DIFFERENTIAL IMpACTS OF EChINOID gRAzERS ON CORAL RECRUITMENT

It has been argued that recent phase-shifts towards algal dominance in the Caribbean could have a profoundly negative influence on the future survival of settling corals. In the lower Florida Keys, USA, we used field surveys in areas with recovering populations of the formerly abundant herbivorous black urchin Diadema antillarum (philippi, 1845) and comparatively stable populations of the urchin Echinometra lucunter (Linnaeus, 1758) to assess differences in the benthic communities associated with these two alternating urchin dominants. Two primary questions were addressed: (1) Are the effects of urchin grazing on benthic community composition differential based on urchin identity? and (2) how are these differences affecting coral recruitment and long-term success? Within the Diadema zone, urchin densities approached those of pre die-off levels and were associated with low macroalgal biomass and elevated coral cover. These findings are consistent with those of numerous prior investigations, and suggest that Diadema grazing is integral to the maintenance of low-biomass high-turnover algal turfs and prostrate coralline algal crusts that characterize healthy reef systems. Conversely, and despite orders of magnitude greater abundance, Echinometra was associated with chemically and morphologically defended macroalgal forms (i.e., Halimeda sp. and Dictyota sp.) and a benthos dominated by algal canopies with virtually no live coral cover. Overall, our results are concordant with those of other recent studies in the Caribbean and suggest that in the lower Florida Keys D. antillarum may be the strongest interactor in a diverse guild of invertebrate and vertebrate grazers. The direct effects of grazer density on the abundance and growth of marine macroalgae are reasonably well understood. For example, trophic cascades leading to alternating densities of urchins and kelp in the Eastern pacific are responsible for large-scale fluctuations in kelp biomass and, consequently, both the biodiversity and functioning of these ecosystems (Estes et al., 1998; halpern et al., 2006). Similarly large impacts on plant biomass have been demonstrated many times for urchins grazing on algae associated with coral reefs (Ogden et al., 1973; Carpenter, 1981; Belliveau and paul, 2002). Less is known, however, about the indirect effects of grazermediated disturbances on alternative space occupiers, such as scleractinian corals. Although removal of algal biomass increases the amount of substrate available for settlement by colonial invertebrates, past investigators have suggested that scouring caused by urchin feeding may also be responsible for appreciable rates of bioerosion (hughes et al., 1985; Carreiro-Silva and McClanahan, 2001; griffin et al., 2003) and might contribute to coral death, particularly for juvenile corals (Sammarco, 1980). Understanding the density-dependent relationship between grazing pressure and community composition is particularly important for coral reef systems, where recent declines in both grazing pressure and coral cover have been associated with increases in algal biomass stocks (McCook et al., 2001; Szmant, 2002). In the Caribbean, declining coral cover and increasing macroalgal biomass have prompted considerable debate regarding the resilience of coral communities that are exposed to anthropogenically-driven disturbance regimes (hughes et al., 1999; McCORAL REEF pApER BULLETIN OF MARINE SCIENCE, VOL. 85, NO. 2, 2009 122 Cook, 1999; Nyström et al., 2000; Szmant, 2002; Nedimyer and Moe, 2003; Aronson and precht, 2006). McCook et al. (2001) have argued that phase-shifts towards algal dominance could have had a profoundly negative influence on the survival of recently settled corals, through shading, overgrowth, and physical disturbance, and may be a factor limiting the diversity and abundance of juveniles beneath algal canopies. Despite recent reports that recovering populations of the herbivorous black urchin, Diadema antillarum (philippi, 1845) can precipitate small-scale reductions in algal cover (Aronson and precht, 2000; Edmunds and Carpenter, 2001; Moses and Bonem, 2001; Miller et al., 2003; Nedimyer and Moe, 2003; Carpenter and Edmunds, 2006), surprisingly few studies have examined the impacts of: (1) localized increases in grazing intensity, (2) grazer identity or (3) protracted periods of algal dominance on the recruitment dynamics of hermatypic corals. Edmunds and Carpenter (2001) supplied a first-order approximation of these influences by examining the effect of recently recovered populations of D. antillarum on several stretches of forereef in Jamaica. They reported substantial reductions in algal cover and an eleven-fold increase in juvenile coral abundance where D. antillarum was abundant and suggested that further recruitment by this urchin could eventually lead to a reversal of algal dominance and a return to scleractinian health (Edmunds and Carpenter, 2001). Recent surveys conducted by this group in 2003 and 2004 supported these findings with nearly identical patterns of echinoid density and algal abundance emerging at sites scattered throughout the Caribbean basin (Carpenter and Edmunds, 2006). Using a similar approach in the lower Florida Keys, USA, we assessed differences in benthic communities associated with two co-occurring species of urchin: D. antillarum and Echinometra lucunter (Linnaeus, 1758). By haphazardly choosing adjacent patches with alternating urchin dominance, we were able to indirectly assess the influence