Dependence of stream predators on terrestrial prey fluxes ...angelo.berkeley.edu/wp-content/uploads/Atlas_2013_Ecosphere.pdf(Holt 1984, Polis et al. 1997, Henschel et al. 2001, Murakami

Dependence of stream predators on terrestrial prey fluxes:food web responses to subsidized predation

WILLIAM I. ATLAS,� WENDY J. PALEN, DANIELLE M. COURCELLES, ROBIN G. MUNSHAW,

AND ZACHARY L. MONTEITH

Earth to Ocean Research Group, Department of Biological Sciences, Simon Fraser University,Burnaby, British Columbia V5A1S6 Canada

Citation: Atlas, W. I., W. J. Palen, D. M. Courcelles, R. G. Munshaw, and Z. L. Monteith. 2013. Dependence of stream

predators on terrestrial prey fluxes: food web responses to subsidized predation. Ecosphere 4(6):69. http://dx.doi.org/10.

1890/ES12-00366.1

Abstract. Resource subsidies in the form of energy, materials, and organisms can support the

productivity of recipient ecosystems. When subsidies increase the abundance of top predators, theory

predicts that top-down interactions will be strengthened. However, the degree to which subsidies intensify

predation should be constrained by the strength of interactions between predators and their prey. To test

the potential for subsidies to drive strong top-down control by two stream predators, steelhead

(Oncorhynchus mykiss) and Pacific giant salamander (Dicamptodon tenebrosus) we reduced terrestrial prey

and manipulated the presence of predators in 32 stream reaches. Prey subsidies supported elevated growth

of predatory steelhead in our study system and in the absence of allochthonous prey steelhead experienced

a 187% reduction in growth. Despite the high biomass of subsidized predators, there was little support for

strong top-down control of herbivore biomass, or a trophic cascade as measured by changes in AFDM and

chlorophyll-a. This result was consistent across subsidy treatments, suggesting that predatory steelhead are

unable to increase exploitation of aquatic prey in the absence of terrestrial prey subsidies. The potential for

top-down control was apparently limited by the fact that most (82%) herbivores in our study system were

armored and relatively invulnerable to predation. These results demonstrate the potential importance of

behavioral and morphological adaptations that can temper predator prey interactions in highly subsidized

ecosystems.

Key words: Dicamptodon tenebrosus; Eel River, California, USA; experimental ecology; food webs; Oncorhynchus mykiss;

resource subsidies; stream ecosystems; terrestrial-aquatic linkages; trophic cascades.

Received 27 November 2012; revised 6 March 2013; accepted 11 March 2013; final version received 13 May 2013;

published 6 June 2013. Corresponding Editor: W. Cross.

Copyright: � 2013 Atlas et al. This is an open-access article distributed under the terms of the Creative Commons

Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the

original author and source are credited. http://creativecommons.org/licenses/by/3.0/

� E-mail: [email protected]

INTRODUCTION

Resource subsidies, the movement of nutrients,organic material, and prey from adjacent habi-tats, support elevated primary and secondaryproduction in many ecosystems, and play animportant role in mediating the dynamics ofrecipient food webs (Polis et al. 1997). Theresponse of food webs to resource subsidies will

depend in part on the trophic level at whichsubsidies enter the recipient food web (Marczaket al. 2007). Subsidies of nutrients and organicmaterial that support higher productivity at thebase of the food web can influence communitycomposition (Hocking and Reynolds 2011) andpatterns of food web regulation (Polis and Hurd1996, Spiller et al. 2010). Subsidies of prey cansupport elevated biomass of predators (Rose and

v www.esajournals.org 1 June 2013 v Volume 4(6) v Article 69

Polis 1998, Hilderbrand et al. 1999, Sabo andPower 2002, Wipfli and Gregovich 2002, Kawa-guchi et al. 2003), creating the potential forstrengthened top-down control of in situ preythrough the process of apparent competition(Holt 1984, Polis et al. 1997, Henschel et al.2001, Murakami and Nakano 2002). In suchinstances, subsidized predators may elicit chang-es in the biomass and composition of multipletrophic levels within the recipient food web (Polisand Hurd 1996, Henschel et al. 2001, Leroux andLoreau 2008).

Although the ability of prey subsidies toinfluence patterns of top-down control is welldocumented in the literature, the response ofcommunities to prey subsidies will depend onprey traits as well as the demographic andbehavioral responses of predators to resourcesubsidies. For a cross ecosystem flux of prey toinfluence the dynamics of a recipient food webvia apparent competition, subsidies must elicit ademographic response within the predator guildsupporting higher biomass of predators thanwould otherwise be sustained by in situ produc-tivity, and increased predator biomass mustresult in greater predation on in situ prey (Poliset al. 1997). Thus, the ability of subsidizedpredation to elicit changes in trophic levelbiomass will also depend acutely on the compo-sition of the community, and ability of predatorsto exploit local prey. Alternatively, if subsidizedpredators exhibit strong selectivity for highquality allochthonous prey (Nakano et al.1999a, Marcarelli et al. 2011), the effect ofsubsidies may be positive for in situ prey andtop-down control may be weaker than expectedin recipient food webs. In such cases, interrupt-ing the availability of terrestrial prey subsidiesmay result in prey switching with increasedexploitation of in situ prey (Sabo and Power2002, Spiller et al. 2010) triggering trophiccascades (Nakano et al. 1999b, Baxter et al. 2004).

In small tributary stream ecosystems wherelight availability is often limited by denseoverhead tree canopies, terrestrial resource sub-sidies play an important role in supporting bioticcommunities (Vannote et al. 1980, Richardson etal. 2010). Inputs of dissolved organic carbon, andleaf litter from the surrounding terrestrial envi-ronment increase the productivity of streamsfrom the bottom-up (Wallace et al. 1997, Finlay

2001), and subsidies of terrestrial invertebratesoften support elevated biomass of predatory fish(Nakano et al. 1999a, Kawaguchi and Nakano2001). Predatory fish can play a key role inmediating dynamics in river food webs, and mayexert strong top-down control resulting introphic cascades (Power et al. 1985, Power 1990,Nakano et al. 1999b). However, the top-downeffects of predation by fish are spatially andtemporally variable (Power 1992, Wootton et al.1996), and this variability may be partially due tothe composition of the primary consumer guild.Several previous studies of fish-induced trophiccascades in river food webs have been conductedin large mainstem rivers (Power et al. 1985,Power 1990) where food webs are primarilysupported by in situ primary production (Finlay2001). Theory predicts that trophic cascadesshould be particularly strong in highly subsi-dized tributaries, where the majority of ecosys-tem energy is derived from allochtonous sources(Finlay 2001), resulting in predator biomass thatis disproportionately high relative to the in situproductivity of the stream ecosystem (Polis et al.1997, Leroux and Loreau 2008).

