BEAVERHEAD NATIONAL FOREST FISHERIES: SECOND ANNUAL REPORT COVERING THE PERIOD JANUARY TO DECEMBER 1986 Bradley B. Shepard Montana Department of Fish, Wildlife and Parks c/o Beaverhead National Forest 610 North Montana Street Dillon, Montana 59725 A cooperative study between the Beaverhead National Forest and the Montana Department of Fish, Wildlife and Parks April, 1987 TABLE OF CONTENTS TABLE OF CONTENTS ii LIST OF TABLES *\ , v LIST OF FIGURES x APPENDICES . xi ACKNOWLEDGEMENTS ........... xii EXECUTIVE SUMMARY ....... . . xiii INTRODUCTION ..... ...... 1 STUDY AREA DESCRIPTION 2 DRAINAGE . . 2 GEOLOGY . 2 BIOTIC COMMUNITY ..... 4 METHODS 5 HABITAT .... ....... 5 Reach Sur.yeya 5 Detailed Habitat Surveys of Sample Sections .... 12 Streambed Sampling 13 FISH ABUNDANCE ESTIMATES 13 Catch per Unit Effort . 14 Tributary Population Estimates ... . 14 Big Hole River Population Estimate ..... 15 LENGTH, WEIGHT, AND CONDITION FACTOR . 15 FISH MOVEMENT 15 Tributary Trapping During the Spring 15 Rainbow Trout Redd Counts 17 Arctic Grayling Radio telemetry. 17 ii MONTANA STATE LIBRARY S 567.55045 F2bnff 1987 c.1 v.1 Shepard Beaverhead National Forest Fisheries :se 3 0864 00062265 7 TABLE OF CONTENTS (continued) Plastic Coded Tagging .... 19 STATISTICAL ANALYSES 19 Tests for. Normality 19 Comparisons Betwep.n Methodologies 22 Habitat Data 22 Fish Data 22 Coxxelations Between Habitat Variables and Fish Densities 22 Test of the COWFISH Model . 22 RESULTS , , 24 HABITAT ....... ....... 24 Reach Characteristics 24 Map Derived Information ......... 24 Habitat Composition 24 Frequency of Debris and Spawning Habitat ..... 26 Detailed Habitat Survey within Sample Sectior^ 26 Physical Character of Stream Channel 26 Cover • 31 Sfcre.ambed Condition . . . . 31 Uc u 1 a r f'Sti ma t e s ■.«•>«•..»«...«».«•.<..».. 38 Embeddedness Measurements ... ... 38 Embeddedness Variation within a Reach .... . 38 Comparison Between Visual and Measured Embeddedness 44 Hollow Core Samples ........... ..... 44 Comparison Between Measured Embeddedness and Hollow Core "Fines" . 48 Principle Component Analysis (PCA) ...... ..... 48 FISH ABUNDANCE 51 iii TABLE OF CONTENTS (continued) Catch per Unit Effort (CPUE) . 51 Tributary Population Estimates 51 Evaluation of Different Estimation Techniques I ..... 60 Presence of "Sensitive Species" by District 60 Big Hole River Population Estimate 67 LENGTH, WEIGHT, AND CONDITION FACTOR ...... 67 Length Frequencies 67 Length-Weight Relationships ..... 72 FISH MOVEMENT 72 Spring Trapping • « 72 Bjg Lake Creek 72 Sandhollow Creek 76 Steel, Creek 76 Sjgamp. Cr_e_gk_ 76 Arc tic Gr ay, 1 lag. EasLig. t-el£me_t£X 77 Plastic Tag Returns 79 RELATIONSHIPS BETWEEN HABITAT VARIABLES AND FISH DENSITIES 81 Spearman Rank Correlations 81 Uptransformed Habitat Variables versus Fish Densities 81 Test of COWFISH Model . 35 LITERATURE CITED 89 MDFWP WATER CODES AND KEYWORDS 91 APPENDICES Separate document iv LIST OF TABLES Page Table 1. Description of stream reaches surveyed during 1986 including stream, reach, landmarks at lower and upper boundaries of each reach, legal descriptions of lower and upper boundaries of each reach, elevations (ft) at lower and upper boundaries of each reach, sample site legal description and length (ft), length of stream channel (mi), length of valley (mi), channel gradient (%) , channel type (from Rosgen 1985), stream order, and channel sinuosity ....... 6 Table 2. Location (stream and legal description) and date sampled for hollow core sampling conducted in waters draining the Beaverhead National Forest during 1985-86 11 Table 3. Areas searched for radio tagged arctic grayling in the upper Big Hole River drainage during the fall of 1986 including date of search, area searched, how search was conducted, water temp- erature (F, n.d. signifies no data), and general comments .... 20 Table 4. Percentage of each main habitat type within each reach of streams draining the Beaverhead National Forest surveyed during 1986 . . 25 Table 5. A comparison between the percentage of stream habitat in pools and riffles estimated by surveying the entire reach versus a 400 to 1200 foot sample section and the results of a Wilcoxon sign- ranked test (Daniel 1978) testing between the different surveys . 27 Table 6. Percentage of each type and class of pools (class V is the highest quality pool) within each reach of streams draining the Beaverhead National Forest surveyed during 1986 .... 28 Table 7. Percentages of each type of riffle habitat for stream reaches draining the Beaverhead National Forest surveyed during 1986 . . 29 Table 8. Frequency (number per mile) of large (six inches in diameter or larger) and small (less than six inches in diameter) debris, frequency which these large and small debris cross the wetted channel, and amount of spawning gravel observed (square feet of gravel per stream mile) by reach in streams draining the Beaverhead National Forest surveyed during 1986 .30 Table 9. Date of survey, temperature at time of survey (F) , flow (cfs), total length of surveyed section (ft), and mean estimated wetted width (ft), channel width (ft), average depth (in), thalweg depth (in), and average depth (in) at the streambank for stream reaches draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses ............. 32 LIST OF TABLES (continued) Table 10. Mean estimates of cover availability including percentage under- cut banks, canopy density over the water's surface (%) , instream cover (%), low (1.0 foot or less above the water's surface) over- head cover (%), high (more than 1.0 foot above the water's surface) overhead cover (%), and depth of undercut banks (in) for stream reaches draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses 34 Table 11. Mean estimates of percentage of streambank altered, streambank vegetation stability rating, percentage of streambank vegetation utilized, and bank angle (degrees) for stream reaches draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses ..... 36 Table 12. Mean estimates of substrate composition (percent by size class), embeddedness (%) , substrate score, and D-90 (in) for stream reaches draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses 39 Table 13. Average percent embeddedness, "free matrix particles", and average size of particles measured (range in inches) from measurements made of streambed particles in riffle habitat types of stream reaches draining the Beaverhead National Forest surveyed during 1986 42 Table 14. Comparison between visual and measurement estimates of embed- dedness in riffles made during 1986 in streams draining the Beaverhead National Forest 45 Table 15. Average percentage of material (by dry weight) less than 0.37 inch (9.5 mm), 0.25 inch (6.34 mm), and less than 0.03 inch (0.85 mm) from hollow core samples taken from typical spawning areas during 1985-86 and predicted survivals of westslope cutthroat trout (WCT), rainbow trout (RB), and eastern brook trout (EBT) embryos from egg deposition to fry emergence based on laboratory studies by Irving and Bjornn (1984) for WCT and RB and survival relationships developed by the author using data from Witzel and MacCrimmon (1985) ..... .... 46 Table 16. Principle components analysis for habitat variables measured in the "detailed habitat survey section" across 442 transects in 17 stream sections surveyed during 1986 50 Table 17. Relative fish abundance by species for fish 3.0 inches and longer in streams draining the Beaverhead National Forest with- in the Dillon Ranger District derived from single pass electro- fishing catches using a Coffelt BP-1C backpack electrof isher during 1985 and 1986 52 vi LIST OF TABLES (continued) Table 18. Relative fish abundance by species for fish 3.0 inches and longer in streams draining the Beaverhead National Forest with- in the Wise River Ranger District derived from single pass electrof ishing catches using a Coffelt BP-\1C backpack electro- fisher during 1985 and 1986 53 Table 19. Relative fish abundance by species for fish 3.0 inches and longer in streams draining the Beaverhead National Forest with- in the Wisdom Ranger District derived from single pass electrof ishing catches using a Coffelt BP-1C backpack electro- fisher during 1985 and 1986 54 Table 20. Estimated fish populations in stream reaches draining the Beaverhead National Forest surveyed during 1985 and 1986. Population estimates were calculated using a two-pass estimator (2P) (Seber and LeCren 1967), maximum-liklihood estimator (ML) (Van Deventer and Platts 1985), mark -recapture estimator (MR) (Ricker 1975), or snorkel counts (SNOKK) 56 Table 21. Comparison of depletion (two or more consecutive electrof ishing catches) and mark-recapture estimates and their associated efficiences (probability of capture for depletion = p; and number of recaptures divided by the total number marked for mark- recapture) from estimates made during 1985 and 1986 in streams draining the Beaverhead National Forest 61 Table 22. Relative numbers and population estimates (where available) of westslope cutthroat trout (identified from external morphological characteristics) ranked from highest to lowest catch per unit effort (CPUE from one electrof ishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Dillon District, Beaverhead National Forest surveyed during 1985-86 63 Table 23. Relative numbers and population estimates (where available) of westslope cutthroat trout (identified from external morphological characteristics) ranked from highest to lowest catch per unit effort (CPUE from one electrof ishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Wise River District, Beaverhead National Forest surveyed during 1985-86 . . 64 Table 24. Relative numbers and population estimates (where available) of westslope cutthroat trout (identified from external morphological characteristics) ranked from highest to lowest catch per unit effort (CPUE from one elec trof ishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Wisdom District, Beaverhead National Forest surveyed during 1985-86 ....... 65 vii ,^MW£W0'^^M* '"''• 'tf:'W'«": :^mW'f- LIST OF TABLES (continued) Table 25. Relative numbers and population estimates (where available) of arctic grayling ranked from highest to lowest catch per unit effort (CPUE from one electrof ishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Beaverhead National Forest surveyed during 1985-86 ... 66 Table 26. Estimated populations of arctic grayling and eastern brook trout in the McDowell section of the Big Hole River above Wisdom, Montana in late June, 1986 using a mark -recapture estimator ... 68 Table 27. Regression equations for log transformations of fish length versus fish weight by species and water from data collected during 1986 in waters draining the Beaverhead National Forest . . 73 Table 28. GM regression equations (Ricker 1975) for log transformations of fish length versus weight by species and water from data collected during 1986 in waters draining the Beaverhead National Forest 73 Table 29. Redd surveys conducted during the spring of 1986 by stream, type of survey, date of survey, and number of redds located 74 Table 30. Summary of results of up- and downstream traps located in Swamp, Steel, and Big Lake creeks and fry drift nets located in Sandhollow Creek during the spring of 1986 75 Table 31. Comparison between the condition factors of two radio tagged arctic grayling versus six non-radio tagged arctic grayling approximately two months (July and August) after radio implants . 78 Table 32. Tag return information for fish tagged during 1985 and 1986 and recaptured during 1986 from waters in the upper Big Hole River drainage • 80 Table 33. Spearman rank correlation coefficients between habitat variables and the number of eastern brook trout per acre by length class. Significance levels are P < 0.05 (**) and P < 0.10 (*) . Nineteen stream sections were sampled. The five highest correlations are in boldface type ........................82 Table 34. Spearman rank correlation coefficients between habitat variables and the number of westslope cutthroat trout per acre by length class. Significance levels are P < 0.05 (**) and P< 0.10 (*) . Eight stream sections were sampled. The five highest cor- relations are in boldface type « 83 Table 35. Stepwise multiple regression results for regressions of habitat variables against density of eastern brook trout by size class from data collected in drainages draining the Beaverhead National Forest during 1985 and 1986 84 viii LIST OF TABLES (continued) Table 36. Stepwise multiple regression results for regressions of habitat variables against density of westslope cutthroat trout by size class from data collected in drainages draining the Beaverhead National Forest during 1985 and 1986. . .V 86 ix . (.0 LIST OF FIGURES Figure 1. Map of the upper Big Hole River drainage \ showing study reaches (shaded areas) studied during 1986 as part of the cooperative fisheries study between the BNF and MDFWP 3 Figure 2. Schematic diagram of upstream and downstream fish traps and orientation of traps in the stream for trapping of migrating fish in the spring of 1986 as part of the cooperative fisheries study between the BNF and MDFWP . . 16 Figure 3. Location of fish trapping sites in the upper Big Hole River drainage in the vicinity of Wisdom, Montana which were operated during the spring of 1986. Sites in Big Lake, Steel, and Swamp creeks had up- and downstream traps, while the site in Sandhollow Creek had drift nets <. . . 18 Figure 4. Pie chart diagrams of substrate composition which was estimated ocularly for two reaches in Elk, Joseph, and Sheep creeks which are all tributaries of Trail Creek. Percentages of fine material (less than 0.25 inch) are offset . 41 Figure 5. Relationship between "percent fines" (material less than 0.25 inch) sampled by hollow core sampling and percent embeddedness sampled by measuring at least 100 individual particles at the same sample site 49 Figure 6. Length frequency histograms for fish captured in streams on Dillon District during 1986 ......... 69 Figure 7. Length frequency histograms for fish captured in streams on Wise River District during 1986 ..... 70 Figure 8. Length frequency histograms for fish captured in streams on Wisdom District during 1986 . ..... 71 Figure 9. Relationship between the actual number of fish estimated and the existing number of fish predicted by the COWFISH habitat model at 13 sites sampled during 1986 which contained brook trout 87 Figure 10. Relationship between the actual number of fish estimated and the existing number of fish predicted by the COWFISH habitat model at four sites sampled during 1986 which contained cut- throat trout 87 APPENDICES (separate document) Appendix A. Description and explaination of information obtained from USGS maps (scale: 1:24,000). Appendix B, Description and explaination of information collected during habitat surveys conducted by walking the entire reach (or a minimum of one mile). Appendix C. Description and explaination of information collected during detailed habitat surveys conducted within the sample section of each reach. Appendix E. Appendix D. Mean estimates of temperature (F) , streamflow (cfs), length of each habitat unit (ft), wetted width (ft), channel width (ft), average depth (in), thalweg depth (in), and average depth at each bank (in) for streams draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses . Mean estimates of cover availability including percentage undercut banks, canopy density over the water's surface (%), instream cover (%), low (1.0 foot or less above the water's surface) overhead cover (%), high (more than 1.0 foot above the water's surface) overhead cover (%), and depth of undercut banks (in.) by habitat type for waters draining the Beaverhead National Forest surveyed during 1986. Appendix F. Embeddedness data by sample (hoop) and stream reach for streams sampled during 1986. Appendix G. Mean lengths and weights, condition factors and sample sizes by stream, reach, and species for all fish captured in streams draining the Beaverhead National Forest surveyed during 1986. xi E&FTZM "' -'**,»; *. ■ ' ■: ifi:-^:.: "i ACKNOWLEDGEMENTS The Beaverhead National Forest (BNF) and Montana Department of Fish, Wildlife and Parks (MDFWP) provided funding and administrative and technical support for this study. Personnel of the BNF who helped make this study possible were Mike Rath and Dan Pence who provided administrative support and District Rangers Joe Spehar, Ed Levert, and Barry Hicks and their staffs who provided logistical field support. Jim Lloyd, Zone Fisheries Biologist, helped with technical and administrative support. Jerry Wells and Dick Oswald of the MDFWP helped with technical and administrative support. Mr. Oswald also helped with fieldwork and study design. Bob McFarland, MDFWP, helped with the statistical anaylses and computer graphics. Greg Gibbons and Shawn Latham assisted in the collection of field data. Bill Schrader, Bill Gould, and Bob White of the Montana Cooperative Fisheries Research Unit assisted with radio tagging by helping implant tags and loan of receivers and several transmitters. xii EXECUTIVE SUMMARY This report documents the activities of the cooperative fisheries program between the Beaverhead National Forest (BNF) and the Montana Department of Fish, Wildlife and Parks (MDFWP) from 1985 through 1986. Data were collected on fish habitat, abundance, and spawning and overwinter movements. Electrof ishing catch per unit effort (standardized as the number of fish 3.0 inches and longer captured in one electrof ishing pass per 1,000 feet of stream length) ranged from 1 to 33 for arctic grayling, 1 to 177 for cutthroat trout (based on external morphological characteristics, some were "pure" westslope, some were Yellowstone, and some were probably introgressed with rainbow trout), 2 to 35 for rainbow trout, and 3 to 660 for brook trout in sections where the above species were captured. Streams (Ranger District coded by BNF number) in which fluvial arctic grayling have been documented by this or any other MDFWP surveys are Big Lake (D-3) , Deep (D-2) , Fishtrap (D-2) , Francis (D-3), Governor (D-3), LaMarche (D-2), Rock (D-3), Sandhollow (D-3), Steel (D-3), and Swamp (D-3) creeks, and the North Fork Big Hole River (D-3). Fluvial arctic grayling distribution appears to be limited to the Big Hole River and the lower portions of its tributaries within the main river valley bottom above Divide, Montana. Streams on the Dillon District which have been found to contain westslope cutthroat trout are Andrus , Brown's Canyon, Fox, Governor, Painter, Pole, Reservoir, and Thayer creeks. Streams on the Wise River Ranger District where westslope cutthroat trout have been found are Adson, Harriett Lou, Lacy, Meadow, Mono, and Wyman creeks. Only Doolittle and the South Fork Steel creeks have been found to support westslope cutthroat trout on the Wisdom District. Cutthroat trout distribution appeared to be limited to small headwater and/or high gradient tributaries. Frequently, cutthroat trout populations were found above a fish migration barrier. Rainbow trout were found in reach 1 (Rl) LaMarche and R2 Wyman creeks. The likely source of the Wyman Creek rainbow was past releases of hatchery rainbow into Lake of the Woods between 1941 and 1960. The source of rainbow in LaMarche Creek could be either hatchery releases made into LaMarche Creek between 1928 and 1954 or from fluvial Big Hole River populations. Brook trout (charr) were the most commonly found trout species. High densities (at least 150 fish per 1,000 feet of stream length based on a population estimate) of brook trout 6.0 inches and longer have been documented in R2 Governor, R2 LaMarche, and R2 Wyman creeks. R2 Elk, R2 Joseph, Rl LaMarche, R2 Old Tim, Rl Steel, and R2 Trail creeks all had high densities (at least 180 fish per 1,000 feet of stream length based on a population estimate) of brook trout 3.0 to 5.9 inches in length. Extremely low densities of brook trout were observed in Rl Adson, R2 Cow Cabin, R2 Morrison, R2 Pole, R2 Ruby, Rl Sheep, R2 Steel, Rl Trail, and Rl Wyman creeks. All of these reaches except Rl Adson and R2 Pole creeks had received moderate to high livestock impacts. The streams in R2 Cow Cabin and R2 Morrison as well as the above two reaches were small headwater type streams and low fish densities are to be expected in these types of reaches. Depletion-type estimators (two or more consecutive electrof ishing passes) appeared to consistently underestimate fish numbers when compared to a mark- recapture estimator, and this bias seems to be high when probability of capture xiii values are lower than 0.75. Underwater census techniques appear to have value in certain types of waters and may be able to provide good estimates when applied as the recapture technique using a mark-recapture estimator, provided an easily visible external mark or tag can be found. A population estimate made in the Big Hole River immediately above Wisdom during late June yielded an estimate of 35 arctic grayling 8.5 inches and longer (Age 11+) and 282 brook trout 9.0 inches and longe . This number of arctic grayling is