North American Journal of Fisheries Management
ISSN: 0275-5947 (Print) 1548-8675 (Online) Journal homepage: http://www.tandfonline.com/loi/ujfm20
Livestock Grazing, Golden Trout, and Streams in the Golden Trout Wilderness, California: Impacts and Management Implications
Roland A. Knapp & Kathleen R. Matthews
To cite this article: Roland A. Knapp & Kathleen R. Matthews (1996) Livestock Grazing, Golden Trout, and Streams in the Golden Trout Wilderness, California: Impacts and Management Implications, North American Journal of Fisheries Management, 16:4, 805-820, DOI: 10.1577/1548-8675(1996)016<0805:LGGTAS>2.3.CO;2 To link to this article: https://doi.org/10.1577/1548-8675(1996)016<0805:LGGTAS>2.3.CO;2
Published online: 08 Jan 2011.
Submit your article to this journal
Article views: 101
Citing articles: 36 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ujfm20 North American Journal of Fisheries ManaKenient 16:805-820. 1996 €' Copyright by the American Fisheries Society 1996
Livestock Grazing, Golden Trout, and Streams in the Golden Trout Wilderness, California: Impacts and Management Implications ROLAN . KNAPDA P Sierra Nevada Aquatic Research Laboratory, University of California Star Route 198. 1.Box Mammoth Lakes. California 93546;and Marine Science Institute. University of California Santa Barbara. California 93106.USA KATHLEEN R. MATTHEWS' U.S. Forest Sen'ice. Pacific Southwest Research Station Box 245. Berkeley. California 94701. USA
Abstract.—Impacts of livestock grazing on California golden trout Oncorhynchus rnykiss agua- bonita and their habitat were studied inside and outside of livestock exclosures in the Golden Trout Wilderness, California consecutivo tw n I . e years e majoritth , f streao y m physical charac- teristics showed large differences between graze ungrazed dan ddirectione areasth d an , f thesso e differences were consistent wit recovere hth f excloseo y d stream ripariad an s n areas from impacts caused by livestock grazing. Ungrazed areas consistently had greater canopy shading, stream depths, and bank-full heights and smaller stream widths than grazed areas. California golden trout were very abundant in the study sites; their densities and biomasses were among the highest ever recorded for stream-dwelling trout in the western United States. California golden trout density biomasd an r unipe s t area were significantly highe ungrazen i r d tha grazen ni d area thren si f eo four comparisons. Differences between grazed and ungrazed areas were less consistent when densit biomasd yan s were calculate basie th f strea so n do m length resultr .Ou s suggest that current level f livestocso k grazin degradine gar g the strea ripariad man n component stude th f yso meadows to the detriment of golden trout populations.
Grazin f domestigo c livestoc mose th s kti wide- review of conditions on U.S. Forest Service lands sprea dwestern i lan e dus n North America (Wagner also concluded that most riparian ecosysteme ar s 1978)westere th n I . n United States, grazing occurs in need of restoration (USGAO 1988). on the majority of federal lands, including national Livestock grazing directly affects three general forests, national wildlife refuges, lands adminis- components of stream and riparian ecosystems: tered by the U.S. Department of Interior's Bureau streamside vegetation; stream channel morpholo- of Land Management (BLM), and some national , includingy e watee shapth gth f ro e columd nan parks. In the western states, grazing affects 64 mil- streambank structure; and water quality (Platts lion hectares administered by the BLM and 53 1979; Kauffman and Krueger 1984). These im- million hectares administere e U.Sth y .d b Fores t pact alten populatioe sca rth n structur f resideneo t Service (Armou t alre . 1994). Impact f livestocso k fish, particularly salmonids (Platts 1991)- Al . grazin n streao g d ripariaman n ecosysteme ar s thoug e spatiath hd temporaan l l variabilitf o y widespread (for recent reviews Kauffmae se , d an n stream salmonids may often obscure any popula- Krueger 1984; Platts 1991; Fleischner 1994) and tion changes caused by land management practices are particularly acute in the arid western states. In (Hall and Knight 1981; Platts and Nelson 1988), arid climates, lush vegetation is often found only a recent review reported tha9 studie1 t f 1o 5s near stream corridors and as a result, livestock tend showed that stream fish were diminishee th n di to congregate in these areas (Roath and Krueger presenc f livestoco e k grazing (Platts 1991). 1982; Gillen et al. 1984). Although western ri- In 1993 and 1994, we conducted a study of graz- parian zones are the most productive habitats in ing impacts to streams in the Golden Trout Wil- North America (Johnson et al. 1977), at least 50% derness, California, using a series of livestock ex- of these ecosystems are degraded as a consequence closures 133,500-he Th . a Golden Trout Wilderness f livestoco k grazing (Armou . 1994)al t r e recen.A t (GTW) was created in 1977, in part to protect the habitat of the two subspecies of golden trout On- o whoT 1 m reprint requests shoul e sentdb . corhynchus mykiss ssppe LittlTh . e Kern golden 805 806 KNAP MATTHEWD PAN S
trou m.. O t whitei nativ s Littl2i e th o et Kern River 90% of the annual precipitation falls as snow (Ma- an currentls di y liste threatenea s da d species under remaindee r 1977)th jo d an , r mostly occurs during the federal Endangered Species Act (Behnke summer thunderstorms. 1992). The California golden trout O. m. agua- Livestock have grazed the area now contained bonita nativs Soute i th o het Fork Kern Rived ran within the GTW since at least 1860, and there are Golden Trout Creek (Behnke 1992) mosd an f , o t report f 200,00so 0 sheearee th an pi durin yeaga r its native range lies within the GTW. The basic between 1860 and 1890 (Felando, unpublished re- ecology of the California golden trout remains f 10,00o port d an )0 late cattlth e n ei 1800 s (Inyo poorly understood, although recent research on National Forest 1982). Past overgrazing has re- stream populations in the GTW shows that indi- sulte widesprean di d riparian degradation (Albert viduals are long lived, slow growing, and exist at 1982; Felando, unpublished report) d largean , - high densities (Knap Dudled pan y 1990) Cale .Th - scale restoration efforts have been implemented by ifornia golden trou s bee subjece ha t nth f muco t h the U.S. Forest Service (Inyo National Forest) dur- management interest becaus status it f s Caleo a s - pasine years0 gth 7 t . ifornia's state fish, its limited natural distribution, Mulke d Ramshaan y w meadow e currentlar s y and several perceived threat s viabilityit o t s - in , grazed by approximately 950 cow-calf pairs (in cluding introduction of nonnative brown trout Sal- 1993. 235 in Mulkey and 700 in Ramshaw; in mo trutta and habitat degradation caused by live- 1994. 235 in Mulkey and 730 in Ramshaw). Mulk- stock grazing. Becaus f theseo e threats Calie th , - ey Meado s wtypicalli y graze r severadfo l weeks fornia golden trou beins i t g considere r federadfo l in July and again in September. In Ramshaw Mead- listin threatenea s ga d species. ow, cattl generalle ear y trailed throug latn hi e July Although much attention has been focused on with only light grazing and are gathered into the damage cause y pasb d t livestock grazine th n i g meado a wee r f high-intensitwfo o k y grazinn gi . Felando(e.g.A W . T GT , , Inyo National Forest, October. unpublished report, 1982), very little is known fise hTh faun f thiao s watershe composes di f do about whether current level f livestoco s k grazing two native species Californie th , a golden troud an t e causinar g additional degradatio f streano d man Sacramento sucker Catostomus occidentalis. How- riparian ecosystems. Therefore, we quantified a se- ever, we encountered only California golden trout rie riparianf so , stream fisherd an , y variables inside durin r surveysou g . These population e selfar s - and outside three grazing exclosure addreso st e sth sustaining, are not subject to management activi- following questions: (1) Are stream and