the primacy of top-down effects in shallow benthic ecosystemsof trophic cascades came fiom studies...

Estuaries and Coasts Vol. 30, No. 3, p. 371-381 June 2007

THE H. T. ODUM SYNTHESIS ESSAY

The Primacy of Top-down Effects in Shallow Benthic Ecosystems

IC L. HECKJR.* andJ . F. VALENTINE

Dauphin Island Sea Lab, University of Soutk Alabama~ t0 t Bienvilte Boulevard, Dauphi~ Island, Ala.bama 36528

ABSTRACT: Individual scientists, scientific organizations, and goxerament agendes have all concluded that eutrophication is among the most detrimental of all human activities in coastal ecosystems; very large amounts of funding have been earmarked to study the negative consequences of nutrient pollution. Most studies of eutrophication have been conducted long after the mtmbers and diversity of larger marine cm~sumers were dramatically reduced by centmqes of intense harvesting. It is now lmderstood that these once abundant predators played pivotal roles in regulating ecosystem structure and fm~c~on, and thai the widespread overharves-ting of large consumers can trigger indirect e f feos that alter spet:ies compositions in x*%vs that are very similaa- to those reported to res-ult from eutrophication. All of this suggests thal we should reev',duate whether the many negative effects attributed to eutrophication m'e ac/xrally a resldt of nutrient additions or whether they may be the restdt of the intiirect effects of drmnatically uitered coastal food webs. In tiffs essay, we rexdm~' experhnental assessments of the degree to which changes in con&nner abm~damces have indirectly altered the slruetm-e of bentlffc ecosystems hi coastal waters, and on the relative importmace of top-down and bottom-up effects on coral reefs, rocky shoros, -and seagq-ass meadows. We find that the evidence dearly indicates that indirect consumer effects axe the prhnm3v drive** of coastal benthic ecosystem structure and fmlcfion.

In t roduct ion

OVERVIEW

In the past two decades, individual scientists, scientific organizations, arid govemInen t agencies have all concluded that eut rophicat ion is among the most detr imental of all h u m a n activities to coastal ecosystems (Nixon 1995; Br icker et al. 1999; Howarth et al. 9000; NAS 2000; NSF 2000). As a z~ very large remounts of ~hncling have been, and cont inue to be, ea rmarked to study ~lle ef~kcis of eut rophicat ion in coastal waters. Among the most compel l ing evidence of the negative consequences of eu t roph ica t ion are r ecur ren t and pers is tent periods of hy-poxia and anoxia (Diaz and Rosenberg 1995; Rabalais and Turne r 2001), ti-equent harmihl algal b looms, and the overgrowth of seagrass meadows and coral reeis by- macroalgae (Howarth et at. 2000). Most studies of nutr ient pol lut ion have been conducted long afier the numbers and diversity of larger consumers were dramatically reduced by centuries of intense harvesting (Dayton et al. 1995, 1998; Pinnegar et al. 2000; Myers and Worm 2003; Pandolti et al. 2005). We now know that the harvesting of large consumers, in many cases to fhncfional extinction, can trigger indirect ettects that result in al tered species composi t ion and abundance at several trophie levels, and can be

*Corresponding author; tele: 251/861-7533; fax: 251/861- 7540; e-mail: kheck@dist org

very similar to those r epo r t ed to result t}om eutrophicat ion (Heck et at. 2000; Williams and Heck 2001). This suggests to us that the extent to which the negative e~tects at tr ibuted to the eutrophicat ion of coastal waters are a resuh of nutr ient additions, or whether they are due to the indirect efl'ects of dramatically altering food webs, is unclear. In ore opinion, the evidence, which we review below, strongIy suggests that indirect consumer e{{ects are o~ien the pr imary drivers of coast~ ecosystem structure and fhnction.

Indirect eftects, by their nature, are complex and di~icult to identify. In part, this is because they include interactions among three or more species and are def ined as "how one species alters the effect that ano the r species has on a th i rd" (Strauss 1991, p. 206). Indirect effects include a host of diRerent interactions, including apparen t competi- tion, apparen t mutualism-facilitation, exploitative compet i t ion, and most famously, t rophic cascades (eL, Strauss 1991; Woot ton 1994). The difficulty in recognizing indirect effects, a long with the tradi- tional focus of fisheries managers on changes in the populat ions of single species rather than on the indirect ecosystem e~ects of fishiag, may have delayed study of the impor tance of the effects of losing apex predators f i om coastal waters. This delay in assessing the indirect consequences of removing large predators is somewhat surprising, given that p reda t ion is a fundamenta l process that has long been known to shape the structure of

�9 2007 Estuarine Research Federation 33'1

:372 K.L. Heck Jr. and J. F. Valentine

marine ecosystems (see Pauly and Watson 2005), and the pr imary means by which energy is trans- fer red f rom pr imary producers to other t rophic levels. The re also is a large body of work that d o c u m e n t s the s t rong eftects of smaller sized predators on macroinvertebrates in coastal waters (e.g., see reviews by Peterson 1979 and Wilson 1990) and the well known literature on the indirect ef~ects of predators in rocky intertidal and subtidal areas, including tile papers of ConnelI (I961), Paine (1966), and Estes and Palmism~o (1974).

It may- also be that the impor tance of the indirect eftects of removing large predators was over looked by estuarine ecologists who spent much of their t ime evaluating tile mediat ing el]leers of s tructured habitats on the tbraging etticiencies of predators (e.g., seagrass meadows: Heck et al. 2003, salt marshes: Minello et al. 2003, mangrove forests: Sheridan and Hays 2003, and o?.~ter reefs: Coen et al. 1999). Higher survival rates of juvenile finfish and shellfish and reduced foraging efficiencies of larger predators are a lnong the most consistently t o u n d characteristics of estuarine nursery habitats (Beck et al. 2001), as are tow smvival rates of epifaunal organisms on unstructured sand and m u d bottoms. Even though the present day effects of predat ion are probably greatly reduced f lom what they once were, these many studies leave no doub t tha t ma r ine p reda to rs con t inue to d e t e r m i n e species abundances , composi t ions , and habitat utilization patterns. Additional examples, some not yet suppor ted by exper imenta l evidence, are reviewed by Jackson et al. (2001), Dulvy et al. (2004), and Steneck and Sala (9005), but it remains true that most coastal ecologists continue to emphasize the pr imacy of bo t tom-up /',actors, most notably nut r ien t availability, in de te rmin ing the dynamics of coastal ecosystems (Howarth et al. 2000; LaPointe et al. 2005; Hauxwell et al. 2006).

Al though it is unders tood by many investigators that bo th top-down and bo t tom up factors can act in concer t to de te rmine the structure and funct ion of coastal ecosystems (Lotze et aL 9006), exper imenta l assessments of the relative impor tance of these two factors remain rare. Our view is that estuarine ecologists have been so heavily focused on bot tom- up factors as regulators of ecosystem s t ructme and fkmction that the very large simultaneous effects that consLIlTIeFs can have on estuarine ecosystems remain unNppreciated. \~ la tever the reason, our goal here is to summm-ize the results of recent exper iments and recta-analyses of exper imenta l studies that have assessed the strength of indirect consumer effects. We also evaluate an additional set of studies that compared the relative impor tance of top-down and bot tom-up t3ctors on benthic t bod webs in coral reef, rocky shore, and seagrass systems.

Seagrass meadows are emphasized because we know them best, and because their global decline and their clear connect ion to a vast array of impor tan t ecosystem services (Orth e t a [ . 2006) have made them a subject of many recent studies, The final section of this essay contmns recommenda t ions tor research projects that wig allow a bet ter under- standing of how an thropogenic manipulat ions of consumers can fundamental ly alter estuarine ecosystems.

DATA SOURCES AND A CAVEAT

We tocus our discussion on studies that relied on manipulat ive exper iments because we believe that exper imenta l evidence is the only conclusive way- of evaluating the magni tude of indirect eftects and of quantifying the relative impor tance of top<town and bo t tom-up factors on coastal ecosystems. Many papers that r epor t observat ional or correlative evidence are not considered here, even though they contain impor tan t information. We also chose not to discuss in detail the most familiar examples of cascading t rophic effects, those involving kelp forests, sea urchins and sea otters, and sea urchins, algae, and coral reefs, ei ther because they are inc luded in the recta-analyses or because they already appear in most basic marine biolog 7 text- books (e.g., Valiela 1995; Nybakken 2000; Levinton 2001).

