Rehabilitation of the Middle Fork Willamette Bull Trout Population

REHABILITATION OF THE MIDDLE FORK WILLAMETTE BULL TROUT POPULATION

RISK ANALYSIS AND MONITORING PLAN

Oregon Department of Fish and Wildlife USDA Forest Service/Rigdon Ranger District

1998

INTRODUCTION

Bull trout populations have declined throughout their range in the western United States and Canada. Some of the factors that caused the decline were overharvest, water quality changes because of intensive land management practices, and removal of fish by chemical application. On June 11, 1997, USFWS proposed to list the Columbia River bull trout population segment (including the Willamette populations as Threatened. A public comment period ended August 12, 1997, and currently (November 1997) USFWS is determining how to proceed. The status of populations in the Willamette Basin is varied (Buchanan et al., 1997). Bull trout have been extirpated from Santiam systems and most likely the Clackamas. McKenzie bull trout have been fragmented into several populations by dams and their status varies from “Of Special Concern” to “High Risk.” The population in the Middle Fork Willamette basin is at an extremely low level with a low probability of persistence.

Several specific factors are identified in the decline of bull trout in the Middle Fork Willamette. Rotenone poisoning in 1960 to remove undesirable fish above Hills Creek Dam killed bull trout. Subsequent to the Rotenone treatment, the river was heavily stocked with rainbow trout. A large number of anglers fished this area of the river and overharvest of the remaining bull trout undoubtedly occurred. In addition, timber harvest and road building practices damaged bull trout spawning and rearing habitat or precluded access to suitable habitat. Construction of Dexter, Lookout Point, and Hills Creek dams modified stream temperatures and restricted migrations to and from spawning areas. Table 1 displays the last reliable observations of bull trout in the Middle Fork Willamette basin.

Table 1. Bull trout observations in the Middle Fork Willamette basin.

WATERBODY LAST YEAR OBSERVED Middle Fork Willamette below Hills 1953 Creek Reservoir Middle Fork Willamette above Hills 1990 Creek Reservoir North Fork of Middle Fork 1962 Willamette Hills Creek Reservoir 1988 Salt Creek 1960

Bull trout habitat remains in spring-fed portions of the Middle Fork Willamette and tributaries. However, repeated electrofishing and snorkel surveys of the Middle Fork Willamette and tributaries have not detected bull trout. ODFW has electrofished margin areas of the Middle Fork Willamette and major portions of all tributaries from Chuckle Springs to Swift Creek, the areas with the most likely bull trout habitat on the Middle Fork Willamette. Most of the pools from Paddy’s Valley to Staley Creek, as well as much of Swift Creek, have been snorkeled by ODFW and/or USFS personnel.

2 Potential bull trout habitat remains in spring-fed portions of the MFW and tributaries above Hills Creek Reservoir (Rigdon Ranger District 1996). However, despite occasional angler reports, repeated electrofishing and snorkel surveys of the Middle Fork Willamette and tributaries have not detected bull trout presence. Therefore, the Upper Willamette Bull Trout Working Group believes that Middle Fork Willamette bull trout will not continue to persist in the subbasin without rehabilitation efforts.

As you may recall at the SCCS meeting at McKenzie Bridge, Paul Spruell said the only way of holding on to the remaining genetic material was to provide a carrier for that material. Under the definition of Supplementation in the OAR's, 635-07-501 (53) we believe this is what we are doing. If we are not using this definition of supplementation, or if there are some non-policy rules that we don't have, please let us know (We're unsure if supplementation means an ongoing program. If that is the case, perhaps we are not supplementing because our plan is to go in for one life cycle and let the chips fall). We do have another alternative for you to consider. Under the Operating Principles for Wild Fish Management, OAR 635-07-527 (3) outlines special rehabilitation programs. We have always been very leery of calling this population extinct. If it was easy to call them extinct, we would probably have had this program going three years ago. If there is a chance that they are not extinct, our program fits into this category. This project is designed to comply with Wild Fish Management Policy (WFMP) Oregon Administrative Rules 635-07-527 (3), addressing rehabilitation of wild fish populations.

PROJECT BACKGROUND

In 1997, ODFW and USFS/Rigdon Ranger District began a cooperative effort to reintroduce bull trout to the Middle Fork Willamette basin. Release sites were chosen principally on criteria and information provided in Goetz (1994).

In order to develop and evaluate rehabilitation procedures, ODFW transferred 178 bull trout fry (age 0+) from Anderson Creek on the McKenzie River, to three tributaries of the Middle Fork Willamette in June 1997. The release sites included Chuckle Springs, Indigo Springs and Skunk Creek. Of the three pilot program sites, only Chuckle Springs will be utilized in future restocking efforts. After further analysis, Skunk Creek was determined to be too warm to provide optimal rearing habitat for juvenile bull trout, and spawning and access were limited in Indigo Springs. Therefore, we chose new rehabilitation sites for 1998.

Rehabilitation Sites The McKenzie basin contains the only verifiable populations of bull trout in western Oregon. These fish may have ranged occasionally into the mainstem Willamette and spawned with bull trout from the Middle Fork Willamette. However, warm water temperatures in the mainstem Willamette and the bull trout’s strong homing instinct probably precluded common genetic exchange between the populations.

