MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 520: 1–20, 2015doi: 10.3354/meps11170

Published February 3

INTRODUCTION

Coral reefs support fisheries that are essential forthe well-being of millions of people (Pauly et al.2002). However, these fisheries often target specieswhich perform important ecological functions, suchas herbivorous fishes, whose grazing prevents theproliferation of algae that compete with corals forspace (Koslow et al. 1988, Bellwood et al. 2004,McClanahan & Cinner 2008). Because removal ofherbivores can shift the competitive balance be -tween algae and reef-building corals, fishing couldreduce the capacity of corals to withstand or recoverfrom perturbations, including those resulting from

© Inter-Research 2015 · www.int-res.com*Corresponding author: thomas.adam@lifesci.ucsb.edu

FEATURE ARTICLE: REVIEW

Herbivory and the resilience of Caribbean coralreefs: knowledge gaps and implications for

management

Thomas C. Adam1,4,*, Deron E. Burkepile1, Benjamin I. Ruttenberg2, Michelle J. Paddack3

1Marine Sciences Program, Department of Biological Sciences, Florida International University, North Miami, FL 33181, USA2NOAA Fisheries, Southeast Fisheries Science Center, Miami, FL 33149, USA and Biological Sciences Department, California

Polytechnic State University, San Luis Obispo, CA 93410, USA3Biological Sciences Department, Santa Barbara City College, Santa Barbara, CA 93109, USA

4Present address: Marine Science Institute, University of California, Santa Barbara, CA 93106, USA

ABSTRACT: Herbivory is a key process on coral reefsthat can facilitate reef-building corals by excluding algaethat otherwise negatively impact coral settlement,growth, and survivorship. Over the last several de cades,coral cover on Caribbean reefs has declined precipitously.On many reefs, large structurally complex corals havebeen replaced by algae and other non-reef-building or-ganisms, resulting in the collapse of physical structureand the loss of critical ecosystem services. The drivers ofcoral decline on Caribbean reefs are complex and varyamong locations. On many reefs, populations of key her-bivores have been greatly reduced by disease and over-fishing, and this has resulted in the proliferation of algaethat hinder coral recovery following major disturbances.Yet, evidence that increases in herbivory can promotecoral recovery on Caribbean reefs has been mixed. Here,we discuss key contingencies that will modify the rela-tionships between herbivores, algae, and corals andidentify critical knowledge gaps that limit our ability topredict when and where herbivores are most likely tofacili tate coral persistence and recovery. Impacts of her-bivores on coral reef ecosystems will vary greatly inspace and time and will depend on herbivore diversityand species identity. While there are still a large numberof knowledge gaps, we make several management rec-ommendations based on our current understanding ofthe processes that structure reef ecosystems. Reversingthe fate of Caribbean coral reefs will require the de -velopment of integrated management strategies that simultaneously address multiple stressors in addition tothe impacts of fisheries on herbivore assemblages.

KEY WORDS: Phase shift · Grazing impacts · Macroalgae ·Parrotfish ·Fishing ·Diadema ·Climate change ·Restoration

Resale or republication not permitted without written consent of the publisher

Blue parrotfish Scarus coeruleus cropping algae on a reef.

Photo: T. C. Adam

FREEREE ACCESSCCESS

Mar Ecol Prog Ser 520: 1–20, 2015

coastal development and global climate change (Bell -wood et al. 2004, Mumby 2006, Hughes et al. 2010).Thus, there are likely to be important trade-offsbetween short-term fisheries goals of maximizingyield and profit and the long-term sustainability ofcoral reef ecosystems and the services they provide(Brown & Mumby 2014).

In light of these trade-offs, it is critical to under-stand how fishing impacts reef-building corals byaltering the process of herbivory on coral reefs. In -tense feeding by herbivorous fishes and sea urchinsfavors corals by excluding upright macroalgae andfacilitating the development of benthic communitiesdominated by taxa tolerant to grazing, includingcrustose coralline algae and closely cropped filamen-tous turfs (Steneck 1988, Hay 1991, Carpenter 1997),which tend to have neutral or positive impacts oncorals (Harrington et al. 2004, Price 2010, Barott et al.2012). In contrast, the macroalgae and dense turfsthat develop under low grazing regimes negativelyimpact settlement of coral larvae as well as growthand survivorship of adult corals (River & Edmunds2001, Nugues & Szmant 2006, Box & Mumby 2007,Rasher & Hay 2010, Vega Thurber et al. 2012).

However, these broad generalizations about her -bivore−algae−coral interactions are likely context- dependent. For example, herbivores differ greatly intheir feeding preferences and behaviors, and conse-quently different species have unique impacts onbenthic communities (Choat et al. 2004, Mantyka &Bellwood 2007, Burkepile & Hay 2008). Herbivoresalso vary widely in their susceptibility to fishing andother anthropogenic stressors (Bellwood et al. 2012,Edwards et al. 2014). In addition, other factors be sidescompetition with algae can limit coral success, suchas sedimentation, nutrient loading, and disease (Mc-Clanahan et al. 2002). Therefore, it is critical to under-stand the roles that particular types of herbivores playin limiting harmful algae and facilitating corals undera range of environmental conditions to improve sus-tainable management of coral reef ecosystems.

In this manuscript, we identify critical knowledgegaps which limit our ability to predict when herbi-vores are likely to promote coral persistence andrecovery. We also explore what types of managementinterventions may be most effective for maintaininghealthy coral reef ecosystems. In particular, we askquestions centered around 3 main topics: (1) Whatprocesses operate to prevent or facilitate coral per-sistence and recovery, and how are these influencedby herbivory? (2) What are the independent andcombined effects of different species of herbivores inlimiting algae and facilitating reef-building corals?

(3) What factors limit herbivore populations and theprocess of herbivory on coral reefs? We focus on coralreefs throughout the wider Caribbean basin, wherefishing pressure is high, coral decline has beensevere, and algae have increased in abundance inrecent decades (Gardner et al. 2003, Bruno et al.2009, Schutte et al. 2010, Jackson et al. 2014).

Coral cover began declining throughout much ofthe Caribbean in the early 1980s following severalmajor hurricanes, a coral disease epidemic, and theproliferation of macroalgae following the mass die-off of the herbivorous sea urchin Diadema antillarum(Gardner et al. 2003, Schutte et al. 2010, Jackson etal. 2014). In subsequent years, corals have continuedto decline in response to a complex combination ofchronic and acute drivers, including declining waterquality (Rogers 1990, Fabricius 2005, Sutherland etal. 2011), increases in the frequency and severity ofmass bleaching events (Baker et al. 2008, Eakin et al.2010), and impacts from several of the most activehurricane seasons on record (Gardner et al. 2005,Williams & Miller 2012). During this time, large reef-building corals have experienced especially steepdeclines, resulting in a rapid collapse in structuralcomplexity (Alvarez-Filip et al. 2011) and cascadingindirect effects on reef-associated biota includingreef fishes (Paddack et al. 2009). Declines in coralcover have been accompanied by abrupt shifts incommunity structure, with once-dominant corals be -ing replaced by smaller weedy coral species, non-reef-building invertebrates, and/or fleshy macroalgae(Green et al. 2008, Norström et al. 2009, Schutte et al.2010, Burman et al. 2012, Ruzicka et al. 2013, Tothet al. 2014). These persistent changes are unprece-dented in the recent geologic past (Jackson 1992,Aronson & Precht 1997, Pandolfi & Jackson 2006),giving rise to the paradigm that human activity hascaused a regime shift throughout the Caribbean(Knowl ton 1992, Mumby & Steneck 2008, Hughes etal. 2010).

A central tenet of this paradigm is that the reefs ofthe 1980s were unexpectedly vulnerable to distur-bance due to the serial depletion of important con-sumers—especially herbivorous fishes—in the de -cades prior to ecosystem collapse (Jackson 1997,Jackson et al. 2001, Pandolfi et al. 2003). As a conse-quence, the die-off of D. antillarum triggered a rapid‘phase shift’ to algal dominance by removing the lastremaining abundant herbivore in many places in theCaribbean (e.g. Hughes 1994). In addition to being asymptom of the larger regime shift, algae are widelybelieved to play a critical role in preventing coral recovery, prompting strong calls for the protection of

2

Adam et al.: Herbivory and Caribbean coral reefs

herbivorous fishes (Bellwood et al. 2004, Mumby &Steneck 2008, Hughes et al. 2010, Jackson et al.2014). However, evidence that herbivorous fishes canpromote coral recovery on Caribbean reefs has beeninconsistent. Whereas localized recovery of D. antil-larum has been strongly associated with heightenedcoral recruitment and enhanced growth and survivor-ship of juvenile corals (Edmunds & Carpenter 2001,Carpenter & Edmunds 2006, Myhre & Acevedo-Gutiérrez 2007, Furman & Heck 2009, Idjadi et al.2010), studies throughout the Caribbean have oftenfailed to find strong links between the biomass of her-bivorous fishes and coral cover (e.g. Newman et al.2006, Sandin et al. 2008, Burkepile et al. 2013). Fur-ther, a recent meta-analysis suggests that protectionof herbivorous fishes within marine protected areashas not been associated with increases in coral coverrelative to fished locations (Guarderas et al. 2011; butsee Mumby & Harborne 2010, Selig & Bruno 2010).

The emerging picture suggests that impacts of herbivores on coral recovery are likely to be highlycontext-dependent. A deeper understanding of therelationships between different types of herbivores,algae, and corals is therefore needed to facilitatemanagement. With this goal in mind, we identify critical knowledge gaps regarding the ability of her-bivores to promote coral recovery under a range ofenvironmental conditions. These gaps are presentedas a series of broad research questions, which, ifaddressed, would improve the capacity of managersto prioritize strategies for recovering and/or main-taining healthy coral reef ecosystems.

KNOWLEDGE GAPS

Question 1. What processes operate to prevent orfacilitate coral persistence and recovery and how

are these influenced by herbivory?

Understanding the factors that confer resilience tocoral reefs is critical for understanding the recent tra-jectory of Caribbean reefs and for predicting theirfuture dynamics. Here, we define resilience as thecapacity of a reef to recover (i.e. return to its previousstate) following an acute perturbation such as a hur-ricane or coral bleaching event. Healthy herbivorepopulations may be an important factor leading toresilience by preventing the establishment and proliferation of algae following disturbances. Whilemany Caribbean reefs have failed to recover fromrecent disturbances (but see Manfrino et al. 2013),

reefs in the Indo-Pacific often recover rapidly fromacute pulse disturbances (Connell 1997, Sheppard etal. 2008, Adjeroud et al. 2009, Gilmour et al. 2013),indicating a high level of resilience that is oftenlinked to robust herbivore populations.

