its objectives are eligible forAnnual membership dues area college or university; $13.50ng Members; $250 for Officialit, and Retired Members mayer subscriptions, which includeis, are $70 per year in Canada,ust be made in United States:anadian funds), made payableive Director.

), published at Lawrence, KansasPrice per copy: $12. Undelivereded by North American subscribershin eight months of publication.

:E, KANSASders of this journal, and nonprofitof the material in it. Articles not

at and may be copied without limit;r teaching purposes. This journalcment of the source. To reprint ae original authors and notification.ic or multiple reproduction of anyunder license from the Society; inI from one of the authors. Address

TRANSACTIONS of theAMERICANFISHERIES SOCIETY

July 1981VOLUME 110

NUMBER 4

Trawaaihh, thr Ahreriran Fisherirc Society 110:409-478, 1981Copyright by the American Fisheries Society 1981

Effects of Canopy Modification and AccumulatedSediment on Stream Communities

MICHAEL L. MURPHY,' CHARLES P. HAWKINS, AND N. H. ANDERSON

Department of Entomology, Oregon Slate UniversityCorvallis, Oregon 97331

AbstractSmall streams differing in sediment composition were compared in logged arid forested reach-

es to determine effects of accumulated fine sediment on stream communities under differenttrophic conditions. Three stages of forest communit y succession were studied in the CascadeMountains: recently clear-cut areas without forest canopy (5-10 years after logging); second-growth forest with deciduous canopy (30-40 years after logging); and old-growth coniferousforest (>450 years old). One stream with mostly coarse sediment (56-76% cobble) and one withmore fine sediment (5-14% sand and 23-53% gravel) were contrasted for each successionalstage. In general, streams traversing open clear-cuts had greater rates of microbial respiration,and greater densities or biomasses of aufwuchs, benthos, drift, salamanders, and trout than didthe shaded, forested sites regardless of sediment composition. We conclude that for these smallCascade Range streams, changes in trophic status and increased primary productivity resultingfrom shade removal may mask or override effects of sedimentation.

IETYid was incorporated in 1911iervation, development, andromotion of all branches ofion of knowledge about fish,

Fishing InstituteMinistry of

sity of Wisconsin-n Illinois Universityla Department of

osvenor Lane, Bethesda,

31Office Box 1150, Columbia,

of Fisheries and Wildlife Sci-tic Insitute and State Univer-i 24061

DAVID J. HANSENCALVIN L. GROENSTEPHEN A. FLICKINGERDANIEL J. FABERWILLIAM D. DAVIESDONALD F. AMENDNORMAN J. ABRAMSON

469

Fine inorganic sediment as a nonpoint-sourcepollutant of streams is a major concern to re-source managers. Sediment occurs naturally instreams, but some land-use practices increasesediment loads of streams to the point of de-grading aquatic resources (Cordone and Kelly1961; Gibbons and Salo 1973; Iwamoto et al.1978). It often is difficult to assess effects ofsedimentation, however, because of the inher-ent large natural variation in stream character-istics and because watershed practices oftenhave multiple effects on stream ecosystems.The purpose of this study was to evaluate therole of trophic characteristics of stream systemsin modifying effects of sedimentation on thestream community.

The relative importance of allochthonous

' Present address: National Marine Fisheries Ser-vice, Auk Bay Laboratory, Auk Bay, Alaska 99825.

and autochthonous energy sources is one of theprimary determinants of structure and func-tion of stream communities (Cummins 1974).To a large degree, availability of different en-ergy sources is controlled by forest canopy. Thefood base of shaded streams is dominated byallochthonous organic matter from leaf fall,whereas in open streams the dominant foodsource arises from autochthonous productionof periphyton. Amounts of these inputs influ-ence the kinds of functional groups (for ex-ample leaf shredders, aufwuchs scrapers, fine-particle collectors) that can develop in a stream(Cummins 1974).

Earlier work has shown that both fauna andflora often are more abundant in sections ofstreams with open canopies than in forestedsections (Hughs 1966; Thorup 1966; Albrecht1968; Lvford and Gregory 1975; Aho 1976;Erman et al. 1977; Hunt 1978; Gregory 1980;Newbold et al. 1980; Murphy and Hall 1981).Removal of streamside vegetation during log-

NOTICETM!: MATEp !AL MAYBE PRBIC: Ev?COPIIANT 614!

intlE /7 Os

•OREGON

OLD-GROWTH

0 C L EARCUT

Li SECOND-GROWTH

MackJr..;

Mc Kenzie River

Cougar Cr

N. ForkCr Wycof

Cr.

•

MillCr

Fawn

km

ST REAP

ging appears to increase aquatic production atall trophic levels by increasing periphyton pro-duction (Gregory 1980; Murphy and Hall1981). On the other hand. man y reports of log-ging impacts emphasize the destructive poten-tial of accumulated sediment that adversel y af-fects stream habitat (Cordone and Kell y 1961;Gibbons and Salo 1973; Iwamoto et al. 1978).Thus, logging may have two opposing effects,canopy removal tending to increase stream pro-ductivity and sedimentation tending to degradephysical habitat. Is the increased aquatic pro-duction observed in clear-cut areas a generalphenomenon, or it is restricted to watershedswhere sediment is not a problem? Resourcemanagers need a better understanding of theroles and relative importance of sedimentationand trophic status in the overall response ofstream ecosystems to watershed practices.

This study was an attempt to resolve some ofthese contradictory results by focusing on thepositive stimulus of canopy removal that mightoffset detrimental effects of sedimentation as-sociated with clear-cut logging in the westernCascade Range of Oregon. We comparedstreams within logged and forested sections ofwatersheds in the Oregon Cascades. We sam-pled sediment, organic detritus, algae, insects,and vertebrates. We show that for small Cas-cade Range streams, changes in energy flow re-sulting from shade removal may mask or over-ride effects of sediment pollution.

MethodsThe study area was located in the upper

McKenzie River drainage of the heavily forest-ed western Cascade Mountains (Fig. 1). Siteswere located on the Little Butte and Sardinevolcanic series that are a mosaic of tuffs, brec-cias, and basalt of varying resistance to weath-ering (Beaulieu 1971). Clear-cut logging iscommon in the area. Typicall y , all trees are re-moved and slash is burned, usually withoutstreamside buffer strips.