riparian ties (e.g., fish stocking), and experience very light habitat variables different between areas inside angling pressure. and outsid densite th f exclosurese o yAr ) (2 d ?an The Inyo National Forest constructed exclosures and biomas f goldeso n trout different betwee- nar in several GTW meadows in 1983 and 1991 to s insid ea outsidd ean f exclosureseo ? protect stream segments from grazing impacts. Our study sites were inside and outside three grazing Study Area exclosures in Ramshaw Meadow (2,660 m; Figure The GTW is at the southern end of the Sierra 1) and Mulkey Meadow (2,850 m; Figure 2). Cattle Nevada, California (1I8°I5'N, 36°22'W). This rarely trespass inside these exclosures (D. Hubbs, study was confined to the eastern portion of the Inyo National Forest, unpublishede w data) d an , GTW in the Inyo National Forest. This area was consider stream sections inside the exclosures to largely unaffecte y b Pleistocend e glaciation have been ungrazed since exclosures were con- (Odion et al. 1988) and is characterized by large structed. All stream reaches used in our study were subalpine meadows (up to approximately 7.5 km2). typical of those found in low-gradient meadows These meadows are found primarily along the (types C-4 and E-4 of Rosgen 1994). South Fork Kern a majoRive d an rr tributary, Mulkey Creek. Meadows are dominated by sage- Study Design brush Artemisia cana. but streamside zones are An inherent problem with studie- sex thae us t typically dominated by sedge Carex spp. and wil- closure o investigatt s e impactth e f livestoco s k low Salix spp. (Odio . 1988al t ne ; Sarr 1995). Over grazing is treatment (i.e., exclosure) replication. e mosTh t statistically robust study design would 2 Although the systematics of western salmonids re- incorporate numerous randomly placed exclo- mains controversial mose th te recenus e w ,t classifica- sures. Suc desigha difficuls ni achievo t t e because tiocommod nan n name describes sa Behnky db e (1992). limitee th f o d availabilit f grazeyo d sites with sim- LIVESTOCK GRAZIN GOLDED GAN N TROUT 807
Ramshaw Meadow
FIGURE I.—Map of Ramshaw Meadow. The upper exclosure is depicted by the dotted rectangle and the lower drift fence exclosure is depicted by the dotted line across the stream and meadow. Arrows show study sites inside and outside exclosures. Shaded are foress ai t surroundin meadowe gth . ilar site potentia cose constructinf th o t d an l g mul- agencies to protect particular stream sections from tiple exclosures (Platt Wagstafd an s f 1984)a s a ; grazing impacts (Rinne 1988). These studies fre- aware resultar e f verstudieeo w w , yfe s that have quently are based on comparisons of stream char- used such a design (Buckhouse et al. 1981; Kauff- acteristics insid outsidd e an singl a f eo e exclosure . 1983maal t ne ; Platt Nelsod san n 1985a). Much (Keller and Burnham 1982; Platts and Nelson more common are grazing studies that take ad- 1985b; Odio. 1988al t ne ; Kondolf 1993). How- vantag f exclosureeo s place lany db d management ever, statistical analyses base differencen do - in s
Mulkey Meadow
FIGURE 2.—Map of Mulkey Meadow. The exclosure is depicted by the dotted rectangle. Arrows show study sites below, inside, and above the exclosure. Shaded area is forest surrounding the meadow. 808 KNAP MATTHEWD PAN S
side and outside of single exclosures are generally affect only a very small portion of the exclosure. pseudoreplicated, making interpretations of dif- Study sites wer m0 belo10 e w (ungrazed sited an ) ferences problematic (Hurlbert 1984). Despite this 100 m above (grazed site) the upstream end of the flaothed wan r shortcomings inheren exclosuro t t e exclosure. The exclosure in upper Ramshaw studies (Rinne 1988), unreplicated exclosures of- Meadow was built in 1983. The ungrazed site was ten provid onle eth y avenue availabl stude th r yefo jus exclosurte insidth e f th lowee o d eth d ean en r of grazing impacts. Indeed, such studies provide grazed site was 100 m downstream of the exclosure bule th f availablko e informatio effecte th n f no o s (Figure 1). We located the ungrazed site at the livestock grazing on stream and riparian ecosys- exclosure fence lin upstreaem instea0 10 mo f t do tems (Plaits 1991). allow comparisons with data that had been col- The design of our study was similarly con- lecte t thida s locatio 1984n ni . strained by the availability of grazing exclosures. e exclosurTh Mulken ei y Meado s builwwa n i t Although we could have used our three exclosures 1991. We located study sites 100 m downstream as replicate treatments, we chose not to because of the exclosure (grazed site), 100 m above the the exclosures were placed nonrandomly (violat- downstream end of the exclosure (ungrazed site), ing assumptions underlying statistical tests) and and 100 m above the upstream end of the exclosure because of the low statistical power resulting from (grazed site) (Figure 2). a sample size of three. This low statistical power Stream physical characteristics.—We quantified would have increase likelihooe dth - f typdo er I eI stream physical characteristics and surveyed fish rore findin significano th ,n f o g t difference when populations within each study site during August a true difference exists resulta s chose A . w , o et 20-30, 1993, and August 16-24, 1994. All sites make separate grazed versus ungrazed compari- contained f streamt o eac12a m 5d h an ,sit e w e sons for each exclosure. measured characteristics along each of 25 transects e locatioTh f exclosureno s also constrainer dou spaced 5 m apart and arranged perpendicular to placement of grazed and ungrazed study sites. Al- stream flow (Simonson et al. 1994). At each tran- though it may have been statistically more appro- sect, we measured channel width, channel depth, priat o plact e e habitat transects alon entire th g e stream width, stream depth, bank-full height, bank stream length inside the exclosures and over sim- overhang, bank angle, and bank water depth. These ilar distances outside exclosures e werw , e pre- variables are potentially sensitive to land use ac- vented from using this design because the exclo- tivities, suc s livestoca h k grazing, that influence sures contained several very different stream chan- channel stability. We defined channel width as the nel types. Because comparison of stream and ri- cross section containing the stream that is distinct parian conditions across disparate channel types from the surrounding area due to breaks in the could obscure any differences in the sites due to general slope of the land, lack of upland vegeta- grazing, we reduced the spatial heterogeneity of tion changed compositioe an , th n i s f substratno e sites used in our grazed versus ungrazed compar- materials (Platt . 1983)al t e s . Channel widts hwa isons by locating study sites as close together as measured to the nearest 10 cm. Channel depth, the possible (only Rosgen's C-4 and E-4 channel types distance from the top of the channel to the water were included). This design reduced the spatial surface measures wa ,nearese . Streath o cm dt 5 tm scale of the study, however. width presene widte th th , f ho t water surfact no e Despit e shortcomingth e r studou f y o s design, including islands s measure wa ,nearese th o dt 5 t the exclosures represent the only means of as- cm. We measured unvegetated steam width, the sessing grazing impact streamn so Californid san a total stream width minus the width of any bands golden trout in the GTW. We believe our data con- of emergent vegetation along each bank, to the tribute meaningfull e managementh o t y f Calio t - . nearesThicm s 5 tvegetatio s typicalli n e th y fornia golden trout population d theisan r habitat. beaked sedge Car ex rostrata . utriculata).C = ( a species that rapidly colonizes stream margins in Methods the absence of grazing in Mulkey and Ramshaw Study sites.—Ramshaw Meadow containo tw s meadow play thad importann an sma ya t t roln ei grazing exclosures (Figur elowee 1)Th . r exclosure channel stabilization and stream narrowing (Ros- was built in 1991 to keep cattle from grazing the gen 1994; Sarr 1995). Unvegetated stream width entire lower portion of Ramshaw Meadow. Ap- was measured only in 1994. Although bank-full proximately 700 cattle are trailed through the ex- heigh s generalli t y defineheighe th s a dt reached closure every yea earln i rd lat an ye summet bu r by a stream on average very 1.5 years (Gordon et LIVESTOCK GRAZING AND GOLDEN TROUT 809