Before proceeding, we should make it clear that we share the opinions o~ ~ Dayton et al. (1995), Jackson et al. (9001), and Steneck and Sala (2005) who have argued that virtually all of the publ ished preda t ion studies have been carried out unde r conditions that are fundamental ly dif terent tkom those that existed when most p r eda to r -p rey rela- tionships evolved. This is because humans have removed extraordinary quantities of large consumers f rom the world's coastal oceans, so that many are ecologically extinct, and some have been so tor as long as several hund red years (see examples in Dayton et al. 1995, 1998;Jackson et al. 2001; Myers and Worm 2003; Steneck and Sala 2005). Because we believe that most coastal ecosystems are now devoid, or nearly devoid, of apex predators, as well as many mid and lower order consumers (Panly et al. 1998), it is unlikely that any of the publ ished studies were conducted in ecosystems whose species composi t ion was simile" to that pr ior to human intervention. In most instances we can only specu- late on the impor t ance of apex p reda to r s in esmarine and coastal ecosystems, al though it is very likely that top-doxsll pressure on estuarine and coastal food webs is much lower than it once was (Pauly and MacClean 2003). Many intermediate predators remain (perhaps in recently elevated numbers ) , because of release fi-om control by apex

Top-down Effects in Benthic Ecosystems 373

preda to r s (see S teneck et al. 2004) a n d the i r a b u n d a n c e s a nd composi t ions have b e e n manipu- lated in b o t h i n t e n d e d and u n i n t e n d e d ways by commerc ia l interests and g o v e r n m e n t agencies with jur i sd ic t ion over these resources (Pauly and Mac- Clean 2003). It is with these caveats tha t we evaluate the evidence tor consumer-dr iven indirect effects, and the relative impor t ance o f nu t r i en t supplies a nd c o n s u m e r s on ben th i c ecosystem s t ructure a n d tunct ion .

EVIDENCE FOR VERY STRONG INDIRECT CONSUMER EFFECTS ON THE MARINE BENTHOS

T h e Shur in et al. (2002) pape r is significant because it sheds light o n the relative s t rength o f c o n s u m e r effects in mar ine ecosystems. T h e authors c o m p a r e d the s t rengths o f t rophic cascades docu- m e n t e d in six very d i f ferent types o f f o o d webs: lentic, ma r ine a n d s t ream ben thos , lentic a n d mar ine p l ank ton , and terrestrial grasslands plus agricul tural fields. T h e y inc luded 102 studies in their analysis and f o u n d that the s t rongest evidence o f t rophic cascades came f i o m studies d o n e in mar ine ben th ic systems (Fig. 1). Tha t is, manipula- t ions o f the density- o f ben th ic p reda tors p r o d u c e d s t ronger indirect effects on p r imary p roducers , via shifts in herbivore a b u n d a n c e , than did p r e d a t o r manipu la t ions in terrestrial, mar ine o r lake plank- tonic, o r s t ream or lake ben th ic systems. A m o r e recen t meta-analysis by Borer et al. (2004), which i nc luded the 102 studies used by Shur in et al. (2002) plus 12 others, also f o u n d that t rophic cascades were s t rongest in the mar ine benthos . While these meta-analyses were based primari ly on rocky intert idal studies, they clearly demons t r a t e tha t indirect c o n s u m e r effects are c o m m o n a nd s t rong in mar ine ben th ic t o o d webs. This also seems to be t rue in salt marshes where recen t studies (Silliman and Bertness 2002; Silliman et al. 2005) have suggested that f isher ies- induced reduc t ions in b lue crab (Callimctes sapidus) density have allowed o n e o f their ma in prey items, herb ivorous l i t torine snails (Littoraria irrorata), to increase their abundance dramatically. This increase in snail density in t u rn led to increased grazing pressure on, and reduc t ions in, cordgrass (Spartina alterniflora), m u c h in the same way- tha t r educ t ions in sea otters (E~#,ydra [utris) released u rch ins ( Stro~zgytocentrotus purpuratus) f i o m p reda to ry control , and once released, the u rch ins dec ima ted kelp (Macrocystis pyriJera) torests (see review by- Estes 2005).

A n o t h e r oppor tun i ty to evaluate evidence o n the indi rec t effects o f man ipu la t ing h ighe r t rophic level p reda tors comes t}om compar i sons of t b o d web interact ions in repl icated no take Marine Pro tec ted Areas (MPAs) and nearby f ished sites. T h e estab- l i shmen t o f MPAs represents a for in o f whole

3

c~

J~

2 -

1 -

",1:1

i

I Lentio

E~r~ho~

I

I_entic "C "~176

O ~ Temasldal

I I I I - 5 - 4 - 3 - 2 - 1 0

~arflha~

Herbivore Log Ratio

Fig. 1. Effect size of predators, as measured by the log ratio of herbivore and plant density in the presence and absence of predators (-- confidence interval). The effect of predators is significant if the confidence interval does not overlap zero. Predators with relatively little effect on lower trophic levels are found on the lower right, while those with large effects (e.g., marine benthos) occur on the upper left. The linear regression relating plant and herbivore effect sizes is shown by the solid line and the 1:1 relationship is shown by the dotted line (source: Shurin et al. 2002).

ecosystem man ipu l a t i on (Estes a nd Pe te rson 2000) in which mf fo rcemen t o f no take regulat ions can allow the recovery o f over f i shed h i g h e r o r d e r consumer s (e.g., Ha lpe rn and Warne r 2002; Russ a nd Alcala 2004; McClanahan a nd G r a h a m 2005). T h e res tora t ion of h i g h e r t roph ic levels, when s t u d i e d in c o n j u n c t i o n with n e a r b y r e f e r e n c e (fished) areas, provides the oppor tun i ty to assess the effects o f once diverse assemblages o f h igher o r d e r consumers o n lower t rophic levels.

Evidence of s t rong c o n s u m e r effects on lower t roph ic levels in MPAs has b e e n mixed, with some investigators f inding s t rong evidence o f cascading indi rec t effects on lower t rophic levels while others have not. For the mos t part, these cascades have b e e n t b u n d in structurally simple t o o d webs where slow moving invertebrates (e.g., sea urchins) are the key condui t s for the t ransfer o f p r imary p r o d u c t i o n to h ighe r o r d e r consumers (e.g., McClanahan 1998; Shears and Babcock 2003). W h e n the key t roph ic in termedia tes are fishes a nd species r ichness is greater , t rophic cascades seem to be diffuse (Jen- nings a nd Po lun in 1997; Jenn ings and Kaiser 1998; M u m b y et al. 2006). Perhaps this is because h igh diversity- systems c o n t a i n m o r e o m n i v o r e s a n d grea ter amoun t s o f dietary- overlap. T roph i c cascades do occur in h igh diversity MPAs (McClanahan 2005), and some of the best examples o f dramat ic indi rec t effects of diverse assemblages o f h igher o r d e r consumers are the MPA studies of McClana- hall and his col leagues in eas tern Afi-ica (McClana- hall 1998). In this work, the p resence o f large an d

374 K.L. Heck Jr. and d. F. Valentine

diverse assemblages of fishes controlled large-scale habitat distributions by consuming large numbers of sea urchins, reducing rates of coral reef erosion fi-om urchin feeding and preventing replacement of reefs by seagrasses (McClanahan and Kurtis 1991). Such findings are commonplace worldwide for coral reefs within MPAs.

THE RELATIVE IMPORTANCE OF TOP-DO\gN AND

BOTTOM-UP FACTORS IN COASTAL V(ATERS CORAL

REEFS AND ROCKY INTERTIDAL S~STEMS

A growing number of investigators have evaluated the effects of different types of consumers versus nutrient supplies on algal biomass, and concluded that the effects of consumers are greater than, or equal to, those of nutrients. Grazing has been found to have a greater effect on the biomass of macroalgae on coral reefs than nutrient enr ichment by several teams of investigators (Larkum and Koop 1997; Miller et al. 1999; Koop et al. 2001; Szmant 2002; McManus and Polsenberg 2004). While no meta-analysis of the relative importance of top-down and bottom-up forces on macroalgal abundances on reefs has been published, and there is disagreement on the experimental methods used to manipulate nutrients on reefs (see Littler et al. 2005), the preponderance of the evidence ~Svors both strong and concomitant consumer and nutrient effects on the abundance of reef algae.

On rocky shores, Hillebrand et al. (2000) found that benthic algal biomass was determined by strong and balancing effects of amphipod and gastropod grazing and nutr ient delivery. Worm et al. (2000) also found that invertebrate grazers were able to buffer the effects of moderate nitrogen enr ichment on algae in the western Baltic Sea. Another recent meta-analysis provides further evidence in support of the important roles that both consumers and nutrients pl W in controlling algal abundance. In an analysis of 54 marine benthic studies that manipu- lated both herbivore pressure and nutrient loading, Burkepile and Hay (2006; Fig. 2) found that both produced significant effects, but that decreasing herbivore abundance had s t ronger effects on benthic marine macroalgae than did increasing rates of nutrient loading.

The papers reviewed above constitute a large body of evidence that includes most of the experimental work that has been done. They show quite clearly that benthic consumers often have very large effects on their food supplies, as well as indirect effects on community composition, and that these effects are as large, and often larger, than those of nutrients. Taken in its entirety, this literature suggests that the accumulation of plant biomass in shallow benthic habitats is more likely to be controlled by consumer effects than by nutrients.