3 Three streams have been selected as sites for rehabilitation of bull trout fry: Chuckle, Shadow, and Iko springs. These spring-fed streams emerge from rock on the north side of the Middle Fork Willamette between river miles 258 and 260.5 (Appendix A). The sites were chosen following surveys of several potential rehabilitation streams (see Appendix B). Once chosen, the food base and physical attributes of the rehabilitation sites were assessed (Appendix C). Cutthroat trout inhabit the lower reaches of each stream; there have been no brook trout sightings in the vicinity. Chuckle and Shadow springs contain minimal spawning habitat but are adjacent to good spawning habitat in the Middle Fork Willamette. For 1998, we propose to place 20% of the bull trout fry in Chuckle and 20% in Shadow springs. The remaining 60% will be placed in Iko Springs. Iko has the best combination of potential bull trout rearing and spawning habitat observed in the Middle Fork Willamette watershed. Spawning habitat could be improved dramatically at all three sites with gravel supplementation. Percentages may be adjusted as deemed appropriate in the future. We assume that fry released over the four-year period will disperse downstream and that adults will return to spawn in any suitable habitat.

ACTIONS

We plan to capture bull trout fry from Anderson Creek in the downstream migrant trap located below the OR Highway 126 culvert. We plan to remove 25% or 2,000 (whichever is less) of the potentially- trapped age 0+ bull trout fry (where potential assumes sampling 100% of the time, but does not account for trap efficiency) during the peak migration period in March and April. During the trapping week, the trap will be checked daily and captured fry will be held in mesh-lined plastic containers in the trap live box (Goetz 1989). Twice a week, we will combine all fry into one of the plastic containers, which will be suspended in a liberation tank. Ice will be added to the tank as needed along the route to the Middle Fork Willamette in order to maintain water temperatures at 4-6oC. The fry will be transported as close as possible to the release sites in a truck and carried the remaining distance in covered buckets. Because of the long-distance transport required to move fry to the Middle Fork Willamette, we propose transporting fry twice a week, every other week, through the months of March and April, with a goal of ten transfers per year from 1998-2001.

RISK ANALYSIS

Risks to the McKenzie River Bull Trout Population This project involves the risk of decreasing the size of the McKenzie River bull trout population by removing fry from Anderson Creek. The WFMP limits the number of fish that can be removed to 25% or less of the breeding population. Assuming the trap will operate an average of 50% of each week, removing all of the fry trapped in one week and not removing any fry the following week should result in removal of approximately 25% of the potentially trapped Anderson Creek fry over the migration period. Because some of the fry do not migrate during the trapping period and Anderson Creek trap efficiency

4 is estimated at 60-80 % (Pers. comm. Mark Wade, ODFW), we will actually remove far less than 25% of the migrating fry. Fry will be removed from Anderson Creek based upon fry migration totals for a single week. Therefore, we will need to monitor the number of fry removed from Anderson Creek to ensure that the number of fry removed never exceeds 25% of the potential fry totals at any point in the season. We believe that many of these migrating fry would not find adequate habitat in the lower reaches of Anderson Creek and would eventually enter the McKenzie River. We believe that fry migrating at this age rather than at age 1 or older have a lower survival rate because of a lack of fry habitat in the McKenzie and predation from other fish. We expect that a decrease in the density dependent mortality will somewhat compensate for removal of a portion of the fry population. Thus, we believe that we will capture less than 25% of the future breeding population.

During 1994-1996, the number of fry potentially trapped in Anderson Creek from mid-February to May 31 hovered at approximately 3,500 per year. However, the 1997 total exceeded 13,000. The reasons for the large increase are not entirely clear, but an increased number of spawning bull trout combined with favorable incubation conditions are probably the biggest factors. The number of bull trout redds increased from 30 redds in fall 1994 to 77 redds in fall 1995, but this redd increase did not result in a corresponding increase in the number of fry in spring 1996. We hypothesize that the flood event of February 1996 scoured redds, significantly reducing fry survival. Subsequent flood surveys noted many areas of scour on Anderson Creek. Without the flood, fry numbers in 1996 may have been much closer to those in 1997. In the absence of a major high water event, we believe the 1998 fry numbers should approximate those of 1997 because the 1997 redd count slightly exceeded the 1996 count. The decreased substrate embeddedness noted in Anderson Creek after the high water events of 1996 may have resulted in more favorable spawning conditions. Because of the uncertainty of predicting future trends with these data, we will take a conservative approach and remove no more than 2000 fish or 25% of the fry potentially trapped in 1998. Significant changes in the Anderson Creek bull trout population or any additional plans to transfer bull trout fry from Anderson Creek to the South Fork McKenzie or Sweetwater Creek will require re-evaluation of this number, while always maintaining the transfer limit of 25% of the fry potentially trapped per year.

Removing adult bull trout from the McKenzie could have a major effect on the small McKenzie population because bull trout are repeat spawners and one female may be responsible for tens of thousands of eggs over several years. In 1989-1990, the California Department of Fish and Game attempted to reintroduce bull trout to the McCloud River by trapping approximately 60 small adults from the Klamath River watershed for brood stock. Mortality rates were high in the adults, resulting in a release of only 270 fingerlings available for rehabilitation (Buchanan and Howell 1992). Another source of bull trout for rehabilitation would be eggs. However, the only source of bull trout eggs would be McKenzie adults, which would subject valuable breeding fish to unacceptable risk. In addition, using eggs would result in significantly less genetic diversity than removing only fry from the population.

Risks to the Middle Fork Willamette Bull Trout Population By introducing genetically different fish into the system, we run the risk of compromising the genetic

5 integrity of any remnant Middle Fork Willamette bull trout population. However, any bull trout remaining in the Middle Fork Willamette system exist in very limited numbers and are probably closely related to McKenzie River fish due to possible genetic exchange before dam construction.