We selected streams that provided a contrastin size of streambed sediments. Three pairs ofsites, each pair at a different stage of forestcommunity succession, were sampled: recentlyclear-cut areas without forest canopy (5-10years after logging); second-growth forest withdeciduous canop y of Alnus rubro (30-40 yearsafter logging); and old-growth coniferous for-est of P.seudotsuga menziesii and Thuga helerophila

FIGURE 1.-Locations of the study sites in the CascadeMountains. The H. J. Andrews Experimental Forest isoutlined. Stippled areas are reservoirs. H and L indicatehigh- and low-gradient sites, respectively.

(>450 years). Each pair within a successionalstage had a stream with relatively coarse sedi-ment and one with more fine sediment in thestream channel. On first inspection, streambedsediment composition in the study area ap-peared to be closely correlated with channelgradient. Thus, sections were selected that hada channel gradient of about 10% to providesites with relatively coarse streambeds, and thathad a gradient of about 1% to provide sites withmore fine sediment.

All sites were on small perennial streams of

4-8-km2 drainage areas and minimum dis-charges of 0.02-0.09 m 3/second (Table 1). Max-imum stream temperature was less than 21 Cin all sites and was correlated with canopy den-sity in that the clear-cut sites had the highestsummer temperatures. Streams at the lower el-evations with southerl y aspect were slightlywarmer than other streams with similar canopydensity.

The high-gradient old-growth and clear-cutsites were on Mack Creek in the H. J. AndrewsExperimental Forest. This watershed is about90% undisturbed old-growth forest upstreamof the sites. The clear-cut site was located 200in downstream from the old-growth forest andwas relativel y undisturbed except for canopy

TABLE characterioalong Cascade Mountain st-rea

Riparian age (years)Watershed area (km')Percentage of watershed loggedCanopy density (%)Discharge (minimum m3/second)Temperature (maximum C)Elevation (m)Aspect

removal. The high-gradieiwas on North Fork Wycofmuch of the upper area of1920's and about 32% oflogged in the 1940's. Nohas occurred in the past 3°client old-growth site was (contained about 35% oldonly a small area (8%) logglow-gradient second-growtCreek. Approximately 30was old-growth forest. INdone during the 1970'sclear-cut site, Fawn Creekmost heavily disturbed of t5 years preceding this stttershed was clear-cut. Sloptributary channels contrittines of sediment to the m:1978.

For each site we measudensity (an index of shadieter (see Brown 1969). Tisisted of a convex mirror dallowed a visual estimatestreambed shaded by sunStream discharge was detemer low flow by measuringand current velocity. Stre;measured with maximum-eters from June to Novem

Six core samples, three iriffles, were taken at eachto sample sediment and dhider (0.25 m 2 ) was driventorn and a constant volumters) was removed. Particmm were sieved in theclasses: sand (0.05-1.0 numm); pebbles (16-50 mm

470

MURPHY ET AL.

STREAM PRODUCTIVITY: EFFECTS OF CANOPY AND SEDIMENTS 471

TAB LE 1.-Physical characteristics of the recent clear-cut (CC), second-growth (SG), and old-growth (OG) study sitesalong Cascade Mountain streams.

High-gradient streams (10%) Low-gradient streams (1%)

CC SG OG cc SG OG

Riparian age.(years) 10 35 450 7 35 450Watershed area (km2) 5.5 4.0 5.4 6.8 8.2 6.4Percentage of watershed logged 10 32 10 36 70 65Canopy density (%) 0 85 75 0 85 75Discharge (minimum m3/second) 0.07 0.03 0.07 0.06 0.09 0.02Temperature (maximum C) 18.5 15.5 15.5 20.0 14.0 18.0Elevation (n) 730 500 760 500 500 36))Aspect S E W E S

of the study sites in the Cascade(. Andrews Experimental Forest isas are reservoirs. H and L indicate71 sites, respectively.

pair within a successionalwith relatively coarse sedi-

t more fine sediment in thefirst inspection, streambed

:ion in the study area ap-qy correlated with channelLions were selected that hadt of about 10% to providecoarse streambeds, and that)out 1% to provide sites witht.small perennial streams ofareas and minimum dis-

9 (Table 1). Max-ierature was less than 21 Ccorrelated with canopy den-ar-cut sites had the highestres. Streams at the lower el-:herly aspect were slightl ystreams with similar canopy

nt old-growth and clear-cutCreek in the H. J. Andrewsst. This watershed is about)1(1-growth forest upstreamear-cut site was located 200n the old-growth forest andsturbed except for canopy

removal. The high-gradient second-growth sitewas on North Fork Wycof Creek. Fire burnedmuch of the upper area of the watershed in the1920's and about 32% of the watershed waslogged in the 1940's. No new logging activityhas occurred in the past 35 years. The low-gra-dient old-growth site was on Mill Creek, whichcontained about 35% old-growth forest withonly a small area (8%) logged in the 1970's. Thelow-gradient second-growth site was on CougarCreek. Approximately 30% of this watershedwas old-growth forest. No clear-cutting wasdone during the 1970's. The low-gradientclear-cut site, Fawn Creek, appeared to be themost heavily disturbed of the study sites. In the5 years preceding this study, 36% of the wa-tershed was clear-cut. Slope failures and sluicedtributary channels contributed massive quan-tities of sediment to the main channel in 1977-1978.

For each site we measured angular canopydensity (an index of shading) with a densiom-eter (see Brown 1969). The densiometer con-sisted of a convex mirror divided into grids andallowed a visual estimate of the percent ofstreambed shaded by surrounding vegetation.Stream discharge was determined during sum-mer low flow by measuring cross-sectional areaand current velocity. Stream temperature wasmeasured with maximum-minimum thermom-eters from June to November 1978.