"13

03 "O G) "I-

N

"d O9

:~= LIJ

Mean effects Individual effects

0

Overall

n = 54

T a

ab tab ta b l a

6 -

5 -

4 -

3 -

2 -

1 o

Tropical macroalgae

n = 1 5

II

II I II

Seagrasses

n = 3

t a

i l a

i iab .

q b tab

i l a

ia

~,9> x-, ~" \e - .§

Fig. 2. Results of meta-analyses on m e a n and indixidual effects (left panel and r ight pane l , respectively) for all p r imary producers, tropical macroalgae, and seagrasses f rom Burkepile and Hay (2006). Effect sizes are Hedges ' d • 95% confidence intelxals. Effects are significant (p < 0.05) if confidence intervals do not overlap 0. A positive d indicates an increase and a negative d indicates a decrease in pr imm T producer abundance . Different lowercase le*ters designate differences a m o n g categories within an analysis as based on 95% confidence intmwals, i.e., data points with d i f ferent letters do no t have over lapping conf idence intervals. Graphs with no letters had no significant differences. Note the consistent pa t tern of greater effect size in the no herbivore t reatments than in the en r i chmen t treatments.

SEAGRAS~ MEADOWS

A prevail ing view, which is par t of the overall conce rn abou t nu t r ien t enr ichment , is that the increasing eu t rophica t ion of bays and estuaries has indirectly tr iggered global reduct ions of seagrass meadows via the overgrowth of seagrasses by the nut r ien t - induced prol i ferat ion of last-growing algae (Duarte 1995; Bricker et al. 1999; Howarth et al. 2000; NAS 2000; Hauxwell et al. 2001). This explm~ation is most often p roposed to account for the ioss of seagrasses in Nor th kanerica (Orth and Moore 1983; Neundor fe r and Kemp 1993; Short et al. 1995; Tomasko et al. 1996), Europe (Giesen et al. 1990; den Har tog 1994), and Australia (Cam- br idge and McComb 1984; Shepherd et al. 1989). An impor t an t point is that, for the most part , these studies were conduc ted in the absence of consumers. As no ted by Heck et al. (2000) and Heck and Valent ine (2006), when algal grazers (primarily mesograzers and some small herbivorous fishes) were inc luded in study designs, grazing eftiects always expla ined at least as much, or more , of the variance in algal a b u n d a n c e than did nut r ien t e n r i c h m e n t (Necktes et al. 1993; Williams and Ruckelshaus 1993; L i n e t al. 1996; see summar ies in Valent ine and Duffle _9006; and Heck and Valent ine 2006). Nixon et al. (~001) summar ized their work in mesocosms over a span of many years and r epo r t ed that there were no signiflcmlt increases in epiphyte biomass following nutr ient enr ichment , a l though they- did observe changes in epiphyte composi t ion.

In aggregate, resuhs i>om these exper imenta l studies clearly showed that mesograzers most often control led the abundance of epiphFtes, even in enr iched conditions, a conclusion clearly at odds with the paradigm of nut r ient -enr ichment based seagrass decline summarized by Duarte (1995) and many others (Bricker et al. 1999; Howarth et al. 2000; NAS 2000; Hanxwell et al, 2001). These results are consistent with the hypothesis that top-down control, via cascading trophic effects like those associa ted with overharves t ing large p reda to r s (Jackson et al. 2001), has impor tan t consequences tor the f lora and tauna of seagrass meadows (Heck et al. 2000; Heck a n d Valent ine 2006). This hypothesis was t~r ther suppor ted by a meta-analysis (Hughes et al. 2004) of the results of the studies cited above, along Mth all others that r epor ted exper imenta l results. [-{ughes et al. (2004) found that among the studies that compa red the relative effects of nutrients a~<l grazers on epiphytic biomass on seagrass leaves, grazers were a key de- te rminant of the extent to which epiphytes overgrew living seagrass leaves. As ~hey pu~ it " T h e positive effects of epiphyte grazers were comparable in


1 . 0 -

0 . 5 -

, . - , 0 . 0

r,.) OO

-0.5 - cn

+1

" - " - 1 . 0 -

U

0 3

- 1 . 5 - o (D

U.I - 2 . 0 -

- 2 . 5 -

- 3 . 0 I I G r a z e r s N u t r i e n t s

Fig, 3. The relative effect of gra~ers and nur~ients on epiphyte biomass growing on seagn~ss leaves~ Note that the magni tude of the effect of removing g r ~ e r s is significantly greater than that of the nut r ien t addition effect (so'urce: Hughes et al. 2004).

magni tude to the negative impacts of water column nutr ient enrichments, suggesting that the 2 thctors should not be considered in isolation of each o ther . " (Hughes et al. 2004, p. 87). The i r pape r (see Fig. 3) actually showed that consumer effects were greater than those of nutrients and explained more of the variance in algal biomass than their cautious s ta tement reflected. At locations where epiphyte loads on seagrass leaves are large, it seems to us that the most impor tant questions to ask are why a ren ' t grazers controll ing epiphytic algae or what h a p p e n e d to the grazers (Heck and Valentine 2006).

A recent stud}, of shoalgrass (Hak~d'u~ arrightii) ineadows in the Gulf of Mexico (Heck et al. 2006) found significant consumer and nutr ient effects oll seagrasses, a l though there were fewer consumer eflects and more nut r ien t effects than in our

376 K.L. Heck Jr. and J. F. Valentine

previous study- (Heck et al. 2000). Because epiphyte prol i fera t ion did not occur in nu t r ien t e n r i c h m e n t t rea tments in this exper iment , algal overgrowth could not explain the negative effects os nu t r ien t l oad ing on seagrass b iomass tha t we f o u n d . Nut r ien t loading p r o d u c e d ni t rogen-r ich shoalgrass leaves, and this high-quality t ood s t imulated increased pinfish (Lagodon rhomboides) herbivory. Elevated pinfish c o n s u m p t i o n of the en r i ched shoalgrass then resul ted in the decline of seagrass biomass in e n r i c h m e n t t reatments . These unex- pec t ed results are similar to those os Williams (1988), McGlathery (1995), and Goecker et al. (2005), who all observed that elevated n i t rogen conten t in seagrass leaves t r iggered increased fish grazing, and they demons t ra te that there is m u c h m o r e to learn abou t the interact ing effects of nu t r i en t supplies and the feeding pa t t e rns os consumers in seagrass meadows.

CHANGING PARADIGMS FOR SHALLOW BENTHIC ECOSYSTEMS?

Research on s u b m e r g e d aquatic vegeta t ion in lakes in the Uni ted Kingdom has gone th rough a t rans format ion that appears to be similar to what we have descr ibed tor seagrasses. In British lakes, nut r ients were first t hough t to be the driving force leading to the demise of roo ted macrophytes (see Phillips et al. 1978). Jones and Sayer (2003) conc luded tha t the key d e t e r m i n a n t of shallow water p lan t b iomass was actually fish p reda t ion on invertebrates , which, t h rough a t r o p h i c cascade, indirectly inf luenced the biomass of leaf per iphy- ton a n d r o o t e d mac rophy te s . While n u t r i e n t e n r i c h m e n t is an ex t remely i m p o r t a n t issue in coastal areas t h r o u g h o u t the world, because of its association with phy t op l ank t on and macroalgal b looms, so too are t ood web interact ions, which mus t no t be ove r looked in fu tu re studies os eu t rophica t ion . The no t ion that top-down control is o f p r imary impor t ance in seagrass meadows worldwide requires tu r the r co r robora t ion in o the r parts os the world (see below), bu t the evidence for this is a b u n d a n t and growing (cf., Heck and Valent ine 2006), and it comes with a n u m b e r os appl ied implicat ions (e.g., reduc ing nu t r i en t inputs to coastal waters may not b r ing abou t full seagrass recovery if epiphyte grazers are artificially low in abundance ) . Evidence for the p r imacy of consumer eftects in de t e rmin ing algal b iomass is also growing s t ronger for coral reefs, kelp forests, and rocky shores, as discussed above, and we believe it likely that top-down factors will soon be general ly acknowledged to be of primary- impor- tance in mos t mar ine ben th ic ecosystems.

FUTURE NEEDS

There is a clear need to evaluate the generality os the results discussed above at sites beyond temper- ate Nor th America, Europe, and Australia. We also need to de te rmine more clearly the extent to which the findings of small scale studies conducted in the laboratory or in the field can be extrapolated to predict the relative effects os top-down and bot tom- up factors at the ecosystem level, because most informat ion on tile effects os higher order consumers on lower t rophic levels comes f rom experiments conducted at scales of one to tens of meters. Processes operat ing at such small spatial scales often difter f rom those opera t ing at larger scales (Thrush et al. 1995; Crowder et al. 1997; Sih et al. 1998; Estes and Peterson 2000). Conclusions f rom small-scale exper iments cannot sately be extrapolated to entire ecosystems without validation (Walters and Holling 1990; Eberhard t and Thomas 1991; Menge 1992; Carpen te r 1998), and o ther approaches to evaluating the effects os larger, highly mobile , apex predators are needed. To date, large-scale manip- ulations (e.g., at the size of an embayment ) of consumers and nutrients have not been conducted in coastal waters. Limnologists have learned a great deal by manipula t ing entire lake ecosystems (e.g., Schindler 1998; Carpen te r et al. 2001), even though repl ica t ion and controls are of ten difficult to include in such study designs. We feel strongly that ma r ine and es tuar ine ecologists could benef i t greatly t}om employing this approach.

