Taxonomic and functional assessment of mesopredatordiversity across an estuarine habitat mosaic

COLLIN GROSS,1,� CINDE DONOGHUE,2 CASEY PRUITT,2 ALAN C. TRIMBLE,1 AND JENNIFER L. RUESINK1

1Department of Biology, University of Washington, Seattle, Washington 98195 USA2Aquatics Division, Aquatic Assessment and Monitoring Team, Washington Department of Natural Resources,

Olympia, Washington 98504 USA

Citation: Gross, C., C. Donoghue, C. Pruitt, A. C. Trimble, and J. L. Ruesink. 2017. Taxonomic and functional assessmentof mesopredator diversity across an estuarine habitat mosaic. Ecosphere 8(4):e01792. 10.1002/ecs2.1792

Abstract. A long-standing rule in ecology is that structural complexity increases abundance and diver-sity of organisms, but this paradigm glosses over potential trait-specific benefits of habitat structure acrossdifferent regional species pools. We tested this idea using multiple response variables emphasizing taxo-nomic and functional diversity in seagrass-vegetated, edge, and unvegetated habitats across three estuariesin Washington State (USA). We also used these variables in tandem to evaluate functional redundancy as aproxy for ecosystem resistance and resilience. The estuaries spanned a twofold range in richness of meso-predatory fishes and decapods. Increases in per-sample abundance with habitat structure were confined tothree of seven functional groups, specifically those occupying the water column or directly associated withseagrass shoots. Consequently, seagrass reduced mesopredator diversity by reducing evenness. Habitatsdiffered in mesopredator assemblages despite their spatial proximity (~3 m), supported by multivariateanalyses performed at functional and taxonomic resolutions, but site differences were less apparentfunctionally than taxonomically. Functional redundancy did not differ by habitat, but increased with therichness of the regional species pool. Edge habitats were generally intermediate in community structure,per-sample abundance, and diversity between seagrass and unvegetated habitats. Structural complexityprovided a trait-specific enhancement of abundance, and this pattern applied across species pools. Becauseseagrass benefits species with certain traits, management focused on the low-intertidal estuarine habitatmosaic, rather than a particular habitat type, and on places where redundancy is already low, bestsupports mesopredator diversity and function.

Key words: functional diversity; mesopredators; seagrass; structured habitat; Zostera marina.

Received 10 January 2017; revised 14 March 2017; accepted 22 March 2017. Corresponding Editor: Hunter S. Lenihan.Copyright: © 2017 Gross et al. This is an open access article under the terms of the Creative Commons AttributionLicense, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.� E-mail: colligross@gmail.com

INTRODUCTION

An open question exists concerning the valueof using taxonomic (species-level) vs. functional(trait-level) information to describe responses ofspecies assemblages to environmental conditionsand their effects in a food web context (Thrushet al. 2006, Lefcheck and Duffy 2015). Functionalapproaches are arising in ecology as ways to findbroad patterns across idiosyncratic local assem-blages, to understand functional redundancy and

the risks associated with species losses acrossassemblages of different richnesses, and becausefunctional diversity may augment ecosystem func-tion (Hooper et al. 2005, Micheli and Halpern 2005,Mouillot et al. 2013a). A long-standing rule in ecol-ogy is that abundance and species richness increasewith habitat structural complexity (MacArthur andMacArthur 1961), which makes habitat structurean ideal gradient for comparing assemblagesusing taxonomic and functional lenses. Here, weexamine mesopredator community structure along

❖ www.esajournals.org 1 April 2017 ❖ Volume 8(4) ❖ Article e01792

gradients of seagrass, an ecologically relevantbiogenic structure widely acknowledged to under-lie fisheries production (Heck et al. 2003, Blandonand Ermgassen 2014) and to harbor strong foodweb interactions (Duffy et al. 2005, 2015a, b).

Species assemblages are typically character-ized in two basic ways to examine the influenceof habitat characteristics on functional attributesof an assemblage: indices of functional diversityor functional groups. Rao’s quadratic entropy (Q)is one of the most commonly used functionaldiversity indices and incorporates both trait rich-ness and evenness. A decrease in Q may indicatedominance by a few traits conferring specializa-tion to a particular habitat type, while anincrease in Q may indicate greater niche avail-ability, as in structured habitats (Botta-Dukat2005). While diversity indices like Rao’s are typi-cally based on suites of traits that characterizespecies, species diversity indices may also beapplied to functional groups to produce a relatedindex (Micheli and Halpern 2005). Functionalgroups are often derived a priori based on grossmorphology, feeding guild, reproductive mode,or habitat use (Nagelkerken and van der Velde2004, Elliott et al. 2007, Villeger et al. 2008), butmay also be derived a posteriori based onclusters in multivariate trait space (e.g., Aguilar-Medrano and Calderon-Aguilera 2016). Group-ing species into functional groups, especiallythose based a posteriori on traits of interest,allows researchers to examine how particularhabitat types may facilitate or exclude specificsets of traits (Verberk et al. 2013).

