Stream foodweb d13C and geomorphology are tightly coupled inmountain drainages of northern Idaho

S. Mazeika P. Sullivan1

School of Environment and Natural Resources, The Ohio State University, 2021 Coffey Road, Columbus,Ohio 43210 USA

Abstract. The relative influence of landscape vs local processes on variation in stream C and energysources to lotic consumers remains unresolved, but is a key component in understanding pathwaysthrough stream food webs. I investigated potential linkages between the distribution of d13C signatures ofprimary producers (periphyton), terrestrial organic matter (detritus), and secondary consumers (aquaticinvertebrates and fish) and: 1) drainage area and 2) stream reach geomorphology in 3 mountainwatersheds of northern Idaho, USA. When considered alone, drainage area explained variation in the d13Csignatures of periphyton (R2

= 0.30) and sculpin (R2= 0.15) only. Subsequent General Linear Models

(GLMs) showed that stream morphology explained significant variation in d13C of periphyton (79%),Diptera (93%), sculpin (51%), and trout (63%). Drainage area was not significant in these models, a resultindicating that once d13C signatures are explained as a function of geomorphic type, the contribution ofdrainage area becomes less important. On the whole, consumer d13C signatures related more strongly toperiphyton d13C than detrital d13C, though this relationship varied among geomorphic types. Collectively,results indicate that the morphology of a stream reach (including factors such as substrate size and storageelements, floodplain development, etc.) may be an important factor in the assimilation and distribution ofC in food webs in forested mountain streams.

Key words: ecosystem size, streams, carbon stable isotopes, food web, patch dynamics.

The complexity of stream ecosystems continues tochallenge ecologists to identify environmental gradi-ents governing lotic communities and to link spatiallyexplicit patterns and processes (Power et al. 1988,Lowe et al. 2006). In particular, a cogent understand-ing of the processes that govern transfers of organicmatter through stream food webs is lacking, but iscritical to our understanding of stream foodwebproductivity and the subsequent dynamics of streamcommunities.

Fluvial geomorphology has provided a physicaltemplate and an overarching conceptual frameworkfor the genesis of many guiding paradigms in streamecology. For example, the hydraulic and geomorphicstream channel adjustments along a river proposed byLeopold and colleagues (Leopold and Maddock 1953,Leopold et al. 1964, Leopold and Langbein 1966)formed the basis for the physical construct of theRiver Continuum Concept (RCC) proposed by Van-note et al. (1980) and complemented their predictionsabout energy flow along the drainage network (i.e., ashift from terrestrial C sources in headwater streams

to increasingly autotrophic C sources in mid-orderstreams and a decreasing trend thereafter). However,many alternative concepts, many of which havecontradicted the RCC (see reviews by Townsend 1996,Thorp et al. 2006, Poole 2010), also have been proposed.

In particular, significant support has emerged for apatch-dynamics view that recognizes discontinuouspatterns along the stream network (Pringle et al. 1988,Townsend 1989, Wu and Loucks 1995). Notably, therole of fluvial geomorphology is central to many ofthese patch perspectives. For instance, in the conceptof Process Domains, Montgomery (1999) hypothe-sized that spatial variability in geomorphic processesinfluences physical habitat, disturbance, and ultimate-ly, the structure and function of stream ecosystems.Poole (2002) proposed a marriage of patch dynamicsand hierarchy theory that would integrate multiplecharacteristics of stream ecosystems (e.g., directional-ity, heterogeneity, spatial scale) and recognize theuniqueness of each stream system. The RiverineEcosystem Synthesis (RES; Thorp et al. 2006) portraysrivers as downstream arrays of hydrogeomorphicpatches that form functional process zones resultingfrom differences in physicochemical habitats.1 E-mail address: sullivan.191@osu.edu

Freshwater Science, 2013, 32(2):606–621’ 2013 by The Society for Freshwater ScienceDOI: 10.1899/12-101.1Published online: 18 April 2013

606

The fundamental links between physical and bioticpatterns in streams (Poole 2002) have served as theunderpinning of many stream classification systems(see reviews by Mosley 1987, Thorne 1997). Forexample, Montgomery and Buffington (1997) devel-oped an intuitive, channel-reach classification schemein mountain drainage basins that classifies streammorphologies into 7 reach types: colluvial, bedrock,and 5 alluvial channel forms (cascade, step-pool,plane-bed, pool–riffle, and dune-ripple). Their systemuses a process-based approach that focuses onsediment transport through the drainage network, inwhich reach-level channel processes are linked to thespatial arrangement of reach morphologies, hillslopeerosional processes, channel confinement characteris-tics, riparian vegetation, and large woody debris.

Whereas process-based classification systems, suchas that of Montgomery and Buffington (1997), can berelated to biological patterns (e.g., diversity, distribu-tion, community composition; Sullivan et al. 2006,Cianfrani et al. 2009), they also may be coupled withfunctional biotic components (e.g., feeding, energyflow, habitat selection). Measurements of C stableisotopes (d13C), for instance, have proven fruitful inunderstanding C flow and in identifying trophic basesin stream ecosystems, largely because of the smallfractionation of C associated with trophic transfer oforganic C and the potential for organic matter sourcesto have distinct d13C signatures (Finlay 2001, Schind-ler and Lubetkin 2004, Delong and Thorp 2006). Thestructural–mechanistic link between stream morphol-ogy and variation in d13C signals of stream taxa hasnot been fully investigated, but is consistent withgeomorphically based patch-dynamics perspectivesand would provide a valuable comparison with otherwork on this topic. For instance, Finlay (2001) showedthat epilithic algae d13C and watershed area werepositively correlated along a gradient in watershedarea. Finlay et al. (1999; variability in d13C betweenpools and riffles) and Walters et al. (2007; d13Cenrichment in biofilm and consumers at rocky vssandy stream reaches) suggested that mechanismsunderlying spatial variation in d13C probably areexplained by small-scale processes.

Multiple mechanisms underlying both longitudinaland local-scale patterns in d13C distribution have beenproposed, and most are linked to differences in thesource of dissolved CO2 (weathering, atmosphere,biotic respiration). For example, in small headwaterstreams, algal 13C depletion probably is influenced bysoil CO2 production and transport and by weatheringprocesses on dissolved inorganic C (DIC) (Finlay2003). 13C depletion at high flow velocity has beenrelated to the photosynthetic demand of CO2 by algae

relative to its supply (Trudeau and Rasmussen 2003,Finlay 2004, Hill and Middleton 2006). Downstreamincreases in temperature, algal growth rates, andchlorophyll biomass also can affect algal d13C (Daw-son et al. 2001, Finlay 2003, Hill and Middleton 2006).In more productive downstream reaches, Finlay(2001) hypothesized that C limitation reduces dis-crimination against algal 13C. Finlay (2004) illustrateda negative relationship between algal d13C anddissolved CO2 (CO2(aq); 1 of the 2 forms of DIC usedby aquatic plants, the other being bicarbonate,HCO3

2), a pattern explained as a result of the jointeffects of decreasing d13C of DIC and increasingphotosynthetic fractionation in CO2(aq) distributionalong the drainage network.

I undertook my study to establish the relationshipsbetween reach-scale stream geomorphology andstream food webs in mountain drainages. I quantifiedthe relationships between stream geomorphology andd13C signatures of stream periphyton, detritus, aquat-ic invertebrates, and fishes in 3 high-relief, forestedheadwater drainages of northern Idaho, USA. Myprimary goal was to assess C sources and flow to theprimary producer and consumer components ofaquatic food webs among geomorphic patches, asrepresented by Montgomery and Buffington’s streamgeomorphic types (vs along the continuum represent-ed by ecosystem size). My secondary goal was toquantify allochthonous (terrestrial) vs autochthonous(originating from the stream) use of C by streamconsumers. Given that 13C of algae and consumersshows a downstream enrichment pattern (Finlay 2001,2004), I hypothesized that differences in d13C ofperiphyton and consumers would be based largelyon ecosystem size (i.e., stream drainage area [DA]),but that geomorphic type also would explain sub-stantial variation. Likewise, the increased role of in-stream primary producers in downstream food webs(Minshall 1978, Vannote et al. 1980, Finlay 2001, Rosi-Marshall and Wallace 2002) suggests that the relativecontribution of autochthonous resources to streamconsumers would be greater in stream reaches withmorphologies typified by less confined channels withmore open canopies and some floodplain develop-ment (e.g., plane-bed, pool–riffle; see Fig. 1 fordetailed predictions).