We caution that most of the studies analyzed by Shur in et al. (2002) and Borer et al. (2004) manipu la ted only a single species os predator . Most food webs contain a diversity- of consumers whose interaction strengths vary greatly, and their com- b ined effects can be unpredic table and non-additive (Crowder et al. 1997; McCollum et al. 1998; Sih et al. 1998). T h e r e is a n e e d to s imul taneous ly manipula te mult iple variables to bet ter simulate conditions in nature and more realistically evaluate the relative impor tance os top-down and bot tom-up factors. Prior studies have manipula ted nutrients and usually only one type os grazer. This design should be expanded to manipula te multiple phys- icochemical variables (e.g., salinity-, nutrients, and tempera ture ) , along with multiple combinat ions os grazers (Estes and Peterson 2000; Ibar ra-Obando et al. 2004).

ADDITIONAL TOP-DOWN AND BOTTOM-UP QUESTIONS

There are other questions about the opera t ion of top-down and bot tom-up processes that remain to be answered. If algal grazers pre te r nitrogen-rich plants, as demons t ra ted experimental ly tor seagrasses by Williams (1988), McGlathery (1995),

and Goecker et al. (2005) and tbr algae by H e m m i and Jornialainen (2002) and Boyer et al. (2004), how can nitrogen-rich f i lamentous green algae, which are characteristic o~ eutrophic waters, accu- mulate when g r~ e r s are present? One likely answer is that since most Elmnentous green ~gae are palatable to a wide range of grazers, persistent accmnulat ions of green ~gae are only possible if there are tew grazers present. This can easily be tested experimentally, using both ni t rogen-enriched and unenr i ched algae ,~ith a variety of consumers.

Are there latitudinal differences in the stimulato- ry effects of elevated nutr ient inputs on epiphyfic algae (e.g., more negative effects in cold than warm climates because grazer abnndance may not be able to catch up to algae that begin spring-time growth before animals in areas with short growing seasons)? This can be addressed by compar ing the results of latitudinally distant studies, and this can be done simply by separating the studies in the meta-analyses discussed above by latitude,

Are there positive eftects of nutr ient en r i chment on the consumers of algae with high ni t rogen content and altered carbon:n i t rogen:phosphorus ratios? Eco]ogical stoichiometry predicts that ele- mental ratios of consumers will r e m ~ n constant, despite ~he makeup of their fbod sources. To achieve this constancy, consumers must adjust their assimilation m~d excretion efficiencies in accord with the e lemental composi t ion of their tood (Elser and Urabe 1999). Because tbod quality- can play- a major role in d e t e r m i n i n g the growth and iiecundity of consmners, we might expect positive eftects on these factors in eut rophic waters contain- ing nitrogen-rich algae. To date, we are not aware that this has been tested with coastal consumers.

Can chemically deterided algae (e.g., red and brown algal species) becom e abundan t on the surfaces of substrates (e.g., rocks, bivalve shells, or seagrass leaves) even in locations where grazing is intense? This could h a p p e n when fast-growing palatable algae (e.g., f i lamentous green algal species) are kept in check by grazers, which then allows slow-growing unpala table species to prolitierate. DriR algal mats of brown and red algae commonly ibund in North and Cen.tral American seagrass meadows may be an example , as may 5e the macroalgal accumulat ions often associated with eunoph i c waters.

TP~MT-MED I&TE D ]INTER&C1 IONS

Until very- recently, trmt-umdiated effects of predat ion on communi ty structure and funct ion have been virtually unexp lo red in marine environments (Dill et al. 2003). These non le tha l eftects of predators on prey, primarily expressed th rough changes in prey behavior, can be as impor tan t as

Top-down Effects in Benthic Ecosystems :377

density-mediated (consumptive) preda t ion in terrestrial and treshwater systems (see meta-analyses by Schmitz et al. 2004; Preisser et al. 2005). Marine examples that do exist include the work of Hei thans and Dill (2002; see ~so Heithans et ~. 2002), who have shown that predat ion threats by tiger sharks (C~ff~'.ocerdo c.uvier) ,altered the fbraging behavior of bot t lenose dolphins (Tursi@s aduncus), such that they avoided the food-rich seagrass habits that are also f~tvored by the tiger sharks. Dolphins trade off access to abundan t food in the seagrass beds for increased survival rates. This trade-off benefits the fish as predat ion rates by dolphins are reduced (Heithaus and Dill 2002). In the New England rocky intertidal, trai t-mediated effects were found to be as eftective as densiW-mediated effects in regulating snail density that in turn regulated ephemera l green algal abundance in tide pools (Trussell et al. 2002, 2004). In the New England rocky subtidal, Freeman (2006) showed that small green sea urchin (Strca~gy- torch, trot'us droebachiensis) g r ~ i n g on macroalgae was significantly reduced in the presence of echinovor- ous sea stars (PTozopodia helian, thoides).

Dill et al. (2003) list a n u m b e r of cases of bo th positive and negative eitects of behaviorally-mediated interactions, one type of t ral t lmediated interactions (TMI) a m o n g mar ine organisms. While the n u m b e r of existing cases is not large, TMIs are likely to be c o m m o n in mar ine ecosystems, and can have impor tant , unant ic ipa ted consequences for ecosystem structure and funct ion (Dill et al. 2003).

EFFEC2"S O1= PREDATORY [NVASIVE SPECIES

We briefly address the unant icipated, indirect eftects that can aaise f rom the in t roduct ion of nonnat ive consumers. This increasingly c o m m o n p h e n o m e n o n , with examples known in every- type of shallow water habitat (Steneck and Carlton 2001) has led to some of the bet ter known examples of mar ine trophic cascades. The introduct ion of the normat ive green crabs (Carcinus maenas) t~om Europe to the Gulf of Maine p roduced tundamenta l changes in the species composi t ion and abundance of organisms in both subtidal flats and the rocky interfidaI of New England (Steneck and Carltun 2001). Where perixdnkles (Littorh~a 5~to~'eca) are present, green crabs reduce periwinkle %ecling by bo th direct and trait-mediated means, and this allows ephemera l green algae to proliferate. In the absence of green crab, periwinkles preferential ly consume green algae and the result is dominance of the substrate by less palatable brown and red algae (Lnbchenko 1978; Vadas and Elner 1992). Another location where normative consumers have had large effects on ecosystem structure and funct ion is San Francisco BW, where the Asian clam (Potamocorbuta

3 7 8 K.L. Heck Jr. and J. F, Valentine

amuro,sis) has greatly a l tered p h y t o p l a n k t o n species compos i t i on a n d a b u n d a n c e , improv ing water clar~ it}, tha t fi~w~rs tile t?roliferation o~ ben th i c p l an t species (Car l ton 1999). O t h e r examples inc lude the ettects of the n o n n a t i v e c t e n o p h o r e Mnemiopsis leidyi in the Black Sea (Malyshev a n d Arkhipov 1992) a nd the exotic seastar Asterias amurensis in Austral ia (Bn t t e rmore et al. 1994), In b o t h cases the n o n nat ive species had large eKects on ene rg T flow a n d food web s t ruc ture (Steneck mid Car l ton 2001). T h e o n g o i n g process of m a r i n e i n t roduc t ions (Ruiz et N. 1997; Car l ton 1999) cons tant ly provides ot?t?or- tuni t ies to investigate how c o m m o n l y nonnaf fve p reda tors p r o d u c e cascading e~tects tha t drastically change the food web s t ructure of b e n t h i c ecosys~ terns.

PLANKTON VERSUS BENTHOS

We po in t ou t the surpr i s ing f ind ing that there are relatively few cascading t rophic effects in m a r i n e p l ank ton ic food webs (Micheli 1999). This was u n e x p e c t e d because the s emina l work on the i m p o r t a n c e of t r o p h i c cascades was d o n e on p l a n k t o n i c food webs in lakes ( C a r p e n t e r a n d Kitchell 1993). S h u r i n et al. (2002) c o n f i r m e d that in m a r i n e p l a n k t o n i c food webs, p h y t o p l a n k t o n showed lower responses to p reda to r removals t h a n those in freshwater. W h e t h e r this lack of cascading eftects reflects a f u n d a m e n t a l d i f ference be tween the p l a n k t o n i c systems of fresh and salt water, a d iKerence be tween the m a r i n e p l ank ton ic mad b e n t h i c systems, or whe ther the ettects of r emov ing consumers f rom m a r i n e p l a n k t o n i c food webs has a t t e n u a t e d s t rong cascading eftects, is debatable . St ibor et al. (2004) a n d Dufty- a n d Stachowicz (2006) suggest tha t apex p r e d a t o r losses o f t en do cascade to m a r i n e phy top [ank ton , b u t whe ther the effect is posit ive or negat ive is d e t e r m i n e d by whe ther the fbod cha in has three or f~Jur levels, which is a f u n c t i o n of the cell size of the d o m i n a n t phytoplankters . These au thors suggest that because earl ier studies (like those s n m m a r i z e d by- Micheli 1999), i n c l u d e d results f rom b o t h th ree a n d four l ink food chains, the i r eftects cance l led out , l ead ing to the incor rec t c o n c l u s i o n that cascading eftects in m a r i n e p l a n k t o n i c assemblages were weak (Duffv a n d Stachowicz 2006), T h e r e is m u c h that r emains to be exp l a ined a b o u t the o rgan iza t ion of plank~ tonic and ben th i c fbod webs in shallow coas t s waters.