Through a multivariate lens, these functionalperspectives are expected to give a more consis-tent test of community structure across environ-mental gradients than taxonomic perspectives(Nagelkerken and van der Velde 2004, Villegeret al. 2008, Wong and Dowd 2015, Aguilar-Medrano and Calderon-Aguilera 2016). Historicalbiogeographic and evolutionary processes deter-mine the broad-scale distributions of speciesacross sites, often comprising functionally identi-cal but taxonomically distinct local assemblages,while the effects of habitat structure and speciesinteractions tailor the traits in the species poolwithin a given habitat to produce the final assem-blage (Pavoine and Bonsall 2011, Adler et al.2013, Mouillot et al. 2013a, Kraft et al. 2015).Thus, by looking at species’ functional traits

across geographically distinct areas, one may beable to specifically pick out the role of structurallycomplex habitats like the interiors of seagrassbeds in forming distinct communities.Functional redundancy, defined as “a charac-

teristic of species within an ecosystem where cer-tain species contribute in equivalent ways to anecosystem function such that one species maysubstitute for another” (Froese and Pauly 2015),has often been used as a measure of ecosystemresiliency and resistance to disturbance (Folkeet al. 2004, Tilman et al. 2006, Downing et al.2012); disturbances that remove particular spe-cies might have a smaller impact on an ecosys-tem function if the remaining species perform theecosystem function in the same capacity as theextirpated species. The relationship between spe-cies and functional diversity determines thedegree of functional overlap and redundancyand informs hypotheses about resilience andresistance to disturbance. At one extreme, a 1:1relationship between functional and speciesdiversity indicates that each species performs aunique function. At the other end, if functionaldiversity increases minimally with species diver-sity, species exhibit a high degree of overlap infunctional traits. These relationships are expectedto plateau as additional species have similarfunctional traits to those already present in thecommunity (Micheli and Halpern 2005). Mea-sures of ecosystem resiliency such as functionalredundancy are useful in the context of resourcemanagement, in that they can help distinguishpriority species, and guide species conservationplans and habitat protection policies (Micheliand Halpern 2005).The purpose of this research was to test the

paradigm that habitat structure fosters greaterdiversity and abundance by examining functionaland taxonomic diversity in tandem, and usingthe relationships between these values to drawconclusions about community resilience. Ourfocal organisms were estuarine mesopredators,which represent a key trophic link, feeding onmesograzers (gastropods and crustaceans) thatcan limit algal blooms, and consumed by toppredators of larger body size that are oftenexploited commercially (Heck et al. 2003, Hugheset al. 2014, Duffy et al. 2015a). We asked (1)“Do structurally complex habitats differentiallybenefit mesopredators with some functional

❖ www.esajournals.org 2 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

traits over others, and is this consistent acrossregional species pools?” and (2) “Are structurallycomplex and/or higher-richness sites or habitatsmore functionally redundant, predicting greaterresiliency?”

Over the past 130 yr, an estimated >50,000 km2

of seagrass meadow has been lost throughout theworld (Waycott et al. 2009). Natural resource agen-cies are attempting to rectify this trend throughconservation and restoration efforts. Understand-ing how structure and function of mesopredatorcommunities change along a gradient of structural

complexity is particularly important for managersto design restoration projects and prioritizeconservation sites that target specific species forprotection.

METHODS

Study sitesSampling for mesopredators was done at three

sites spanning 186 km in Washington State, USA(Fig. 1; Appendix S1: Table S1). These sites differin depth ranges and influence by freshwater, and

Fig. 1. Maps of sampling region and sampling sites. Points on the larger map represent sampling sites shownin the smaller maps at right; points on smaller maps represent sampling subsites, each containing unvegetated,edge, and eelgrass–vegetated habitats. (A) N fjord; (B) S fjord; (C) Coastal.

❖ www.esajournals.org 3 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

consequently in their complement of mesopreda-tor species. Willapa Bay (hereafter referred to asCoastal; 46.52° N, 123.99° W) is a coastal estuaryformed at a drowned river mouth, with an averagedepth of 3.2 m (Hickey and Banas 2003). Severalrivers fed by winter rains influence its salinity,which rarely exceeds 27 psu at the study site. PortGamble Bay (N fjord, 47.84° N, 122.58° W) andthe Skokomish River delta (S fjord, 47.36° N,123.13° W) are at the north and south, respectively,of Hood Canal, an estuarine fjord that is long(100 km), deep (54 m), and narrow, usually <2 kmwide (Williams 2001). These fjord sites occur attwo distinct positions along a strong gradient inenvironmental conditions, with salinity and dis-solved oxygen lower at the S fjord than at the Nfjord site, and temperature higher (Kawase andBang 2013, Keister and Tuttle 2013). All study siteshave a mesotidal regime, given a mean diurnaltidal range of 2.8–3.5 m, which exposes intertidaleelgrass beds (Zostera marina) on extreme lowtides. Replication occurred within each site, wherewe selected three subsites, each containingthree intertidal habitat types (eelgrass, edge, andunvegetated tideflat). These habitats represent agradient of structural complexity; eelgrass shootdensity was roughly half as dense on edges asinside patches (Appendix S1: Table S2, Fig. S1).Aboveground biomass of eelgrass was similaracross sites (63 g DW/m2; Appendix S1: Table S2,Fig. S2).

Sampling of fishes and decapodsSamples were collected in July and August

2015 in a custom beach seine when the water levelon the target habitat patches was below 1 m deep.The seine sampled a circular area of 11 m2

(Appendix S1: Fig. S3). Animals were counted,identified to the lowest possible taxonomic level(typically species), and released. Of these, 10 indi-viduals per species were measured to the nearest0.5 cm (total length for fish and shrimp, carapacewidth for crabs) to be incorporated into functionaltrait analyses. Two seines were carried out in eachhabitat patch.

Functional traits and functional groupsEach mesopredator was evaluated for a suite

of nine functional traits based on published liter-ature including FishBase (Froese and Pauly 2000,2015). These traits included Estuarine Use,Reproductive Mode, and Feeding Mode Func-tional Groups (EUFG, RMFG, and FMFG, Elliottet al. 2007), and others emphasizing morphol-ogy, feeding habits, and body size (Table 1), thatcould be broadly applied across fishes and deca-pods. Additional references for trait data are pro-vided in Appendix S2. We used similaritiesbased on the nine functional traits to arrange spe-cies into a functional dendrogram based onGower distances and cut the dendrogram at aGower height of 0.475 to produce seven discretespecies clusters as functional groups (Fig. 2).