Six core samples, three in pools and three inriffles, were taken at each site during Augustto sample sediment and detritus. A metal cyl-inder (0.25 m 2 ) was driven into the stream bot-tom and a constant volume of sediment (12 li-ters) was removed. Particles larger than 0.05mm were sieved in the field into four sizeclasses: sand (0.05-1.0 mm); gravel (1.0-16.0mm); pebbles (16-50 mm); and cobbles (>50

mm). Fine silt (<0.05 mm) and water thatpassed through the sieves were retained in abucket. Two subsamples (1%) of this suspen-sion were filtered (0.45 pm pore size), dried,weighed, and burned (500 C, 12 hours) in afurnace to determine amounts of inorganic andorganic matter. Density of dried silt was mea-sured to convert weight data to volume. Vol-umes of sand and gravel fractions were mea-sured in the field by water displacement, thenthe materials were dried and burned in the lab-oratory to determine inorganic and organicportions. For pebble and cobble fractions, or-ganic matter was decanted or picked from theinorganic material, volume of inorganic sedi-ment was measured by water displacement, andthe organic portion was dried and weighed.Because silt never accounted for more than0.2% by volume of the total sediment sample,we combined data for silt and sand in the dataanalyses.

Microbial respiration associated with organicdetritus was measured monthly from June toSeptember. Sediment was randomly scoopedfrom a pool at each site and sieved into twofractions, 0.45-50.0 p.m and 0.05-1.0 mm. Wemeasured respiration on both fractions in aGilson Respirometer (Gilson 1963), adjustingtemperature of the respirometer to approxi-mate ambient stream temperature for the sam-ple period. Samples were dried, weighed, andashed, and respiration (p.I 0 2 . hour-1 ) was ex-pressed per AFDW (ash-free dry weight).

Standing crop of aufwuchs (periphyton, bac-teria, fungi, and associated organic matter onrock surfaces) was estimated monthly from Julyto October. We used a wire brush to scrape ad-hering matter from three cobbles from bothpools and riffles from each site each month.Scrapings were collected, dried, weighed, and

HIGH GRAD/ENT SITES

<I Ell r-I6 ■6-50 '50

LOW GRAD/ENT SITES

fF-

STREAM

VFPOM

SIZE FR,

FIGURE 3.—Amount of organic in

streambed sediment cores (meanfine particulate organic matterAm; fine particulate organic mamm; coarse particulate organic r,mm.

p.m fraction (r = 0.68) or t(r = 0.59).

Respiration associated '4%,

from the clear-cut sites wasgreater than that from the4). Comparison of means340 ± 101 Al 02 - ' 1sites compared with 134in the forested sites (t-test,ration did not differ signifigrowth and second-growthno consistent relationshirrespiration and percentalstreambed.

Clear-cut sites containedaufwuchs as did foresteding 3.1 ± 1.2 g- comFg• m -2 in forested sites (t-teswuchs standing crop did notbetween old-growth and stNo consistent relationshipstanding crop of aufwuchsfine sand. Periphvton cons:crusted diatoms in shadedmentous and other macro-alabundant in open sites.

Densit y of benthic invertecut sites averaged 1.5 and

MURPHY ET AL.472

ashed. We measured the three longest orthog-onal axes of each rock and calculated approx-imate total surface area by the formula for arectangular solid. Aufwuchs AFDW on eachrock was expressed per unit surface area of thecobble.

Benthic invertebrates were sampled duringJune, August, and October. We took three sam-ples (0.1 m 2 ) in both pools and riffles using amodified Surber sampler (Erman et al. 1977).Organisms retained by a 1-mm mesh werecounted and their lengths measured. Biomass(dry weight) was estimated from length-weightrelationships. A detailed analysis of inverte-brate communities will be the subject of a forth-coming paper so only major trends in densityand biomass are discussed here.

Drifting invertebrates in each stream weresampled five times from June to November. Asingle net (opening 23 cm by 18 cm, net 1 mlong, mesh 330 inn) was placed in a riffle ateach site and removed 24 hours later. Driftsamples were sieved into two fractions (>1.0mm and <1.0 mm). Three subsamples (2%) ofthe finer fraction and all of the coarser fractionwere examined under a microscope to countinvertebrates. Density (number • m-3) of drift-ing invertebrates was calculated based on thevolume of water filtered.

Biomass of all aquatic vertebrates at each sitewas estimated by the removal method (Zippin1956) three times at bimonthly intervals fromJune through October. Vertebrates in a 30-mreach of each site were collected with an elec-troshocker, anesthetized, identified, measured,weighed, and released within the site. Produc-tion of young-of-the-year trout (age 0+) wasestimated by the instantaneous-growth method(Chapman 1971). Based on length measure-ments, older trout (age 1+) were difficult toseparate into age-classes, and thus their pro-duction could not be accurately estimated. Scalesamples were not taken. Salamanders and frogsseemed sensitive to electroshocking, and pop-ulations declined through the study period.Hence we estimated population densities andbiomasses of amphibians only for the first sam-pling period and made no estimate of produc-tion.

ResultsSediment in the high-gradient streams con-

sisted mostly of loose cobble and almost nothing

I00

80

60

40

20

80

60

40

20

SEDIMENT SIZE FRACTION( mm )

FIGURE 2.—Size composition of inorganic sediment in corestaken from the old-growth, clear-cut, and second-growthsites (mean ± SE; N = 6).

was finer than 1 mm. The low-gradient streamscontained 36-58% sand and gravel, whichembedded larger rocks, compared to 9-25%sand and gravel in the high-gradient streams(Fig. 2). The difference between high- and low-gradient sites was significant for both sand(t-test, P < 0.01) and gravel (P < 0.001).

The amount of organic matter in thestreambed corresponded to channel gradientand not to forest successional stage. The low-gradient sites contained more organic detritusthan did the high-gradient sites (Fig. 3): anaverage of 274 138 (±2 SE) g • m-2 versus115 48 g• m-2 (t-test, P < 0.025). We foundsignificant differences (P < 0.01) betweenhigh- and low-gradient sites for the two smallestsize fractions, but the coarsest fraction did notdiffer significantly among the sites. Compari-son of means also showed no significant differ-ence between open and forested sites for anysize fraction of organic detritus.

