This article was downloaded by: [Montana State University Bozeman] On: 03 October 2011, At: 09:48 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Transactions of the American Fisheries Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/utaf20 Geomorphic Influences on the Distribution of Yellowstone Cutthroat Trout in the Absaroka Mountains, Carter G. Kruse a , Wayne A. Hubert a & Frank J. Rahel b a U.S. Geological Survey Wyoming Cooperative Fish and Wildlife Research Unit, University of Wyoming, Laramie, Wyoming, 82071, USA b Department of Zoology and Physiology, University of Wyoming, Larande, Wyoming, 82071, USA Available online: 09 Jan 2011

To cite this article: Carter G. Kruse, Wayne A. Hubert & Frank J. Rahel (1997): Geomorphic Influences on the Distribution of Yellowstone Cutthroat Trout in the Absaroka Mountains, Wyoming, Transactions of the American Fisheries Society, 126:3, 418-427 To link to this article: http:// dx.doi.org/10.1577/1548-8659(1997)126<0418:GIOTDO>2.3.CO;2

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and- conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 Transactions of the American Fisheries Society 126:418-427, 1997 €> Copyright by the American Fisheries Society 1997

Geomorphic Influence Distributioe th n o s f o n Yellowstone Cutthroat Trout in the Absaroka Mountains, Wyoming CARTER G. KRUSE AND WAYNE A. HUBERT U.S. Geological Survey, Wyoming Cooperative FishWildlifeand Research Unit1 University Wyoming.of Laramie. Wyoming 82071.USA FRANK J. RAHEL Department of Zoology and Physiology, University of Wyoming Laramie, Wyoming 82071, USA

Abstract.—Influences of large-scale abiotic, geomorphic characteristics on distributions of Yel- lowstone cutthroat trout Oncorhynchus clarki bouvieri arc poorly understood samplee W . d 151 perennia6 5 site n o s l stream thn i se Greybull-Wood river drainag northwestern ei n Wyomino gt determin effecte eth geomorphif so c variable Yellowslonn so e cutthroat trout distributions. Channel slope, elevation, stream size, and barriers to upstream movement significantly influenced the presenc absencd an e f Yellowstono e e cutthroat trout. Wild population f Yellowstonso e cutthroat trout were not found upstream of barriers to fish migration, at sites with channel slopes of 10% or greater, or at elevations above 3,182 m. Based on channel slope alone, logistic regression models correctly classified presence or absence of Yellowstone cutthroat trout in 83% of study sites. The addition of elevation and stream size in the models increased classification to 87%. Logistic models tested on an independent data set had agreement rates as high as 91% between actual and predicted fish presence. Large-scale geomorphic variables influence Yellowstone cutthroat trout distributions, and logistic functions can predict these distributions with a high degree of accuracy.

Yellowstone cutthroat trout Oncorhynchus clarki well 1995). Hanzel (1959), Behnke and Zarn bouvieri inhabit a larger geographic range than any (1976), Scarnecchi d Bergersean a n (1986)d an , other cutthroat trout subspecies except coastal cut- Varley and Gresswell (1988) noted that most ge- throat trout O. c. clarki (Varley and Gresswell netically pure population f Yellowstono s e cut- 1988). As in other interior stocks of cutthroat trout, throat trout occur in high-elevation headwaters in hybridization (Leary et al. 1984; Campion and Ut- remote areas. Additionally, Behnke (1992) sug- ter 1985; Gresswell 1995) and competition with geste e cutthroath d t trouy hav ma a teselectiv e exotic fishes (Griffith 1988; Young 1995) s wela , l advantage over nonnative trout in these areas be- as anthropogenic influences (logging, mining, ag- cause y thefunctioma y n bette coln i r d environ- riculture, irrigation; Hanzel 1959; Allendorf and ments. Leary 1988; Thuro . 1988al t we ; Gresswell 1995), Persistence of fish species in a stream system is have reduced the distribution of Yellowstone cut- influence abilite fise successfullth o th h t y f db yo y throat trou 70-90y b t % since settlemen Eury b t - use the habitat for cover, to acquire food, to in- opeans (Hadley 1984; Varle Gressweld yan l 1988; teract with othe e communityr th fisn i ho t d an , Behnke 1992; Young 1995). reproduce (Bozek and Hubert 1992). Constraints Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 Decline n Yellowstoni s e cutthroat trout distri- imposed by physical habitat can limit species per- butions have been most sever high-ordern i e , low- sistence; thus, identification of the abiotic factors elevation streams where human impacts are great- that limit species distribution s importanti s . Cat- est (Hanzel 1959; Gresswell 1988; Young 1995). astrophic events (Lamberti et al. 1991) and fish Limited acces remoteo st , high-elevation drainages migration barriers (Stube . 1988al t e r; Riemad nan and public ownership of most of these systems Mclntyre 1993) can have major impacts on fish have contribute protectioo dt f Yellowstonno e cut- surviva distributionsd an l . throat trout population theid an s r habitats (Gress- Habitat features that characterize trout distri- butions and abundance have been a common focus of research (e.g., Fausc . 1988al t he ; Nelso. al t ne e Uni jointlTh 1s i t y supporte e Universitth y b d f yo Wyoming, Wyoming Gam Fisd an eh Department, U.S. 1992; Riema d Mclntyrnan e 1995) s YellowA . - Department of the Interior, and Wildlife Management stone cutthroat trout populations continue to de- Institute. cline s imperativi t i , e that fishery manager- un s 418 GEOMORPHIC INFLUENCES ON TROUT DISTRIBUTION 419