Table 1. Traits used for quadratic entropy (Q) analyses and functional group designation.

Trait Type Values/Range

Trophic level Continuous numerical 2.99–4.40Vertical distribution Categorical Benthic, epifaunal/eelgrass associate, pelagicMorphology Categorical Anguilliform, shrimp, compressed, fusiform, majoid, cottoid, depressed,

cancroidMedian size Continuous numerical Total length for fishes and shrimps, carapace width for crabs; 0.5–33 cmMaximum size Continuous numerical Total length for fishes and shrimps, carapace width for crabs; 0.5–33 cmPrey phyla Discrete numerical 1–6Estuarine UseFunctional Group

Categorical Semi-anadromous (SA), marine straggler (MS), estuarine resident (ER),estuarine-dependent marine migrant (MMD), opportunist marine migrant(MMO), freshwater migrant (FM), estuarine migrant (EM), anadromous(AN)

Reproductive ModeFunctional Group

Categorical Viviparous (V), egg-guarder (Og), egg-brooder (Os), unguarded benthiceggs (Ob), unguarded eggs attached to structure (Ov), pelagic eggs (Op)

Feeding ModeFunctional Group

Categorical Opportunistic feeder (OP), epibenthivore (ZB-E), generalist zoobenthivore(ZB), detritivore (DV), zooplanktivore (ZP), hyperbenthivore (ZB-H),omnivore (OV), infaunal feeder (ZB-I)

Notes: Estuarine Use, Reproductive Mode, and Feeding Mode Functional Groups are based on Elliott et al. (2007). Speciesfor which trait values were not available were assigned values based on species with similar diets, taxonomic position, and/orsize.

❖ www.esajournals.org 4 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

Mesopredator analyses—abundanceMesopredator abundance per seine was ana-

lyzed in aggregate, including all fishes and deca-pods, and as the number in each functional group.Abundance per seine (total and subdivided byfunctional group) was the response variable in lin-ear mixed-effects models including site, habitat,and their interaction as fixed effects and subsite asa random effect to account for two seines in eachreplicate and the spatial juxtaposition of habitattypes within subsites. To meet normality assump-tions, abundance was log-transformed. Due to thelinear mixed-effects structure, significance of fixed

effects was determined by likelihood ratio testscomparing models with and without each maineffect and the interaction. Post hoc comparisons ofhabitats were made using Tukey’s tests on modelsincluding fixed effects only.

Univariate analyses of taxonomic and traitdiversityDiversity was initially calculated for each seine

as the Shannon-Weiner diversity index (H0), appliedboth at the species level and to functional groups.Further, we calculated functional trait diversity asRao’s quadratic entropy (Q, Botta-Dukat 2005).

0.6 0.5 0.4 0.3 0.2 0.1 0.0Gower distance

Amphisticus rhodoterusCymatogaster aggregatusOncorhynchus tshawytschaGasterosteus aculeatusHypomesus pretiosusSyngnathus leptorhynchusHippolytidaePugettia productaPlatichythys stellatusParophrys vetulusUnidentified flatfishLumpenus sagittaApodichthy flavidusPholis ornataUnidentified gunnelClevelandia iosOligocottus maculosusLepidogobius lepidusUnidentified sculpin 1Unidentified sculpin 2Oregonia gracilisTelmessus cheiragonusCrangon sp.Cancer productusPandalus danaePandalus hypsinotusLeptocottus armatusPorichthys notatusHemigrapsus oregonensisMetacarcinus gracilisMetacarcinus magister

Generalist crabLarge cottoidSmall zoobenthivoreFlatfishCryptic epifaunaFusiform pelagicSurfperch

Fig. 2. Functional dendrogram of mesopredators observed in seines, clustered by Gower distance. The dendro-gram was cut at an arbitrary distance of 0.475 to form seven functional groups.

❖ www.esajournals.org 5 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

A linear mixed-effects model was applied to eachdiversity index (species, functional group, trait),with habitat, site, and their interaction as fixedeffects and subsite as a random effect. Signifi-cance was determined using likelihood ratio teststo compare more vs. less complex models.

Functional redundancy analysesWe measured functional redundancy by using

ANCOVAs to evaluate the relationships betweenfunctionally derived and taxonomically derivedShannon-Weiner indices, using each seine as asample. Separate analyses were carried out totest for differences in these relationships withingroups of seines grouped by habitat and by site.Model intercepts were fixed at zero. We alsocompared the slopes of these relationships to aslope of 1 (zero functional redundancy) usingone-tailed t tests for each habitat and site. Theseanalyses were conducted under the assumptionthat the relationships between diversity indiceswere in their linear phases, below the plateaupredicted by Micheli and Halpern (2005).

Multivariate analyses of mesopredatorsNon-metric multidimensional scaling (NMDS)

analyses were conducted on Bray-Curtis dissimi-larities to describe and visualize community vari-ation. Because certain samples exhibited highdominance of one or two species, we log(n + 1)-transformed the species abundance data per seineto downweight the most abundant species andallow rare species to exert more influence on dis-tance calculation (Clarke and Warwick 2001). Apermutational multivariate ANOVA (PERMA-NOVA, maximum permutations = 9999) was car-ried out on the sample-by-species abundancematrix (each seine as a sample) to test whethermesopredator communities differed according tohabitat, site, and their interaction. Subsites weretreated as strata within sites to ensure that ran-domizations occurred only within habitat patchesper subsite.

In addition to the above analyses based on tax-onomy, trait-based approaches were taken byperforming similar NMDS analyses using data as(1) log(n + 1)-transformed per-seine abundanceby functional group and (2) log(n + 1)-trans-formed per-seine abundance by trait (Table 1;Hewitt et al. 2016). We expected that the PER-MANOVAs for traits and functional groups

would show an increased amount of variationattributed to habitat, and reduced amount ofvariation attributed to site, relative to the species-based analyses. All analyses were carried outusing the lme4, FD, and vegan packages in R(Bates et al. 2014, Lalibert�e et al. 2014, Oksanenet al. 2015, R Core Development Team 2015).