Conchts ion

T h e papers reviewed here cons t i tu te a large body of e~qdence that inc ludes most e x p e r i m e n t a l work d o n e o n the ind i rec t eftects of a l ter ing c o n s u m e r a b u n d a n c e , a n d o n the relative i m p o r t a n c e of top~

down a n d b o t t o m - u p factors o n coral reefs, rocky shores, a n d seagrass meadows. It clearly shows, b u t only because manipu la t ive expe r imen t s were used, tha t the indi rec t effects of ben th i c consumers are s t rong a n d m u c h m o r e widespread than mos t have believed, T h e cumula t ive e~tects of a diverse array of coastal consumers are, on average, as s t rong or s t ronger t h a n tile of ten r epo r t ed eftects of eutro- ph i ca t i on , T a k e n in its entirety-, this l i t e ra ture indicates that the a c c u m u l a t i o n of p lan t b iomass in shallow b e n t h i c habitats is m o r e likely con t ro l l ed by c o n s u m e r eKects ~han by nut r ien ts , and fllat ind i rec t c o n s u m e r eKects are o~ien the pr imary drivers of coastal ecosystem s t ructure a nd iunc- t ion, We note that the impl ica t ions of this con- c lus ion ta r m a n a g e m e n t are s ignif icant , as ou r analyses make cleat- that r e d u c i n g n u t r i e n t i n p u t to coastal waters is unlikely- to res tore b e n t h i c habitats such as seagrass meadows or coral reeis if there have b e e n pervasive a l tera t ions to f0od webs that have resul ted in r educed mesograzer popula- t ion sizes.

ACKNOWLEDGMENTS

We thank Carlos D'uarte for gMng us the opportunity to write this essay, D. Byron and C. Wood for help in preparing this manuscript, and three anonymous reviewers for their constructive criticism. We also acknowledge our funding sources for support- ing the development of these ideas: the Environmental Ptvtection Agency through its fi.mdmg of the Alabama Center for Esmarine Studies; the Nationat Marine Fisheries Service through its support cf the Alabama Oyster Reef and Fisheries Habitat Enhaneetaen~ Program, and for funding from the MAre.FIN Program; the Nature Conservancy's Ecosystem Research Program; and the Dauphin Island Sea Lab. This is Dauphin Island Sea Lab Contribution Number 386.

LITERATURE CITED

BEta% M. W., g_ L. HEcx JR., K. W. ABLE, D, L. CHXLnFa~S, D. B. EGGLESTON~ B. ~vI. GILLANDERS, B. HALP~a'4, C. G. HAYS, K. HOSHINO, T .J . ~IINEI-LO, R.J. OKTHj P, F. SHERIDANj AND M~ P \~EINSTEIN. 2001, The identk-tca.~icm, conservation, and management of esmarine and marine nurse*{es for fish and invertebrates. Biosdence 51:633-641.

Bom~R, E. T., E. W. SZABLOO~a, J. B. SHURIN, K, E. ,~DE~SON, C.A. BLANCHETTE, B. B,v, OITMAN, S. D. COOPER, AND ]~. S. HALPERN. 2004. What detem~ines the strength of atrophic cascade? Ecology 86:528-537

Bo'mR, K. E., P. FONG, A. R. ARMITACE, ANn R. A. COHEN. 2004. Elevated nutrient co utent of tropical macroalgae increases rates of herbivory in coral seagmss, mangrove habitats. Carall~@ 23: 530-538.

BKICKNR~ S B~ C. G. CLEMENT s ]-) F~ PIE2tALLA, S. P. ORLANDO, AND D. R. G. F.~,Row 1999. National Esruarine Eutrophicatb~n Assessment: Effects of Nutrient Enrichment in the Nation's Estuaries. National Oceanic and Atmospheric Administration, National Ocean Service, Special Projects Office and the National Centers for Coastal Ocean Science Silver Spring, Maryland.

BLrRKEPILE, D. E. AND M. E. HAY. 2006. Herbivore vs. nutrient control of marine prmmry producers: Context-dependent effects. Ecology 87:3128-3139.

Top-down Effects in Benthic Ecosystems 3 7 9

Btrrr~,aom~, R E., E TURNER, AND M. G. Moo, rata< 1994. The in t roduced no r t he rn Pacific seastar Aste~as arnurensi~ in Tasmania . Memoox oJ'~he @eensla~d Museum 36:21-25.

CAMBRIDGE, ]~[. L. AND A. J. McCob~m. 1984. The loss of seagcasses in Cockbuca Sonnd, westero Australia. I. The time course and magni tude of seagrass decline in relativn to industrial de velopment . A~z~ahc Bo!a.~3~ 20:22q-24B

C.~d~TON, .I" W. 1999. T he scale and ecological consequences of biological im, asions in the world's oceans, p. 195-212. In O. Sandlund, P Schei, and A Viken (eds.), Invasive Species and BiodiversiB" Management . Kluwet; Dordrecht, The Nether lands.

CAI12ENTER, S. R 199& Ecosystem ecology: Integrated ph3~ical, chemical a~d biological processes, p. 123-162. t*t S. I. Dodson fed.), Ecology. Oxford University Press, London , England.

Cam, zr, r r~ , S. R. A~,mJ. F. t~TC:H'~LL. 1993. The T m p h i c Cascade ir~ Lakes. Cambridge University Press, Cambridge, U.I~

Cam'z~rE~., 8 R, J. J. COLE, J. R. HODGSON, J. F. I~awc~zLL, M. L. PACE, D. B.~.o~, K. L. COTr~NGHAM, T. E. ESSaNGTON, J. N HouszR, AND D. E. SCHINDLER. 2001. T roph ic cascades, nut r ients , and lake productivity: Whole-lake exper imen t s Ecologi.caI Monog,z~phs 71:163-186.

Cozy, L. D., IVI W. LUC%2ENBACH, AND D. L. Bm~tT~mG. 1999. The role of oyster reef~ as essential fish habitat: A review of curretxc knowledge at~d some new perspect)ves, p. 438-454. tn L. R. Benaka fed.) , Fish Habitat: Essential Fish Habi ta t a~ad Restor'ation./American Fisheries Society Syrnpesia 22. Washing- ton, D.C.

CONN~a~L, J. H. 1961. The influence of interspecific contpetition and other factors con the distribution of the barnacle Chth.t~.mol~s std~atus. Eco/o~ ~t2:710--723.

C~owI}r~, L. B., D. D. Sqmm~s, ANDJ. A. PdCZ. 1997. Nonadditi~e effects of terrestrial and aqnatic predators on jnvenile esmar ine fish. Ecd%�9 78:1796-1804.

DAYFON, P. K., M.J. TEGNER, P. B. EDWARDS, AND K. L. PdS~R. 1998 Sliding baselines, ghost, and reduced expectations in kelp forest communit ies . Ecological Applicatiom 8:309-322.

DAYFON, P. K., S. F. THRUSH, M. T. AGARDY, AND R. J. HOt'NAN 1995. Viewpoint: Environmenta l efffeets of marine fishing. Aquatic Comemahorl: Mature a~d Freshwater Ecosystems 5:205-23~

DEN H~TOC;, C. 1994. SuffOcation o f a littoral Zostera bed by Enteromo~pha india ta. Aquatic Botan~ 47:21-28.

DIAz, R.J. AND R. ROSE~,n~ER& 1995~ Marine benth ic hypoxia: A review of its ecological effects and tl?e bettaviuural response of ben th ic macrofanna . Ocen.~z~g~p~ 3, ~a4 Mezri'ize Biology: A~z A'mnee~t Review 33:245-303.

DILL, L.J., M. R. Ht~T~aUs, aND C. J. WAL~RS. 2003. Behaviorally media ted indirect interactions in mar ine communi t ies and their conseB, ation implications. Ecology 84:1151-1167.

DUART~, C. M. 1995. Submerged aquatic vegetation in relatioD to different nu t r ien t regimes. Ophelia 41:87--112.