Methods

Study area

I worked at 15 sites (reaches) distributed across 3watersheds—Beaver Creek Watershed (BCW, n = 5),Elk Creek Watershed (ECW, n = 5), Mica CreekWatershed (MCW, n = 5)—in the Northern Rockies of

2013] GEOMORPHOLOGY AND d13C 607

Idaho, USA (Fig. 2). All 3 watersheds are in theNorthern Rockies Ecoregion (USEPA 2002), character-ized by mountainous, rugged topography with amaritime-influenced climate. In each watershed, reach-es (headwater to confluence) were selected to representMontgomery and Buffington’s morphologies most

common to the study watersheds (Fig. 2). All studyreaches were steep (§0.002 m/m), confined channelsdominated by gravel, cobble, and boulder substrateswith limited floodplains and, thus, met the criteria ofmountain stream channels outlined by Wohl andMerritt (2008). To avoid potentially confounding

FIG. 1. Stylized representations of Montgomery and Buffington’s (1997) geomorphic types and hypothesized predictionsregarding organic C and trophic linkages relative to proportion of autochthonous to allochthonous energy in stream consumers.Each prediction is followed by potential mechanisms in parentheses. Predictions and potential mechanisms are based onpublished literature and field observations. C predictions are primarily for periphyton (based on d13C), but are expected to extendto stream consumers. LW = large wood, DIC = dissolved inorganic C, CO2(aq) = dissolved CO2.

608 S. M. P. SULLIVAN [Volume 32

effects of reach location along the watershed gradient, Iattempted to find reaches of each geomorphic type atmultiple locations along the drainage network. Onceselected, I tested for spatial autocorrelation amonggeomorphic types using Moran’s I (Fortin and Dale2005) in ArcGISH (version 9.3; Environmental SystemsResearch Institute, Redlands, California). Results con-firmed that the geomorphic types were distributedalong the drainage networks without significantclustering or dispersion (I = 0.014, p . 0.05). To avoidtemporal variation, I constrained sampling to the latesummer months, when discharge was relatively stable.Each reach was sampled once across 3 field seasons(2006–2008), so sampling periods were randomized bygeomorphic structure.

Geomorphic surveys

To link biological and physical characteristics ofcatchments in a functionally meaningful way, Idefined reaches as arbitrary units equal to 15 to 203

bankfull width (Harrelson et al. 1994, Kondolf andMicheli 1995) and subsequently used breaks ingeomorphic types to delineate reach boundaries moreprecisely. At each reach, I established 10 equidistantlateral transects (across the stream) and 1 longitudinaltransect (bisecting the stream, running down itslength). I characterized the channel geomorphologyof each reach with longitudinal and cross-sectionalsurveys conducted along each transect, where Imeasured slope, bankfull width, and depth with a

FIG. 2. Locations of Beaver Creek Watershed (BCW), Mica Creek Watershed (MCW), and Elk Creek Watershed (ECW) innorthern Idaho, USA. In each watershed, study reaches are shown from upstream (smaller numbers, e.g., MC1) to downstream(larger numbers, e.g., MC5). PB = plane-bed, PR = pool–riffle, SP = step-pool.

2013] GEOMORPHOLOGY AND d13C 609

stadia rod, laser level, and measuring tape followingthe methods of Cianfrani et al. (2004). I usedWolman’s (1954) pebble-count method to assess bedgrain size at each of the 10 lateral transects (1000sediment clasts/reach). I used pebble-count data toidentify the predominant substrate size (D50 and D95).

I surveyed all pieces of large wood (LW) .0.10 mdiameter and .1.0 m length (Montgomery et al. 1995).I conducted diagnostic visual surveys of each streamreach, with a focus on dominant roughness elementsand sediment sources, sediment storage elements,floodplain development, and confinement character-istics to classify reaches according to Montgomeryand Buffington’s classification system. I measured %

canopy cover over the stream with a sphericaldensitometer and derived drainage area using geo-processing tools in ArcGIS.

Classification of stream reaches

I used diagnostic features outlined by Montgomeryand Buffington (1997) to classify each reach by itsdominant morphology (Table 1). When classifyingreaches, I noted that northern Idaho streams deviatedin some cases from the pattern established byMontgomery and Buffington. Some reach morpholo-gies differed from the expected longitudinal sequence(e.g., pool–riffle upstream of plane-bed, Fig. 2; seeWohl and Merritt 2008). Where appropriate, I usedMontgomery and Buffington’s forced step-pool andforced pool–riffle classifications for channels thatsupported flow obstructions (e.g., LW) that forcedeither step-pool or pool–riffle morphologies. Giventhe unique characteristics of each study system,diagnostic features of the reach may not be fullyconsistent with predicted ranges in some of thesecases (Table 1).

I attempted to include §3 study reaches in all 5geomorphic types, but was unable to do so in somecases. All 3 study watersheds were in remote areas,and this factor limited accessibility of potential studyreaches. Finding free-formed pool–riffle morphologywas difficult (n = 2). I initially conducted allcomponents of the study at 3 plane-bed reaches.However, one of the reaches contained a split channelin the upper quadrant. This characteristic set it apartfrom the other plane-bed reaches, so I decided toexclude it from the analysis.

Sample collection

I collected larval stages of common aquatic inver-tebrates at 6 longitudinally distributed locations/reach with a 600-cm2 Surber sampler with a 500-mm-mesh net. I kept aquatic invertebrates in unfiltered

TA

BL

E1.

Dia

gn

ost

icta

ble

of

ind

icat

ors

for

Mo

ntg

om

ery

and

Bu

ffin

gto

n’s

(199

7)g

eom

orp

hic

clas

sifi

cati

on

of

stre

amre

ach

esfo

rn

ort

her

nId

aho

stu

dy

wat

ersh

eds.

Rel

ativ

ero

ug

hn

ess

=90

thp

erce

nti

leb

edse

dim

ent

dia

met

er/

stre

amd

epth

,D

50

=m

edia

nb

edse

dim

ent

dia

met

erfo

rst

ream

reac

h(m

m),

LW

D=

larg

ew

oo

dy

deb

ris.

Rea

chID

Slo

pe

(m/

m)

Rel

ativ

ero

ug

hn

ess

D5

0(m

m)

Do

min

ant

sub

stra

tety

pe

Do

min

ant

sto

rag

eel

emen

tF

loo

dp

lain

dev

elo

pm

ent

Co

nfi

nem

ent

Mo

ntg

om

ery

/B

uff

ing

ton

clas

sifi

cati

on

BC

10.

050

0.20

028

Co

arse

gra

vel

LW

Dst

eps

No

Co

nfi

ned

Fo

rced

step

-po

ol

BC

20.

033

0.29

732

Co

arse

gra

vel

LW

Dst

eps

Yes

Un

con

fin

edF

orc

edst

ep-p

oo

lB

C3

0.01

50.

284

81C

ob

ble

Bar

s,L

WD

step

s,b

ou

lder

sY

esU

nco

nfi

ned

Fo

rced

po

ol–

riff

leB

C4

0.05

30.

223

8M

ediu

mg

rav

elL

WD

step

sN

oC

on

fin

edF

orc

edst

ep-p

oo

lB

C5

0.02

70.

263

97M

ediu

mco

bb

leB

ou

lder

s,o

ver

ban

kY

esU

nco

nfi

ned

Pla

ne-

bed

EB

10.

013

0.01

24

Ver

yfi

ne

gra

vel

Bar

sL

imit

edS

emic

on

fin

edP

oo

l–ri

ffle

EB

20.

020

0.09

83

Ver

yfi

ne

gra

vel

LW

Dst

eps

Lim

ited

Sem

ico

nfi

ned

Fo

rced

step

-po

ol

EB

30.

040

0.08

45

Fin

eg

rav

elL

WD

step

sL

imit

edS

emic

on

fin

edF

orc

edst

ep-p

oo

lE

B4

0.01

30.

400

11M

ediu

mg

rav

elB

ars,

bo

uld

ers,

LW

Dp

oo

lsL

imit

edU

nco

nfi

ned

Fo

rced

po

ol–

riff

leE

B5

0.00

20.

056

3V

ery

fin

eg

rav

elB

ars,

bed

form

sY

esU

nco

nfi

ned

Po

ol–

riff

leM

C1

0.07

60.