However, the extent that herbivores can promoteresilience will depend on the range of factors limitingboth coral and algal populations and impacting theability of herbivory to mediate interactions betweenthe two. Reef ecosystems are heterogeneous sea-scapes consisting of a mosaic of individual reefs, eachwith different physical and biological characteristicsand at different successional stages. Consequently,the factors limiting coral, algae, and herbivore popu-lations will vary greatly in space. Acute pulse pertur-bations such as hurricanes and bleaching events, aswell as chronic press-perturbations such as pollutionand sedimentation, can further modify herbivore−algae− coral interactions. Impacts of herbivores canalso depend on the presence of feedbacks that canreduce herbivory, facilitate algae, and impede coralrecovery following a perturbation (Knowlton 1992,Mumby & Steneck 2008, Nyström et al. 2012). Theability of herbivores to facilitate the maintenance andrecovery of resilient coral reefs will therefore behighly context-dependent (McClanahan et al. 2002,Aronson & Precht 2006). Understanding how ex

ternal drivers such as hurricanes, coral bleachingevents, and local anthropogenic impacts can interactwith feedbacks to reduce herbivory and preventcoral recovery requires careful consideration of thewide range of scales over which these processesoperate. Even large perturbations initially operate onlocal scales to disrupt ecological interactions on indi-vidual reefs. However, as perturbations become morefrequent and/or spatially extensive, they will beginto alter patterns of connectivity among reefs in thewider seascape (Fig. 1). In this section we ask howthe impacts of herbivores on coral persistence andrecovery will vary in space and time due to variationin human impacts and the natural factors that shapebenthic communities on coral reefs.

How do ecological feedbacks mediate the impacts ofherbivores on reef dynamics?

Understanding the factors responsible for recentcoral decline on Caribbean reefs is critical for thedevelopment of effective management strategiesbecause corals are unlikely to recover if the chroniccauses of coral mortality are not addressed. However,on many reefs, factors in addition to those respon -

3

Mar Ecol Prog Ser 520: 1–20, 2015

sible for initial mortality are likely oper-ating to prevent coral recovery (i.e. resil-ience of the coral system has been lost).Thus, the factors limiting the recovery ofcorals will often be different from thosethat initially caused their decline. Wherechronic anthropogenic impacts such assedimentation or nutrient loading haveled directly to coral decline, eliminatingthose impacts will be necessary to facili-tate coral recovery. However, becausefeed backs can prevent coral recoveryeven after the initial causes of coral de -cline have been removed, intense man-agement intervention may also be re -quired to break feedbacks that operate toslow or prevent coral recovery. A numberof feedbacks may prevent coral recoveryon Caribbean reefs, and many of theseare directly dependent on the ability ofherbivores to control algae (Fig. 1).

One of the key feedbacks that likelyreinforces an algae-dominated state isthe dilution of grazing intensity as coralsdecline and create large amounts ofspace for the establishment of benthicalgae (Williams et al. 2001, Paddack et al.2006, Mumby et al. 2007). Even in theabsence of fishing pressure, loss of coralswill decrease grazing intensity as herbi-vores graze over a larger area of reef.Macroalgae that were once excludedunder higher grazing pressure may thenbecome established. This can result in areef shifting from a state dominated byearly successional stages of highly palat-able and productive algal turfs to a statedominated by later successional stages ofless palatable and less productive macro-algae (e.g. Carpenter 1986). Dense standsof macroalgae can deter herbivory (Hoey& Bellwood 2011) and prevent the re -establishment of corals by inhibitingcoral recruitment (Birrell et al. 2008).Over longer time scales, declines in coralcover will reduce habitat complexity, po -tentially reducing habitat for fishes anddriving declines in herbivore populations(Graham et al. 2006, Paddack et al. 2009,Adam et al. 2014), further reinforcing theshift to an algae-dominated state. Finally,as coral decline becomes more spatiallyextensive, coral populations will reach

4

Fig. 1. (A) Illustration of the external drivers (red arrows) and internal feed-backs (gray arrows) that can lead to coral decline and prevent coral recoveryon degraded reefs (modified from Nyström et al. 2012). Feedbacks generatedby local competitive interactions between corals and algae are influenced bylocal factors such as nutrient loading, sedimentation, and overfishing, as wellas global increases in CO2 emissions which can interact to directly kill coralsand slow coral growth relative to algae (1). Feedbacks are strengthenedthrough time as 3-dimensionally complex coral reefs degrade, reducinghabitat for herbivorous fishes and further reinforcing an algae-dominatedstate (2). As the number of degraded reefs increases, fewer corals remain foreffective reproduction, potentially leading to region-wide recruitment failure(3). (B) Managing coral reefs in an era of rapid climate change requires actionat multiple levels of governance to address drivers of coral decline thatemerge from human activities at different spatial scales. Note that manage-ment at all levels will be most effective if enacted early. OA: ocean acidification

Adam et al.: Herbivory and Caribbean coral reefs

critically low densities, limiting their ability to suc-cessfully reproduce and ultimately leading to region-wide recruitment failure (Knowlton 1992, Hughes &Tanner 2000). While all of these processes (and oth-ers) may operate to inhibit coral recovery on manyCaribbean reefs, the relative importance of each isunknown and will likely depend strongly on environ-mental context. Managing herbivores to facilitatecorals requires that we can identify when and whereparticular feedbacks are operating to reduce her-bivory and/or prevent coral recovery.

How does the frequency, intensity, and scale ofdisturbance mediate the impacts of herbivores on

reef dynamics?

Coral reef ecosystems vary widely in their suscep-tibility to natural disturbances. At one extreme, manyshallow, exposed reefs undergo rapid cycles of dis-turbance and recovery, routinely losing almost all oftheir coral cover every decade or so due to tropicalcyclones (e.g. Connell et al. 1997, Trapon et al. 2011).At the other end of the spectrum, massive coralcolonies growing in less disturbance-prone locationscan live hundreds of years without being severelyimpacted by disturbance (e.g. Brown et al. 2009).Because disturbances result in the rapid liberation ofspace for the establishment and growth of benthicalgae, herbivores may be more important on reefsfrequently subjected to intense disturbances (e.g.Adam et al. 2011). In addition, some herbivores maybe more capable of responding rapidly to increasesin algae and therefore may be particularly importantfor preventing persistent phase shifts on disturbance-prone reefs (see Questions 2 and 3).

One of the most important consequences of globalclimate change (GCC) for herbivore−algae−coral in -teractions will be an increase in the frequency, sever-ity, and spatial scale of disturbances caused by massbleaching events (Hoegh-Guldberg et al. 2007). Dur-ing the past 2 decades, reefs throughout the Carib-bean basin have experienced warm water eventsthat are unprecedented in at least the last 150 yr,with major bleaching episodes in 1998 and again in2005 (Eakin et al. 2010). Both events caused severecoral decline on the most heavily impacted reefs (e.g.Aronson et al. 2002, Whelan et al. 2007, Miller et al.2009). As global CO2 emissions continue to increase,coral bleaching is expected to become an even greaterthreat to coral reefs in the future (Donner et al. 2007,Hoegh-Guldberg et al. 2007). Increases in the fre-quency and severity of bleaching will also interact

with chronic environmental drivers such as oceanacidification that will slow coral growth (Hofmann etal. 2010) and potentially increase the competitiveabilities of algae relative to corals (Diaz-Pulido etal. 2011).

While herbivores may be able to offset the impactsof GCC to some degree, the resilience of reefs willdepend in part on the availability of coral larvae torecolonize reefs following major disturbance events.Reefs which are less prone to disturbance—such asthose occurring in deep water or in upwelling zoneswhere cool water can prevent coral bleaching—willlikely be important refugia for some coral speciesand may be important sources of larvae for recoloniz-ing impacted reefs following large bleaching events(Bongaerts et al. 2010, Bridge et al. 2013, Chollett &Mumby 2013). Projected increases in disturbancedue to GCC highlight the need to manage reefs forresilience by maintaining healthy herbivore popula-tions. However, these management actions will onlybe effective if we can identify and conserve reefs thatcan act as a source of coral recruits following cata-strophic disturbances. Herbivory will matter little ifthere are too few corals present to produce larvae tocolonize disturbed reefs. Thus, it will be important toidentify reefs that are more/less prone to disturbancein order to appropriately focus management efforts.

How does water quality mediate theherbivore−algae−coral interaction?

Coastal pollution has likely been an importantdriver of coral decline throughout much of the Carib-bean. Poor land-use practices associated with agri-culture and urban development increase the deliveryof sediments, nutrients, and a wide range of contam-inants to near-shore ecosystems (Fabricius 2005). Inaddition, large amounts of nutrients, organic contam-inants, and harmful microbes are introduced to coralreefs directly through the disposal of sewage (e.g.Lipp et al. 2002, Lapointe et al. 2005). Collectively,these factors reduce water quality and adverselyaffect reef-building corals via both direct and indi-rect pathways (Fabricius 2005). Increases in nutrientscan have direct negative effects on corals by decreas-ing their growth rates (Shantz & Burkepile 2014) andincreasing the prevalence of coral disease andbleaching (Vega Thurber et al. 2014), as well as indi-rect effects by increasing growth rates of their algalcompetitors (Jompa & McCook 2002). Thus, like manystressors, impacts of nutrient pollution on corals arecomplex and operate through a variety of different

5

Mar Ecol Prog Ser 520: 1–20, 2015

pathways. It is likely that herbivores can offset someof these negative impacts by consuming excess algalproduction, but they may have limited effects on others, such as those leading to coral bleaching anddisease.

In addition, multiple stressors often interact in com-plex non-linear ways such that the impact of any sin-gle stressor is contingent on the presence of others.For example, sedimentation is a major stressor oncoral reefs that can lead to coral decline by smother-ing and killing adult corals and can impede coralrecovery by preventing the successful recruitment ofcoral larvae (Rogers 1990). By trapping sediments,algal turfs can interact with sedimentation to in -crease coral mortality, reduce coral growth, andimpede coral settlement (Nugues & Roberts 2003,Birrell et al. 2005, Gowan et al. 2014). Fishing couldtherefore act synergistically with sedimentation tosuppress corals by reducing herbivory and allowingthe development of mats of sediment-trapping algalturfs. However, once sedimentation levels becomesufficiently high, sediments will begin to smotherand kill corals regardless of the presence of algae(e.g. Nugues & Roberts 2003), negating any positiveimpacts of herbivores.