derstand the environmental processes that control with high channel slopes, unstable substrates, and the habitat and, ultimately, fish populations in ar- large fluctuation dischargn i s e from sprin lato gt e eas where native fish still persist. Until recently, summer (Hanse Gloved nan r 1973). Elevationf so biologists have attempte characterizo dt e habitats streams in the study area range from 2,300 to 3,250 through traditional, small-scale approaches such as m above mean sea level. Channel slopes range stream reac r habitaho t unit inventories (Rieman fro mwit % 0.5 meah25 a %o t f 8.5% no , whics hi and Mclntyre 1995). But, as Nelson et al. (1992) generally considered steep (Kondol . 1991al t e f; and Rieman and Mclntyre (1995) pointed out, it Rosgen 1994). Snowmelt dominate annuae sth - hy l may be more important to understand trout distri- drograp d resultan h n extremeli s y high spring butio responsd nan habitao et t base spatialla n do y flows (Hanse d Glovean n r 1973; Martner 1982; large-scale, geomorphic approach. ZaffAnnead an t r 1992). Wetted stream widtf o h Trout habitat has been linked to large-scale geo- mainstreae th m Greybull River during late summer morphic variables (Lank . 1987al t ae ; Huberd an t glaciae varieth t a ls m headwaterfro 5 m3. 5 2 o t s Kozel 1993). Several researchers have shown that m at the boundary with measures of elevation, channel slope, and stream most tributary streams ranging 2.5- widthn i 5m . size are useful indicators of trout occurrence The volcanicly derived watershed (Keefer 1972) (Platts 1979; Lanka et al. 1987; Kozel and Hubert is steep and rugged with uplifted peaks and deep 1989; Nelso . 1992al t ne ; Riema d Mclntyran n e valleys, resulting in mountain streams that have 1995). Incidence functions, base thesn do e types steep longitudinal profiles and low biological pro- of variables, that predict presence or absence of ductivity. Stream substrates and banks are pre- trout populations can be useful tools, particularly dominately erodible volcanic material (Hansen in areas where little detailed habitat or population Gloved an r 1973; ZaffAnnead an t r 1992), which, information is available (Rieman and Mclntyre coupled with high spring flow steed an s p channel 1995). slopes, result in channels that shift regularly (Kent Although generalized distribution f Yellowo s - 1984), are strewn with large angular rocks, are stone cutthroat trou Wyominn i t e knowngar - ac , poorly defined d providan , e limited fish habitat. curate description f theio s r specific locatione ar s e GreybulTh l River, historic Yellowstone cut- not available. Several areas have been identified throat trout range s currentli , y managee th y b d e Wyominbth y g Gam d Fisan eh Departmens a t Wyoming Game and Fish Department as a native having high potentia contaio t l n remnant, geneti- sport fishery for cutthroat trout. Mountain white- cally pure Yellowstone cutthroat trout populations, including the Greybull-Wood river drainage in fish Prosopium williamsoni, mountain suckers Ca- northwestern Wyoming. tostomus platyrhynchus, longnose dace Rhinicthys e purposTh f thieo sdetermino t stud s ywa e eth cataractae, and brook trout Salvelinus fontinalis distribution of Yellowstone cutthroat trout in the are also present in the drainage (Yekel 1980). Rain- Greybull-Wood river drainag o determint d an e e bow trout O. mykiss were stocked in the drainage whether geomorphic features and stream size af- but are no longer present (electrophoretic analysis fect these distributions developee W . d logisti- cre failed to detect rainbow trout alleles in the pup- gression models based on large-scale geomorphic ulation). McBride Lakd LeHardan e y Rapids features and stream size to predict Yellowstone strains of Yellowstone cutthroat trout have been cutthroat trout distributions withi drainagee th n . periodically stocke n foui d r previously barren Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 stream reaches above fish migration barriers since Study Area 1985 (Kruse 1995). e GreybulTh l River, includin s primargit y trib- Anthropogenic influences on the Greybull River utary, the Wood River, is the third largest tributary watershed are relatively minor. Fishing pressure is e Bighorth o t n River t drainI . s more than 2,900 low becaus f limiteeo d access (Yekel 1980)e Th . km2 of the eastern in northwest- Greybull River drainage, especially the Wood Riv- ern Wyoming stude Th . y area included that portion er near Kirwin, has been prospected for minerals of the Greybull-Wood river drainage within the sinc e 1890seth . Mining activit highess wa y- be t Shoshone National Forest (Figure 1). Fifty-si- xpe tween 1892 and 1907, but viable operations con- rennial tributarie sf tota o (35 m l 5k strea m length) tinued inte 1960oth s (Hanse Gloved nan r 1973). 650-kme occuth n i r 2 headwater drainage. Becaus f limiteeo d human influence Greybule th , l The Greybull River and its tributaries tend to River drainage is considered to be essentially un- be torrential, high-elevation mountain streams impacted. 420 . KRUSAL T EE

Greybull River

Wood River

Meeteetse 34km FIGURE 1.—The upper Greybull River drainage withi Shoshone th n e National Forest (inset, blackened arean i ) northwestern Wyoming, east of Yellowstone National Park (inset, crosshatched area). Spots indicate streams on whic t leasha sample on ts taken ewa . Dashed line indicates Shoshone National Forest boundary.

