originate in Yellowstone National Park (YNP) along the western boundary (figure 1). Combined, these two rivers provide 1,031 km (640 mi.) of stream habitat for both westslope cutthroat trout and Arctic grayling in YNP. Indigenous westslope cutthroat trout currently occupy 2 km (1.2 mi.) of stream within their historic range in the park, while resident grayling were extirpated from the park by 1935 (Vincent 1962). Within the Madison River drainage, westslope cutthroat trout and grayling occupied the Madison River and lower portions of the Gibbon River (up to Gibbon Falls) and the Firehole River (up to Firehole Cascades; figure 1). The Gallatin River begins on the northwest side of YNP and flows for approximately 27 km (17 mi.). Westslope cutthroat trout were historically found throughout the Gallatin River and its tributaries, while grayling were confined to the main stem (figure 1).
Status of Westslope Cutthroat Trout
In Montana, genetically pure westslope cutthroat trout occupy less than 3% of their historic range and are confined to isolated headwater streams. Extensive stocking and subsequent establishment of populations of non-native competing species, including brook and brown trout, and interbreeding with rainbow trout led to a serious reduction in the park’s resident westslope cutthroat trout, and their near extinction from most park streams by the 1930s (Varley and Schullery 1998).
In 2005 park fisheries biologists discovered two previously unknown populations of westslope cutthroat trout. That June, a population was discovered in an unnamed tributary of Grayling Creek (Madison River drainage), which was later given the name Last Chance Creek (figure 1). Testing confirmed these fish were 100% genetically pure westslope cutthroat trout. Since their discovery, gametes (reproductive cells) from this population have been used to restore trout populations into other areas of the upper Madison and Gallatin river drainages. Then, in August 2005, a second population was discovered in Geode Creek, a tributary of Yellowstone River located in northern Yellowstone. These fish are not native to this drainage and were most likely stocked in the 1920s. The discovery of this genetically pure population in a location with quick and easy access was exciting because it has enabled biologists to collect and move large numbers of fish or gametes for re-establishing westslope cutthroat trout to restored habitats elsewhere.
Status of Fluvial Arctic Grayling
Grayling are comprised of two distinct strains depending on life history: fluvial (stream dwelling) and adfluvial (living in lakes and spawning in streams). Historically, fluvial grayling were indigenous to the park in the headwaters of the Madison and Gallatin rivers. Grayling within the upper Gallatin River drainage disappeared around 1900, while grayling in the upper Madison River drainage disappeared by 1935 (Vincent 1962). Although grayling had disappeared in the Madison River, anglers continued to report catching them in the Gibbon River (figure 1). Intensive sampling in 2005-2006 found grayling in low numbers (Steed et al. 2011). Genetic analyses provided conclusive evidence that the grayling observed were fish which had strayed downstream from Grebe and Wolf lakes. These grayling descended from an adfluvial population introduced in the early 1900s and are not native to YNP. An additional introduced population also exists in nearby Cascade
Lake in the Yellowstone River drainage (Kaya 2000).
Approach to Native Fish Restoration in Streams and Lakes
To successfully restore native westslope cutthroat trout and fluvial grayling, high priority watersheds were identified; restoration areas were protected via a waterfall or other in-stream barrier; non-native and/or hybridized fish were removed from the area above the barrier using an approved fish toxin (rotenone); and native fish were reintroduced from genetically-unaltered sources.
Prioritizing Watersheds—Watersheds must be large enough to support a fish population and resilient enough to withstand natural disturbances such as fires, droughts, and/or floods. Because the largest threat to native fish is invasion by non-natives, we also initially looked for the presence of barriers that would serve to isolate the restoration area and create a headwater refuge for native species. Information used in this prioritization process included fish species composition, genetic integrity, presence of barriers, road and trail access, and watershed complexity. This information was used to create a prioritization matrix designed to rank each stream based on its potential for successful restoration. The absence of brook trout was a major deciding factor in choosing which streams in the Gallatin and Madison river drainages would be initially restored, as brook trout can be incredibly challenging to extirpate from complex stream systems. Grayling Creek and the upper Gallatin River and all of its tributaries (e.g., Fan and Specimen creeks) do not contain brook trout and, therefore, were considered locations with a high probability of restoration success.