278

34V

ery

coar

seg

rav

elS

tep

s,b

ou

lder

sN

oC

on

fin

edS

tep

-po

ol

MC

20.

032

0.37

350

Ver

yco

arse

gra

vel

Ste

ps,

bo

uld

ers

No

Co

nfi

ned

Ste

p-p

oo

lM

C3

0.03

30.

364

56V

ery

coar

seg

rav

elS

tep

s,b

ou

lder

sN

oC

on

fin

edS

tep

-po

ol

MC

40.

012

0.32

056

Ver

yco

arse

gra

vel

Ste

ps,

bo

uld

ers

Lim

ited

Sem

ico

nfi

ned

Fo

rced

po

ol–

riff

leM

C5

0.01

60.

290

85M

ediu

mco

bb

leO

ver

ban

k,

bo

uld

ers

Yes

Un

con

fin

edP

lan

e-b

ed

610 S. M. P. SULLIVAN [Volume 32

stream water for 6 to 8 h to evacuate their guts beforepreserving them. I used a nylon brush to collectdetritus samples (at the same locations as inverte-brates) from deposits of coarse benthic organic material(CBOM) and periphyton (epilithic algae) from cobblesalong the longitudinal length of each reach. I sampledfish with a backpack electrofisher (Smith-RootH LR12;Smith-Root, Inc., Vancouver, Washington) and dipnets. At each reach, I collected §8 individual adulttrout (.150 mm) and 6 adult sculpin to represent thedominant species present at the sampling reaches (seeStable-isotope analysis below).

Stable-isotope analysis

In the laboratory, I sorted all aquatic invertebratesto order, and I identified the most dominant taxa tofamily with the keys in Merritt and Cummins (1996). Igrouped aquatic invertebrates by the dominant orders(Ephemeroptera, Plecoptera, Trichoptera, and Dip-tera) for stable-isotope analysis. I freeze-dried aquaticinvertebrates for 48 to 72 h, ground them to a finepowder with a mortar and pestle, and packed thepowder in tin capsules. I assembled tissue frommultiple individuals (6–15) in a single compositesample for stable-isotope analysis to minimize within-site variance (Lancaster and Waldron 2001). Compos-ite samples for each study reach comprised individ-uals grouped by taxonomic order.

I filtered and dried periphyton in the laboratoryand combined several samples per reach to createcomposite samples for each reach. I removed fineparticulate organic matter (FPOM; particles ,1 mm)from the detritus, leaving primarily terrestrial leaves.I dried detritus and periphyton samples in a 60uCoven for 48 h, ground the samples in a ball mill, andpacked the powder in tin capsules for analysis.

The fish I selected for analysis represented 6 speciescommon to northern Idaho mountain streams: BrookTrout (Salvelinus fontinalis), Westslope CutthroatTrout (Oncorhynchus clarkii lewisi), Rainbow Trout(Oncorhynchus mykiss), Westslope Cutthroat/Rainbowhybrid (O. mykiss 3 O. clarkii lewisi), Slimy Sculpin(Cottus cognatus), Shorthead Sculpin (Cottus confusus),and Mottled Sculpin (Cottus bairdi). Given the broadlongitudinal distribution of the various stream mor-phologies included in my study, I did not anticipatefinding all trout or all sculpin species in everygeomorphic type. Because of this spatial segregationof the target species, a direct comparison of isotopicsignatures across geomorphic classifications by spe-cies was not always feasible. However, the highdietary overlap of sculpin species (Wydoski andWhitney 2003) suggested that combining sculpin

species for analysis of C isotopes would be anacceptable approach. Dietary studies related to BrookTrout, Rainbow Trout, and Cutthroat Trout haveyielded mixed results (Dunham et al. 2002, Hilder-brand and Kershner 2004, McGrath and Lewis 2007).However, investigators generally agree that signifi-cant dietary overlap occurs among these fishes(aquatic invertebrates, terrestrial invertebrates, andamorphous organic detritus), and they are oftenclassified as primarily invertivores (Raleigh et al.1984, Behnke 2002, Wydoski and Whitney 2003).

At the 5 reaches where .1 trout species werefound, I also ran preliminary a priori analyses (t-testor analysis of variance [ANOVA]) to test for potentialdifferences in d13C values among Brook, Rainbow,and Cutthroat Trout. I found no significant differenc-es among species (all p . 0.05), and I used this resultas justification for grouping trout species together forthe purposes of stable-isotope analysis. I used adultfish of similar size in a replicate to avoid potentialconfounding effects of age-specific diets on stable-isotope signatures. I removed plugs of skinless dorsalmuscle from each individual (Pinnegar and Polunin1999) and then freeze-dried, pulverized (to ensuresample homogeneity), and packed them in tincapsules.

I analyzed all replicate samples for C by continu-ous-flow isotope-ratio mass spectrometry (EA-IRMS)at the University of Washington Stable Isotope Core(Pullman, Washington). I expressed the isotopiccomposition of samples with d13C notation defined as:

d13=

Rsample

Rstandard

� �{1

� �|100

where R is 13C/12C and parts per thousand (%) arethe atomic ratios of the number of atoms in the sampleor standard. Typical analytical precision was 0.2% ford13C determination.

Statistical approach

I was unable to achieve a balanced design becauseof the remote location and the logistic difficulty inaccessing many of the study reaches. Furthermore,patterns related to differences in presence/absence orabundance of the target taxonomic groups resulted inan unbalanced number of observations per reach.Therefore, my primary analytical approach focusedon weighted linear regression and General LinearModels (GLMs; adjusted type III) (Shaw and Mitchell-Olds 1993). First, I used linear regression to analyzethe relationships between drainage area (log[x]-transformed) and d13C of target taxonomic groups

2013] GEOMORPHOLOGY AND d13C 611

(i.e., aquatic invertebrates [Ephemeroptera, Plecop-tera, Trichoptera, Diptera], periphyton, detritus, andfish [sculpin and trout]). Then, I used GLMs to test forrelationships between geomorphic type and d13Cvalues of my biological endpoints, with drainage areaas a covariate in the analysis. Where significant maineffects were detected, I conducted Tukey–KramerHonestly Significant Difference mean comparisontests.

I used simple linear regression to analyze relation-ships between periphyton and 2 groups: aquaticinvertebrates and fish. I used Bonferroni adjustmentfor a (a/k = 0.05/8 = 0.006, where k is the number oftests/treatments; Wright 1992) for these analyses. Iused a = 0.05 for all other tests. Except for oneextreme case (Ephemeroptera, forced step-pool), Iavoided removing statistical outliers from the analysisgiven the potential ecological importance of variabil-ity in d13C signatures. I ran all analyses in MinitabH(version 16; Minitab, State College, Pennsylvania) andJMP (version 9.0; SAS Institute, Cary, North Carolina).

Results

Detailed data and summary statistics related tostream physical characteristics are presented in Tables 1and 2. Drainage area of the reaches ranged from 1.8 km2

to 161.4 km2. Plane-bed reaches drained larger areasthan the other morphologies, were lowest in the

watersheds, and had the least canopy coverage (x=

7.5%). Step-pool and forced step-pool reaches had thegreatest slopes, and the steepest channel was at MC1(step-pool, 0.076 m/m). LW abundance was greatest inforced step-pool morphologies (x = 0.071 6 0.036 [SD]).Step-pool and forced step-pool reaches were the mostheavily canopied morphologies (Table 2).

Mean d13C values of periphyton had the greatestvariability among biological samples and ranged from235.3% (pool–riffle) to 225.0% (forced pool–riffle;Table 2). Aquatic invertebrates were dominated byherbivores, and mean d13C values ranged from229.5% (Plecoptera, pool–riffle) to 222.2% (Diptera,plane-bed). Among the aquatic invertebrate orderssampled, Trichoptera exhibited the greatest overallvariability, with a range of 10.9% (231.4% to220.5%) across all reaches. Trichoptera made up thegreatest percentage of the aquatic invertebrate assem-blage at pool–riffle reaches (34%) and the lowest atstep-pool reaches (10%). Overall, Trichoptera werepredominantly Glossosomatidae, Hydropsychidae,and Limnephilidae. Ephemeroptera were predomi-nantly Baetidae, Ephemerellidae, and Heptageniidaeand had the greatest relative abundance (22%) atpool–riffle reaches. Plecoptera (largely Chloroperli-dae) ranged from 4% of the aquatic invertebrateassemblage at plane-bed reaches to 17% at step-poolreaches. Diptera was the predominant aquatic inver-tebrate order at 3 of the 5 geomorphic types and made

TABLE 2. Means (61 SD) of measures of stream geomorphology and C signatures of aquatic insects, stream periphyton,detritus, and fish by morphology and for all study reaches combined. D50 = median bed sediment diameter for stream reach,D95 = 95th percentile bed sediment diameter for stream reach, – indicates missing data.