Risks to Other Fish Species in the Middle Fork Willamette System Dispersal of the introduced bull trout throughout the Middle Fork Willamette and areas such as Hills Creek Reservoir is essential to ensure the food base required by adult bull trout. Bull trout eat terrestrial and aquatic insects (Shepard and others 1984), macrozooplankton, myriads, and fish (Rieman and Lukens 1979). Fish are common in the diet of individuals 110 millimeters long or longer (Shepard and others 1984). Large bull trout may feed almost exclusively on fish (Fraley and Shepard 1989, Rieman and Lukens 1979, Shepard and others 1984).

Wild cutthroat trout, mountain whitefish (Prosopium williamsoni), and sculpins (Cottus spp.) are present in the rehabilitation area, while stocked and wild rainbow trout are present in the Middle Fork Willamette and tributaries farther downstream. Excess adult chinook salmon (Oncorhynchus tshawytscha) from Willamette Hatchery are released each year into the Middle Fork Willamette above Hills Creek Reservoir. Bull trout were historically sympatric with all of these species; therefore, we anticipate no major risk to any of these fish populations by reintroducing bull trout to the system.

Table 2. Spring chinook salmon released in the Middle Fork River above Hills Creek Reservoir.

Adults Juveniles Jacks 1992 49,950 1993 764 32 1994 176 1 1995 515 50,142 7 1996 341 50,160 19 1997 956 0 Source: Willamette Fish Hatchery (ODFW)

Other Risks to the Project As stated earlier, several factors led to the demise of bull trout in the Middle Fork Willamette, including overharvest, lack of fish passage, and lack of habitat due to poor land management practices. These factors are risks to the project if they have not been adequately addressed.

Overharvest

Angling regulations, changed in 1997, allow use of only artificial flies and lures and take of only hatchery trout in the Middle Fork Willamette above Hills Creek Reservoir. In addition, hatchery rainbow trout

6 are now stocked only from the reservoir upstream to the Rd. 2134 bridge, approximately seven miles below the lowest bull trout rehabilitation site. Finally, the number of fish stocked has been reduced from a high of 33,000 in the early 1980’s to a planned 17,000 in 1998. These regulations reduce angling pressure and the risk of mortality if a bull trout is accidentally hooked.

Lack of Fish Passage

The US Army Corps of Engineers (USACE) is continuing to evaluate anadromous fish passage around Corps dams (USACE 1996). As noted above, ODFW has provided passage for adult chinook salmon by moving hatchery salmon above the dams. The USFS has removed or modified or is planning to remove or modify culverts that block fish passage.

Lack of Habitat

The largest risk to the success of this project is that there may not be enough high-quality habitat to sustain a bull trout population in the Middle Fork Willamette. Various projects have been completed or are underway to improve instream habitat, fish passage, and overall watershed health (Appendix D). Spawning habitat is particularly limited in the few streams cold enough to support bull trout, including the Middle Fork Willamette itself. However, the conditions in the Middle Fork Willamette system appear to be similar to conditions on the upper McKenzie River above Trail Bridge Reservoir. Although spawning habitat is also limited on the upper McKenzie, bull trout have survived there.

Trail Bridge Dam and Oregon Highway 126 were constructed in 1963, isolating a population of bull trout above the dam and blocking passage to a prime spawning area in Sweetwater Creek. Until passage was re-established on Sweetwater in 1992, only the mainstem waters of the upper McKenzie were available for Trail Bridge bull trout to spawn. As of 1997, Sweetwater has not yet been utilized by spawning bull trout. Therefore, for 34 years, bull trout have spawned in a one-mile section of river, using small pockets of gravel behind boulders. There is little or no large woody debris associated with redds, and the flow fluctuates more than any other known bull trout spawning area in the McKenzie watershed, including Anderson Creek, Olallie Creek, and Roaring River. In the last four years, a maximum of eight bull trout redds have been observed on the upper McKenzie. In addition, an unknown number of bull trout migrating downstream through the Dam are lost to the spawning population. Other impacts on the Trail Bridge population include heavy angling pressure from a put- and-take rainbow trout fishery and a competing brook trout population in the reservoir.

In spite of the previously mentioned disadvantages for bull trout above Trail Bridge Dam, the population persists. This evidence provides us confidence that bull trout can survive in marginal habitat despite some serious negative pressures. Because bull trout habitat in the Middle Fork Willamette has been determined to be at least as good as the habitat above Trail Bridge Dam, we believe it will support a population of bull trout.

7 Limited Number of Bull Trout Available for Rehabilitation

If present trends in Anderson Creek bull trout recruitment continue, we should be able to release at least 10,000 bull trout fry in the Middle Fork Willamette watershed over the five-year life of this project. We will attempt to maintain a fry density in the release sites comparable to Sweetwater and Anderson creeks. However, because we will not know the rate of survival to adults of these fry, we will not know how many bull trout fry it will take establish a sustainable population in the Middle Fork Willamette. If we assume a one percent survival from fry to adult, 20 adults will be available to spawn in five or six years. In 10 years, up to 100 spawning adults would be available from the releases, with the potential of additional fish from spawning in the Middle Fork Willamette.