Here we experimentally tested the response ofa highly subsidized tributary stream food webcomprised of three trophic levels to changes inthe short term availability of resource subsidies.Using a three-way factorial experiment we testedthe strength of top-down control by two preda-tors, juvenile steelhead trout (Oncorhynchus my-kiss) and Pacific giant salamander (Dicamptodontenebrosus), how the effects of these predators aremediated by the availability of terrestrial prey,and the degree to which predator biomass wassupported by the availability of terrestrial preyresources. We predicted that over the course ofthe short-term experiment (;9 weeks) the pres-ence of both predator species would depress thebiomass of herbivorous aquatic invertebrates,releasing primary producers from grazer control,thereby inducing a trophic cascade as indicatedby an increase in primary producer biomass inthe presence of predators (Fig. 1). Furthermore,we predicted that, if terrestrial prey subsidiessupported elevated biomass of steelhead result-ing in stronger top down control of aquaticconsumers, we would see reductions in steelheadgrowth in the absence of terrestrial prey subsi-dies which should result in a weakening of


ATLAS ET AL.

trophic cascades over time. Alternatively, if

predatory steelhead preferentially exploit terres-

trial prey resulting in a positive effect of resource

subsidies on in situ consumers, we predicted that

experimental reductions in terrestrial prey in-fall

would necessitate a shift in predation by trout

towards a more aquatic prey base, depressing

aquatic herbivores and amplifying trophic cas-

cades relative to treatments where terrestrial prey

were available. Patterns of top-down regulation

by salamanders, which feed primarily on benthic

aquatic prey (Parker 1994), were not expected to

Fig. 1. Food web diagram depicting the predicted and observed effects of terrestrial resource subsidies in the

Fox Creek food web. Gray arrows represent the degree of predator dependence on terrestrial prey subsidies and

black arrows depict the strength of top-down control over a given biomass pool. The authors predicted that

subsidized predators would exert strong top-down control over aquatic consumer biomass, such that predator

removal would result in a release of primary consumers from predator control triggering a trophic cascade with

reductions in primary producer biomass.


ATLAS ET AL.

change in the absence of terrestrial prey subsi-dies. However, the predicted shift in resource useby steelhead trout was expected to increasecompetition between steelhead and salamanders,reducing the growth of steelhead and the benthicdwelling salamanders. Contrasting these predic-tions, if morphological and behavioral attributesof the aquatic invertebrate community limitpredation by steelhead and salamanders, wepredicted that neither predator species wouldinitiate a trophic cascade, and steelhead wouldexperience dramatic reductions growth in theabsence of terrestrial prey.

METHODS

Study siteWe manipulated 32 reaches of Fox Creek, a

tributary of the South Fork Eel River (2.8 km2

drainage area, 3984304500 N, 12383804000 W) pro-tected within the Angelo Coast Range Reserve inMendocino Co., California, and part of theUniversity of California Natural Reserve system.The creek is relatively high gradient and isdominated by step-pool channel morphology.Rainfall is highly seasonal with most rainfalloccurring during winter and a protracted sum-mer dry season. Consequently, winter base flowsare typically an order of magnitude higher thansummer base flow. Peak stream temperatures areobserved from late July to early August. Verte-brate predators within Fox Creek include Pacificgiant salamander (D. tenebrosus) as well as bothyoung of the year (YOY) and age 1 and older (1þ)juvenile steelhead (O. mykiss). While both resi-dent and anadromous O. mykiss may be presentin Fox Creek, the majority are thought to spend 2years in freshwater before migrating to sea.However, the life-history of any individual fishcould not be visually distinguished and all O.mykiss in Fox Creek are therefore referred to assteelhead.

Experimental protocolFor nine weeks (July–August) during summer

2010, we manipulated replicate reaches of FoxCreek to test the role of terrestrial subsidies inmediating the top-down effects of juvenilesteelhead trout and Pacific giant salamanderson trophic dynamics in a stream food web. Weselected 32 comparable pools throughout the

length of Fox Creek that is accessible to fish (1.3km). These pools represented habitat withsufficient area (.7 m2) and with adequate depth(.0.25 m) to support 1þ steelhead and largesalamanders throughout the summer. Pools wererandomly assigned to one of 8 treatmentsresulting from the factorial combination of thepresence or absence of both predator species andthe availability of terrestrial prey subsidies (Table1). Each treatment was replicated 4 times.However, due to an initial assignment error onetreatment, reduced terrestrial subsidy with bothpredators was replicated 5 times and another,reduced terrestrial subsidy with salamanderpredators only, was replicated 3 times. Fencesburied in the stream substrate and extendingabove the surface of the water were constructedat the top of the upstream riffle, and below eachpool (3 mm Vexar mesh) to limit the movementof animals during the study. Upstream rifflehabitats were included to ensure that driftingbenthic invertebrates from immediately up-stream of the focal pool remained available topredators. For experimental reaches assigned toreduced terrestrial subsidy treatments, we in-stalled covers extending over the entire reachconstructed of transparent polyethylene plasticand window screen stretched over PVC hoops.Covers were designed to allow maximum lightpenetration (,8% reduction in visible light) andventilation, while blocking terrestrial organicmatter and invertebrate in-fall. Upstream fencesprevented the downstream drift of most organicmaterial and terrestrial invertebrates howeversmall invertebrates may have been able to passthrough the mesh. Two small (,10 cm indiameter) ventilation holes were cut into theapex of each cover to allow emerging aquaticinsects to escape.

Experimental contrasts in the vertebrate pred-ator community were established by first remov-ing all O. mykiss and D. tenebrosus through acombination of snorkel, hand capture, andelectrofishing until no new animals were cap-tured. All animals were weighed (mg), measured(mm), tagged, and released according to theassigned treatment. Pools were then visuallysearched via snorkeling in the weeks followingthe start of the experiment to confirm therobustness of the predator contrasts. The smallestsize at which salamanders regularly consume


ATLAS ET AL.

YOY steelhead is thought to be 100 mm totallength (TL) (Parker 1993), and this was chosen asthe minimum size cut off in our predatorysalamander treatments. Salamanders smallerthan this threshold were not manipulated.Steelhead were divided into two age groupsbased on size; young of year (YOY) and 1þsteelhead (.85 mm), both marked with smalladipose fin clips. Predatory salamanders (.100mm TL) and 1þ steelhead and were taggedindividually with passive integrated transponder(PIT) tags (HPT8, 8.4 mm, Biomark; Boise, ID,USA). Experimental densities for YOY (1.25 fish/m2) and 1þ steelhead (0.26 fish/m2) were stan-dardized to mean densities previously observedin Fox Creek (W. Palen, unpublished data), andsalamanders density (.100 mm) was set basedon densities observed during the first two days ofsampling (0.52 salamanders/m2). At the conclu-sion of the 9-week experiment, each experimentalunit was searched as above, and all O. mykiss,and D. tenebrosus were weighed, measured, andreleased. Experimental contrasts in 1þ steelheaddensity were well maintained throughout thecourse of experiment with final densities closelymatching initial treatments (þsteelhead treat-ment: 0.29 fish/m2; �steelhead treatment: 0.06fish/m2). However, salamanders showed a highdegree of emigration and immigration possiblydue to their ability to move short distances overland and final densities did not reflect initialtreatment contrasts (þsalamander treatment: 0.34salamanders/m2; �salamander treatment: 0.39salamanders/m2). Key response variables forvertebrates included batch growth estimated asthe change in the mean mass for young-of-yearsteelhead in each pool, as well as the growth ofindividually marked 1þ steelhead and salaman-ders.