Methods effects on fish presence. Barriers were defined as Study streams withi Greybule nth l River drain- geologic structures at least 1.5 m high (Stuber et e werag e selecte y usin b do criteria tw g : avail- al. 1988 r reache)o f verso y high channel slopr eo abilit f permaneno y t stream flow (base n U.Sdo . water velocity. Sampling progressed upstream un- Geological Survey [USGS] 1:24,000-scale topo- til trout were no longer present; then an additional graphic maps potentiad )an supporo t l trouta t pop- upstream s samplesitewa ensuro dt e fish absence ulation (Wyoming Game Fish Department, unpub- further upstream in the drainage and to obtain ad- Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 lished data). Perennial streams were sampled from ditional site information from areas where fish June to September 1994 with Smith-Root 12-B were not found. Sampling sites were grouped into programmable output waveform, battery-powered four categories: (1) fish present with no down- backpack electroshockers. stream barrie migrationo t r ) fis(2 ;h present above Samplin initiates gconfluencwa e th t da f eaceo h downstream barrier (s) to migration; (3) fish absent tributary within the Greybull and Wood river with no downstream barrier to migration; and (4) drainag progressed ean d upstrea mt approximatea - fish absent abov ebarriea fiso t r h migration. ly 1-km intervals. At each location, a one-pass e locationTh s (latitude, longitude elevad an ) - electrofishing run was performed across a 100-m tions (meters) of sampling sites and barriers to fish stream reach. Additional stream reaches were sam- migration were identified with global positioning pled immediately downstrea upstread man m from units and USGS topographic maps. Barrier heights permanent barrier fiso st h migration (Stube t alre . (nearest 0.1 m) and composition were recorded. 1988; Rieman and Mclntyre 1993) to assess barrier Wetted stream width (nearest 0.05 m) was mea- GEOMORPHIC INFLUENCES ON TROUT DISTRIBUTION 421

sured wit tapha e perpendicula streao t r m flot wa dicted by the logistic models and compared with four transects spaced equally within the 100-m observed values. Predicted probabilitie 0.5f so r 0o reach. Channel slope (%) was estimated with a greater indicated cutthroat trout were presentd an , clinomete e entirth r e fo r 100-m reach. Stream probabilities less than 0.50 indicated absence. lengths (kilometers) were measured wit elecn ha - Kappa values were tested to determine whether tronic map wheel from 1 :24,000-scale USGS to- trout classification logistiy b s c functions were sig- pographic maps. nificantly different than random classifications. Normal distribution date th a f alloweso f o e dus The kappa statistic (Titus et al. 1984) expresses a one-way analysis of variance (ANOVA) to test the proportio f siteno s correctly classifiee th y b d for differences in geomorphic and stream size fea- model after removing the effect of correct clas- tures amon foue gth r categories. When significant sification by chance (Beauchamp et al. 1992). differences were found, Tukey's multiple-compar- The logistic models from the Greybull River ison test was used to assess differences among drainage were applied to an independent data set categories. Pearson correlations were used to de- fro Wooe mth d River drainag teso et t model per- termine relations among channel slope, elevation, formanc n predictini e g fish presence. Predicted and wetted width (Krebs 1989). Logistic regres- probabilitie f 0.5so r greate0o r indicated probable sios use develono wa dt p incidence function- be s fish presence, whereas probabilities less than 0.50 binomiacause th f eo l naturdependene th f eo t vari- indicated fish absence. Statistical analyses were able (fish presence-absence d becausan ) t proi e - performed with SPSS/PC+ (SPSS 1991). Signif- vide a probabilistis c prediction- alloo in T .r wfo icanc determines l testsewa al 0.0 < r .5P fo t da dependent testins modele th wa date f t go th ,se a divided by subdrainage and the logistic models Results were developed with sites fro Greybule mth l River distributione Th f Yellowstonso e cutthroat trout drainage and tested on sites from the Wood River were describe usiny db g samples from 151 loca- drainage. Because barriers influence fish move- tions on 56 streams throughout the Greybull-Wood men distributionsd an t performee w , d logisti- cre river drainage. Cutthroat trout were presen5 1 n i t gression analysis for only those sites unaffected Greybul3 o3 f l River tributariesites9 8 f o ) 8 (5 s by fish migration barriers (i.e., sites below barriers and in 7 of 23 Wood River tributaries (31 of 62 or sites with Yellowstone cutthroat trout above sites). Cutthroat trout were sampled from 116 of barriers). Models were also developed withou (49%m k t f 6 perennia)o 23 l stream Greybule th n si l sites having gradients higher than 10%. The lo- River drainage and from 44 of 119 km (37%) of gistic function fors mi streams withi e Wooth n d River drainage. Within the entire Greybull River drainage (16) % 45 ,0km P = e"/(\ + e"), of the 355 total km of perennial streams contained wher probabilit= P e f Yellowstono y e cutthroat cutthroat trout, but the remaining 55% lacked fish trout presence or absence, e = the inverse natural because of barriers to fish migration or inadequate linealogarithd u— an , r1 model f mo : habitat associated with stream gradients steeper than 10%. u = b\X\ wilo N d cutthroat trout populations were found where / = regression constant, bn = regression above fish migration barriers. Barriers to upstream coefficients, and X = independent variables. The migratio f Yellowstono n e cutthroat trout were Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 m maximum (—2) log-likelihood method was used found on 17 of the 56 study streams. Fish were to estimate regression coefficients (Hosmer an d1e 7th stream f preseno 5 1 sn i t blocke barriersy db ; Lemeshow 1989). in f thes1o 1, fis 15 eh wer ebase presenth eo t p u t Appropriatenes f eaco s h mode s evaluatewa l d of the barrier and absent upstream. The other four with the (-2) log-likelihood statistic (hereafter re- stream d recentlha s y been stocked with Yellow- log-likelihooe th ferres a o dt d statistic); statistical stone cutthroat trout above the barrier by the Wy- significanc s assessewa e d wit chi-squara h e test. oming Game and Fish Department. Introduced Reduction log-likelihooe th n si d indicate improved Yellowstone cutthroat trout upstream from barriers model fit because the statistic is analogous to the occupied 20 of 160 km (12.5%) of stream reaches residual sum f squareso linean si r regressiond an , where cutthroat trout were found. it measures observed value deviation from the Sites with cutthroat trout presen d channeha t l model (Hosme Lemeshod an r w 1989). Cutthroat slopes less than 10% and were at elevations of trout presence-absence probabilities were pre- 3,182 m or less. Elevations of the study sites 422 . KRUSAL T EE