Hybridization

An additional, less significant, risk to the success of the project is bull trout hybridization with the brook trout currently inhabiting lakes in the Swift Creek/Bear Creek drainage, Beaver Creek, and the Middle Fork Willamette above Beaver Creek. Brook trout are not known to inhabit the rehabilitation area. Currently, we believe brook trout dispersal is limited by sections of excessively cold water and other barriers. Similarly, transplanted bull trout may be blocked from upstream movement to brook trout habitat. However, if brook trout and bull trout have overlapping spawning habitat, hybridization could occur, compromising the spawning potential of the bull trout population.

Monitoring Plan

To effectively monitor the impact of our actions on the McKenzie River and Middle Fork Willamette bull trout populations, we plan to continue to collect data on the following:

McKenzie River

Monitoring Activity Method Responsible Agency

Measure adult Bull trout Standard pool counts, Anderson ODFW, USFS/McKenzie RD abundance and Olallie creek redd surveys, angler surveys

Measure juvenile Bull trout Anderson Creek downstream ODFW, USFS/McKenzie RD abundance migrant trap

Middle Fork Willamette River

Monitoring Activity Method Responsible Agency

8 Measure juvenile Bull trout Snorkel surveys of release sites USFS/Rigdon RD abundance and habitat use and adjacent pools of Middle Fork Willamette

Measure juvenile Bull trout Screw trap in Middle Fork ODFW abundance Willamette just below Found Creek

Monitor water quality changes USFS/Rigdon RD

Monitor adult Bull trout Angler surveys ODFW population (future)

9 APPENDIX A: MAP OF REHABILITATION SITES

10 APPENDIX B: SURVEY OF POTENTIAL REHABILITATION SITES

Considering prior experiences and the Upper Middle Fork of the Willamette Watershed Analysis (Willamette National Forest 1996), we identified streams in the Middle Fork Willamette watershed that had potential for bull trout use. These potential streams had maximum temperatures of <13oC and were downstream of any known barriers to fish passage. To evaluate these potential sites, in the summer of 1997, we walked and/or snorkeled all or sections of Swift, Bear, Echo, Beaver, and Tumblebug creeks, as well as the Middle Fork Willamette from Beaver Creek to Swift Creek. These surveys included evaluations of substrate, canopy cover, fish migration barriers, water temperature, available gravel for spawning adults, rearing habitat for juveniles, and coarse and fine woody debris. Except for temperature, no measurements were taken. Surveyors independently rated each stream according to the above criteria, and then the ratings were combined to produce a single numerical rating for each stream. Slide photographs and maps documenting each survey are available at ODFW, Springfield.

Swift Creek Swift Creek was surveyed on June 11-12, with a return trip on August 6. We snorkeled/walked Swift Creek from approximately 50 m above the confluence with Bear Creek to the mouth. At the time of the survey, Swift Creek was fairly high with snowmelt (inflow into Hills Creek Reservoir was 769 cfs on 6/11 and 854 cfs on 6/12) but was generally clear throughout the survey. A rainstorm added some sediment to the stream near Minnehaha Creek. The weather was cloudy and showery throughout the survey. Cutthroat trout were the only fish species observed.

There was a probable barrier to fish passage at a logjam just upstream from the confluence with Bear Creek. Below the barrier, Swift Creek ran through alternating sections of young forest (20 to 30 years old) and mature forest. There were piles of rock and large woody debris throughout the canyon, with high water marks 5 m. or more above the current stream level. Most woody debris was piled above the current stream level. Throughout the survey, we saw evidence of debris torrents and extensive slope failures in tributary canyons. In the first section surveyed, the stream bottom and adjacent riparian areas consisted mainly of large, loose boulders. As we continued downstream of Baboon Creek, the stream gradient decreased. However, most of the channel substrate remained large cobbles, with occasional pockets of gravel behind large boulders. Very few pools were >1 m deep. Just above and below Rd. 21, there were several deeper pools and extensive woody debris in the stream.

The only potential release sites we noted were three side channels, two above Rd. 21 and one below. One of the upper side channels was located between the mouths of Chako and Coulee creeks, and the other was located approximately halfway between Minnehaha and Baboon creeks. Near these channels, spawning habitat was limited to small pockets of gravel behind large boulders. We returned in August and found that the side channels had insufficient flow to be considered as bull trout release sites. The third side channel, below Rd. 21, was longer and showed greater potential as bull trout rearing habitat.

11 Overall assessment: While stream temperatures are suitable for bull trout, Swift Creek appears to have poor bull trout habitat due to a lack of spawning and rearing habitat and the flashy hydrology of the watershed.

Bear Creek We surveyed sections of two branches of Bear Creek and spot-checked additional reaches. A priori, these sections showed promise as potential spawning/rearing areas due to their slope and positions in the upper watershed.

The first survey included the stream from Rd. 2149 upstream to just below Happy Lake. The culvert under Rd. 2149 appeared passable. Bear Creek descended in a stair-like fashion from Happy Lake through uncut, mature forest. The stream consisted of small pools and riffles through bedrock and cobbles, with many downed logs spanning the stream channel. Canopy coverage was nearly 100% for the length of the surveyed reach. However, there was very little gravel for spawning and very little woody debris in the stream. Also, temperatures were elevated, apparently due to flow from Happy Lake: Water temperatures were 17o C at both ends of the survey reach.