Pan traps and leaf buckets were used toquantify the in-fall of terrestrial prey and leaflitter into Fox Creek. The aerial flux of terrestrialprey into experimental units was quantifiedusing pan traps deployed five times at fivelocations across the longitudinal extent of theexperiment (1.3 km). Paired traps (37 cm 3 26.5cm) were set inside and outside of the experi-mental enclosures above the stream surface witha few centimeters of water and 2–3 drops ofsurfactant to capture any falling invertebrates.The percentage of leaf litter and other organicmatter excluded from covered treatments wasquantified using buckets (23 cm dia.) deployedover an 11 day period near the end of theexperiment at five locations with six buckets ateach location, three inside and three outsidecovered experimental units. Leaf litter sampleswere dried for 48 hours at 608C to obtain dryweight, then placed in a muffle furnace at 5508Cfor 4 hours, and measured immediately after toestimate ash free dry mass and carbon content.

Responses by the aquatic invertebrate commu-nity to experimental contrasts were estimatedusing sticky traps which indexed aquatic insectemergence, and by sampling benthic aquaticinvertebrates from rocks within each experimen-tal pool. To sample the emergence of aquaticinsects from experimental pools, three stickytraps were deployed within each experimentalunit three times at approximately two-weekintervals throughout the duration of the experi-ment (9 weeks). Traps were constructed of 21.63

27.9 cm clear overhead transparencies whichwere sprayed on both sides with the agriculturaladhesive Tangle-Trap (Contech; Victoria, BC,Canada) and were deployed perpendicular tothe direction of flow and left for 48 hours. Trapscollected at uncovered pools likely captured

Table 1. Factors included in three-way factorial experiment including the presence/absence of 1þ steelhead

predators, salamanders (.100) and terrestrial prey in-fall. Food web arrangement within 32 individual pool

habitats was manipulated with a combination of levels across each of the three factors resulting in 4 replicates

of 8 unique experimental treatments.

Level Factor Background density or rate Treatment density or rate

þ/� 1þ steelhead (.85 mm) 0.26 fish/m2 0 fish/m2

þ/� Salamanders (.100 mm TL) 0.52 salamanders/m2 0 salamanders/m2�þ/� Terrestrial prey subsidy 523.1 mg�m�2�d�1 18.86 mg�m�2�d�1

� Tagged salamanders exhibited high degree of movement during the course of the experiment, such that final salamanderdensity was unrelated to the initial treatment Experimental contrasts in salamander presence were not included in the analysis.


ATLAS ET AL.

insects from adjacent areas and were thereforedropped from the analysis of emergence re-sponse. Within covered pools sticky traps wereassumed to have equal capture efficiency, allow-ing us to compare the effects of our predatortreatments on aquatic insect emergence. Thebiomass and composition of benthic aquaticinvertebrates in our experimental pools wereestimated by sampling 6 randomly selectedbenthic rocks from each unit at the end of theexperiment as in McNeely et al. (2007). Each rockwas moved quickly from the stream bottom intoa net positioned downstream and placed in a trayto minimize the escape of more mobile aquaticinsects. The entirety of each rock was thencleaned with a garden sprayer and visuallyinspected to ensure the capture of all aquaticinvertebrates. Invertebrate biomass was thenstandardized by the surface area of the 6 rockssampled. Invertebrates captured in pan traps orsampled from benthic rocks were stored in 70%ethanol, identified to family and genus whenpossible, measured to the nearest 0.1 mm, andconverted to dry mass estimates using taxonspecific length-weight relationships (Hodar 1996,Benke et al. 1999, Sabo et al. 2002). Sticky trapswere frozen for later identification to order, andbiomass was estimated as above. Average per-cent canopy cover was estimated for each poolusing a spherical densitometer as a proxy forlight availability. Key response variables foraquatic invertebrates included the biomass ofbenthic invertebrates, partitioned into groupsbased on functional feeding groups, as well asthe biomass of emerging aquatic insects asindexed by sticky traps.

For the purpose of our analysis all members ofthe scraper functional group were consideredherbivores. This assumption was based onprevious work demonstrating that members ofthis functional group disproportionately rely onin situ production in our study system (McNeelyet al. 2007). Invertebrates were further parti-tioned into groups based on their relativevulnerability to predation, with armored inver-tebrates such as cased caddisflies classified asinvulnerable, and soft-bodied invertebrates suchas mayflies classified as vulnerable.

To test for experimentally induced effects onprimary producer biomass, we incubated 12unglazed ceramic tiles (4.8 cm 3 4.8 cm) in each

pool over the duration of the experiment. Fourtiles were destructively sampled during thecourse of the experiment at weeks 5, 7, and 9 totest for changes in algae and biofilm standingstock using ash-free dry mass (AFDM) andchlorophyll-a concentration. Algae and biofilmswere sampled by scrubbing each tile with atoothbrush and filtered stream water. The result-ing slurry was sub-sampled for further analysis,with 20 ml filtered through pre-combusted 0.7lm glass fiber filters (Whatman GF-F, 47 mm),and then ashed (5508C for 24 hrs) to estimateAFDM. A 4 ml sub-sample was filtered onto 0.7lm glass fiber filters (Whatman GF-C, 25 mm)and frozen for later estimation of chlorophyll-ausing ethanol extraction and fluorometry (Stein-man and Lamberti 1996). To evaluate whetherthe observed responses to our covered treatmentswere due to changes in temperature we placedtemperature loggers (ibutton, MAXIM: Sunny-vale, CA, USA) in 12 experimental units; 6covered and 6 in uncovered pools along thelongitudinal extent of the experiment. Tempera-tures were then compared by one-way ANOVAto test for an effect of experimental covers.