al. 1992), this measur s unlikeli e e stronglb o t y y size-fractions wit mechanicaha l shaker n 1994I . , influenced by land management practices. Instead, we measured substrate particle size distributions we defined bank-full height as the height above by measuring the substrate contacted by the up- e currenth t water leve t whica l e bankth h s lose stream edge of the wading rod base at each of the their abilit contaio t y e streath n m (Gordo. al t ne five equally spaced points along each transect. Par- 1992)s measure wa nearese t . i ;th cm o dt 5 t ticles 1 mm or larger in diameter were measured We quantified streambank morpholog meay yb - to the nearest millimeter. Particles smaller than 1 suring bank angle, bank overhang, and bank water mm were classifie s fina d e san r silo d t) (0.5mm depth of both banks along each transect. We mea- (0.1 mm). We used this technique in 1994 instead sured bank nearese anglth o usine° t 5 t clinomga - shovee th f o l sampler becaus difficulte th f eo f o y eter on a 1.5-m rod placed against the bank slope transporting heavy samples out of the remote study (Platt . 1983)al t se . Overhanging banks have bank areas and the time-consuming nature of the sifting angles of 0-89°, and laid-back banks have bank process. angles of 90-180°. If banks were overhanging, the To quantify the size-structure of riparian wil- extent of overhang was measured to the nearest 5 lows countee w , measured dan d height l wilal -f so fro m deepese c mth t bank undercu furthese th o t t t lows withi f eacm2 no h streamban t eacka h site point of bank protrusion. We defined bank water in 1994. Willow heights were measured to the depth as the water depth 15 cm from each bank nearest 1 cm. and measured it to the nearest 5 cm. Of the 14 measured variables, most were ex- To quantify instream characteristics, we mea- pected to respond relatively quickly and in a pre- sured vegetative canopy shading, substrate com- dictable direction to removal of livestock from the position, and, at equally spaced points along each streamside zone (Platts 1991). For these variables, transect, water depth, water velocity d heighan , t e predictew e followinth d g change ungrazen i s d of any aquatic vegetation. Canopy shading was areas inside exclosures relativ grazeo et d areas out- measured at each bank and in midstream by facing side exclosures. up and downstream with a densiometer (Platts et al. 1983). Water depth, water velocity, and height (1) Total stream widt d unvegetatean h d stream f submergeo d vegetation were measure0 1 t a d width will have decreased as vegetation in- equally spaced points along each transec 1993n i t , vaded the channel and banks stabilized. and at 5 equally spaced points in 1994. The re- (2) Bank-full height will have increased as veg- duced numbe f pointo r n 199i ss necessitate 4wa d etation colonized point bars and captured sed- by extremel floww lo y s that would have caused iment. points to be too close together (often <10 cm). (3) Canopy shading will have increased as willow Water depth and height of submerged vegetation and sedge species recolonized streambanks. were measure nearese . Wateth o cm dt 1 tr velocity (4) Bank angle will have decreased and bank over- s measurewa d wit a hMarsh-McBirney 3 model hang and bank water depth will have increased 2000 current meter, and each measurement rep- banks a s stabilize werd dan e transformed from resente da 10- s average. Velocitie t eaca s h point laid-back to undercut configurations. were measured at 60% of the water depth with a (5) Stream water depth will have increased as top-setting wading rod. stream width decreased and constricted stream In 1993, we quantified substrate particle size flow. distributions (geometric mean diameter—Dg of Several additional measured variables wer- eex Platts et al. 1979—and percent fines) by taking a pecte chango dt e very slowl responsn yi liveo et - core sample wit a shovee hdeepes th t a l t point stock exclosure or were not anticipated to change along each transect. The shovel was inserted into in a predictable direction. We included these in the substrate to a depth of 15 cm and then lifted our study to allow post hoc determinations of from the stream (Grost et al. 1991). Samples were whether sites insid outsidd ean excolsuref eo s were placed into resealable plastic bags for transport to similar. Large difference f theso l mosn eal i s r o t a laboratory, where they were dried for a minimum variables would suggest that paired sites differed of 72 h to a constant weight and separated into 11 before exclosures were built. Channel width and channel depth were expecte- respono re dt e th o dt 3 Trade name commerciad an s l enterprise mene ar s - mova f livestoco l k grazing thest bu , e changes were tioned solel r informationfo y endorsemeno N . y thb te likely to be measurable only over a period of sev- U.S. Forest Servic s impliedei . eral decades and not the 2-11 year time scale used 810 KNAPP AND MATTHEWS
in this study (Kondolf 1993). For two additional in pass 1), the number of fish from the second pass variables, substrate sizwated ean r velocityd di e w , in sections 1-5 (= total number of fish in pass 2), not predic directioy an t f changeno . Although sub- numbee th d f fisan o r h fro thire mth d pas f secso - strate size woul expectee db increaso dt e aftee th r tions 1-5 (= total numbers of fish in pass 3) (Van remova f livestoco l consequenca s ka reduca f eo - Deventer and Platts 1985). Capture probabilities tion in delivery of fine sediments to the stream, were similar among size-classes—in 1993 they livestock grazin s continuegha d upstreae th f mo were 0.50 for adults (> 100 mm) and 0.36 for age-0 exclosures. Therefore, eve finf ni e sediment inputs fish (55 mm) (paired /-test, N = 7, P > 0.09); in were reduced withi exclosurese nth expectee w , d 1994 they were 0.50.52d 9an , respectivel> P y( these changes to be masked by continued inputs 0.2)—an size-classel dal s were poolepope th r - dfo from upstream (Rinne 1988). ulation estimates. The number of fish per square Although we do not have preexclosure data for meter was calculated for each 125-m site by di- mos havtd sitesdi detaileea e w , (accuratp dma e vidin estimatee gth d numbee th f r fissit ro y hpe eb ) draw e cm strea n 198th i 0 n 1 f m 4o o t section stream surface area (average stream 5 widt12 X h inside the upper Ramshaw exclosure (T. L. Dudley m). The estimated total weight of fish in each sec- and R. A. Knapp, unpublished data). To determine tion was extrapolated from the mean weight of fish whether stream width insid e uppeeth r Ramshaw captured. We calculated fish weight per square me- exclosure had changed since the time of exclosure ter by dividing the estimated total weight per site construction in 1983, we compared the mean by the stream surface area. stream widt witp meae h ma hth fro e n mth strea m basie recena th f sn o O t revie f livestocwo - kim width at the site in 1993 and 1994. To extract total pacts on stream fish populations (Platts 1991), we stream widths fro e 198mth 4 map e drew ,5 w2 predicted that California golden trout density transects on the map, each perpendicular to stream (number/12 , number/m5m biomasd 2an ) s (g/125 flospaced wapartan m t deac5 A . h transecte w , m, g/m2) should be greater in ungrazed than grazed measured the total stream width with a ruler and areas. Although recent evidence shows that trout converted these measurements to their actual di- of some species move considerable distances f streaI . mm) t 1 no widtmension = d hha m c 1 ( s (Young 1995a thad )an t movemen confouny ma t d changed since 1984 (i.e., no recovery had oc- studie f habitat-relateso d difference troun i s t pop- curred) meae th , n stream widt s measureha d from ulation structure when site close sar eaco et h other the 198p shoulma 4 d have been similae th o t r (Young 1995b), other research indicates that adult stream width measured in 1993 and 1994. California golden trout rarely move more than a Fish population characteristics.