Variables Plane-bed Step-poolForced

step-pool Pool–riffleForced

pool–riffle All reaches

Geomorphic characteristics

Canopy cover (%) 7.5 6 3.5 73.3 6 29.3 83.0 6 19.9 40.0 6 49.5 26.7 6 2.9 54.0 6 35.9D50 (mm) 91.9 6 8.5 46.7 6 11.4 15.2 6 13.7 3.5 6 0.7 49.3 6 35.5 36.9 6 32.5D95 (mm) 305.1 6 106.1 141.0 6 41.9 104.6 6 56.2 23.0 6 22.6 186.7 6 46.2 144.1 6 95.1Drainage area (km2) 133.6 6 39.2 8.7 6 6.0 8.0 6 3.2 56.1 6 48.6 78.0 6 32.9 45.3 6 50.8Large wood (no./m) 0.004 6 0.006 0.049 6 0.061 0.071 6 0.036 0.014 6 0.004 0.006 6 0.004 0.04 6 0.04Mean bankfull depth (m) 1.1 6 0.5 0.4 6 0.1 0.6 6 0.1 0.6 6 0.1 0.6 6 0.2 0.6 6 0.3Mean bankfull width (m) 20.8 6 8.3 3.6 6 1.0 5.4 6 3.2 7.5 6 3.2 9.8 6 3.2 8.2 6 6.4Slope (m/m) 0.021 6 0.008 0.047 6 0.025 0.039 6 0.013 0.008 6 0.008 0.013 6 0.002 0.029 6 0.020Stream order 3.5 6 0.7 1.7 6 0.6 1.4 6 0.5 2.5 6 0.7 3.3 6 0.6 2.3 6 1.0

d13C (%)

Ephemeroptera 223.8 6 0.6 227.6 6 1.5 228.0 6 1.3a 229.1 6 0.1 224.2 6 0.8 226.1 6 2.7Plecoptera 224.8 6 1.0 228.0 6 2.2 226.4 6 1.0 229.5 6 0.2 224.9 6 1.3 225.6 6 2.0Trichoptera 223.5 6 1.0 227.4 6 1.4 225.9 6 2.1 227.7 6 5.3 225.4 6 0.2 226.0 6 2.3Diptera 222.2 6 0.3 227.9 6 0.0 225.2 6 0.6 228.0 6 0.1 222.7 6 0.9 224.7 6 2.2Periphyton 225.1 6 0.1 233.7 6 1.4 231.1 6 2.1 235.3 6 1.0 225.0 6 1.8 230.2 6 4.3Detritus 226.6 6 0.2 226.8 6 1.0 227.3 6 0.7 – 226.2 6 1.6 226.8 6 1.0Sculpin 222.5 6 1.0 226.1 6 1.3 224.6 6 1.3 227.2 6 0.3 223.4 6 0.9 224.8 6 1.9Trout 222.0 6 0.4 224.6 6 0.6 224.4 6 0.4 225.8 6 0.5 225.0 6 0.1 224.4 6 1.2

a One outlying observation of 220.5% was excluded from summary statistics and analysis

612 S. M. P. SULLIVAN [Volume 32

up 26%, 32%, and 44% of the invertebrate assemblag-es at forced pool–riffle, plane-bed, and step-poolmorphologies, respectively. Chironomidae was thedominant Dipteran family at all reaches. Simuliidaeand Tipulidae were present at all reaches, buttypically made up ,5 to 10% of the assemblage.d13C signatures of sculpin and trout had parallelranges (227.9% to 221.3% and 228.9% to 221.7%,respectively) and means (Table 2).

Overall, d13C values of stream periphyton, aquaticinvertebrates, and fish consumers had a distributionpattern more tightly linked to stream morphologythan to drainage area (Fig. 3A–D, Table 3). Alone,drainage area was a significant predictor of d13Cvalues for only periphyton and sculpin, explaining30% and 15% of the variation, respectively (Fig. 3A,D). The regression slopes of drainage area andEphemeroptera d13C and drainage area and troutd13C were suggestive, but nonsignificant (Fig. 3B, D).However, when included as a covariate with geomor-phic type in the full GLM models, drainage area wasnot significant in any models, a result indicating thatonce d13C signatures were explained as a function ofgeomorphic type, the contribution of drainage areabecame more subtle. In contrast, d13C of streamperiphyton, aquatic Diptera, sculpin, and trout werestrongly associated with geomorphic types (p , 0.006;R2

= 79%, 93%, 51%, 63%, respectively), withEphemeroptera and Plecoptera marginally significantafter Bonferroni correction (Table 3).

For periphyton, Ephemeroptera, Diptera, and trout,greatest 13C enrichment was observed in plane-bedmorphologies (Table 2), but d13C values of these sametaxonomic groups were consistently lowest in pool–riffle reaches. Terrestrially derived detrital d13C wasinvariant across all geomorphic types (Figs 3A, 4A).d13C values of sculpin and trout were similar bymorphology, but sculpin d13C was most variable instep-pool reaches, whereas trout d13C was mostvariable in pool–riffle morphology (Fig. 4B). Overallmodels for d13C of all consumers except Trichopterawere driven by multiple pairwise differences in d13Csignatures (Table 3).

FIG. 3. Relationship between drainage area and d13C ofperiphyton and detritus (A), Ephemeroptera and Plecoptera(B), Diptera and Trichoptera (C), and sculpin and trout (D).The regression slope was nonsignificant for detritus (p =

r

0.523, R2= 0.07), Ephemeroptera (p = 0.052, R2

= 0.30),Plecoptera (p = 0.138, R2

= 0.14), Trichoptera (p = 0.203, R2

= 0.10), Diptera (p = 0.128, R2= 0.18), and trout (p = 0.057,

R2= 0.08), but significant for periphyton (p = 0.009, R2

=

0.30) and sculpin (p = 0.015, R2= 0.15). Regression lines

shown in the figure were based on pooled means for each ofthe 15 study reaches, but statistical results were based onweighted regressions.

2013] GEOMORPHOLOGY AND d13C 613

TABLE 3. General linear models (GLMs) for d13C of aquatic insects, periphyton, detritus, and fish followed by Tukey–KramerHonestly Significant Difference (HSD) comparisons. Groups with the same letter are not significantly different (a = 0.05). p ,

0.006 indicates significance after Bonferroni adjustment. – indicates missing data.

Taxon Response/model F p HSD Mean (%)

Ephemeroptera Geomorphic type 6.25 0.014Drainage area 0.04 0.839

Plane-bed A 223.8Forced pool–riffle A B 224.2Step-pool B C 227.6Forced step-pool C 228.0Pool–riffle C 229.1

Plecoptera Geomorphic type 3.80 0.036Drainage area 0.27 0.613

Plane-bed A 224.8Forced pool–riffle A 224.9Step-pool A B 226.4Forced step-pool A B 228.0Pool–riffle B 229.5

Trichoptera Geomorphic type 1.33 0.313Drainage area 2.79 0.121

Plane-bed A 223.5Forced pool–riffle A 225.4Step-pool A 225.9Forced step-pool A 227.4Pool–riffle A 227.7

Diptera Geomorphic type 20.59 ,0.001Drainage area 0.16 0.700

Plane-bed A 222.2Forced pool–riffle A 222.7Step-pool B 225.2Forced step-pool C 227.9Pool–riffle C 228.0

Periphyton Geomorphic type 9.72 ,0.001Drainage area ,0.01 0.954

Plane-bed A 225.1Forced pool–riffle A 225.0Step-pool B 231.1Forced step-pool B 233.7Pool–riffle B 235.3

Detritus Geomorphic type 0.59 0.621Drainage area 0.51 0.477

Plane-bed A 226.2Forced pool–riffle A 226.6Step-pool A 226.8Forced step-pool A 227.3Pool–riffle –