The previous example illustrates that managementactions targeted at increasing levels of herbivory arelikely to benefit reefs greatly over some ranges ofstressors but relatively little or not at all once thosestressors reach some upper threshold level. A bettermechanistic understanding of how corals and theirsymbionts respond to multiple stressors, includingnutrients, sediments, increases in temperature, andcompetition with algae, will help identify the factorsthat limit coral populations. By identifying the factorsmost strongly limiting corals, managers will be ableto more efficiently allocate limited resources to pro-mote coral persistence and recovery by targeting thespecific locations that are most likely to benefit fromparticular management interventions, such as pro-tection of herbivorous fishes.

How do the effects of herbivores on coral and algaevary with productivity?

The standing stock of algae on a coral reef repre-sents a dynamic balance between algal productionand mortality, with more grazing needed to controlalgae on highly productive reefs. Productivity ofalgae on many shallow coral reefs is likely limited bythe availability of nutrients (nitrogen, phosphorus,and micronutrients such as iron; McCook 1999).

Large-scale differences in nutrient availability occurwithin and among ocean basins (Kleypas et al. 1999),and these differences may influence the level of her-bivory necessary to control algae across geographicregions. Indeed, high levels of algal production onCaribbean reefs may make these reefs more suscep-tible to coral−algal phase shifts compared to manyPacific reefs (Roff & Mumby 2012).

While regional differences in productivity areimportant for understanding the capacity of herbi-vores to control algae, productivity also varies greatlyon much smaller spatial scales. For example, turbu-lent reef zones (e.g. shallow exposed forereefs) aregenerally more productive than calmer reef zones(e.g. lagoonal backreefs and fringing reefs; Falter etal. 2004), and shallow, well-lit reefs are more pro -ductive than deeper reefs (Carpenter 1985). Reefs inmore productive environments may be less resilientto disturbances due to their capacity to support rapidalgal growth. However, herbivorous fishes stronglytrack spatial variation in productivity across the reeflandscape, such that grazing intensity is often high-est in the most productive habitats (Van den Hoek etal. 1978, Russ 2003, Fox & Bellwood 2007). Indeed, onCaribbean reefs, macroalgae are often least abun-dant in shallow locations and more abundant indeeper, less productive habitats where grazing pres-sure is lower (Hay et al. 1983, Morrison 1988, Brugge -mann et al. 1994b). In addition, because corals canalso be limited by light and flow, shallow reefs oftensupport the highest capacity for coral growth andreef accretion (Schlager 1981). Thus, managementefforts to maintain herbivory may have the biggestimpacts on shallow reefs where the capacity for bothalgae and coral growth are often the highest.

How does the outcome of herbivore−algae−coralinteractions vary among different reef zones and

habitats?

Prior to recent collapse of coral populations, coralreefs throughout the Caribbean exhibited distinctzonation patterns, with the major framework-build-ing corals found in characteristic reef zones deter-mined by physical factors such as exposure, depth,and distance from shore (Jackson 1991). In general,the coarse-branching coral Acropora palmata domi-nated reef crests and shallow spur and groove reefs,the fine-branching A. cervicornis dominated mid-depth forereefs, and massive Montastraea spp. dom-inated deeper forereefs, with both A. cervicornisand Montastraea forming lagoonal patch reefs (e.g.

6

Adam et al.: Herbivory and Caribbean coral reefs

Goreau 1959, Adey & Burke 1976, Geister 1977).While these reef zones appear to be highly determin-istic, physical factors such as depth and exposure caninteract in complex ways to modify patterns fromlocation to location (Geister 1977). Therefore, site-specific historical information on the distribution ofdifferent coral species will be invaluable for managerswhen determining places for targeted restoration.

A large fraction of hard-bottom habitat throughoutthe Caribbean is relatively flat limestone pavementdominated by gorgonians, sponges, and/or calcifyingand non-calcifying algae. While these habitats areoften not suitable for the growth of major framework-building corals, they can support abundant anddiverse fish communities and contribute significantlyto fisheries production (Wolff & Grober-Dunsmore1999, Garrison et al. 2004). Integrating informationon the physical factors influencing the distribution ofcorals with knowledge of the ecological impacts ofherbivores could create opportunities for spatialmanagement strategies that can limit trade-offs be -tween different management goals (such as sustain-ing a fishery while also maintaining high grazingrates to control algae and facilitate corals). For exam-ple, fisheries managers may be able to limit the takeof herbivorous fishes on structurally complex reefswhile allowing fishing of herbivores in the low-reliefhabitats which appear to be largely unsuitable formajor framework-building corals (Mumby 2015). Inte -grating knowledge about the current and historicaldistributions of major framework-building coralscould allow for targeted management actions aimedat facilitating coral recovery (including augmentingherbivory) in the locations where corals are mostlylikely to benefit.

Question 2. What level of herbivory is needed tocontrol algae and how important is herbivore

diversity and identity?

Most experiments testing the impacts of herbivoreson benthic communities on coral reefs have usedexclusion cages to compare treatments with andwithout ambient levels of herbivory (e.g. Lewis 1986,Miller et al. 1999, Hughes et al. 2007). These experi-ments have demonstrated that algae are controlledby top-down forces (e.g. Burkepile & Hay 2006), butthey do little to determine the level of herbivoryneeded to control algae or to differentiate impacts ofdifferent types of herbivores. The fact that algae areabundant on many Caribbean reefs suggests thatcurrent levels of herbivory are too low to adequately

control algae, and exclusion experiments do little toinform us how much more herbivory is required.

Determining the amount of herbivory needed tosuppress algae is further complicated by the species-specific impacts that herbivores have on algal com-munities. Early studies investigating how differentherbivores affect Caribbean reefs focused on fishesversus D. antillarum (Carpenter 1986, Morrison 1988).These studies found that D. antillarum can suppressa wide variety of filamentous turf algae and macro-algae, many of which are apparently unpalatable tofishes. However, recent studies indicate that thereare also large functional differences among herbivo-rous fishes (Burkepile & Hay 2008, 2010, 2011). Usingexperimental enclosures to partition the effects ofdifferent fishes, Burkepile & Hay (2008, 2010) foundthat different species of fishes had unique and complementary impacts on algal communities, withthe relative importance of each depending on the successional state of the community. The parrotfishSparisoma aurofrenatum controlled macroalgae incommunities with high existing cover of macroalgae,while the parrotfish Scarus taeniopterus and sur-geonfish Acanthurus bahianus suppressed filamen-tous algal turfs and prevented the establishment ofmacroalgae in early successional stage communities.Because upright macroalgae and filamentous turfsboth have negative effects on corals, no single spe-cies was capable of preventing the spread of harmfulalgae and facilitating coral recruitment and growth.Thus, herbivore diversity appears to be critical formaintaining ecosystem function. Understanding theimportance of herbivore diversity and identity incomplex species-rich coral reef ecosystems requiresgreater knowledge of patterns of complementarityfor the entire herbivore guild. Herbivores that feedon different types of algae are likely to be comple-mentary because they will control algae that candominate in different places or at different pointsin time. In contrast, herbivores with high overlap indiet are more likely to have similar impacts on algalcommunities and may therefore be functionallyredundant.

Complementarity among herbivores is importantfor suppressing algae and facilitating corals, butfunctional redundancy may also confer resilience toreefs by making them less susceptible to the loss of asingle species (Walker 1992). For example, when aspecies declines due to overfishing or disease, spe-cies with similar functions may be able to fill thevacated role in the herbivore guild. Thus, responsediversity, when functionally redundant species re -spond differently to perturbations, is important for

7

Mar Ecol Prog Ser 520: 1–20, 2015

the resilience of an ecosystem (Elmqvist et al. 2003).Therefore, it is critical to determine the extent thatherbivores with similar functions vary in their re -sponse to different types of perturbations.

Managing herbivores for their impacts on coralreef ecosystems requires knowledge of the specificroles played by particular species, and ultimately thelevels of complementarity and response diversitypresent within the herbivore guild. Levels of comple-mentarity will be a function of (1) the amount, types,and life stages of algae that different herbivores eat,(2) the habitats they select while foraging, and (3)the particular substrates they feed from within thosehabitats. Consequently, understanding the level offunctional redundancy and complementarity presentwithin the herbivore guild will require detailed infor-mation on the diet, foraging behavior, and grazingrates of individual species.

How are herbivore traits distributed across taxa?

Diet selection and feeding mode. Herbivorousfishes can be classified into functional groups accord-ing to their diet and feeding mode (Bellwood & Choat1990, Choat et al. 2004). These functional groups canhelp identify patterns of complementarity and redun-dancy within the herbivore guild in order to predictthe unique impacts of different herbivore species oncoral reef ecosystems (Bellwood et al. 2004, Bonaldoet al. 2014). Unfortunately, a functional group frame-work has not been well-developed for herbivores onCaribbean reefs. While Caribbean reefs have lowerherbivore diversity than many Pacific reefs, they stillharbor over 3 dozen species of fish from 10 familiesand several species of sea urchins that feed primarilyon algae and/or sea grass (Randall 1967).

Quantitative data on the diets of most Caribbeanherbivores are rare, but we can make some generali -zations. Parrotfishes (Scaridae), surgeonfishes (Acan -thu ridae), and the sea urchin D. antillarum are likelythe most important herbivores in the region based ontheir abundance, size, and feeding behavior. WhileD. antillarum has a very wide diet breadth (Ogden1976), most herbivorous fishes are more selective,with specific feeding preferences likely driven by avariety of algal traits (e.g. chemical and physicaldefenses and nutritional content; Hay 1997). Parrot-fishes in the genus Sparisoma tend to browse macro-algae, while those in the genus Scarus primarily tar-get turf algae (Randall 1967, Lewis 1985, McAfee &Morgan 1996, Burkepile & Hay 2010). In contrast, the3 species of surgeonfishes, Acanthurus bahianus, A.

chirurgus, and A. coeruleus, appear to target a com-bination of turf algae, detritus, and macroalgae (Ran-dall 1967, Lewis 1985, Burkepile & Hay 2011). Theseobservations suggest some level of complementarityand redundancy within and between the majorgroups of herbivores on Caribbean reefs, but moredetailed data on the specific types of algae targetedby these herbivores are needed.