Statistical analysesWe evaluated the response of predators,

herbivores, and primary producers to our exper-imental treatments by fitting a range of compet-ing linear and linear mixed effects models in thestatistics program R version 2.13.2 (R Develop-ment Core Team 2011). We used an informationtheoretic approach based on Akaike’s informa-tion criterion adjusted for small sample sizes(AICc) (R-package AICcmodavg) (Mazerolle2011) to examine the response of the aquaticfood web to experimental treatments and back-ground environmental variability such as lightavailability. For each response variable weconsidered all possible combinations of our maineffects (presence or absence of predatory steel-head, predatory salamander density, presence orabsence of terrestrial prey) and their interactions,as well as percent canopy cover as a proxy forlight availability. For benthic invertebrate re-sponses we modeled total biomass as well asthe biomass of two categories of herbivores(vulnerable, armored) with invertebrate taxapartitioned into these groups based on thepresence of physical armoring such as rock or


ATLAS ET AL.

stick cases. Community response metrics whichwere sampled on a single date such as benthicinvertebrates, as well as predator and YOYgrowth were modeled using linear regression inR. Response metrics with repeated samplingsthroughout the summer such as chlorophyll-a,AFDM, and aquatic insect emergence were fitusing linear mixed effects models (R packagenlme) (Pinhero et al. 2012) estimated usingmaximum likelihood, including sample date asa random effect to account for repeated mea-surements. Data for the biomass of emergingaquatic invertebrates did not meet the assump-tions of normality and were therefore logtransformed. Models were then ranked basedon their relative likelihood (xi ). To avoid overfitting we removed ‘‘pretending variables’’, var-iables which receive support in the AICc frame-work but do not change model deviance. Wereport only model’s within 2 DAICc units(Burnham and Anderson 2002) of our top model,and we report weighted coefficients based on allmodels within this threshold (Appendix A).

There was a high degree of immigration andemigration among predatory salamanders dur-ing the experiment, and final salamander densi-ties did not reflect initial treatment contrasts.Consequently, we considered models that includ-ed either initial salamander treatment (presenceor absence) or salamander density at the conclu-sion of the experiment in our candidate modelset. Experimental pools without covers providedlittle inference about the emergence of aquaticinsects since capture efficiency on sticky trapswas lower than in units with covers, and becauseof the potential for capturing invertebratesemerging from areas outside of the experiment.As a result, we excluded uncovered pools fromthe analysis of aquatic insect emergence, andreport only the results for covered pools.

RESULTS

Terrestrial subsidyCovered units experienced a dramatic reduc-

tion in the flux of terrestrial material during theexperiment. The mean daily flux of terrestrialprey to uncovered units was 261.5 6 50.6 mgDM�m�2�d�1, and covered pools experienced a27-fold reduction in this prey subsidy (9.43 6

2.53 mg DM�m�2�d�1). Similarly, terrestrial leaf

litter subsidies were reduced more than 90 fold,from 800.5 6 121 mg DM�m�2�d�1 in open poolsto 8.7 6 5.05 mg DM�m�2�d�1 in covered. Streamtemperature was not affected by the presence ofcovers (covered, 14.208C 6 0.178C [mean 6 SE];uncovered, 14.288C 6 0.158C; p ¼ 0.755).

Steelhead and salamandersOn average 1þ and YOY steelhead grew

approximately 1.23 g and 1.12 g, respectively,during the duration of the experiment. Growth ofjuvenile steelhead from both age classes (1þ,YOY) was best predicted by the availability ofterrestrial prey; however models of YOY growthwhich included light availability (% canopycover) also received some support with YOYgrowth declining across a gradient of increasingcanopy closure (�2.31 6 1.96 g). In the absence ofterrestrial prey, most individual fish (1þ steel-head and YOY) experienced negative growthover the course of the experiment (Fig. 2A, C,Table 2; coefficients:�2.30 6 0.77 g and�0.51 6

0.18 g, respectively). There was no indication thatthe growth of 1þ steelhead was altered by thepresence of salamander predators and neitherpredator species limited the growth of YOYsteelhead. The growth of predatory salamanderswas not reduced in the absence of terrestrialsubsidies (Fig. 2B), and was instead best predict-ed by a model that included only the presence of1þ steelhead. The presence of steelhead facilitat-ed higher growth in salamanders over the courseof the experiment (1.86 6 0.67 g; Fig. 2B), thoughthere was limited support for an intercept onlymodel as well (Table 2).

Food web responsesNeither predator depressed the total biomass

of aquatic herbivores, and AICc indicated a lackof support for top-down control of herbivorebiomass regardless of the availability of subsidies(Fig. 2D, Table 2). The majority (65% of biomass)of aquatic invertebrates sampled from benthicrocks in Fox Creek were comprised of relativelypredator-invulnerable taxa such as armored andcase building caddisflies (Order Trichoptera),which are rarely found in the diets of steelheadand salamanders (Parker 1994, Wootton et al.1996). When the biomass of vulnerable andinvulnerable (armored) taxa were analyzed sep-arately there was support for top-down control


ATLAS ET AL.

of vulnerable herbivore biomass in experimentalunits with steelhead (�11.3 6 5.4 mg DM�m�2,Fig. 2E). However total aquatic invertebratebiomass increased in the presence of predatorysteelhead (369.4 6 323.9 mg DM�m�2) anddecreased along a gradient from 0 to 100% incanopy cover (�340.5 6 193.1 mg DM�m�2).There was also support for a positive interactionbetween steelhead predators and light availabil-ity (% canopy cover) such that the slope of thepositive relationship between total benthic bio-mass and light availability was steeper in the

presence of steelhead (�603.9 6 193.1 mgDM�m�2; Fig. 3).The top model for aquatic insectemergence included an interaction between %canopy cover and steelhead as well as the maineffect of salamander density, however a modelwhich included an interaction between canopyand salamander density also fell within the 2DAICc unit threshold (Table 2). The biomass ofaquatic insects emerging from covered experi-mental units showed a similar pattern to totalbenthic invertebrate biomass. The biomass ofinsect emergence was higher in the presence of

Fig. 2. Food web response to experimental manipulation of predator assemblage and availability of terrestrial

prey. Terrestrial prey availability treatment is depicted on the x-axis of each plot, grey areas indicate experimental

treatments with steelhead predators. Dark horizontal bars represent mean values for each treatment, and curves

represent the distribution of response data. (A) Response of 1þ steelhead growth to reductions in terrestrial prey

availability. (B) Growth of recaptured salamanders (.100 mm), (C) growth of young of year steelhead, (D) final

herbivore biomass, (E) final biomass of vulnerable herbivores, (F) primary producer biomass indexed by AFDM

and (G) chlorophyll-a response to experimental changes in terrestrial prey availability and 1þ steelhead presence.


ATLAS ET AL.

steelhead (16.45 6 7.07 mg DM�m�2) and therewas a positive interaction between steelhead andlight availability (�19.58 6 8.2 mg DM�m�2) suchthat the slope of the relationship between lightavailability and emergence biomass steepened inthe presence of steelhead. Because of the highdegree of movement in and out of experimentalenclosures by tagged salamanders, the trueresponse of the stream food web to the presenceof salamanders was difficult to discern andestimated model coefficients were suspect.

The amount of algae and biofilm were stronglyinfluenced by the availability of light in our studysystem. The top model for ash free dry mass(AFDM) included only a single factor, % canopycover, with biofilm AFDM declining across anincreasing gradient in canopy cover from 0 to100% (�3.78 6 0.91 g AFDM�m�2). However, amodel which included an interaction between %canopy cover and steelhead received somesupport, with AFDM increasing in the presenceof steelhead (�1.653 6 1.93 g AFDM�m�2; Fig. 2F,

Table 2. AICc model selection of linear regression models for response variables including R2 values for each

model. Variables include % canopy cover, steelhead predators (presence/absence), final salamander density,

and the availability of terrestrial prey.