Isolating the Project Area—For native fish restoration, the best option for long-term persistence of species is to create isolated headwater refuges. Some stream systems in Yellowstone have natural waterfalls that can be altered, if necessary, to ensure upstream passage by fish is prevented and a headwater native refuge is secured. Stream systems not having waterfalls naturally allow unimpeded access from fish downstream. In these systems, if a suitable location exists, an artificial barrier can be created to isolate the drainage. Artificial barriers are created with logs, rocks, and mortar, and are designed to completely preclude all upstream fish movement while ensuring structural integrity and function across a wide range of water flows. Structures of this type are commonly at least 1.8 m (6 ft.) in height, with a vertical or near vertical drop onto a concrete splash pad preventing a plunge pool from forming at the base of the barrier. Ideal locations for fish barriers are in high-gradient stream reaches with steep banks and exposed bedrock. Fish barriers would be built in the most downstream location suitable for construction, to provide the largest possible area upstream for native fish restoration.
Pre-Project Assessments—Once a watershed has been selected for a fish restoration project, biologists map the watershed; conduct surveys to document fish, amphibian, and aquatic invertebrate populations; and determine the upper extent of water in each tributary. Knowing fish and amphibian distributions are important because it helps biologists make better informed decisions during the planning phases of the project. For example, if some waters in the extreme upper extent of the watershed are isolated above falls or steep cascades and do not have fish, then those waters do not need to be chemically treated. To achieve a complete removal of non-native fish, all connected surface waters having the potential to support fish must be treated with rotenone.
Non-native Fish Removal—Fish toxins (piscicides) are commonly used to control or eradicate non-native and undesirable fish species in both standing and flowing water. Rotenone, the only piscicide currently available, is toxic to gill breathing organisms, affects fish at the cellular level, and is relatively nontoxic to humans or wildlife. All piscicide applications follow applicable permitting requirements and guidelines set forth by regulatory authorities.
Before applying rotenone to remove fish, we conduct biological assessments on the water to be treated. Bioassays are done on-site to evaluate how environmental factors such as water temperature, sunlight, and pH influence the effects of rotenone on fish. Bioassays ensure the lowest concentration of rotenone is used to achieve a complete fish kill, while minimizing impacts to other aquatic organisms.
Rotenone is applied using drip stations during stream treatments and a boat bailer pump system during lake treatments. Backpack sprayers with hand-held wands are used to apply highly diluted rotenone in stagnant water areas along streams and shorelines of lakes. Drip stations consist of a five-gallon container that dispenses a dilute concentration of rotenone to flowing waters at a constant rate. Placement of drip stations along a stream is determined by conducting travel time studies using nontoxic dye. Because rotenone breaks down and loses its toxicity quickly in flowing water, multiple drip stations are needed to maintain concentrations lethal to target fish. The amount of rotenone used in each drip station is calculated from stream flow measurements taken prior to treatment and from results of bioassays.
For lake applications, we use inflatable rafts with an 80-gallon collapsible tank. The tank is filled with concentrated rotenone which is pumped directly into the motor wash during application. The amount of rotenone applied to a lake is determined by estotal water volume in the lake. Lake water volume, along with results from our bioassays, is used to determine how much rotenone is needed to achieve a complete fish kill.
At the downstream reach of the project area, a neutralization station is set up to detoxify rotenone. The neutralization station consists of a volumetric feeder that applies potassium permanganate at a predetermined rate. Potassium permanganate neutralizes rotenone, eliminating its toxic effects. Potassium permanganate is applied until the last of the rotenone has theoretically flowed pass the neutralization station, as calculated from our travel time study, and stopped after sentinel fish placed above the station remain alive for an additional four hours. To ensure all fish are removed, treatments are typically conducted at least twice within the same year or over successive years. Monitoring by electrofishing is conducted following rotenone treatments to ensure all fish have been removed and the restoration area is ready for reintroduction of native fish.
Reintroduction of Native Fish—Two methods are used for introducing native fish. One method is the introduction of gametes using remote site incubators (RSIs). After fertilization, embryos are reared in hatchery settings to enhance survival during early growth and development. The developed embryos are then transported to the field and placed in RSIs positioned in streams and tributaries. Over a 2-3 week period, the embryos hatch and swim out of the incubator into the stream system. A second method for introducing native fish is to stock them as fry (young fish capable of self-feeding), juveniles, and/or adults directly into project waters. These fish are acquired from existing wild populations or from a hatchery facility.
Both methods have distinct advantages and limitations. The use of RSIs allows fish to hatch in the stream and imprint to its waters, which theoretically results in these fish later returning to spawn as adults. In addition, RSIs make it possible to stock large numbers of genetically diverse fish with relatively little transportation effort. The main limitations of RSIs are their susceptibility to failure from clogging, or disturbance by wildlife or people during the period the embryos are completing final development and hatching.