—We surveyed few meters (Matthews 1996). In addition, numer- fish populations using standard electrofishin- gde ous grazing studies have documented differences pletion techniques (Van Deventer and Plans 1983). in trout populations in adjacent study sites inside To facilitate electrofishing, we divided each 125-m and outside exclosures (Platts 1991). site into five 25-m sections with block nets. We Statistical analysis.—To provide an overview of conducted three passes through each section with the differences between sites inside and outside a Smith-Root type XII electrofisher that produced the three exclosures e talliew , e numbeth d d an r 400 V and a pulse frequency of 90 Hz. The length direction of differences in physical stream char- f timo e thaelectrofishee th t s runninwa r eacn go h acteristics f livestocI . k grazing outside exclosures pass was similar within a section to ensure a sim- had not influenced stream characteristics, an equal ilar electrofishing effor n eaco t h pass. Captured number of changes in variables should have agreed fish were measure r fordfo k lengte nearesth o ht t and disagreed with expectations. Differences from 1 mm and weighed on an electronic balance to the equality were evaluated with the binomial test. To nearest 0.1 g. Fish were released into the section determine the magnitude of differences between from which they were captured afte finae th r l pass paired grazed and ungrazed sites, we also treated within the section was completed. uppee th r Ramshaw, lower Ramshaw Mulked an , y e numbeTh f fis o reacn i h h s estimatesitewa d exclosures as separate analyses. Most physical from the rate of depletion by maximum-likelihood stream variables could not be normalized for par- estimation techniques (Microfish, versio soft0 n3. - ticular sites, so we relied primarily on nonpara- ware; Van Deventer and Platts 1985). The deple- metric Kruskal-Wallis one-way analyse f vario s - tion dat ae maximum-likelihoo th use n i d d esti- ance (ANOVA) to test for differences in values mations were obtaine addiny db l fisgal h froe mth insid d outsidan e f exclosureso e . One r twoo - - first pas sectionn i s= tota ( 5 l 1- numbes f fiso r h tailed tests were used accordin e - nulth hy l o t g LIVESTOCK GRAZIN GOLDED GAN N TROUT 811
pothesis being tested. Because of the pseudo- TABLE I.—Difference strean i s m physical characteris- replication problems inherent in statistical com- tics in relation to predicted changes between paired sites parisons of single sites inside and outside of an inside (ungrazed) and outside (grazed) exclosures in 1993 and 1994. exclosure (e.g., artificially inflated sample sizes; Hurlbert 1984) presene w , t P-values r comparfo - Differ- ison f sitso e characteristics onl provido yt relea - Differ- ence ence opposite ative measure of the magnitude of differences, and consistent No from no drao t t w conclusions based solel statistican yo l Exclosure comparison with differ- predic- significance (P < 0.05). (ungrazed versus grazed) prediction ence tion Fish data were analyzed as estimated density Upper Ramshaw. 1993 6 1 0 (number/125 m and number/m2) and biomass Upper Ramshaw. 1994 8 () 0 2 Lower Ramshaw. 1993 3 2 2 (g/125 m and g/m ) per site with modified /-tests. Lower Ramshaw, 1994 7 0 1 estimatee Th d numbe coms f fisr 12wa ro -hpe 5m Mulkey (versus site below). 1993 7 0 0 pared between graze ungrazed dan d sites wite hth Mulkey (versus site below), 1994 5 0 3 Mulkey (versus site above). 1993 7 0 0 following /-test formula (Kelle d Burnhaan r m Mulkey (versus site above), 1994 8 0 0 1982): All. 1993 23 3 2 All. 1994 28 0 4 [(number) ) - / = V(SE,) (SE2+ 2)2 associated with comparisons of fish population subscript 1 refers to the grazed site of a pair, sub- characteristics onl provido yt erelativa e measure script 2 to the ungrazed site; standard errors were e magnitudth f o f differencedrao t o e t wno d an s calculated by the maximum-likelihood model. De- conclusions based solely on statistical significance grees of freedom were calculated with the formula: (P < 0.05). 1 var2 Results \n n df = ni Stream Physical Characteristics protectee th Alf o l d sites showed differencen i s var2 stream physical characteristics that were consis- tent with changes expected following the removal numben= variances i f sections o r va d .an , Fish of livestock from streamside zones. In 1993, of 28 weight per 125 m was compared between sites with comparisons between paired ungrazed and grazed the /-test formula: sites for variables that we predicted would change l(weighti (weight2)~ ) | in a particular direction, 23 (82%) differed in the / = predicted direction, 3 (11%) showed no change, V(SE X ,) (SE+ 2 'X )2 r wt 2 wt2 and 2 (7%) changed in directions opposite to what where (weigh (weightd t jestimatee an ) th e 2ar ) d s predictewa d (Tabl com2 e3 e 1)1994n -I th . f o , weight of fish in site 1 and site 2, (SEj) and (SE2) parisons made, 28 (88%) differed in the predicted standare th e ar d error (number]f so (numberd )an 2) direction and 4 (12%) changed opposite to pre- calculated by the maximun-likelihood model, and diction se differenc(TablTh . 1) e e betweee th n Xwti and Xwt2 are the average individual fish number confirmef so contradicted dan d predictions weight . Multiplyin2 d siten si an 1 s standare gth d was statistically significant in both years (1993, error by the average individual fish weight was 23 0.001< versu P ; ;2 s 1994 < versu8 ,2 P ; 4 s necessary to scale the standard error to fish weight. 0.001). Canopy shading and bank water depth Comparisons of the number of fish per square me- showed the greatest number of predicted differ- ter and weight of fish per square meter were made ences, showing difference l comparisonal n i s n i s with formulas simila thoso t r e given above except 1993 and 1994. Water depth, stream width, and that (number/ d (SE/) an d (weight,) an , d an ) bank-full height showed difference e preth -n i s (SE/-XWU) were first divided by the stream surface dicted directions in 75-100% of the comparisons, area of site / to scale these variables to area. d unvegetatean d stream widt d banan h k angle As with the habitat data, analysis of the fish data showed the predicted differences in 75% of the involved statistical comparisons of single sites in- comparisons. Bank overhang showe predictee dth d side and outside of an exclosure and are therefore differences in 50-75% of the comparisons. pseudoreplicated resulta s presene A . w , t P-values Of the variables with no predicted direction of 812 KNAPP AND MATTHEWS
change, only channel dept widtd han h showed con- exclosure (inside, 5-22 ; below0cm , 10-70 cm). sistent differences between insid outsidd ean e eth Also, willows 5-20 cm tall were abundant inside exclosures. Channel depth was shallower in grazed the exclosure but nearly absent below the exclo- sites in all of the 1993 comparisons and in 75% sure. of the 1994 comparisons. Channel width was greate graze n i re 199 th d 199 f d3an o 4areal al n si Lower Ramshaw Meadow comparisons. Substrate size, percent fine sediment, One of the four variables for which there was and water velocity showe o consistenn d t differ- no predicted directio f changno e showe da larg e ences between insid outsidd ean f exclosureseo . difference between sites insid d outsidan e e th e Stream physical habitat results were very similar lower Ramshaw exclosure (Table 2): substrate size between 199 d 1993an 4 (Tabl ; therefore1) e e w , was 59% larger above than inside the exclosure. present detailed stream physical characteristics foeighe rth f t O variables thaexpectee w t chango dt e each site for 1994 only. in a predicted direction as a result of livestock exclusion, seven showed difference e preth -n i s Upper Ramshaw Meadow dicted directions, and four of these differences foue Ith n r f 1994variableo e ,on whicr sfo h there were larger (Tabl% tha . 