much lower than a previous estimate obtained by Oswald (1984) of 105 per river mile and is cause for concern. Arctic grayling appear to be very suseptable to angling with easily recognizeable hooking scars observed on 15% of all captured grayling 10.0 inches and longer. Rainbow trout redds were found in Jerry and Big Lake creeks, but not found in Bryant Creek. A large mature rainbow (19.0 inches long) was captured by an angler at the mouth of Steel Creek during the spawning season which may indicate Steel Creek is used for spawning. An effort to document arctic grayling movement into or out of four Big Hole River tributaries (Big Lake, Sandhollow, Steel, and Swamp creeks) using . fish traps and drift nets captured only three grayling moving downstream out of Swamp Creek. These three grayling were captured immediately after the opening of the 1986 fishing season and it is likely they were displaced due to the stress of being captured and released by anglers. Seven arctic grayling were radioed by implanting radio transmitters during September. It was hoped that fall movements to overwinter habitat could be documented. Relocations were obtained for all but one fish. All subsequent relocations, but one, indicated the fish moved downriver. This downriver movement was either a slow staged movement from large pool to large pool or very fast active movement of up to six miles in eleven days. Unfortunately, no confirmed signals were received after October 21 even though the river from five miles above Wisdom down to Divide and lower Steel and Swamp creeks were searched . Plastic coded tag return information indicated that tagged grayling and rainbow trout moved very little during the summer. The longest recorded movements were made by one juvenile grayling which moved 4.2 miles downstream out of Big Lake Creek into the Big Hole River between May 15 and September 4 and another juvenile grayling which moved approximately 2.8 miles up Steel Creek from the east channel of the Big Hole River between May 21 and August 27. Tag return rates were 8%, 24%, and 63% for juvenile grayling, adult grayling, and adult rainbow trout, respectively. Almost all of these tag returns were from fish captured during the course of this and other MDFWP studies indicating anglers are not a sufficient source of tag return information at present. Habitat data were collected throughout study reaches and within each sample section. Pool habitat was abundant in R2 Mono and R2 Wyrnan creeks. Pools were sparse in Rl Wyman and moderately low in number in R2 Elk, Rl LaMarche, R2 Meadow, Rl Mono, and Rl and R2 Sheep creeks. A comparison between the percentages of pool and riffle habitat types estimated within the 400 to 1200 foot sample sections versus those estimated within the entire reach found no significant difference between the two methods (P > 0.10), however, in xiv individual cases there were large differences. Habitat condition appeared to be related to livestock use, especially for those habitat parameters which were related to streambank condition and cover. Streambed condition was assessed visually, by measuring embeddedness , and by sampling with a hollow core sampler. Embeddedness estimates found that R2 Sheep and R2 Elk creeks both had high embeddedness'Umore than 60% embedded). No difference was observed between embeddedness estimates in two different riffles within the same reach, especially when the two sites were located near each other. No significant difference (P > 0.10) was found between embeddedness estimated visually versus that estimated by measurements, however, individual pairs of estimates did appear to differ. Hollow core sampling found that few sampled sites contained less than 25% "fines" (material less than 0.25 inch) with most sites between 30 and 40 percent. Several sites (two in Trail Creek, one in Blacktail Creek, and one in Adson Creek) had more than 40% fine material which would indicate potential problems. It was believed the sampling biased the Adson samples by sampling in silts underneath the streambed gravels in seven of the ten samples. The distribution of fine sediment within the Trail Creek drainage on the Wisdom Distict is discussed. Linear regression between measured embeddedness and percentage of fine material in hollow cores found a moderately good correlation (r = 0.67), but the spread of values at the higher levels of impact was disturbing. Principle component analysis (PCA) was used to group habitat variables. The PCA function which explained the most variability in all habitat parameters measured at all sites weighted streambed variables most heavily. The relationship between habitat variables and fish abundance was evaluated using Spearman rank correlations. There were significant positive correlations (P < 0.10) between the density of brook trout 6.0 inches and longer and the percentage of high class pools and between the density of cutthroat trout 6.0 inches and longer and the percentage of small gravel in the streambed. There were significant negative correlations between the percentage of low class pools (P -4. 0.10) and the density of brook trout 6.0 inches and longer, between the percentage of large gravel (P < 0.05), the amount of spawning habitat (P < 0.10), and channel sinuosity (P < 0.10) and the density of cutthroat trout 3.0 to 5.9 inches in length, and between the density of cutthroat trout 6.0 inches and longer and stream order (P < 0.10). For all cutthroat trout 3.0 inches and longer there was a highly significant negative correlation (P< 0.05) with stream order and channel sinuosity which suggests that cutthroat trout are more abundant in smaller, straighter stream channels which are usually associated with headwater portions of tributaries. Stepwise multiple regressions between habitat variables and fish densities provided little insight into habitat variables which influenced brook trout densities (R ranged between 0.31 and 0.36), but showed promise assessing the influence of habitat variables on cutthroat trout (R ranging between 0.58 to 0.95). The habitat variables comprising this regression were percentage of the streambed in large gravel, bank angle, and channel sinuosity. Extremely small sample sizes presently limits the utility of these relationships, however, this area represents a fertile area for future development. Evaluation of the COWFISH model found that the model appears to have limited utility when applied to streams supporting brook trout, but may have xv utility for streams containing cutthroat trout. The comment on sample size above is also pertinent here. Another finding was that it appears that, at least in the case of cutthroat trout, the COWFISH model underestimates the number of catchable (6.0 inches and longer) by a factor between 2.0 and 3.0. xvi INTRODUCTION \- This report documents the activities of the cooperative fisheries program between the Beaverhead National Forest and the Montana Department of Fish, Wildlife and Parks for the period January 1, 1986 to December 31, 1986. This program was initiated during the late summer of 1985. During 1985 preliminary data collection was begun and data interpretation for that data is done in this report. This year, results have been separated into two reports. This report details the methodology and describes the statistical analyses used to describe relationships between fish abundance and habitat condition, test methodologies, and contrast differences between different stream reaches. A companion report entitled "Beaverhead National Forest Fisheries - Streams Surveyed During 1985-86" summarizes the fish and fish habitat information by stream reach. The objectives of this cooperative program are: 1> Collect baseline fisheries and hydrologic information in areas that are designated for intensive timber harvest activities. 2. Collect baseline fisheries information on various grazing allotments to evaluate grazing strategies and help calibrate the Forest Service's COWFISH model. 3. Determine fish populations in selected streams. 4. Determine the present condition of game fish habitat and identify factors which may be presently limiting game fish populations in streams draining Forest lands . 5. Identify tributary streams which provide spawning and/or rearing habitat for mainstem riverine fish populations. 6. Cooperatively work with the Forest Service Zone Fisheries Biologist to develop a positive fisheries program regarding habitat protection and enhancement opportunities . Objective 2 was added to evaluate grazing impacts on fisheries resources (in addition to impacts from timber related activities) in response to the present updating of several allotment management plans and a desire to include fishery objectives in those plan updates. STUDY AREA DESCRIPTION The primary study area includes the Big Hole River drainage above Divide, Montana (Figure 1). The area includes the upper 100 miles of the 156 mile long Big Hole River drainage covering an area of approximately 1,635 square miles. The upper Big Hole River flows approximately 90 miles through a wide, high-altitude basin surrounded by the Beaverhead Mountains, Pioneer Mountains, and the Anaconda Range before entering a narrow canyon at Wise River which contains the river for the lower ten miles within the study area. The Big Hole River joins with the Beaverhead River near Twin Bridges, Montana to form the Jefferson River. The Jefferson River is a tributary to the Missouri River. Much of the following description is from Levings (1986). DRAINAGE Tributaries to the Big Hole River along the west and north sides of the study area generally contribute the majority of the water to the river with the exception of the Wise River drainage which joins the Big Hole River at Wise River. The U.S. Geologic Survey has gauging stations on the Big Hole River at Melrose, Montana (river mile 31.1 or approximately 24 miles below Divide, Montana) for which there is data from 1923 to the present and on Wise River near Wise River, Montana for which there is data from 1972 to the present. Peak flows generally occur in the late May to mid-June period with east- and north-side tributaries usually peaking slightly earlier than west- and south- side tributaries. For the water year 1985-86 the Big Hole River discharge at the Melrose gauging station recorded a total annual flow 95% of the average annual discharge for the period of record (USGS preliminary data, Helena, Montana) which made this year a near normal year. Flood irrigation has been the accepted practice in the upper Big Hole valley for many generations. A large number of the tributaries to the Big Hole and Wise rivers are partially or entirely diverted near the valley sidewalls to provide water to hay and pasture lands. Portions of the Big Hole and Wise rivers are also diverted at various points along their lengths to provide irrigation water. Normally these diversions began diverting water to the fields at the onset of spring runoff and remain open until sometime in July. At that time the water is shut off to allow the hayfields time to dry before the hay is cut. Because of the relatively short growing season ranchers only cut hay once a year. After the harvest diversions are again opened and remain open until the late fall. GEOLOGY From the divide to the valley floor the upper west side of the basin (from Governor Creek north to Ruby Creek) is underlain by basement sedimentary rocks (primarily fine-grained impure quartzites). The west side of the basin from Trail Creek to Fishtrap Creek is underlain primarily by intrusive rocks with a few large isolated areas of glacial till. From Fishtrap Creek to Deep Creek the north side of the valley is underlain by a mixture of intrusives, Precambrian belt rocks, coarse valley fill, and alluvial deposits. From Deep Creek to Divide Creek the north side of the valley is underlain by intrusives, Precambrian belt rocks, alluvial deposits, volcanics, Precambrian quartzites, BiG HOLE DRAINAGE Figure 1. Map of the upper Big Hole River drainage showing study reaches (shaded areas) studied during 1986 as part of the cooperative fisheries study between the BNF and MDFWP. '■;£■ •;■■•'-:?: siltitesj and argillites, and shales , sandstones and limestones. The upper valley bottom, from the headwaters to Fishtrap Creek is filled with glacial and alluvial deposits. The lower valley bottom is dominated by alluvial deposits. The east side of the drainage from the headwaters to Wise River is underlain by intrusives, Precambrian quartzites, siltites, and argillites, coarse valley fill, and several large isolated areas of glacial till. The south side of the valley from Wise River to Divide Creek is underlain^ by intrusives, Precambrian quartzites, siltites, and argillites, and shales, sandstones and limestones. The intrusives and glacial and alluvial deposits are generally the most erosive followed by the shales, sandstones, and limestones. Quartzites, siltites, and argillites generally are resistent to rapid erosion. BIOTIC COMMUNITY The upper Big Hole River drainage supports populations of arctic grayling (Thymallus areticiis) , brown trout (Salmo trutta) , burbot (Lota lota) , rainbow trout (Salmo gairdneri) , brook trout (Salvelimis fontinalis). westslope cutthroat trout (Salmo c.larki Ip.wisi). Yellowstone cutthroat trout (Salmo clarki honvieri ) . mountain sucker (Catostomus platyrhynrhus) . white sucker (Catostornus conmip.rsom" ) , longnose sucker (Catostomus catostomus) , longnose dace (Rhinichthys catarac tap.) ■ and mottled sculpins (Cottus bairdi) . Tailed frogs (Ascaphus truei) have been found in several tributaries. METHODS HABITAT A total of ten streams were selected to inventory based on recommendations made by Forest Service personnel on the Wise River and Wisdom Districts. LaMarche Creek was selected as a control stream. Prior to the field season these streams were seggregated into relatively homogeneous reaches based on ^channel gradient, valley shape, and area drained using 1:24,000 USGS maps. This resulted in a total of 17 stream reaches in the ten streams. Reaches were numbered consecutively from the mouth upstream. The following data were recorded from these maps: stream order; reach length; channel gradient; acres drained by the reach; acres drained by the entire stream; lower and upper reach boundary landmarks; lower and upper reach boundary legal descriptions; lower and upper elevations of the channel; valley length; channel sinuosity; landtype association; and channel type according to methods described by the Fish Habitat Relationships System. Stream ordering was not done using the contour crenulation method. In addition, descriptive information for land use in the drainage will be obtained from the Beaverhead National Forest's database. This information was not all available at the time of this report, so it will be included next year. Detailed descriptions of each of these variables are provided in Appendix A. Fish habitat was surveyed in 406 to 1184 foot long sample sections of 17 stream reaches where fish population data were collected to correlate fish numbers to quantity and quality of available habitat. These 17 reaches were located in Adson, LaMarche, Meadow, Mono, and Wyman creeks on the Wise River Ranger District, and Elk, Johnson, Joseph, Sheep, and Steel creeks on the Wisdom Ranger District (Table 1). In addition, entire reaches were surveyed in 14 of the above 17 reaches to further quantify available habitat. Entire reach surveys were not done in reaches 1 and 2 of Johnson Creek and reach 1 of Steel Creek • Streambed samples were taken from potential spawning habitats in seven streams including Adson, Big Swamp, Jerry, Joseph, LaMarche, and Trail creeks and Wise River (Table 2). Reach Surveys Reach surveys were conducted by walking the entire length of each reach and tallying the occurence of the following habitat parameters: main habitat types (pools, riffles, runs, and pocketwa ters ) , habitat sub-types ( ie . for riffles - low gradient, rapids, and cascades), amount of spawning gravel (arbitrarly defined as areas larger than four square feet predominated by streambed material in the 0.5 to 3.0 inch category), accumulations of small (less than 6.0 inches in diameter) and large (6.0 inches and larger) organic debris (accumulations had to cover four square feet to be tallied), the number of these debris accumulations which crossed the entire wetted stream channel by size class, and the percentage of these debris considered to be stable (would not normally be moved in an average high flow year). In addition to tallying the above parameters various features within the reach were located including Table 1. Description of stream reaches surveyed during 1986 including stream, reach, landmarks at lower and upper boundaries of each reach, legal descriptions of lower and upper boundaries of each reach, elevations (ft) at lower and upper boundaries of each reach, sample site legal description and length (ft), length of stream channel (mi), length of valley (mi), channel gradient (%), channel type (from Rosgen 1985), stream order, and channel sinuosity. RANGER DISTRICT Stream Reach description Channel Valley length length (mi) (mi) Channel Channel Stream Channel Lower Upper gradient type order sinuosity elevation elevation (%) (ft) (ft) WISE RIVER DISTRICT Adson Ck Re ac h : 1 Lower landmark: MOUTH AT WISE RIVER Lower legal description: T1S R11W SECTI0N19AB Upper landmark: HEADWATERS Upper legal description: T1S R11W SECTION33DB Sample site legal description (length in feet): 3.3 3.1 5.1 T 1SR11WSEC28BC (406) 1.06 5930 6810 LaMarche Ck Re ac h : 1 Lower landmark: MOUTH AT BIG HOLE RIVER Lower legal description: T2N R13W SECTI0N34DD Upper landmark: FOREST SERVICE BOUNDARY Upper legal description: T2N R13W SECTI0N21DA Sample site legal description (length in feet): T 2NR13WSEC22CC (1,018) 2.9 2.5 4.4 B 1.16 5815 6050 Reach: 2 Lower landmark: FOREST SERVICE BOUNDARY Lower legal description: T2N R13W SECTI0N21DA Upper landmark: JUNCTION OF MIDDLE AND WEST FORKS Upper legal description: T2N R13W SECTI0N6DA Sample site legal description (length in feet): T 2NR13WSEC16BB (1,184) 4.5 3.6 0.8 1.26 6050 6235 Table 1. (continued) RANGER DISTRICT S t re am Reach description Channel length (mi.) Valley length (mi.) Channel Channel Stream gradient type order (%) Channel Lower sinuosity elevation (ft.) Upper elevation (ft.) Meadow Ck Reach: 2 Lower landmark: FOREST SERVICE BOUNDARY Lower legal description: TIN R12W SECTION36AD Upper landmark: HEADWATERS Upper legal description: T1S R12W SECTION10CD Sample site legal description (length in feet): T 1NR12WSEC36AC (500) 3.3 3.3 1 1 . 0 1.00 6040 7980 Mono Ck Re ac h : 1 Lower landmark: MOUTH AT JUNCTION WITH JACOBSON CREEK Lower legal description: T3S R12W SECTION33AA Upper landmark: BRIDGE CROSSING OF F.S. ROAD NUMBER 484 Upper legal description: T4S R12W SECTION4BC Sample site legal description (length in feet): T 4SR12WSEC 4BS (496) 1.5 1.4 8.3 1.03 6880 7525 Reach: 2 Lower landmark: BRIDGE CROSSING OF F.S. ROAD NUMBER 484 Lower legal description: T4S R12W SECTION4BC Upper landmark: CULVERT CROSSING OF F.S. ROAD NUMBER 484 Upper legal description: T4S R12W SECTION8AB Sample site legal description (length in feet): T 4SR12WSEC 5DA (700) 1.0 1.0 1.3 1.03 7525 7597 Table 1. (continued) RANGER DISTRICT Stream Reach description Channel Valley length length (mi.) (mi.) Channel Channel Stream Channel Lower Upper gradient type order sinuosity elevation elevation (%) (ft.) (ft.) Wyman Ck Re ac h : 1 Lower landmark: MOUTH AT WISE RIVER Lower legal description: T3S R12W SECTI0N17CA Upper landmark: LOWER END OF LOWER ANDERSON MEADOWS (SM 1.88) Upper legal description: T3S R13W SECTION24CA Sample site legal description (length in feet): T 3SR13WSEC24DA (621) 1.9 1.8 3.8 B 1.06 6715 7100 Re ac h : 2 Lower landmark: LOWER END OF LOWER ANDERSON MEADOWS (SM 1.88) Lower legal description: T3S R13W SECTION24CA Upper landmark: MOUTH OF DEER CREEK Upper legal description: T4S R13W SECTION9AA Sample site legal description (length in feet): T 3SR12WSEC17CC (500) 5.2 4.0 0.9 1.29 7100 7350 WISDOM DISTRICT Elk Ck Re ac h : 1 Lower landmark: MOUTH OF ELK CK AT TRAIL CK Lower legal description: T2S R18W SECTION9BD Upper landmark: 1.52 MILES UPSTREAM FROM MOUTH Upper legal description: T2S R18W SECTION33DA Sample site legal description (length in feet): 1.5 1.5 1.7 B T 2SR18WSEC 4DB (571) 1.02 6420 Reach: 2 Lower landmark: 1.52 MILES UPSTREAM FROM MOUTH Lower legal description: T2S R18W SECTION33DA Upper landmark: HEADWATERS Upper legal description: T1S Rl 8W SECTION21BA Sample site legal description (length in feet): 6560 T 1SR18WSEC33BD (528) 3.4 2.9 2.7 £ 1.15 6560 7060 Table 1. (continued) RANGER DISTRICT Stream Reach description Channel length (mi.) Val ley length (mi.) Channel Channel Stream Channel Lower Upper gradient type order sinuosity elevation elevation C%) (ft.) (ft.) Johnson Ck Reach: 1 Lower landmark: MOUTH AT NORTH FOEK BIG HOLE RIVER Lower legal description: T2S R16W SECTION4BC Upper landmark: 10.3 MILES ABOVE MOUTH Upper legal description: T1S R17W SECTION16AA Sample site legal description (length in feet): T 1SR17WSEC25AA (597) 10.3 6.1 0.7 1.68 6085 6460 Reach: 2 Lower landmark: 10.3 MILES ABOVE THE MOUTH Lower legal description: T1S R17W SECTION16AA Upper landmark: HEADWATERS Upper legal description: TIN R17W SECTION32AA Sample site legal description (length in feet): 3.9 2.8 4.4 T 1SR17WSEC 5CD (534) 1.35 6460 7360 Joseph Ck Re ac h : 1 Lower landmark: MOUTH AT JUNCTION WITH TRAIL CREEK Lower legal description: T2S R18W SECTION15BD Upper landmark: MOUTH OF RICHARDSON CREEK Upper legal description: T2S R18W SECTION7BC Sample site legal description (length in feet): T 2SR18WSEC16BC (575) 3.7 3.4 0.8 Reach: 2 Lower landmark: JUNCTION OF RICHARDSON CREEK Lower legal description: T2S R18W SECTION7BC Upper landmark: HEADWATERS Upper legal description: T2S Rl 9W SECTION2CA Sample site legal description (length in feet): 2.2 1.9 2.9 B 1.08 6380 6540 T 2SR19WSEC12BC (592) 1.15 6540 6880 -uM i.v.