Sculpin Geomorphic type 4.68 0.004Drainage area 0.43 0.516

Plane-bed A 222.5Forced pool–riffle A B 223.4Step-pool B C 224.6Forced step-pool C D 226.1Pool–riffle D 227.2

Trout Geomorphic type 14.68 ,0.001Drainage area 0.06 0.800

Plane-bed A 222.0Forced pool–riffle B 224.4Step-pool B C 224.6Forced step-pool B C 225.0Pool–riffle C 225.8

614 S. M. P. SULLIVAN [Volume 32

Regression relationships for aquatic invertebratesindicated that d13C of Diptera and Ephemeroptera,Plecoptera, Trichoptera (EPT; combined) closelytracked periphyton d13C across all study reaches (p ,

0.05, Fig. 5A). Sculpin d13C also closely aligned withperiphyton d13C, a result indicating that fish feeding onaquatic invertebrates were assimilating primarilyautochthonous C. The weaker relationship betweenperiphyton d13C and trout d13C (Fig. 5B) suggested thatallochthonous C was contributing significantly to troutconsumers. When I removed plane-bed and pool–rifflereaches (predicted to have high autochthonous Ccontributions to consumers) from the analysis, theinfluence of periphyton became weaker for all con-sumers: Diptera (R2

= 0.44, F = 5.43, p = 0.053), EPT(R2

= 0.59, F = 11.77, p = 0.009), sculpin (R2= 0.60, F =

10.52, p = 0.014), trout (R2 , 0.01, F = 0.0001, p =

0.994).

Discussion

Numerous investigators have studied the associa-tions between stream geomorphology and bioticcommunity structure (Peterson and Rabeni 2001, Walterset al. 2003, Sullivan et al. 2004, 2006, D’Ambrosio et al.2009), and recent work has shown considerable promisefor linking geomorphology with functional streamresponses (McNeely et al. 2006, Thorp et al. 2006, 2008,Walters et al. 2007, Poole 2010). I used stable C isotopes asa functional link between multiple stream taxa andgeomorphology. My results suggest that different streammorphologies may represent distinct patches (Poole 2002,Thorp et al. 2006) that govern the distribution andassimilation of organic-matter food sources in local streamfood webs. For stream consumers, the relative contribu-tions of autochthonous and allochthonous resources mayalso be partially mediated by stream geomorphology.

FIG. 4. Box plots of d13C by Montgomery and Buffington’s (1997) geomorphic classification types for stream periphyton anddetritus (i.e., terrestrial leaf litter) (A) and sculpin and trout species (B). Lines in boxes are medians, box ends are quartiles, andwhiskers show minimum and maximum values. * indicates outlying data points, which were not excluded from analyses.

2013] GEOMORPHOLOGY AND d13C 615

Variation in d13C and trophic links by drainage area

Variation in the isotope ratio of the inorganic Csource and the degree of fractionation during theuptake and assimilation of CO2 largely determinesd13C of C3 plants (Farquhar et al. 1982). In freshwaterecosystems, d13C of DIC can range from 226 to 0%(Mook and Tan 1991) and contributes to highlyvariable algal d13C signatures, as reported by severalinvestigators (Finlay et al. 2002, McCutchan andLewis 2002, McNeely et al. 2006). In my study,periphyton d13C values also were quite variable(Table 1), although I found a positive correlation

between periphyton 13C enrichment and drainagearea (Fig. 3A). This result is consistent with previousresearch that has linked patterns in algal stable-isotope ratios to a gradient in watershed area (Finlay2001, 2004, McNeely et al. 2006).

The C isotopic signature of an organism reflects andintegrates its assimilated diet (Peterson and Fry 1987),so predictable patterns in consumer d13C also mightbe expected along the drainage network (Finlay et al.2002, Coat et al. 2009). 13C of consumers usuallyshows a downstream enrichment pattern (Finlay 2001,2004) thought to be a function of the increased role ofalgae in downstream food webs and an increase inprimary productivity (Vannote et al. 1980, Lambertiand Steinman 1997, Finlay 2001). Among the aquaticinvertebrate and fish consumers I considered, onlysculpin (Fig. 3D) had a significant pattern of down-stream 13C enrichment, a pattern also observed byGray et al. (2004) in slimy sculpin. Ephemeroptera(Fig. 3B) and trout (Fig. 3D) showed a nonsignificantdownstream trend in 13C enrichment.

C isotopes reflect the recent feeding history of anorganism, but consumer movement can result in pooralignment between their d13C signatures and those oflocal food sources when site fidelity is low (Rasmus-sen et al. 2009). For trout in my study, this factor maybe an important caveat, although I attempted tominimize the potential confounding influence ofmovement by sampling only in late summer, duringwhich time trout are thought to be relativelysedentary (Jackson and Zydlewski 2009). Sculpinexhibit high site fidelity and limited upstreammovement (few meters) (Gray et al. 2004), andRasmussen et al. (2009) attributed high correlationwith invertebrate d13C signatures to this sedentarybehavior.

Variation in d13C among stream morphologies

My primary hypothesis was only partially support-ed by the data. My results indicate that localizedcharacteristics related to stream geomorphologyexplained more variation in d13C of stream producersand consumers than did drainage area (Table 2).Other investigators also have suggested the impor-tance of local controls on d13C of aquatic biota. Forexample, Walters et al. (2007) observed that endoge-nous physical factors related to stream geomorphol-ogy governed isotopic signatures and found consum-er 13C enrichment in high-gradient, rocky reaches(particles .64 mm) relative to in low-gradient, sand/gravel reaches (particles ,64 mm). In my study,changes in slope and substrate composition createdmarkedly different habitat and flow conditions. To a

FIG. 5. Relationship between periphyton d13C andDiptera (p = 0.0004, R2

= 0.70) and Ephemeroptera,Plecoptera, Trichoptera (EPT; combined) d13C (p = 0.0001,R2

= 0.72) (A) and d13C of trout (p = 0.024, R2= 0.36) and

sculpin (p = 0.0002, R2= 0.74) (B) across all study reaches.

616 S. M. P. SULLIVAN [Volume 32

large extent, periphyton and consumer d13C trackedthese changes.

Variability in groundwater inputs might also beexpected to influence DIC and, in turn, d13C signa-tures of periphyton and stream consumers. Forexample, Finlay (2003) found that CO2(aq) waspredominantly controlled by groundwater CO2(aq) insmall forested streams. In turn, algal d13C has beenstrongly tied to CO2(aq) (Finlay 2004). Differences ingeomorphic features among morphologies (e.g., sub-strate, flow obstacles including boulders and logs,topographic irregularities such as meanders) result indifferent pressure-head profiles and distinct poten-tials for surface–subsurface water exchange (Buffing-ton and Tonina 2009). In my study systems, step-poolreaches contained the most 13C-depleted periphyton,whereas plane-bed reaches contained the most 13C-enriched organisms (Tables 2, 3, Fig. 4A, B). Theobserved 13C enrichment in plane-bed stream reaches(characterized as relatively open with hyporheicexchange limited to localized circulation aroundobstructions, such as boulders; Buffington and Tonina2009) may be explained by a greater influence of 13C-enriched atmospheric CO2 than 13C-depleted ground-water CO2 (Rounick and Winterbourn 1986, Doucettet al. 1996). Conversely, all taxonomic groups were13C depleted in pool–riffle morphologies, which arerelatively unconfined reaches typified by increasedsubstrate porosity, more developed floodplains, andgreater hyporheic flow pathways (Montgomery et al.1996, Montgomery and Buffington 1997).

Water velocity can strongly influence d13C of DIC(Hecky and Hesslein 1995, Finlay et al. 1999, Singer etal. 2005). For instance, in laboratory (Trudeau andRasmussen 2003) and field (Rasmussen and Trudeau2007) studies, periphyton 13C was depleted at higherwater velocities. The investigators attributed thispattern to the lack of a stagnant, nutrient-depletingboundary layer in faster flows. I did not measurevelocity directly in my study, but step-pool reacheshad a mean slope of 0.047 m/m, markedly higherthan for any other morphology (Table 2), a character-istic suggesting higher velocity and potentially re-flecting depleted periphyton 13C.