Amounts of sand and organic detritus weredirectly related. For individual samples, thecorrelation between amount of organic matterand percentage of sand was significant for allthree size fractions of organic matter. The cor-relation was stronger for the 0.05-1.0-mm frac-tion (r = 0.97, N = 36) than for the 0.45-50.0-

CLEARCUT OLD-GROWTH

SECOND-GROWTH

V GRAD/ENT SITES

OLD- GROWTH

SECOND-GROWTH

GRADIENT SITES

200

I 00

200

100

0

HIGH GRADIENT SITES

SECOND-GROWTH

LOW GRADIENT SITES

aLL

cr,

0 300

Z 2000Hacc 100

LL)Cr

2

0

400

HIGH GRADIENT LOW GRADIENT

A

B

1-16 1116-50 >50

RENT SIZE FRACTIONmm

4tion of inorganic sediment in coresmoth, clear-cut, and second-growth= 6).

VFPOM FPOM CPOM

SIZE FRACTION

FIGURE 3.—Amount of organic matter by size fraction instreambed sediment cores (mean SE; N = 6). Veryfine particulate organic matter (VFPOM) = 0.45-50gm; fine particulate organic matter (FPOM) = 0.05-1mm; coarse particulate organic matter (CPOM) = 1-16

S I REAM PRODUCTIVITY: EFFECTS OF CANOPY AND SEDIMENTS 473

I I CC I SG OG

gm fraction (r = 0.68) or the >1-mm fraction(r = 0.59).

Respiration associated with organic matterfrom the clear-cut sites was nearly three timesgreater than that from the forested sites (Fig.4). Comparison of means showed an average340 ± 101 gl 02 . hour-IL g- ' in the clear-cutsites compared with 134 ± 44 ttl 02 ' hour- ' •in the forested sites (t-test, P < 0,001). Respi-ration did not differ significantly between old-growth and second-growth sites. We also foundno consistent relationship between detritalrespiration and percentage of sand in thestreambed.

Clear-cut sites contained about twice as muchaufwuchs as did forested sites (Fig. 4), averag-ing 3.1 ± 1.2 g • m-2 compared to 1.4 ± 0.2g- m-2 in forested sites (t-test, P < 0.001). Auf-wuchs standing crop did not differ significantlybetween old-growth and second-growth sites.No consistent relationship was found betweenstanding crop of aufwuchs and percentage offine sand. Periphyton consisted mostly of en-crusted diatoms in shaded sites, whereas fila-mentous and other macro-algae were seasonallyabundant in open sites.

Density of benthic invertebrates in the clear-cut sites averaged 1.5 and 2.3 times greater

SITE

FIGURE 4.—Respiration of organic matter taken fromstreambeds of the study sites (A) and standing crop ofaufwuchs on rock surfaces (B). Data are mean ± SE;N = 8 measurements of respiration; N = 24 auf-wuchs samples. CC = clear-cut, SG = second-growth,and OG = old-growth sites; AFDW = ash free dryweight.

than in the forested sites during June and Au-gust, respectively (t-test, P < 0.005, Fig. 5).Benthic density in October, however, did notdiffer significantl y among successional stages.In October, following leaf fall and a reductionin shading, all forested sites showed peaks indensity of benthos. The increase was especiallypronounced in the second-growth sites, whichhad mostly deciduous canopies. During Octo-ber, the increased density of benthos in siteswith deciduous canopies may have resultedfrom greater abundance of food sources due toboth increased primary production and thelarge influx of leaves. We found no significantrelationship between mean density of benthosand average sediment composition.

Biomass of invertebrates did not always par-allel observed trends in density. Biomass in low-gradient sites often was dominated by the largesnail Juga plicifera, which was very abundant in

n. The low-gradient streamssand and gravel, which

rocks, compared to 9-259the high-gradient streams

mce between high- and low-significant for both sand

Id gravel (P < 0.001).f organic matter in themded to channel gradientuccessional stage. The low-ined more organic detritus-gradient sites (Fig. 3): an138 (±2 SE) g- m-2 versustest, P < 0.025). We foundnces (P < 0.01) betweenem sites for the two smallestle coarsest fraction did notamong the sites. Compari-iowed no significant differ-and forested sites for any

tnic detritus.and organic detritus were

ir individual samples, theamount of organic matter

;and was significant for all)f organic matter. The cor-r for the 0.05-1.0-mm frac-36) than for the 0.45-50.0-

HIGH GRADIENT SITES

SECOND -GROWTH

CLEARCUT

OLD -GROWTH

OCTJUN AUG'

HIGH GRADIENT 57 TES

SECOND-GROWTH

CLEARCUT

OLD-GROWTH

LOW GRADIENT SITES

IC

0 14

12

/0

8

6

4

2

12

10

8

6

4

2

0 102LO0

6u_

4

O 2

LOW GRADIENT SITES

JUN ['AUG22 15

1-111 T o FE,AUG 3 SEP

31 29 IOCT26

SAL MO

D1CAMPTODON

HIGH GRADIENT25

20

15

10

CC SG OG

SAMPLING PERIOD

FIGURE 5.--Density of invertebrates in benthic samples bysampling period (mean ± SE; N = 6).

pools. Open sites and high-gradient sites hadfew snails. Among riffle samples, biomass wasabout 2.3 times greater in open sites than inshaded streams (P < 0.005). However, shadedpools had 2.2 times more biomass than pools inopen sections of streams, although this differ-ence was not statistically significant. There wasno difference in biomass between riffles of low-and high-gradient streams. Among pools, how-ever, low-gradient streams had 3.4 times thebiomass of high-gradient streams (P < 0.06).

Density of drifting invertebrates varied ac-cording to sampling date and successional stage(Fig. 6). Two-way analysis of variance with com-parison of treatment means showed significanteffects for both sample date (P < 0.005, F-test)and successional stage (P < 0.025, t-test). Allsites exhibited peaks in density of drift in Au-gust with minima in September and October.Differences among successional stages weremost pronounced during August. The maxi-mum values for the two August sampling pe-riods were both recorded in the low-gradientclear-cut site. Overall, the two clear-cut sites av-eraged 1.5 ± 1.3 individuals/m 3 more than didthe forested sites. We found no consistent dif-

6

4E2 2

SAMPLE DATE

FIGURE 6.—Deatity of invertebrates in 29-hour drift sam-ples by sampling period. °Lost sample.

ference in density of drift between high- andlow-gradient sites.