In addition to diet preferences, we also need infor-mation about grazing rates of individual species.Grazing capacity of herbivores can be estimatedbased on bite rate data combined with algal yield perbite, or extrapolated from models that scale the meta-bolic requirements of herbivorous fishes to their bodysize (e.g. van Rooij et al. 1998, Mumby 2006, Paddacket al. 2006). The relationship between fish mass andgrazing capacity is likely a non-linear, deceleratingfunction. The result is that biomass-specific grazingrates may in fact be greater for smaller fish, whichagrees with both metabolic theory (West et al. 1997)and field data on the grazing capacity of Sparisomaviride (van Rooij et al. 1998). However, there are spe-cific functions, such as bioerosion and the removal ofsome types of macroalgae, that are only achieved bylarge fishes (Bellwood et al. 2012). Thus, altering thesize structure of herbivore populations, without alter-ing overall biomass of fishes, may fundamentallyalter the grazing capacity of the herbivore guild.Future research should address how different met-rics of grazing capacity (e.g. consumption of differenttypes of algae, bioerosion, areal grazing rates, etc.)change as a function of the size structure and speciescomposition of the herbivore guild. These data willbe critical for understanding how fishing impacts thegrazing capacity of the herbivore guild via changesin the size structure and abundance of fish assem-blages.

In addition to exhibiting preferences for particulartypes of algae, herbivores use a variety of feedingmodes that will influence benthic communities differ-ently. For example, many surgeonfish feed by crop-ping algal filaments, while many parrotfish and seaurchins scrape or excavate the substrate (Bellwood etal. 2004). Parrotfishes and sea urchins often removepart of the calcium carbonate reef framework inaddition to epilithic algae, thereby also removing allalgal tissue and inhibiting algal regrowth. Conse-quently, parrotfish and sea urchins often open upnew space that can be colonized by crustose corallinealgae and/or coral. However, bioerosion by parrot-fishes and sea urchins also impacts the long-termdynamics of coral and fish habitat by influencing thebalance between reef accretion and erosion (reviewed

8

Adam et al.: Herbivory and Caribbean coral reefs

by Glynn 1997). Bioerosion by sea urchins and par-rotfishes often accounts for a major fraction of the cal-cium carbonate budget on a coral reef, and high ratesof erosion combined with moderate or low levels ofaccretion (like those currently experienced on manyCaribbean reefs) can result in the net erosion of thereef structural framework with cascading indirecteffects on corals and fishes that depend on the archi-tectural complexity of reef habitats (Perry et al. 2013).In addition, bioerosion by fish and sea urchins couldbe magnified by ongoing local and global anthro-pogenic drivers, such as nutrient enrichment andocean acidification that will weaken the reef frame-work (Hallock 1988, Manzello et al. 2008, Wisshak etal. 2012).

S. viride is likely the most important bioerodingfish on Caribbean reefs. S. viride is functionallyunique among the Sparisoma parrotfishes, targetingendolithic algae by using its strong jaws and hardteeth to excavate the substrate (Bellwood 1994,Bruggemann et al. 1996, Streelman et al. 2002).Direct estimates of bioerosion by S. viride indicatethat where abundant, they can erode several kg ofcarbonate material m−2 yr−1 (Bruggemann et al.1996), an amount that is similar to that removed byvery high densities of the sea urchin D. antillarum(Ogden 1977). While S. viride removes much morecarbonate material than other similar sized parrot-fishes such as Scarus vetula, the amount of carbonatematerial removed by an individual parrotfish in -creases with body size. Thus the 2 largest parrot-fishes in the Caribbean, S. guacamaia and S. coelesti-nus, are likely important bioeroders on the reefswhere they occur. However, no published estimatesof erosion by these fishes exist, and little is knownabout their basic feeding preferences, foraging habits,and historical abundance.

Many scraping and excavating parrotfishes, andsome sea urchins, also feed on benthic invertebratesincluding coral (Bak & Eys 1975, Sammarco 1980,Miller & Hay 1998, Burkepile 2012). Sea urchins andparrotfishes can both kill small coral recruits, andsome parrotfishes, such as S. viride, can kill largecoral colonies (e.g. Bruckner & Bruckner 1998, Rotjan& Lewis 2005). Other Caribbean parrotfishes, in -cluding Scarus vetula, S. guacamaia, and Sparisomaaurofrenatum also feed on live coral (Rotjan & Lewis2008). No studies to date have quantified the relativecontribution of different species to the total level ofcorallivory occurring on Caribbean reefs. Yet recentstudies indicate that total predation intensity by parrotfishes on corals could increase as coral densitydeclines (Burkepile 2012), potentially creating a

feedback that could prevent coral recovery (Glynn1985, Knowlton et al. 1990, Jayewardene et al. 2009).Therefore, understanding the direct negative im -pacts of parrotfish on corals is a major priority, espe-cially on reefs that already have low coral cover.

Spatial patterns of herbivory. The scale over whichherbivores move and the habitats they select whileforaging will influence how their grazing impacts aredistributed in space. Herbivores that forage over dif-ferent spatial scales can have fundamentally differ-ent impacts on benthic communities. For example,recent modeling work suggests that the high inten-sity, spatially constrained foraging behavior of thesea urchin D. antillarum is more likely to facilitatecoral recruitment and coral dominance than the morespatially diffuse herbivory of large, roving fishes(Sandin & McNamara 2012). However, herbivorousfishes also vary greatly in how they use space(e.g. Nash et al. 2013), and we know little about theforaging scales of most herbivorous fishes in the Caribbean.

Herbivorous fishes also feed in different habitatsand target different substrates (Hay 1985, Robertson& Gaines 1986, Bellwood & Choat 1990, Bruggemannet al. 1994a, Brandl & Bellwood 2014). Because habi-tat quality for corals varies greatly in space, these for-aging choices will influence how herbivores impactcoral recruitment and growth. Understanding howherbivores affect coral persistence and recovery willtherefore require knowing precisely where differentspecies of herbivores are exerting their impacts inaddition to knowing the types of algae they are eating.

Are some grazers more important than others?

For years, coral reef scientists have debatedwhether fishes or sea urchins are more important forcontrolling algae and facilitating corals on Caribbeanreefs. This question is complex and the answers havebeen shaped by where research has been done.For example, much of the work on the impacts ofD. antillarum on algae and corals was conducted onreefs with sustained heavy fishing pressure (e.g.Jamaica: Sammarco 1982, Morrison 1988; Virgin Is -lands: Carpenter 1986). Hay (1984) suggested thatthe im portance of D. antillarum may have been over-estimated by studies in places with few herbivorousfishes or D. antillarum predators. He showed thatmost herbivory was from fishes in areas with minimalfishing pressure while urchins were most importantin areas with high fishing (see also Jackson et al.2014). Historically, reefs in the Caribbean may have

9

Mar Ecol Prog Ser 520: 1–20, 2015

had relatively more herbivory from fishes and lessherbivory from urchins than was documented shortlybefore the D. antillarum die-off in the early 1980s.This suggests that D. antillarum may have been themost important herbivore in many parts of the Carib-bean, not because fish were incapable of performingthe same ecological function, but rather because D.antillarum provided a level of response diversity tofishing (i.e. fishing had indirect positive effects on D.antillarum, causing it to increase in abundance whileother herbivores were fished out).

This is not to say that D. antillarum are unimportant.Coral recovery has coincided with recovery of D. an-tillarum in recent years, strongly supporting the ideathat D. antillarum are important for coral re covery inthe Caribbean (Edmunds & Carpenter 2001, Carpenter& Edmunds 2006, Myhre & Acevedo-Gutiérrez 2007,Furman & Heck 2009, Idjadi et al. 2010). D. antillarumhave several traits that may make them particularlyeffective herbivores, in cluding having a wide dietbreadth (Ogden 1976), a scraping/excavating feedingmode (Ogden 1977), and foraging behavior that re-sults in intensive grazing over a very small spatialscale (Carpenter 1984). Thus, the functional impact ofmany different species of fishes was likely approxi-mated well by 1 species of urchin, at least for a limitedamount of time. Nonetheless, the catastrophic impactof the D. antillarum die-off on many Caribbean reefshighlights the danger of relying on a single species toperform a critical ecological function.

Given that D. antillarum have not recovered fol-lowing their mass mortality in most places across theCaribbean, it is critical to identify which combina-tions of herbivorous fishes can facilitate coral persist-ence and recovery in its absence. It is likely that a dif-ferent suite of herbivores is needed to facilitate coralrecovery on degraded reefs compared to those thatcan maintain the resilience of intact reefs. Speciesthat feed on adult macroalgae are likely to be partic-ularly important for reversing phase shifts on macro-algae-dominated reefs. These include macroalgae-browsing parrotfishes in the genus Sparisoma, aswell as other browsing herbivores such as chubs(Kyphosidae). Knowledge of the diets and feedingbehaviors of different herbivores is an important firststep in predicting their impacts on coral persistenceand recovery. However, predictions of ecologicalfunction based on behavioral and morphologicaltraits will be most powerful when combined withmechanistic models (e.g. Sandin & McNamara 2012)and field experiments (e.g. Burkepile & Hay 2008,2010) that can better determine the combinedimpacts of multiple herbivore species.

Question 3. What limits the process of herbivory oncoral reefs?

Many of the feedbacks that could prevent coralrecovery in the Caribbean and elsewhere operatethrough their impacts on herbivory (Fig. 1). There-fore, understanding what limits herbivory is essentialfor predicting how these ecosystems will respondto acute and chronic perturbations, including coralbleaching events, anthropogenic nutrient loading,and fishing. Feedbacks impacting herbivore− algae−coral interactions are realized over a range of spatialand temporal scales. While some feedbacks originatefrom changes in the behavior of individual herbi-vores, others occur due to longer-term impacts onpopulation dynamics that operate over large spatialscales. Consequently, understanding what limits theprocess of herbivory on coral reefs will require a syn-thetic approach that integrates information on indi-vidual behavior with population-level data. Herbi-vore populations will be limited by a combination oftop-down and bottom-up processes, with predatorsand resources also impacting individual behavior.

To what extent is herbivory limited by fishing,natural predation, and disease?