In the study watersheds, forced morphologies(forced step-pool, forced pool–riffle) arose primarilyfrom LW either partially (forced pool–riffle) or fully(i.e., channel-spanning, forced step-pool) obstructingthe channel. d13C signatures from fish consumers inforced morphologies were intermediate values andoften significantly different from d13C values in fishfrom free-flowing morphologies. Unique flow char-acteristics related to the transitional nature of forcedmorphologies (Cianfrani et al. 2009) may be related to

greater within-site variation in flow and, hence, toperiphyton d13C via effects on DIC d13C. Finlay et al.(1999) showed that variation in algal C isotopes wasassociated with variation in water velocity andsurface characteristics (greater 13C enrichment inalgae in low-velocity pool habitats than in higher-velocity riffle habitats). In my study, forced pool–riffle(SD = 1.8%) and forced step-pool (SD = 2.1%)morphologies exhibited the greatest variability inperiphyton d13C signatures (Table 2, Fig. 4A), per-haps in part because of the abundance and distribu-tion of LW, which is a key element related to storageand flow in forced morphologies.

I found limited evidence that relative roughnessmay be linked to differences in d13C distribution inforced morphologies. For example, 13C enrichment ofperiphyton and sculpin (225.0%, 223.4%; respec-tively) was higher in forced pool–riffle morphologiesof higher average relative roughness (0.335) than intheir free flowing pool–riffle counterparts (235.5%,227.2%; respectively; relative roughness = 0.034;Tables 1, 3). The pattern between relative roughnessand d13C was consistent (but weaker) between forcedstep-pool and step-pool morphologies.

Variation in trophic links among stream morphologies

Periphyton and detritus often have similar d13Cvalues, which confound attempts to quantify relativecontributions to stream food webs. In my study, anaverage detrital d13C value of 226.8% was indicativeof allochthonous/terrestrial origin (France 1995),although there was not significant separation betweenperiphyton and detrital d13C at all geomorphic types(Table 2, Fig. 4A).

Plane-bed reaches were lowest in the watershedand had the largest drainage areas (Table 2, Fig. 2).Given that autotrophs are thought to be more criticalto downstream food webs (Finlay 2001, Rosi-Marshalland Wallace 2002), I anticipated that stream consum-ers in these reaches would have the greatest relianceon autochthonous food sources. I also anticipated thatconsumers in pool–riffle reaches would show agreater reliance on autochthonous C. Exclusion ofthese 2 geomorphic types from the regression analy-ses of d13C of periphyton and stream consumers(Fig. 5A, B) significantly weakened the relationships(e.g., R2 of Diptera decreased by 0.26). For trout,removal of these 2 geomorphic types led to anonsignificant relationship (F = 0.0001, p = 0.994),indicating an increasing reliance on allochthonousC at other geomorphic types. In addition to greateravailability of detrital matter, the importance ofterrestrially derived arthropod fluxes to surface-

2013] GEOMORPHOLOGY AND d13C 617

feeding fishes in canopied streams (e.g., trout species)is well documented (Garman 1991, Kawaguchi et al.2003, Romaniszyn et al. 2007, Sweka and Hartman2008). For example, in my study, step-pool reacheswere typically heavily canopied and found in theupper M of the watershed, whereas pool–rifflereaches were characterized by more open canopiesand were more variable relative to watershed position(Fig. 2).

Finlay et al. (1999) and Finlay (2003) succeeded indiscriminating invertebrate and fish d13C signaturesbetween pools and riffles in streams. In my study, thescale of interest was the channel reach, whichspanned up to 600 m and in which fish consumersfed across multiple flow habitats. In this case, fishmight be expected to smooth out (sensu Rasmussenand Trudeau 2007) the within-site variability in Csignals of food resources. However, seasonal shifts indiet probably are significant (Woodward and Hildrew2002, McCarthy et al. 2009). The role of detritusappeared to be minimal in certain geomorphic types,but the fine particulate portion of the organic materialwas not sampled (FPOM) and probably would havebeen important for aquatic invertebrates.

Synthesis and conclusions

Walters et al. (2007) presented a useful synthesis ofthe ecological significance of d13C shifts as they relateto the interplay of C trophic enrichment, spatialeffects, and the assimilation of resources by consum-ers. In my study, the maximum difference in d13C(Dd13C) among morphologies for all taxonomicgroups was observed between pool–riffle and plane-bed morphologies. Expected trophic fractionationfactors would suggest a 0.3% increase for aquaticinvertebrates and a 1.3% increase for fish (McCutchanet al. 2003). Sculpin and trout muscle tissues fromplane-bed reaches were enriched by 4.7% and 3.8%,respectively, relative to in pool–riffle reaches andtracked increased periphyton enrichment at plane-bed reaches. Differences in C isotopes observed insculpin and trout are particularly noteworthy giventhat the variability among habitat patches at the baseof the food web might be expected to be overwhelmedby the time they are transferred to larger-bodiedconsumers. Aquatic invertebrates were similarlyenriched, ranging from Dd13C of 4.2% betweenplane-bed and pool–riffle reaches for Trichoptera, toDd13C of 5.8% for Diptera. Differences at finer levelsof taxonomic resolution for aquatic invertebratesmight be expected to enhance these results. Forexample, dominant Trichoptera families in my studywere variable in their functional feeding roles

(scrapers, collector-gatherers, shredders), and uniquepatterns in d13C signatures might have been observedat the family level.

Improved understanding of and ability to relatestream morphology and processes to biologicalfunction in mountain drainages represents an impor-tant contribution to our conceptual understanding ofstream ecosystems and the intricate, spatially explicitrelationships between aquatic biota and their envi-ronment (see Isaak et al. 2007, Murray and Innes2009). In my study, periphyton and consumer (otherthan Trichoptera) d13C appeared to be tied morestrongly to stream-reach geomorphology than to agradient of ecosystem size. My results stronglysupport geomorphic patch dynamics perspectives(Pringle et al. 1988, Montgomery 1999, Poole 2002,Thorp et al. 2006) and offer evidence related to howstream geomorphology may be linked to ecosystemprocesses.

My results also reinforce the utility of Montgomeryand Buffington’s classification scheme. Substrate,flow, and the spatial arrangement of geomorphicpatterns are integral to many stream classificationsystems (see Cianfrani et al. 2009). Thus, otherclassification systems may be helpful in elucidatingvariation in stream C and energy sources to loticconsumers. An intrinsic synthesis takes place when areach is categorized by its dominant morphology.This synthesis incorporates myriad variables that arenot easily measured and that lose functional signifi-cance when separated from the whole. Fish and otheraquatic organisms probably experience their environ-ment by integrating component patches, so classifica-tion systems may capture and synthesize criticalinformation related to spatial effects on C distribution.

The spatial variability of mountain drainagesheightens the importance of understanding ecologicalpatterns and processes at a local, reach scale. Myresults add important resolution and additionalnuance to patch dynamics concepts, but widerapplication of these results will require testing acrossa broad range of settings. Empirically linking thesefindings to pattern and process across spatial scaleswill represent a critical step in an integrated physico-biological framework for mountain drainages and willaid in forecasting the effects of anthropogenicalterations on stream food webs.

Acknowledgements

I would like to thank Jeff Braatne; PotlatchCorporation; and the Department of Fish and Wild-life Resources, University of Idaho for support duringthe initial stages of the project. My research was

618 S. M. P. SULLIVAN [Volume 32

supported by the National Research Initiative of theUS Department of Agriculture Cooperative StateResearch, Education, and Extension Service, grantnumber 2003-01264, the University of Idaho, Collegeof Natural Resources, and The Ohio State University,School of Environment and Natural Resources. Ithank all coworkers who assisted in field andlaboratory work, especially Adam Kautza, RyanMann, Danielle Vent, Paul Charpentier, and MatthewMason. I also thank Edith Seier for her valuable inputinto the statistical approach and James Bauer andanonymous referees for their comments on earlierdrafts of the manuscript.

Literature Cited

BEHNKE, R. J. 2002. Trout and salmon of North America. TheFree Press, New York.

BUFFINGTON, J. M., AND D. TONINA. 2009. Hyporheic exchangein mountain rivers II: effects of channel morphology onmechanics, scales, and rates of exchange. GeographyCompass 3:1038–1062.

CIANFRANI, C. M., W. C. HESSION, AND M. C. WATZIN. 2004.Evaluating aquatic habitat quality using channel mor-phology and watershed-scale modeling techniques.World Water and Environmental Resources Congress.American Society of Civil Engineers, Reston, Virginia.