Dl;ggY, J. E. ,bND J~ J. STAC~OVaCZ. 2006. Why biodiversity is impor tan t to oceanography: Potential roles of genetic, species and trophic diversity in pelagic ecosystem processes. Marine Ecology Probzre~s Series 311:179-180.

DHEvY, N. K., R. P. FRECKI~ETON, AND N. V. C. POLUNIN. 2004. Corral reef ca~scades and the indirect effects of predator removal b}' exploitation. Ecolo~, Letter 7:410--416.

EB~Am*T, L. L. ~,YD J. M. THONAS. 1991. Designing envirot~ menta l field studies. Ecological Mo~*ographs 61:53-73.

EL51~R, J. j . AND J~ u ~ . 1999. The stoichiometry of consumer- driven nut~xent rec)c[ing: Theory, observations, and corpse ~ vluences. Ecology 80:745-751~

E s ~ s , J. A. 2005. CamivoP)~ and wophic connectivity in kelp forests, p. 61-81. btJ . C. Ray, I{- H. Redford, R. S. Steneck, at?d J. Berger (eds0, Large Caznivores and the Conservation of Biodiversity. Island Press, Washington.

ESTES, J. A. a>aD j . I r P~>alSANO. 1974. Sea otters: Their role in s t ructur ing nearshore communi t ies . Science 185:1058-1060.

E s ~ s , J . A. AND C. H. PZTERSON. 2000. Marine ecological research in seashore and seafloor systems: Accompl ishments and furore directions. Marius Ecology Progress 5en'es 195:281-289.

F~EEMAN, A. 2006. Size-dependent tcait~mediated indirect interne- dons a m o n g sea urchin herbivores. Behawo~zd Ecology 17:182- 187.

GIESEN, W. B..]. T , ivl. M VA~,' t~ws<JK, AND C. DZN H.~,XTO~3 lqq0. Eelgrass condit ion and t~,'bidity it? ~he Dutch Wadden Sea. Aquatic Botar~ 37:71-85.

GOEClCKR, M. E., K. L. HzcN, JR, a s p J. F. Va~ENnNZ~ 2005. EffEcts of ni t rogen concentrat ions in tur t leg~ss, ThMassia testudinurn, on c o n s u m p t i o n by the bnck too th parrotf ish, Spar isoma tadians. Marine Ecology Progress Series 286:239-248.

HAta'ERN, B. S. AND R. R. WARNER. 2002. Marine reserves have vapid and lasting effects. Ecology Letters 5:361-366.

HAUXWEH~, J., j . C~B~dAN, C. FtremoNC:, .a~,a) I. V~U~LA. 2001. Macroalgal canopies contr ibute to eelgrass (Zostera maria*a) decline in tempera te estnarine ecosystems. Ecology 82:1007- 1022.

HAUXWELL, J., j . CEBRIAN, AND I. VALIEL&. 9-006. Light dependence of Zostera marina annua l growth dynamics in estuaries subject to different degrees o f eutrophicat ion. Aq~abc Botany 84:17-25.

HECK, JR., K. L., C. G. HA~, AND R. J ORTH~ 2003. Critical evaluation of the nursery hypothesis for seagrass meadows. Marcus Eco/og~ Progre~ ,bb'~e~ 253:12 g~ 136

HECX, JR., ts L., .] R PENNOCa-~,J. F. VALs L. D. COEN, .aNn S. A. S~CLZN~ 2000 Effects of n~trie~at en r i chmen t and small predator density on seagrass ecosystems: An exper imenta l assessment. Lirn~*ology (znd Oce~n%>rap]o: 45:1041-1057.

HECx, JR., K. L. AND J. F. ~,.',~a~?,mNE 2006 Plant-herbivore interactions in seagrass meadows. Journal of Ex~)ecirnental Marius Biology and Ecology 330:420-4.36.

HECK, JR., F~ L., J. F. VALENTINZ, J. R. PI;NNOCX, G. CHAPtaN, AND P. M. SprrzzR. 2006. Effects o f nu t r ien t en r i chmen t and grazing on shoalgrass (Halodule w~ghti~) and its epiphytes: Results of a field exper iment . Marine Ecology t>r%~ress Series 326:145-156.

HHTHAUS, M. R. AND L. M. Dmt. 2002. Food availability and tiger shark predat ion risk influence bolt /chose dolphin habitat use. Ecology 83:480-491.

H~aTHAUS, M. R., L. M. DmL, G.J . MA~51"tALL, AND B. Btn-~]EK 2002. Habitat use and foraging behavior o f tiger sharks (Galeocerdo c~uvier) in a seagrass ecosystem. ),lat-b~e Biology 140: 237-248.

HEMMI, A. AND V JOR2CIALAINEN. 2002. Nutr ient e n h a n c e m e n t increases peNotTnance of a marine herbivore via quality of its food a lga Ecol%;~ 83:1052-1064.

H~U~_a~,aN~, H., B. Worn% AN> H. K. LOTZZ. 2000. Marine mic rohen th ic communig~ s t ructure regulated by n i t rogen loading and grazing pressure. Morme Ecology IbogreB Se~4es 204: 27-38.

HOWA~TH, R., D. ANDERSON, j. cL~m-~, c. ELFRlNG, C. HOPK1NSON, B. LAPOINTE, T. MALONE, N. M~~ K. MCGLATHERY, A. SHARPLEY, AND D. WALKER. 2000. Nutt ' ient pollut ion of coastal rivers, bays, and seas. Issnes in Ecology 7. Ecological Society of America, Washington, D.C.

HUGHES, A. R., K.J. BANDO, L. F. RoDmCkmz, AND S. g. WILLIAMS. 2004. Relative effects of grazers and n u u i e n t s on seagrasses: A meta-analysis approach. Marine Ecology PrQ, ores~- Series 282:87-99.

IBARRA-OBANDO, S. E., K. L. H'~CX, Js., AN~ P. M SP]TZZR. 2004. Effects o f s imul taneous changes in Ligltt, nut r ients , and herbivory levels, on the structure and fi.mction of a subtropical rurdegrass meadow. Jo~rr~al oj Expemne~ad Ma~'nte Biology and E~vlogy 301:19B-224.

J a c x s o ~ , I. B. C., M X. t{am~u W. H. Bz~cz~, K A. BJo~,'~.~J_, L W. BoTsso~2,, B. J Bov,~q2uz, R. H. Br, a n ~ u ~ ; R. Coo}ca, J. ~RLANDSON,J. A ESTES, T. P. HUGHES, S. I~D\~rkLL, C. B. LANGZ, H~ S. LZN]HAN, J. M. PANDOLN, C. H. PI~TERSON, R. S. STENE~_~, M. J. TEGNE~, AND R. R. W.~,NZ~.. 2001~ Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629- 638.

3 8 0 K.L. Heck Jr. and J. F, Valentine

JENNINGS, S. AND M. J~ l~asln~ 1998. The effects of f ishing on mar ine ecosystems. Advmwes i~ Ma.rine Biology 34:201-352.

JENNINGS, S. aND N. V. C. PO1.V-N~N. 1997. hnpacts of predator depletion by fishing con the biomass and di~.ersity of non-target reef fish comm unities. Cco-a~ Re@ 16:71-82.

JoJ',~s, j . I~ a ~ c~ D sA,~wx~ 2003. Does the fish-inver'tebr'ate- periphyton cascade precipitate plant loss in shallow lakes? Eml~gy 84:2155-216Z

I(oo~', K, D BOOTH~ A. B~.OAOI~I~NT, J. BRODm, D. B~cPm~, D. CaeON~, J. COL~, W. DENNISON, M. ERDMAN, P. HARRISON, O. HOEGH-GULDm~RC;, R HUTCa-IING8, G. B. JONL% A. W. D. LARXUM, J. O'NEILL, A. STEVEN, T. T~yrom, S. W ~ , J. Wnm~AMSON, AND D. YELLOWLEES. 2001. ENCORE: T h e effect of n u t r i e n t en r i chmen t on coral reefs. 2. Synthesis of results and conclusions. Marine Pollution Bulletin ~t~:91-120.

Laro~N~, B. E., P.J. B a r d ~ , J J L r r r ~ r , ann D. S. L~Tr~ra~. 2005. Macroalgal b looms on soa theas t Florida coral reefs: II. Cross- she l f d i sc r imina t ion of n i t rogen sources indicates widespread assimilation o f selvage ni t rogen. Harmful Algae 4: 1106-1122.

Lam~tr~a, A. W. D. Arm K Koo~' 1997r ENCORE: Algal productivity and possible paradigm shifts. 2aroceedings 8th International Coral Reej Syrnposiurn 1:881-884.

L~WNWON, J. s. 2001. Marine Biology: Function, Biodiversity, Ecology, 2rid edition. Oxford University Press, New York.

LIN, H.J . , s. w. N~xoN, D. 1. TA~LOg, S. L. GRANGER, AND B. A. BUCRLEY. 1996~ Responses of epiphytes on eelgrass, Zostera marina L., to separate and combined ni t rogen and phosphorus enr ichment . Ag~a~ir Bo~a~y 5~:243-258.