of grazer identity on benthic species composition. past studies have demonstrated that these two species differ markedly in their foraging range and control of benthic production, with E. lucunter exhibiting limited potential for grazer control even at densities as high as several hundred per square-meter (grunbaum et al., 1978; Lawrence and Kafri, 1979; Schneider, 1985). We predicted that these differences would result in benthic assemblages that were differentially affected by the urchin species present, and that coral recruitment success would be negatively associated with algal standing stock. In this study, two primary research questions were addressed: (1) Are the effects of urchin grazing on benthic community composition differential based on urchin identity? and (2) how are these differences affecting coral recruitment and long-term success? Materials and Methods Study Site and Depth.—This study was conducted during July 2004 in Bahia honda State park on the south side of Little Bahia honda Island (24°39 4́7 ̋N, 81°15 ́50 ̋W, located 0.8 km southwest of Bahia honda Key). here, consolidated substrate resembled a small-scale spur and groove arrangement with low-relief platforms separated by basins of accumulated sand and detritus. platforms varied from 1 to 4 m in depth and were frequently exposed to high rates of flow and surge as tidally driven exchanges with Florida Bay occurred to the west, through Bahia honda Channel. Benthic community structure was dominated by macroalgae (predominantly, Halimeda spp. and Dictyota spp.) with sparse cover of hard coral, mostly FURMAN AND hECK, JR.: DIFFERENTIAL IMpACTS OF EChINOID gRAzERS ON CORAL RECRUITMENT 123 Porites astreoides Lamarck, 1816. Within this bathymetric range, and extending some 20 m to the southeast, there existed small, largely monospecific aggregations of D. antillarum or E. lucunter. These urchin aggregations were readily visible and are hereafter termed “Diadema” and “Echinometra” zones, respectively. Three such patches per species were selected for study. however, due to inclement weather and strong wave action, only two Diadema patches were successfully surveyed. Depth was determined as an average of both transect ends (see below) and varied between 1.5 and 2.3 m, with an average depth (± SE) of 1.91 ± 0.38 and 1.73 ± 0.10 m for Diadema and Echinometra zones, respectively. Echinoid Size Distributions.—Urchin abundance was visually estimated along a 5-m transect that was haphazardly positioned within each of the patches (N = 5); these transects were used in all subsequent benthic assessments (see below). All urchins within 0.5 m of the transect line were identified and enumerated. The test diameters of 55 D. antillarum were measured from the oral hemisphere using a Scienceware Measy 2000 Type 5921 caliper. These 55 individuals were haphazardly collected from the site prior to delineation of the urchin zone and were removed for use in another experiment. Reductions of this nature are not believed to have been significant to the remaining distribution or abundance of D. antillarum for two reasons: (1) hand-collections were made over a larger area than the one considered here, and (2) subsequent identification of zones was based on visual assessments made at the time of survey, making any artificially defaunated patches unlikely to have been chosen as representative of our treatment levels. Benthic Cover.—Aerial coverage of macroalgae and live coral were directly assessed with 0.0625-m2 photoquadrats positioned on both sides of the transect at 1-m increments (n = 10 per transect). Images were taken with an Olympus Stylus 400 digital camera set to its highest resolution and steadied for parallax error with a pVC-constructed tripod. Raw images were adjusted using Auto Levels in photoshop 7.0 (Mac OS X) and outlined groupings of macroalgae or live coral cover were created using the available selection techniques. Resulting layers were converted to two-tone images (black and white) and exported individually to Carnoy 2.1 (Mac OS X) for area calculation. Juvenile and Small Coral Communities.—We operationally defined juvenile corals to be colonies < 4.0 cm in diameter (Bak and Engel, 1979; Carpenter and Edmunds, 2006) and of taxa not known to sexually reproduce at that size (Edmunds and Carpenter, 2001; Carpenter and Edmunds, 2006). Colonies in violation of the latter property were termed ‘small’ species and considered separately during anlaysis. All colonies meeting the size criterion were visually located within 0.0625-m2 quadrats (discussed above), identified to the lowest possible taxonomic level, enumerated, and measured (in situ). Size was estimated as the average of the two longest colony diameters (± 0.01 cm). Statistical Analyses.—Multivariate analyses of benthic cover, juvenile coral abundance, small coral abundance, and total colony abundance (i.e., “small” and “juvenile” taxa) were performed independently using the pRIMER 5 (Windows 98) software for ANOSIM (analysis of similarities) and SIMpER (similarity percentages). Multidimensional scaling (MDS) of species-level abundance data was based on resemblance matrices of square-root transformed, Bray-Curtis similarities. global R statistics were assessed for significance at p = 0.05. parametric (univariate) statistical analyses were not performed due to the uneven sample sizes. however, factor plots (mean ± SE) of individual metrics were used to assess relative effect size and general trends. Wherever possible, one-way comparisons were conducted using nonparametric Mann-Whitney U tests.