TABLE 1.—Physical habitat characteristics for the four categories of cutthroat trout presence or absence. Means along a row not sharing a common lowercase letter indicate significant difference among categories (Tukey, P ^ 0.05).

Stream characteristic Fish present, Fish absent, Fish present Fish absent, and statistic no barrier barrieo n r above barrier barrier present P Channel slope {%) Mean 4.4 z 9.6 x 4.0 zy 6.7 y

ranged from 2,09 o 3,20t 6 , channem 6 l slopes correlated, makin t difficulgi determino t t e which ranged from 0.5% to 25%, and wetted widths characteristi e greatesth d ha ct influenc n prei e - ranged from 1.3 to 15.8 m (Table 1). Cutthroat dicting fish distributions (Figure 2). trout were e sitfouneon t wita d channeha l slope Logistic models were developed by using stream of 17%. However, vertical falls were situated only variables observe significantle b o dt y differen- be t 1.5 km upstream from this creek's confluence with tween sites where fish were present and absent the Greybull River, and only three adult cutthroat (Table 1). Models developed by using only channel trout were collected, suggesting this was not a res- slope classified 83% of the sites in the Greybull ident population. This site was dropped from the River drainage correctly when barrier influences logistic regression analysis. were not considered (Table 2). High correct clas- Mean elevatio s significantlnwa y lowe t sitea r s sificatio lowese th d tn an log-likelihoo d value- soc where Yellowstone cutthroat trout were present curred with the inclusion of channel slope, ele- without downstream barriers than at sites where vation measura d an , f streaeo m modele sizth n ei ; Yellowstone cutthroat trout were stocked above however, the stream size regression coefficient was barriers or where fish were absent (Table 1). El- not significan y modelan n i t. Mode 2 (Tabll ) 2 e evation did not differ among sites void of Yellow- containing channel slop elevationd ean , correctly stone cutthroat trout, with or without barrier log-likelihooa ef d - ha sitel d classifieal f san o % d d87 fects d sitean ,s with stocked cutthroat trou- up t value only slightly higher than model 3. All mod- stream from barriers. significand ha s el t chi-square value predicted san d Channel slopes were significantly higher at sites presence-absence of cutthroat trout with 80-87% where Yellowstone cutthroat trout were absent overall correct classification. without influence from barrier o fist s h migration Kappa values (Tabl ) indicatee2 d tha modell al t s Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 compared with both categories of fish presence. classified Yellowstone cutthroat trout presence sig- The proportion of sites occupied by Yellowstone nificantly better than chance correct classifica- cutthroat trout declined linearly with increasing tions. Classifications from all models were at least channel slope from 0% to 10%. 32% better than chance alone, and most models Stream size, indicated by mean wetted width, were more than 50% better. was significantly greater at sites where Yellow- All logistic models except d modehigha h 4 l stone cutthroat trout were present with no barrier rates of agreement between actual and predicted than at sites where Yellowstone cutthroat trout fish presence when teste siten do s fro Wooe mth d were absent wit r withouo h a barriet r (Table 1). River drainage (Tabl . Althouge3) h model 1, con- However, mean width was highly variable among taining only channel slope, had a higher log-like- all sites sampled. Channel slope, elevation, and lihoo lowed dan r overall correct classification than strea msignificand ha siz l eal t effect troun o s t dis- e highesth d modelhad3 ha t r t agreemeni o , 2 s t tributions; however, these variables were highly between actual and predicted fish presence (91%). GFOMORPHIC INFLUENCES ON TROUT DISTRIBUTION 423

3,400

3,200

3.000 •O 8 <

2,800 ••• > 2,600 Q) LJJ 2,400 vj*.*

2,200

2,000 0 2 5 1 10 25 30 Channel Gradient (%)

30 B

^ 20 0) ^ I15

10

10 15 20 Stream Width (m) FIGURE 2.—Presenc absenc d f cutthroao an ) ) e(O e(• t trou relation i t nelevatio) lo(A channed nan l slopd ean (B) channel slope and stream size. All observations are included.