The second survey covered a portion of the Bear Creek branch flowing from Blue Lake. Above Rd. 2149, this branch of Bear Creek emerged from uncut, mature forest. The stream temperature was 11oC where it emerged from the culvert, which was likely passable by even small fish. We surveyed from the Rd. 2149 culvert to the beginning of a steep descent, which likely marks a barrier to fish passage. The surveyed reach was completely clear-cut through the riparian zone within the last 20 years, with current canopy coverage less than 10%. The stream bottom consisted entirely of large cobbles and bedrock, with no large woody debris. The stream temperature increased to 16oC at the base of the clear-cut, approximately 250 m downstream of the road crossing. One, approximately 20- cm cutthroat trout was observed at the base of the culvert.

The northernmost branch of Bear Creek merges with the Happy Lake branch just below Rd. 2149. This branch flows through a ~20-year-old clear-cut in which Sitka alders have become densely established in the riparian zone, resulting in nearly 100% canopy coverage. Although the stream was only 8oC at the (likely passable) culvert, it was very small, with large cobble substrate and little or no instream wood.

We spot-checked the area below the confluence of the upper two branches of Bear Creek. The stream temperature was 11o C where the stream emerged from uncut forest into a clear-cut. Even in the mature forest, the canyon was open and bright, with canopy coverage ~50%. Most substrate consisted of large cobbles with occasional gravel pockets. We observed very little fine or coarse woody debris. This section was electrofished in 1996 along with approximately 100 m of stream immediately above the confluence with Swift Creek.

Overall assessment: Bear Creek shows some promise as a possible bull trout spawning area in the reach between the junction of the main branches and the confluence with Swift Creek. However, the

12 Swift/Bear watershed is composed of highly unstable, large substrate, with high potential for scour. Little rearing habitat exists in the watershed, and most bull trout fry would likely be washed very quickly to the Middle Fork Willamette, where their survival would be questionable.

Echo Creek We began our survey of Echo Creek in a clear-cut, approximately 250 m upstream of the largest (unnamed) tributary on the east side of the stream at a known barrier to fish passage [approximate 3.7 river miles (RM), 6 river kilometers (RK)]. The barrier was an approximately 2-m falls. As we moved downstream, we encountered another probable barrier at the mouth of the unnamed tributary. Then we passed through the only section of Echo Creek that flowed through unharvested forest, a 500-m section immediately downstream of the unnamed tributary. This section contained many small pockets of potential spawning gravel as well as one lengthy side channel with some rearing habitat. Large woody debris was common in the stream channel, though there were few pools. For the remainder of the distance to the Middle Fork Willamette, Echo Creek flowed through recent (<30 years) clear-cuts with primarily cobble substrate and minimal alder canopy. Both the road culvert and a falls ~100 m downstream from the culvert appeared to be barriers to passage. High flows will likely eliminate the lower barrier at some point, but the 1.5-m falls could grow higher if debris accumulates at the site. The section below the road culvert contained more, deeper pools, with occasional large woody debris. There was one short side channel below the culvert, with a temperature of 12oC (on 8/7). The main channel of Echo was 11o C at the mouth, when the Middle Fork Willamette was 9oC on 8/7.

Overall assessment: While stream temperatures appear suitable for bull trout, only a short section of Echo Creek contains suitable spawning and rearing habitat. In addition, this minimal habitat is upstream of at least two barriers to fish passage.

Beaver Creek We spot-checked Beaver Creek from just above Rd. 21 to the mouth. Above the road, the creek flows through a brushy clear-cut with little canopy cover. The stream had little flow and was 15°C at the start of the survey. Considering the warm temperature, the lack of adequate flow, and a probable passage barrier just upstream from the Middle Fork Willamette, we ruled out Beaver Creek as a potential rehabilitation site.

Tumblebug Creek We surveyed Tumblebug Creek from just above Tumblebug Gorge to the confluence with the Middle Fork Willamette. We reached the stream via Rd. 2144, approximately 250 m below the mouth of West Branch Tumblebug Creek. In Tumblebug Gorge, the stream cut through a crack in a lava flow, resulting in a canyon with vertical rock walls of > 100 m. The walls of the gorge contained many cracks and caves, with occasional small springs. These springs did not modify the temperature of the stream, as Tumblebug temperatures remained at 13oC throughout the survey. Large rocks and logs that fell from the gorge walls and rim formed many barriers to fish passage, the most dramatic of which was a 6- 7 m falls at the mouth of the first tributary encountered in the survey. Substrate consisted of mostly

13 boulders and large cobbles, with numerous pockets of gravel. Trees have become established on occasional islands and rock outcrops. Elk tracks were observed in the gorge.

A debris dam formed at the mouth of the first major tributary below Tumblebug Gorge during a high water event in 1996. Logs, rocks, and soil were piled over 10 m high, damming the stream for at least 200 m upstream. With the higher water table, numerous trees were dying along the impounded section of stream. These trees will likely fall and reinforce the dam in the future.

Below the debris dam, the stream flowed through mature forest until near the confluence with the Middle Fork Willamette, where there were two clear-cuts above the riparian area on the west side of the stream. Substrate primarily consisted of cobbles, with occasional gravel. We observed no side channels or main channel habitat suitable for juvenile bull trout rearing. The lowest tributary to Tumblebug was a spring-fed stream of 11oC. However, the stream appeared to be impassable for fish within 100 m of Tumblebug. Tumblebug was 13oC at the mouth, where the Middle Fork Willamette was 10oC.

Overall assessment: Tumblebug Creek is an interesting, isolated, largely pristine (where surveyed) stream with little potential for bull trout rehabilitation. The stream is probably warmed in clear-cuts upstream of the survey area. The reaches of Tumblebug that are passable to migratory fish are marginal in terms of temperature, have little spawning habitat, and little or no juvenile rearing habitat. Tumblebug should be monitored over time to see how it responds to improved land management practices.