20 nUnvegetate 2) e d was no predicted direction of change (channel stream widt totad han l stream widthd weran 1 e3 width, channel depth, substrate size, and water ve- 19% narrower insid exclosure eth e than above eth locity) showe dlarga e difference (definee on s a d exclosure, respectively d streaan , m depth, bank exceeding 20%) between sites inside and outside water depth, and canopy shading were 30-50% the upper Ramshaw exclosure (Tabl : channee2) l greater inside than abov exclosuree eth e difTh -. width was 47% greater below than inside the ex- ferenc ban n oppositi e th k n angli s e edirectiowa n closuree variableth f O . s thae expectew t o t d of wha predictee w t d (bank angl larges e th ewa n ri chang predictea n ei d directio resula s na f live o t - ungrazed site), but this difference was small (5%). stock exclusion, all showed differences consistent uppen i s A r Ramshaw Meadow, willows were with changes expected afte cessatioe th r f liveno - much more abundant within the lower exclosure; stoc eighe th kf t o grazingdifference x si d ,an s were willow2 22 s were counted insid exclosure eth d ean larger than 20% (Table 2). Bank-full height, bank willow2 2 s were counted abov exclosure eth e (Fig- overhang d banan , k water depth were 75-100% ure 3B). Willows inside the exclosure also showed greater inside than outside the exclosure. Canopy a much wider size range than those outside th e shadin s 250wa g % greater insid e exclosurth e e exclosure (inside, 5-170 cm; above, 10-60 cm). than outside. In 1994. strea narrowem% widt34 s hwa r inside Mulkey Meadow the exclosure than outside (Tabldetermino T . e2) e e fouth rf o variable e On r whicfo s h thers ewa whether this difference was the result of a change no predicted directio f changno e showe da larg e (i.e., recovery) thaoccurred ha t d insid excloe eth - difference between sites inside and below the sure since 1984 e compare,w e streath d m width Mulkey exclosure (Table 2): water velocity was inside the exclosure obtained from the 1984 chan -highe% 40 r belo exclosure wth e than insidf O . eit nel map to the 1993 and 1994 field measurements the variables that we expected to change in a pre- of stream width. Stream width insid exclosure eth e dicted directio resula s na f livestoc o t k exclusion, in 198 s significantl4wa y wider than stream width five differe e predicteth n i d d direction d threan , e in both 199 199d 3an 4; 1993 (1984cm 5 , 2734 , 1 of these differences were larger (Tabltha% n20 e cm ; ANOV 1994cm 0 n log-transforme,A23 o d 2). Canopy shading, bank water depth, and stream data0.0001)< P ; . Difference strean i s m width depth were 25-33% greater inside than below the between 1993 and 1994 were not significant (Tu- exclosure. Three variables differe opposite th n di e key pairwise compariso f meanso n 0.05)> P ; . direction from our predictions, but all of these dif- Therefore, stream width had narrowed signifi- ferences were smaller (Tabltha% n. 20 Strea e2) m cantly inside the exclosure since 1984 (i.e., re- width and unvegetated stream width were 12 and covery was occurring). 10% wider inside the exclosure than outside, and Willows were much more abundant withie th n bank overhan s 17%gwa less insid exclosuree eth . exclosure; 264 willows were counted inside the comparisoe Inth physicaf no l characteristic- sbe exclosur d 11 ean willow s were counted beloe wth tween sites insid abovd ean exclosuree eth , three exclosure (Figure 3A). Willows inside covereda foue oth f r variable r whicfo s hpreo n ther s - ewa much wider rang f heighteo s than those beloe wth dicted direction of change showed large differ- LIVESTOCK GRAZIN GOLDED GAN N TROUT 813
TABLE 2.—Means (SEs) of stream physical characteristics outside (grazed) and inside (ungrazed) the upper Ramshaw, lower Ramshaw Mulked an , y exclosure n 1994si .
Percent Variable Grazed Ungrazed difference11 AT Pb Upper Ramshaw exclosure versus below exclosure Channel depth (cm) 61 (2) 69(2) 13 25 0.008 Channel width (cm) 899 (39) 613 (34) 47 25 <0.001 Substrate size (mm) 4.7 (0.4) 4.6 (0.4) 2 125 0.90 Water velocity (cm/s) 17(1) 19(1) 12 125 0.30 Stream width (cm) 309(19) 230(11) 34 + 25 V(A\MJ ———————————————————— 140 p Upper Ramshaw Exclosure 120 ^ 100 i —— i inside § 80 ^m below S" 60 £ 40 20 _n.nnnnn_ 0 20 40 60 80 100 120 140 140+ (B) 90 I_c)W€*r Ramshaw Exclosure 75 i —— i inside 60 0) ••i above I 45 30 15 • n. • P r- 1400 14 + 0 12 0 10 0 8 0 6 0 4 0 2 0 (C) 60 Mulkey Exclosure 50 40 below CD inside 30 above I 20 10 0 .I.I i . . 0 20 40 60 80 100 120 140 140+ Willow height (cm) FIGURE 3.—Size-frequency histograms of willows (A) inside and below the upper Ramshaw exclosure. (B) inside and abov lowee eth r Ramshaw exclosure ) below(C d ,an , inside abovd an ,Mulkee eth y exclosure. Frequencf yo willow s presentei s 20-cn di m size ranges. ences (Table 2). Above the exclosure, the channel (Figure 3C). The size range of willows inside and was 56% shallower and 78% wider, and substrate belo exclosure wth greates ewa r tha sizee nth range size was 26% larger, than inside the exclosure. Of abov exclosure eth e (inside, 5-14 below; 0cm , 10- the variables that we expected to change in a pre- ; above120cm , 5-60 cm). Small (5-2 wil) 0cm - dicted direction as a result of livestock exclusion, lows were much more abundant inside the exclo- all eight showed differences in the predicted di- sure than either abov r beloeo exclosuree wth . differencee th f rectionso x si sd wer,an e larger than 20% (Table 2). Bank water depth and stream depth Fish Population Characteristics were 100-110% greater inside than - abovex e eth Three electrofishing passes through each 25-m closure additionn I . , canopy shadin s 15 gwa time s section allowed the capture of nearly all fish in a greater and bank overhang was 5 times greater 125-m site. There were small differences between inside than above the exclosure. actua estimated an l d total fisr sectiow hpe lo d nan Willows were more abundant within the exclo- s arounSE d population estimates associE S e -Th . sure than either abov r beloeo exclosuree wth ; 124 ated with the estimated number of fish per site was willows were counted inside the exclosure, 14 estimate 3.6th ; range%f 7 o 199n = ei ,N 3( 2.3 - abov exclosuree eth belo0 7 exclosure d wth an , e 4.8%) and 2.8% of the estimate in 1994 (N = 7; LIVESTOCK GRAZIN GOLDED GAN N TROUT 815 ) 2.2(A 5 800 range, 1.3-6.5). The average capture probability Upper Ramshaw was 0.50 in 1993 (N = 7; range, 0.40-0.59) and (V 0.5; range7 199n 6i = ,4 N 0.41-0.63)( . C1M0 California golden trout were very abundant in 1 2.00 - 700 the study streams. Densities ranged from7 2. 1. o 3t fish/m2 (370-692 fish/125 m) and biomasses ranged from 15.8 to 21.2 g/m2 (3,186-6,779 g/125 1 1.75 600 jg m). Fish population characteristics were similar 3 between 1993 and 1994; therefore, we present only data from 1994 e outcom.Th f statisticaeo l anal- 1.50 yse f fisso h population characteristics dependen di part on whether density and biomass were calcu- XV X%7 XN> X/ lated on a unit-area or unit-stream-length basis .) (B ,.75 -—————————————— 500 When density and biomass were calculated on a Lower Ramshaw unit-area basis, California golden trout densitd yan biomass were significantl ^P ( y 0.05) highe- in r side than outside exclosure e fou thren th i s rf o e comparison t significantlno d san y differene on n i t comparison contrast(Figuren I . 5) , s4 , when den- sity and biomass were calculated on a unit-stream length basis, California golden trout densitd yan biomass were significantly higher inside than out- sid exclosure eth comparisone on n ei , significantly lower inside than outsid e comparisonon n ei d an , t significantlno y differen e remaininth n i t o tw g comparison uppee sth t (FigurerA Ramsha. 5) , s4 w exclosure, the number and weight of California golden trou r squarpe t e meter were significantly higher inside than below the exclosure (Figures . 