- « v; Table 1. (continued) RANGER DISTRICT S t re am Reach description Channel Valley length length (mi.) (mi.) \ Channel Channel Stream Channel Lower Upper gradient type order sinuosity elevation elevation (%) (ft.) (ft.) Sheep Ck Re ac h : 1 Lower landmark: MOUTH AT TRAIL CREEK Lower legal description: T2S R18W SECTI0N14DC Upper landmark: 1.08 MILES UPSTREAM FROM MOUTH Upper legal description: T2S R18W SECTI0N11CD Sample site legal description (length in feet): 1.1 1.0 1.7 B Reach: 2 Lower landmark: 1.08 MILES ABOVE MOUTH Lower legal description: T2S R18W SECTI0N11CD Upper landmark: HEADWATERS Upper legal description: T1S R18W SECTION35 Sample site legal description (length in feet): 2.8 2.6 4.0 B T 2SR18WSEC14BD (555) 1.09 6340 6440 T 2SR18WSEC11BB (534) 1.11 6440 7040 Steel Ck Re ac h : 1 Lower landmark: MOUTH AT BIG HOLE RIVER Lower legal description: T2S R15W SECTION15BB Upper landmark: MOUTH OF FRANCIS Upper legal description: T2S CREEK R15W SECTION3CA Sample site legal description (length in feet): T 2SR15WSEC34A (600) 5.6 4.8 0.6 1.16 5990 6175 10 Table 2. Location (stream and legal description) and date sampled for hollow core sampling conducted in waters draining the Beaverhead National Forest during 1985-86. Year S t re am Date Legal Description 1985 'Doolittle Ck East Fork Ruby River Harriett Lou Ck Meadow Ck Mill Ck S Fk Blacktail Ck S Fk Willow Ck Trail Ck 11/04/85 10/25/85 11/05/85 10/31/85 10/23/85 10/24/85 10/30/85 11/07/85 v. T IS R 14W SEC 28C T US R 3W SEC 5B T IN R 12W SEC 36D T IN R 12W SEC 36A T 2S R 4W SEC 23A T 12S R 5W SEC 30C T 3N R 3W SEC 13A T 2S R 18W SEC 13C 1986 Ad son Ck Big Swamp Ck Jerry Ck Joseph Ck LaMarche Ck Trail Ck Wise River 11/13/86 T IS R 11W SEC 20CA 10/2 9/86 T 5S R 16W SEC 16AA 11/17/86 T IN R 11W SEC 36CD 10/22/86 T 2S R 18W SEC 16BA 10/23/86 T 2N R 13W SEC 5CD 10/20/86 T 2S R 18W SEC 15B 11/07/86 T 3S R 12W SEC 21CA 11 ,» ;.**•** any barriers to upstream fish movement, unique features ( ie . eroding banks, livestock damage, bridges, etc.), irrigation withdrawals or returns, side channels, and areas of abundant high quality spawning habitat. All these features were located by pace. Side channels were further quantified by pacing from the point the side channel left the main channel to the point where the side channel returned to the main channel (this distance was paced along the main channel). For a more detailed description of these parameters and how they were measured consult Appendix B. Reach surveys were conducted in Reach 1 (Rl) of Adson, Rl and R2 of LaMarche, R2 of Meadow, Rl and R2 of Mono, and Rl and R2 of Wyman creeks on the Wise River District, and Rl and R2 of Elk, Rl and R2 of Joseph, and Rl and R2 of Sheep creeks on the Wisdom District. Due to extremely difficult access (downfall timber) or time constraints the entire length of R2 Sheep Creek, R2 Wyman Creek, R2 Meadow Creek, and R2 Joseph Creek were not surveyed. Results of these surveys are reported as the frequency of occurence (number per mile) for debris accumulations, square feet per mile for spawning gravel, and percentage composition for both main and sub- habitat types. In addition, the locations of pertinent features were noted. Detailed... Pabitat Surveys of Sample Sections Detailed habitat surveys of 406 to 1184 foot long segments in each of the 17 reaches were conducted using techniques similar to Fish Habitat Relationships System methodology with the following exceptions: site selection was based on stream reach classification; secondary channel pools were not separated in pool classification (these would be included as side channels); channel gradient was calculated from USGS maps (scale 1:24,000), not in the field; the size classification for stream substrate materials was done using a modified Wentworth scale (Shepard 1986); canopy closure was not estimated; and canopy density was visually estimated as the percentage of the stream's surface overhung by canopy (tree) boughs. Each study section was broken down into habitat types and classed into both main and sub-habitat types. Within each habitat type the following data were collected for the entire habitat type: length of type, length of undercut for both banks, canopy density (visually determined as the amount of overstory which actually overhung the wetted surface), instream cover (which was the total percentage of the water's surface area which had instream cover in the form of actual structure, ie . substrate or debris, water depth or disturbance, etc. - anything which prevented fish from being observed from above the water's surface was considered cover), low (one foot or less above the water's surface) and high (higher than one foot above the water's surface) overhead cover (percent of the water's surface covered), substrate composition (silt, sand, small gravel, large gravel, cobble, small boulder, large boulder), soil alteration rating (Platts et al. 1983), vegetation stability rating (Platts et al. 1983), and vegetation use by animals (Platts et al. 1983). The following data were collected across at least one cross section per habitat type (cross sections were generally done at a frequency of one every ten to twenty feet of stream length): wetted width, channel width, average depth, thalweg depth (the deepest portion of each cross section), water depth 12 at each shoreline averaged for Che cross section, embeddednes classified visually, substrate score (modified from Crouse et al. 1981), D-90 (the diameter of a streambed particle which is larger than 90 percent of the remaining particles), depth of undercut bank (a horizontal measurement averaged for each cross section), and bank angle (also averaged for each cross section). Streamflow was measured at one uniform cross-section within each reach at the time of the above survey. Detailed descriptions of each of these variables and how they were measured can be found in Appendix C. The above data were summarized by main habitat type and the averages and percentage composition are presented by habitat type and for the reach as a whole. Means and standard deviations were calculated for all variables. The amount of undercut bank was converted to a percentage using the formula: Total Length of Undercut (ft) Percentage of Undercut ■ . X 100 (Total Length of Sample Section) X 2 Streambed Sampling In 16 of the 17 sampled reaches embeddedness measurements were made following methodology described by Burns (1984). Reach 2 of Mono Creek was not sampled because its streambed was composed primarily of sand and was considered to be fully embedded. In Rl Steel Creek, Rl Adson Creek, and R2 LaMarche Creek two separate areas were sampled in an attempt to begin to assess the variability of embeddedness sampled within a reach. Ten "hollow core" (McNeil and Ahnell 1964) samples were taken using the same methodology as in 1985 (Shepard 1986) in Adson, LaMarche and Jerry creeks and upper Wise Piver on the Wise River District and in upper Trail, Joseph and Big Swamp creeks on the Wisdom District. Embeddedness measurements (Burns 1984) were also made in each area where core samples were taken. The core sampling was summarized by site and is presented along with estimated egg-to-fry survival values for cutthroat and rainbow trout from Irving and Bjornn (1984). I derived survival curves for brook trout using data from Witzel and McCrimmon (1983). Embeddedness values were summarized by site. FISH ABUNDANCE ESTIMATES Fish abundance was assessed using relative catch per unit effort (CPUE) and by making population estimates. Lengths of the sections censused ranged from 300 to 1184 feet. A Coffelt backpack elec trof isher Model BP-1C was used in all sections. A total of 28 sections in 20 streams were electrof ished . Abbreviations used for fish species throughout this report are: GR ■ arctic grayling; EBT = eastern brook trout (charr); RBXWCT = hybrid between wests lope cutthroat and rainbow trout or unidentified Sal mo spp. (cutthroat trout, rainbow trout, or hybrids between the two); LING = burbot; MWF = mountain whitefish; RB ■ rainbow trout; WCT = westslope cutthroat trout. 13 Catch per Unit Effort (CPUE) Electrof ishing was conducted in section(s) of Adson, Butler, Fishtrap, LaMarche, Meadow, Mono, Swamp, and Wyman creeks on the Wise River District and Bender, Big Lake, Elk, Joseph, Johnson, Mussigbrod, Placer, Plimpton, Sheep, and Steel creeks and Salefsky and Goris gulches on the Wisdom District. The relative abundance of each species of fish for each pf the above sampling sections was expressed as the number of fish for all fish 3.0 inches and longer captured in one electrof ishing pass standardized to a 1,000 foot section of stream. Tributary. Population Estimates Population estimates were made in 17 stream reaches using either depletion or mark -recapture estimators. Each sample section was electrof ished from its upstream boundary downstream to its lowermost boundary. A block net was used at the downstream boundary of the section if there was no reasonable fish blocking feature naturally present in the stream channel. All captured fish were processed after the first pass to allow the section at least one to two hours of "rest" between electrof ishings . The fish captured on the first pass were held in a livecar while the section was electrof ished again in a downstream direction. All fish were marked with a fin clip. If the estimated probability of capture (p) calculated using the formula: a 1 2 P n2 (where n and a, - number of fish captured in the first and second electrof ishing passes, respectively) was higher than 0.60 it was assumed that a reasonable population estimate could be obtained using a depletion estimator (personal communication, 1984, Tom Berggren, Bonneville Power Administration, Portland, Oregon). If the ^ value was less than 0.60, a recapture electrof ishing was conducted from two to seven days later. The only exception to this general rule was in R2 of LaMarche Creek where the subsequent "recapture" was done using underwater observation by a diver snorkeling the stream in a mask and wet suit. Unfortunately, it was difficult for the diver to identify all fish with a clipped fin which designated marked fish (in this case an upper caudal clip). Populations were estimated using the maximum liklihood technique in the MICROFISH software package (VanDeventer and Platts 1985) on a Zeinth AT microcomputer for depletion electrof ishings and/or using Chapman's (1951) mark-recapture formula (cited in Ricker 1975) within a PRESENT query on the BNF's Data General computer system. For the estimate in Old Tim Creek (Dillon District) an equipment malfunction during the second electrof ishing caused and incomplete second pass. An estimate pf the total number of fish captured on the second pass was made (by dividing the actual number caught by the percentage of the sections sampled expressed as a decimal) and using this value within a two-pass estimator (Seber and LeCren 1967). Population estimates were made for fish in the 3-0 to 5.9 inch length class and for fish 6.0 inches and longer. No attempt was made to estimate the numbers of fish under 3.0 inches, 14 m however, the number of these small fish captured were used in constructing length frequency histograms. gig Hole River Population Estimate. A 4.98 mile section of the Big Hole River above Vt he Highway #43 bridge westof Wisdom was elec trof ished in the early summer to estimate the number of arctic grayling and brook trout inhabiting this section of river. Four marking runs were conducted between June 23 - 26 and three recapture runs were conducted between June 30 and July 2. In addition to length and weight data, the presence of obvious hooking scars were noted for all handled fish. The percentage of hook scared fish was estimated by species. This estimated percentage is likely an underestimate due to fact it is likely that some previously hooked fish may not have an obvious scar. The MDFWP mark-recapture computer program (which uses equations described by Vincent 1971 and 1974) was used to estimate the numbers of brook trout, rainbow trout, and arctic grayling in this section of river. LENGTH, WEIGHT, AND CONDITION FACTOR Length in inches to the nearest 0.1 inch was measured from all captured fish. Length frequency histograms were constructed by species for each stream reach where at least 25 fish of the same species were captured. Weight in pounds to the nearest 0.01 pound was obtained using a spring scale (weight range 0.00 to 5.00 pounds) for each fish captured, however, the scale became unreliable near the end of the field season and all unreliable weights were discarded. Condition factors were estimated for all salmonids using the formula (Anderson and Gutreuter 1983): 3 Condition factor = (Weight/Length ) X 10,000. FISH MOVEMENT Tributary Trapping During the Spring Up and downstream box traps were installed in Big Lake, Steel, and Swamp creeks to monitor the movement of fish between the Big Hole River and these tributaries during the spring. Arctic grayling was the primary target species for this trapping effort. The upstream traps were constructed using a frame of 0.5 inch rebar covered with 0.5 inch mesh hardware cloth. A conical fyke was constucted at the downstream end to allow fish to move into the trap and prevent them from moving back out. The downstream traps were constructed using a wooden frame (2 by 4 inch stock) and plywood sides covered with 0.5 inch mesh hardware cloth at the upstream and downstream ends. A V-shaped entrance was constructed at the upstream end to allow fish moving downstream to enter the traps. These traps were placed on either side of the stream with a diagonal fence connecting them. This fence was constucted of steel fence posts supporting four foot high 0.5 inch mesh hardware cloth with a 0.5 foot portion of this hardware cloth buried into the streambed (Figure 2). 15 UPSTREAM TRAP Downstream V5®i DOWNSTREAM TRAP (showing fyi )]; TRAP ORIENTATION IN STREAM (from abovs) ownstraam Trap Figure 2. Schematic diagram of upstream and downstream fish traps and orientation of traps in the stream for trapping of migrating fish in the spring of 1986 as part of the cooperative fisheries study between the BNF and MDFWP. 16 The upstream and downstream traps were installed on April 7, 1986 in Steel Creek. Upstream traps were installed on April 7 in Big Lake Creek and on April 23 in Swamp Creek (Figure 3). The downstream traps were added on May 7 to the Big Lake and Swamp creek sites. These traps were checked at intervals ranging from every eight hours to every four days dependent upon flow conditions and the number of fish moving through the traps. The traps were difficult to maintain due to fluctuations in streamflow and the presence of drifting algae and debris. The Big Lake Creek trap site was moved upstream on May 22 because reduced flows from irrigation withdrawals dewatered the downstream trap. A mink predation problem was evident at the Swamp Creek trap site (partially eaten fish were found in the traps and on the bank adjacent to the traps) and may have occured at the other sites. During the major snowmelt event (from May 28 to June 19) the Big Lake and Steel creek traps were removed and the fence between the Swamp Creek traps was left down. The Swamp Creek trap was reinstalled on June 20 and the trap was operated until June 25 when it was removed . Drift nets (1.0 by 1.5 foot rectangle openings with 80 openings per inch mesh nets, Wildco Supply number 158) were placed at three locations in Sandhollow Creek on the Wisdom District on May 20, 1986. These nets were checked twice daily until their removal on June 2. Each time the traps and drift nets were checked all gamefish were measured to the nearest 0.1 inch and weighed to the nearest 0.01 pound, water temperature was measured, and condition of the traps, leads, and general observation of streamflow noted. Rainbow Trout Redd Counts Redd (trout spawning site) counts were conducted in Jerry and Bryant creeksdraining the Wise River District to document the relative use of these two tributaries as spawning areas by rainbow trout. A portion of Jerry Creek from Forest Service Road #83 down to the mouth of Jerry Creek at the Big Hole River and a portion of Bryant Creek from immediately above the Forest Service boundary down to its mouth at the Big Hole River were surveyed on May 8, 1986. All observed disturbances in the streambed were classified into one of the following classes; sure redd, probable redd, and possible redd, based on criteria established by Shepard et al . (1982). Identified redds in Jerry Creek were further seggregated based on size in an attempt to quantify the number of redds constructed by fluvial Big Hole River adults. Arctic Grayling Rad io telemetry The movement of grayling within the Big Hole River system was evaluated using radio tags. Radio tags were implanted during the early summer (July 1 and 2) and during the early fall (September 22). The early summer work was done by Gould (1986). The fall work was conducted as part of this study. On September 22, 1986 seven radio transmitters were surgically implanted into grayling which had been captured via electrof ishing. The fish were captured and released in a segment of the river immediately above the town of Wisdom (between river miles 116 and 119). The radios weighed approximately 0.2 17 mile Figure 3. Location of fish trapping sites in the upper Big Hole River drainage in the vicinity of Wisdom, Montana which were operated during the spring of 1986. Sites in Big Lake, Steel, and Swamp creeks had up- and downstream traps ( M ) , while the site in Sandhollow Creek had drift nets ( o ) . 