RESULTS

Mesopredator abundanceSites differed substantially in mesopredator

species richness, with 14 species identified at theCoastal site, 15 at S fjord, and 27 at N fjord, butall seven functional groups occurred at all sites.Overall abundance per seine was 8.5 timesgreater in eelgrass than in unvegetated habitatsat all three sites, but edges showed less consis-tency, sometimes grouping with eelgrass and atother sites containing mesopredators at interme-diate abundance per seine. Overall, a significantsite-by-habitat interaction highlighted differencesin mesopredator response to edges across sites(Fig. 3; Appendix S1: Table S3). When examinedby functional group, eelgrass contained more surf-perches (40.39), fusiform pelagic species (41.59),and cryptic eelgrass climbers (4.49) than unvege-tated habitats, on average. Other functional groupsshowed no significant habitat associations (Fig. 3;Appendix S1: Table S4). Functional group repre-sentation also varied across sites, with the Coastalsite relatively more represented by surfperchesand generalist crabs, and the N fjord site contain-ing relatively more small zoobenthivores andcryptic eelgrass climbers (Fig. 3; Appendix S1:Table S4).

Mesopredator diversityDiversity metrics reflected that eelgrass har-

bored high abundance of one or a few dominantspecies per seine, which reduced Q and Shannon-Weiner indices. Consequently, these indices werelowest in eelgrass across taxonomic, functionalgroup, and trait-based approaches to definingdiversity. For species diversity, H0 was signifi-cantly influenced by site, habitat, and their inter-action (Appendix S1: Table S3). H0 was highest inunvegetated, intermediate on edges, and lowestin eelgrass for Coastal and S fjord, while N fjordshowed equal diversity in unvegetated and edgehabitat and low diversity in eelgrass (Fig. 4A).

❖ www.esajournals.org 6 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

When calculated for functional groups, ratherthan species, diversity showed the same patterns(Fig. 4B; Appendix S1: Table S3).

Trait diversity (Q) was significantly related tohabitat, site, and their interaction (Fig. 4C;Appendix S1: Table S3). Across habitats, Q washigher in unvegetated mudflat than in eelgrassor on edges, which were not significantly differ-ent. Q was significantly higher at N fjord than atthe Coastal site.

Functional redundancyThe slope of the relationship between taxo-

nomic and functional Shannon-Weiner diversitydid not differ significantly among habitats, andaveraged 0.85 � 0.017. This slope was signifi-cantly less than 1 (t0.05(1)53 = �9.28, P < 0.001),indicating significant functional group overlapbetween species in all habitats. Sites differed sig-nificantly from each other in terms of the relation-ship between functional and taxonomic diversityindices (F2,51 = 28.12, P < 0.001; Fig. 5); at the

Coastal site, slope was 0.97 � 0.025, S fjord0.89 � 0.021, and N fjord 0.76 � 0.017. Only thefjord sites showed that slopes were significantlydifferent from 1 (S fjord: t0.05(1)17 = �5.054,P < 0.001; N fjord: t0.05(1)17 = �14.27, P < 0.001),indicating significant functional redundancy.

Multivariate analyses of mesopredatorsAcross taxonomic-, functional group-, and trait-

based assemblages, eelgrass and unvegetated habi-tats were distinct, while edge was intermediate(Fig. 6). Both habitat and site were significant fac-tors influencing taxonomic structure in multivari-ate analyses, but site explained nearly twice asmuch of the variation in structure as habitat (habi-tat: R2 = 0.18, pseudo-F2,45 = 9.7, P < 0.001; site:R2 = 0.35, pseudo-F2,45 = 18.6, P < 0.001; Fig. 6A).Site and habitat remained significant in determin-ing community structure at the level of functionalgroups, but contributed more equally to the varia-tion in structure than they did at the species level(habitat: R2 = 0.21, pseudo-F2,45 = 10.23, P < 0.001;site: R2 = 0.29, pseudo-F2,45 = 14.2, P < 0.001;Fig. 6B). Similarly, at the functional trait level,site and habitat contributed equally to variation incommunity structure (R2 = 0.26, pseudo-F2,45 =13.89, P < 0.001, for both site and habitat; Fig. 6C).There was no significant site-by-habitat interactionfor any of these multivariate analyses.

DISCUSSION

Mesopredators share a set of functional traitswithin estuarine environments, which is evidentby virtue of their being grouped together basedon size and collective diet. Despite this, we distin-guished seven functional groups within meso-predators that differed in traits related to feedinghabits, life history, and morphology (Table 1,Fig. 2). Distinguishing functional groups revealedthat the enhanced abundance of mesopredatorsper seine with habitat structure was primarilydue to habitat selectivity of surfperches and pela-gic fusiform fishes. Most other functional groupswere found in statistically indistinguishable abun-dances per seine across the gradient of structuralcomplexity. Because of differential enhancementof a few functional groups by eelgrass, functionaland taxonomic diversity indices actually declinedwith structural complexity (Fig. 4), as assem-blages in eelgrass contained similar numbers of

CO SO NO CE SE NE CI SI NI

Abu

ndan

ce0

100

200

300

400

500

Generalist crabLarge cottoidSmall zoobenthivoreFlatfishCryptic epifaunaFusiform pelagicSurfperch

Fig. 3. Total abundance, subdivided into seven aposteriori functional groups. Bars represent site 9 habi-tat pairings, arranged from left to right by habitat.C = Coastal; S = S fjord; N = N fjord; O = unvege-tated; E = edge; and I = vegetated. Error bars representthe standard error of total abundance (six seines perhabitat at each site).