5A)e numbe4A Th . weighd an r f r fis12o t pe h5 m. however, were significantly lower inside than belo exclosure wth e (Figure , 5A) numbee s4A Th . r and weigh f r fissquaro t pe h e meter were signifi- cantly higher inside than above the lower Ram- shaw exclosure, but the number and weight of fish per 125 m did not differ significantly between sites (Figure , 5B) 4B t sth .A e Mulkey exclosure, Cal- ifornia golden trout densities calculated on a unit- areunit-stream-lengtd an a h basis both were sig- Site nificantly higher insid exclosure th e e than below FIGURE 4.—California golden trout population density the exclosure, but not significantly different from calculated as number of fish/m2 (left-side bars) and num- the densities abov exclosure eth e (Figure 4C). Cal- f fish/12o r f streao be 5m m (right-side bars ) insid)(A e ifornia golden trout biomasses calculated on a unit- and belo uppee wth r Ramshaw exclosure insid) (B ,d ean area and unit-stream-length basis both were sig- above the lower Ramshaw exclosure, and (C) below, nificantly higher insid e exclosurth e e than above inside d abov an e ,Mulke th e y exclosure asterisn A . k between paired bars indicates thadifference th t stas i e - the exclosure but not significantly different from tistically significant (P s 0.05). those below the exclosure (Figure 5C). Discussion Our study was hampered by the small number other portion stude th f yso meadows should there- of exclosures available, their nonrandom place- for made b e e cautiously exclosuree th , s represent ment, and the resulting pseudoreplication of our the only possible means of quantitatively assessing sampling design (Hurlbert 1984). Although we ac- current levels of livestock grazing on California knowledge that extrapolations from our data to golden trout and their habitat at a time when such 816 KNAP MATTHEWD PAN S 7500 and Mulkey meadows strongly suggest thae th t observed difference e resulth f livestoce o t ar s k grazing. The consistent differences in numerous physical characteristics between inside and outside sites, including increased streamside vegetation, stream narrowin d deepeninggan d increasean , d bank stability e consistenar , t with recovery from grazing-induced damage (Platts 1991). Differ- ences were particularly large at the Mulkey and upper Ramshaw exclosures, where recovery from past grazin resultins gi mora n gi e confined, narrow stream (conversion from a type C to a type E chan- nel; Rosgen 1994) lined with abundant mesic and hydric vegetation. e magnitude basiOth th nf o s f differenceo e s insid outsidd ean e exclosure mear eacr ou sfo f - ho sured stream physical characteristics, vegetation apparently responded most rapidly to grazing ex- clusion recoverd an , f streambankyo channed san l morphology proceede slowea t da r rate. Kondolf (1993) reported similar result a subalpin n i s e meadow in the White Mountains, California. Mea- surements taken insid outsidd ean e24-year-ola d exclosure showed significant vegetative recovery, but no detectable recovery in stream channel mor- phology. Amon e mosth g t pronounced difference- be s tween protected and unprotected sites were the numbers and sizes of willows. The number of young (5-40-cm) willows was much higher inside than outside all exclosures, suggesting that live- stock grazin s i impeding g willow recruitment. Odion et al. (1988) also found that willow plant- ings had significantly lower survival outside than inside exclosures in the GTW, and reduction in 5^^ jfjf"*^° willow cover appears to be a common result of livestock grazing (Kauffman and Krueger 1984; Platts 1991). Therefore, although willow cure sar - Site rently quite sparse outside grazing exclosures in FIGURE 5.—California golden trout population bio- the GTW, thiresula e s b f 13rarit o ty 0 yearyma s mass calculated as fish weight (g)/m2 (left-side bars) and f livestoco knaturaa t grazinno ld attributgan f eo fish weight (g)/!2 f streao m 5m (right-side bars) (A ) these meadows. inside and below the upper Ramshaw exclosure. (B) in- Past attempts to establish willows from cuttings sidabovd ean lowee eth r Ramshaw exclosure) (C d an , below, inside, and above the Mulkey exclosure. An as- have resulte survivaw lo n di l evenW insidGT f eo terisk between paired bars indicates thadifference th t e grazing exclosures (Odion et al. 1988). However, is statistically significan ^P ( t 0.05). observee w dlarga e numbe f younro g willow- sin side exclosures. Taken together, these results sug- gest thamose th t t effective mean f establishinso g dat criticalle aar ydetermininn i neede d ai o dt e gth naturay b s i l W recruitmenwillowGT e th n i s t fol- statu f thio s s subspecies. lowing livestock exclusion. streae Oth f m physical habitat variables thae w t Stream Physical Characteristics measured, several were expecte chango dt e only Our comparison of stream physical character- very slowly (20-100 years) or we could not predict istics inside and outside exclosures in Ramshaw in which direction they would change inside versus LIVESTOCK GRAZIN GOLDED GAN N TROUT 817 outside exclosures. These variables were measured surements (e.g., number/5 , g/5m 0; Rinn0m e to allow post hoc determinations of whether sites 1988). Becaus pairer eou d sites were close together insid d outsidan e f exclosureo e s were similat a r and discharge t pairesa d sites should therefore eb the time of exclosure construction. Channel depth similar e founw , e o reasounit-volumn dus o t n e and channel width both showed consistent differ- measurements. Although both unit-area and unit- ences inside versus outside exclosures; channels length measurements are justified in study designs were generally more incised and narrower inside such as ours, unit-area measurements may provide exclosures. Although the differences in channel the more accurate portraya f grazino l g impacts. depth and width between some of our grazed and Trout population size is frequently limited by the ungrazed sites suggest that some aspectr ou f o s autotrophic food base (Murphy and Hall 1981; sites may have been different at the time of ex- Murphy et al. 1981; Hawkins et al. 1983). Because closure construction, differences in channel depth total autotrophic production should increase with and width alon e unlikelar e o accoune t y th r fo t increasing stream width but remain constant on a consistent difference othen i s r stream character- per-area basis, wide stream sections should have istics between inside and outside sites at all ex- largea r total autotrophic food base than narrower closures. Instead, these differences are much more stream sections consequencea s A . l otheal , r fac- likel resule yth f recovero t y from livestock grazing tors being equal numbee th , f fisunir ro hpe t stream inside exclosures. length should be an increasing function of stream Changes in stream physical characteristics sim- width, wherea e numbeth s f r fisunio r pe h t area ilar to those found in our study are reported in should be unaffected by stream width. Therefore, other studie f grazino s g impact n streao s d man unit-stream-length measurements are potentially riparian ecosystems. These include increased confounded by effects associated with stream streamside vegetation (Marcuson 1977 Veln ;Va - width, whereas unit-area measurements remove son 1979; Duff 1983; Platts and Nelson 1985a; these effects. Based on this reasoning and our re- Kondolf 1993), stream narrowing and deepening sults showing that California golden trout density (Duff 1983; Platt d Nelsoan s n - 1985a)in d an , and biomass per square meter were generally creased streambank stability (Kauffman et al. greater in ungrazed than grazed sites, we conclude 1983; Platt Nelsod an s n 1985b; Rinne 1988). that livestock grazing in the study meadows is hav- ing negative effect n Californio s a golden trout Fish Population Characteristics populations. Our comparisons of California golden trout den- Several other studies have also documentee dth sitbiomasd yan s show thamagnitude th t f difeo - negative consequence f livestoco s k grazinn o g ferences between insid outsidd ean e exclosures si trout populations (Platts 1991). For example, Mar- influenced by the method used to calculate these cuson (1977) found that the biomass of brown trout variables. When densit d biomasan y s were