Discussion Martner 1982; Lambert t ale i . 1991 n locallca ) y Geologic barriers to fish migration influence dis- decimate or extirpate trout populations. Swanson Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 tribution f Yellowstono s e cutthroate th trou n i t et al. (1987) suggest debris torrents are episodic Greybull-Wood river drainage e hypothesizeW . d events influency thama t e mountain streams every that barriers might isolate genetically pure Yel- 50-200 years. Therefore relativelo t e du , y short lowstone cutthroat trout from potential hybridiza- stream lengths above the barriers, poor habitat tion or competition; however, Yellowstone cut- conditions, and relatively common occurrences of throat trout did not appear upstream of barriers catastrophic events, it is probably difficult for Yel- except in the four streams where Yellowstone cut- lowstone cutthroat trou recoo t o persis t r t o - n i t throat trout were stocked above barriers f cutI . - lonize areas above barriers. throat trout naturally occurred above barriers, cat- Yellowstone cutthroat trout distributions in the astrophic events probably limited cutthroat trout Greybull-Wood river drainage were significantly persistenc n thesi e e areas. Flood flows, severe influenced by channel slope. Channel slope has drought (30-100-year cycles in ; been found to influence habitat (Chisholm and Mariner 1982), and debris torrents (Keefer 1972; Hubert 1986; Boze d Huberan k t 1992; Rosgen 424 KRUS ALT EE .

TABLE 2.—Logistic regression analysis results with significanc r parameterfo e modeld san , correct classificationd an , kappa statistic f geomorphio s stread an c m parameters predicitng Yellowstone cutthroat trout presenc r absenceo n ei northwestern Wyoming. Difference significane ar s

Classification % Model and Log- (present/absent/ Kappa parameter Value (SE) P likelihood overall) (P) Channel slope -0.485(0.12) <0.0001 Constant 3.840 (0.82) <0.0001 0.580 Model 1 <0.0001 72.44 93/62/83 (<0.001) Channel slope -0.488(0.15) 0.0012 Elevation -0.009 (0.00) 0.0004 Constant 27.830(7.17) 0.0001 0.688 Model 2 <0.0001 49.23 91/77/87 (<0.001) Channel slope -0.510(0.19) 0.0074 Elevation -0.009 (0.00) 0.0004 Mean width -0.064 (0.33) 0.8445 Constant 28.333 (7.66) 0.0002 0.688 Model 3 <0.0001 49.19 91/77/87 (<0.001) Elevation -0.008 (0.00) <0.0001 Constant 22.520 (5.46) <0.0001 0.530 Model 4 <0.0001 69.14 89/62/80 KO.OOl)

1994) and, therefore, distribution (Kozel and Hub- distributions in the study streams; however, ele- ert 1989) and abundance of trout (Chisholm and vation was also positively correlated with channel Hubert 1986 othen i ) r studies. Boze Huberd kan t n independena t no slop d an e t influenc n trouo e t (1992) showed that channel slope was significantly distribution. Because channel slope decreases with relate cutthroao t d t trout occurrenc high-elen i e - decreasing elevation, channel slop probabls ewa y vation systems. Cutthroat trout occupied locations e majoth r variable influencing distributions (Fig- in streams with higher channel slopes than did oth- ure 2). All study streams were at elevations above er species of trout, but they were not found ine classifieb n 209 ca s high-elevatio 0a d man n streams with channel slopes greater than 8%. More streams (Kozel and Hubert 1989). Bozek and Hub- than 60% of the stream sites in the Greybull River t (1992er ) found that elevatio a significan s wa n t basin had channel slopes considered to be steep predictor of cutthroat trout locations within high- (>4%, Rosge wern% 1994)16 e verd an ,y steep mountain streams of the central Rocky mountains. (>10% channel slope). Only one site where cut- Elevatio categorizee b n nca predictoa s da f cliro - throat trout were found had a channel slope greater mate (Bozek and Hubert 1992), which influences than 10%, suggesting that channel slope is limiting several stages of the cutthroat trout life history Yellowstone cutthroat trout distributions in (Gresswell 1995). If a portion of the life cycle is streams withi Greybull-Wooe nth d river drainage. incompatible with climatic conditions resulting In their discussio presencn no f buleo l trout Sal- from elevational differences, elevation woul- ddi veimus confluentus in relation to patch size, Rie- rectly impact fish distribution. Similarly, Nelson man and Mclntyre (1995) also observed a gradient et al. (1992) found that elevation, along with sub- Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 threshold of about 10%. strate embeddedness and streamflow, was a good Elevation significantly influenced cutthroat trout discriminatory variabl n distinguishini e g land classes (geologic), which in turn were related to TABLE 3.—Agreement (percent) between actua pred an l- trout distributions. dicted fish presence when testing logistic model siten so s Mean wetted widt s significantlhwa y greaten i r Wooe inth d Rive rl models) drainagal r fo 3 .4 e (N= sites where Yellowstone cutthroat trout were pres- ent without barrier influence than in both site cat- Number classified egories where fish were absent within the study Model correctly Agreement (%) area. Wetted width is a measure of stream size, but 1 39 90.7 it is also related to climate and stream energy be- 2 37 86.1 cause larger streams ten havo dt e lower gradients 3 38 88.4 and occu t lowea r r elevations (Boze Huberd kan t 4 32 74.4 1992). Furthermore, stream size affects the avail- GEOMORPHIC INFLUENCES ON TROUT DISTRIBUTION 425