Middle Fork Willamette River (Beaver Creek to Swift Creek) We surveyed the Middle Fork Willamette from Beaver Creek downstream to Swift Creek to evaluate barriers to fish passage upstream of Chuckle Springs, and qualitatively evaluate the Middle Fork Willamette and short, spring-fed tributaries for potential bull trout spawning and rearing habitat. Just below the mouth of Beaver Creek, a debris dam and falls that was previously considered a barrier had washed out on one side of the channel, opening passage for fish. Downstream, two or more 2-m falls exist on the Middle Fork Willamette where it passes through a gorge upstream of Tumblebug Creek. These probable barriers to fish passage are located near the middle and lower end of the gorge, approximately 700-1000 m above the mouth of Tumblebug Creek.

14 APPENDIX C: STUDY AREAS

Criteria used in selecting the Middle Fork Willamette release sites were primarily based on the following environmental factors: water temperature, mean velocity, average depth, available cover, stream complexity and macroinvertebrate composition. Because bull trout tolerate a narrow range of temperature conditions to reproduce and survive, and we selected release sites during the warmest period of the year. Only streams with low water temperatures (<10.0°C) and average depths <125 cm were considered favorable. Temperatures in excess of about 15°C are thought to limit bull trout distribution (Allan 1980, Brown 1992, Fraley and Shepard 1989, Goetz 1991, Oliver 1979, Pratt 1985, Ratliff 1992, Shepard and others 1984). Ideal juvenile rearing temperatures are about 7°C to 8°C (Goetz 1989). Optimum temperatures for incubation are between 2°C and 4°C (McPhail and Murray 1979).

Mean velocities of the Middle Fork Willamette sites were analyzed, comparing observations to the McKenzie River bull trout streams. Only streams with low mean velocities, (0-66 cms) and evidence of stability were selected. Studies on juvenile bull trout conducted by Goetz (1994) showed high electivity to low mean stream velocities.

Available cover and stream complexity were evaluated using USFS 1996, Region 6, Stream Survey Methodology. Observations and stream surveys on McKenzie River tributaries showed a high correlation of juvenile bull trout presence near all size classes of woody debris and substrate types. We selected streams with high percentages of woody debris, fine woody debris mixed in the substrate, and low gradients.

On 3 September 1997, benthic invertebrates were collected in the selected McKenzie River tributaries in, or near, areas where juvenile bull trout were observed during prior snorkel surveys. On 4 September, additional invertebrate samples were collected from the three potential Middle Fork Willamette rehabilitation locations and from Middle Fork Willamette mainstem reaches adjacent to possible rehabilitation sites. Within each study stream (i.e. Anderson Creek) invertebrates were collected from three riffle/run habitats using a kick net (500um mesh size) that disturbed 0.9 square meter of stream substrate (total 2.7 square meters per study area). Samples were collected by Willamette National Forest personnel. Aquatic Biology Associates, Corvallis, Oregon, conducted identification and enumeration of taxa and determination of total weight per sample.

Iko Springs Iko Springs is a previously unmapped, spring-fed tributary to the Middle Fork Willamette (Appendix A). Total length is 483 meters with an average width of 5.5 meters. Average depths are between 27.4 cm and 36.5 cm. In a study by Goetz (1994), bull trout used most available total depths up to 125 cm with a peak in use at 25-50 cm both day and night. Mean water temperature during the study period, (August-September 1997) ranged from 6.7°C to 8.9°C. Previously, no annual temperature cycles have been recorded from this stream. A water temperature logger was placed at the mouth in

15 September1997. Discharge at the mouth during the study period was recorded at 12.27 cfs. Annual flow seems to be stable, and very little scour is evident, even after the 1996-1997 high water levels in the Middle Fork Willamette basin. Mean velocity was 34.1 cm/s with a range of 0.0-66.8 cm/s. Average gradient of the valley floor surrounding the stream channel is less than two percent. Most surface runoff apparently is diverted to the surrounding drainages of Found Creek to the northwest and Indigo Creek to the southeast.

A Level II stream survey identified four riffle and four pool habitat units. Substrate was dominated by sand in all eight habitat units (Table 1). A thick layer of moss and fine woody debris also contributed to the substrate structure. Juvenile bull trout are found in close association with the bottom of the channel, often using substrate for cover (Fraley and Shepard 1989, Oliver 1979, Pratt 1984, Shepard and others 1984). In Goetz (1994) juvenile bull trout showed diel preference towards sand substrate, although juveniles were observed feeding during the day over cobble substrate. Snorkel surveys and observations of bull trout in McKenzie River tributaries showed moss and fine woody debris were extensively used as cover.

Table 1. Substrate consistency of eight habitat units in Iko Springs*

Habitat Type % Sand % Gravel % Cobble % Boulder % Bedrock Riffle #1 50 40 10 - - Pool #1 90 10 0 - - Riffle #2 80 10 10 - - Pool #2 100 0 0 - - Riffle #3 80 10 10 - - Pool #3 100 0 0 - - Riffle #4 50 10 40 - - Pool #4 100 0 0 - - *Substrate size classes are consistent with [USFS 1996, Region 6, Stream Survey Methodology].