18 ounces (5 g) and were approximately 1.0 X 0.6 X 0.9 inches (26 X 16 X 24 ram) in size. These radios represented approximately 1.6 to 2.8 percent of the total body weight of the fish in which they were implanted. Hop (1985) found that grayling in Alaska were not impaired when radios weighed between 1.7 to 3.3 percent of total body weight. These radios had a rated life expectancy of 90 days. The 30 and 40 mHz bands were the receivers available from the Montana Cooperative Fisheries Research Unit, however, due to, the moderately low water conductivity found in the Big Hole River (80-90 microsiemens: Levings 1986) higher bands may provide better results (Don Stoneburner, Custom Telemetry Consulting, Athens, Georgia; the source for the radios). The surgical technique consisted of anesthesizing the fish in an approximately 1% solution of 2-phenoxy-ethanol . Fish were placed on an inclined V-shaped platform with their heads and gills within the anesthetic. A 0.5 to 1.0 inch (12.7 to 25.4 mm) incision was made between the pelvic fins and a transmitter which had been sterilized in an alcohol solution was inserted into the body cavity. The incision was sutured closed using Chromic 4-0 collagen gut suture material. The incision normally required three to four sutures to close. Fish were relocated by floating the river or driving or walking the river bank at one week intervals. Relocation searches were done on sixteen separate days from September 23 to November 26. The search areas varied, but coverage included from river mile 55 up to river mile 121 and the lower portions of Swamp and Steel creeks (Table 3). During an extreme cold period from November 14 to November 19 many sections of the river froze completely over, especially in the Wisdom area. Most of the areas which froze over near Wise River opened up after a thaw around November 20 while much of the river in the Wisdom area remained frozen. Plastic Coded Tagging All arctic grayling and rainbow trout longer than 3.0 inches captured during the season were tagged with either a numbered juvenile dangler type (fish between 3.0 and 7.9 inches) or a numbered anchor type (fish 8.0 inches and longer) tag. Tag recoveries were made during the course of sampling and from anglers. These data are summarized by tag for all returns. STATISTICAL ANAYLSES Tests for Normality The data were summarized using mean and standard deviations. Tests for normality were conducted for the habitat data collected within the fish abundance sample sections of Rl LaMarche Creek, Rl Adson Creek, R2 Joseph Creek, and Rl Sheep Creek using the micro-computer version of SAS's UNIVARIATE procedure (SAS 1985). This procedure uses the "W" statistic (Shapiro and Wilk 1965) to test for normality. The data was not normally distributed for most habitat variables. Nonparame trie statistical procedures were utilized to overcome the problems associated with non-normal data analyses. Principal components anaylsis (PCA) was used to transform these intercorrelated habitat values "to allocate the greatest possible variation to 19 Table 3. Areas searched for radio tagged arctic grayling in the upper Big Hole River drainage during the fall of 1986 including date of search, area searched, how search was conducted, water temperature (F, n.d. signifies no data), and general comments. Water Date temperature Area searched Me thod General comments 9-22-86 48 9-23-86 n.d. 9-23-86 n.d. 9-26-86 44 10-3-86 41 Tags implanted From RM 119.9 down to Float Highway 43 bridge (RM 116.0) Below bridge approximately 0.5 miles On foot From RM 116.5 down to RM On foot 116.0 (Highway 43 bridge) From RM 116.0 down to RM Float 111.0 (below where Steel Creek enters east channel) From Highway 43 bridge down Float to below the mouth of Swamp Creek (RM 116.0 to RM 108.0) Recap electro- fishing run Floated the west channel Floated the east channel 10-8-86 n.d. 10-14-86 40 10-21-86 39 10-28-86 38 11-3-86 35 Float From above the Highway 43 bridge (KM 116.8) down to bridge (RM 116.0) From head of McDowell's irrigation diversion (RM 117.5) down to RM 111.0 (below where Steel Creek enters the east channel) From RM 120.9 (Rutledge Rd Float culvert) down to below the cemetary (RM 110.0) On foot Only had receiver to monitor 3 rad ios on 30mHz Floated the east channel Floated the east channel McDowell's diversion and On foot Spot checked area above Highway 43 bridge Floated from Highway 43 bridge (RM 116.0) down to below the cemetary (RM 110.0) Float From access below cemetary Float (RM 110.0) down to Daniel's ranch (Crane ranch at EM 104.2) Floated the west channel Few channel splits 20 Table 3. continued. Water Date temperature 11-14-86 n.d. 11-17-86 n.d, Area searched Method From Highway 43 bridge near Drove Wisdom (KM 116.0) to Highway 43 bridge near Divide (EM 55.7) From Jerry Creek down to Highway 43 bridge near Divide (KM 55-7) Drove General comments Only had receiver to monitor 3 radios on 30 mHz Lots of ice Only had receiver to monitor 3 radios on 30 mHz Lots of ice 11-19-86 n.d. From Wisdom to Divide (KM 116.0 to EM 55.0) Drove Covered highway and dirt roads on both sides of river 11-20-86 11-21-86 n.d , n.d , From McDowell's diversion (KM 117.5) down to below cemetary (KM 110.0) The lower 0.2 miles of Steel Creek On foot Searched the east channel. Still lots of ice From Daniel's ranch (Crane On foot Much of the river ranch at KM 104.2) down to iced over below Wallace Chris tensen's house (KM 89.8) 11-24-86 11-25-86 11-26-86 n.d n.d . n.d. From below Chris tensen 's On foot Much of the river house (KM 89.9) down to iced over Highway 43 bridge (KM 91.6) From Jerry Creek down to Big Hole dam (KM 62.8 down to KM 57.8) Swamp Creek from Norths ide Big„Hole River Road bridge (CM"' 1.8) dowi mouth (CM 0.0) Float River mostly free of ice in this section On foot Creek mostly iced over above spring (CM 1.5) and open below the spring 1/ 2/ EM indicates river mile from "Eiver Mile Index of the Missouri River" Water Eesources Division, Montana Department of Natural Resources and Conservartion, January 1979. CM indicates creek mile as measured from a USGS quad (scale: 1:24,000). 21 the fewest possible new uncorrelated variables" (Green 1979). These computed "variables" are actually linear additive functions which retain all the information in the "old" original data set for use in subsequent regression analyses against fish density information. Comparisons Between Methodologies Habitat Data The Wilcoxon sign-ranked test (Daniel 1978) was used to compare the percentages of each of the two main habitat types (pools and riffles) estimated from the survey of the entire reach and the detailed survey of the fish abundance section in the 14 reaches where both surveys were conducted. This test was also used to compare the embeddedness values obtained from sampling using the Burns (1984) sample technique versus ocular estimation in the 17 reaches where both were done. Fish Pat?. Estimates of fish populations derived from the two electrof ishing estimators (depletion and mark-recapture) and from the underwater count and deplection estimator in R2 LaMarche Creek were compared. The effect of radio implants on the condition of arctic grayling was assessed by comparing September condition factors of arctic grayling which had radios implanted in early July to those that did not using a two-sample t-test (Zar 1984). Correlations Between Habitat Varlabl.es and Fish Densities Fish population estimates were converted to fish densities by calculating the number of fish per surface acre in each sample section. Spearman's rank correlations were done between each habitat variable value and the corresponding fish density value. PCA functions (see above) derived from habitat data were regressed against the corresponding fish densities (SAS 1985). These initial attempts at regressing these PCA functions against fish abundance data by species were inconclusive due to the small sample sizes. Test of the COWFISH Model Fish population estimates were conducted in sample sections of Browns Canyon, Cow Cabin, Morrison, Pass, and Painter creeks within the Dillon District where Range personnel had completed COWFISH habitat sampling (Lloyd 1986). In addition, the reaches surveyed in Rl Elk, Rl Joseph, R2 LaMarche, R2 Mono, Rl Sheep, Rl Steel, and R2 Wyman creeks during 1986 and R2 Governor, R2 Ruby, and R2 Steel creeks during 1985 all were within areas that had various levels of livestock grazing. 22 Data needed for the COWFISH habitat evaluation could be derived from habitat surveys directly except for the percent of streambank with overhanging vegetation. COWFISH estimates the linear percentage of streambank which has vegetation overhanging the water's surface, while the habitat surveys done during this study estimated the percentage of the water's surface covered by overhanging vegetation. I converted the percentage of water surface covered to the percentage of streambank with overhanging vegetation by adding the percentage of low and high coverage of the water's surface times the stream's average width divided by 8. This conversion assumes that much of the cover overhung the stream by 4.0 feet and usually resulted in an increase when converting from percentage coverage to percentage of streambank with cover. Regressions were made between the estimated number of fish longer than 6.0 inches ("catchable") from electrof ishing samples and both the predicted "optimum" and predicted "existing" numbers from the COWFISH model as well as between the electrof ishing estimates and the total "parameter suitability index" for each stream using the "STATGRAPHICS" micro-computer statistcal software package. These regressions were run separately for stream sections which supported cutthroat trout and those which supported brook trout. 23 RESULTS Fish and habitat information is presented in the report "Beaverhead National Forest Fisheries - Streams Surveyed During 1985-86" by stream with a brief discussion of habitat condition. This report will be used to contrast the condition of the aquatic resources between stream reaches which were surveyed during 1986 and to analyze relationships between habitat variables and between habitat variables and fish densities. \ HABITAT Reach Character-; stirft Channel gradient, channel type, stream order, channel sinuosity, and lower and upper elevation of the stream channel within each stream reach were derived from contour maps. The results of surveys conducted throughout the "entire reach" describe the habitat composition, frequency of side channels, frequency of large and small woody debris, amount of available spawning habitat, and locations of potential fish passage barriers. Map Derived Information Information interpreted from maps is presented in Table 1. "A-type" channel reaches (Rl Adson, R2 Johnson, R2 Meadow, and Rl Mono creeks) are typified by relatively high channel gradient, narrow valley bottoms, and low channel sinuosity. "C-type" or typical "meadow" channel reaches (Rl Johnson, Rl Joseph, R2 LaMarche, R2 Mono, Rl Steel, and R2 Wyman creeks) are characterized by relatively low channel gradients, wide valley bottoms, and relatively high channel sinuosity. "B-type" channel reaches have channel and valley characteristics between "A" and "C" type channels. Habitat Composition Pool habitats dominated R2 Mono Creek and R2 Wyman Creek (Table 4). Pool habitats were noticeably sparse in Rl Wyman Creek and moderately low in R2 Elk, Rl LaMarche, R2 Meadow, Rl Mono, and Rl and R2 Sheep creeks. Riffle habitats were especially abundant in Rl Wyman, R2 Meadow, Rl Mono, and Rl Sheep creeks and scarce in R2 Mono Creek. Pocketwaters made up a moderately large percentage of the habitat in R2 Elk, Rl LaMarche, R2 Meadow, Rl Mono, R2 Sheep, and Rl Wyman creeks which reflects the relatively higher gradient in these reaches. Side channels were found along much of Rl and R2 of Joseph Creek. Moderate side channel development was found along Rl Elk, R2 Meadow, R2 Mono, and Rl Wyman creeks. Side channel development in Joseph, Rl Elk, and R2 Mono was caused by both beaver activity and livestock impacts. Side channel development in the other reaches listed above was primarily due to higher channel gradient associated with large debris which formed side channels during high streamflows. 24 Table 4. Percentage of each main habitat type within each reach of streams draining the Beaverhead National Forest surveyed during 1986. RANGER DISTRICT Pocket Side Stream Reach Pools Riffles Runs waters channels WISE RIVER DISTRICT Ad son Ck I 30 39 20 11 4 LaMarche Ck 1 23 32 26 19 5 2 37 33 25 5 3 Meadow Ck 2 22 48 10 20 12 Mono Ck 1 2 2? 53 45 13 7 31 22 3 3 17 Wyman Ck 1 2 9 43 56 24 11 28 24 5 16 1 WISDOM DISTRICT Elk Ck 1 2 35 26 29 37 25 18 10 19 10 3 Johnson Ck 1 2 17 19 38 29 21 19 25 24 10 Joseph Ck 1 2 36 34 36 35 23 21 6 10 27 35 Sheep Ck I 2 23 24 44 34 30 14 2 27 1 6 Steel Ck ll 21 42 37 1/ Habitat compostion in reaches within these streams were based on habitat composition within the detailed sample section. See "Habitat Composition" section of "RESULTS" for a discussion of use of these data. 25 A comparison between habitat composition estimated within the sample sections and habitat composition for the reach as a whole was made by comparing the percentages of pools and riffles estimated within the sample section to counts made throughout the entire reach (Table 5). There was no significant difference between estimates for pools or riffles (P> 0.10). There were cases where relatively large differences between survey techniques were observed (ie. pools in Rl and R2 Joseph, R2 Meadow, and Rl Sheep; and riffles in Rl and R2 Wyman and Rl and R2 Elk). However, there was no consistent bias observed because in some cases higher percentages were estimated in the sample section and in other cases a higher percentage was found in the "entire reach" survey. This could present a problem in any attempt to expand fish population estimates derived using habitat composition data obtain from sample sections to the entire reach and suggests that any reasonable estimate of habitat composition should be based on a sample larger than a 300 to 1,000 foot sample section. Pools were formed primarily by water plunging over debris and/or large streambed material and by lateral scouring of the stream's bank and bed at bends in the channel (Table 6) . Lateral scouring was the predominant pool forming mechanism in lower gradient channels, while plunge pools predominated higher gradient reaches. Beaver dams were responsible for forming many pools in Elk and Joseph creeks. There were numerous high quality pools in LaMarche and Joseph creeks, while low quality pools dominated in Adson, Elk, Meadow, Rl Mono, R2 Sheep, and Wyman creeks. Low gradient riffle types dominated riffles within low gradient reaches, while cascade riffle types dominated in high gradient reaches (Table 7). Frequency of Debris and Spawning Habitat, The frequency of both large and small size classes of debris was relatively high in R2 Elk, R2 Meadow, and R2 Sheep creeks (Table 8). Debris frequencies were relatively low in Rl Joseph, Rl and R2 LaMarche, Rl and R2 Mono, Rl Sheep, and R2 Wyman creeks. More debris crossed narrow channels versus larger channels and large debris was more frequently observed across stream channels than small debris. Both these findings were expected and logical. Spawning habitat was extremely plentiful in R2 LaMarche Creek and probably adequate in all other stream reaches with the exceptions of Rl Mono and Rl Wyman creeks. It is likely that mature fish in Rl Wyman Creek and the upper 0.6 mile of Rl Mono Creek move upstream to spawn. Rl Mono Creek has a total barrier to upstream fish movement located near stream mile 0.4 which suggests that spawning habitat below this barrier may be limited. Detailed Habitat Survey within Sample Sections Detailed habitat surveys within sample sections documented the physical character of the stream channel, the amount and types of cover available to fish, and occular estimates of streambed composition and condition. Physical Character of Stream Channel Table 5. A comparison between the percentage of stream habitat in pools and riffles estimated by surveying the entire reach versus a 400 to 1200 foot sample section and the results of a Wilcoxon sign-ranked test (Daniel 1978) testing between the different surveys. Stream Percentage of pools Percentage of riffles Reach Section Reach D. Rank Section Reach D. Rank i i Adson Ck 1 29 30 - 1 - 2.5 31 39-3 -10.0 Elk Ck 1 44 35 9 8.0 40 2 37 26 11 9.0 26 Joseph Ck 1 54 36 IS 14.0 32 2 42 34 12 10.0 29 LaMarche Ck 1 24 23 1 2.5 38 2 50 37 13 11.0 31 Meadow Ck 2 7 22 -15 -12.0 55 48 7 8.5 Mono Ck 1 2 Sheep Ck 1 2 Wyman Ck 1 2 24 27 - 3 - 5.0 41 54 53 1 2.5 20 39 23 16 13.0 39 20 24 - 4 - 6.0 37 O 9 - 1 - 2.5 42 33 43 ranks - 5 T + - 7.0 38 gative = 70.0 T- - 35.0 29 11 11.5 37 -11 -11.5 36 - 4 - 3.5 35 - 6 - 6.5 32 6 6.5 33 - 2 - 1.0 45 - 4 - 3.5 13 7 8.5 44 - 5 - 5.0 34 3 2.0 56 -14 -13.5 24 14 T+ 13.5 = 50.5 T- = 54.5 Ho: The median of the population of differences is zero ( ie . There is no difference between the two) Ha: The nedian is not zero Pools: T- - 35.0 Riffles: ,T+ = 50.5 P > 0.135 P > 0.445 Therefore, conclude that there is no significant difference between the methods Table 6. Percentage of each type and class of pools (class V is the highest quality pool) within each reach of streams draining the Beaverhead National Forest surveyed during 1936. RANGER DISTRICT Types of pools Classes of pools Lateral Stream Reach Plunge Dammed Beaver Trench scour V IV III WISE RIVER DISTRICT Ad son Ck LaMarche Ck Meadow Ck 1 2 56 49 25 69 .1 1 22 0 .3 32 46 62 1 25 74 2 9 65 27 44 6 15 82 'ono Ck 1 2 93 12 2 8 0 0 0 4 0 76 9 16 40 54 51 30 Wyman Ck 1 2 89 16 6 2 0 2 4 8 0 72 15 10 28 27 57 63 WISDOM WJh TRICT Elk Ck 1 2 21 45 10 10 10 3 2 36 34 9 7 2 0 29 71 65 Joseph Ck 1 2 24 34 5 5 15 22 3 3 52 •"1 f JO 39 55 38 25 24 20 Sheep Ck 1 2 1 0 68 5 10 3 0 13 0 5 5 2 2 24 1 39 r 0 37 93 Table 7. Percentages of each type of riffle habitat for stream reaches draining the Beaverhead National Forest surveyd during 1986. RANGER DISTRICT Stream Reach Low Gradient Rapid Cascade WISE RIVER DISTRICT Ad son Ck LaMarche Ck Meadow Ck Mono Ck Wynian Ck WISDOM DISTRICT Elk Ck Joseph Ck Sheep Ck 13 48 39 1 0 82 18 2 69 27 3 42 56 1 0 7 93 2 61 17 23 1 6 50 44 2 80 20 1 1 62 36 2 2 39 51 10 1 70 29 0 2 53 38 9 1 41 53 6 2 27 51 22 Table 8. Frequency (number per mile) of large (six inches in diameter or larger) and small (less than six inches in diameter) debris, frequency which these large and small debris cross the wetted channel, and amount of spawning gravel observed (square feet of gravel per mile) by reach in streams draining the Beaverhead National Forest surveyed during 1936. Large debris Small debris RANGER DISTRICT Stream Reach Total (#/mi) Cross channel (#/mi) Total (#/mi) Cross channel (#/mi) Spawning gravel (sq. ft/mi) WISE RIVER DISTRICT Ad son Ck 1 153.2 109.1 241.7 124.1 481 LaMarche Ck 1 30.6 3.3 23.0 0.0 1801 2 27.0 2.6 46.7 0.0 24237 Meadow Ck 2 380.2 286.6 335.8 83.3 350 Mono Ck I 49.6 36.1 52.9 0.0 21 2 0.0 0.0 7.7 0.0 261 Wyman Ck 1 40.2 11.1 19.2 0.0 23 2 2.8 1.1 3.9 0.0 223 WISDOM DISTRICT Elk Ck 1 145.1 62.8 110.5 0.0 333 2 243.8 95.5 108.0 7.0 455 Joseph Ck 1 39.0 6.7 82.7 0.0 230 2 101.7 19.4 106.5 0.6 2 58 Sheep Ck 1 38.1 9.5 26.2 0.0 508 1 726.1 455.2 316.2 66.3 239 30 The physical characteristics of the stream channel in reaches surveyed during 1986 is presented in Table 9. A summary of physical characteristics stratified by main habitat type within each sample section is presented in Appendix D. Cover Mean estimates of cover parameters are presented in Table 10 and also segregated by main habitat type in Appendix E. Reaches which had relatively high percentages of undercut banks (greater than 50%) were Rl Adson, R2 Elk, R2 Joseph, R2 LaMarche, R2 Meadow, R2 Mono, and Rl and R2 Sheep creeks. The previous reaches with high percentages of undercut banks which also had relatively deep undercut banks (6.0 inches or deeper measured horizontally) were R2 Elk, R2 Joseph, R2 LaMarche, and Rl and R2 of Sheep creeks. R2 Mono Creek was obviously being impacted by livestock along its streambanks. R2 Meadow Creek was a high gradient stream with moderate livestock use occuring within the sample section. Stream reaches with a relatively low percentage of their streambanks undercut (less than 30 percent) were Rl LaMarche, Rl Mono, Rl Steel, and Rl Wyman creeks. The apparent cause of the low amount of undercut banks in Rl Steel Creek was livestock damage of the streambank within the private landholdings within the sample section. The relatively low percentage of undercutting in the other reaches was probably related to high peak streamflows coupled with the relatively narrow valley bottoms and boulder and cobble material along the streambanks. Livestock impacts to streambanks were observed in Rl Elk, Rl Johnson, Rl Joseph, R2 Mono, Rl Sheep, Rl Steel, and R2 Wyman creeks (Table 11). Instream cover was abundant in Rl Mono Creek and was provided primarily by large streambed material (Table 10). Instream cover was noteably low in Rl Elk, Rl and R2 Joseph, Rl Sheep, and Rl Steel creeks. All of these reaches had moderate to high levels of livestock use with the possible exception of R2 Joseph Creek. Instream cover in the remaining reaches which were surveyed was considered moderate to high and consisted of instream debris, aquatic vegetation, streambed material, and surface disturbance. Canopy coverage of the water's surface was obviously not found in reaches flowing through meadows dominated by grass/forb vegetation types and moderate in relatively small streams which flowed through dense forests. Relatively wide stream reaches such as Rl LaMarche had low canopy coverage even though it flowed through a forested canopy. R2 Johnson Creek had relatively low canopy coverage because the adjacent forested land on one side of the stream had been clearcut. R2 Joseph and Rl Wyman creeks flowed through open forests mixed with small stringer-type meadows. Low overhead cover was related to the amount of woody brush and grasses on the streambank, while high overhead cover was related to the amount of woody brush and low overhanging branches from trees. Streambed Condition The condition of the streambed was assessed by visually estimating the composition of the streambed; ranking the two predominant substrate types, the size of the material surrounding the two dominanat substrate types, and embeddedness of the dominant substrate types into a "substrate score"; visually 31 Table 9. Date of survey, temperature at time of survov fPl f, i * ^ «.» estimated wetted width (ft), channel "dth (ft) average d»nth%"\' "u",1 le"8th °f 8UrV^ed '"tion (ft.), and RANGER DISTRICT Date Str6RLch (n) —d Hf «~ «-.* SS* "SS X"h8e "depth" TdhalthB8 r" (F> Ccfs) (ft) (ft) (f , °?p™ dePth depth depth WISE RIVER DISTRICT Ad son Ck 1 ( 38) 11 16/ 86 LaMarche Ck 1 ( 21) 7/ 31/ 86 54 2 ( 26) 7/ 31/ 86 Meadow Ck 2 ( 29) 11 29/ 86 46 Mono Ck 1 ( 17) 7/ 14/ 86 2 ( 39) 11 9/ 86 Wyman Ck 1 ( 24) 11 25/ 86 59 2 ( 24) 7/ 28/ 86 52 ratio (in) (in) 3.3 29.5 29.5 3.1 3.0 2.6 8.6 7.9 406'° , 6.6 5.7 6.9 H.7 ( 1.5) ( 2.1) ( 2.7) 495'6 , *l'L , 12'8 6-3 28-3 ( 2.6) ( 2.4) ( 3.4) ( 2.3) ( 2.9) ( 4.2) " ( eil) ( 2.8) ( 4.9) ( 2.3) (3.5) 10.8 4.8 ( 3.2) ( 3.0) 1018.0 39.6 31.8 10.0 41.4 195 , . ( 6'8) ( 8-9> (3.0) ?\% (J;J, 1184.0 33.2 26.0 19 9 ,n 1 ,, , < *•» < 3.0) (199:2) 2oa {\i% »;?, 493-0 ( f; « r ;-2, , 5-: 22-3 »-4 2.5 ( 2.2) ( 2.8) ( 1.4) ( 2>3) ( Illy 14.5 4.3 ( 5.4) ( 2.2) 669'9 < l-l , i'l ,13-5. ♦.» 20.9 ( 5.0) 621.0 21.0 15.4 73 70 1 1K , r-.ni / , !., , .-. 28,1 15-3 3.6 ( 2.6) 834-° , "'!, , 18-?