Top-down impacts of human exploitation oncoastal marine ecosystems are pervasive (Jackson etal. 2001). Humans have been exploiting near-shoreenvironments of the Caribbean for thousands ofyears, and hunting of megaherbivores (sea turtlesand manatees) during the past several centuries hasreduced their abundance by orders of magnitude(Jackson 1997). Archeological records from sitesacross the Caribbean indicate that herbivorous reeffishes, especially parrotfishes, have been heavilyfished in some locations for thousands of years(reviewed by Fitzpatrick & Keegan 2007). Compar-isons of fish fauna of inhabited versus uninhabitedislands in the Indo-Pacific demonstrate that even lowlevels of exploitation can rapidly deplete stocks oflarge fishes, including herbivorous parrotfishes (e.g.Friedlander & DeMartini 2002, Bellwood et al. 2012).Given the high levels of current and historical fishingthroughout much of the Caribbean and the obser -vation that even low levels of fishing can limit thepopulations of larger, slower-growing species, fish-ing is almost certainly a major factor limiting theabundance of large parrotfishes in many locations.Indeed, the largest species in the Caribbean, Scarusguacamaia, S. coelestinus, and S. coeruleus, are rare

10

Adam et al.: Herbivory and Caribbean coral reefs

or absent on most reefs (Hawkins & Roberts 2003,Kramer 2003) and are most abundant in places withlittle or no fishing pressure (Debrot et al. 2008, Com-eros-Raynal et al. 2012). In addition, biomass of inter-mediate-sized parrotfishes, such as Sparosoma virideand Scarus vetula are negatively correlated withfishing pressure across much of the region (Hawkins& Roberts 2003).

While some herbivore species are likely limited byfishing throughout large parts of the Caribbean, it isless clear how their loss impacts the total level of her-bivory on these reefs. For example, smaller-bodiedspecies may increase in abundance as they are re -leased from competition or predation, potentiallycompensating for the loss of larger species (Hay1984, Dulvy et al. 2004, Clua & Legendre 2008).Nonetheless, different herbivore species and sizeclasses have distinct diet preferences and feedingmodes, and thus they are unlikely to perform thesame ecological functions (see above). Therefore, it iscritical to identify which herbivore traits are uniqueamongst species most susceptible to fishing, and howlack of redundancy in these traits will impact ecosys-tem function. Finally, while low to moderate levels offishing can lead to the functional elimination of manylarger-bodied herbivore species, it is less clear whatlevels of fishing are sustainable for most other species.

In addition to being heavily exploited by humans,herbivorous fishes and sea urchins are preyed uponby a wide variety of predatory fishes (Randall 1967).Predators can have large impacts on the structureand function of natural ecosystems via their con-sumptive effects on prey populations and cascadingindirect effects on the resources of their prey (e.g.trophic cascades). However, predators also stronglyimpact prey traits (e.g. behavior, growth, and repro-duction), and recent research indicates that thesenon-consumptive effects (NCEs) frequently have far-reaching ecosystem-level consequences (Dill et al.2003, Peckarsky et al. 2008, Schmitz 2008).

NCEs arise due to inherent trade-offs between theneed to eat while avoiding being eaten. As a result ofthese trade-offs, prey often allocate less time to for-aging or shift activity to safer but less profitable habi-tats when predation risk is high (Lima & Dill 1990).NCEs may be of considerable importance in coralreef habitats, where high structural complexity offersmany refugia, potentially increasing the ability ofprey to respond behaviorally to their predators. Inboth the Caribbean and Pacific, herbivorous seaurchins and fishes often concentrate their grazingaround the edges of large patch reefs where they canfind shelter from their predators. Grazing pressure

around the perimeter of these reefs can be so intensethat it creates distinct ‘grazing halos’ easily seen inaerial photographs and satellite images (Randall1965, Ogden et al. 1973, McManus et al. 2000, Madinet al. 2011).

Recent research on reefs in the central Pacific sug-gests that enhanced predation risk may also causeherbivorous fishes to restrict their foraging ranges oncontinuous reefs, thereby affecting the spatial distri-bution of algae and potentially modifying coral−algalcompetition (Madin et al. 2010). This work suggeststhe possibility that predators could create heteroge-neous landscapes where hotspots of grazing activitycreate habitat suitable for coral recruitment. Thus, incontrast to predictions from classic trophic cascades,where predators of herbivores would be expected tohave negative effects on corals by releasing algaefrom top-down control, incorporating impacts ofpredators on herbivore behavior suggests that pre -dators could have indirect positive effects on coralsthrough their influence on the spatial patterning ofherbivory. It is only recently that coral reef scientistshave started to link behavioral changes in prey topredation risk (e.g. Catano et al. 2014), yet under-standing how predators impact habitat use, foragingrate, and the spatial scale of foraging of different herbivore species will be important for predicting theimpacts of fishing on ecosystem function on coralreefs.

In addition to being preyed upon whole, many reeffishes are preyed upon by parasites, including thosethat cause infectious disease. Parasites are difficult tosample, and their impacts on host populations aredifficult to quantify; thus, the role of parasites in lim-iting many animal populations is largely unknown(Tompkins et al. 2011). Many of the parasites thataffect reef fishes are likely to have important impactson their population dynamics and behavior. Forexample, parrotfishes in both the Indo-Pacific andCaribbean spend a significant fraction of their dailytime budget visiting ‘cleaner stations’, where theyhave their ectoparasites removed by specializedcleaner fish and shrimps (Soares et al. 2013). Thus,the distribution of cleaners may create hotspots ofactivity for herbivorous and corallivorous fisheswhich in turn may indirectly affect coral and algae(Adam 2012).

Microbes that cause infectious disease are oftendifficult to identify, yet they can have overwhelminglylarge impacts on the functioning of entire ecosystems.The die-off of D. antillarum and the cascading in -direct effects on Caribbean reefs was apparently a re-sult of infectious disease that spread quickly through-

11

Mar Ecol Prog Ser 520: 1–20, 2015

out the entire Caribbean basin (Lessios 1988). Whilethe causative agent in the D. antillarum epidemic hasnever been identified, the outbreak may have beenexacerbated by unusually high densities of the urchindue to the elimination of its predators and competi-tors. Infectious disease appears to be a common causeof boom and bust dynamics in other echinoderms(Uthicke et al. 2009), suggesting that relying on ur -chins to be the primary herbivores on reefs may setthe stage for repeated, catastrophic losses of herbi-vore populations. Because disease agents are usuallyspecific to particular taxa and transmission is posi-tively density-dependent, other herbivore speciescould also become vulnerable to infectious disease iftheir populations experience large increases in re-sponse to removal of their predators and/or compe -titors. Unfortunately, because we know little about infectious disease in marine organisms, future epi-demics will almost certainly come as a surprise.

Are herbivore populations limited by the availabilityof juvenile habitat?

Like many foundation species, corals provide phys-ical habitat for other organisms. Branching corals inparticular serve as important habitat for juvenilefishes, including herbivores, that shelter in andaround corals to escape from their predators. Manyreef fishes selectively use live branching corals asrecruitment habitat and could become vulnerable tohabitat loss as corals decline (Tolimieri 1998, Jones etal. 2004, Adam et al. 2011). However, fish can alsogain refuge in other common reef substrates, includ-ing macroalgae and sponges, and many fishes arelikely to have some level of plasticity in microhabitatselection. For example, in the Florida Keys, S. viridesuccessfully uses the abundant macroalgae Dictyotaspp. as recruitment substrate in the absence of pre-ferred branching corals (Paddack & Sponaugle 2008).Nonetheless, the capacity of most fishes to recruitto different microhabitats is unknown. In fact, sur-prisingly little is known about recruitment habitatrequirements of several species of Caribbean parrot-fishes (e.g. Scarus coelestinus, S. coeruleus, and S.guacamaia). Clearly, successfully managing herbi-vore populations requires a better understanding oftheir basic habitat requirements.

While some herbivorous fishes are strongly reef-associated throughout their ontogeny, several Carib-bean parrotfishes (e.g. S. iseri, S. guacamaia, S.coeru leus, Sparisoma chrysopterum) use off-reefnursery habitats such as sea grasses and mangroves

before migrating to coral reefs as large juveniles oradults (Nagelkerken et al. 2000, 2002, Mumby et al.2004, Machemer et al. 2012). Thus, it is possible thatthese species are limited by the availability of juve-nile nursery habitat. For example, the largest herbi -vorous fish in the Caribbean, S. guacamaia, appearsto depend on mangroves during its juvenile phase,and mangrove clearing may have led to its localextinction on an isolated atoll in Belize (Mumby et al.2004). The fact that many herbivorous fishes use non-reef habitats as juveniles presents both opportunitiesand challenges for management. While managementefforts often focus on a single habitat, connectivitybetween different habitats should enhance the resili-ence of coral reefs by providing reefs with a source ofherbivores following localized coral decline (Adam etal. 2011). However, we know relatively little aboutthe dependence of particular herbivore species onspecific nursery habitats and even less about thescale of movement that occurs among habitats. Spa-tial planning efforts, including the implementation ofmarine reserves, could benefit tremendously frombetter information on habitat utilization across lifehistory stages and the scale of movement of differentherbivore species.

How do food availability and quality shape herbi-vore abundance and behavior?

Models linking herbivory to the resilience of coralreefs often begin with the assumption that herbivoresare limited by top-down forces, yet herbivores maybe food limited (van Rooij et al. 1998). Indeed, time-series data increasingly suggest that herbivores canbe limited by food, even in locations with moderate tohigh levels of fishing. For example, recent observa-tions of reefs in the Caribbean and Indo-Pacific haverevealed that herbivorous fish can experience largeincreases in abundance and biomass in response tothe increase in turf algae that accompanies a rapiddecline in coral (Hawkins et al. 2006, Adam et al.2011, Gilmour et al. 2013). Similarly, Carpenter (1990)observed large increases in the abundances of juve-nile parrotfish and surgeonfish in the US VirginIslands shortly after the Diadema die-off, suggestingthat these fish had been limited at least in part bycompetition for algae with Diadema.