Stocking live fish is accomplished by transporting them from a hatchery or wild population into the project area using tanks on trucks or suspended below a helicopter. This method requires little or no post-stocking maintenance, can quickly restore recreational fishing, and is not susceptible to disturbance from wildlife or humans. However, stocking live fish is costly in remote locations because of the need for a helicopter to move fish with large amounts of water. Additionally, because fish were not born within or imprinted to project area waters, it is suspected they experience lower survival and reproductive success than their natal counterparts.
Once native fish have been reintroduced into an area, we continue monitoring the population to confirm it has become successfully established. To evaluate fish populations, we conduct electrofishing surveys in streams, and use seines, gillnets, and snorkel surveys to evaluate lake populations. Collecting different age classes of fish helps determine the health of the population and validate natural reproduction, which indicates the population can likely sustain itself.
Westslope Cutthroat Trout and Arctic Grayling Successes To-Date
East Fork Specimen Creek Specimen Creek is a tributary of the Gallatin River located on the northwest corner of YNP. East Fork Specimen Creek is a large watershed that originates in the Gallatin Mountain Range beginning at High Lake at 2,682 m (8,800 ft.) elevation (figure 1). High Lake encompasses 2.9 ha (7.1 ac) and was historically fishless due to a natural waterfall just downstream from the lake outlet. In 1937, however, the National Park Service (NPS) stocked the lake with Yellowstone cutthroat trout, which are not native to the upper Missouri River drainage. Over time, these fish migrated downstream, while rainbow trout from the Gallatin River moved upstream into the watershed. Due to interbreeding among species, the genetic integrity of native westslope cutthroat trout was severely compromised with a hybridized population less than 80% pure. Because these westslope cutthroat trout were not considered a “conservation population” (which requires more than 90% genetic purity), the fish were prime candidates for removal and replacement with genetically pure westslope cutthroat trout.
Because High Lake was isolated by a waterfall, we were able to work on the lake independently of the waters downstream. During August 2006, rotenone was applied to High Lake and its associated waters. Two 14-foot rafts with outboard motors were used to apply the bulk of rotenone, while backpack sprayers and drip stations were used to treat shallow water and inlet streams. Following the first treatment, daily visual surveys of the lake and inlet streams did not detect any live fish; however, to ensure a complete removal of fish, a second treatment was conducted two weeks later.
In July 2007, we surveyed High Lake for any remaining fish. No fish were collected or observed during these efforts, confirming the absence of fish and eliminating the need for an additional piscicide application. As a result, during 2007-2009 westslope cutthroat trout were introduced into High Lake. A total of 5,345 fertilized eggs and 2,963 fish from Geode Creek were moved to High Lake. In 2010, fry were observed within High Lake inlet tributaries, indicating successful reproduction. In 2016 we sampled High Lake by placing a gillnet in the lake for one night. This net yielded 14 fish ranging in size from 174-440 mm (6.8-17.3 in.) This range of sizes again demonstrates successful reproduction and a healthy fish population.
In 2008, a log barrier was completed on lower East Fork Specimen Creek, allowing the restoration project to extend from High Lake downstream near the confluence with North Fork Specimen Creek. During chemical treatment of East Fork Specimen Creek, we divided the drainage into two manageable reaches by placing a portable barrier approximately halfway down the watershed. The upper and lower reaches were each treated twice within a 2-week period down to the barrier. Rotenone was applied using drip stations and backpack sprayers. A third treatment of East Fork Specimen Creek was conducted in 2009. No fish were found during this treatment, and subsequent monitoring provided evidence all fish were successfully removed from the system.
Following two years of treatments and monitoring, the creek was considered free of non-native fish; restocking efforts took place from 2010-2013. During this time, approximately 10,300 eyed-eggs, which are embryos, were placed in RSIs throughout East Fork Specimen Creek. In 2015 and 2016, we conducted several surveys throughout East Fork Specimen Creek. These surveys indicated a natural reproducing population of westslope cutthroat trout with all fish appearing healthy. The long-term goal for this watershed is to integrate East Fork Specimen Creek into a larger westslope cutthroat trout restoration project that includes the North Fork to improve the resilience of this isolated population to natural threats.