> ,i0-2 31.3 19.1 4.0 < 6.3) ( 7.4) ( 5.8) C ilij ( 3.4) Table 9. (continued) RANGER DISTRICT Date Stream Reach (n) surveyed Temp • (F) Flow (cfs) Length (ft) Channel width (ft) Wetted width (ft) Average depth (in) Width 1 depth ratio :o Thalweg depth (in) Bank depth (in) WISDOM DISTRICT Elk Ck 1 ( 25) 8/ 12/ 86 51 2.3 571.0 13.1 ( 2.6) 9.2 ( 3.2) 6.8 ( 3.6) 22.0 12.7 ( 6.1) 5.1 ( 3.7) 2 ( 27) 8/ 13/ 86 57 1.7 528.0 13.2 ( 1.7) 8.7 ( 3.9) 5.6 ( 3.0) 23.3 10.2 ( 4.6) 3.4 ( 2.5) Johnson Ck 1 10/ 7/ 86 45 7.5 ( 24) u-; 597.0 24.4 ( 4.0) 20.4 ( 7.0) 7.9 ( 3.2) 35.9 14.7 ( 5.2) 2.7 ( 1.7) 2 ( 21) 8/ 20/ 86 46 2.2 534.0 27.3 ( 8.4) 18.0 ( 5.4) 4.7 ( 2.1) 57.2 11.0 ( 5.4) 1.5 ( 2.0) Joseph Ck 1 ( 22) 8/ 19/ 86 54 3.4 575.0 16.5 ( 3.4) 10.4 ( 3.7) 8.5 ( 4.1) 19.3 15.2 ( 7.0) 3.0 ( 3.2) 2 8/ 19/ 86 56 2.2 ( 38) 592.0 13.3 ( 2.4) 7.9 ( 3.7) 6.0 ( 3.4) 21.1 10.7 ( 5.5) 3.0 ( 2.3) Sheep Ck 1 8/ 7/ 86 43 1 .9 ( 18) 555.0 12.2 ( 2.8) 7.9 ( 5.2) 8.9 ( 5.2) 15.6 14.1 ( 7.1) 6.9 ( 5.1) 2 8/ 6/ 86 48 1.9 ( 30) 534.0 15.2 ( 2.9) 11.9 ( 2.7) 4.7 ( 2.1) 35.0 9.1 ( 3.2) 2.7 ( 1.8) Steel Ck 1 9/ 11/ 86 50 6.7 ( 19) 855.0 32.9 ( 4.9) 25.9 ( 8.9) 7.0 ( 3.4) 57.8 11.7 ( 5.0) 1.9 ( 2.0) Table 10. Mean estimates of cover availability including percentage undercut banks, canopy density over the water's surface (%), instream cover (%), low (1.0 foot or less above the water's surface) overhead cover (%), high (more than 1.0 foot above the water's surface) overhead cover (%), and depth of undercut banks (in) for stream reaches draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses. RANGER DISTRICT Overhead cover Percent Depth Canopy Instream :___ . Stream n undercut undercut density cover Low High Reach bank bank (%) (%) (%) (%) WISE RIVER DISTRICT Ad son LaNarche 1 21 26 11 ( 12) 3.6 ( 3.1) 10 ( 20) 51 ( 25) 4.8 ( 2.5) 0 ( 0) 38 56 4.8 2 30 13 22 ( 28) ( 3.5) ( 8) ( 30) ( 13) ( 29) 20 ( 16) 3.9 ( 4.3) 4 ( 6) 39 ( 24) 3 8 ( 2) ( 4) 55 ( 14) 6.8 ( 3.1) 3 ( 10) 21 ( ID 6 11 ( 5) ( 8) Me ad ow 2 29 53 4.4 34 49 18 19 ( 24) ( 3.2) ( 34) ( 25) ( 19) ( 17) Mono 1 17 11 3.6 10 87 8 9 (6) ( 7) ( 12) 4.8 17 8 0 ( 17) ( 7) ( 0) Wyman I 24 24 27 ( 16) 2.7 ( 2.4) 2 ( 6) 50 ( 24) 7 ( 6) 17 ( 16) 33 ( 22) 3.2 ( 3.1) 0 ( 0) 30 ( 28) 5 ( 5) 4 ( 6) 34 Table 10. (continued) RANGER DISTRICT Stream Reach WISDOM DISTRICT Elk 1 25 2 27 Percent Depth Canopy undercut undercut density bank bank (%) Ins tream cover (%) 47 ( 24) 6.5 ( 4.1) 1 ( 2) 9 ( 6) 66 ( 18) 6.9 ( 4.5) 18 ( 21) 20 ( IS) Overhead cover Low (%) High (%) 3 5 ( 5) ( 6) 10 8 ( 9) ( 10) Johnson 24 21 36 ( 19) 3.0 ( 2.6) 11 ( 19) 47 ( 19) 10 ( 9) 17 ( 15) 44 ( 30) 6.0 ( 7.7) 5 ( 9) 22 ( 25) 21 ( 25) 20 ( 22) Joseph Sheep 22 38 18 30 45 3.4 0 10 9 IS ( 29) ( 3.2) ( 0) ( 7) ( 8) ( 14) 51 6.4 8 17 7 13 ( 26) ( 5.1) ( 19) ( 15) ( 7) ( 14) 58 7.1 0 12 14 24 ( 27) ( 5.3) ( 0) ( 8) ( 13) ( 24) 53 7.0 30 29 16 24 ( 26) ( 5.6) ( 31) ( 18) ( 15) ( 21) Steel 19 22 ( 24) 2.1 ( 1.9) r ( 0) 14 ( 10) ( 1) ( 5) 15 Table 11. Mean estimates of percentage of streambank altered, streambank vegetation stability rating, percentage of streambank vegetation utilized, and bank angle (degrees) for stream reaches draining the Beaverhead National Forest surveyed during 1986. Standard deviations are in parentheses. DISTRICT Stream Reach Streambank Vegetation Vegetation Bank altered stability use angle (%) (rank) (%) (degrees) WISE RIVER DISTRICT Adson Ck 1 LaMarche Ck 1 Meadow Ck 2 Mono Ck 1 Wyman Ck 1 38 21 26 29 17 39 24 24 15 ( 20) ( 4 0) 6 ( 4) 68 ( 23) 9 ( 7) ( 4 1) 0 ( 1) 54 ( 19) 12 ( 11) ( 4 1) 10 ( 11) 68 ( 15) 4 ( 4) ( 4 0) 1 ( 1) 76 ( 18) 2 ( 3) ( 4 0) 10 ( 1) 76 ( 14) 37 ( 23) ( 4 1) 10 ( 2) 73 ( 18) 28 ( 20) ( 3 1) 60 ( 21) 54 ( 19) 44 ( 25) ( 2 1) 60 ( 17) 57 ( 19) 36 Table 11. (continued) DISTRICT Stream Reach n WISDOM DISTRICT Streambank Vegetation Vegetation Bank altered stability use angle (%) (rank) ' (%) (degrees) Elk Ck 1 Joseph Ck 1 Sheep Ck 1 Steel Ck 1 25 27 Johnson Ck 1 24 21 22 38 IS 30 19 43 2 38 64 ( 22) ( I) ( 19) ( 21) 21 3 3 63 ( 18) ( 1) ( 2) ( 16) 18 3 69 73 ( 12) ( 1) ( 9) ( 14) 10 4 16 75 ( 9) ( 0) ( 8) ( 14) 19 3 13 67 ( 17) ( 1) ( 16) ( 14) 17 4 2 73 ( 16) ( 1) ( 2) ( 12) 15 4 10 70 ( 13) ( 1) ( 5) ( 17) 6 4 1 65 ( 5) ( 0) ( 1) ( 21) 50 1 67 66 ( 21) ( 1) ( 8) ( 17) 37 estimating surficial embeddedness; measuring embeddedness; measuring the diameter of a streambed particle which was larger than 90% of all remaining streambed particles (D-90); and sampling the streambed with a "hollow core" sampler within known or suspected spawning areas to provide a more reliable estimate of streambed composition. Occular Estimates Occular estimates of streambed composition and condition including embeddedness » substrate score, and average D-90 are presented in Table 12 and Figure 4. Fine streambed material ("silts" and "sands") made up a relatively large percentage of the streambed (in decreasing order of percentage of fines) in R2 Mono, R2 Wyman, Rl Joseph, Rl Steel, R2 LaMarche , and R2 Joseph creeks. It should be noted that in Elk Creek R2 contains a higher proportion of "fine" material in the streambed than Rl , probably an indication of erosive nature of the surrounding geology and inability of this portion of the stream to transport sediments out of its stream channel (Figure 4). Fine material was relatively low (in increasing order of percentage of fines) in Rl Mono, Rl LaMarche, Rl Wyman, R2 Meadow, and R2 Sheep creeks. The other stream reaches contained from 20 to 30 percent "fine" material. In general, occular estimates of embeddedness reflected the percentage of "fine" sediments within the streambed with reaches having high percentages of "fines" also having relatively high embeddedness values. The exception to this general rule was R2 of Sheep Creek where the streambed appeared to be highly embedded even though surficial "fines" were estimated to make up only 13% of the streambed. Substrate scores were inversely related to percentage of "fines" and embeddedness values which is due to the fact that both these measures are used to define substrate score (Table 12). Average D-90 values ranged from 0.6 to 29.4 inches. Embeddedness Measurements., The majority of the sample sites had an average embeddedness of 40 to 50 percent (Table 13) . The characteristics of each embeddedness sample site is presented in Appendix F. Those reaches with embeddedness averaging 30 to 40 percent were Rl Jerry, upper R2 LaMarche, Rl Steel, R2 Big Swamp, and lower R2 Trail creeks, and R3 Wise River. The only reaches with embeddedness values higher than 60 percent were R2 Elk and R2 Sheep creeks. Neither of these reaches have much land-use development, but both overlay erosive geological types. It should be noted that timber harvest activities are planned in lands adjoining these two reaches. The proposed timber harvest activities and scheduling of these activities have been addressed in the Trail Creek Area Analysis. The percentage of free matrix particles (those particles which were not embedded at all) ranged from zero to 30 percent. Embeddedness Variation Within a Reach A cursory examination of the difference between sampling in two separate riffles in Rl Steel and Rl Adson creeks found little difference between average 38 : I'Vo'dl; szszszz 3"^i^i-i-^:-ss;t;i— ..i-s^ts^.^'^-i;;^""— J- -~t <»■ -*•— — • - in parentheses. * eavernead National Forest surveyed during 1986. Standard deviations arf RANGER DISTRICT Substrate composition (%) Stream 6 ( 10) 0 ( 1) 25 ( 4) 16 ( 7) 1 ( 4) 0 ( 0) 24 ( 12) 16 ( 12) 31 ( 9) 28 ( 23) 0 ( 0) 1 ( 6) 22 ( 7) 10 ( 9) 2 ( 3) 2 ( 5) ( 10) 25 ( 6) 62 ( 18) 36 ( 18) 15 ( 6) 98 ( 7) 48 ( 14) 65 ( 24) 2 ( 3) 0 ( 0) 5 ( 5) 5 ( 7) ( 13 4) ( 21 1) ( 12 2) ( 19 3) ( 21 0) ( 5 1) ( 19 2) ( 13 3) Keach „ silt Sand 8ravelgr^ CobMe ^ UJJ. Embeddedness Substrate M, "" *•*•; score (in) WISE RIVER DISTRICT Ad son Ck 1 38 10 12 17 34 21 C 14) ( 13) ( 17) ( 22) ( 19) LaMarche Ck 1 21 1 5 9 16 28 ( 2) ( 2) ( 4) ( 4) (5) 2 26 9 23 17 44 6 < 6) ( 10) ( 5) ( 15) ( 4) Meadow Ck 2 29 3 10 11 is 18 ( 3) ( 6) ( 6) ( 6) (6) Mono Ck 1 17 1 3 13 0 24 < 2> C 3) ( 9) ( 0) ( 11) 2 39 36 32 30 o o < 29) ( 21) ( 29) (1) (o) Wyman Ck 1 U 7 5 2 12 41 < *> C 5) ( 3). ( 5) ( 14) 2 2<* 18 20 17 30 12 ( 8) ( 6) ( 7) ( 7) (9) WISDOM DISTRICT Elk Ck 1 25 10 17 19 28 24 < 5)( 6) ( 8) < 9) (,) (£, ,"*, ( l'l 73 14 15.3 < U> ( 3) ( 7.4) ( 6.0 2.9) ( 28.7 4.7) ( 3.9 2.2) ( 21.1 5.5) ( 29.4 7.4) { 0.6 2.8) ( 16.9 3.7) ( 6.4 5.4) 27 6 18 27 25 15 ( 7) ( 5) < » < 9) ( 4) ( 5) (7) I , , *t , »•? Table 12. (continued) RANGER DISTRICT Substrate composition (%) St ream D , Small Large Reach n Sih s=„j „--.._, ___. Silt Sana gravel grarvgeei Cobble bo^er boulf^ ^°%?™° S»»— °7'° _ uuuiuei ^/aj score (in) 21 6 18 13 21 22 4> Joseph Ck ° 1 22 14 23 17 25 21 ( 6) ( 5) f 5) ( 7) ( 9) ( 1) (Si f fL , »* . 6.1 38 11 19 15 21 22 ( 5) ( 5> < 5) ( 8) ( 8) (6) f « ■ , ?L , " . ?•? < 16> ( 2) ( 1.3) 55 16 9.2 < 18) ( 3) ( 2.8) Sheep Ck 1 18 9 18 20 31 21 ( 9) ( 9> < « ( 13) ( 10) ( 2) ( 01 , 11 , l* . *•* Steel Ck 1 19 11 26 21 28 13 < » < » c s) < o < «, t o) co) c ?59, ( »}, ( ;:;, 20 ( 6) ( 6 4) 10 ( 5) ( 10 12) 0 ( 1) ( 0 0) 9 ( 6) ( 2 3) 1 ( 2) ( 0 0) 12 ( 8) ( 7 7) 0 ( 0) ( 0 0) (in) Johnson Ck 1 24 6 21 13 15 21 c 3) < 7> c « ( 3) ( 4) ( z6) r » t I , ll , »!•? < 18) C 3) ( 4.3) . 5\ 17 18.8 ( 16> ( 3) ( 9.7) 30 A 9 20 22 27 ( 2) ( A) < » < 12) ( 9) c 8) ( ) , f i , " , "•; < 20) ( 3) ( 2.0) . « 16 16.1 t«) ( 3) ( 7.5) SMALL GRAVEL /""""""l L A9* lp3 SAND LARGE GRAVEL SMALL GRAVEL SAND SILT 10% "LARGE BOULDER 0% SMALL BOULDER 2% COBBLE ELK R1 LARGE GRAVEL SILT 5.9% LARGE BOULDER 55! SMALL BOULDER 55? COBBLE 14.9% ELK R2 SMALL GRAVEL SAND SAND 23% m mm. S 14% |H Mtell«fi LARGE BOULDER 0% SMALL BOULDER 0% SMALL GRAVEL 15.2% LARGE GRAVEL JOSEPH R1 LARGE GRAVEL COBBLE JOSEPH R2 LARGE BOULDER 2" SMALL BOULDER 9.1! SMALL GRAVEL SAND LARGE GRAVEL SILT 9% LARGE GRAVEL LARGE BOULDER 0% SMALL GRAVEL COBBLE SHEEP R1 LARGE BOULDER 6.9% SMALL BOULDER 1 1 .9% SHEEP R2 Figure 4. Pie chart diagrams of substrate composition which was estimated ocularly for two reaches in Elk, Joseph, and Sheep creeks which are all tributaries of Trail Creek. Percentages of fine in te rial (less than 0.25 inch) are offset. 41 Table 13. Average percent embed dednes s , "free matrix particles", and average size of particles measured (range in inches) from measurements made of streambed particles in riffle habitat types of stream reaches draining the Beaverhead National Forest surveyed during 1986. Stream Adson Creek Percentage of Average particle "free matrix" size (mm) Reach n Embeddedness particles (range) 1/ 1A 110 1B!/ 107 52 47 10 56 (24 - 113) 54 (18 - 121) Elk Creek 1 102 50 8 54 (28 - 190) Jerry Creek Johnson Creek Joseph Creek 100 72 2 146 115 108 63 34 43 42 41 3 14 17 10 12 72 (40 - 175) 64 (28 - 164) 66 (23 - 200) 41 (19 - 84) 50 (20 - 135) LaMarche Creek 1 105 11 112 2A 2/ 2B 105 33 48 33 23 11 14 71 (28 - 243) 51 (24 - 111) 59 (20 -225) Meadow Creek 2 3 01 47 60 (27 - 144) Mono Creek 1 101 50 11 67 (24 - 196) 42 Table 13. (cont,) Stream Sheep Creek Steel Creek Swamp Creek Trail Creek Wise River Wyman Creek Percentage of Average particle "free matrix" size (mm) Reach n Embeddedness particles (range) 1 96 44 2 106 69 ia3' 95 39 iB3' 99 38 95 2 104 94 1 101 103 35 36 37 40 41 17 2 9 19 11 16 18 12 57 (27 - 137) 63 (30 - 205) 39 (18 - 104) 39 (13 - 75) 60 (28 - 135) 41 (16 - 74) 57 (28 - 133) 85 (33 - 231) 49 (29 - 80) 1/ 2/ 3/ Sample 1A was immediately above the sample section near stream mile 2.0. Sample IB was in lower Adson Creek near stream mile 1.0. Sample 1A was within the sample section near stream mile 5.0 while sample IB was near the hollow core site near stream mile 7.1. Sample 1A better typifies the reach's embeddedness. Sample 1A was in one riffle while sample IB was in another riffle approximately 200 yards below the IA site. 43 sampled embeddedness between the two riffle sites (Table 13). There was a relatively large difference between the two riffles sampled in R2 LaMarche Creek, however, it should be noted that the riffle sampled at 2A was located approximately two miles below the riffle sampled at 2B. Gradient differences existed between these two sites. Site 2B was in a higher gradient portion of the reach near its uppermost boundary, while site 2A was in a more "typical" area of R2 in a low gradient meadow. Comparison Between Visual and Measured Embeddedness A comparison between visual and measurement techniques for estimating embeddedness in riffles found that although there were differences between several pairs of estimates there was no significant difference between the two methods (P ~> 0.10; Table 14) using a Wilcoxon matched-pairs sign ranked test (Daniel 1978). Individual differences in Rl Adson, R2 Meadow, Rl Mono, and R2 Sheep creeks were due to sampling in different riffle sites which were spatially separated by distances of from 0.1 to 1.0 mile. Other large differences, for example R2 Wyman Creek, could not be explained by differences due to sampling in different locations. Hollow Core Samples Hollow core sampling conducted during 1985 and 1986 found that few sites contained less than 25% of material smaller than 0.25 inch (Table 15). Several sites contained more than 40% of material smaller 0.25 inch which indicates a potential sedimentation problem. Survival predictions for brook, cutthroat and rainbow trout embryos are presented to show relative health of the spawning gravels at these sites. It must be remembered that these survival predictions are based on laboratory data and field survival rates may be quite different dependent upon micro-habitat characteristics. The sites in East Fork Ruby River and Harriett Lou, Mill, South Fork Willow, Big Swamp, Jerry, and LaMarche creeks were not ideal spawning sites because these sample sites contained moderate to high amounts of cobble and/or boulder. Variability within sample sites (standard deviation divided by the mean expressed as a percentage) ranged from 15 to 47 percent and a higher variation was generally observed in sites which had higher percentages of material larger than 2.0 inches in diameter. The exception to this was in Adson Creek which had high variability between samples within the sample site, but had relatively low amounts of cobble within the streambed. A problem encountered in sampling Adson Creek was that the streambed material appeared to be "perched" on a layer of valley bottom silts which were encountered in seven samples at depths of approximately four to six inches. These silts were included in these samples and biased the samples with respect to the percentage of fine material. The inclusion of these valley bottom silts in the samples also inflated the amount of "fine" material estimated to occur in Adson Creek. A separate discussion of the results obtained in the Trail Creek drainage is warranted because the Beaverhead Forest is presently in the process of completing an area anaylsis to help schedule timber harvest within the drainage. In 1985 ten hollow core samples were taken from Trail Creek downstream from the May Creek Campground. Those samples estimated that approximately 44% of the streambed material was comprised of material smaller than 0.25 inch (Table 15). The Beaverhead Forest's Management Team wanted to Table 14. Comparison between visual and measurement estimates of embeddedness in riffles made during 1986 in streams draining the Beaverhead National Forest. S t re am Reach Vis Ad son Ck 1 66 Elk Ck 1 2 51 63 Johnson Ck 2 55 Joseph Ck 1 2 53 44 LaMarche Ck 1 2 21 47 Meadow Ck 2 27 Mono Ck 1 2 12 53 Sheep Ck 1 2 53 54 Steel Ck 1 53 Wyman Ck 1 2 39 63 Measured D. Rank i 11 6.0 3 4.0 - 12 - 7.5 - 1 - 2.0 52 14 9.0 50 1 2.0 63 0 43 12 7.5 42 41 33 48 47 - 20 - 12.0 50 - 38 - 14.0 38 15 10.5 44 9 5.0 69 - 15 - 10.5 44 9 5.0 40 - 1 - 2.0 41 22 13.0 Sum of positive and negative ranks T+ = 57.0 T- ~ 48.0 Ko: Bedisn of difference is zero (ie. There is no difference between methods) lis: Median of differences is not zero P > 0.104 Therefore, conclude that there is no significant difference between methods. Table 15. Average percentage of material (by dry weight) less than 0.37 inch (9.5 ram), 0.25 inch (6.34 mm) and less than 0.03 inch (0.85 mm) from hollow core samples taken from typical spawning areas during 1985-86 and predicted survivals of westslope cutthroat trout (WCT), rainbow trout (RB) , and eastern brook trout (EBT) embryos from egg deposition to fry emergence based on laboratory studies conducted by Irving and Bjornn (1984) for WCT and RB and survival relationships developed by the author using data from Witzel and MacCrimmon (1985). Year Stream (n) Reach Percentage 9.5 mm 6.34 less than mm 0.85 mm P surv WCT A- ► A1' redicted ival (%) of RB EBT 1985 Doolittle Creek (10) 1 37 32 10 19 29 49 E. Fork Ruby River (10) 1 30 24 7 31 45 59 Harriett Lou Creek ( 5)2/ 1 27 27 15 27 15 56 3/ Meadow Creek (1C) 2 37 31 11 17 23 50 Mill Creek (10) 2 32 26 7 29 48 57 S. Fk. Blacktail Ck (10) 1 58 48 24 50 3 28 S. Fk. WilloT? Creek (10) 2 29 24 5 42 61 60 Trail Creek (10) 1 49 44 15 9 4 33 1986 Adson Creek (9) 1 48 40 19 0 2 42 Big Swamp Creek (10) 2 27 22 6 40 56 63 Jerry Creek (1C) 1 30 25 7 33 48 59 Joseph Creek (10) 1 41 36 12 12 16 43 LaMarche Creek (1C) 2 40 27 7 32 8 56 Trail Creek (10) 2 64 56 18 8 0 17 Wise Rivev (1C) .•> 34 30 8 25 39 52 Table 15. (continued - footnotes) Many of these streams do not support populations of cutthroat or rainbow trout; however, these relative survival values are presented to indicate the relative condition of the spawning habitat. 2/ Harriett Lou Creek contained a streambed composed of large angular boulder and cobble surrounded by fine material. A cursory of the lower portion of the stream did not locate any spawning habitat; therefore, streambed sampling was done in a boulder/cobble habitat which was extremely difficult to sample. These data are of questionable value. 3/ The best spawning site, and therefore sample site, in Meadow Creek was was located immediately downstream from an old bridge site. This area may have contained an abnormally high level of fine sediment and may not accurately reflect the streambed condition of the entire stream. 47 further investigate the possible source of this fine sediment to document whether it originated from highway construction, livestock and timber activities within the Forest, or mining activity. While it is not possible to accurately determine the source of sediment from hollow core sampling, it was hoped that sampling two additional sites (one near the mouth of Joseph Creek and one in Trail Creek immediately above the mouth of Joseph Creek) would shed some light on where this sediment originated. It can be seen that the Joseph Creek sample contained an estimated 36% material smaller than 0.25 inch, while upper Trail Creek contained an estimated 56% material smaller than 0.25 inch. These data suggest that past management activities within the upper Trail Creek drainage (most noteably past livestock damage to streambanks) probably were the primary sources of fine sediment seen in the lower drainage. This sediment appears to be slowly "migrating" down the stream channel with the main "pulse" of sedimentation presently located near the mouth of Joseph Creek. Further sampling in upper Trail Creek near the mouth of Sunshine Creek during 1987 should help quantify any possible sediment "recovery" of the upper channel. It is likely that since Trail Creek has a relatively low gradient and numerous beaver ponds throughout its length, making its ability to transport sediment relatively low, and presently has large quantities of fine sediment "stored" within the streambed "flushing" of these fine sediments from the streambed will take a long time. Comparison Between Measured Embeddedness and Hollow Core "Fines" Simple linear regression as used to compare average measured embeddedness estimates with percentage of material less than 6.34 mm (0.25 inch) estimated from hollow core samples taken at or near the same locations (Figure 5) . The regression was calculated using both the untransformed data and after transforming both the embeddedness and hollow core percentages using the arcsin square root transformation recommended by Zar (1984). Figure 5 shows the correlation obtained using the untransformed data. The transformed data also yielded an "r" = 0.82. While this correlation shows promise, the increasing scatter of data points from the predictive line at the higher levels of embeddedness and percent fines would present a problem if one were trying to predict results from one measure using the other. Further testing of correlations between these two methods needs to be done to ensure that results obtained using one technique could be compared to results from the other. Principle Component Analysis (PCA) PCA was used in an attempt to consolidate all the measured habitat variables into several functions which could then be regressed against fish abundance variables. The habitat surveys conducted during 1986 resulted in 442 separate observations for each of the 25 variables in 17 sample sections. All 11,050 observations (442 observations times 25 variables) were used in the PCA. The five factors which explained most of the variance observed within stream habitat are listed in Table 16 along with the coeffecients assigned to each individual variable. The variables are segregated into overhead cover, streambed, streambank, channel shape, and instream cover classes. Coeffecients larger than 0.5 are highlighted in bold type. It can be seen that Factor 1 explained approximately 33% of the variation in habitat and this factor relied heavily on streambed related variables. This suggests that the streambed component is an important component in explaining stream habitat and any 48 PERCENT, FINES = 0. t 12* 0. 