Response variable Model DAICc xi R2

Predator growth1þ steelhead growth Terrestrial prey (þ/�) 0 0.917 0.346

Intercept 4.8 0.083Salamander growth Steelhead (þ/�) 0 0.815 0.406

Intercept 2.96 0.185YOY growth Terrestrial prey (þ/�) 0 0.569 0.229

% Canopy cover þ terrestrial prey (þ/�) 0.7 0.413 0.259Intercept 5.27 0.037

Aquatic invertebratesTotal invertebrate biomass Steelhead (þ/�) 3 % Canopy cover 0 0.554 0.305

% Canopy 0.811 0.369 0.196Intercept 5.75 0.031

Total herbivore biomass Intercept 0 1Vulnerable herbivore biomass Steelhead (þ/�) 0 0.721 0.105

Intercept 1.9 0.279Total emergence biomass % Canopy cover 3 steelhead (þ/�) þ salamander

density0 0.680 0.292

% Canopy cover 3 salamander density þ % canopycover 3 steelhead (þ/�)

1.718 0.298 0.311

Intercept 6.112 0.032Primary producers

Ash free dry mass (AFDM) % Canopy cover 0 0.686 0.153% Canopy cover 3 steelhead (þ/�) 1.564 0.314 0.180Intercept 13. 9 0.001

Chlorophyll-a % Canopy cover þ terrestrial prey (þ/�) 0 0.530 0.235% Canopy cover þ steelhead (þ/�) þ terrestrial prey(þ/�)

1.494 0.251 0.243

% Canopy cover 1.765 0.219 0.205Intercept 16.476 0.000

Fig. 3. Effect of % canopy cover in mediating

response of total benthic biomass to manipulations in

the presence of predatory steelhead. Solid line and

circles indicate experimental units with steelhead.


ATLAS ET AL.

Table 2). Like AFDM, the concentration ofchlorophyll-a was positively related to lightavailability (0–100% canopy: 17 6 3.99 lg�m�2),which was included in every model receivingsupport (Table 2, Fig. 2G).

DISCUSSION

We found clear evidence that terrestrial preysubsidies support the growth of predatorysteelhead in tributary stream food webs duringthe course of our 9-week experiment. Through-out the summer, fish in pools with experimentalcovers experienced dramatic reductions ingrowth, with many losing mass during the 2-month experiment. While this finding suggestshigh potential for apparent competition betweenterrestrial prey and in situ prey species, the top-down effect of steelhead on the biomass ofbenthic invertebrates was limited to a smallsubset of the aquatic community (vulnerableherbivores), and was not strengthened in theabsence of terrestrial prey. This weak top-downcontrol is likely due to the abundance ofmorphologically predator-resistant invertebratesin this tributary stream food web (McNeely et al.2007), with approximately 65% of the totalbenthic invertebrate biomass and 82% of herbi-vore biomass classified as armored and invul-nerable to predation. Though predatory fisheshave also been shown to induce trophic cascadeson detrital processing rates (Konishi et al. 2001,Boyero et al. 2007,, but see Ruetz et al. 2002), wedid not estimate changes in the processing ofterrestrial detritus as part of this study. We expectthat the processing rates by aquatic invertebratesmay be similarly insensitive to top-down controlin our study system due to the high abundance ofarmored taxa within shredder and scraperfunctional feeding groups. Given the relativeinvulnerability of most aquatic prey in our studysystem, it appears that rather than subsidizingfish predation on local invertebrates, terrestrialprey may serve as the primary source of energyin supporting predatory fish in the Fox Creekfood web. Consequently, when terrestrial subsi-dies were interrupted, there was little scope forintensified predation on in situ prey as observedin other studies (Nakano et al. 1999b, Baxter et al.2004).

Previous work in stream food webs has found

that the degree to which predatory fish exert top-down control depends in part on the vulnerabil-ity of the aquatic invertebrate community (Power1992). In the South Fork Eel River, inter-annualvariability in the strength of trophic cascades hasbeen linked to flood-pulse events that scour thestream bottom and remove large armored cad-disflies (Genus Dicosmoecus). These large winterfloods leave behind an aquatic invertebratecommunity comprised of more predator vulner-able taxa resulting in dramatic changes in year-to-year patterns of top-down control by preda-tory steelhead trout (Wootton et al. 1996, Poweret al. 2008). In tributary streams in northernCalifornia, the smaller, multivoltine Glossosomapentium are the dominant armored caddisfly, andthis taxon maintains high standing biomass formuch of the year despite the low primaryproductivity of these streams. Persistent highdensities of Glossosoma in tributaries such as FoxCreek result in strong herbivore control of algalbiomass and limit the availability of in situproduction to other aquatic consumers that aremore readily available to predatory fish andsalamanders (McNeely et al. 2007). The food webconsequences of reduced prey vulnerability arewell documented, and prey species that escapepredation by attaining large body size may beresource limited despite high predator biomass(Chase 1999, Sinclair et al. 2003). Similarly,species which employ heavy armored casesmay be largely invulnerable to predation andmay serve as a trophic cul-de-sac (Bishop et al.2007), limiting the biomass of other vulnerablegrazers via competition, reducing the trophictransfer of algal biomass to higher trophic levels,and diminishing the ability of top-down preda-tion to propagate through the food web.

Despite the prevalence of armored aquaticprey in Fox Creek and the lack of strongexperimentally induced trophic cascades, wedid identify several notable effects of predatorysteelhead on the food web. When we partitionedherbivores into armored versus more vulnerablegroups and analyzed food web responses sepa-rately, we found strong evidence for top-downcontrol of the biomass of vulnerable herbivoresover the course of the experiment (Table 1, Fig.2E). Light availability (% canopy cover) was themost important factor influencing indices ofprimary production (AFDM, chl-a), models that


ATLAS ET AL.

included steelhead as well as light availabilityalso received support (Table 2), with steelheadhaving a positive effect on AFDM and appearingto reduce the concentration of chl-a. Despite thelack of top-down control by steelhead on thetotal biomass of herbivores, this interaction(predator 3 light) suggests a positive effect ofsteelhead on primary producer biomass such thatthe positive effect of increasing light availabilitywas greater in the presence of steelhead, possiblydue to nutrient recycling by steelhead (Vanni etal. 2006, Munshaw et al. 2013) (Fig. 3). Similarly,there was support for a positive effect ofpredatory steelhead on the total biomass ofbenthic invertebrates as well as the emergenceof aquatic insects from experimental pools, aswell as an interaction between steelhead andlight availability. The conformity of the responseof primary producers and benthic invertebratesto the presence of predatory fish and lightavailability suggest that fish and other predatorsmay mediate food web productivity in tributarystreams from the base of the food web vianutrient recycling, with fish excretion ultimatelyinfluencing the magnitude of reciprocal subsidiesinto the riparian forest.