L~TTc~, M M , D~ S. L~Tr~.~Z, aND Br Lr BROOXSr 2005r Har'mful algae on tre~pical corral reefs: Bottom-up eutrophicat ion and top-down he~%ivr~r>~ Har~zd Algae 5:5~5-585

LOTZE, H~ K., H~ S~ LENIHAN, B J, BOIAa~.QIAZ, R. H. ~dRADBUKY, R. G. COOgs M. C. 1,2~ L S. M. t{1BWELL, M. X. I{am~" 6 C. H. PETEKSON, ANn J~ B. C. Jac~csON. 2006. Depletion, degredat ion, and recovery poteuttal of estuaries and coastal seas. Science 312: 1806-1809.

Lt~C~q~NRO, J. 1978~ Plant species diversity in a mar ine intertidal community: Impor tance of herbivore food preference and algal competitive abilities. Am,~*can A.'aturalist 112:23-39.

MAL'eSH~v, V. I. ANn A. G .~.~.m~ov. 1992. The c tenophore Mnemiopsis lddyi in the western Black Sea. Hydrologicalfournal 28:33--39.

MCCLANAHAN, Z. 1998. Predat ion and the distr ibution and a b u n d a n c e o f tropical sea u rch in populat ions . Journal of Experimental Marine Biology a.nd Ecology 218:231-255.

MCCLANAHAN, T. R. 2005. Recove U of carnivores, t rophic cascades and diversity in coral reef marble parks, p. 247-267./:~J. C. Ray, K. H. Redford, R. S. Steneck, and J. Berger (eds.), Large Carnivores and the Conservation of Biodivet~ity. Island Press, Washington, D.C.

MCCLANAHAN, T. R. aND N. A. J. GRAHAM. 2005. Recovery trajectories of coral reef fish assemblages within Kenyon marine protected areas. Ivlar~e Ecolo~ Yr%,ress &,ties 294:241-248.

McCLANA2qAN, T. ANDJ. I(URTIS. 1991. Population regulation of the reck-bering sea re'chin Edd~,zo~etra. "magka.d (de Blainville) .[o a raal of E~e~-.~e~z~al Marine ~oioIo~ a~zd Ecology 147:121 - 146.

M(_'COLLLQI, E. W., C. B. CKO~.~ER, AND S. A. MGCOLLU~'I. 1998 Complex interactions of 6sh, s~ails and littoral zone periphyton. Ecology 79:1980-1994.

MCGLATHERY, K 1995. Nut r ien t and grazing inf luences on mhtropical seagmss commut~ity, Marine Ecology Progress Series 122:239-252.

MCMANUS, J. W. ~NI)J. F. Pol.s~rq~a~3. 2004. Coral-algal phase shirts on coral reefs: Ec.oMgica[ and envi ronmenta l aspects. Progress in Oceanography 60:26"3-279.

M~NGE, B. A. 1992. Commtmit3, regulation: Unde r what conditions are hot tom-up factors impor tant on rocky shores. Ecology 73:755-765.

MICHELI, F. 1999. Eutrophication, fisheries, and consumer-re- source dynarnics in mar ine pelagic ecosystems. Scie~*ce 285: 1396-1398.

M~LI~L% M. W, M E. HaY, Sr L. M]~.~.ra~, Dr /vL,~OyE, E. E. SOTXA, ANn A. M SZMAbaT. 1999r Effects of nutr ients ver'sus her'bivores on reef algae: A new me thod for manipu la t ing ncmients on co ral reefs. L.,rmzdo~ a~zd Ocea~zogr@ky 44:1847-1861.

MIiVELLO, T.J. , K W. 2~LZ, M. P. WLINS~IN, ANn C. G. Ha~ . 2003~ Salt marshes as nurseries for nekton: Testing hypotheses on density, gro*~th :and survival th rough reels-analysis. Ma*'ine Ecology l~vgr~s Series 246:39-59.

MUMBY, P. J., C. P. DA/4LGR~N, A. R. Ham~om,~E, C. V. KArrrm, F. MlCar~LI, D. R. Bro_q~m.~rjcH, K. E. HOL~,~S, J. M. MENDES, I~ BROAD, J. N. SANc~Ivaco, K B~CH, S. Box, R. W. STOrfLI% aND a . g. (JILL. 2006. Fishing, t rophic cascades, and the process of grazing on coral reef~. Science 311:98-101.

Msz~s, R. A. AND B. Worn,.1. 2003. Rapid worldwide depletion of predatory fish comnmai t ies . Nature 423:280-283.

NATIONAL ACADEMYOF SCIENCES (NAg). 2000. Clean coastal waters: Unders tand ing and reducing the effects of nu t r ien t pollution. Nat ional Academy of Sciences, Nat ional Academy Press, Washington, D.C.

NATIONAL SCIENCE FOLrNDAT1ON (NSF). 2000. Ocean sciences at the new mil lennium. Proceedings f rom the University Corpo- ration fbr Atmospher ic Research Joint Office for Science Support. Washington, D.C.

N'~C:KL~, H. A., R. L. Wzwz~, ANP R. J. ORTH. 1993. Relative effects of nu t r i en t en r i chmen t and grazing on epiphyte- macrophyte (Zosle~a marbTa L.) dynamics. OecoI%&* 93:285-295.

NEtr~no~'~R, J. V~ a ~ W. M. I~,',.~'. 1993. Nitrogen ver'sus phc~sphcwc~s enr i chmen t of br'ackish waters: Responses of the submersed plant Pota~nogeto~z pe~Jotiat~,c and its associated algal commanky . Mar.me Ecotogy P~'ogres~ ,Se~es 94:71-82.

NIXON, S. W. 1995. Coastal masine eutrophicat ion: A definition, social causes, and furore concerns. Ophega 41:199-219.

NIXON, S., g~ Btrca~tEY, S GRANGER, ANDJ. BINTZ. 2001. Responses o f very shMlow mar[ae ecosystems to nu t r ien t enr ichment . Human and Ecological Risk A~es~ment 7:1457-1481.

N',daX~N, J. W. 2000. Marine Biology: An Ecological Approach, 5th editiom Benjam[tl Cummings , San Francisco, California.

ORTH, R.J., T. J~ B. C~.r, trruzR% W. C. D~NN~SON, C. M. DUARTE, J. W. FouxquXraN, K L H~c~, J~, A. R. Hvovms, G. A. t~NI)mC~, w . j . KENWORTHY, S. OL'~q~RNIK, E. T. SHORT, M. Wa'oCorr, AND S. L. WILLIAMS. 2006. A global crisis for seagrass ecosystems. Biosde~tce 56:987-996

ORTH, R. J. AND K~ A MOOI~E. 1983. Chesapeake Bay: An Unpreceden ted Declir'~e in Submerged Aquatic Vegetation. Science 222:51-53.

PAINE, R. T. 1966. Food web complexity and species diversity. American Na~ural~st 100:65-75.

PANDOLn, J. j., j . B. C. jaca~SON, N. B.~ON, R. H. B~D~tmY, H. M. GUZMAN, T. P. Hur;l-~S, C r Vr Is F. MaCHELI, J. c . OGDEN, H. P. POSg~q~;HaM, AN~ Er SALA. 2005r Are UrS. coral reefs on the slippery slope to slime? &ience 307:1725-1726.

PAULY, D., V. CHRISTENSEN,J. I).ado%~.~.~RD, R. FROESE, AND F. TORRES Ja. 1998 Fishing down marine f~od webs ,%&'tzc,: 279:860-863.

PA~_rLV, D. AN~_I. MacCsmaN. 2003. In a PetTect Ocean: The State ,?f Fisheries and Ecosystems in the N~r'th Atlantic Ocean Island Press, Washington, D.C.

PAISLu D. AND R. ~4~TSON. 2005. Background and interpretat ion of the "mar ine troph[c index" as a measure of biodiversity. Philosopfli~z~l Tm.*~'aet~ons of the Royal Society B: Biological Scier~ces- 360:415-423.

P~7~SON, C. H. 1979 Predation, competitive exclusion, and diversity in the soft-sediment benth ic communi t ies of estnaries and lagoons, p. 233-261. tn R. J. Livingston (ed.), Ecological Processes in Coastal sod Marine Systems. P lenum Publishing Corp., New York.


PHILLIPS, G. L., D. ElVlINSON, AND B. Moss. 1978. A mechan i sm to account for macrophytes decline in progressively eutrophicated freshwaters. Aquatic Botany 4:103-126.

P~NNXG*U% J. K., N. V. C. PotmvIN, P. Fa~,NCOCm, F. B~maLA~mN~, R. C~'zMELLO, M.-L. HARMELIN-VIVIEN, B ['-[EREtr, M. MILAZZO, M. ZABALA, G. D'ANNA, AND G. PtPtTONE. 9000. Trop[fic cascades in ben th ic mar ine ecosystems: Lessons for' fisheries and protected- area managemen t . Enwronmental Co.nsova~w.a 27:179-200.