784( PERCENT EMBBEONESS) r a 0.87 Percent "fines" 2 0 40 60 Percent embeddedness (measured) Figure 5. Relationship between "percent fines" (material less than 0.25 inch) sampled by hollow core sampling and percent embeddedness sampled by measuring at least 100 individual particles at the same sample site. 4Q Table 16. Principal components analysis for habitat variables measured in the "detailed habitat survey section" across 442 measured transects in 17 stream sections surveyed during 1986. Habitat Factors variable 1 2 3 4 5 Overhead cover Canopy density 0.28388 0.33354 - 0.10441 0.09041 0.06647 Low overhead veg. 0.14411 0.51195 - 0.04891 0.24565 0.54523 High overhead veg. 0.23471 0.44501 - 0.05166 0.45111 0.51418 Silt - 0.58286 0.06916 0.13728 - 0.40189 0.31313 Sand - 0.67058 0.04448 0.08234 - 0.12119 0.07420 Small gravel - 0.45222 0.12519 - 0.27031 - 0.04549 - 0.37720 Large gravel - 0.12440 - 0.22720 - 0.18430 0.78116 - 0.09747 Cobble 0.60574 - 0.18535 - 0.13294 0.20525 0.13543 Small boulder 0.81873 0.09928 0.23210 - 0.25801 - 0.01038 Large boulder 0.66566 0.15138 0.23966 - 0.37264 - 0.05586 Embeddedness - 0-83389 0.12326 - 0.08600 - 0.08596 0.14915 Substrate score 0.90698 - 0.07155 0.10452 0.10908 - 0.00045 D-90 0.84089 0.07645 0.22507 - 0.17867 - 0.11697 Streambank Bank alteration - 0.39872 - 0.64645 - 0.05388 - 0.21547 0.30242 Vegetation stability 0.15998 0.74646 - 0.04219 - 0.03573 - 0.30741 Vegetation use - 0.04498 - 0.63240 0.18284 - 0.03537 0.43132 Undercut bank - 0.20902 0.24272 0.38943 0.37868 - 0.05433 Depth of undercut - 0.12364 0.53517 0.10685 0.26850 0.05923 Bank angle - 0.06250 0.59962 0.05413 - 0.02066 - 0.00557 Ch.ajD.ne 1_ shape. Wetted width 0.17611 - 0.44377 0.68149 0.25651 - 0.12203 Channel width 0.17155 - 0.58290 0.57898 0.32008 - 0.1 191 7 Average depth - 0.51367 0.20537 0.74773 - 0.00872 - 0.03926 Thalweg depth - 0.39467 0.11409 0.81920 - 0.02870 - 0.02533 Near shore depth - 0.46507 0.43989 0.46484 0.06105 - 0.04970 Ins tr earn cover 0.55808 0.25300 0.37591 - 0.26785 0.34344 VARIANCE EXPLAINED 6.128 3.668 FINAL COMMUNALITY ESTIMATES: TOTAL = 18.371 2.938 1.869 1.409 50 changes in streambed resulting from management activities could change stream habitat. It is unclear at this time what- effect changes in habitat has on fish populations from the data collected to date. A discussion of these results will be presented later. Factors 2, 3, 4, and 5 relied heavily on streambank, channel shape, large gravel (related to spawning gravel), and overhead cover variables, respectively. These factors explained approximately 20, 16, 10, and 8 percent of the variability observed in stream habitat, respectively. FISH ABUNDANCE Catch per Unit Effort (CPUE) Catch of fish 3.0 inches and longer by species made in a single electrofishing pass standardized to 1,000 feet of stream length are presented in Tables 17 through 19. CPUE ranged from 3 to 660 fish per 1,000 feet for brook trout in sections where brook trout were captured, 1 to 177 fish per 1,000 feet for cutthroat trout in sections where cutthroat trout were captured, 2 to 35 fish per 1,000 feet for rainbow trout in sections where rainbow trout were captured, and 1 to 33 fish per 1,000 feet for arctic grayling in sections where grayling were captured. Tributary Population Estimates Population estimates made for tributary reaches electrof ished during 1985 and 1986 are presented in Table 20. R2 Governor, R2 LaMarche, and R2 Wyman creeks had the highest densities of brook trout 6.0 inches and longer, while R2 Elk, R2 Joseph, Rl LaMarche, R2 Old Tim, Rl Steel, and R2 Trail creeks all had relatively high densities of brook trout 3.0 to 5.9 inches long. Reaches which had extremely low densities of brook trout were Rl Adson , R2 Cow Cabin, R2 Morrison, R2 Pole, R2 Ruby, Rl Sheep, R2 Steel, Rl Trail, and Rl Wyman creeks. Rl Adson, R2 Cow Cabin, R2 Morrison, and R2 Pole are all reaches where the streams are very small and the low density of brook trout was to be expected. R2 Ruby, Rl Sheep, and R2 Steel appeared to be impacted by livestock grazing. Rl Trail appeared to be impacted by high levels of fine sediment (see above for a discussion of the source of this sediment) . Densities of westslope cutthroat trout (identified using external morphological characteristics - see below for a discussion of genetic analysis conducted on some of these populations) were high for R2 Brown's Canyon, R2 Painter, and R2 Reservoir creeks. Streams within the Wise River District generally had lower densities of cutthroat. It appears that dewatering near the mouths of tributaries to Horse Praire Creek and Grasshopper Creek limited rainbow arid/or brook trout from entering these tributaries and competing and/or introgressing with the cutthroat trout populations native to these tributaries. In general, cutthroat trout populations are limited to small headwater and high gradient tributaries or are above some type of fish passage barrier. Rainbow trout were found in Rl LaMarche and R2 Wyman creeks. The rainbow in LaMarche Creek could have orginated from either releases of hatchery rainbow trout between 1928 and 1954 or from fluvial Big Hole River rainbow populations. SI Table 17. Relative fish abundance by species for fish 3.0 inches and longer in streams draining the Beaverhead National Forest within the Dillon Ranger District derived from single pass electrof ishing catches using a Coffelt BP-1C backpack electrof isher during 1985 and 1986. Section Number per 1.000 feet Legal length 1/ S t re am description Year (ft.) EBT WCT RB GR LING OTHER Andrus Ck T 7SR14WSec 5CB 85 565 183 4 - - 2 - Browns Canyon T 8SR13WSec30AA 86 300 - 177 Cow Cab in Ck T 6SR14WSec24EC 86 300 30 - Fox Ck T 7SR14WSecl2AC 85 550 171 13 T 6SR14WSec33DC 85 450 129 _ Governor Ck T 6SR14WSec 6BA 85 500 84 2 - 32 40 MWF T 7SR14WSec 6DA 85 1,375 168 5 - - 3 T 7SR14WSec32BA 85 325 123 28 Morrison Ck T13SR12WSecl5CC 86 300 14 — — — — _ T13SR12WSecl0DC 86 300 No f ish captured except 2 sculps. Old Tim Ck T 4SR13WSec33DD 86 300 220 ----- Painter Ck T 8SR14WSec25AB 86 300 - 120 Pass Ck T12SR12W3ec33CC 86 300 144 _____ Pole Ck T 5SR13WSec34AD 86 300 24 4 Reservoir Ck T 8SR13WSecl6AB 86 300 - 97 Saginaw Ck T 7SR15WSecl0 No t s amp led - flows very little water Thayer Ck T 7SR14WSec26BB 85 320 225 4 - - - _ Abreviations for species are: EBT = eastern brook trout; WCT - westslope cutthroat trout; RB = rainbow trout; GR = arctic grayling; LING - burbot; and under the other - IIB _ hybrids between RB end WCT; MWF = mountain whitefish. Table 18. Relative fish abundance by species for fish 3.0 inches and longer in streams draining the Beaverhead National Forest within the Wise River Ranger District derived from single pass electrof ishing catches using a Coffelt BP-1C backpack electrof isher during 1985 and 1986. Legal description Year Section length (ft.) Nuiabejr_ per 1 ,000 feet S t re am EBT 1/ WCT RB GR LING OTHER Ad son Ck T lSRllWSec28BC T !SRllWSec28CB 86 86 406 275 5 11 12 4 - - - - Bryant Ck T lNR12WSec 8AD T lNR13WSec25AB 85 85 500 200 150 145 - - - - - Butler Ck T lSRHWSec30CD 86 200 No f ish c aptu red California Ck T 3NRllWSec30DB 35 580 61 - 35 - - - Fishtrap Ck T 2NR13WSec32DD 86 200 660 - 15 5 10 - Harriet Lou T lSR12WSecl2BB T lSR12WSec 1AC 85 85 300 250 - No i _ _ fish observed - - Lacy Ck T 2SR12WSec 6DA T 3SR12V/Sec 2 AD 85 85 250 500 7 6 42 4 36 - - 20 - LaMarche Ck T 2NR13WSec22CC T 2NR13WSecl6BB 86 86 1,018 1,184 131 82 - 22 11 - 1 MWF Meadow Ck T lNR12WSec36AC T lNR12WSec36BA 85 86 500 493 - 28 14 - - - - Mono Ck T 4SR12WSec 4BA T 4SR12WSec 5DA T 4SK12WSec 3AB 86 86 86 496 700 300 - 34 24 9 - «• _ — O'Dell Ck T 3SR13WSec25AC 85 500 88 - - - - 18 HB Sevenmi le Ck T 3NR12WSec23AC 85 400 245 - - - - - Wyman CI; T 3SR13WSec24DA T 3SB12WSecl7CC 86 86 621 834 40 155 7 7 1 6 8 HB 2 HB Abrevii; t: ons fox species f.ie: F.BT = eastern brook trout; WCT = westslope cutthroat trout; KB = rainbow trout; GR = arctic grayling; LING «= burbot; and under the othwr - "»! - hybrids between ^B ; nd WCT; MWF - mountain wh it of ish . 9/ Arctic groyliix were observed and angled from the lower .segment of LaMntche fVcok. 5 3 Table 19. Relative fish abundance by species for fish 3.0 inches and longer in streams draining the Beaverhead National Forest within the Wisdom Ranger District derived from single pass electrof ishing catches using a Coffelt BP-1C backpack electrof isher during 1985 and 1986. Legal description T lSR17WSecl2CB T 3SR15WSecl8DD T 3SR15WSecl9DC T 4SR16WSec32AA Year 86 36 85 85 Section length (ft.) Number oer 1.000 feet Stream 1/ EBT WCT RB GR LING OTHER Bender Ck Big Lake Ck 7002/ 1,550 1,000 350 8 10 11 146 - 0 2 20 17 MWF Doolittle Ck T lSR14WSec23CD 85 640 50 2 Elk Ck T 2SR18WSec 4DB 86 571 96 T lSR18WSec33BD 86 528 208 - Goris Gulch T lSR14WSec 8AA 86 - Very little flow - no fish Johnson Ck T lSR17WSec25AA 86 597 103 39 T lSR17WSec 5CD 86 534 110 Joseph Ck T 2SR18WSecl6BC 86 575 113 2 T 2SR19WSecl2BC 86 5 92 381 T 2SR19WSecllAA 86 150 153 2/ Kussigbrod Ck T lSR16WSec 9BA 86 700 3 Placer Ck T 2SR17WSecl6DA 86 650 17 3/ Plimpton Ck T 1SR1 5WSec22BD 86 300 87 Rock Ck T 3SR15WSecl9CC 85 400 35 - - 33 Ruby Ck T 3SR1 8WSec25AD 85 1,000 10 T 3SR17WSec30BC 85 350 23 Sales fsky C T lSR14WSec 8CA 86 500 No fish captured Sheep Ck T 2SR18WSec 14BD 86 555 43 9 T 2SR18WSeollRtf 86 534 60 Squaw Ck T lNR14WSec2 7DC 85 500 70 - - - 10 Steel Ck T 2SR15WSec 1 5BD 86 1,400 7 - - 5 1 27 KWF T 2SR15WSuc34AB 86 85 5 180 - - 2 5 17 MWF T 2SR15WSp.c34AB 85 600 149 - - 3 2 13 MWF t 3SR14WSrc 5CB 85 880 30 1 - - 6 2 HB 54 Table 19. (continued) Legal description Year Section length (ft.) __HsMober_ per 1.000 feet Stream 1/ EBT \ WCT RB GR LING OTHER Swamp Ck T 2SR15WSecl6CA 85 500 90 - - 4 22 4 MWF Tie Ck T 2SR17WSec 2BC T lSR17WSec34CA 85 85 500 350 56 146 - - ~ 10 3 — Trail Ck T lSR18WSec31AB T 2SR17WSec22D 85 85 1,000 500 184 12 - - - 17 14 4 MWF 1/ 2/ 3/ 4/ Abreviations for species are: EBT ■ eastern brook trout; WCT • westslope cutthroat trout; RB = rainbow trout; GR = arctic grayling; LING = burbot; and under the other - HB = hybrids between RB and WCT; MWF = mountain whitefish. These streams were electrof ished early in the spring when water temperatures were near 35° to 40° F. These values should be considered low due to the low efficiency. An additional 1,000 feet were electrof ished in an effort to capture grayling, but no grayling were found. A local resident claimed to have angled grayling out of the creek in the recent past. This portion of Steel Creek was within a channel of the Big Hole River which captured the lower 3.0 miles of Steel Creek. Several burbot were electrof ished, but not netted. 55 Table 20. Estimated fish populations in streams surveyed on the Beaverhead National Forest during 1985 and 1986. Population estimates calculated using a two-pass estimator (2P) (Seber and LeCren 1967), maximum-liklihood estimator (ML) (Van Deventer and Platts 1985), mark -recapture (MR) (Ricker 1975), and snorkel counts (SNOEK). Estimator Stream (Section Size range Reach length ft) Species (inches) Number Number Estimated 80 % per per population C.I. 1,000 ft acre DILLON DISTRICT Browns Canyon Cow Cabin 2 Governor 2 Morrison 2 Old Tim 2 Painter 2 Pass 2 Pole 2 3/ Reservoir 2 ML WCT ' 3.0 - 5.9 (300) 6.0 + ML EBT 3.0 - 5.9 (300) 6.0 + ML EBT 3.0 - 5.9 (1,375) 6.0 + 2P EBT 3.0 - 5.9 (300) 6.0 + 2P EBT 3.0 - 5.9 (300) 6.0 + ML WCT 3.0 - 5.9 (300) 6.0 + ML EBT 3.0 - 5.9 (300) 6.0 + ML EBT 3.0 - 5.9 (300) 6.0 + WCT 3.0 - 5.9 6.0 + ML WCT 3.0 - 5.9 (300) 6.0 + 41 40-44 137 850 23 23-24 77 477 5 6 7 3 12 17 5-6 6-7 "2/ 26 26-27 14 14-15 33 17 186 20 223 73 66-80 53 196 213 209-217 155 572 13 7-9 3-4 2-4 2/ "2/ 87 64 62-67 213 1387 16 15-18 53 347 87 353 47 190 33-34 . 110 622 14 - z/ 47 264 23 141 10 61 40 40 379 57 537 56 Table 20. (continued) Estimator Stream (Section Size range Reach length ft) Species (inches) Estimated 80 % population C.I. Number Number per per 1,000 ft acre WISDOM DISTRICT Elk 1 ML (571) EBT 3.0 6.0 + 5.9 48 15 47-502/ 84 26 398 124 2 ML (528) EBT 3.0 6.0 + 5.9 99 22 98-101 22-23 188 42 939 209 Doolittle 2 MR EBT 3.0 6.0 + 5.9 76 21 49-103 14-29 119 33 450 124 Johnson 1 ML (597) EBT 3.0 6.0 + 5.9 41 31 41-43 31-33 69 52 147 111 2 ML (534) EBT 3.0 6.0 + 5.9 69 14 63-76 14-15 129 26 313 63 Joseph 1 ML (575) EBT 3.0 6.0 -i- 5.9 44 39 41-49 38-41 88 78 381 336 2 ML (592) EBT 3.0 6.0 + 5.9 229 35 224-234 35-36 458 71 2660 407 Ruby 2 ML (1,000) EBT 3.0 6.0 + 5.9 12 6 6-18 6-7 12 6 31 16 Sheep 1 MR (555) EBT 3.0 6.0 + 5.9 15 12 15 12 27 22 149 119 2 MR (534) EBT 3.0 6.0 + 5.9 38 10 30-46 8-13 71 19 260 69 57 Table 20. (continued) Estimator Stream (Section Size range Reach length ft) Species (inches) Estimated 80 % population C.I. Number Number per per 1 ,000 ft acre Steel 1 Trail 1 ML (855) ML (880) SNOIK (500) ML (1,000) EBT EBT EBT EBT 3.0 - 5.9 270 201-339 316 531 6.0 + 62 61-66 73 122 3.0 - 5.9 33 28-41 38 172 6.0 + 12 11-15 14 63 3.0 - 5.9 11 _ 22 34 6.0 + 14 - 28 43 3.0 - 5.9 193 174-212 193 701 6.0 + 83 82-85 83 301 WISE RIVER DISTRICT Ad son 1 LaMarche 1 Meadow 2 ML WCT 3.0 - 5.9 (406) 6.0 + EBT 3.0 - 5.9 6.0 + MR EBT 3.0 - 5.9 (1,018) 6.0 + RB 3.0 - 5.9 6.0 + SN0J3C EBT 3.0 - 5.9 (1,184) 6.0 + MR WCT 3.0 - 5.9 (493) 6.0 + 7 3 7-13 3-4 17 7 132 56 207 171-243 203 279 84 68-97 83 113 50 29-71 49 67 6 6 6 8 222 — 188 314 189 ~* 160 267 8 1 7"f 2/ 16 2 79 10 58 Table 20. (continued) Estimator Stream (Section Size range Reach length ft) Species (inches) Estimated 80 % population C.I. Number Number per per 1,000 ft acre Mono 1 MR (496) WCT 3.0 - 5.9 6.0 + \ 31 6 20-42 . 63 213 - Z/ 12 41 Wyman 1 ML MR (621) MR (834) WCT EBT RBXWCT EBT 3.0 - 5.9 6.0 + 3.0 - 5.9 6.0 + 3.0 - 5.9 6.0 + 3.0 - 5.9 6.0 + 13 12-15 21 302 10 8-13 17 238 24 22-27 39 109 18 17-20 29 82 7 6-9 11 32 2 2 3 9 226 174-278 271 625 157 140-174 188 434 1/ 2/ 3/ Species codes are: WCT = westslope cutthroat trout; EBT = eastern brook trout; RB = rainbow trout; and RBXWCT ■ unidentifiable westslope cutthroat, rainbow trout and/or hybrids between these two species. All captured fish were captured on the first pass making a confidence estimate impossible. A complete second pass was not completed in Old Tim Creek. The catch in the second pass was estimated, based on the length of stream fished versus the total section length. 59 I believe the rainbow trout in R2 Wyman Creek originated from past releases of hatchery rainbow trout made into Lake of the Woods between 1941 and 1960. Evaluation of Different Estimation Techniques A comparison between depletion type (two or metre consecutive electo- fishing passes) and mark-recapture estimators found that depletion type estimates were lower than mark-recapture estimates , especially when capture probabilities (p) were less than 0.75 (Table 21). This result suggests that two-pass estimates may be underestimates and underestimation is more likely to increase as probability of capture decreases. All two-pass probability of capture estimates reported in Table 20 were 0.75 or higher except for brook trout 3.0 to 5.9 inches in R2 Governor Creek (0.68), cutthroat trout 6.0 inches and longer in R2 Pole Creek (0.67), brook trout 3.0 to 5.9 inches in R2 Johnson Creek (0.69), brook trout 3.0 to 5.9 inches in R2 Ruby Creek (0.55), brook trout 3.0 to 5.9 inches in Rl Steel Creek (0.38), brook trout 3.0 to 5.9 inches and 6.0 inches and longer in R2 Steel Creek (0.59 and 0.65, respectively), brook trout 3.0 to 5.9 inches in R2 Trail Creek (0.59), and cutthroat trout 3.0 to 5.9 inches and 6.0 inches and longer in R2 Mono Creek (0.55 and 0.57, respectively). It is likely that severe underestimates were made in any case where these probability of capture values were less than 0.60. A comparison between the depletion estimate and a snorkel count in R2 LaMarche Creek found that the snorkel count observed more fish than estimates made using the depletion estimator (222 versus 143 for brook trout 3.0 to 5.9 inches long and 189 versus 132 for brook trout 6.0 inches and longer). The capture probabilities for the depletion estimator were low (p = 0.25 and 0.47 for brook trout 3.0 to 5.9 inches long and 6.0 inches and longer, respectively) making the depletion estimate less reliable. An attempt was also made to mark fish using electrof ishing and conduct the recapture data using snorkel observation. It was difficult to observe the mark (a clipped dorsal lobe of the caudal fin) underwater, however, this type of approach appears to hold promise providing a tag or other mark which is easily applied and identifiable underwater can be found. Presence , of .".Sensitive Species" by District Tables 22 through 25 highlight the streams where westslope cutthroat trout and arctic grayling have been documented. Both these species have been classified as "sensitive species" by the Forest Service and are "species of special concern" within the state of Montana. Further quantification of the genetic status of suspected "pure" westslope cutthroat trout populations in five tributaries (Brown's Canyon, Reservoir, South Fork Steel, and Mono creeks) was made by sending seven to eleven fish to the University of Montana's Genetics Laboratory for electrophoretic analyses. From these analyses it appears that the cutthroat trout in both Brown's Canyon and Reservoir creeks are "pure" westslope cutthroat trout (letter dated September 1, 1986 from Robb Leary, Genetics Laboratory, University of Montana to Brad Shepard , MDFWP) . The cutthroat from Mono and Fox creeks were certainly introgressed cutthroat- rainbow trout populations. Introgression within the South Fork Steel Creek population was less clear and more sampling from this stream would be necessary to verify if this population has been introgressed with rainbow trout. 60 Table 21. Comparison of depletion (two or more consecutive electrof ishing catches) and mark-recapture estimates and their associated efficiencies (probability of capture for depletion = p; and number of recaptures divided by the total number marked for mark-recapture) from estimates made during 1985 and 1986 in streams draining the Beaverhead National Forest. Stream Species Estimate Depletion Mark -recap Reach Size c~. (inches .ass 0 Depletion Mark -recap \\ (p) (R/M) Doolittle Ck 1/ EBT ' 1 _2/ 3.0 - 5. 9 76 - .33 6.0 + 15 21 .79 .30 LaMarche Ck 1 EBT 3.0 - 5 9 130 207 .45 .37 6.0 + 76 84 .67 .52 RB 3.0 - 5 9 26 50 .463/ .25 6.0 + 6 6 .50 Meadow Ck 2 WCT 3.0 - 5 9 7 8 .873/ .67 6.0 + 1 I 1.00 Sheep Ck 1 EBT 3.0 - 5 9 15 15 .88 .53 6.0 + 12 12 .92 .58 2 EBT 3.0 - 5.9 32 38 .79 .44 6.0 + 8 10 .89 .29 Wyman Ck 1 EBT 3.0 - 5.9 23 24 .61 .60 6.0 + 1? 18 .74 .50 RBXWCT 3.0 - 5.9 6 7 • 603/ .60 6.0 + 2 2 1.00 At 2 ' EBT 3.0 - 5.9 203 226 .24 .20 6.0 + 130 163 .60 .30 61 Table tS, (continued - footnotes) 1/ 2/ 3/ 4/ Species abbreviations are: EBT = eastern brook trout; WCT = westslope cutthroat trout; RBXWCT = westslope cutthroat, rainbow trout, and hybrids between the two species. \ No estimate was possible because the same number of fish were captured in the first and second electrof ishings . No estimate of p* possible because all captured fish were captured on the first electrof ishing. More than two electrof ishing passes were made for the depletion estimate. 62 Table 22. Relative numbers and population estimates (where available) of westslope cutthroat trout (identified from external morphological character- istics) ranked from highest to lowest catch per unit effort (CPUE from one electrofishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Dillon District, Beaverhead National Forest surveyed during 1985-86. CPUE per 1.000 fa at Section Population Number Cutthroat Other length estimate per Year trout salmonids (ft) (80% CI) 1,000 ' Stream (reach) Browns Canyon Creek 86 177 - 300 Painter Creek 86 120 - 300 Reservoir Creek 86 97 - 300 Governor Creek (3) 85 28 123 325 Fox Creek (1) 85 13 171 550 Governor Creek (2) 85 b 168 1375 Thayer Creek 85 4 225 320 Andrus Creek 85 4 183 565 Pole Creek 86 h 24 300 64 (4) 214 40 (2) 134 29 (0) 97 8 (2) 2 (1) 63 Table 23. Relative numbers and population estimates (where available) of westslope cutthroat trout (identified from external morphological character- istics) ranked from highest to lowest catch per unit effort (CPUE from one electrofishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Wise River District, Beaverhead National Forest surveyed during 1985-86. CPUE PP.r 1.000 f*Pt Section Population Number Cutthroat Other length estimate per Stream (reach) Year trout salmon ids (ft) (80% CI) 1,000 ' Lacy Creek (2) 85 36 42 500 - - Mono Creek (1) 86 34 - 496 21 (1) 42 Meadow Creek (2) 85 28 _ 500 - - 86 14 - 493 9 (2) 18 Mono Creek (2) 86 24 — 700 28 (4) 40 86 9 - 300 - - Adson Creek (1) 86 12 5 406 7 (2) 18 86 4 11 275 - - Wyman Creek (2) 86 I1' 163 834 - - Harriett Lou Creek 85 3 - 300 - - It is probable that these cutthroat have hybridized to some extent with rainbow present in the Wyman Creek drainage. 