Counter-intuitively, herbivores may be food-limitedeven on algae-dominated reefs since macroalgae arefrequently physically and chemically defended andcan exclude more palatable filamentous turf algae(e.g. McClanahan et al. 2000). Thus, as reefs become

12

Adam et al.: Herbivory and Caribbean coral reefs

dominated by unpalatable macroalgae, they mayshift to a detrital based food-web with much of themacroalgae going uneaten by herbivorous fishes, po-tentially transferring energy away from reef fishesand other herbivores (Carpenter 1986). Increasingherbivore food limitation would further reinforce analgae-dominated state. This suggests that in order toavoid a persistent phase shift to an algae-dominatedstate, herbivores would need to respond rapidly to increases in algal production caused by the liberationof space for algal growth following a major dis -turbance (Scheffer et al. 2008, Adam et al. 2011,Blackwood & Hastings 2011). While herbivores canrespond behaviorally to in creased production onshort time scales, rates of population change will beslower and will depend on growth rates, generationtimes, and the scale of self-recruitment, all of whichwill vary among species and locations.

When are herbivore populations likely to be limitedby larval supply?

The availability of larvae can determine patterns ofrecruitment of reef fishes. The processes controllinglarval supply of reef fishes are complex, involving acombination of biological processes that determinereproductive output and initial larval quality, andoceanographic processes that shape larval transportand nourishment (Cowen et al. 2006). The continuedfailure of D. antillarum to recover after the 1982− 1983mass mortality event is likely a result of insufficientlarval availability due to an Allee effect limiting re-productive output (Lessios 2005). Recruitment failureof herbivorous fishes has not been reported, but littleis known about the recruitment dynamics of manyspecies, and recruitment on some heavily fished reefsis potentially persisting due to upstream sources oflarvae (Cowen et al. 2006). Given the low abundancesof many species of larger parrotfishes on most Carib-bean reefs, it would not be surprising if these specieswere limited by larval supply in some locations. Toavoid recruitment failure in the future, we need moreinformation on the relationship between local adultdensity and larval connectivity to identify thresholdlevels of adult biomass and vulnerable areas.

SUMMARY

Herbivores control algae that otherwise have neg-ative impacts on the recruitment, growth, and sur-vivorship of reef-building corals. Consequently, we

expect that enhancing herbivore populations shouldincrease the resilience of coral-dominated reefs andfacilitate coral recovery on degraded reefs. However,impacts of herbivores will vary greatly in space andtime, and understanding when and where herbivoreswill be most important for facilitating coral persist-ence and recovery requires a better mechanisticunderstanding of the multiple processes that are currently limiting coral populations. This could bebest achieved by forging stronger linkages betweenconceptual thinking, mathematical modeling, andempirical data.

For example, many of the conceptual argumentsfor strong positive indirect effects of herbivores oncoral have focused on coral recruitment (e.g. Mumby& Steneck 2008, Mumby et al. 2013). Yet, recentfield-parameterized demographic models of 2 of themost important framework-building corals in theCaribbean, Montastrea annularis and Acropora pal -mata, indicate that ecologically relevant increases inrecruitment rates cannot drive population recovery ifthe conditions limiting the growth and survivorshipof adult colonies are not addressed (Edmunds & Elahi2007, Vardi et al. 2012). Factors limiting the growthof adult colonies are not well understood, but likelyinclude interactions amongst a suite of chronic stres-sors, including increased water temperature, declin-ing water quality, competition with algae, and in thecase of A. palmata, chronic predation by coralli -vorous snails (Edmunds & Elahi 2007, Vardi et al.2012, Williams & Miller 2012). While coral recruit-ment is clearly important for the replenishment ofcoral populations, demographic models underscorethe need to better understand how macroalgae andfilamentous turf algae limit the growth and survivor-ship of adult colonies (e.g. Lirman 2001, Vermeij etal. 2010), and especially how algae can interact withthe multiple other stressors that impact coral growthand survivorship. These insights would allow us tobetter predict the extent that increases in herbivorycan benefit corals facing a range of stressors.

In addition to knowing the role that algae play inlimiting coral, we need to understand how diversitywithin the herbivore guild impacts algal assemblages.Mathematical models designed to predict coral andalgae dynamics in response to local and global stres-sors have generally treated herbivores as a singlefunctional group (e.g. Mumby 2006, Hoegh-Guldberget al. 2007, Mumby et al. 2007, Anthony et al. 2011,Fung et al. 2011, Baskett et al. 2014). However,because herbivores performing different functionswill exhibit a diversity of responses to anthropogenicstressors such as fishing (Bellwood et al. 2012,

13

Mar Ecol Prog Ser 520: 1–20, 2015

Edwards et al. 2014), this approach may be too sim-ple. Empirical research should focus on identifyingthe ecological functions performed by different species of herbivores, and determine whether somefunctions are more susceptible to particular pertur-bations than others. Likewise, future modeling workaimed at predicting the impacts of anthropogenicstressors on coral persistence and recovery shouldexplore the extent that complementarity and re -sponse diversity within the herbivore guild altermodel predictions.

Developing effective management strategies toincrease herbivory on Caribbean reefs requires iden-tifying the factors limiting herbivore populations andthe process of herbivory on particular reefs. Herbi-vore populations can be limited by a number of fac-tors, including natural predation, fishing, food, theavailability of juvenile habitat, and larval supply.Better information on larval connectivity and juvenilehabitat associations in particular could facilitatemanagement by helping managers identify and tar-get key locations for protection. In addition, herbi-vores may avoid degraded reefs with high cover ofmacroalgae or low levels of physical structure, creat-ing a reinforcing feedback that keeps herbivory lowon degraded reefs even in locations with healthy her-bivore populations. Thus, a better understanding ofthe behavioral responses of herbivorous fishes tocoral decline, increases in algae, and the loss ofstructural complexity could aid efforts to restore reefsby identifying habitat characteristics that need to berestored to enhance grazing levels on reefs.

MANAGEMENT RECOMMENDATIONS

We currently have an incomplete understandingof the multiple factors limiting corals on Caribbeanreefs. To complicate matters, many of these factorsoperate on regional or global scales managers cannotcontrol. However, there are many actions that localmanagers can take. First, local management effortsshould focus on minimizing direct sources of coralmortality, such as sedimentation and pollution, aswell as restoring ecological processes, such as her-bivory, that are important for coral persistence andrecovery. Since nutrient loading can increase theseverity and prevalence of coral disease and bleach-ing, reducing nutrient inputs may mitigate the nega-tive impacts of ocean warming (e.g. Vega Thurber etal. 2014). Furthermore, reducing nutrient loadingmay also reduce algal productivity, facilitating thecontrol of algae by herbivores.

Maintaining healthy herbivore populations is alsolikely to mitigate the negative impacts of oceanwarming since abundant herbivores can control algaethat inhibit coral recovery following coral decline(Edwards et al. 2011). Parrotfishes are particularlyimportant herbivores on Caribbean reefs, but theyalso frequently support local fisheries. Therefore,complete bans on the harvesting of parrotfishes maynot always be politically viable or economically desir-able. However, better spatial management of fishingcould minimize trade-offs between the need to main-tain high levels of grazing while supporting sustain-able fisheries. For example, managers could restrictfishing for herbivores on the reefs that are most suitable for major framework-building corals, whileallowing fishing in other locations where corals areunlikely to benefit from herbivory (Mumby 2015).Effective spatial management requires detailedknow ledge on the distribution of suitable habitat fordifferent species of coral; at the coarsest level, thisknowledge already exists in many locations in theCaribbean due to recent advances in remote sensingand large-scale efforts to map coral reef habitats (e.g.Pittman et al. 2013).

Implementation of marine protected areas or otherspatial restrictions on herbivore fishing will only beeffective if we can sustainably manage herbivorepopulations outside of protected areas. Different spe-cies of parrotfishes have different life-history traitsand different impacts on benthic communities. There -fore, they should not be managed as a single speciescomplex. In particular, the largest parrotfishes in theCaribbean, Scarus guacamaia, S. coelestinus, and S.coeruleus, are highly susceptible to overexploitationand are not functionally equivalent to smaller spe-cies, and thus should be fully protected from fishing.Further, Caribbean parrotfishes belong to at least 3different functional groups (scrapers that feed pri -marily on turf algae, bioeroders that remove largeamounts of carbonate material, and browsers thatfeed largely on upright macroalgae). Consequently,managers will need to ensure that reefs have theright mix of herbivores to carry out the full set offunctions normally performed by the herbivore guild.Finally, it is critical to protect seagrasses and man-groves, which are important nursery habitats for several species of Caribbean herbivores.

While protection of herbivorous fishes is likely nec-essary for the maintenance of resilient coral reef systems, it is unlikely to be sufficient for the recoveryof reefs, particularly where degradation has beensevere and feedbacks are operating that could slowor prevent coral recovery. Under these conditions,

14

Adam et al.: Herbivory and Caribbean coral reefs

management actions targeted specifically at break-ing feedbacks that maintain reefs in a degraded stateare necessary. Appropriate interventions are likely tobe highly context-dependent. For example, in loca-tions with high levels of fishing, restoration effortsaimed at augmenting densities of D. antillarum couldbe an important tool for increasing local grazingintensity on specific reefs. In contrast, in locationsthat are lightly fished, it may be more appropriate tofocus on attracting herbivorous fishes by manuallyremoving macroalgae, using artificial structures toincrease habitat, or engaging in active restoration ofcorals. Restoration efforts targeted at specific reefswill be most successful under strong management ofthe wider seascape that limits chronic sources ofcoral mortality and protects key functional groupssuch as herbivorous fishes. Many of the feedbacksthat prevent coral recovery may strengthen throughtime, so efforts to restore reefs likely have the highestchance of success (and will be least costly) whenenacted early, before loss of living coral and struc-tural complexity are severe.

Acknowledgements. This work was supported by a grantfrom NOAA’s Coral Reef Conservation Program to D.E.B.and B.I.R. We thank M. W. Miller, J. Schull, and 2 anony-mous reviewers for helpful feedback on an earlier version ofthis manuscript.