Goose Lake Chain-of-Lakes
Goose Lake and two other small, historically fishless lakes lie within the Firehole River drainage, but are not connected to the river by surface waters (figure 1). Their proximity to a service road makes the lakes easily accessible most of the year. Yellow perch, stocked in the lake early in the 20th century, were eradicated from the lake in 1938. The lake was then stocked with non-native rainbow trout which established a self-sustaining population. In September 2011, the Goose Lake chain-of-lakes were treated with rotenone to remove rainbow trout. In 2012, fisheries staff completed surveys within the three lake system. These surveys did not yield any sign of fish, a good indication that complete removal of rainbow trout was achieved. For three consecutive years (2013- 2015), more than 10,300 westslope cutthroat trout fry were stocked into Goose Lake. In 2016 we placed a gillnet in Goose Lake to asses the fish population. This overnight net yielded 12 fish ranging in size from 170- 200 mm (6.7-7.8 in.). We will continue to monitor the success of these stocking efforts over the coming years; plans are to one day use this pure westslope population as a brood source, providing offspring for restoration projects elsewhere within the upper Missouri River system.
Grayling Creek, a tributary of the Madison River, was historically home to westslope cutthroat trout and fluvial grayling (figure 1). Two waterfalls existed on the lower portions of Grayling Creek, but were not complete barriers because fish could move past them during certain times of year. By the 1950s, grayling had disappeared entirely from the watershed due to non-native fish introduction and completion of Hebgen Dam (Kaya 2000), which submersed the stream’s lower reaches where grayling were most abundant. Westslope cutthroat trout fared little better, with most of the population being eliminated or hybridized with rainbow trout by the early 2000s.
In 2007, federal and state fisheries biologists assessed Grayling Creek for a potential westslope cutthroat trout and fluvial grayling restoration project. The plan included modification of the existing waterfall near Highway 191 to create a complete barrier to upstream fish movement. In 2012, the NPS partnered with technical blasters from Gallatin National Forest, a private contractor, and a Montana Conservation Corps crew to create the barrier. The completed modification elevated the barrier to a height of more than 1.8 m (6 ft.) and filled deep pools in order to create a large concrete “splash pad” at the barrier base, making it a complete barrier to upstream fish movement.
The upper Grayling Creek watershed is located in a remote section of the park with limited road and trail access. Access to the upper portions of the watershed involves hiking off trail, through downed timber and thick vegetation. The Grayling Creek restoration area includes 95 km (59 mi.) of connected stream habitat. Actions to remove fish from upper Grayling Creek took place in August 2013 and 2014 with assistance from the USDA Forest Service; U.S. Fish and Wildlife Service; Montana Fish, Wildlife & Parks; and Turner Enterprises, Inc. Because of the remote nature of the watershed, most equipment and supplies had to be flown in via helicopter. More than two dozen fish biologists and technicians worked for several weeks to remove non-native and hybrid trout from the restoration area using rotenone. Electrofishing surveys conducted after the rotenone treatments in 2014 did not yield any fish, indicating fish removal was a success.
In 2015, we began introducing westslope cutthroat trout and fluvial grayling into the Grayling Creek project area. In April 2015, approximately 680 westslope cutthroat trout of varying sizes from Geode Creek were captured and moved to the lower portions of Grayling Creek above the barrier. In May 2015, more than 100,000 fluvial grayling eggs were placed in RSIs throughout the South Fork Grayling Creek watershed. In addition 4,800 westslope cutthroat trout eggs were placed in RSIs along the lower portion of Grayling Creek while 5,000 westslope cutthroat trout eggs were placed in the North Fork of Grayling Creek.
Restoration of westslope cutthroat trout and fluvial grayling continued in 2016. During May, 263 westslope cutthroat trout were captured and moved from Geode Creek to lower portions of Grayling Creek, and 50,000 fluvial grayling eggs were placed in RSIs throughout the South Fork Graying Creek. During June, approximately 27,800 westslope cutthroat trout were placed throughout the Grayling Creek drainage above the barrier. Restoration efforts are scheduled to continue in 2017-2018.
Potential Future Westslope Cutthroat Trout and Arctic Grayling Restoration Projects
Upper Gibbon River
The upper Gibbon River (above Virginia Cascades) and connecting lakes in YNP will be used as a refuge for native fish threatened with a warming climate. The refuge would include 16 km (10 mi.) of stream, three fish bearing lakes totaling 92 ha (228 surface ac), and extensive tributary networks, representing the largest and most logistically feasible location for westslope cutthroat trout and fluvial grayling restoration in the species’ historic range (figure 1). High-elevation aquatic systems, such as the upper Gibbon River, may be the only chance to protect sensitive, cold water species such as westslope cutthroat trout and fluvial grayling against climate change. The project will begin with the removal of fish from Ice Lake. The project will continue with the complete removal of fish from the Gibbon River above Virginia Cascades upstream to Grebe Lake. Westslope cutthroat trout and fluvial grayling will be introduced immediately afterwards. This project is expected to take three years to complete.