As predicted, a two month long reduction inthe influx of terrestrial prey subsidies did notappear to negatively affect the growth of Pacificgiant salamanders. This finding is consistent withprevious observations that salamander diets aredominated by benthic aquatic invertebrates(Parker 1994). Further, given the high biomassof predatory and largely benthic feeding Pacificgiant salamanders that we found in Fox Creek,occurring at twice the density of 1þ steelhead(0.52 salamanders/m2 vs. 0.26 steelhead/m2), wepredicted that benthic invertebrate biomasswould be depressed in their presence. However,low recapture rates and movement by individu-als outside of our study reaches made it difficultto detect the food web effects of salamanders. Wealso predicted that YOY steelhead would havereduced growth in the presence of large 1þsteelhead and salamanders, but did not detectany effect of either predator on their growth.Despite considerable diet overlap between sala-manders and 1þ steelhead (Parker 1994, Mun-shaw et al. 2013), the presence of salamandershad no detectable effect on the growth of 1þsteelhead, regardless of the availability of terres-

trial prey. Limited isotope samples collected inconjunction with the experiment suggest thatsteelhead relied on terrestrial prey to a muchgreater degree during the summer than large(.100 mm SVL) salamanders (Appendix B).Contrary to the prediction that competitionbetween predator species would increase in theabsence of terrestrial subsidies, the presence ofpredatory steelhead led to increased growthamong recaptured salamanders regardless ofthe availability of terrestrial prey. While thisconclusion is tempered by the low sample size ofsalamanders recaptured at the end of ourexperiment (n ¼ 19), we propose that thepresence of 1þ steelhead allowed salamandersto more efficiently exploit aquatic prey. Steelheadare mobile predators that feed in the watercolumn, and are known to elicit changes in thebehavior of their prey (Douglas et al. 1994, Postet al. 1998). If behavioral changes by prey(including young of year steelhead) in responseto the threat of steelhead predation increasedtheir vulnerability to predation by more seden-tary, benthic salamanders, it may explain in-creased salamander growth in the presence ofpredatory steelhead (Sih et al. 1998).

The nine week duration of our experimentalmanipulation during the peak window of sum-mer productivity limits our inference about theresponse of the Fox Creek food web to subsi-dized predators over longer time scales. Giventhe short duration of our experiment, we wereunlikely to capture the full range of possibleinvertebrate population responses to our im-posed food web manipulations. However, ourexperimental treatments do provide importantinference about how short-term changes in theintensity of predation can affect local aquaticinvertebrate biomass, invertebrate movement,and growth. The invertebrate community presentwithin our experimental stream reaches certainlyreflected the legacy of interactions with bothpredators, and the pre-existing herbivore assem-blage may have limited the potential for exper-imentally induced changes in the strength oftrophic cascades in our study over the two monthtime period we ran the experiment. Despite thislimitation, previous work conducted over asimilar duration has resulted in strong commu-nity level responses. For example, when McNeelyet al. (2007) experimentally removed the caddis-


ATLAS ET AL.

fly Glossosoma, the dominant armored grazer inFox Creek, chlorophyll-a in Fox Creek doubled,suggesting that short-term manipulations caninduce changes in grazer control and primaryproductivity. The importance of terrestrial re-source subsidies for predator populations likelyalso manifests over longer time scales. Forexample, subsidies may increase the carryingcapacity of the recipient ecosystem for predatorsincreasing population productivity and carryingcapacity for salamanders and trout. Subsidies ofleaf litter and other organic material also serve toincrease food web productivity from the base ofthe food web, however given the relatively longturn over time of this organic material in thestream food web we cannot draw inference aboutthe long-term consequences of reduced detritalsubsidies. Our experiment was not designed toexplicitly address long-term population levelresponses of predators to resource subsidies,however dramatic reductions in growth experi-enced by 1þ steelhead in the absence of terrestrialprey suggest that current population sizes aresupported in large part by terrestrial preysubsidies. Due to the nature of our experimentalenclosures, fish were not allowed to move inresponse to manipulations in resource availabil-ity. However, the reductions in fish growthobserved in our study coupled with previousexperimental work demonstrating high rates ofemigration by trout in the absence of terrestrialprey (Kawaguchi et al. 2003) suggest that ingeneral terrestrial prey subsidies support preda-tory stream fish biomass in tributary stream foodwebs.

Our results support the prediction that thepresence of armored herbivores in Fox Creekserve to compartmentalize tributary food websinto two parts. One, which consists of a closedloop between algal primary production andarmored herbivores, and another in whichaquatic consumers are tied to the productivityof the surrounding terrestrial environment, bothfrom the bottom of the food web via terrestrialdetritus and from the top by prey subsidies.While predators in recipient ecosystems un-doubtedly benefit from the influx of allochtho-nous prey, the degree to which subsidiespropagate through multiple trophic levels de-pends upon the ability of predators to exploit insitu prey. In our study system herbivores are

known to limit primary producer biomass(McNeely et al. 2007), yet the invulnerability ofmuch of the herbivore guild ultimately limits theimportance of predation. Even in the absence ofterrestrial prey, aquatic predators did not appearcapable of expanding their exploitation of localherbivore biomass. This finding is perhaps notsurprising given the co-evolution of the aquaticpredators and herbivore communities in thesesubsidized food webs. Because predators arelargely decoupled from in situ productivity inthese systems, the persistence of the aquaticherbivore guild through time has likely beendependent on morphological and behavioraltraits that limit their vulnerability to predation.Our results highlight the degree to which thetrophic consequences of prey subsidies areinfluenced by the composition of the in situ preycommunity, and raise interesting questions aboutthe degree to which morphological and behav-ioral traits conferring low vulnerability to preda-tion may be ubiquitous in highly subsidized foodwebs.

ACKNOWLEDGMENTS

We thank Peter Steel for facilitating access to theUniversity of California Natural Reserve System,Angelo Coast Range Reserve, Brooke Weigel, GraceWilkinson, Angela Rosendahl, Wil Torres and HiromiUno for field assistance, Jill Welter and Mike Limm foradvice on laboratory analyses and field protocols,Camille McNeely, Kathrine Stewart, Greg Lee, JodieMcCormick, Rylee Murray, and Janet Rickards forassistance processing macroinvertebrate samples, theEarth to Ocean research group for feedback onstatistical analyses. Mary Power, Jacques Finlay, andNick Dulvy provided helpful comments on an earlierdraft. Thanks to Selena Popovic for offering her designexpertise in the creation of the figures. This work wasconducted under California Department of Fish andGame (#11077), NOAA (#14904), and Simon FraserUniversity Animal Care (920B-09) permits, and wassupported by the National Science and EngineeringResearch Council of Canada, and the Canada ResearchChairs program.

LITERATURE CITED

Baxter, C. V., K. D. Fausch, M. Murakami, and P. L.Chapman. 2004. Fish invasion restructures streamand forest food webs by interrupting reciprocalprey subsidies. Ecology 85:2656–2663.

Benke, A. C., A. D. Huryn, L. A. Smock, and J. B.