Only five species of aquatic vertebrates werecollected and no consistent difference in num-ber or relative dominance of species was foundamong successional stages or between low- andhigh-gradient sites (Fig. 7). Dominant specieswere neotenic . Pacific giant salamander Dicamp-todon ensatus, cutthroat trout Salmo clarki, andrainbow trout S. gairdneri. The tailed frog As-caphus truei and the reticulate sculpin Cottus per-plexus were present in some of the sites. Effectsof shade and sediment on vertebrates depend-ed on the species (Fig. 7). Salamanders weremore abundant in high-gradient sites than inlow-gradient sites (paired 1-test, P < 0.025).Biomass in the low-gradient sites averaged only55% of that in the high-gradient sites. Sala-mander biomass also was inversely related tothe percentage of sand in the streambed (r =—0.90, N = 6, P < 0.05), but biomass was notsignificantly correlated with amount of shade(r = —0.48).

Trout were more abundant in clear-cut sitesthan in forested sites (Table 2). Biomass wassignificantl y greater in clear-cut than in forest-ed sites for all sampling periods except August,when freshets reduced sampling efficiency(t-test; P < 0.025, < 0.25, < 0.0025 for July,August, and October, respectively). For thethree sampling periods, trout biomass averaged

STREAM

SI

FIGURE 7.—Biomass of stream yewith amounts of salamandertrout Salmo spp., reticulate saand frog Ascaphus truei indiclear-cut, SG = second-growthsites.

about three times more ithan in the forested sites,4.4 g•m-2 in the clear-cut2.9 ± 0.7 g • m- 2 in the fotrout biomass (June—Octcorrelated with mean motlector-gatherer invertebraCummins 1978) and mewith total invertebrate biotpies (Fig. 8). Correlation'mass and both mean denserers and mean density cfrom riffles were also sign0.89, respectively; N = 6; .no consistent relationshipmass and percentage of saMean weight of age-I+ tripsignificantly between higlsites nor among successior

Young-of-the-year (age-Ionly 0-26% of total trouttheir small size (Tables 2biomass of age-0+ trout inbiomass was greatest, was

979

MURPHY ET AL.

D1CAMPTODON

LOW GRADIENT

E

CC SG OG C SG OG

ASCAPHUS

COTT US

a

200

0 COLLECTOR-GATHERERS

TOTAL INVERTEBRATES12 0

0 •

00 •

0•

0 0 40 BO 120 160

475STREAM PRODUCTIVITY: EFFECTS OF CANOPY AND SEDIMENTS

'H GRADIENT SITES

SECOND-GROWTH

OLD-GROWTH

NILW GRADIENT SITES

SOS AUG El sEp OCT31 29 26

SAMPLE DATE

invertebrates in 24-hour drift sam-5d. °Lost sample.

of drift between high- and

of aquatic vertebrates wereinsistent difference in num-iinance of species was found1 stages or between low- and(Fig. 7). Dominant species

fic giant salamander Dicarnp-roat trout Salmo clarki, andg irdneri. The tailed frog As-reticulate sculpin Cottus per-

. in some of the sites. Effectslent on vertebrates depend-(Fig. 7). Salamanders werehigh-gradient sites than in(paired /-test, P < 0.025).

-gradient sites averaged onlye high-gradient sites. Sala-Iso was inversely related tosand in the streambed (r =0.05), but biomass was not

ated with amount of shade

abundant in clear-cut sitestes (Table 2). Biomass was- in clear-cut than in forest-sling periods except August,luced sampling efficiency< 0.25, < 0.0025 for July,ber, respectively). For theods, trout biomass averaged

SITE

FIGURE 7.-Biomass of stream vertebrates during summer,with amounts of salamander Dicamptodon ensatus,trout Salmo spp., reticulate sculpin Cottus perplexus,and frog Ascaphus truei indicated for each site. CC =clear-cut, SG = second-growth, and OG = old-growthsites.

about three times more in the clear-cut sitesthan in the forested sites, with a mean of 9.0 ±4.4 g- m -2 in the clear-cut sites compared with2.9 ± 0.7 g • M -2 in the forested sites. Averagetrout biomass (June-October) was stronglycorrelated with mean monthly biomass of col-lector-gatherer invertebrates (see Merritt andCummins 1978) and moderately correlatedwith total invertebrate biomass from riffle sam-ples (Fig. 8). Correlations between trout bio-mass and both mean density of collector-gath-erers and mean density of total invertebratesfrom riffles were also significant (r = 0.88 and0.89, respectively; N = 6; P < 0.01). We foundno consistent relationship between trout bio-mass and percentage of sand in the streambed.Mean weight of age-I + trout also did not differsignificantly between high- and low-gradientsites nor among successional stages (Table 2).

Young-of-the-year (age-0+) trout constitutedonly 0-26% of total trout biomass because oftheir small size (Tables 2 and 3). Density andbiomass of age-0+ trout in October, when theirbiomass was greatest, was significantly greater

COLLECTOR- GATHERERS (mg • rn -2 I

200 400 600 800 1000

TOTAL INVERTEBRATES (mg • rn -2 )

FIGURE 8.-Observed relationships between trout biomassand riffle invertebrate standing crops. Values for bothtrout and invertebrates are means of 3 months (June,August, and October) for each site (r = 0.99, P < 0.01for trout versus collector-gatherers; r = 0.83, P < 0.05for trout versus total invertebrates).

in the clear-cut sites than in the forested sites(t-test, P < 0.05). Growth and production weremeasured for two periods, July-August andSeptember-October (Table 3). Growth rate wasdirectly related to the average amount of insectdrift during the corresponding period (r =0.76; P < 0.05; N = 8; sites and periods withdensities <0.05 . M -2 of age-0+ trout excluded).Production through the study period was sig-nificantly greater in the clear-cut sites than inthe forested sites (t-test, P < 0.05) and was par-ticularly high in the low-gradient clear-cut site.