Of the three rehabilitation sites, Iko Springs has the greatest potential for the development of future spawning habitat. Bull trout spawning sites are often found near springs (Allan 1980, Ratliff 1992, Shepard and others 1984) and are often found in the coldest streams in the basin. In Iko, spawning gravel could be added to the entire area with relative ease. Access is available via the Middle Fork Trail or an abandoned logging road that parallels the streambed.

Available cover includes large woody debris (LWD), rootwads, undercut banks and many short pieces of large diameter wood. In-stream wood correlates with the distribution and abundance of bull trout, habitat complexity in any form can be important (Mullan and others 1992).

LWD was classified and grouped into three categories. Small woody debris (SWD) had at least a 12- inch diameter, 25 feet from the largest end. Medium woody debris (MWD) had at least a 24-inch

16 diameter, 50 feet from the largest end. And large woody debris (LWD) had at least a 36-inch diameter, 50 feet from the largest end (Table 2). Short pieces of woody debris probably influence flows and provide complexity and cover as much as the available LWD but were not considered when LWD was tallied. However, even without considering these short pieces of woody debris, the per mile ratio of LWD can be extrapolated to 103 pieces/mile, still well above USFS recommended standards of 80 pieces/mile. The short pieces of woody debris were undoubtedly rolled into the streambed from past logging in the area. The surrounding riparian zone consists of a variety of tree size classes and should provide adequate recruitment of additional LWD. Additional information from the Level II survey can be acquired through the Willamette National Forest, Rigdon Ranger District, Oakridge, Oregon.

Table 2. Classification of large woody debris for eight habitat units in Iko Springs.

Habitat Type (SWD) (MWD) (LWD) Riffle #1 8 0 0 Pool #1 2 0 0 Riffle #2 9 0 2 Pool #2 0 0 0 Riffle #3 2 0 0 Pool #3 3 0 0 Riffle #4 1 0 0 Pool #4 3 1 0

Shadow Springs Shadow Springs is another previously unmapped spring-fed tributary to the Middle Fork Willamette. Two springs actually flow into a large cold-water pool to form Shadow Springs. Both springs flow out of the hillside just north of the Middle Fork Trail (Appendix A). Total length of both springs is about 60 meters and will provide minimal rearing habitat for juvenile bull trout. However, the pool area is approximately 468 square meters and should supply adequate rearing habitat. Several out-flows connect the pool to the Middle Fork Willamette. The riparian zone consists of mature Douglas-fir and hemlock trees.

Water temperatures of the Shadow Springs pool ranged from 7.2°C to 8.3°C during the study period. Average temperatures of the two feeder springs were 6.6°C and 10.0°C. Stream survey data were only collected from the coldest feeder spring. Discharge at the mouth of the spring was 7.41 cfs with an average depth of 42.7 cm. Flows of both springs and the connecting pool show limited scouring and seem stable.

Substrate is primarily sand and silt with a thick covering of moss and fine woody debris. Very little gravel is available in either the spring channels or the pool. Little probability exists in creating spawning habitat in the area, mainly because of the limited area of the springs and pool.

Available cover consists of LWD, rootwads, undercut banks, short pieces of large-diameter woody debris, and extremely dense woody vegetation throughout the pool area.

Chuckle Springs In June 1997, 96 of 178 bull trout fry transferred from Anderson Creek were released into Chuckle Springs. Fry were positively identified during a snorkel survey on 29 July 1997. In August 1997, probable bull trout fry were once again observed near the release site.

Chuckle Springs could again be utilized as a release site in 1998. Water temperatures were between 5.6°C and 7.8°C during the study period. Discharge at the 1997 release site was 37.9 cfs and the average depth was 49.4 cm.

Substrate is mainly sand and silt with a thick layer of moss and detritus. There is currently very little gravel present, but possibilities of adding spawning substrate do exist. Large wood, undercut banks, and fine woody debris are available throughout the spring to supply cover and refuge area. Access is available from the Chuckle Springs Trail or Middle Fork Trail. The riparian zone consists of mature Douglas-fir and hemlock trees, which should supply an adequate source of LWD for future recruitment.

Aquatic Invertebrates At all study locations aquatic invertebrate communities were predominately composed of intolerant, cold-water adapted taxa within the orders Ephemeroptera (mayflies), Plecoptera (stoneflies), Trichoptera, (caddisflies) and Diptera, (true flies). This reflects the thermal regime of cold, spring-fed systems that are preferred bull trout spawning and rearing habitat in Oregon. Because the focus of the aquatic invertebrate section of this report is on composite food base parameters, a more detailed taxonomic discussion is not included herein. Additional taxonomic information is available through the Willamette National Forest, Rigdon Ranger District, Oakridge, Oregon.

Macroinvertebrate mean standing crop (wet weight biomass), abundance, and taxa richness found during the study are presented in Figure 2. Comparisons are based on a mean value of all samples taken from the McKenzie Basin reference stations versus a mean value of all samples collected at Middle Fork Willamette stations, compared to mean values at each potential rehabilitation site. A more detailed presentation of the data, which includes a measure of data variability, is available in Figures 3-5. However, the observed variability in the data set does not obscure overall biological patterns between sites, and for the sake of clarity, the following discussion will focus on information presented in Figure 2.