64 Table 24. Relative numbers and population estimates (where available) of westslope cutthroat trout (identified from external morphological character- istics) ranked from highest to lowest catch per unit effort (CPUE from one electrofishing pass with a Coffelt BP-lC backpack electrof isher) in streams draining the Wisdom Districts Beaverhead National Forest surveyed during 1985-86. CPUE per 1 .000 feet Section Population Number Cutthroat Other length estimate per Stream (reach) Year trout salmonids (ft) (80% CI) 1,000 ' Doolittle Creek (1) 85 2 50 640 Steel Creek (2) 85 l1' 32 880 It is probable that the cutthroat in Steel Creek had hybridized to some extent with the rainbow in the drainage. 65 Table 25. Relative numbers and population estimates (where available) of arctic grayling ranked from highest to lowest catch per unit effort (CPUE from one electrofishing pass with a Coffelt BP-1C backpack electrof isher) in streams draining the Beaverhead National Forest surveyed during 1985-86. CPUE per 1 .000 fspt Section Population Number Arctic Other length estimate per Stream (reach) Year grayling salmonids (ft) (80% CI) 1,000 ' Rock Creek (1) Steel Creek (1) Fishtrap Creek (1) Swamp Creek (1) Big Lake Creek (1) Wyman Creek (2) LaMarche Creek (1) 85 33 35 400 86 51' 34 1400 85 3 149 600 86 2 197 855 86 5 675 200 85 4 94 500 86 2 27 1550 86 12/ 171 834 86 _3/ 153 1018 3 (2) The channel electrof ished was in lower Steel Creek where the Big Hole River has presently cut a side channel which has captured this lower portion of Steel Creek. 2/ It is believed that the grayling captured in reach 2 of Wyman Creek originated from Lake Odell stock. 3/ No grayling were electrof ished from the sample section in reach 1 of LaMarche Creek, however, several grayling were observed and angled from the lower 0.5 mile of the creek. 66 Big Hole River Population Estimate The total number of arctic grayling 8.5 inches and longer during late June was estimated to be approximately 35 fish per river mile in a 4.9 mile long sample section of the Big Hole River located immediately above the Highway 43 bridge at Wisdom (Table 26). This segment of the papulation represents all fish age 2 and older. An estimate of all fish 6.0 inches and longer yielded an estimate of 71 fish per mile. Only one fish larger than 13.0 inches was captured. Grayling appeared to be very susceptable to anglers. The incidence of easily recognizeable hooking scars was 10% for all grayling handled and increased to 15% for grayling longer than 10.0 inches. Approximately 6.7% of the rainbow trout handled and 1.5% of the brook trout handled had recognizeable hooking scars. It appears that the population of grayling in this portion of the river has declined since 1983 when Oswald (1984) estimated approximately 105 age 2 grayling per mile. Oswald (personal communication) estimated grayling numbers in the same section of the river during the fall of 1986. He had difficulty obtaining reliable estimates due to the initiation of fall downstream movements between his marking and recapture electrof ishings , however, using two estimation techniques, he estimated that this section contained somewhere between 51 and 98 grayling 6.0 inches and longer per mile. One noteworthy finding of Oswald's sampling during the fall was the presence of numerous age 0 grayling in several side channel areas which may indicate that the 1986 year class had better than average survival since in previous sampling very few age 0 grayling had been captured. The number of eastern brook trout 9.0 inches and longer was estimated to be 282 fish per river mile in this same sample section, while the number of brook trout under 9.0 inches was estimated to be 152 per river mile (Table 22). LENGTH, WEIGHT, AND CONDITION FACTOR Average lengths, weights (where reliable weight information vias available), and condition factors are presented in Appendix G. Length Frequencies. Length frequency data suggest that the upper reaches of tributaries are important spawning and rearing areas while the lower reaches support a higher percentage of adult and "catchable" size fish (Figures 6 through 8). The exceptions to this general rule appear in data for brook trout in LaMarche Creek and cutthroat trout in Mono Creek (Figure 7). The large number of brook trout less than 3.0 inches in Johnson Creek (30) probably occurred because this reach was sampled late in the year (August 27) after the young of the year were suseptable to the electrofishing. R2 of Wyman Creek supported a higher number of larger fish than Rl , probably because of the presence of more high quality pool habitats in R2 . Adult cutthroat trout in Mono Creek appeared to spawn and spend the early summer in R2, while the juveniles seemed to prefer the higher gradient and large substrate found in Rl • More brook trout in all length classes 6.5 inches and less were found in R2 of Joseph versus Rl . Rl appeared £.7 Table 26. Estimated populations of arctic grayling and eastern brook trout in the McDowell section of the Big Hole River above Wisdom, Montana in late June, 1986 using a mark -recapture estimator. Estimated Number Species Length class number (80% CI) per mile Arctic grayling 8.5 - 10.4 75 (39 - 111) 16 10.5 + 93 (62 - 124) 19 8.5 + 168 35 6.0 + 348 70 Eastern brook less than 9.0 741 (593 - 889) 152 9.0 - 11.9 1,170 (1,056 - 1,284) 239 12.0 + 209 (146 - 272) 43 9.0 + 1,379 282 68 BROWNS CANYON 20, 18 16 14 CUTTHROAT TROUT o p, 10 § 8 ■: 2 0 PAINTER CREEK - CUTTHROAT TROUT c u. 4 o CD 3 2 => z 2 RESERVOIR 20 CREEK - CUTTHROAT TROUT PASS CREEK - BROOK TROUT 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 LENGTH CUSSES (INCHES) (GROUPED BY HALF INCH CLASSES) 12.0 + 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 LENGTH CLASSES (INCHES) (GROUPED BY HALF INCH CLASSES) 12.0 + Figure 6. Length frequency histograms for fish captured in streams on the Dillon District during 1986. 69 LAMARCHE CREEK 40 35 30 £25 O *20 Lul m lis z 1 1 ■i F 10 I 5 4 1 0 A tt j > > 1 ^ 1 >_, TROUT LEGEND HI REACH 1 ! HI REACH 2 LAMARCHE CREEK - RAINBOW WYMAN CREEK - BROOK TROUT LEGEND H REACH 1 ■ REACH 2 20 IB 16 I * '* in C 12 U. I o fjj II 8 z 6 4} 2 0 MONO CREEK 40 T 35 30} x C 25 K 20 m §15 z 10 5 0 CUTTHROAT TROUT LEGEND H REACH HI REACH ik^ 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 12.0 + LENGTH CLASSES (INCHES) (GROUPED BY HALF INCH LENGTH CLASSES) 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 12.0 + LENGTH CLASSES (INCHES) (GROUPED BY HALF INCH LENGTH CLASSES) Figure 7. Length frequency histograms for fish captured in stream reaches sampled on the Wise River District during 1986. 70 ELK CREEK - BROOK TROUT LEGEND H REACH 1 1 REACH 2 1 M mn 1 1 , , , , i SHEEP CREEK 20 13 16 in tz 12 u. * 10 m § 8 z 6 4 2 a 80 70 SO x 53 sn C 50 BROOK TROUT LEGEND HI REACH 1 Hi REACH 2 JOHNSON CREEK - i -4 1 >— t— I 1 t— ( BROOK TROUT LEGEND H REACH 1 Hi REACH 2 wu JOSEPH CREEK BROOK TROUT LEGEND STEEL CREEK - BROOK TROUT REACH 1 REACH 2 n _» — pt-q — *l— t — «H — q «t q 80 r 70 SO x H 50 a40 UJ m § 30 t 20 10| 0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 12.0 + LENGTH CLASSES (INCHES) (GROUPED BY HALF INCH LENGTH CLASSES) 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 LENGTH CLASSES (INCHES) (GROUPED BY HALF INCH CLASSES) 12.0 + Figure 8. Length frequency histograms for fish captured in stream reaches sampled on the Wisdom District during 1986. 71 to be much more impacted by livestock than R2 , however, Rl had more high class pools than R2 and that was reflected in the numbers of captured fish larger than 6.5 inches. Length-Weight Relationships \ Regressions on log transformations of both length and weight values yielded length-weight predictive equations for grayling and rainbow trout from the Big Hole River and brook trout and cutthroat trout from all tributaries combined (Table 27). In general the predictive ability of these equations was good 5 however, use of the GM regression technique recommended by Ricker (1975) yielded somewhat better predictive capability, particularly for smaller fish (Table 28). FISH MOVEMENT Rainbow Trout Redd Surveys A preliminary survey of Big Lake, Bryant, and Jerry creeks were made during May, 1986 to document the presence of rainbow trout redds (spawning sites). Redds were observed in Big Lake and Jerry creeks (Table 29). No redds were observed in lower Bryant Creek. An angler captured a 19.0 inch rainbow at the mouth of Steel Creek on May 20, 1986 which may indicate Steel Creek is used for spawning. Further survey work is needed to confirm this possibility. Spring Trapping The spring trapping effort captured only three arctic grayling (Table 30) . All of these grayling were captured in the downstream trap located in Swamp Creek. Two were captured on May 18 and one was captured on May 21. The general fishing season opened on May 16 and it is likely these fish may have been caught and released by anglers fishing the opening weekend and moved downstream after their release. These fish ranged in length from 9.3 to 13.4 inches and all appeared to be males. None of these fish were obviously spent (spawned recently), however, the sexual condition of grayling is notoriously difficult to determine from external examination. One fish had died in the leads (it was believed a victim of angler catch and release) and internal examination found that it was a male with one half of the testes in a mature, but not ripe, stage. Below is a summary by creek. Big Lake Creek The upstream trap captured nine fish in 53 days of operation from April 4 to May 30 during which the leads remained up only 11 days. No game fish were included in the catch. The catch consisted of eight longnose dace (Rhinichthys cataractae) > one white sucker (Catostomus r.oTnmersoni) , and one mountain sucker (Catostomus platvrhynchus) . 72 Table 27. Regression equations for log transformations of fish length versus fish weight by species and water from data collected during 1986 in waters draining the Beaverhead National Forest. 2 Species Water Equation r GR ' Big Hole River Log(weight) = 2.78*Log( length)-3.21 0.97 RB Big Hole River Log(weight) = 2.66*Log( length)-3.05 0.99 EBT Big Hole River Log(weight) = 2. 79*Log( length)-3. 16 0.95 EBT Big Hole tribs Log(weight) = 2.34*Log( length)-2. 91 0.90 WCT Big Hole tribs Log(weight) = 2. 12*Log( length)-2.79 0.90 1/ Species abbreviations are: GR = arctic grayling; RB = rainbow trout; EBT = brook trout (charr); and WCT = westslope cutthroat trout. Table 28. GM regression equations (Ricker 1975) for log transformations of fish length versus weight by species and water from data collected during 1986 in waters draining the Beaverhead National Forest. Species Water Equation Log(weight) = 2. 82*Log( length) - 3.21 Log(weight) ■ 2. 68*Log( length) - 3.05 Log(weight) = 2. 86*Log( length) - 3.16 Log(weight) = 2 .46*Log( length) - 2.91 Log(weight) = 2. 23*Log( length) - 2.79 1/ Species abbreviations are: GR = arctic grayling; RB = rainbow trout; EBT = brook trout (charr); and WCT = westslope cutthroat trout. GRX/ Big Hole River RB Big Hole River EBT Big Hole River EBT Big Hole tribs WCT Big Hole tribs 73 Table 29. Redd surveys conducted during the spring of 1986 by stream, type of survey, date of survey, and number of redds located. Type of Date of \ Number of rgdds by size Stream survey survey Location Small Large Big Lake Creek Cursory 5-8-86 CM 4.5 Noted a few redds Bryant Creek Detailed 5-8-86 CM 0-1.0 No redds found Jerry Creek Detailed 5-8-86 CM 0-1.7 21 7 Steel Creek One 19.0 inch rainbow was caught by angler 5-20-86 74 Table 30. Summary of results of up- and downstream traps located in Swamp, Steel » and Big Lake creeks and fry drift nets located in Sandhollow Creek during the spring of 1986. Parameters Big Lake - Mountain whitefish - Other ' Streams- Steel 15 3/ Swamp UPSTREAM TRAPS i \ Date in 4-7-86 4-7- 86 4-23-86 Date removed 5-30-86 5-30 -86 6-25-86 Number days trapped 53 53 64 Number of days the leads were down 32 25 52 Total catch - Artie grayling 0 0 0 - Brook trout 0 2 0 Sandhollow DOVJNSTREAM TRAPS Date in 5- -6-86 4-7-86 5- -7-86 5-20-86 Date removed 5 -30-86 5-30-86 6 -25-86 6-2-86 Number days trapped 24 53 49 13 Number of days the leads were down 9 25 42 0 Total catch - Artie grayling 0 0 3 > - Brook trout 10 1 2 - Mountain whitefish - Other ' 8 1 1 0 43 15 11 34 1/ 2/ 3/ 4/ A day was considered to have had the leads down if the leads were found down during any time of the day. Other species included mountain suckers, white suckers, longnose dace and burbot. Also found numerous additional fry. One brook trout 5.2 inches long was captured. 75 The downstream trap captured 61 fish in 24 days of operation from May 6 to May 30 during which the leads remained up 15 days. The catch included ten brook trout ranging in length from 4.1 to 14.0 inches, eight mountain whitefish ranging in length from 4.5 to 12.1 inches, seven burbot, four longnose dace, 20 mountain suckers, and 12 white suckers. Mink predation was believed to be a problem at the trap site. Much of the downstream movement observed appeared to be a response to decreasing streamflows associated 'with the initiation of irrigation withdrawals. Sandhollow Creek The drift nets placed in this small intermittent stream captured 34 fry and one 5.2 inch brook trout during 13 days of operation from May 20 to June 2. One fry captured on May 27, 1986 appeared to be a recently emerged burbot fry which was approximately 0.6 inches long. The remaining fry were sucker fry. Steel Creek The upstream trap in Steel Creek (actually an east channel of the Big Hole River near Steel Creek) was operated for 53 days from April 7 to May 30, but the leads were down during 25 days of that operation. This trap captured two brook trout, 15 white suckers, two mountain suckers, two longnose dace, one mountain whitefish, two burbot, and 34 sucker fry. Several burbot were found in the leads. One large rainbow trout (19.0 inches) was measured from an anglers catch made on May 20, 1986. This rainbow trout was captured in the pool immediately below where Steel Creek and the east channel of the Big Hole River converge. The downstream trap operated for 49 days from April 7 to May 30 during which the leads were down 25 days. This downstream trap captured one brook trout, seven white suckers, two mountain suckers, 10 burbot, and one sucker fry. Swamp Creek The upstream trap in Swamp Creek was operated for 64 days from April 23 to June 25 during which the leads were down 52 days. The trap and leads had to be removed for 20 days (June 1 to 20) due to the extremely high spring streamflows. This trap captured two mountain whitefish, three mountain suckers, and three longnose dace. Two burbot were found stuck in the leads. The downstream trap was operated for 49 days from May 7 to June 25, but the leads were down 42 of these days (see explaination above for 20 of the days). This trap captured three arctic grayling, two brook trout, one mountain whitefish, two white suckers, nine mountain suckers, and one fry. A discussion of the grayling catch was presented above. Predation (probably by mink) was a problem at this trap site as partially eaten fish were found in the traps and against the leads. 76 Arctic Grayling Radiotelpmp.try Eight arctic grayling larger than 12.0 inches, two which had been radio tagged in early July, were captured during September electrof ishing. Condition factors which indicate length to weight relationship were compared between radioed and non-radioed fish to explore the possibility that the radio tags interferred with a fish's growth during the summer a These data suggest that condition factors for radioed fish were lower than non-radioed fish (Table 31). While this difference was significant (P < 0.05), the extremely small sample size may have influenced the findings. Of the seven grayling radiotagged on September 22, 1986, six were subsequently relocated. Five of these six relocations indicated the fish had moved in a downriver direction with the sixth fish moving 0.1 miles upriver. All relocated fish were found in deep pools or backwaters. The longest recorded movement was approximately 6.0 miles downriver in 11 days. None of the fish were positively relocated after October 21 even though the river was searched from five miles above the release sites to 61 miles below the release sites. The lower portions of two tributaries were also searched with no positive relocations. There was a problem with the reception of a good strong signal at all times. Often a single signal would be received with no subsequent signals. The telemetry history of each individual is fish is described below. Grayling number 1 (a female 13.2 inches long) was relocated one day after being tagged in the same pool she was released into. From September 23 to October 21 the areas searched with the receivers were all downstream from her location. She was relocated again on October 21 approximately 200 yards upstream from her release site in a large deep pool. She was not relocated again after October 21 even though the area upstream and downstream of that same pool were searched on October 28 and several searches were made from her last relocation downriver. The total documented movement for this grayling was approximately 0.1 miles upstream in 29 days. Grayling number 2 (a female 10.7 inches long) was recaptured via electrofishing one day after being tagged in the same pool she was released into. At that time she appeared healthy and behaved normally. On September 26 (four days after tagging) she was relocated approximately 350 yards downstream of her release site in a backwater off the main channel. She was relocated again in this backwater on October 8. She was not relocated again in this backwater in a search of this area on October 21. On November 20 received several signals on this channel at a deep hole immediately above Highway 43 bridge near Wisdom (approximately 200 yards below the last relocation). I could not confirm that this was grayling number 2 because even with repeated searching, I did not receive any more signals in this area. The confirmed movement of this grayling was approximately 0.2 miles downstream in 16 days. Grayling number 3 (a male 10.6 inches long) was never relocated after its release ■ Grayling number 4 (a female 10.4 inches long) was relocated on October 14 (22 days after being tagged) approximately 350 yards below her release site in a pool in the main river channel. She was relocated again in this same pool on October 21, but was not relocated again after that time even with repeated 77 Table 31. Comparison between the condition factors of two radio tagged arctic grayling versus six non-radio tagged arctic grayling approximately two months (July and August) after radio implants. Radio vs. Non-radio Length Weight Mean Sum of Fish number (sex) (in) (lb) Condition condition Squares Radio 1 (rj: N ■ 3.66/6 + 3.66/2 = 1.56 (.mean difference) t, . . - (32.67 - 29.57)/1.56 - 1.99 (sample) '(tabled) = 1'86° Therefore, reject the Ho and say that condition factor of radioed fish is poorer than non-radioed fish (0.025 < P< 0.05). 