TABLE 2.-Trout standing stock (g -m-2) by site and sam-pling period and mean weight (g) of age-1+ individualsin Cascade Mountain streams. Values are wet weights.CC = clear-cut, SG = second-growth, and OG = old-growth sites.

Samplingperiod

High-gradient streams Low-gradient streams

CC SC OG CC SG OG

Standing stock

Jun 8.3 4.5 4.2 10.0 1.9 2.7Aug 2.4 4.9 1.9 7.7 1.4 1.9Oct 7.0 2.1 3.2 18.7 1.5 3.1Mean 5.9 3.8 3.1 12.1 1.6 2.6

Mean weight of age-l+ trout

Jun 22.4 29.9 15.0 18.0 11.1 16.8Aug 18.5 15.5 17.0 18.5 10.6 16.8Oct 21.4 18.6 18.5 17.8 13.7 16.8Mean 20.8 24.7 16.8 18.1 11.8 16.8

IT

TABLE 3.-Standing stock (g y m -2), instantaneous growth rate (G), and production (mg •nr- 2 •day-') of age-0+ trout 1,1site and sampling period. CC = clear-cut, SC = second-growth, and OG = old-growth sites.

Samplingperiod

High-gradient sit cams Lom -gradient so eat Its

CC SC O(; CC Sc Ot:

Standing stockJul 0.4 0.0 2.0 0.1 0.5Aug 0.3 0.0 0.1 1.2 0.2 0.2Oct 1.0 0.0 0.2 4.9 0.1 0.4

Growth rateJul-Aug 0.0090 - 0.0018 0.0136 0.0026 0.0027Sep-Oct 0.0065 0.0081 0.0015 0.0007 0.0096

ProductionJul-Aug 3.6 0.0 21.5 0.4 1.0Sep-Oct 4.2 0.0 1.0 4.5 0.1 3.3

° Fry not completely emerged.° No age-0+ trout present.

476 STREAMMURPHY ET AL.

Total biomass of aquatic vertebrates duringJuly was about 1.5 times greater in the clear-cutsites than in the forested sites, averaging22.0 4.3 g• m -2 in the clear-cut sites com-pared to 12.8 4.3 g • M-2 in the forested sites(Fig. 7). Total density in the clear-cut sites wasabout double that in the forested sites, with anaverage of 2.6 0.1 individuals • m- 2 in theclear-cut sites compared with 1.3 ± 0.5 in theforested sites. The difference between clear-cutand forested sites was significant for both totalbiomass (t-test, P < 0.05) and total density(P < 0.025). We found no consistent differencein biomass or density between second-growthand old-growth sites.

Discussion

Two major patterns emerge from the data:(1) streams in open clear-cut sites had highermicrobial respiration associated with organicmatter; greater standing crops of aufwuchs,benthic invertebrates, trout, and other verte-brates (all species combined); greater density ofdrifting invertebrates; and greater productionof trout than did the shaded old-growth andsecond-growth sites; (2) low-gradient streamscontained more fine sediment and organic mat-ter and less salamander biomass than did thehigh-gradient streams. Our data corroboratestudies by Albrecht (1968), Erman et al. (1977),Hunt (1978), and Murphy and Hall (1981) thatshowed greater abundance of aquatic species inclear-cut or open sections of streams than inshaded stream sections. In addition, our dataindicate that strong linkages exist between light

levels reaching the stream, primary production,microbial respiration, invertebrate production,and ultimately vertebrate production.

Periphyton production in small streams inthe study area is light-limited (Gregory 1980).Chapman and Knudsen (1980) use similar ar-guments to suggest that fish production in somestreams of the Pacific northwest is indirectlylight-limited. Increased amounts of periphytonin clear-cut zones become available to othertrophic levels both as live algae and as organicdetritus. Because of its low UN ratio, periph-yton may be a higher quality food than alloch-thonous organic matter (Boyd 1973; Andersonand Cummins 1979). Microbial respiration inclear-cut sites was relatively high probabl y be-cause of live algae associated with detritus andbecause dead algae enriched the organic detri-tus in the streambed providing a higher qualitysubstrate for microbes.

Few detrimental effects of fine sediment wereobserved, perhaps because levels of accumulat-ed fine sediment in the study streams were be-low thresholds that are detrimental to thestream community. Generally, sediments 0.1-3.3 mm have the most disruptive effects onstream communities and fish spawning habitat(Iwarnoto et al. 1978). In our study, sedimentless than 1 mm never accounted for more than14% of the streambed, a level that did not ap-pear to adversely affect the community. Exceptfor the abundance of the salamander (D. ensa-

tus), which was inversel y related to the amountof accumulated sand, most other componentsof the stream community varied according to

the amount of shade over t.,of sediment composition.

Sediment load and canoFboth may have adverse oftions not encountered in tin sediment load may camcome wide and shallow, ochannel (Leopold et al. 19with sediment ma y reducetrout (Bjornn et al. 1974;served in our sites wherefrom logging and naturalstairstep channel profilepools downstream of debriKeller and Swanson 1970may cause increases in stre;can be lethal to trout (MoriStream temperature did ncsites, but temperature mayiting factor for salmonidser streams (Brown 1974) or

Small Cascade Range stra high capacity to flushdownstream because of tlEven with the relatively m.iment in Fawn Creek, thecut site, most sediment les:ently was removed from t.the streambed within 1 toshowed large accumulaticmm), and accumulated gra,in the second-growth sitesging. However, instead offish production, accumulatacted to enhance productittimal substrate for retentihabitat for some species cRabeni and Minshall 1977;

Streams of the Cascadescompared to those in otherMean annual yield of suspean undisturbed forested yvern Cascades was only 13and from a clear-cut waters(1963-1968, Fred ricksen Pthree Coast Range streams53, 97, and 102 t- km-2.turbed forest (1959-1964one of these nearly triple1978). Coastal streams ofmore erosive, annually exit • km 2- (Karlin 1979). Rescades may not be applicat

STREAM PRODUCTIVITY: EFFECTS OF CANOPY AND SEDIMENTS 477

(mg • m -2 • day -s) of age-O+ trout b)moth sites.