cal- was more than three times higher in an ungrazed culate numbee th s weighd da an r f goldeo t n trout stream section than in one that was grazed. Platts per square meter, densities and biomasses were (1981) reported that fish densities were more than generally significantly higher inside than outside 10 times highe streaa n i r m section subjec ligho t t t the exclosures. Whe e densitth n d biomasan y s grazin grazino n r go g relativ heavila o et y grazed were expressed as the number and weight per 125 section. Similarly, densitie f rainboo s w trout On- f streammo , however, there wer o consistenn e t corhynchus mykiss and brook trout Salvelinus fon- differences inside versus outside exclosures. The tinalis were higher in an ungrazed than in a grazed contrasting results obtained when fish density and stream section (Kelle d Burnhaan r m 1982)- Al . biomass were calculated based on unit area versus though these effects are generally attributed to re- unit stream lengta consequenc e ar h e difth -f eo duction qualite th n i sf physica yo l stream habitat, ferent stream widths (and therefore stream areas) cumulative effect f grazino s n furthegca r reduce inside and outside exclosures. trout population alteriny b s g stream discharg- ere Comparisons of fish populations are generally increasiny gimeb d san g water temperatures (Van made based on unit-area measurements (e.g., num- Velson 1979; Platt d Nelsoan s . nal t 1989e i L ; ber/m2, g/m2, kg/ha; Kelle d Burnhaan r m 1982; 1994). Platts and Nelson 1985b; Beard and Carline 1991; The results of our fish population surveys show Larscheid and Hubert 1992), but authors have also e thaeffect n spitth i t f f livestoco o es k grazing. used unit-volume measurements (e.g., g/m3; Platts California golden trout exist at extremely high and Nelson 1985a) and unit-stream-length mea- densities in Mulkey and Ramshaw meadows (1.3- 818 KNAP MATTHEWD PAN S 2.7 fish/m2). In comparison with salmonid densi- tional Forest consider these management options ties in 277 streams reviewed by Platts and Mc- to reduce impacts on stream and riparian ecosys- Henry (1988), the California golden trout popu- e GTWtemth e restoration i s. Th f theso n e eco- lations in our study sites were among the densest systems will increase meadow stability (Odiot ne ever reporte wester e trour th d fo n i t n United States . 1988)al , improve habita r nativfo t e California and wer orden ea f magnitudo r e more dense than golden trout, and enhance conditions for a wide average th e trout densitl ecoregional r yfo e th n si range of other riparian-dependent species (Johnson western United States (0.25 fish/m- 2; PlattMc d san et al. 1977; Szaro and Rinne 1988). Henry 1988). Becaus uncleas i t ei r whether Platts d McHenran y (1988) included age-0 fis thein i h r Acknowledgments density estimates made w ,same eth e comparison Phil Pister deserves accolades for his extensive after removing age-0 fish fror densitmou y cal- wor Californin ko a golde r bringinn fo trou d an tg culations. Densities of California golden trout in attentio theio nt r plight. Roland Knap s suppwa - Mulkey and Ramshaw meadows (1.2-2.0 fish/m2) ported during this researc cooperativa y hb e agree- remained among the greatest in the western United ment wit e U.Shth . Forest Service Pacific South- States. Biomass of California golden trout in our west Research Station (agreement PSW-92- study streams (15.8-21.2 g/m2 als)s i o amone gth 0041CA). The Fish Habitat Relationship Program highes t4 timerecorde3- s i s d highedan r thae nth throug regionae hth l U.Se officth f . o eFores t Ser- average trout biomas f stream o se wester th n i s n vice provided additional financial assistanced an , United States (5.4 g/m2; Platt d McHenran s y we are especially grateful to Lynn Decker, Pacific 1988). Southwest Region Fish Program Manager, for her support. We also acknowledge Donald Sada for Livestock Grazing Wildernessand Management his role in designing portions of the stream phys- r studOu y provides evidence that area Ramn si - ical habita d fisan ht sampling protocol thane W . k sha d Mulkewan y meadows graze y livestocb d k d BallardE Muckm Ji , , Randall Osterhuberd an , pooren i e ar r condition than areas inside exclosures Vance Vredenbur U.Se th f .g o Fores t Servicd ean and that this degradatio s negativelni y impacting Melanie Paquie Universitth f o n f Californiao y , California golden trout. If stream condition in our Davis, for assisting with the field work. Sara grazed site s representativi s conditioe th f eo f no Chubb, former Inyo National Forest fish biologist, streams subjecte grazino dt g throughouW GT e th t l HubbsanDe d , Inyo National Forest range con- (and we believe it is), our study raises the question servationist, were helpful in setting up the study. of whether such degradation is appropriate in a e commentTh f Sylviso a Mori (Pacific Southwest designated wilderness. Under the Wilderness Act Research Station Statistician) and four reviewers of 1964, "an area of wilderness ... which is pro earlien -a f o r versio manuscripe th f no t greatl- yim tecte managed preservo t dan s a naturas o s dit e l proved our presentation. Dennis and Jody Win- conditions and which . . . generally appears to have chester of Cottonwood Packers and their mules been affected primarily by the forces of nature, made wilderness research possible. with the imprint of man's work substantially un- noticeable" (Kloepfer et al. 1994). Although na- References tional t havforest authorite no eth o d s curtaio yt l Albert . 1982P . C ,surve A . f factoryo s influencine gth livestock grazing solely because of wilderness des- conditio streae th f mno zon Golden ei n Trout Wil- ignation (Kloepfer et al. 1994), they do have the derness. Master's thesis. Sonoma State University. authority to make changes in livestock grazing Sonoma, California. programs to reduce unacceptable impacts. Th . Elmore eW . Duff A d . an Armour,D . , 1994L. e . Th .C , effects of livestock grazing on western riparian and Inyo National Forest apparently recognize- im e sth stream ecosystems. Fisheries 19(9):9-12. pact of livestock grazing on stream and riparian Beard, T. D. J., and R. F. Carline. 1991. Influence of ecosystem GTWe th n t pasi s ,bu t effort rehao t s - spawning and other stream habitat features on spa- bilitate degraded strea reliemW habitatdGT e th n si tial variabilit f wilyo d brown trout. Transactionf so primarily on expensive structural remedies (Laituri the American Fisheries Society 120:711-722. et al. 1987; Felando, unpublished report) that have Behnke, R. J. 1992. Native trout of western North met with very limited success. One of the simplest America. American Fisheries Society, Monograph 6, Bethesda, Maryland. and most cost-effective means of reducing grazing Buckhouse, J. C, J. M. Skovlin, and R. W. Knight. 1981. impact reso t s ts i area reducr so e livestock numbers Streambank erosion and ungulate grazing relation- (Platts 1991), and we suggest that the Inyo Na- ships. Journal of Range Management 34:339-340. LIVESTOCK GRAZIN GOLDED GAN N TROUT 819 . Duff1983A . D .. Livestock grazing impact aquatin so c to livestock exclosure. White Mountains. Califor- habitat in Big Creek, Utah. Pages 129-142 in J. W. nia. Restoration Ecology 1:226-230. Menke. editor. Proceedings, workshop on livestock . ErlichLaituriF . . GiffenE R . . J..d R M ,, 1987an , A . and wildlife-fisheries relationships in the Great Ba- history of erosion control efforts in the Golden sin. Universit f Californiayo . Agricultural Sciences Trout Wilderness. Proceedings of the International Special Publication 3301. Berkeley. Erosion Control Association Conference 18 Reno. Fleischner . 1994L . T . . Ecological cost f livestoco s k Nevada. grazin n westeri g n North America. Conservation Larscheid, J. G., and W. A. Hubert. 1992. Factors in- Biology 8:629-644. fluencing the size structure of brook trout and brown Gillen. Miller. F KruegerC . . . R L.R ,W . ,d 1984an . . trout in southeastern Wyoming mountain streams. Cattle distribution on a mountain rangeland in North American Journa f Fisherieo l s Management northeastern Oregon. Journa Rangf o l e Management 12:109-117. 