❖ www.esajournals.org 7 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

species but with less evenness across species incomparison with mesopredators in bare areas(Fig. 3). Overall, in this study, functional traitsprovided novel insight into how organisms withdifferent traits respond to eelgrass, relative tocombining mesopredators as a whole. On theother hand, functional and species diversityshowed similar patterns in their response to thestructural complexity of lower intertidal habitats.

A broadly accepted paradigm is that seagrassharbors more fish and decapods than adjacentunvegetated habitats (Heck et al. 1989, Dean et al.2000, Murphy et al. 2000, Hughes et al. 2002, Fer-raro and Cole 2010, Boyer et al. 2015). Our resultssupport this idea despite the close spatial proxim-ity of habitats forming a gradient of structuralcomplexity (Fig. 3). We found no additional bene-fit of habitat edge; in both multivariate and

univariate analyses, this habitat either groupedwith fully vegetated, or was intermediate (Fig. 6).In contrast, our results in terms of diversity high-light that some mesopredators may benefit muchmore from structural complexity than others(Fig. 4). These tended to be cryptically colored spe-cies or those that inhabit the water column, forwhich floating aquatic vegetation would providethe most shelter from predators as well as easilyaccessible food in the form of canopy mesograzers.Some studies from Washington State have pre-

sented exceptions to the paradigm of greatermesopredator abundance inside eelgrass thanoutside. Two studies from Willapa Bay reportedno difference in fish and decapod multivariatestructure among habitats differing in above-ground structure (Hosack et al. 2006, Dumbauldet al. 2015), one of which also showed no

H' sp

ecie

s

0.0

1.0

2.0 A

CoastalS fjordN fjord

H' fu

nctio

nal g

roup

s

0.0

1.0

2.0 B

Bare Edge Eelgrass

Q

0.00

0.10

0.20 C

Fig. 4. Diversity metrics in seines. (A) Shannon-Weiner diversity in seines, based on individual species; (B)Shannon-Weiner diversity in seines, based on functional groups (Fig. 2); (C) Rao’s quadratic entropy in seines,based on functional traits (Table 1). Quadratic entropy and species-based Shannon-Weiner responded signifi-cantly to site, habitat, and their interaction; functional group-based Shannon-Weiner responded significantly tohabitat and a habitat-by-site interaction. Error bars represent standard error (six seines per habitat at each site).

❖ www.esajournals.org 8 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

difference in abundance (Hosack et al. 2006).These studies used methods that sampled largerareas of the tideflat (freestanding fyke net andtow net, respectively), and they did not specifi-cally sample across the vegetated–unvegetatedecotone. Sampling over large areas of patchyhabitat might have obscured finer-scale differ-ences in assemblage structure across habitats.

The functional approach of our study helpeddisentangle the use of habitat types by specificfunctional groups, and we were able to determinethat while the pattern of no difference in abun-dance per seine holds for some groups, otherswere significantly more abundant in eelgrassseines than in seines in unvegetated habitats.Overall, our study forms an important contribu-tion to the understanding of mesopredator usageof eelgrass in the Pacific Northwest, which ispoorly known relative to other regions (Dean et al.2000, Murphy et al. 2000, Ferraro and Cole 2010).

Functional redundancy and mesopredatorresilience

A common finding in structurally complexhabitats is that functional and taxonomic diversity

are highly correlated (Jackson et al. 2001, Micheliand Halpern 2005, Bustamante et al. 2014) andthat functional redundancy is correspondinglylow. These relationships make sense in light of theidea that structural complexity increases speciesrichness by increasing available niches andthereby reducing competition, and that rare spe-cies that contribute significantly to species rich-ness are often functionally distinct (MacArthurand MacArthur 1961, Mouillot et al. 2013b). Incontrast, unstructured habitats or habitats inwhich structure has been removed might be pop-ulated by more generalist species with broad over-lap in traits, conferring functional redundancy tothe community (Thrush et al. 2006, Handley et al.

0.0 0.5 1.0 1.5 2.0

0.0

0.5

1.0

1.5

2.0

H 'species

H' fu

nctio

nal g

roup

sCoastalS fjordN fjord1:1

Fig. 5. Plot of relationships between species-basedand functional group-based Shannon-Weiner diversityindices across sites. The slopes of these lines representthe degree of functional overlap among species, andare compared to a solid 1:1 line representing zero func-tional redundancy. Sites differed significantly fromeach other, although only S and N fjord differed signif-icantly from the 1:1 line.

−1.5

−0.5

0.5

1.5

NM

DS

2

Stress = 0.159A

−1.5

−0.5

0.5

1.5

NM

DS

2

Stress = 0.177BCoastalS fjordN fjord

BareEdgeEelgrass

−3 −2 −1 0 1 2 3

−1.5

−0.5

0.5

1.5

NM

DS

2

Stress = 0.127C

Fig. 6. Non-metric multidimensional scaling (NMDS)plots of species (A), functional groups (B), and func-tional traits (C) in seines.

❖ www.esajournals.org 9 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

2014). These patterns may increase the vulnerabil-ity of structurally complex ecosystems to distur-bance; redundancy of species within functionalgroups is often associated with increased ecosys-tem resilience, and greater numbers of rare, func-tionally distinct species in habitats like kelp forestsand coral reefs decrease this resilience (Folke et al.2004, Tilman et al. 2006, Downing et al. 2012).