LITERATURE CITED

Adam TC (2012) Mutualistic cleaner fish initiate trait-medi-ated indirect interactions by influencing the behaviour ofcoral predators. J Anim Ecol 81:692−700

Adam TC, Schmitt RJ, Holbrook SJ, Brooks AJ, Edmunds PJ,Carpenter RC, Bernardi G (2011) Herbivory, connectiv-ity, and ecosystem resilience: response of a coral reef to alarge-scale perturbation. PLoS ONE 6:e23717

Adam TC, Brooks AJ, Holbrook SJ, Schmitt RJ, Washburn L,Bernardi G (2014) How will coral reef fish communitiesrespond to climate-driven disturbances? Insight fromlandscape-scale perturbations. Oecologia 176:285−296

Adey WH, Burke R (1976) Holocene bioherms (algal ridgesand bank-barrier reefs) of the eastern Caribbean. GeolSoc Am Bull 87:95−109

Adjeroud M, Michonneau F, Edmunds PJ, Chancerelle Yand others (2009) Recurrent disturbances, recovery tra-jectories, and resilience of coral assemblages on a SouthCentral Pacific reef. Coral Reefs 28:775−780

Alvarez-Filip L, Côté IM, Gill JA, Watkinson AR, Dulvy NK(2011) Region-wide temporal and spatial variation inCaribbean reef architecture: Is coral cover the wholestory? Glob Change Biol 17:2470−2477

Anthony KRN, Maynard JA, Diaz-Pulido G, Mumby PJ,Marshall PA, Cao L, Hoegh-Guldberg O (2011) Oceanacidification and warming will lower coral reef resili-ence. Glob Change Biol 17:1798−1808

Aronson RB, Precht WF (1997) Stasis, biological disturbance,

and community structure of a Holocene coral reef. Paleo-biology 23:326−346

Aronson RB, Precht WF (2006) Conservation, precaution,and Caribbean reefs. Coral Reefs 25:441−450

Aronson RB, Precht WF, Toscano M, Koltes K (2002) The1998 bleaching event and its aftermath on a coral reef inBelize. Mar Biol 141:435−447

Bak RPM, Eys G (1975) Predation of the sea urchin Diademaantillarum Philippi on living coral. Oecologia 20:111−115

Baker AC, Glynn PW, Riegl B (2008) Climate change andcoral reef bleaching: an ecological assessment of long-term impacts, recovery trends and future outlook. EstuarCoast Shelf Sci 80:435−471

Barott KL, Rodriguez-Mueller B, Youle M, Marhaver KL,Vermeij MJA, Smith JE, Rohwer FL (2012) Microbial toreef scale interactions between the reef-building coralMontastraea annularis and benthic algae. Proc R SocLond B Biol Sci 279:1655−1664

Baskett ML, Fabina NS, Gross K (2014) Response diversitycan increase ecological resilience to disturbance in coralreefs. Am Nat 184:E16−E31

Bellwood DR (1994) A phylogenetic study of the parrotfishfamily Scaridae (Pisces: Labroidei), with a revision ofgenera. Rec Aust Mus 20 (Suppl):1−86

Bellwood DR, Choat JH (1990) A functional analysis of graz-ing in parrotfishes (family Scaridae): the ecologicalimplications. Environ Biol Fishes 28:189−214

Bellwood DR, Hughes TP, Folke C, Nyström M (2004) Con-fronting the coral reef crisis. Nature 429:827−833

Bellwood DR, Hoey AS, Hughes TP (2012) Human activityselectively impacts the ecosystem roles of parrotfishes oncoral reefs. Proc R Soc Lond B Biol Sci 279:1621−1629

Birrell CL, McCook LJ, Willis BL (2005) Effects of algal turfsand sediment on coral settlement. Mar Pollut Bull 51:408−414

Birrell CL, McCook LJ, Willis BL, Diaz-Pulido GA (2008)Effects of benthic algae on the replenishment of coralsand the implications for the resilience of coral reefs.Oceanogr Mar Biol Annu Rev 46:25−63

Blackwood JC, Hastings A (2011) The effect of time delayson Caribbean coral-algal interactions. J Theor Biol273:37−43

Bonaldo RM, Hoey AS, Bellwood DR (2014) The ecosystemroles of parrotfishes on tropical reefs. Oceanogr Mar BiolAnnu Rev 52:81−132

Bongaerts P, Ridgway T, Sampayo EM, Hoegh-Guldberg O(2010) Assessing the ‘deep reef refugia’ hypothesis: focuson Caribbean reefs. Coral Reefs 29:309−327

Box SJ, Mumby PJ (2007) Effect of macroalgal competitionon growth and survival of juvenile Caribbean corals. MarEcol Prog Ser 342:139−149

Brandl SJ, Bellwood DR (2014) Individual-based analysesreveal limited functional overlap in a coral reef fish com-munity. J Anim Ecol 83:661−670

Bridge TCL, Hughes TP, Guinotte JM, Bongaerts P (2013)Call to protect all coral reefs. Nat Clim Change 3:528−530

Brown CJ, Mumby PJ (2014) Trade-offs between fisheriesand the conservation of ecosystem function are definedby management strategy. Front Ecol Environ 12:324−329

Brown DP, Basch L, Barshis D, Forsman Z, Fenner D, Gold-berg J (2009) American Samoa’s island of giants: massivePorites colonies at Ta’u island. Coral Reefs 28:735

Bruckner A, Bruckner R (1998) Destruction of coral by Spari-soma viride. Coral Reefs 17:350

15

http://dx.doi.org/10.1007/s003380050138http://dx.doi.org/10.1007/s00338-009-0494-8http://dx.doi.org/10.1038/nclimate1879http://dx.doi.org/10.1111/1365-2656.12171http://dx.doi.org/10.3354/meps342139http://dx.doi.org/10.1007/s00338-009-0581-xhttp://dx.doi.org/10.1016/j.jtbi.2010.12.022http://dx.doi.org/10.1016/j.marpolbul.2004.10.022http://dx.doi.org/10.1098/rspb.2011.1906http://dx.doi.org/10.1038/nature02691http://dx.doi.org/10.1007/BF00751035http://dx.doi.org/10.3853/j.0812-7387.20.1994.51http://dx.doi.org/10.1086/676643http://dx.doi.org/10.1098/rspb.2011.2155http://dx.doi.org/10.1016/j.ecss.2008.09.003http://dx.doi.org/10.1007/BF00369023http://dx.doi.org/10.1007/s00227-002-0842-5http://dx.doi.org/10.1007/s00338-006-0122-9http://dx.doi.org/10.1111/j.1365-2486.2010.02364.xhttp://dx.doi.org/10.1111/j.1365-2486.2010.02385.xhttp://dx.doi.org/10.1007/s00338-009-0515-7http://dx.doi.org/10.1130/0016-7606(1976)87%3C95%3AHBARAB%3E2.0.CO%3B2http://dx.doi.org/10.1007/s00442-014-3011-xhttp://dx.doi.org/10.1371/journal.pone.0023717http://dx.doi.org/10.1111/j.1365-2656.2011.01943.x

Mar Ecol Prog Ser 520: 1–20, 2015

Bruggemann JH, Kuyper MWM, Breeman AM (1994a)Comparative analysis of foraging and habitat use by thesympatric Caribbean parrotfish Scarus vetula and Spari-soma viride (Scaridae). Mar Ecol Prog Ser 112:51−66

Bruggemann JH, van Oppen MJH, Breeman AM (1994b)Foraging by the stoplight parrotfish Sparisoma viride. I.Food selection in different, socially determined habitats.Mar Ecol Prog Ser 106:41−55

Bruggemann JH, van Kessel AM, van Rooij JM, BreemanAM (1996) Bioerosion and sediment ingestion by theCaribbean parrotfish Scarus vetula and Sparisomaviride: implications of fish size, feeding mode and habitatuse. Mar Ecol Prog Ser 134:59−71

Bruno JF, Sweatman H, Precht WF (2009) Assessing evi-dence of phase shifts from coral to macroalgal domi-nance on coral reefs. Ecology 90:1478−1484

Burkepile DE (2012) Context-dependent corallivory by parrotfishes in a Caribbean reef ecosystem. Coral Reefs31:111−120

Burkepile DE, Hay ME (2006) Herbivore vs. nutrient controlof marine primary producers: context-dependent effects.Ecology 87:3128−3139

Burkepile DE, Hay ME (2008) Herbivore species richnessand feeding complementarity affect community structureand function on a coral reef. Proc Natl Acad Sci USA105:16201−16206

Burkepile DE, Hay ME (2010) Impact of herbivore identityon algal succession and coral growth on a Caribbeanreef. PLoS ONE 5:e8963

Burkepile DE, Hay ME (2011) Feeding complementarity versus redundancy among herbivorous fishes on a Caribbean reef. Coral Reefs 30:351−362

Burkepile DE, Allgeier JE, Shantz AA, Pritchard CE,Lemoine NP, Bhatti LH, Layman CA (2013) Nutrient supply from fishes facilitates macroalgae and suppressescorals in a Caribbean coral reef ecosystem. Sci Rep 3:1493

Burman SG, Aronson RB, van Woesik R (2012) Biotic homog-enization of coral assemblages along the Florida reeftract. Mar Ecol Prog Ser 467:89−96

Carpenter RC (1984) Predator and population density con-trol of homing behavior in the Caribbean echinoidDiadema antillarum. Mar Biol 82:101−108

Carpenter RC (1985) Relationships between primary pro-duction in coral reef algal communities and irradiance.Limnol Oceanogr 30:784−793

Carpenter RC (1986) Partitioning herbivory and its effects oncoral reef algal communities. Ecol Monogr 56:345−364

Carpenter RC (1990) Mass mortality of Diadema antillarum.II. Effects on population densities and grazing intensityof parrotfishes and surgeonfishes. Mar Biol 104:79−86

Carpenter RC (1997) Invertebrate predators and grazers. In:Birkeland C (ed) Life and death of coral reefs. Chapman& Hall, New York, NY, p 198−226

Carpenter RC, Edmunds PJ (2006) Local and regional scalerecovery of Diadema promotes recruitment of scleractin-ian corals. Ecol Lett 9:271−280

Catano LB, Shantz AA, Burkepile DE (2014) Predation risk,competition, and territorial damselfishes as drivers ofherbivore foraging on Caribbean coral reefs. Mar EcolProg Ser 511:193−207

Choat JH, Robbins WD, Clements KD (2004) The trophic status of herbivorous fishes on coral reefs. II. Food pro-cessing modes and trophodynamics. Mar Biol 145:445−454

Chollett I, Mumby PJ (2013) Reefs of last resort: locating andassessing thermal refugia in the wider Caribbean. BiolConserv 167:179−186

Clua E, Legendre P (2008) Shifting dominance among scaridspecies on reefs representing a gradient of fishing pres-sure. Aquat Living Resour 21:339−348

Comeros-Raynal MT, Choat JH, Polidoro BA, Clements KDand others (2012) The likelihood of extinction of iconicand dominant herbivores and detritivores of coral reefs:the parrotfishes and surgeonfishes. PLoS ONE 7:e39825