ability and quality of physical habitat and abun- cutthroat trout presence (89-93%) than absence dance of trout species (Kennedy and Strange 1982; (62-77%; Table 2), suggesting that additional fac- Chisholm and Hubert 1986; Bisson et al. 1988; tors may limit the ability of Yellowstone cutthroat Kozel and Hubert 1989; Heggenes et al. 1991). trou survivo t t givea n i e n system. After removal Several studies have found stream size, discharge, f siteo s influence channey b d l slope greater than and variation in discharge to be important factors 10%, model prediction remained high, indicating influencing fish distributions and abundance (Min- that these variable e useb classifo t dn ca s y cut- shall et al. 1983; Kozel et al. 1989; Bozek and throat trout presenc absencd an e e acros a relas - Hubert 1992; Fausch and Northcote 1992; Nelson tively small range of channel slopes (1-9%). . 1992)al t e . However Greybull-Wooe th n i , d river Models including variables of elevation or drainage, stream sizs onl eha a ymino r influence stream sizslightld eha y lower agreement rates (Ta- cutthroan o t trout distributions. ; however3) e bl , they were actually better models Modeling trout distributions base n habitao d t as shown by lower log-likelihood values and high- parameter commoa s i s n approac understando ht - er correct classifications and kappa values (Table d managininan g g fisheries; however, most mod- 2). Examination of sites that were incorrectly clas- elin bees gha n don relativeln eo y small scales (e.g., sifie y modelb d 3 indicate1- s d other anomalies stream reach level). Additionally s Fausca ,. al t he present that influenced fish presence. Three th f eo (1988) pointed out, these type f modelso dife sar - four sites incorrectly classified by model 1 in the ficul realisticallo t t y appl fisherien yi s management Wood River drainage were influenced by either a becaus laca f f generalitkeo o r precisionyo . More nonpermanent fish barrier (large-woody-debris recently, with recognitio e linth k f o between n jam) to upstream movement or inadequate stream large-scale geomorphology and stream habitat, flow (sinking into stream bed). Thus, agreement distribution models have been developed on larger coul e increaseb d eliminatiny b d g unusual cases scales (Lanka et al. 1987; Nelson et al. 1992; Rie- fro date mth a set t anomalie,bu s suc thess ha d ean mad Mclntyran n e 1995) r exampleFo . , Rieman others (e.g., land use, upwellings, habitat changes) and Mclntyre (1995) found consistent relation- will complicate model application. ships between bull trout occurrence in a stream Additionally, it is important to remember, as system and patch size and stream width in south Nelso . (1992al t ne ) pointed out, habitat features central Idaho. affecting trout occurrence vary across broad land Whereas other models have employe a comd - classes. Without additional testin ground-trud gan - bination of geomorphic and stream size variables, thing, application of our models to areas other than our logistic functions indicated that fluvial Yel- volcanic watersheds within native range of Yel- lowstone cutthroat trout distribution n northi s - lowstone cutthroa t advised trouno s i t . western Wyoming can be predicted with a few geo- Platts (1979) and Parsons et al. (1982) have morphic variables (Table 2). Although channel shown that trout habitat on a stream reach basis slope is a large-scale variable, we actually mea- can be related to drainage basin geomorphology; sured channel slopsite-specifia n eo c scale; how- additionally, trout distribution a functio e ar s f o n ever, this variable can easily be approximated by the habitat. The ability to predict Yellowstone cut- using longitudinal stream profiles derived from throat trout distributions with some degre f cereo - USGS l:24,000-scale topographical maps. Com- e basitaintth f geomorphio sn o y c variables will

Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 parisons between channel slope estimates derived be useful for fisheries managers. These large-scale on site with a clinometer and from topographical models can be incorporated into geographic in- maps showed difference f les . Isaakso s(D tha% , n1 formation system coverages, allowing managers to Universit f Wyomingo y , personal communica- determine potential Yellowstone cutthroat trout tion). Channel slope correctly classified 83% of all distributions without extensive field surveys but sites sampled (Table 2), and in conjunction with with onl neeya r simpldfo e ground-truthing cen- elevation, classification increased to 87% when the suses. Models of this type provide managers pre- influence of barriers was removed. Although it is dictive and diagnostic tools allowing for efficient difficul o separatt t e independenth e t effectf o s watershed plannin managementd gan . channel slope, elevation, and stream size on Yel- lowstone cutthroat trout distributions (Figure 2), Acknowledgments channel slope appears to be the primary factor gov- We thank C. Ewers, K. Harris, C. Griffith, K. erning trout occurrence. Geomorphic variables Krueger . McLean . R William, M d r an assis, fo s - were generally bette t classifyina r g Yellowstone tance in collecting field data; and S. Mullner, R. 426 KRUSE ET AL.