In contrast to these assumptions, our datashowed increased functional redundancy withincreased species richness; richness and redun-dancy were dependent on site, and habitat com-plexity did not influence their relationship. Anumber of factors may contribute to this discrep-ancy. First, from a practical perspective, thedegree of functional redundancy observed in asystem is dependent on the scale at which func-tional groups are defined. Finer-scale functionalresolution increases the number of functionalgroups and decreases the number of specieswithin each group, thereby increasing the slopebetween taxonomic and functional group diver-sity; had we defined more functional groups, wewould have observed less functional redundancy(Micheli and Halpern 2005). Second, assemblageconnectivity may contribute to redundancybecause species that are both functionally redun-dant and far ranging may contribute to richnessat a site while increasing functional redundancyin a particular subsite or habitat (Auster andLink 2009). The habitats we sampled were patchyand had broad connections to the surroundingregions, and many of the species we observedare only part-time residents of these estuariesand are considered to use both eelgrass andunstructured tideflats at juvenile stages (Hugheset al. 2014). High mesopredator mobility andhabitat connectivity might have contributed tothe relatively high functional redundancy weobserved at our most species-rich sites becausethese factors increase the likelihood that multiplespecies with similar functional traits or from dif-ferent habitats will end up in the same samplingarea. Although this was likely not the case in ourstudy, a third explanation for high functionalredundancy in species-rich areas is that theseareas may have a greater number of rare species,which are often functionally distinct (Mouillotet al. 2013b), creating greater functional redun-dancy in these areas by sampling probability. Inour study, rare species did not form distinct

functional groups and thus had little influenceon functional redundancy.The different degrees of functional overlap

between species observed in our study sitessuggest differing levels of ecosystem resiliency(Auster and Link 2009; Fig. 5), potentially neces-sitating site-specific management plans for main-taining ecosystem function. It is also importantthat management decisions be based not only onpotential ecosystem resiliency but also on theidentities of specific community members,especially given the cultural and economicvalue of some community members comparedto others that might be functionally equivalent.For example, Metacarcinus magister and Hemi-grapsus oregonensis are both in the generalistcrab functional group, but only M. magister iscommercially valuable.

Species- and trait-based evaluation of habitatdifferencesUsing different response variables at taxo-

nomic and functional resolutions generally sup-ported the notion that a trait-based perspectiveis more robust in explaining assemblage varia-tion across sites varying in species pool. From ataxonomic perspective, site explained nearlytwice as much of the multivariate variation incommunity structure as did habitat, but site andhabitat were equally important in explaining thevariation in community structure through afunctional lens. The species in assemblages areconstrained by stochastic, broad-scale historicalprocesses like vicariance and dispersal leading tounique assemblages in particular geographicareas. Functional traits in assemblages, however,are more often determined by focused, smaller-scale processes, like environmental filtersimposed by a particular habitat, or interspecificinteractions (Pavoine and Bonsall 2011, Adleret al. 2013, Mouillot et al. 2013a, Kraft et al.2015). In multivariate analyses, site explainedless variation in functional group and trait com-position than species composition, but notwith-standing, the role of habitat was relatively small.This may reflect that environmental filteringeffects differed more among sites than amonghabitats within sites, that species were allocatedto functional groups that did not accurately char-acterize their role in the ecosystem, or that thesmall spatial scale of the habitat mosaic makes it

❖ www.esajournals.org 10 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

more likely that all functional groups and traitsare present in all habitats.

Seagrass beds and estuaries worldwide pro-vide “nursery habitat” for juvenile fishes anddecapods before moving to offshore waters. Thisnursery role is one of the most oft-cited servicesof eelgrass beds (Costanza 1997, Orth et al. 2006,Cullen-Unsworth and Unsworth 2013). A surveyof estuaries on the west coast of the United Statesrevealed eelgrass beds and unvegetated tideflatsas the most commonly used estuarine nurseryhabitats among 15 ecologically and economicallyimportant species (Hughes et al. 2014). Twelve ofthese species are documented to use our focalestuaries as juveniles, and seven of them—sandshrimp (Crangon sp.), Dungeness crab (M. magis-ter), Chinook salmon (Oncorhynchus tshawytscha),English sole (Parophrys vetulus), starry flounder(Platichthys stellatus), shiner perch (Cymatogasteraggregata), and staghorn sculpin (Leptocottusarmatus)—were caught in seines. Two of thesespecies—C. aggregata and O. tshawytscha—belongto functional groups with highest abundanceobserved in eelgrass patches. The data on theseand other species collected in this study canbe used to quantitatively assess the value andecosystem services of eelgrass and other habitatsin Washington’s tideflats, and will ultimatelyhelp inform management and policy decisions.Because eelgrass benefits species with certaintraits, management focused on the low-intertidalestuarine habitat mosaic, rather than a particularhabitat type, and on places where redundancy isalready low, best supports mesopredator diver-sity and function.

ACKNOWLEDGMENTS

This project was supported by Washington Depart-ment of Natural Resources through an interagencyagreement with the University of Washington (IAA16-19) and by Washington Sea Grant Program, pur-suant to National Oceanic and Atmospheric Admin-istration Award NA14OAR4170078. The viewsexpressed herein are those of the authors and do notnecessarily reflect the views of funding agencies orsub-agencies. Help in the field was provided byB. Dumbauld, J. Borin, T. Leach, A. Lowe, A. Klemmer,M. Ware, D. Sare, and A. Dowty. The manuscript wasimproved by comments from M. Dethier, E. Grason,and one anonymous reviewer.

LITERATURE CITED

Adler, P. B., A. Fajardo, A. R. Kleinhesselink, and N. J.B. Kraft. 2013. Trait-based tests of coexistencemechanisms. Ecology Letters 16:1294–1306.

Aguilar-Medrano, R., and L. E. Calderon-Aguilera.2016. Redundancy and diversity of functional reeffish groups of the Mexican Eastern Pacific. MarineEcology-an Evolutionary Perspective 37:119–133.