The study areas ranked consistently for values of standing crop, abundance and taxa richness. In general, the McKenzie River tributary reference streams and the mainstem Middle Fork Willamette exhibited the highest values, with progressive decrease in the order of Iko Springs, Shadow Springs and Chuckle Springs. Of the three possible Middle Fork Willamette rehabilitation sites, Iko Springs compared most favorably to McKenzie reference stations with a mean wet weight of 50% of the reference stations, abundance of 100% and taxa richness of 98%. September data suggest that the

18 invertebrate populations at both Shadow and Chuckle Springs are relatively depauperate both in terms of abundance and community richness. Although the invertebrate biomass measured at Shadow Springs was approximately 1/3 that of the McKenzie study areas, the mean abundance was only 5%, and taxa richness 48% of reference streams. Relative to McKenzie streams, Chuckle Springs had an even lower standing crop (1%), abundance (2%) and taxa richness (38%). Aquatic macroinvertebrate populations within Shadow and Chuckle Springs may be limited by a combination of factors which include: 1) Few or no riffle areas composed of cobble/gravel substrates, which are the most productive instream habitats for macroinvertebrates (Keup 1988), 2) Limited immigration from upstream reaches that were nearly vertical falls, 3) Instream primary production dominated by mosses that are not immediately palatable by most stream invertebrates, 4) Chuckle Springs has limited leaf energy inputs from the riparian vegetation due to the shading effects of streamside mature conifers further limiting food availability, 5) Seasonality in macroinvertebrate abundance and biomass that is typically lowest in late summer and highest in the spring. In contrast, Iko Springs had considerably more riffle habitat, a 1/4 mile low-gradient reach prior to confluence with the Middle Fork Willamette, and an inner riparian zone of hardwoods capable of providing a food source for macroinvertebrates derived from seasonal leaf fall.

19 APPENDIX D: MIDDLE FORK WILLAMETTE HABITAT AND PASSAGE IMPROVEMENT PROJECTS

20 WORKS CITED

Allen, J. H. 1980. Life history notes on the Dolly Varden char (Salvelinus malma) min the Upper Clearwater River, Alberta. Manuscript Report. Red Deer, AB: Energy and Natural Resources, Fish and Wildlife Division. 58p.

Brown, L. G. 1992. Draft Management guide for the bull trout Salvelinus confluentus (Suckley) on the Wenatchee National Forest. Wenatchee, WA: Washington Department of Wildlife. 75p.

Fraley, J. J. and B.B. Shepard. 1989. Life history, ecology and population status of migratory bull trout (Salvelinus confluentus) in the Flathead lake river system, Montana. Northwest Science. 63(4):133-143.

Goetz, F. 1989. Biology of bull trout Salvelinus confluentus, a literature review. Eugene, OR.: U.S. Department of Agriculture, Forest Service, Willamette National Forest. 53p.

Goetz, F. 1991. Bull Trout life history and habitat study. Corvallis, OR: Oregon State University.49p. Thesis.

Goetz, F. 1994. istribution and juvenile ecology of bull trout. (Salvelinus confluentus) in the Cascades Mountains. Masters Thesis, Oregon State University, Corvallis, Oregon.

Howell, P.J. and D.V. Buchanan. 1992. Introduction: Proceedings of the Gearhart Mountain bull trout workshop. In: Howell, P.J. and D.V. Buchanan, eds. 1992. Proceedings of the Gearhart Mountain bull trout workshop. Corvallis, OR: Oregon Chapter of the American Fisheries Society: vii-x.

Howell, P.J. and D.V. Buchanan, eds. 1992. Proceedings of the Gearhart Mountain bull trout workshop. Corvallis, OR: Oregon Chapter of the American Fisheries Society.

Keup, L. E. 1988. Invertebrate fish food resources of lotic environments. Instream flow information paper No. 24. USFWS. Biol. Rep. 1988. (13): 96pp.

McPhail, J. D. and C.B. Murray. 1979. The early life history and ecology of Dolly Varden (Salvelinus malma) in the Upper Arrow Lakes. Vancouver, BC: University of British Columbia, Department of Zoology and Institute of Animal Resources. 113p.

Oliver, G. 1979. A final report on the fisheries of the Wigwam River with an emphasis on the migratory life history and spawning behavior of the Dolly Varden charr Salvelinus malma (Walbaum). Fisheries investigations in tributaries of the Canadian portion of Libby Reservoir. Victoria, BC: British Columbia Fish and Wildlife Branch. 27p.

Pratt, K. L. 1984. Habitat use and species interactions of juvenile cutthroat (Salmo clarki

21 lewisi) and bull trout (Salvelinus confluentus) in the Upper Flathead River basin. Moscow, ID: University of Idaho. 95p. Thesis.

Pratt, K. L. 1985. Pend Oreille trout and char life history study. Boise, ID: Idaho Department of Fish and Game. 105p.

Ratliff, D. E. 1992. Bull trout investigations in the Metolius River-Lake Billy Chinook system. In: Howell, P.J. and D.V. Buchanan, eds. Proceedings of the Gearhart Mountain bull trout workshop; 1992 August; Gearhart Mountain, OR. Corvallis, OR: Oregon Chapter of the American Fisheries Society: 37-44.

Rieman, B. E.and J.R. Lukens. 1979. Lake and reservoir investigations: Priest Lake creel census. Job Completion Rep., Proj. F-73-R-1, Subproj. III, Study I, Job I. Boise, ID: Idaho Department of Fish and Game. 105p.

Shepard, B.; Pratt, K.; Graham, P. 1984b. Life history of westslope cutthroat and bull trout in the upper Flathead River basin, Montana. Kalispell, MT: Montana Department of Fish, Wildlife and Parks. 85p.

US Army Corps of Engineers. 1996. Middle Fork Willamette River, Oregon: Fishery restoration initial draft reconnaissance study.