ATLAS ET AL.

Wallace. 1999. Length-mass relationships for fresh-water macroinvertebrates in North America withparticular reference to the southeastern UnitedStates. Journal of the North American Bentholog-ical Society 18:308–343.

Bishop, M. J., B. P. Kelaher, R. Alquezar, P. H. York, P. J.Ralph, and C. G. Skilbeck. 2007. Trophic cul de-sac,Pyrazus ebeninus, limit trophic transfer through anestuarine detritus-based food web. Oikos 116:427–438.

Boyero, L., P. A. Rincon, and R. G. Pearson. 2008.Effects of a predatory fish on a tropical detritus-based food web. Ecological Research 23:649–655.

Burnham, K. P., and D. R. Anderson. 2002. Modelselection and multimodel inference: A practicalinformation-theoretic approach. Second edition.Springer, New York, New York, USA.

Chase, J. M. 1999. Food web effects of prey size refugia:variable interactions and alternative stable equilib-ria. American Naturalist 154:559–570.

Douglas, P. L., G. E. Forrester, and S. D. Cooper. 1994.Effects of trout on the diel periodicity of drifting inbaetid mayflies. Oecologıa 98:48–56.

Finlay, J. C. 2001. Stable-carbon-isotope ratios of riverbiota: implications for energy flow in lotic foodwebs. Ecology 82:1052–1064.

Finlay, J. C., and V. T. Vredenburg. 2007. Introducedtrout sever trophic connections in watersheds:consequences for a declining amphibian. Ecology88:2187–2198.

Henschel, J. R., D. Mahsberg, and H. Stumpf. 2001.Allochthonous aquatic insects increase predationand decrease herbivory in river shore food webs.Oikos 93:429–438.

Hilderbrand, G. V., C. C. Schwartz, C. T. Robbins, M. E.Jacoby, T. A. Hanley, S. M. Arthur, and C. Servheen.1999. The importance of meat, particularly salmon,to body size, population productivity, and conser-vation of North American brown bears. CanadianJournal of Zoology 77:132–138.

Hocking, M. D., and J. D. Reynolds. 2011. Impacts ofsalmon on riparian plant diversity. Science331:1609–1614.

Hodar, J. A. 1996. The use of regression equations forestimation of arthropod biomass in ecologicalstudies. Acta Oecologia 17:421–433.

Holt, R. D. 1984. Spatial heterogeneity, indirectinteractions, and the coexistence of prey species.American Naturalist 124:377–406.

Kawaguchi, Y., and S. Nakano. 2001. Contribution ofterrestrial invertebrates to the annual resourcebudget for salmonids in forest and grasslandreaches of a headwater stream. Freshwater Biology46:303–316.

Kawaguchi, Y., Y. Taniguchi, and S. Nakano. 2003.Terrestrial invertebrate inputs determine the localabundance of stream fishes in a forested stream.

Ecology 84:701–708.Konishi, M., S. Nakano, and T. Iwata. 2001. Trophic

cascading effects of predatory fish on leaf litterprocessing in a Japanese stream. Ecological Re-search 16:415–422.

Leroux, S. J., and M. Loreau. 2008. Subsidy hypothesisand strength of trophic cascades across ecosystems.Ecology Letters 11:1147–1156.

Marcarelli, A. M., C. V. Baxter, M. M. Mineau, andR. O. Hall. 2011. Quantity and quality: unifyingfood web and ecosystem perspectives on the role ofresource subsidies in freshwaters. Ecology 92:1215–1225.

Marczak, L. B., R. M. Thompson, and J. S. Richardson.2007. Meta-analysis: trophic level, habitat, andproductivity shape the food web effects of resourcesubsidies. Ecology 88:140–148.

Mazerolle, M. 2011. AICcmodavg: Model selection andmulti-model inference based on (Q) AIC(c). Rpackage version 1.19.

McNeely, C., J. C. Finlay, and M. E. Power. 2007.Grazer traits, competition, and carbon sources to aheadwater-stream food web. Ecology 88:391–401.

Miller, T. W. 2006. Tissue-specific response of d15N inadult Pacific herring (Clupea pallasi ) following anisotopic shift in diet. Environmental Biology ofFishes 76:177–189.

Munshaw, R. G., W. J. Palen, J. C. Finlay, and D. M.Courcelles. 2013. Predator-driven nutrient recy-cling in a California stream ecosystem. PLOSONE in press.

Murakami, M., and S. Nakano. 2002. Indirect effect ofaquatic insect emergence on a terrestrial insectpopulation through predation by birds. EcologyLetters 5:333–337.

Nakano, S., Y. Kawaguchi, Y. Taniguchi, H. Miyasaka,Y. Shibata, H. Urabe, and N. Kuhara. 1999a.Selective foraging on terrestrial invertebrates byrainbow trout in a forested headwater stream innorthern Japan. Ecological Research 14:351–360.

Nakano, S., H. Miyasaka, and N. Kuhara. 1999b.Terrestrial-aquatic linkages: riparian arthropodinputs alter trophic cascades in a stream foodweb. Ecology 80:2435–2441.

Parker, M. S. 1993. Predation by Pacific giant salaman-der larvae on juvenile steelhead trout. Northwest-ern Naturalist 74:77–81.

Parker, M. S. 1994. Feeding ecology of stream-dwellingpacific giant salamander larvae (Dicamptodon tene-brosus). Copeia 1994:705–718.

Pinhero, J., D. Bates, S. DebRoy, and D. Sarkar. and theR Development Core Team. 2012. nlme: Linear andnon-linear mixed effects models. R package version3.1-104.

Polis, G. A. and S. D. Hurd. 1996. Allochthonousinputs across habitats, subsidized consumers, andapparent trophic cascades: examples from the


ATLAS ET AL.

ocean-land interface. Pages 275–285 in G. A. Polisand K. O. Winemiller, editors. Food webs: integra-tion of patterns and dynamics. Chapman & Hall,London, UK.

Polis, G. A., W. B. Anderson, and R. D. Holt. 1997.Toward an integration of landscape and food webecology: the dynamics of spatially subsidized foodwebs. Annual Review of Ecology and Systematics28:289–316.

Post, J. R., E. A. Parkinson, and N. T. Johnston. 1998.Spatial and temporal variation in risk to piscivoryof age-0 rainbow trout: patterns and populationlevel consequences. Transactions of the AmericanFisheries Society 127:932–942.

Power, M. E., W. J. Matthews, and A. J. Stewart. 1985.Grazing minnows, piscivorous bass, and streamalgae: dynamics of a strong interaction. Ecology66:1448–1456.

Power, M. E. 1990. Effects of fish in river food webs.Science 250:811–814.

Power, M. E. 1992. Habitat heterogeneity and thefunctional significance of fish in river food webs.Ecology 73:1675–1688.

Power, M. E., M. S. Parker, and W. E. Dietrich. 2008.Seasonal reassembly of a river food web: floods,droughts, and impacts of fish. Ecological Mono-graphs 78:263–282.