Auster, P. J., and J. S. Link. 2009. Compensation andrecovery of feeding guilds in a northwest Atlanticshelf fish community. Marine Ecology ProgressSeries 382:163–172.

Bates, D., M. Maechler, B. Bolker, S. Walker, R. H. B.Christensen, H. Singmann, and B. Dai. 2014. lme4:linear mixed-effects models using Eigen and S4. Rpackage version 1.1-7. http://CRAN.R-project.org/package=lme4

Blandon, A., and P. S. E. Z. Ermgassen. 2014. Quantita-tive estimate of commercial fish enhancement byseagrass habitat in southern Australia. EstuarineCoastal and Shelf Science 151:1–8.

Botta-Dukat, Z. 2005. Rao’s quadratic entropy as ameasure of functional diversity based on multipletraits. Journal of Vegetation Science 16:533–540.

Boyer, K., C. Pinnell, J. Miller, J. Moderan, andK. Stockmann. 2015. Fauna use and interactionsamong restored oyster reefs and eelgrass at the SanFrancisco Bay Living Shorelines: nearshore Link-ages project. Journal of Shellfish Research 34:613.

Bustamante, M., J. Tajadura, J. M. Gorostiaga, and J. I.Saiz-Salinas. 2014. Response of rocky invertebratediversity, structure and function to the verticallayering of vegetation. Estuarine Coastal and ShelfScience 147:148–155.

Clarke, K. R., and R. M. Warwick. 2001. Change inmarine communities: an approach to statistical anal-ysis and interpretation. Second edition. PRIMER-E,Plymouth, UK.

Costanza, R., et al. 1997. The value of the world’secosystem services and natural capital. Nature387:253–260.

Cullen-Unsworth, L., and R. Unsworth. 2013. Seagrassmeadows, ecosystem services, and sustainability.Environment: Science and Policy for SustainableDevelopment 55:14–27.

Dean, T. A., L. Haldorson, D. R. Laur, S. C. Jewett, andA. Blanchard. 2000. The distribution of nearshorefishes in kelp and eelgrass communities in PrinceWilliam Sound, Alaska: associations with vegeta-tion and physical habitat characteristics. Environ-mental Biology of Fishes 57:271–287.

Downing, A. S., E. H. van Nes, W. M. Mooij, andM. Scheffer. 2012. The resilience and resistance of

❖ www.esajournals.org 11 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

an ecosystem to a collapse of diversity. PLoS ONE7:e46135.

Duffy, J. E., J. P. Richardson, and K. E. France. 2005.Ecosystem consequences of diversity depend onfood chain length in estuarine vegetation. EcologyLetters 8:301–309.

Duffy, J. E., et al. 2015a. Biodiversity mediates top-down control in eelgrass ecosystems: a global com-parative-experimental approach. Ecology Letters18:696–705.

Duffy, J. E., S. L. Ziegler, J. E. Campbell, P. M. Bippus,and J. S. Lefcheck. 2015b. Squidpops: a simple toolto crowdsource a global map of marine predationintensity. PLoS ONE 10:e0142994.

Dumbauld, B. R., G. R. Hosack, and K. M. Bosley.2015. Association of juvenile salmon and estuarinefish with intertidal seagrass and oyster aquacul-ture habitats in a northeast Pacific Estuary. Trans-actions of the American Fisheries Society 144:1091–1110.

Elliott, M., A. K. Whitfield, I. C. Potter, S. J. M. Blaber,D. P. Cyrus, F. G. Nordlie, and T. D. Harrison. 2007.The guild approach to categorizing estuarine fishassemblages: a global review. Fish and Fisheries8:241–268.

Ferraro, S. P., and F. A. Cole. 2010. Ecological periodictables for nekton usage of four US Pacific North-west estuarine habitats. Canadian Journal ofFisheries and Aquatic Sciences 67:1957–1967.

Folke, C., S. Carpenter, B. Walker, M. Scheffer, T. Elmq-vist, L. Gunderson, and C. S. Holling. 2004. Regimeshifts, resilience, and biodiversity in ecosystemmanagement. Annual Review of Ecology Evolutionand Systematics 35:557–581.

Froese, R., and D. Pauly. 2000. FishBase 2000: concepts,design and data sources. ICLARM, Laguna,Philippines.

Froese, R., and D. Pauly. 2015. FishBase. http://www.fishbase.org/search.php

Handley, S. J., T. J. Willis, R. G. Cole, A. Bradley, D. J.Cairney, S. N. Brown, and M. E. Carter. 2014. Theimportance of benchmarking habitat structure andcomposition for understanding the extent of fish-ing impacts in soft sediment ecosystems. Journal ofSea Research 86:58–68.

Heck, K. L., K. W. Able, M. P. Fahay, and C. T. Roman.1989. Fishes and decapod crustaceans of Cape Codeelgrass meadows – species composition, seasonalabundance patterns and comparison with unvege-tated substrates. Estuaries 12:59–65.

Heck, K. L., G. Hays, and R. J. Orth. 2003. Criticalevaluation of the nursery role hypothesis for sea-grass meadows. Marine Ecology Progress Series253:123–136.

Hewitt, J. E., J. Norkko, L. Kauppi, A. Villnas, andA. Norkko. 2016. Species and functional traitturnover in response to broad-scale change and aninvasive species. Ecosphere 7:11.

Hickey, B. M., and N. S. Banas. 2003. Oceanographyof the US Pacific Northwest Coastal Ocean andestuaries with application to coastal ecology. Estu-aries 26:1010–1031.

Hooper, D. U., et al. 2005. Effects of biodiversity onecosystem functioning: a consensus of currentknowledge. Ecological Monographs 75:3–35.

Hosack, G. R., B. R. Dumbauld, J. L. Ruesink, and D. A.Armstrong. 2006. Habitat associations of estuarinespecies: comparisons of intertidal mudflat, seagrass(Zostera marina), and oyster (Crassostrea gigas)habitats. Estuaries and Coasts 29:1150–1160.