CIANFRANI, C. M., S. M. P. SULLIVAN, W. C. HESSION, AND M. C.WATZIN. 2009. Mixed stream channel morphologies:implications for fish community diversity. Aquatic Con-servation: Marine and Freshwater Ecosystems 19:147–156.

COAT, S., D. MONTI, C. BOUCHON, AND G. LEPOINT. 2009.Trophic relationships in a tropical stream food webassessed by stable isotope analysis. Freshwater Biology54:1028–1041.

D’AMBROSIO, J. L., L. R. WILLIAMS, J. D. WITTER, AND A. WARD.2009. Effects of geomorphology, habitat, and spatiallocation on fish assemblages in a watershed in Ohio,USA. Environmental Monitoring and Assessment 148:325–341.

DAWSON, J. J. C., M. F. BILLETT, AND D. HOPE. 2001. Diurnalvariations in the carbon chemistry of two acidic peat-land streams in north-east Scotland. Freshwater Biology46:1309–1322.

DELONG, M. D., AND J. H. THORP. 2006. Significance ofinstream autotrophs in trophic dynamics of the UpperMississippi River. Oecologia (Berlin) 147:76–85.

DOUCETT, R. R., G. POWER, D. R. BARTON, R. J. DRIMMIE, AND

R. A. CUNJAK. 1996. Stable isotope analysis of nutrientpathways leading to Atlantic salmon. Canadian Journalof Fisheries and Aquatic Sciences 53:2058–2066.

DUNHAM, J. B., S. B. ADAMS, R. E. SCHROETER, AND D. C.NOVINGER. 2002. Alien invasions in aquatic ecosystems:toward an understanding of brook trout invasions andpotential impacts on inland cutthroat trout in westernNorth America. Reviews in Fish Biology and Fisheries12:373–391.

FARQUHAR, G. D., M. H. O’LEARY, AND J. A. BERRY. 1982. Onthe relationship between carbon isotope discriminationand the intercellular carbon-dioxide concentration inleaves. Australian Journal of Plant Physiology 9:121–137.

FINLAY, J. C. 2001. Stable carbon isotope ratios of river biota:implications for energy flow in lotic food webs. Ecology84:1052–1064.

FINLAY, J. C. 2003. Controls of streamwater dissolvedinorganic carbon dynamics in a forested watershed.Biogeochemistry 62:231–252.

FINLAY, J. C. 2004. Patterns and controls of lotic algal stablecarbon isotope ratios. Limnology and Oceanography 49:850–861.

FINLAY, J. C., S. KHANDWALA, AND M. E. POWER. 2002. Spatialscales of carbon in a river food web. Ecology 83:1845–1859.

FINLAY, J. C., M. E. POWER, AND G. CABANA. 1999. Effects ofwater velocity on algal carbon isotope ratios: implica-tions for river food web studies. Limnology andOceanography 44:1198–1203.

FORTIN, M. J., AND M. R. T. DALE. 2005. Spatialanalysis: a guide for ecologists. Cambridge UniversityPress, Cambridge, UK.

FRANCE, R. 1995. Critical examination of stable-isotopeanalysis as a means for tracing carbon pathways instream ecosystems. Canadian Journal of Fisheries andAquatic Sciences 52:651–656.

GARMAN, G. 1991. Use of terrestrial arthropod prey by astream-dwelling cyprinid fish. Environmental Biologyof Fishes 30:325–331.

GRAY, M. A., R. A. CUNJAK, AND K. R. MUNKITTRICK. 2004. Sitefidelity of slimy sculpin (Cottus cognatus): insights fromstable carbon and nitrogen analysis. Canadian Journalof Fisheries and Aquatic Sciences 61:1717–1722.

HARRELSON, C. C., J. P. POTYONDY, AND C. L. RAWLINS. 1994.Stream channel reference sites: an illustrated guide tofield technique. General Technical Report RM-245. USDepartment of Agriculture Forest Service, Fort Collins,Colorado.

HECKY, R. E., AND R. H. HESSLEIN. 1995. Contributions ofbenthic algae to lake food webs as revealed by stableisotope analysis. Journal of the North AmericanBenthological Society 14:631–653.

HILDERBRAND, R. H., AND J. L. KERSHNER. 2004. Influence ofhabitat type on food supply, selectivity, and diet overlapof Bonneville cutthroat trout and nonnative brook troutin Beaver Creek, Idaho. North American Journal ofFisheries Management 24:33–40.

HILL, W. R., AND R. G. MIDDLETON. 2006. Changes in carbonstable isotope ratios during periphyton development.Limnology and Oceanography 51:2360–2369.

ISAAK, D. J., R. F. THUROW, B. E. RIEMAN, AND J. B. DUNHAM.2007. Chinook salmon use of spawning patches: relativeroles of habitat quality, size, and connectivity. Ecolog-ical Applications 17:352–364.

JACKSON, C. A. L., AND J. ZYDLEWSKI. 2009. Summer move-ments of sub-adult brook trout, landlocked Atlantic

2013] GEOMORPHOLOGY AND d13C 619

salmon, and smallmouth bass in the Rapid River,Maine. Journal of Freshwater Ecology 24:567–580.

KAWAGUCHI, Y., Y. TANIGUCHI, AND S. NAKANO. 2003. Terres-trial invertebrate inputs determine the local abundanceof stream fishes in a forested stream. Ecology 84:701–708.

KONDOLF, G. M., AND E. M. MICHELI. 1995. Evaluating streamrestoration projects. Environmental Management 19:1–15.

LAMBERTI, G. A., AND A. D. STEINMAN. 1997. A comparison ofprimary production in stream ecosystems. Journal of theNorth American Benthological Society 16:95–103.

LANCASTER, J., AND S. WALDRON. 2001. Stable isotope values oflotic invertebrates: sources of variation, experimentaldesign, and statistical interpretation. Limnology andOceanography 46:723–730.

LEOPOLD, L. B., AND W. B. LANGBEIN. 1966. River meanders.Scientific American 214:60–70.

LEOPOLD, L. B., AND T. MADDOCK. 1953. The hydraulicgeometry of stream channels and some physiographicimplications. U.S. Geological Survey Professional Paper252. US Geological Survey, Reston, Virginia.

LEOPOLD, L. B., M. G. WOLMAN, AND J. P. MILLER. 1964. Fluvialprocesses in geomorphology. W. H. Freeman and Co.,San Francisco, California.

LOWE, W. H., G. E. LIKENS, AND M. E. POWER. 2006. Linkingscales in stream ecology. BioScience 56:591–597.

MCCARTHY, S. G., J. J. DUDA, J. M. EMLEN, G. R. HODGSON, AND

D. A. BEAUCHAMP. 2009. Linking habitat quality withtrophic performance of steelhead along forest gradientsin the South Fork Trinity River Watershed, California.Transactions of the American Fisheries Society 138:506–521.

MCCUTCHAN, J. H., AND W. M. LEWIS, JR. 2002. Relativeimportance of carbon sources for macroinvertebrates ina Rocky Mountain stream. Limnology and Oceanogra-phy 47:742–752.

MCCUTCHAN, J. H., W. M. LEWIS, JR, C. KENDALL, AND C. C.MCGRATH. 2003. Variation in trophic shift for stableisotope ratios of carbon, nitrogen, and sulfur. Oikos 102:378–390.

MCGRATH, C. C., AND W. M. LEWIS, JR. 2007. Competition andpredation as mechanisms for displacement of greenbackcutthroat trout by brook trout. Transactions of theAmerican Fisheries Society 136:1381–1392.

MCNEELY, C., S. M. CLINTON, AND J. M. ERBE. 2006. Landscapevariation in C sources of scraping primary consumers instreams. Journal of the North American BenthologicalSociety 25:787–799.

MERRITT, R. W., AND K. W. CUMMINS (EDITORS). 1996. Anintroduction to the aquatic insects of North America. 3rd

edition. Kendall/Hunt, Dubuque, Iowa.MINSHALL, G. W. 1978. Autotrophy in stream ecosystems.

BioScience 28:767–770.MONTGOMERY, D. R. 1999. Process domains and the river

continuum concept. Journal of the American WaterResources Association 35:397–410.

MONTGOMERY, D. R., T. B. ABBE, J. M. BUFFINGTON, N. P.PETERSON, K. M. SCHMIDT, AND J. D. STOCK. 1996.

Distribution of bedrock and alluvial channels in forestedmountain drainage basins. Nature 381:587–589.