37:549-553. Li, H. W., and five coauthors. 1994. Cumulative effects Gordon, N. D.. T. A. McMahon. and B. L. Finlayson. of riparian disturbances along high desert trout 1992. Stream hydrology: an introduction for ecol- streams of the John Day Basin, Oregon. Transac- ogists. Wiley, New York. tions of the American Fisheries Society 123:627- Grost, R. T., W. A. Hubert, and T. A. Wesche. 1991. 640. Field compariso f threno e devices use samplo dt e Major . J 1977, . California climat relation ei vegeo nt - substrat smaln ei l streams. North American Journal tation. Page . Barbous G . Major11-7. J M d n 4an ri , of Fisheries Management 11:347-351. editors. Terrestrial vegetation of California. Wiley, Hall, J. D.. and N. J. Knight. 1981. Natural variation New York. in abundance of salmonid populations in streams Marcuson, P. E. 1977. The effect of cattle grazing on and its implications for design of impact studies: a brown trout in Rock Creek, Montana. Montana De- review. U.S. Environmental Protection Agency, partmen f Fiso d Gamet an h , FederaFisn i hd Ai l Corvallis Environmental Research Laboratory, Restoration, Special Report, Project F-20-R-2I. EPA-600/S3-81-021, Corvallis. Oregon. Helena. . MurphyL Hawkins. . AndersonM H . . . N P. .M . d C , an . Matthews, K. R. 1996. Diel movement and habitat use . WilzbachA . 1983. Densit f d salamanfisyo an h - of California golden trout in the Golden Trout Wil- der relation si ripariao nt n canop physicad yan l hab- derness, California. Transaction e Americath f o s n itastreamn i e northwestert th f o s n United States. Fisheries Society 125:78-86. Canadian Journal of Fisheries and Aquatic Sciences . HallMurphyD . J . d . L.1981M an , . Varied effectf o s 40:1173-1185. clear-cut logging on predators and their habitat in Hurlbert . 1984H . S , . Pseudoreplicatio desige th d nnan small streams of the Cascade Mountains. Oregon. of ecological field experiments. Ecological Mono- Canadian Journal of Fisheries and Aquatic Sciences graphs 54:187-211. 38:137-145. Inyo National Forest. 1982. Golden Trout Wilderness Murphy . Anderson. HawkinsH P . . . L.N M C , d an ., management plan. Inyo National Forest, Bishop, 1981. Effects of canopy modification and accu- California. mulated sedimen n streao t m communities. Trans- Johnson, R. R., L. T. Haight, and J. M. Simpson. 1977. actions of the American Fisheries Society 110:469- Endangered species vs. endangered habitats: a con- 478. cept. U.S. Forest Service General Technical Report Odion . D'Antonio . DudleyM L . . C. . C D , . T d an , . 1988. RM-43:68-79. Cattle grazin n southeasteri g n Sierran meadows: Kauffman . KruegerC . W . d B.J ., an , 1984. Livestock ecosystem change and prospects for recovery. Pages impacts on riparian ecosystems and streamside man- 277-292 in C. A. J. Hall and V. Doyle-Jones, edi- agement implications a review: . Journa f Rango l e tors. Plant biolog f easteryo n California. University Management 37:430-438. of California, Los Angeles. Kauffman, J. B., W. C. Krueger, and M. Vavra. 1983. . PlaitsS 1979 . W , . Livestock grazin d riparianan g / Impact f cattlo s n streambanko e northeastern i s n stream ecosystems—an overview. Pages 39-45 in Oregon. Journa f Rango l e Management 36:683- Cope. B . O , editor. Proceeding forum—graze th f so - 685. d riparian/streaingan m ecosystems. Trout Unlim- . BurnhamP Keller . K . d R.C . an .. 1982. Riparian fenc- ited, Denver. ing, grazing troud an , t habitat preferenc Summin eo t Plaits, W. S. 1981. Effects of sheep grazing on a ri- Creek. Idaho. North American Journal of Fisheries parian-slream environmeni. U.S. Foresl Service Re- Management 2:53-59. search Nole INT-307. Kloepfer, D.. J. Watson, and P. Byrnes. 1994. The Wil- Plans, W. S. 1991. Liveslock grazing. Pages 389-423 derness Act handbook. The Wilderness Society. in W. R. Meehan, editor. Influences of forest and Washington. D.C. rangeland management on salmonid fishes and their . DudleyL Knapp. T . d A..R ,an ,1990 . Growtd an h habitats. American Fisheries Society, Special Pub- longevity of golden trout, Oncorhynchus aguabo- lication 19, Bethesda, Maryland. nita. in their native streams. California Fish an d. McHenryL . PlaitsM d . S.an W ,., 1988. Densitd yan Game 76:161-173. biomas f trou so chad westeran n li r n streams. U.S. Kondolf, G. M. 1993. Lag in stream channel adjustment Forest Service General Technical Report INT-241. 820 KNAP MATTHEWD PAN S Plaits. W. S., W. F. Megahan, and G. W. Minshall. 1983. duction, riparian communities d ecosystean , - mre Method r evaluatinfo s g stream, riparian biotid an , c covery in the southern Sierra Nevada. California. conditions. U.S. Forest Service General Technical Master's ihesis. Universily of California. Sanla Bar- Report INT-138. bara. Plaits, W. S.. and R. L. Nelson. 1985a. Impacts of rest- Simonson, T. D., J. Lyons, and P. D. Kanehl. 1994. rotation grazing on stream banks in forested water- Quanlifying fish habila streamsn i l : transecl spac- shed Idahon si . North American Journa Fisherief lo s ing, sample size, and a proposed framework. North Management 5:547-556. American Journal of Fisheries Managemenl 14: . NelsonL . PlatlsR d . S.. an ,W ,1985b . Slream habiiai 607-615. and fisheries response lo liveslock grazing and in- . RinneN Szaro. J .d . C.R 1988.an , . Ecosysle- map stream improvemenl structures g CreekBi . , Utah. proac o management h f southwestero t n riparian Journal of Soil and Water Conservaiion 40:374- communiiies. Transaciion f iho s e North American 379. Wildlif Nalurad ean l Resources Conference 53:502- . NelsonL . PlansR d . S. an W ,1988, . Fluctuation i n 511. trout populations and their implications for land- USGAO (U.S. General Accounling Office). 1988. Pub- evaluatione us . North American Journa Fisherief lo s lic rangelands: some riparian areas restored bul Management 8:333-345. widespread recovery will be slow. USGAO, GAO/ Platts. W. S., and R. L. Nelson. 1989. Stream canopy RCED-88-80. Washingion. D.C. an s relationshidit salmonio pt d- biomasin e th n i s Van Devenier. PlansS . W . d . S.J 1983, an , . Sampling termountain West. North American Journa f Fisho l - and estimating fish populations from sireams. Trans- eries Management 9:446-457. aciion f Ihso e North American Wildlif Nalurad ean l . LewisH . Shirazi. PlattsA D . . .S. d M W ,1979, an , . Resource Conference 48:349-354. Van Devenier, J. S.. and W. S. Plalls. 1985. A computer Sediment particle sizes used by salmon for spawn- software sysle r enieringmfo . managing anad an ,- ing with methods for evaluation. U.S. Environmen- lyzing fish capiure dala from sireams. U.S. Forest tal Protection Agency, Ecological Research Series, Service Research Nole INT-352. EPA-600/3-79-043. Van Velson . 1979R , . Effec f livesloco l k grazing upon . WagslaffJ . PlatlsF d . S.an .W , . 1984. Fencin cono gl - rainbow iroul in Oiler Creek, Nebraska. Pages 53- trol livestock grazing on riparian habitals alon CopeB. g O. , edilorin 55 . Proceedings, grazingand sireams s iviabla i : e allernalive? North American riparian-stream ecosyslems forum. Troul Unlimited. Journa f Fisherieo l s Management 4:266-272. Vienna. Virginia. Rinne. J. N. 1988. Grazing effects on stream habital Wagner . 1978H . F , . Liveslock grazin ihd egan liveslock and fishes: research design considerations. North induslry. Pages . Brokaw121-14P . H n 5i . edilor. American Journa f Fisherieo l s Management 8:240- Wildlif d Americaan e . Counci n Environmenlao l l 247. Qualily. Washingion, D.C. Roalh. L. R.. and W. C. Krueger. 1982. Gallic grazing Young, M. K. 1995a. Mobilily of brown irout in south- influence on a mounlain riparian zone. Journal of central Wyoming sireams. Canadian Journal of Zo- Range Managemenl 35:100-103. ology 72:2078-2083. Rosgen, D. L. 1994. A classification of natural rivers. Young, M. K. 1995b. Residenl iroul and movemenl: Galena 22:169-199. consequences of a new paradigm. U.S. Deparlmenl . 1995SarrA . D . Grazing, graminoid hysleresisd san : of Agricullure Foresi Service Fish Habilal Rela- investigating relationships belween liveslock pro- lionships Technical Bulletin 18:1-5.