Cougar Creek is a small stream that, prior to reaching the Madison River, flows underground (figure 1). Cougar Creek is an ideal candidate for westslope cutthroat trout restoration because it is physically isolated from downstream reaches of the drainage (Kaya 2000). In the early 1990s, this stream was documented as having pure westslope cutthroat trout and mottled sculpin, two species native to the drainage. However, genetic testing of the westslope cutthroat trout in 2011 indicated these fish were highly hybridized with rainbow trout. Future plans for the Cougar Creek drainage are to chemically remove the hybridized trout and reintroduce pure westslope cutthroat trout and mottled sculpin.
North Fork Specimen Creek
Fisheries staff in YNP plan to continue native fisheries restoration work in Specimen Creek. In 2013, plans were developed to construct a concrete barrier on the lower portion of the creek near Highway 191. This barrier would isolate the entire drainage and prevent non-native fish from moving upstream into Specimen Creek from the Gallatin River. The barrier would allow for native fish restoration of the entire watershed including the North Fork Specimen Creek, and would allow for the East Fork Specimen Creek to be reconnected to the rest of the drainage.
Creating secure habitats for the conservation of native inland fish has become a common fisheries management practice in recent years. Most commonly, the goals of these efforts are to restore and preserve native fish biodiversity through the exclusion and removal of non-native fish and the reintroduction of genetically unaltered native species. The need for projects of this nature is largely driven by competition, predation, and/or hybridization by non-native fish; however, other factors including habitat alteration, disease, and climate change may also contribute to the need for action. A common project model that has evolved and become widely used follows three basic steps: 1) ensure isolation of the project area, 2) completely remove all non-native species, and 3) re-introduce genetically unaltered native species. The body of scientific information that has accrued around this model, conventional wisdom, and on-the-ground experience is used to carry out the project.
In the park’s early history, non-native trout were readily stocked into park waters to provide additional fishing opportunities for visitors. Over the last century, these non-native trout have eroded native trout populations to a small fraction of their historic range. Once occupying hundreds of stream kilometers, indigenous westslope cutthroat trout now occupy only 2 km (1.2 mi.) of streams in the Grayling Creek drainage, while fluvial grayling disappeared entirely from the park by 1934. As park biologists, we have worked diligently to restore these native fish back into their historic ranges using approaches described above. Over the past decade, genetically pure westslope cutthroat trout have been reintroduced into the headwaters of Specimen and Grayling creeks, with a local brood source being developed in the Goose Lake complex. Our restoration efforts have added 74 km (46 mi.) of stream that are now occupied by native westslope cutthroat trout and/or fluvial grayling. The grayling introduced to upper Grayling Creek in 2015 were the first fluvial grayling to swim in park waters in more than 80 years. Future projects for westslope cutthroat trout and fluvial grayling include native fish restoration in North Fork Specimen and Cougar creeks as well as the upper Gibbon River. Once these projects are completed, an additional 61 km (38 mi.) of stream will be restored to native fish. As the NPS enters its next century, we continue to work to preserve and protect native fishes of Yellowstone for future generations.
Kaya, C.M. 2000. Arctic grayling in Yellowstone: status, management, and recent restoration efforts. Yellowstone Science 8:12-17.
Steed, A.C., A.V. Zale, T.M. Koel, and S.T. Kalinowski. 2011. Population viability of Arctic grayling in the Gibbon River, Yellowstone National Park. North American Journal of Fisheries Management 30:1582-1590.
Varley, J.D., and P. Schullery. 1998. Yellowstone fishes: ecology, history, and angling in the park. Stackpole Books, Mechanicsburg, Pennsylvania, USA. Vincent, R.E. 1962. Biogeographical and ecological factors contributing to the decline of Arctic grayling, Thymallus arcticus (Pallas), in Michigan and Montana. Dissertation. University of Michigan, Ann Arbor, Michigan, USA.
Jeff Arnold is a Fisheries Biologist with Yellowstone’s Native Fish Conservation Program where he began working in 2002. He leads the various aquatic monitoring programs in the park, which include sampling water quality, amphibians, aquatic invertebrates, and fish. He serves as the native fish restoration biologist for westslope cutthroat trout and Arctic grayling. He received a MS in 1990 from Western Illinois University, Macomb, Illinois, studying aquatic invertebrate communities in the Mississippi River. After graduating, Jeff worked with state agencies in both Illinois and Florida. In Illinois, he was employed by the Illinois Natural History Survey where he worked on a variety of projects studying large river ecology, including aquatic invertebrates, fish, and plant communities. In Florida, Jeff was employed by Florida Fish and Wildlife Conservation Commission where he was involved in a fisheries management program which enhanced local fishing opportunities for urban residents in Orlando, Florida.