MONTGOMERY, D. R., AND J. M. BUFFINGTON. 1997. Channel-reach morphology in mountain drainage basins. Geo-graphical Society of America Bulletin 109:596–611.

MONTGOMERY, D. R., J. M. BUFFINGTON, R. D. SMITH, K. M.SCHMIDT, AND G. PESS. 1995. Pool spacing in forestchannels. Water Resources Research 31:1097–1105.

MOOK, W. G., AND F. C. TAN. 1991. Stable carbon isotopes inrivers and estuaries. Pages 245–264 in E. Degens, S.Kempe, and J. Richey (editors). Biogeochemistry ofmajor world rivers. John Wiley, New York.

MOSLEY, M. P. 1987. The classification and characterization ofrivers. Pages 295–320 in K. Richards (editor). Riverchannels, environment, and process. Basil BlackwellPublishing, Oxford, UK.

MURRAY, S., AND J. L. INNES. 2009. Effects of environment onfish species distributions in the Mackenzie Riverdrainage basin of northeastern British Columbia, Can-ada. Ecology of Freshwater Fish 18:183–196.

PETERSON, B., AND B. FRY. 1987. Stable isotopes in ecosystemstudies. Annual Review of Ecology and Systematics 18:293–320.

PETERSON, J. T., AND C. F. RABENI. 2001. Evaluating thephysical characteristics of channel units in an Ozarkstream. Transactions of the American Fisheries Society130:898–910.

PINNEGAR, J., AND N. POLUNIN. 1999. Differential fractionationof d13C and d15N among fish tissues: implication for thestudy of trophic interactions. Functional Ecology 13:225–231.

POOLE, G. C. 2002. Fluvial landscape ecology: addressinguniqueness within the river continuum. FreshwaterBiology 47:641–660.

POOLE, G. C. 2010. Stream hydrogeomorphology as aphysical science basis for advances in stream ecology.Journal of the North American Benthological Society 29:12–25.

POWER, M. E., R. J. STOUT, C. E. CUSHING, P. P. HARPER, F. R.HAUER, W. J. MATTHEWS, P. B. MOYLE, B. STATZNER, AND I.R. WAIS DE BADGEN. 1988. Biotic and abiotic controls inriver and stream communities. Journal of the NorthAmerican Benthological Society 7:456–479.

PRINGLE, C. M., R. J. NAIMAN, G. BRETSCHKO, J. R. KARR, M. W.OSWOOD, J. R. WEBSTER, R. L. WELCOMME, AND M. J.WINTERBOURN. 1988. Patch dynamics in lotic systems: thestream as a mosaic. Journal of the North AmericanBenthological Society 7:503–524.

RALEIGH, R. F., T. HICKMAN, R. C. SOLOMON, AND P. C. NELSON.1984. Habitat suitability information: rainbow trout.FWS/OBS-82/10.60. US Fish and Wildlife Service,Washington, DC.

RASMUSSEN, J. B., AND V. TRUDEAU. 2007. Influence of velocityand chlorophyll standing stock on periphyton delta d13Cand d15N in the Ste. Marguerite River system, Quebec.Canadian Journal of Fisheries and Aquatic Sciences 64:1370–1381.

RASMUSSEN, J. B., V. TRUDEAU, AND G. MORINVILLE. 2009.Estimating the scale of fish feeding movements in rivers

620 S. M. P. SULLIVAN [Volume 32

using delta C-13 signature gradients. Journal of AnimalEcology 78:674–685.

ROMANISZYN, E. D., J. J. HUTCHENS, AND J. B. WALLACE. 2007.Aquatic and terrestrial invertebrate drift in southernAppalachian Mountain streams: implications for troutfood resources. Freshwater Biology 52:1–11.

ROSI-MARSHALL, E. J., AND J. B. WALLACE. 2002. Invertebratefood webs along a stream resource gradient. FreshwaterBiology 47:129–141.

ROUNICK, J. S., AND M. J. WINTERBOURN. 1986. Stable carbonisotopes and carbon flow in ecosystems. BioScience 36:171–177.

SCHINDLER, D. E., AND S. C. LUBETKIN. 2004. Using stableisotopes to quantify material transport in food webs.Pages 25–42 in G. A. Polis, M. E. Power, and G. R. Huxel(editors). Food webs at the landscape level. Universityof Chicago Press, Chicago, Illinois.

SHAW, R. G., AND T. MITCHELL-OLDS. 1993. ANOVA forunbalanced data: an overview. Ecology 74:1638–1645.

SINGER, G. A., M. PANZENBOCK, G. WEIGELHOFER, C. MARCHE-

SANI, J. WARINGER, W. WANEK, AND T. J. BATTIN. 2005. Flowhistory explains temporal and spatial variation ofcarbon fractionation in stream periphyton. Limnologyand Oceanography 50:706–712.

SULLIVAN, S. M. P., M. C. WATZIN, AND W. C. HESSION. 2004.Understanding stream geomorphic state in relation toecological integrity: evidence using habitat assessmentsand macroinvertebrates. Environmental Management34:669–683.

SULLIVAN, S. M. P., M. C. WATZIN, AND W. C. HESSION. 2006.Influence of stream geomorphic condition on fishcommunities in Vermont, U.S.A. Freshwater Biology51:1811–1826.

SWEKA, J. A., AND K. J. HARTMAN. 2008. Contribution ofterrestrial invertebrates to yearly brook trout consump-tion and growth. Transactions of the American FisheriesSociety 137:224–235.

THORNE, C. R. 1997. Channel types and morphologicalclassification. John Wiley and Sons, Chichester, UK.

THORP, J. H., M. C. THOMS, AND M. D. DELONG. 2006. Theriverine ecosystem synthesis: biocomplexity in rivernetworks across space and time. River Research andApplications 22:123–157.

THORP, J. H., M. C. THOMS, AND M. D. DELONG. 2008. Theriverine ecosystem synthesis. Academic Press, Boston,Massachusetts.

TOWNSEND, C. R. 1989. The patch dynamics concept of streamcommunity ecology. Journal of the North AmericanBenthological Society 8:36–50.

TOWNSEND, C. R. 1996. Concepts in river ecology: pattern andprocess in the catchment hierarchy. Archiv fur Hydro-biologie 113:3–21.

TRUDEAU, V., AND J. B. RASMUSSEN. 2003. The effect of watervelocity on stable carbon and nitrogen isotope signa-tures of periphyton. Limnology and Oceanography 48:2194–2199.

USEPA (US ENVIRONMENTAL PROTECTION AGENCY). 2002. LevelIII Ecoregions. US Environmental Protection Agency,Washington, DC. (Available from: http://www.epa.gov/wed/pages/ecoregions/level_iii.htm)

VANNOTE, R. L., G. W. MINSHALL, K. W. CUMMINS, J. R. SEDELL,AND C. E. CUSHING. 1980. The river continuum concept.Canadian Journal of Fisheries and Aquatic Sciences 37:130–137.

WALTERS, D. M., K. M. FRITZ, AND D. L. PHILLIPS. 2007. Reach-scale geomorphology affects organic matter and con-sumer d13C in a forested Piedmont stream. FreshwaterBiology 52:1105–1119.

WALTERS, D. M., D. S. LEIGH, M. C. FREEMAN, B. J. FREEMAN,AND C. M. PRINGLE. 2003. Geomorphology and fishassemblages in a Piedmont river basin, USA. FreshwaterBiology 48:1950–1970.

WOHL, E., AND D. M. MERRITT. 2008. Reach-scale channelgeometry of mountain streams. Geomorphology 93:168–185.

WOLMAN, M. G. 1954. A method of sampling coarse river-bed material. Transactions of the American GeophysicalUnion 35:951–956.

WOODWARD, G., AND A. G. HILDREW. 2002. Food web structurein riverine landscapes. Freshwater Biology 47:777–798.

WRIGHT, S. P. 1992. Adjusted p-values for simultaneousinference. Biometrics 48:1005–1013.

WU, J., AND O. LOUCKS. 1995. From balance of nature tohierarchical patch dynamics: a paradigm shift inecology. Quarterly Review of Biology 70:439–466.

WYDOSKI, R. S., AND R. R. WHITNEY. 2003. Inland fishes ofWashington. 2nd edition. American Fisheries Society,Bethesda, Maryland.

Received: 1 July 2012Accepted: 11 February 2013

2013] GEOMORPHOLOGY AND d13C 621