Low-gradient streams

SG OG

0.1 0.50.2 0.20.1 0.4

0.0026 0.00270.0007 0.0096

0.4 1.00. 1 3.3

stream, primary production,on, invertebrate production,tebrate production.duction in small streams inight-limited (Gregory 1980).udsen (1980) use similar ar-t that fish production in someicific northwest is indirectlyased amounts of periphytonbecome available to otheras live algae and as organic

3f its low C/N ratio, periph-ier quality food than alloch-atter (Boyd 1973; Anderson

7 9). Microbial respiration inrelatively high probably be-associated with detritus andenriched the organic detri-

d providing a higher qualityibes.effects of fine sediment werebecause levels of accumulat-e the stud y streams were be-at are detrimental to the. Generally, sediments 0.1-most disruptive effects ons and fish spawning habitat78). In our study, sedimenter accounted for more thanled, a level that did not ap-Tect the community. Exceptof the salamander (D. ensa-rsel y related to the amount[d, most other componentsnunity varied according to

the amount of shade over the stream regardlessof sediment composition.

Sediment load and canop y removal, however,both may have adverse effects in some situa-tions not encountered in this study. Increasesin sediment load may cause the stream to be-come wide and shallow, often with a braidedchannel (Leopold et al. 1964). Filling of poolswith sediment may reduce suitable habitat fortrout (Bjornn et al. 1974). This was not ob-served in our sites where large wood y debrisfrom logging and natural blowdown created astairstep channel profile and formed plungepools downstream of debris accumulations (seeKeller and Swanson 1979). Canopy removalmay cause increases in stream temperature thatcan be lethal to trout (Moring and Lantz 1975).Stream temperature did not exceed 21 C in oursites, but temperature may be an important lim-iting factor for salmonids in smaller or shallow-er streams (Brown 1974) or in warmer climates.

Small Cascade Range streams generally havea high capacity to flush introduced sedimentdownstream because of their steep gradient.Even with the relatively massive inputs of sed-iment in Fawn Creek, the low-gradient clear-cut site, most sediment less than 1 mm appar-ently was removed from the surface layers ofthe streambed within 1 to 2 years. Fawn Creekshowed large accumulations . of gravel (1-16mm), and accumulated gravel was still apparentin the second-growth sites 30 years after log-ging. However, instead of adversely affectingfish production, accumulated gravel may haveacted to enhance production by supplying op-timal substrate for retention of detritus andhabitat for some species of invertebrates (seeRabeni and Minshall 1977; Williams 1978).

Streams of the Cascades are sediment-poorcompared to those in other geologic provinces.Mean annual yield of suspended sediment froman undisturbed forested watershed in the west-ern Cascades was only 13 t • km -2 (1960-1968)and from a clear-cut watershed it was 68 t • km-2(1963-1968, Fredricksen 1970). In comparison,three Coast Range streams in Oregon averaged53, 97, and 102 t km -2. year from undis-turbed forest (1959-1964), and clear-cuttingone of these nearly tripled its yield (Beschta1978). Coastal streams of California are evenmore erosive, annuall y exporting 2,000-3,000t- km -2 (Karlin 1979). Results from the Cas-cades may not be applicable to other regions

where effects from sedimentation may be moresevere.

The effects of sedimentation in a stream,however, always should be viewed in the con-text of the whole drainage network. Sedimentdeposition depends on stream competence andcapacity, which vary along the longitudinal pro-file of a stream (Leopold et al. 1964). Fine sed-iment exported from high-gradient headwaterchannels deposits downstream where gradientis lower. These inputs from a large number ofdisturbed tributaries might overload down-stream reaches with sediment and reduce waterquality and aquatic productivity. Althoughheadwater channels of Cascade Range streamsappear minimally affected by sediment, theirresistance to sedimentation may be purchasedat a cost to productivity downstream.

AcknowledgmentsWe thank J. D. Hall for his support and use-

ful suggestions during the formative stages ofthe study. D. C. Erman and C. J. Cederholmcritically reviewed an early draft. Peggy Wilz-bach reviewed the manuscript and gave enthu-siastic help with the fieldwork. The RosburoLumber Company kindly granted permissionto sample Fawn Creek and North Fork WycofCreek. This study was supported by a grantfrom the United States Environmental Protec-tion Agency (R806087) to N.H.A. This is tech-nical paper number 5703 of the Oregon Agri-cultural Experiment Station.

ReferencesArlo, R. S. 1976. A population study of the cutthroat

trout in an unshaded and shaded section ofstream. Master's thesis. Oregon State University,Corvallis, Oregon, USA.

ALBRECHT, M. 1968. Die Wirkung des Lichtes aufdie quantitative Verteilung der Fauna in Fliess-gewasser. Limnologica 6:71-82.

ANDERSON, N. H., AND K. W. CUMMINS. 1979. Theinfluence of diet on the life histories of aquaticinsects. Journal of the Fisheries Research Boardof Canada 36:335-342.

BEAULIEU, J. D. 1971. Geologic formations of west-ern Oregon. Oregon Department of Geologyand Mineral Industries, Bulletin 70, Portland,Oregon, USA.

BESCHTA, R. L. 1978. Long-term patterns of sedi-ment production following road constructionand logging in the Oregon Coast Range. WaterResources Research 14:1011-1016.

BJORNN, T. C., M. A. BRUSVEN, M. M. MOLNAU, F. J.WATTS, R. L. WALLACE, D. R. WALLACE, D. R.

NEILSON, M. F. SANDINE, AND L. C. STUEHREN-

BERG. 1974. Sediment in streams and its effecton aquatic life. Idaho Water Resources ResearchInstitute, Moscow, Idaho, USA.

Bow), C. E. 1973. Amino acid composition of fresh-water algae. Archiv fur Hvdrobiologie 72:1-9.

BROWN, G. W. 1969. Predicting temperatures ofsmall streams. Water Resources Research 5:68-75.

BROWN, G. W. 1974. Forestry and water quality.School of Forestry, Oregon State University,Corvallis, Oregon, USA.