Connell JH (1997) Disturbance and recovery of coral assem-blages. Coral Reefs 16:S101−S113

Connell JH, Hughes TP, Wallace CC (1997) A 30-year studyof coral abundance, recruitment, and disturbance at sev-eral scales in space and time. Ecol Monogr 67:461−488

Cowen RK, Paris CB, Srinivasan A (2006) Scaling of connec-tivity in marine populations. Science 311:522−527

Debrot D, Choat JH, Posada JM, Roberstson DR (2008) Highdensities of the large bodied parrotfishes (Scaridae) attwo Venezuelan offshore reefs: comparison among fourlocalities in the Caribbean. Proceedings of the 60th Gulfand Caribbean Fisheries Institute, Punta Cana, Domini-can Republic, p 335−338

Diaz-Pulido G, Gouezo M, Tilbrook B, Dove S, AnthonyKRN (2011) High CO2 enhances the competitive strengthof seaweeds over corals. Ecol Lett 14:156−162

Dill LM, Heithaus MR, Walters CJ (2003) Behaviorally medi-ated indirect interactions in marine communities andtheir conservation implications. Ecology 84:1151−1157

Donner SD, Knutson TR, Oppenheimer M (2007) Model-based assessment of the role of human-induced climatechange in the 2005 Caribbean coral bleaching event.Proc Natl Acad Sci USA 104:5483−5488

Dulvy NK, Polunin NVC, Mill AC, Graham NAJ (2004) Sizestructural change in lightly exploited coral reef fish com-munities: evidence for weak indirect effects. Can J FishAquat Sci 61:466−475

Eakin CM, Morgan JA, Heron SF, Smith TB and others(2010) Caribbean corals in crisis: record thermal stress,bleaching, and mortality in 2005. PLoS ONE 5:e13969

Edmunds PJ, Carpenter RC (2001) Recovery of Diademaantillarum reduces macroalgal cover and increasesabundance of juvenile corals on a Caribbean reef. ProcNatl Acad Sci USA 98:5067−5071

Edmunds PJ, Elahi R (2007) The demographics of a 15-yeardecline in cover of the Caribbean reef coral Montastraeaannularis. Ecol Monogr 77:3−18

Edwards HJ, Elliott IA, Eakin CM, Irikawa A and others(2011) How much time can herbivore protection buy forcoral reefs under realistic regimes of hurricanes andcoral bleaching? Global Change Biol 17:2033–2048

Edwards CB, Friedlander AM, Green AG, Hardt MJ andothers (2014) Global assessment of the status of coral reefherbivorous fishes: evidence for fishing effects. Proc RSoc Lond B Biol Sci 281:20131835

Elmqvist T, Folke C, Nyström M, Peterson G, Bengtsson J,Walker B, Norberg J (2003) Response diversity, eco -system change, and resilience. Front Ecol Environ 1:488−494

Fabricius KE (2005) Effects of terrestrial runoff on the eco -logy of corals and coral reefs: review and synthesis. MarPollut Bull 50:125−146

Falter JL, Atkinson MJ, Merrifield MA (2004) Mass-transferlimitation of nutrient uptake by a wave-dominated reefflat community. Limnol Oceanogr 49:1820−1831

16

http://dx.doi.org/10.4319/lo.2004.49.5.1820http://dx.doi.org/10.1016/j.marpolbul.2004.11.028http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21764740&dopt=Abstracthttp://dx.doi.org/10.1890/1540-9295(2003)001[0488%3ARDECAR]2.0.CO%3B2http://dx.doi.org/10.1098/rspb.2013.1835http://dx.doi.org/10.1111/j.1365-2486.2010.02366.xhttp://dx.doi.org/10.1890/05-1081http://dx.doi.org/10.1073/pnas.071524598http://dx.doi.org/10.1371/journal.pone.0013969http://dx.doi.org/10.1139/f03-169http://dx.doi.org/10.1073/pnas.0610122104http://dx.doi.org/10.1890/0012-9658(2003)084[1151%3ABMIIIM]2.0.CO%3B2http://dx.doi.org/10.1111/j.1461-0248.2010.01565.xhttp://dx.doi.org/10.1126/science.1122039http://dx.doi.org/10.1890/0012-9615(1997)067[0461%3AAYSOCA]2.0.CO%3B2http://dx.doi.org/10.1007/s003380050246http://dx.doi.org/10.1371/journal.pone.0039825http://dx.doi.org/10.1051/alr%3A2008036http://dx.doi.org/10.1016/j.biocon.2013.08.010http://dx.doi.org/10.1007/s00227-004-1341-7http://dx.doi.org/10.3354/meps10921http://dx.doi.org/10.1111/j.1461-0248.2005.00866.xhttp://dx.doi.org/10.1007/BF01313160http://dx.doi.org/10.2307/1942551http://dx.doi.org/10.4319/lo.1985.30.4.0784http://dx.doi.org/10.1007/BF00392768http://dx.doi.org/10.3354/meps09950http://dx.doi.org/10.1038/srep01493http://dx.doi.org/10.1007/s00338-011-0726-6http://dx.doi.org/10.1371/journal.pone.0008963http://dx.doi.org/10.1073/pnas.0801946105http://dx.doi.org/10.1890/0012-9658(2006)87[3128%3AHVNCOM]2.0.CO%3B2http://dx.doi.org/10.1007/s00338-011-0824-5http://dx.doi.org/10.1890/08-1781.1http://dx.doi.org/10.3354/meps134059http://dx.doi.org/10.3354/meps106041http://dx.doi.org/10.3354/meps112051

Adam et al.: Herbivory and Caribbean coral reefs

Fitzpatrick SM, Keegan WF (2007) Human impacts andadaptations in the Caribbean Islands: an historical eco -logy approach. Earth Environ Sci Trans R Soc Edinb98:29−45

Fox RJ, Bellwood DR (2007) Quantifying herbivory across acoral reef depth gradient. Mar Ecol Prog Ser 339:49−59

Friedlander AM, DeMartini EE (2002) Contrasts in density,size, and biomass of reef fishes between the northwest-ern and the main Hawaiian islands: the effects of fishingdown apex predators. Mar Ecol Prog Ser 230:253−264

Fung T, Seymour RM, Johnson CR (2011) Alternative stablestates and phase shifts in coral reefs under anthro-pogenic stress. Ecology 92:967−982

Furman B, Heck KL (2009) Differential impacts of echinoidgrazers on coral recruitment. Bull Mar Sci 85:121−132

Gardner TA, Côté IM, Gill JA, Grant A, Watkinson AR (2003)Long-term region-wide declines in Caribbean corals.Science 301:958−960

Gardner TA, Côté IM, Gill JA, Grant A, Watkinson AR(2005) Hurricanes and Caribbean coral reefs: impacts,recovery patterns, and role in long-term decline. Ecology86:174−184

Garrison V, Rogers C, Beets J, Friedlander A (2004) Thehabitats exploited and the species trapped in a Carib-bean island trap fishery. Environ Biol Fishes 71:247−260

Geister J (1977) The influence of wave exposure on the eco-logical zonation of Caribbean coral reefs. Proc 3rd IntCoral Reef Symp 2:23−29

Gilmour JP, Smith LD, Heyward AJ, Baird AH, Pratchett MS(2013) Recovery of an isolated coral reef system follow-ing severe disturbance. Science 340:69−71

Glynn PW (1985) Corallivore population sizes and feedingeffects following El Niño (1982–1983) associated coralmortality in Panama. Proc 5th Int Coral Reef Symp 4:183–188

Glynn PW (1997) Bioerosion and coral reef growth: adynamic balance. In: Birkeland C (ed) Life and death ofcoral reefs. Chapman & Hall, New York, NY, p 68−94

Goreau T (1959) The ecology of Jamaican coral reefs. I. Species composition and zonation. Ecology 40:67−90

Gowan JC, Tootell JS, Carpenter RC (2014) The effects ofwater flow and sedimentation on interactions betweenmassive Porites and algal turf. Coral Reefs 33:651−663

Graham NAJ, Wilson SK, Jennings S, Polunin NVC, BijouxJP, Robinson J (2006) Dynamic fragility of oceanic coralreef ecosystems. Proc Natl Acad Sci USA 103:8425−8429

Green DH, Edmunds PJ, Carpenter RC (2008) Increasingrelative abundance of Porites astreoides on Caribbeanreefs mediated by an overall decline in coral cover. MarEcol Prog Ser 359:1−10

Guarderas AP, Hacker SD, Lubchenco J (2011) Ecologicaleffects of marine reserves in Latin America and the Caribbean. Mar Ecol Prog Ser 429:219−225

Hallock P (1988) The role of nutrient availability in bioero-sion: consequences to carbonate buildups. PalaeogeogrPalaeoclimatol Palaeoecol 63:275−291

Harrington L, Fabricius K, De’Ath G, Negri A (2004) Recog-nition and selection of settlement substrata determinepost-settlement survival in corals. Ecology 85:3428−3437

Hawkins JP, Roberts CM (2003) Effects of fishing on sex-changing Caribbean parrotfishes. Biol Conserv 115:213−226

Hawkins JP, Roberts CM, Dytham C, Schelten C, NuguesMM (2006) Effects of habitat characteristics and sedi-mentation on performance of marine reserves in St.

Lucia. Biol Conserv 127:487−499Hay ME (1984) Patterns of fish and urchin grazing on Carib-

bean coral reefs: Are previous results typical? Ecology65:446−454

Hay ME (1985) Spatial patterns of herbivore impact andtheir importance in maintaining algal species richness.Proc 5th Int Coral Reef Congr 4:29–34

Hay ME (1991) Fish-seaweed interactions on coral reefs:effects of herbivorous fishes and adaptations of theirprey. In: Sale PF (ed) The ecology of fishes on coral reefs.Academic Press, San Diego, CA, p 96−120

Hay ME (1997) The ecology and evolution of seaweed-her-bivore interactions on coral reefs. Coral Reefs 16:S67−S76

Hay ME, Colburn T, Downing D (1983) Spatial and temporalpatterns in herbivory on a Caribbean fringing reef: theeffects on plant distribution. Oecologia 58:299−308

Hoegh-Guldberg O, Mumby PJ, Hooten AJ, Steneck RS andothers (2007) Coral reefs under rapid climate change andocean acidification. Science 318:1737−1742

Hoey AS, Bel