PxxISS~R, E. L., D. I. BOLNIr_N, A?,�9 M. F BENARD. 2005~ Scared ro death? The effects o f in t imidat ion and consumpti,_~n in predator-prey interactions. Ecolo~: 80:501-509.

RA~aL~S, N. N. m',~ R. E. 7k~'~v:. 200I. Hypoxia in the nor thern Gulf o f Mexico: Description, causes and change, p~ 1-36. tn N. N. Rabalais and R. E. Tin'Her (eds.), Coastal Hypoxia: Consequences for Living Resources and Ecosystems, Coastal and Estuarine Studies, Volume 58 ~,~nerican Geophysical Union, Washington, D.C.

Rtnz, G. M., J. T. CAm.TON, E. D. G~OS~OLZ, AND A. H. HINES. 1997. Global invasions o f mar ine and estuarine habitats by non- ind igenous species: Mechanisms, extent, and consequences. A rnsrican Zoologist 37:621-632.

Russ, G. R. AND A. C. ALCALA. 2004 Marine reserves: Long-term protect ion is required for full recovery o f predatory fish populat ions. Oecologia 138:622-627

SC~mDL~, D. W. 1998. Replication versus realism: The need for ecosystem-scale experiments . EcosyS'~o,~" 1:323.324.

SCHMITZ, O., K. KIIIVAN, AND O. OV&.DIA ~004~ Trophic cascades: The primacy of trait-mediated indirect interactions. Ecology Letters 7:153-163.

SHEARS, N. T. AND R. C. B~a~cor 2003. Cont inu ing trophic cascade effects after 25 years o f so-rake ruarine reserve pro tection. Marine Ecology Progress' ,5'enos 246:1-16.

Sn~en~a), S. A., A. j, McCo~m, D. A. BtmTmnS, V. N m . z ~ , u s ~ s , D A ST~F~ZrqS~tq, ~,r~ R. W~ST. 1989. Decline of seagrasses, p 34t3-393. l'~ A. W. D. Larkum, A J. MeCornb, and S. A Shephe rd (eds.), Biology of Seagrasses. Elsevier, Amsterdam, The Netherlands.

SHEKI73AN, P. A,xrC~ C. HA't~. 2003..M'e rnangroves nursery bahkat for t ransient fishes and decapods? Wetlands 93:449-458.

SHORT, F. T., D. M. Btnm]cx, AND J. E IL~VLDV. 1995~ Mesocnsm exper iments quantify the effects of eutrophicat ion on eelgrass, Zostera marina. Limnology a~t.d Ocaanograpky 40:740-749.

SHURIN, j. B., E. T. BORER, E. W. S~.~-OO?~, 14..ANDERSON, C. A. BLANCatETTE, B. R. BROITMAN, S. D. COOPER, AND B. HALPERN. 2002. A cross-ecosystem comparison of the s t rength of t rophic cascades. Ecology Letters 5:785-791.

SIH, A., G. ENGLUND, AND D. WOOSTER. 1998. Emergent impacts o f mult iple predators on prey. Tre~ds in Ecology and Evolution 13: 350-355.

SILLIMAN, B. R. AND M. D. BERTNE~R. ~009. A trophic cascade regulates salt marsh primary production~ Proceedings oj the National Academy oj Sciences (U&4) 99:10500-10505.

SILLIMAN, B. R., J. VAN DE KOI'PEL, iV[. D B~TN~SS, L. E. STANTON, AND I. A. M~ND~LSSOHN. 2005. Drought , snails, and large-scale die-off of southern U.S. salt marshes Science 310:1803-1806.

ST~N~Caq R. S. AND J. T. CARLTON. 2001. H u m a n alterations o f mar ine communit ies : Students beware! p~ 445-468. In M. Bertness, S. Gaines, and M. Hay (eds.), Marine Communi ty Ecology. Sinauer Press, Sunderland, Ivlassachusetts.

STENE/_'a<~ R. S. AND E. SALA. 200~. Large marine camb,ores: Trophic cascades and top-down controls in coastal ecosystems pas~ and present, p. 110-173. t~zJ. C Ray, K H. Redford, R. S. Steneek, and J. Berger (eds.), Large Carnivores and the Conservat ion of Biodi~ersib. Island Press, Washington, D.C~

STENEO% R. S., j . VAwaNx< AND A V~ LELAND. 2004. Accelerating trophic-leve[ dysfunction ha kelp forest ecosystems of the western Nor th Atlantic. Ecosystems 7:323-332.

STIBOR, H., O. VADSTTAN, S. DIFA-IL, A. GELZLEIGHTER, f . HANSEN~ F. HANTZSf_THE, A. Ig, ATEG-HAKIS, B. LIPPERT, K. L~SETH, C. PETERS, W. ROEDERER~ M. SANDOW, L. SUNDT-HANSEN, AND f . OLSEN. 2004. Copepods act as a switch between alternative trophic cascades in mar ine pelagic food webs. Ecolo~, Letletx 7:321-328.

STRAUSS, S. Y. 1991. Indirect effects in communib . ecology: The i r definition, study and importance. T~e)zds i~,z Ncdogy a~zd Evoh,.zio~z 6:20@210.

SZ~.IANT, A. M. 2009. NutI ient en r ichment on coral reefs: Is it a major cause of coral reef decline? Estuaries 25:7"t3-766.

THRUSH, S. F., J. E. Hm~TT, V. J. CUMMINGS, AND P. K. DAYTON. 1995. The impact of habitat disrut:bance by scallop dredging on mar ine benthic communit ies: \,~That can he predicted from the results o f exper iments? Mar;He Ecology P~ogress Series 129:141- 150.

TOMASKo, D. A., C.J. DAWES, AN~O M. O. HALL. 1996. The effects of an thropogenic nutr ient e n r i c h m e n t on turtlegrass (Thalassia testudinum) in Sarasota Bay, Florida. Estuaries 19:448-456.

TRUSSLLL, G. C., P.J. EWANCHV~, ,~>m M. D. BERTN~S. 2002. Field evidence of trait media ted indirect effects in a rocky intertidal food web. Ecology Letters 5:241 945~

TRUSSm.L, G. C., P. j. EWANC-HU'K, M. D. BERTNESS, AND B. R. SILUMAN. 2004. Trophic cascades in rocky shore tide pools: Distinguishing lethal and nonlech:~l effects. Oecologia 139:427-- 432.

VADaS, R. L. AND R. W. ELN~. 1992. Pkmt-animal interactions in the north-west Atlan6c, p. 3"3-60 In D. M. John , S.J. Hawkins, andJ . H. Price (eds.], Ptant=',knknal Iureracdons in the Marine Benthos. Clarendon Press, Oxford, U.K

VALENTINE, J. AND J. E. DUFFY. 2006. The central rote o f gt:~zing in seagrass eculogy, p. 463-501. ~rn A~ W~ D. Larknrn, R.J. ot'fl" b and G M Duarte (eds.), Seagrass Biology: A Treatise. CRC Press, Boca Raton, FDdda.

\ZALIELa, Ir I995. Marine EcoIog{caI Processes, 2nd ed i t ion Springer Verlag, New York.

WALTERS, C.J. AND C. S. HOL~aNG. 1990. Large-scale m a n a g e m e n t exper iments and learning by doing~ Ecology 7:2060-2068.

WILLIAMS, S. L. 1988r T]wluss4a testudinum productivity and grazing by green turtles in a highly, disrurbed seagrass bed. Marine Biology 98:447-455.

WmuAMS, S. L. AND K. L. HECX jx. 2001. Seagrass communi ty ecology, p. 317-337. / :n M. D Bermess, M. E. Hay, and S. D. Gaines (eds.), Marine C o m m n m t y Ecology. SinauerAssociates, Sunderland, Massachusetts

WILLIAMS, S. g. AND M. H. RUC-KgLSI<r 1993. Effects o f ni t rogen availability and herbivory on eelgrass (Zostera marina) and epiphytes. Ecology 74:904-918

WILSON, F. S. 1990. Tempora l and ~patial patterns of sett lement: A field study of molluscs in Bogue Sound, Nor th Carolina.Journal oj E~perimental Marine Biolo~ a*~d Ecology 139:201-220.

WOOTFON, J. T. 1994. The nature aud consequences of indirect effects in ecological coninll_lni{ies~ Annual Review oj Ecology and S~stematics 25:443-466.

Woma, B., H. K. DOETZE, AND U. SO~,IMEm 2000. Coastal food web structure, carbon storage, and nkrogen retent ion regolated by c o n s u m e r pressure and nu t r i en t loading. L.~molv�9 a'nd Ocsezr~ogr@h)" 45:339-349.

I~eceived, Nove~e~ 8, 2006 Accepted, Fdoua~y 8, 2007

the primacy of top-down effects in shallow benthic ecosystemsof trophic cascades came fiom studies...

Documents