R Development Core Team. 2011. R: A language andenvironment for statistical computing. R Founda-tion for Statistical Computing, Vienna, Austria.

Richardson, J. S., Y. Zhang, and L. B. Marczak. 2010.Resource subsidies across the land-freshwaterinterface and responses in recipient communities.River Research and Applications 26:55–66.

Rose, M. D., and G. A. Polis. 1998. The distribution andabundance of coyotes: the effects of allochthonousfood subsidies from the sea. Ecology 79:998–1007.

Ruetz, C. R., R. M. Newman, and B. Vondracek. 2002.Top-down control in a detritus-based food web:fish, shredders, and leaf breakdown. Oecologia132:307–315.

Sabo, J. L., J. L. Bastow, and M. E. Power. 2002. Length-mass relationships for adult aquatic and terrestrialinvertebrates in a California Watershed. Journal of

the North American Benthological Society 21:336–343.

Sabo, J. L., and M. E. Power. 2002. River-watershedexchange: effects of riverine subsidies on riparianlizards and their terrestrial prey. Ecology 83:1860–1869.

Sih, A., G. Englund, and D. Wooster. 1998. Emergenceimpacts of multiple predators on prey. Trends inEcology and Evolution 13:350–355.

Sinclair, A. R. E., S. Mduma, and J. S. Brashares. 2003.Patterns of predation in a diverse predator-preysystem. Nature 425:288–290.

Spiller, D. A., J. Piovia-Scott, A. N. Wright, L. H. Yang,G. Takimoto, T. W. Schoener, and T. Iwata. 2010.Marine subsidies have multiple effects on coastalfood webs. Ecology 91:1424–1434.

Steinman, A. D., and G. A. Lamberti. 1996. Biomassand pigments of benthic algae. Pages 295–313 inR. A. Hauer and G. A. Lambert, editors. Methodsin stream ecology. Academic Press, San Diego,California, USA.

Vanni, M. J., A. M. Bowling, E. M. Dickman, R. S. Hale,K. S. Higgins, M. J. Horgan, L. B. Knoll, W. H.Renwick, and R. A. Stein. 2006. Nutrient cycling byfish supports relatively more primary productionas lake productivity increases. Ecology 87:1696–1709.

Vannote, R. L., G. W. Minshall, K. W. Cummins, J. R.Sedell, and C. E. Cushing. 1980. The river contin-uum concept. Canadian Journal of Fisheries andAquatic Sciences 37:130–137.

Wallace, J. B., S. L. Eggert, J. L. Meyer, and J. R.Webster. 1997. Multiple trophic levels of a foreststream linked to terrestrial litter inputs. Science277:102–104.

Wipfli, M. S., and D. P. Gregovich. 2002. Export ofinvertebrates and detritus from fishless headwaterstreams in southeastern Alaska: implications fordownstream salmonid production. Freshwater Bi-ology 47:957–969.

Wootton, J. T., M. S. Parker, and M. E. Power. 1996.Effects of disturbance on river food webs. Science273:1558–1561.


ATLAS ET AL.

SUPPLEMENTAL MATERIAL

APPENDIX A

Table A2. Model averaged coefficient estimates for the response of aquatic invertebrates to experimental

manipulations for all models within 2 DAICc threshold.

Response variable Coefficient SE Rel. Import.

Total benthic invertebrate biomass (mg/m2)Intercept 481.6 168% Canopy cover �340.5 193.1 1Steelhead (þ/�) 369.4 323.9 0.71% Canopy cover 3 Steelhead (þ/�) �603.9 303.8 0.47

Vulnerable herbivores (mg/m2)Intercept 25.6 3.8Steelhead (þ/�) �11.3 5.4 NA

Emergence biomass (mg/m2)Intercept �6.96 7.7% Canopy cover 11.25 9.3 1Salamander density �9.5 13.6 1Steelhead (þ/�) 16.5 7.1 1% Canopy cover 3 Steelhead (þ/�) �19.6 8.2 1% Canopy cover 3 Salamander density 22.9 21.8 0.3

Table A3. Model averaged coefficient estimates for primary producer responses to experimental manipulations

for all models within 2 DAICc threshold.

Variable Coefficient SE Rel. Import.

AFDM (g/m2)Intercept 4.7 0.85% Canopy cover �3.7 0.98 1Steelhead (þ/�) 0.6 1.15 0.59% Canopy cover 3 Steelhead (þ/�) �1.65 1.9 0.19

Chlorophyll-a (lg/m2)Intercept �11324 4904% Canopy cover 21304 5666 1Terrestrial (þ/�) 1204 637 0.74Steelhead (þ/�) �611.5 637 0.25

Table A1. Model averaged coefficient estimates for the response of vertebrate growth to experimental

manipulations for all models within 2 DAICc threshold.

Variable Coefficient SE Rel. Import.

Salamander (.100 mm) growth (g)Intercept 0.85 0.57Steelhead (þ/�) 1.86 0.67 NA

1þ steelhead growth (g)Intercept 1.23 0.55Terrestrial subsidy (þ/�) 2.30 0.77 NA

YOY steelhead growth (g)Intercept 2.07 1.54% Canopy cover �2.31 1.96 0.47Terrestrial subsidy 0.34 1.24 1% Canopy cover 3 Terrestrial subsidy �2.19 4.1 0.1


ATLAS ET AL.

APPENDIX B

At the conclusion of the experiment, small non-lethal tissue samples were taken from a subset ofrecaptured 1þ steelhead and D. tenebrosus (.100mm TL) for comparisons of d13C and d15N acrosstreatments and between species. d13C values foralgae and terrestrial detritus have been shown tobe distinct and stable isotope analysis can allowfor an estimate of the contribution of terrestrialversus aquatically derived energy in animaltissues (Finlay 2001). Fin and salamander tailtissue was dried in the laboratory (,608C),ground, weighed to the nearest 0.001 mg, andenclosed in tin capsules. Isotope analysis wasconducted at the University of California Davis,Stable Isotope Facility using an isotope ratiomass spectrometer. Isotope ratios are expressedas d13C and d15N values, which represent thelevels of enrichment of the heavier isotoperelative to the standard (N2, Pee Dee Belemnite).Fin and tail tissues are thought to reflect theisotopic signature of the diet within a fewmonths of sampling (Miller 2006), roughly thesame duration as the experiment.

Fig. B1. Isotopic comparison of d13C and d15Nsignatures of salamanders and steelhead trout sampled

at the conclusion of the experiment in Fox Creek. Open

triangles represent steelhead trout, closed circles

represent Pacific giant salamanders. More negative

d13C values for steelhead are indicative of a greater

reliance upon terrestrially derived carbon.


ATLAS ET AL.

Documents

Dependence of stream predators on terrestrial prey fluxes ...angelo.berkeley.edu/wp-content/uploads/Atlas_2013_Ecosphere.pdf(Holt 1984, Polis et al. 1997, Henschel et al. 2001, Murakami