Hughes, J. E., L. A. Deegan, J. C. Wyda, M. J. Weaver,and A. Wright. 2002. The effects of eelgrass habitatloss on estuarine fish communities of southernNew England. Estuaries 25:235–249.

Hughes, B. B., M. D. Levey, J. A. Brown, M. C. Fountain,A. B. Carlisle, S. Y. Litvin, C. M. Greene, W. N.Heady, and M. G. Gleason. 2014. Nursery functionsof U.S. west coast estuaries: the state of knowledgefor juveniles of focal invertebrate and fish species.The Nature Conservancy, Arlington, Virginia,USA.

Jackson, J. B. C., et al. 2001. Historical overfishing andthe recent collapse of coastal ecosystems. Science293:629–638.

Kawase, M., and B. Bang. 2013. Seasonal variability ofsalinity and circulation in a silled estuarine fjord: anumerical model study. Continental Shelf Research71:109–126.

Keister, J. E., and L. B. Tuttle. 2013. Effects of bottom-layer hypoxia on spatial distributions and commu-nity structure of mesozooplankton in a sub-estuaryof Puget Sound, Washington, U.S.A. Limnologyand Oceanography 58:667–680.

Kraft, N. J. B., P. B. Adler, O. Godoy, E. C. James,S. Fuller, and J. M. Levine. 2015. Communityassembly, coexistence and the environmental filter-ing metaphor. Functional Ecology 29:592–599.

Lalibert�e, E., P. Legendre, and B. Shipley. 2014. FD:measuring functional diversity (FD) from multipletraits, and other tools for functional ecology. Rpackage version 1.0-12. http://CRAN.R-project.org/package=FD

Lefcheck, J. S., and J. E. Duffy. 2015. Multitrophic func-tional diversity predicts ecosystem functioning inexperimental assemblages of estuarine consumers.Ecology 96:2973–2983.

MacArthur, R., and J. W. MacArthur. 1961. On birdspecies-diversity. Ecology 42:594–598.

❖ www.esajournals.org 12 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.

Micheli, F., and B. S. Halpern. 2005. Low functionalredundancy in coastal marine assemblages. Ecol-ogy Letters 8:391–400.

Mouillot, D., et al. 2013a. Rare species support vulner-able functions in high-diversity ecosystems. PlosBiology 11:11.

Mouillot, D., N. A. J. Graham, S. Villeger, N. W. H.Mason, and D. R. Bellwood. 2013b. A functionalapproach reveals community responses to distur-bances. Trends in Ecology & Evolution 28:167–177.

Murphy, M. L., S. W. Johnson, and D. J. Csepp. 2000. Acomparison of fish assemblages in eelgrass andadjacent subtidal habitats near Craig, Alaska.Alaska Fishery Research Bulletin 7:11–21.

Nagelkerken, I., and G. van der Velde. 2004. A com-parison of fish communities of subtidal seagrassbeds and sandy seabeds in 13 marine embaymentsof a Caribbean island, based on species, families,size distribution and functional groups. Journal ofSea Research 52:127–147.

Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R.Minchin, R. B. O’Hara, G. L. Simpson, P. Solymos,M. H. H. Stevens, and H. Wagner. 2015. vegan:community Ecology Package. R package version2.2-1. http://CRAN.R-project.org/package=vegan

Orth, R. J., et al. 2006. A global crisis for seagrassecosystems. BioScience 56:987–996.

Pavoine, S., and M. B. Bonsall. 2011. Measuring biodi-versity to explain community assembly: a unifiedapproach. Biological Reviews 86:792–812.

R Core Development Team. 2015. R: a language andenvironment for statistical computing. R Founda-tion for Statistical Computing, Vienna, Austria.

Thrush, S. F., J. S. Gray, J. E. Hewitt, and K. I. Ugland.2006. Predicting the effects of habitat homogeniza-tion on marine biodiversity. Ecological Applica-tions 16:1636–1642.

Tilman, D., P. B. Reich, and J. M. H. Knops. 2006. Bio-diversity and ecosystem stability in a decade-longgrassland experiment. Nature 441:629–632.

Verberk, W., C. G. E. van Noordwijk, and A. G.Hildrew. 2013. Delivering on a promise: integratingspecies traits to transform descriptive communityecology into a predictive science. FreshwaterScience 32:531–547.

Villeger, S., J. R. Miranda, D. F. Hernandez, A. S. Lopez,and D. Mouillot. 2008. Stable trophic structure acrosscoastal nekton assemblages despite high species turn-over. Marine Ecology Progress Series 364:135–146.

Waycott, M., et al. 2009. Accelerating loss of seagrassesacross the globe threatens coastal ecosystems.Proceedings of the National Academy of Sciencesof the United States of America 106:12377–12381.

Williams, G. D., et al. 2001. Features of Puget Soundregion: oceanography and physical processes.Pages 20–32 in J. S. Brennan, editor. Reconnais-sance assessment of the state of the nearshore eco-system: eastern Shore of Central Puget Sound,including Vashon and Maury Islands (WRIAs 8and 9). King County Department of NaturalResources, Seattle, Washington, USA.

Wong, M. C., and M. Dowd. 2015. Patterns in taxo-nomic and functional diversity of macrobenthicinvertebrates across seagrass habitats: a casestudy in Atlantic Canada. Estuaries and Coasts 38:2323–2336.

SUPPORTING INFORMATION

Additional Supporting Information may be found online at: http://onlinelibrary.wiley.com/doi/10.1002/ecs2.1792/full

❖ www.esajournals.org 13 April 2017 ❖ Volume 8(4) ❖ Article e01792

GROSS ET AL.