CHAPMAN, D. W. 1971. Production. Pages 199-214in W. E. Ricker, editor. Methods for assessmentof fish production in fresh water. IBP Handbook3. Blackwell Scientific Publications, Oxford, En-gland.

CHAPMAN, D. W., AND E. KNUDSEN. 1980. Channel-ization and livestock impacts on salmonid habitatand biomass in western Washington. Transac-tions of the American Fisheries Society 109:357-363.

CORDONE, A. J., AND D. E. KELLY. 1961. The influ-ence of inorganic sediment on the aquatic life ofstreams. California Fish and Game 47:189-228.

CUMMINS, K. W. 1974. Structure and function ofstream ecosystems. BioScience 24:631-641.

ERMAN, D. C., J. D. NEWBOLD, AND K. B. ROBY. 1977.Evaluation of streamside bufferstrips for pro-tecting aquatic organisms. California Water Re-sources Center 165, University of California, Da-vis, California, USA.

FREDRICKSEN, R. L. 1970. Erosion and sedimentationfollowing road construction and timber harveston unstable soils in three small western Oregonwatersheds. United States Forest Service Re-search Paper PNW-104.

GIBBONS, D. R., AND E. 0. SALO. 1973. An annotatedbibliography of the effects of logging on fish ofthe western United States and Canada. UnitedStates Forest Service General Technical ReportPNW-10.

GILSON, W. E. 1963. Differential respiration of sim-plified and improved design. Science (Washing-ton, District of Columbia) 141:531-532.

GREGORY, S. V. 1980. Effects of light, nutrients, andgrazing on periphyton communities in streams.Doctoral dissertation. Oregon State University,Corvallis, Oregon, USA.

HUGE'S, D. A. 1966. Mountain streams of the Bar-berton area, eastern Transvaal. Part 11, the effectof vegetational shading and direct illuminationon the distribution of stream fauna. Hydrobio-logia 27:439-459.

HUNT, R. L. 1978. Instream enhancement of trouthabitat. Pages 19-27 in K. Hasagen, editor. Anational symposium of wild trout management.California Trout, San Francisco, California,USA.

IWAMOTO, R. N., E. 0. SALO, M. A. MADEJ, AND R. L.

McComAs. 1978. Sediment and water qualitya review of the literature including a suggestedapproach for water quality criteria. Fisheries Re-search Institute, Universit y of Washington, EPA910/9-78-048, Seattle, Washington, USA.

KARLIN. R. 1979. Sediment sources and clay mineraldistributions off the Oregon Coast: evidence fora poleward slope undercurrent. Master's thesis.Oregon State University, Corvallis, Oregon,USA.

KELLER, E. A., AND F. J. SWANSON. 1979. Effects oflarge organic material on channel form and flu-vial processes. Earth Surface Processes 4:361--380.

LEOPOLD, L. B., M. G. WOLMAN, AND J. P. MILLER.1964. Fluvial processes in geomorphology. W.H. Freeman, San Francisco, California, USA.

LYFORD, J. H., AND S. V. GREGORY. 1975. The dy-namics and structure of periphyton communitiesin three Cascade Mountain streams. Internation-ale Vereinigung far Theoretische and Ange-wandte Limnologie, Verhandlungen 19:1610-.1616.

MERRITT, R. W., AND K. W. CUMMINS, editors. 1978.An introduction to the aquatic insects of NorthAmerica. Kendall/Hunt, Dubuque, Iowa, USA.

MORING, J. R., AND R. L. LANTZ. 1975. The Alseawatershed study: effects of logging on the aquat-ic resources of three headwater streams of theAlsea River. Oregon Department of Fisheriesand Wildlife Fisheries Research Report 9, Port-land, Oregon, USA.

MURPHY, M. L., AND J. D. HALL. 1981. Varied effectsof clear-cut logging on predators and their habitatin small streams of the Cascade Mountains, Ore-gon. Canadian Journal Fisheries and Aquatic Sci-ence 38:137-145.

NEWBOLD," D., D. C. ERMAN, AND K. B. ROBY. 1980.Effects of logging on macroinvertebrates instreams with and without buffer strips. CanadianJournal Fisheries and Aquatic Science 37: 1076-1085.

RABENI, C. F., AND G. W. MINSHALL. 1977. Factorsaffecting microdistribution of stream benthic in-sects. Oikos 29:33-43.

THORUP, J. 1966. Substrate type and its value as abasis for the delimitation of bottom fauna com-munities in running waters. Pages 57-74 in K.W. Cummins, C. A. Tryon, Jr., and R. T. Hart-men, editors. Organism-substrate relationshipsin streams. Pvmatuning Laboratory of Ecology,Special Publication 4, University of Pittsburgh,Pittsburgh, Pennsylvania, USA.

WILLIAMS, D. D. 1978. Substrate size selection bystream invertebrates and the influence of sand.Limnology and Oceanography 23:1030-1033.

ZIPPIN, C. 1956. An evaluation of the removal meth-od of estimating animal populations. Biometrics12:163-189.

4 78

MURPHY ET AI.. TranAnnoort■ of Mr .1 rnrrscon Fitherir4 Sociele Cop y right hi . the American Fisheries

Energetics,Yellc

A study to determinethen prey, measure meducted in Oneida Lakeyears were similar in eatfed almost exclusively o:diverse. In early summeassimilated about 68%observed growth and eestimates of food intakecollapsed in late summVariation in first-year gof yellow perch cohorts

The objective of this solithe energy content of younperch and their prey, measrelate energy supply and itsearch focused on youngflavescens in Oneida Lake,period when they fed onranged in length from 2(estimates of consumptiorstomach analyses and rate cResponse of age-0 yelloschanges in daphnid abundtween years were examine(

In recent years, bioenerbeen applied to patterns ofcesses controlling productpopulations (Webb 1978).has been estimated and aevaluated for adult brow!(Elliot 1976a, 19766), largeterus saint (Niimi and 131

perch (Solomon and Bradfeye Stizostedion vitreum nitAlthough the energy feyblueback herring Alosa1974) and white bass Morc1974) have been reporte.hioenergetics of voung-oreceived little attention (Ki

In Oneida Lake, young 1portant in the diet of the