Wiley . YoungM , . Zubi . R Zaffl r critD ,d kfo an -, por Montano t t a Departmen f Fisho t , Wildlifed an , ical review of the manuscript. This project was Parks, Helena. supported by the Wyoming Game and Fish De- Hansen, R. P., and F Glover. 1973. Environmental in- ventory and impact analysis at Kirwin area. Park partment and the U.S. Forest Service. County, Wyoming, for America Climax, Inc. (AMAX). Rocky Mountain Center on Environment, References Denver. Hanzel, D. A. 1959. The distribution of the cutthroat Allendorf . LearyF . R . W.d F ,1988 an , . Conservation trout (Salmo clarki) in Montana. Proceedings of the and distribution of genetic variation in a polytypic Montana Academy of Sciences 19:32-71. species, the cutthroat trout. Conservation Biology Heggenes . Northcote,G J.. . PeterT , A d . an . 1991. Sea- 2:170-184. sonal habitat selection and preferences by cutthroat Beauchamp . Christensen W . Smith P . . J.J S ,. t E .d an , trout (Oncorhynchus clarki) a small n i , coastal 1992. Selection of factors affecting the presence of stream. Canadian Journa f Fisherieo l Aquatid san c brook trout (Salvelinus fontinalis) in Adirondack Sciences 48:1364-1370. lakes casa : e study. Canadian Journa f Fisherieo l s Hosmer . LemeshowS . d W.D ,an , . 1989. Applie- lo d and Aquatic Sciences 49:597-608. gistic regression. Wiley, Toronto. Behnke, R. J. 1992. Native trout of western North Hubert, W. A., and S. J. Kozel. 1993. Quantitative re- America. American Fisheries Society, Monograph lations of physical habitat features to channel slope , Bethesda6 , Maryland. dischargd an unalteren ei d mountain streams. Jour- Behnke. ZarnM d . J. an R ,,1976 . Biolog mand yan - naf Freshwateo l r Ecology 8:177-183. agement of threatened and endangered western Keefer . 1972 R geologie . Th W ., c stor f Yellowstonyo e trouts. U.S. Forest Service General Technical Re- National Park. U.S. Geological Survey Bulletin port RM-28. 1347. Bisson A. . . SullivanP , K , ,. Nielse anL . dJ n 1988. Chan- Kennedy, G. J. A., and C. D. Strange. 1982. The dis- nel hydraulics, habitat used bodan , y- forju f mo tribution of salmonids in upland streams in relation venile coho salmon, steelhead cutthroad an , t trout to depth and gradient. Journal of Fish Biology 20: in streams. Transaction Americae th f so n Fisheries 579-591. Society 117:262-273. Kent . 1984R , . Fisheries management investigationn si Bozek, M. A., and W. A. Hubert. 1992. Segregation of the upper Shoshone River drainage, 1978-1982. resident trout in streams as predicted by three hab- Wyoming Gam d Fisan eh Department, Fish Divi- itat dimensions. Canadian Journa f Zoologo l : y70 sion Completion Report, Cody. 886-890. Kondolf, G. M., G. F. Cada, M. J. Sale, and T. Felando. Campion, D. E., and F. M. Utter. 1985. Natural hy- 1991. Distributio d stabilitan n f potentiao y l sal- bridization between steelhead trout (Salmo gaird- monid spawning gravels in steep boulder-bed neri) coastad an l cutthroat trout (Salmo clarki clarki) stream eastere th f so n Sierra Nevada. Transactions Pugeo itw n t Sound streams. Canadian Journaf o l Americae oth f n Fisheries Society 120:177-186. Fisherie Aquatid san c Sciences 42:110-119. Kozel, S. J., and W. A. Hubert. 1989. Factors influ- Chisholm, I. M., and W. A. Hubert. 1986. Influence of encing the abundance of brook trout (Salvelinusfon- stream gradient on standing stock of brook trout in tinalis) in forested mountain streams. Journal of the Snowy Range, Wyoming. Northwest Science 60: Freshwater Ecology 5:113-122. 137-139. . ParsonsG Kozel . . HubertA M . . J.d S .W ,, an ,1989 . Fausch . Parsons . HawkesG L . . D.. M K , C , d . an ,1988 . Habitat features and trout abundance relative to gra- Models that predict standing cro f streapo m fish dient in some Wyoming streams. Northwest Science from habitat variables: 1950-85. U.S. Forest Ser- 63:175-182. vice General Technical Report PNW-213. Krebs,C.J. 1989. Ecological methodology. Harper Col-

Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011 . NorthcoleFauschG . T . D.d K ,an , . 1992. Large woody lins, New York. debris and salmonid habitat in a small coastal Brit- Kruse. 1995G . C ,. Genetic purity, habitat popud an , - ish Columbia stream. Canadian Journa Fisherief o l s lation characteristics of Yellowstone cutthroat trout and Aquatic Sciences 49:682-693. Greybule inth l River drainage, Wyoming. Master's Gresswell . E.R ,, editor. 1988. Statu managemend san t thesis. Universit f Wyomingyo , Laramie. f interioo r stock f cutthroaso t trout. American Fish- Lamberti, G. A., S. V. Gregory, L. R. Ashkenas, R. C. eries Society, Symposium 4, Bethesda, Maryland. Wildman, and K. S. Moore. 1991. Stream ecosys- Gresswell, R. E. 1995. Yellowstone cutthroat trout. tem recovery following a catastrophic debris flow. Pages 36-54 in M. K. Young, editor. Conservation Canadian Journal of Fisheries and Aquatic Sciences assessmen r inlanfo t d cutthroat trout. U.S. Forest 48:196-208. Service General Technical Report RM-256. Lanka, R. P., W. A. Hubert, and T. A. Wesche. 1987. Griffith, J. S. 1988. Review of competition between Relation f geomorphologo s streao yt m habitad an t cutthroat trou othed an t r salmonids. Pages 134-140 trout standing stoc n smali k l Rocky Mountain i/i Gresswell (1988). streams. Transactions of the American Fisheries So- Hadley . 1984K , . Status Yellowstone reporth n o t e cut- ciety 116:21-28. throat trout (Salmo clarki bouvieri) Montanan i - Re . . Allendorf. W L . . F . PhelpsLeary R K , . F d S . , an R ,, GEOMORPHIC INFLUENCES ON TROUT DISTRIBUTION 427

Knudsen. 1984. Introgression between westslope SPSS. 1991. The SPSS guide to data analysis for SPSS/ cutthroa rainbod an t w trou thn i te Clark Fork River editiond 2n P , C+ . SPSS Inc. Chicago. drainage, Montana. Proceeding e Montanth f o s a Stuber, R. J., B. D. Rosenlund, and J. R. Bennett. 1988. Academy of Sciences 43:1-18. Greenback cutthroat trout recovery program: man- Mariner, B. 1982. Wyoming climate atlas. University agement overview. Pages 71-74 in Gresswell of Nebraska Press, Lincoln. (1988). Minshall, G. W., and six coauthors. 1983. Interbiome Swanson coauthorsx si d . J.F an ,, . 1987. Mass failures comparison of stream ecosystem dynamics. Eco- and other processes of sediment production in Pa- logical Monographs 53:1-25. cific Northwest forest landscapes. Pages 9-38 in E. . JensenPlaits. S Larsen. E . . P Nelson. S W D ,, d . L. ,an . R , . CundyW . T . Sal ,O d editorsoan . Streamside man- 1992. Trout distributio habitad an n relation i t o nt agement: forestr fisherd yan y interactions. Univer- geology and geomorphology in the North Fork sit f Washingtono y , Institut f Foreso e t Resources, Humboldt River drainage, northeaslern Nevada. Seattle. Transaclion Americae th f so n Fisheries Society 121: Thurow . K. CorsiV .E Moored . . F.R C an ,, . 1988. 405-426. Status, ecology d managemenan , f Yellowstono t e Parsons, M. G., J. R. Maxwell, and D. Heller. 1982. A cutthroat trout in the upper Snake River drainage, predictive fish habilal index model using geomor- Idaho. Pages 25-36 in Gresswell (1988). phic parameters. Pages . Armantrout 85-9B . N 1n i , Titus, K., J. A. Mosher, and B. K. Williams. 1984. editor. Acquisition and utilization of aquatic habitat Chance-correlated classification for use in discrim- inventory information. American Fisheries Society, inate analysis: ecological applications. American Western Division, Bethesda, Maryland. Midland Naturalist 111:1-7. Platts. 1979S . W , . Relationships among stream order, Varley, J. D., and R. E. Gresswell. 1988. Ecology, sta- fish populations aquatid an , c geomorphologn a n yi tus, and management of the Yellowstone cutthroat Idaho River drainage. Fisheries 4(2):5-9. trout. Pages 13-24 in Gresswell (1988). Rieman. B. E.. and J. D. Mclntyre. 1993. Demographic Yekel, S. 1980. An evaluation of area fisheries in the d habitaan t requirement r conservatiofo s f bulo n l Greybull River-Wood River drainages, Park Coun- trout. U.S. Forest Service General Technical Report ty. Wyoming Game and Fish Department, Fish Di- INT-302. vision Administrative Report, Cody. . MclntyreRiemanD . J d . E.B an ,, . 1995. Occurrence Young . 1995K . M ,. Conservation assessmen inlanr fo t d of bull trout in naturally fragmented habitat patches cutthroat trout. U.S. Forest Service General Tech- of various size. Transaction Americae th f so n Fish- nical Report RM-256. eries Society 124:285-296. . AnnearC . T Zafft d . .J. an D ,,1992 . Assessmenf o t Rosgen, D. L. 1994. A classification of natural rivers. stream fishery impacts and instream flow recom- Catena 22:169-199. mendations related to the Greybull Valley project. Scarnecchia . BergensenP . E d . L.D an ,, . 1986. Pro- Wyoming Gam d Fisan eh Department, Fish Divi- ductio d habitaan n f threateneo t d greenbacd an k sion Administrative Report, Cheyenne. Colorado river cutthroat trouts in Rocky Mountain headwater streams. Transaction e Americath f o s n Received Apri . 19910 l 6 Fisheries Society 115:382-391. Accepted Novembe . 19910 r 6 Downloaded by [Montana State University Bozeman] at 09:48 03 October 2011