78 searching of the area around the pool of her last relocation and areas up- and downstream. The total documented movement of this grayling was 0.2 miles downriver in 29 days. Grayling number 5 (a male 12.8 inches long) was relocated in a deep pool approximately 150 yards downstream from where he was released the day after being released. He had moved approximately 200 yaitds further downstream on September 26 (four days after his release) and was found at the tail of another deep pool in or near a submerged accumulation of organic debris. On October 3 he was again relocated approximately another 200 yards downstream in another pool where two channels of the river came back together. This fish was not relocated again even though repeated searches of the area above and below his last relocation were conducted. On one subsequent search up- and downstream of this area (made on November 20) a single signal was received several times in areas downstream from the last confirmed relocation on October 3, but none of these signals could be confirmed as this fish's channel. The confirmed movement of this male was approximately 0.3 miles downstream in 11 days. He apparently moved in several stages downriver from pool to pool. Grayling number 6 (a female 11.9 inches long) was relocated in a pool immediately downstream from the pool she was released in the day after her release. She was relocated again at the head of this same pool on October 3 (11 days after her release) and at the tail of this same pool on October 14 (22 days after her release). She was not relocated afterwards even though repeated searches were made in this area. The same discussion of the single signals received on November 20 under grayling number 5 would also apply to this grayling. The confirmed movement of grayling number 6 was only 100 yards downstream over 22 days. Grayling number 7 (a female 11.3 inches long) was not relocated immediately after her release. She was only relocated once (on October 3, 11 days after her release) in a pool at the mouth of Steel Creek approximately six miles below her release site. She was not relocated again after that time even after repeated searches of the river and one search of lower Steel Creek. The total documented movement of this fish was approximately 6.0 miles downriver in 11 days. One final note: During a search of lower Swamp Creek (which enters the Big Hole River at river mile 108.9 or approximately eight miles below the release sites) on November 26 s 1986 a series of signals were received approximately one mile upstream from its confluence with the river on the receiver set up for receiving grayling numbers 4 through 7. No confirmation of channel (and therefore individual fish) could be made even after repeated attempts to receive subsequent signals to document which channel was transmitting. This area was searched for approximately 20 minutes with no further signals being received. Plastic Tag Returns. Tag return information indicated that grayling and rainbow trout moved very little during the course of the summer in 1986 (Table 32). The longest recorded movements were made by juvenile grayling. One moved 4.2 miles downstream out of Big Lake Creek into the Big Hole River between May 15 and 79 Table 32. Tag return information for fish tagged during 1985 and 1986 and recaptured during 1986 from waters in the upper Big Hole River drainage. Species (life stage) Tagging information Recapture,, information Net Tag number Date Length Location Date Length Location Movement Grayling (juveniles) RD 427 7-1-86 RD 437 5-21-86 RD 483 5-15-86 8.7 BHR(116) 9-22-8,6 9.0 7.9 Steel(0.8) 8-27-86 8.7 8.1 B.Lake(4) 9-4-86 9.2 BHR(116) 0 Steel(3.6) + 2.8 BHR(116) - 4.2 TOTAL TAGGED =37 TOTAL RECAPPED =3 8% RETURN Grayling (adults) WF 8653 6-25-86 11.4 BHR(119) 9-3-86 11.6 BHR(118) - 1.0 WF 8658 6-25-86 11.7 BHR(118) 9-3-86 11.9 BHR(118) 0 WF 8672 6-26-86 9.5 BHR(116) 9-3-86 10.2 BHR(116) 0 WF 8679 6-30-86 9.4 BHR(117) 9-3-86 9.5 BHR(117) 0 WF 8680 7-1-86 12.0 BHR(121) 9-22-86 12.2 BHR(121) 0 WF 8683 7-1-86 13.0 BHR(120) 9-22-86 13.1 BHR(120) 0 WF 8684 7-1-86 12.2 BHR(119) 9-3-86 12.3 BHR(120) + 1.0 WF 8686 7-1-86 9.5 BHRU18) 9-22-86 9.8 BHR(117) - 1.0 WF 8691 7-2-86 10.1 BHRU19) 9-3-86 10.4 BHR(120) + 1.0 WF 8693 7-2-86 9.1 BHR(116) 9-3-86 9.8 BHR(117) + 1.0 TOTAL TAGGED = 42 TOTAL RECAPPED - 10 24% RETURN Rainbow trout (adults) WF 8668 6-24-86 14.7 BHRU19) 9-4-86 14.9 BHRU21) + 2.0 WF 8674 6-30-86 13.0 BHR(121) 9-4-86 13.2 BHR(121) 0 WF 8675 6-30-86 17.2 BHR(121) 9-22-86 17.3 BHR(121) 0 WF 8677 6-30-86 11.9 BHR(116) 9-22-86 11.9 BHRU16) 0 WF 8681 7-2-86 13.0 BHRQ20) 9-4-86 12.6 BHR(118) - 2.0 TOTAL TAGGED = 8 TOTAL RECAPPED = 5 63% RETURN 80 September 4. Another moved approximately 2.8 miles upstream from the east channel of the Big Hole River below Steel Creek up into Steel Creek between May 21 and August 27. Tag return rates for juvenile grayling, adult grayling, and adult rainbow trout tagged during the spring and early summer of 1986 and recaptured within the same year were 8%, 24%, and 63%, respectively. \ RELATIONSHIPS BETWEEN HABITAT VARIABLES AND FISH DENSITIES Spearman Rank Corre la t-i nn.fi. Spearman rank correlations between habitat variables and brook trout densities indicated significant (P < 0.10) positive correlations between the density of brook trout 6.0 inches and longer and the percentage of high class pools and negative correlations between percentage of low class pools and frequency of small debris crossing the stream channel (Table 33). For stream reaches containing cutthroat trout significant negative correlations between the density of cutthroat trout 3.0 to 5.9 inches long and the percentage of large gravel (P < 0.05), amount of spawning habitat (P < 0.10), and channel ■ sinuosity (P < 0.10) (Table 34). The negative correlations between small cutthroat and spawning gravel variables is somewhat interesting and suggests that cutthroat rear in areas other than where they were spawned. A significant positive correlation (P < 0.10) was found between cutthroat 6.0 inches and longer and the percentage of small gravel and a negative correlation (P < 0.10) between these large cutthroat and stream order. For all cutthroat 3.0 inches and longer a significant negative correlation (P < 0.05) was found for both stream order and channel sinuosity. This suggests that cutthroat are more abundant in smaller, less sinuous (straighter) stream reaches higher in the drainages . Untransformed Habitat Variables vp.rsus Fish Densities Stepwise multiple regression was used to determine if habitat variables could be used to predict fish density. Results from this type of analysis can be used to help determine which habitat variables are important to fish density and to provide an equation whereby fish density could be predicted using habitat variables. This type of analysis will help in predicting impacts that land management activities might have on fish populations through changes in aquatic habitat. It must be remembered that the sample sizes are presently very low, eleven stream reaches for brook trout and six stream reaches for cutthroat trout. These small sample sizes make this type of analysis difficult and of questionable value at the present time. These results are being included to inform the readers of the study direction and illustrate how these data will be used to aid in the land management decision making process. Seperate regressions were conducted for each species by length class (3.0 to 5.9 inches, 6.0 inches and longer, and 3.0 inches and longer). In each analysis only the five variables which had the highest spearman rank correlation coefficients were regressed against fish densities. The exception to this rule was when two or more variables were obviously related to each other (ie. percentage high quality and percentage low quality pools), then one of those variables was dropped and the variable with the next higher spearman rank correlation coefficient was added. The results for brook trout were 81 Table 33. Spearman rank correlation coefficients between habitat variables and the number of eastern brook trout per acre by length class. Significance levels are P < 0.05 (**) and P< 0.10 (*) . Nineteen stream sections were sampled. The five highest correlations are in boldface type. Number of eastern brook trout per acre Habitat variables 3.0 - 5.9 inches 6.0 + inches 3.0 + inches Channel characteristics Wetted width - 0.1343 Channel width - 0.1272 Average depth - 0.2783 Thalweg depth - 0.1842 Instream cover - 0.3881 Bank characteristics Near shore depth - 0.0985 Bank angle - 0.0816 Soil alteration 0.1493 Percentage undercut bank 0.2098 Depth of undercut bank 0.2344 Vegetation stability - 0.2262 Vegetation use - 0.0595 Low overhead cover - 0.2525 High overhead cover - 0.2084 Canopy density - 0.0489 Streambed characteristics % Silt - 0.0739 % Sand 0.1279 % Small gravel 0.0221 % Large gravel 0.1407 % Cobble - 0.0485 % Small boulder - 0.0855 % Large boulder - 0.1995 Embeddedness 0.0035 Substrate score - 0.2886 D-90 0.1308 Habitat composition % Pools 0.1551 % Class V 0.2779 % Class IV - 0.0366 % Class III - 0.2636 % Riffles - 0.1214 % Puns - 0.1116 % Pocketwaters 0.0449 % Side channels - 0.2068 Organic debris Large debris frequency 0.2471 Freq. cross channel debris 0.1319 Small debris frequency 0.2840 Freq. cross channel debris - 0.1217 0.0913 0.0860 0.0751 0.0658 0.2020 0.0660 0.1389 0.1462 0.1300 0.0316 0.1499 0.0463 0.3684 0.2797 0.3137 0.1285 0.0578 0.0331 0.3127 0.0534 0.1684 0.3898 0.1635 0.3663 0.0518 0.1252 0.5695 * 0.2975 0.5818 * 0.1461 0.1248 0.1942 0.3508 0.1088 0.1437 0.1898 0.4630 * 0.0860 0.0649 0.2713 0.2351 0.2502 0.1336 0.1036 0.0939 0.1694 0.1172 0.1305 0.1189 0.2948 0.2848 0.0512 0.0097 0.0988 0.0909 0.2233 0.0115 0.1154 0.1926 0.0879 0.2975 0.1088 0.0678 0.3872 0.0961 0.3909 0.0699 0.0220 0.0968 0.3088 0.0765 0.0163 0.0530 0.1960 Amount of spawning habitat - 0.1324 Stream order — 0.3016 Channel gradient 0.0800 Channel sinuosity - 0.2452 0.3000 0.3552 0.0480 0.1675 0.2176 0.3627 0.0589 0.1169 82 Table 34. Spearman rank correlation coefficients between habitat variables and the number of westslope cutthroat trout per acre by length class. Significance levels are P < 0.05 (**) and P< 0.10 (*) . Eight stream sections were sampled. The five highest correlations are in boldface type. Number of westslope cutthroat trout per acre Habitat variable 3.0 - 5.9 inches 6.0 + inches 3.0 + inches - 0.4048 - 0.4524 - 0.4524 - 0.5714 0.1429 0.0238 0.1905 0.1667 - 0.2395 0.0838 0.2619 0.1429 0.3189 0.6088 0.0476 - 0.2143 0.5086 0.5911 - 0.2648 - 0.4414 - 0.1429 - 0.2381 . 0.1482 0.2224 - 0.1928 - 0.0723 - 0.2143 0.0952 - 0.4119 0.0588 0.1190 - 0.2381 0.0476 - 0.2857 0.6228 * 0.1078 - 0.0240 - 0.4192 - 0.2381 - 0.1190 - 0.0359 0.3234 - 0.5429 - 0.0857 0.1667 0.0476 - 0.5885 - 0.1471 - 0.2381 0.0714 0.5000 0.2381 0.0286 0.0857 0.4286 0.5429 - 0.4286 - 0.5429 - 0.5030 - 0.1078 - 0.0952 - 0.5476 0.0120 0.46 71 - 0.2619 0.0952 0.0599 0.3234 0.0732 0.3904 - 0.1905 0.1905 - 0.1091 - 0.0273 - 0.2381 - 0.3810 - 0.7307 * - 0.7937 ** 0.1667 0.5238 - 0.2667 - 0.7516 ** - 0.1304 - 0.4956 Channel characteristics Wetted width 0.0476 Channel width - 0.0952 Average depth 0.0476 Thalweg depth 0.3333 Instream cover 0.3234 Near shore depth - 0.0476 Bank angle 0.6667 Soil alteration - 0.1905 Vegetation stability 0.1650 Vegetation use - 0.0883 Percent undercut bank - 0.5238 Depth of undercut bank - 0.0247 Low overhead cover - 0.2771 High overhead cover - 0.2143 Canopy density 0.2942 Streambed characteristics % Silt - 0.2857 % Sand - 0.3810 % Small gravel - 0.2755 % Large gravel - 0.8264 ** % Cobble - 0.1429 % Small boulder 0.3353 % Large boulder 0.6571 Embeddedness - 0.2143 Substrate score 0.3825 D-90 0.3810 Habitat composition % Pools 0.0952 % Class V 0.3714 % Class IV 0.6000 % Class III - 0.6000 % Riffles 0.0719 % Runs - 0.5714 % Pocketwaters 0.4671 % Side channels 0.2143 Organic debris Large debris frequency 0.0719 Freq. of cross channel 0.1464 Small debris frequency 0.0000 Freq. of cross channel - 0.4364 Amount of spawning habitat - 0.6429 * Stream order - 0.4914 Channel gradient 0.3095 Channel sinuosity - 0.6547 * Channel type - 0.2087 81 Table 35. Stepwise multiple regession results for regressions of habitat variables against density of eastern brook trout by size class from data collected in drainages draining the Beaverhead National Forest during 1985 and 1986. Size class (inches) Independent variable Coefficient Standard error t -value Signif . level 6.0 + Constant 519.88 Small debris cross channel - 2.10 Substrate score - 19.79 189.32 2.746 0.025 0.85 - 2.479 0.038 12.23 - 1.619 0.144 R (adjusted) - 0.36 3.0 + Constant 297.34 237.84 1.250 0.243 Percent side channel 37.42 15.90 2.353 0.043 R (adjusted) = 0.31 84 disappointing. No clear predictive capability appeared between the habitat variables and fish density (Table 35). The results for cutthroat trout were more promising, but the small sample size presently limits the utility of these results (Table 36). Stream order and bank angle were the two variables selected as best predictors of total cutthroat trout (3.0 inches and longer) density.. Percentage of large gravel, bank angle, and channel sinuosity were the three variables selected as best predictors of small (3.0 to 5.9 inches long) cutthroat trout density. Test of COWFISH Model The COWFISH model was tested by comparing estimated number of catchable fish (fish 6.0 inches and longer) per 1,000 feet of stream for both brook and cutthroat trout with predicted existing and optimum number of catchable fish (fish 6.0 inches and longer) per 1,000 feet of stream and mean parameter suitability index (PSI) generated by the COWFISH model using habitat data collected at the same sample site. These comparisons were done using simple linear regression. The discussion above concerning the small sample sizes used in these analyses and limitations regarding these small sample sizes applies to these tests as well. Fifteen and four sample sites were tested for brook and cutthroat trout, respectively. It appears that the COWFISH model might have some utility when applied to streams containing cutthroat trout, but its applicability to streams containing brook trout appears limited (Figures 9 and 10). The highest r value for cutthroat (0.96) was obtained by using the predicted existing number of catchable fish per 1,000 feet of stream generated by the COWFISH data which is an encouraging result. One important factor to note is that the coefficient for the existing number of catchable fish per 1,000 feet of stream generated by the COWFISH model (or slope of the line) is 2.53 in the regression between estimated numbers of catchables and predicted existing numbers of catchables. This result indicates that the numbers of catchable fish predicted by the COWFISH model underestimates actual numbers by at least a factor of 2. £5 Table 36. Stepwise multiple regession results for regressions of habitat variables against density of westslope cutthroat trout by size class from data collected in drainages draining the Beaverhead National Forest during 1985 and 1986. Size class (inches) Independent variable Coefficient Standard error t -value Signif « level 3.0 - 5.9 Constant Percent large gravel Bank angle Channel sinuosity 485.81 207.02 - 2.347 0.143 - 5.78 0.82 - 7.080 0.019 5.28 1.36 3.878 0.060 289.17 133.40 2.168 0.163 R (adjusted) ■ 0.95 6.0 + Constant - 24.77 Percent small gravel 5.83 35.71 2.08 - 0.694 2.805 0.526 0.049 R (adjusted) = 0.58 3.0 + Constant Stream order Bank angle 1431.54 - 222.92 - 9.04 R (adjusted) ■ 0.91 373.66 41.57 3.84 3.831 0.031 - 5.363 0.013 - 2.352 0.100 86 200 1 6 0 Act u ■ I estimated number of catichable trook trout per 1,000 feel 120 0 o ACTUAL NUMBERS49.3 + 1.72JPREDICTE0 NUMBER) r * 0.23 <- 40 4- 4- 20 40 60 80 COWFISN predicted number of catchable brook trout per 1,000 feet (existing) Figure 9. Relationship between the actual number of fish estimated and the existing number of fish predicted by the C0WFISH habitat model at 13 sites sampled during 1986 which contained brook trout. 87 03 CO 80 Actual estimated number of cat enable cutthroat trout 60-- 40 ■• 1,000 feet 20 •- ACTUAL NUMBER =■ 1.8 1 * 2. S 3( PREDICTED NUMBER) r* « 0.96 -I 10 20 30 COWFISH predicted number of catchable cutthroat trout per 1,000 feet (existing) Figure 10. Relationship between the actual number of fish estimated and the existing number of fish predicted by the COWFISH model at four sites sampled during 1986 which contained cutthroat trout. LITERATURE CITED Andersons R.O., and S.J. Gutreuter. 1983. Length, weight, and associated structural indices. Pages 283-300 in. L.A. Nielsen and D.L. Johnson, editors. Fisheries techniques. American Fisheries Society, Bethesda, Maryland, USA. Burns, D.C. 1984. An inventory of embeddedness of salmonid habitat in the South Fork Salmon River drainage. Unpublished ^report , Payette and Boise National Forests, Boise, Idaho, USA. Chapman, D.G. 1951. Some properties of the hypergeometric distribution with applications to zoological sample censuses. University of California Publication of Statistics 1: 131-160. Crouse, M.R. , Callahan, C.A., Malueg, K.W., and S.E. Dominguez. 1981. Effects of fine sediments on growth of juvenile coho salmon in laboratory streams. Transactions of the American Fisheries Society 110: 281-286. Daniel, W.W. 1978. Applied nonparametric statistics. Houghton Mifflin Company, Boston, Massachussets , USA. Gould, W.R. 1986. Summer habitat utilized by arctic grayling in the Big Hole River, Montana. DRAFT Report to Montana Wild Trout Foundation, Bozeman, Montana, USA. Green, R.H. 1979. Sampling design and statistical methods for the environmental biologist. John Wiley and Sons, New York, New York, USA. Hop, H. 1985. Winter telemetry study of arctic grayling Thymallus arcticus (Pallus) in the Arctic National Wildlife Refuge, Alaska. Pages 224-231 in. Proceedings of the 65th Annual Conference of the Western Association of Fish and Wildlife Agencies held July 15-18, 1985 in Snowmass, Colorado, USA. Irving, J. S. and T.C. Bjornn. 1984. Effects of substrate size composition on survival of kokanee salmon and cutthroat and rainbow trout. Technical Report 84-6, Idaho Cooperative Fishery Research Unit, Moscow, Idaho, USA. Levings, J.F. 1986. Water resources of the Big Hole Basin, Southwestern Montana. Memoir 59, Montana Bureau of Mines and Geology, Montana College of Mineral Science and Technology, Butte, Montana, USA. Lloyd, J.R. 1986. COWFISH: Habitat capability model. Northern Region Fish and Wildlife Staffs, Fish Habitat Relationships Program, USDA, Forest Service, Missoula, Montana, USA. McNeil, W.J. and W.H. Ahnell. 1964. Success of pink salmon spawning relative to size of spawning bed materials. U.S. Fish and Wildlife Service Special Scientific Report - Fisheries 469, 15pp. Oswald, R.A. 1986. Southwest Montana fisheries study: Inventory and survey of the waters of the Big Hole, Beaverhead, and Ruby River drainages. Job Progress Report for Project Number F-9-R-34, Job Number I-b, Montana Department of Fish, Wildlife and Parks, Bozeman, Montana, USA. 89 LITERATURE CITED (continued) Platts, W.S., Megahan, W.F., and G.W. Minshell. 1983. Methods for evaluating stream, riparian, and biotic conditions. General Technical Report INT-138, USDA, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, Utah, USA. \ Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish populations. Bulletin 191, Fisheries Research Board of Canada, Department of Fisheries and the Environment, Fisheries and Marine Service, Scientific Information and Publication Branch, Ottawa, Canada. Rosgen, D.L. 1985. A stream classification system. Pages 91-95 in the Proceedings of the North American Riparian Conference held April 16-18, 1985 in Tucson, Arizona, USA. SAS Institute. 1985. SAS procedures guide for personal computers. Version 6.0 Edition, Cary, North Carolina, USA. Seber, G.A.F., and E.D. LeCren. 1967. Estimating population parameters from catches large relative to the population. Journal of Animal Ecology 36: 631-643. Shapiro, S.S., and M.B, Wilk. 1965. An analysis of variance test for normality (complete samples). Biometrika 52: 591-611. Shepard, B.B. 1986. Beaverhead National Forest and Montana Department of Fish, Wildlife and Parks Cooperative Fishery Program. First Annual Report covering the period August 15 to December 31, 1985, Beaverhead National Forest, Dillon, Montana, USA. Shepard, B.B., Fraley, J.J., Weaver, T.M., and P. Graham. 1982. Flathead River Fisheries Study - 1982. Study sponsored by the U.S. Environmental Protection Agency and conducted by the Montana Department of Fish, Wildlife and Parks, Kalispell, Montana, USA. Van Deventer, J.S., and W.S. Platts. 1985. A computer software system for entering, managing, and analyzing fish capture data from streams. Research Note INT -352, USDA, Forest Service, Intermountain Research Station, Ogden, Utah, USA. Vincent, R. 1971. River electrof ishing and fish population estimates. Progressive Fish Culturist 33(3): 163-169. Addendum. Progressive Fish Culturist 36(3) : 182. Witzel, L.D., and H.R. MacCrimmon. 1983. Embryo survival and alevin emergence of brook charr, Salvelinus fontinalis. and brown trout, Salmo trutta. relative to redd gravel composition. Canadian Journal of Zoology 61: 1783-1792. Zar, J.H. 1984. Biostatistical analysis. Second edition. Prentice-Hall, Inc., Englewood Cliffs, New Jersey, USA. 90 WATER CODES MDFWP WATER CODES AND KEYWORDS Stream Water code Stream Water code Adson Creek 02-0050 Andrus Creek 02-0125 Bender Creek 02-0375 Big Hole River 02-0475 Big Lake Creek 02-0500 Big Swamp Creek 02-0550 Browns Canyon Creek N.A. Bryant Creek 02-0800 Butler Creek 02-0925 California Creek 02-0950 Cow Cabin Creek 02-1400 Deep Creek 02-1625 Doolittle Creek 02-1750 East Fork Ruby River 01-2520 Elk Creek 02-2075 Fishtrap Creek 02-2200 Fox Creek 02-2275 Francis Creek 02-2325 Goris Gulch N.A. Governor Creek 02-2525 Harriett Lou Creek 02-2650 Jerry Creek 02-2950 Johnson Creek (D-2) 02-2975 Johnson Creek (D-3) 02-3000 Joseph Creek 02-3025 Lacy Creek 02-3150 LaMarche Creek 02-3175 Meadow Creek 02-3800 Mill Creek 01-5020 Mono Creek .02-4000 Morrison Creek 01-5120 Mussigbrod Creek 02-4150 N Fk Big Hole River 02-4275 O'Dell Creek 02-4375 Old Tim Creek 02-4400 Painter Creek 01-5640 Pass Creek 01-5700 Placer Creek 02-4625 Plimpton Creek 02-4650 Pole Creek 01-5940 Reservoir Creek 01-6200 Rock Creek 02-4900 Ruby Creek 02-5000 S Fk Blacktail Creek 01-7220 S Fk Steel Creek 02-5825 S Willow Creek 10-6880 Saginaw Creek N.A. Salesfsky Gulch 02-5075 Sandhollow Creek 02-5128 Sevenmile Creek 02-5275 Sheep Creek 02-5400 Squaw Creek 02-5900 Steel Creek 02-5950 Swamp Creek 02-6175 Thayer Creek 02-6287 Tie Creek 02-6350 Trail Creek 02-6450 Wise River 02-7025 Wyman Creek 02-7075 KEYWORDS arctic grayling, eastern brook trout, grazing, habitat, logging, rainbow trout, sediment, spawning, westslope cutthroat trout. 91