Does depth and sedimentation interact with sea urchins to affect algal assemblage patterns on eastern atlantic reefs?

Journal of Shellfish Research, Vol. 28, No. 4, 947-955, 2009. DOES DEPTH AND SEDIMENTATION INTERACT WITH SEA URCHINS TO AFFECT ALGAL ASSEMBLAGE PATTERNS ON EASTERN ATLANTIC REEFS? LEONOR ORTEGA-BORGES,l FERNANDO TUYA2 ,3* AND RICARDO J. HAROUN1 lBIOGES, Campus de Tajira s/n, University 01Las Palmas de G.c., 35017 Las Palmas de G.c., Canary Islands, Spain; 2 CIlM AR, Rua dos Bragas 289, 450-123 Porto, Portugal; 3Centrelor Marine Ecosystem Research, Edith Cowan Un iversity,Joondalup, 6027 Western Australia, Australia ABSTRAeT Arange of factors may affect the composition and abundance of macroalgae on subtidal rocky reefs. We experimentally determined the interactive effect of the occurrence of the long-spine sea urchin, Diadema antillarum, depth and sedimentation levels on macroalgal assemblage structure on eastern Atlantic rocky reefs. Specifically, we manipulated sea urchin densities (removal of all individuals vs. untouched controls at natural densities) on rocky reefs devoid of erect vegetation, and predicted (1) that removal of sea urchins would differently affect macroalgal assemblage structure between deep (16-18 m) and shallow (8-9 m) reefstrata, and that (2) the effect of sea urchin removal on macroalgae would be altered under different scenarios of sedimentation (ambient vs. enhanced). Experimentalcircularplots (2 min diameter) were set up at 3locations at Gran Canaria (Canarian Archipelago), and were maintained and monitored every 4wk for 1 y. At the end of the experimental period, the structure of the algal assemblages differed between urchin treatments and depth strata, with alarger cover of turf and bushlike algae where urchins were removed and at the shallow reef stratum. More important, differences in algal assemblage structure between urchin treatments were irrespective of sedimentation levels, but shifted from the shallow to the deep stratum. This interactive effect was, in turn, observed for bushlike algae, as aresult of alarger magnitude of response (i.e., larger cover) in the shallow stratum relative to thedeep stratum, but was not detected for either turf or crustose coralline algae. These results highlight the importance of sorne physical conditions (here, differences in depth) to interact with biotic processes (here, urchin abundance) to create patterns in the organization of subtidal and benthic assemblages. KEY WORDS: macroalgae, benthic assemblages, assemblage structure, sea urchins, sedimentation, depth strata INTRODUCTION Subtidal temperate habitats are heterogeneous. Arange of factors may facilitate or disturb the presence and abundance of subtidal macroalgaeliving on hard bottoms, and thereby affect their composition and assemblage structure (Lüning 1990). For example, light, sedimentation, grazing, and water fiow have been repeatedly highlighted as important determinants of macroalgal assemblage structure (Schiel &Foster 1986, Witman & Dayton 2001). In turn, an appreciation of the interactive effects of physical and biological processes may assist in the development of predictive models about the assembly and maintenance of heterogeneity of subtidal communities (Connell 2005). Depth directly affects different abiotic properties, such as light availability, nutrient concentrations, sedimentation, and temperature (Garrabou et al. 2002). For example, an increase in depth is associated with adecrease in light availability and typically in algal growth, which affect the community dominant (Witman &Dayton 2001). Aconsiderable small-scale horizontal spatial variability (from centimeters to meters) in the structure and abundan ce of algal assemblages clearly indicate that other mechanisms are also relevant to infiuence patterns in algal abundance and assemblage structure on subtidal reefs (Fowler-Walker &Connell 2002, Schils &Coppejans 2003, Terlizzi et al. 2007). For example, sediment deposition may disturb, and so alter, algal diversity and abundance at small scales (Schiel &Foster 1986, Airoldi &Virgilio 1998), often indirectly facilitating opportunistic filamentous turfs to replace adversely affected canopy-forming, erect, macroalgae (Airoldi *Corresponding author. E-mail: [email protected]. &Cinelli 1997, lIving &Connell 2002), although sediment disturbance may promote diversity through preventing monopolization of space bycompetitively superior species (Littler et al. 1983b, Airoldi, 2003). The presence of high sea urchin population densities has been widely advocated as akey mechanism determining the organization and functioning of temperate reefs: Overgrazing by sea urchins may cause the elimination of erect macroalgae, ultimately turning algal beds into "urchin-grazed barrens" dominated by algal crusts and sessile invertebrates (e.g., Dayton et al. 1992, Andrew 1993, Andrew &Underwood 1993, Shears &Babcock 2003, Graham 2004, Tuya et al. 2004b). The intensity of sea urchin grazing, and subsequently its effects on the organization of benthic assemblages, is, however, considerably heterogeneous through space and time (Benedetti-Cecchi et al. 1998). Arange of biotic and abiotic mechanisms can alter sea urchin densities, and so grazing rates. For the former, shifts in the behavior (Sala et al. 1998) and recruitment (Ebert 1983, Balch &Scheibling 2000) patterns are two of the main studied mechanisms, whereas for the latter, differences in turbulence, wave action, substrate rugosity, and heterogeneity are factors routinely advocated to infiuence the effect of sea urchin grazing over algal assemblages (Lawrence 2001). For example, water turbulence typically decreases with increasing depth (Denny 1988, Roberts et al. 2006), and may therefore increase the susceptibility of erect algae to sea urchin grazing (Alves et al. 2001, Tuya et al. 2007, Shears et aL, 2008). Similarly, the negative effect of sea urchin grazing over erect macroalgae can be exacerbated under scenarios ofincreased sediment loads that facilitate opportunistic filamentous algae through an inhibition of the recruitment of erect macroalgae (Valentine & Johnson 2005), although the effect of urchins on macroalgae can be 947 948 ORTEGA-BORGES ET AL. reduced when high levels of sedimentation have adverse effects on larval and postsettlement survival of sea urchins (Shears et al. 2008). The long-spine black sea urchin, Diadema antillarum (Philippi), is agregarious echinoid that occurs in almost all marine habitats in the shallow subtidal across the warm temperate waters of the eastern Atlantic, from Madeira to the Gulf of Guinea. In this region, this species plays akey role in structuring subtidal rocky reefs (Alves et al. 2001, Tuya et al. 2004b, because it is directly involved in the transformation of large reef areas previously covered by erect algae to barren s, following the classic trophic cascades paradigm (Sala et al. 1998). Indeed, the contribution of other grazers (e.g., herbivoraus fishes) to the generation of reef barrens is negligible when compared with the grazing caused by D. antillarum (Tuya et al. 2004a). Potential interactions between different factors are often ignored in tests of hypotheses about the structure of benthic assemblages, although they might clarify asubstantial amount of the spatial variability among whole assemblages. In this study, we experimentally determined the interactive effect between the occurrence (absence vs. presence) ofthe long-spine black sea urchin, D. antillarum, and differences in depth and sedimentation (ambient vs. enhanced) to create patterns in the assemblage structure of macroalgae on eastern Atlantic rocky reefs. Specifically, we manipulated sea urchin densities and predicted (1) that removal of sea urchins would differently affect macroalgae assemblage structure between deep and shallow reefstrata, and that (2) the effect of sea urchin removal on macroalgae assemblage structure would be altered under different scenarios of sedimentation (enhanced vs. natural); in particular, we predicted that erect, bushlike algae would be more affected by urchin grazing than filamentous turfs, where sediment loads are heavier. MATERIALS AND METHODS Study Area The study was carried out at 3locations, 1-2 km apart, on the northeast coast cf Gran Canaria (The Canaria nArchipelago, 28°N, eastern Atlantic Ocean), fram March 2007-March 2008. Each location encompasses racky (basaltic) reefs denuded ofvegetation ("urchin-grazed barrens") from approximately 420 mdepth, and are exposed to the prevailing swells and seas from the northwest and northeast, respectively. Oceanographic variables (i.e., current patterns, sea surface temperature, exposure to waves) are typically similar from one location to the other (González Barbuzano 2003). Sea surface temperature fluctuates from l8-l9°C in winter to 23-24°C in summer. In this area, the distribution of benthic communities along the bathymetric axis shows aconsistent vertical zonation pattern. Within the shallowest zone (0-3 m), extensive stands of algal assemblages-principally, bushlike algae (Tuya & Haroun 2006)-dominate the community, and long-spine sea urchins only occur in low densities (0-1 individualsjm2 ). Intensive grazing by D. antillarum produces clear interfaces between these shallow-water algal stand s and deeper areas devoid of vegetation (fram below 3 m of depth, long-spine sea urchin mean densities typically range from 4-6 individualsjm2[Tuya et al. 2004b]). Interactive Effects Betli'een Sea Urchin Presence and Depth: Experimental DesiKn At each location, circular plots (diameter, 2m) were created by either removing a11 urchins (-U) or preserving urchin densities at natural levels (i.e., untouched controls, +U). A11 plots were set up on horizontal reef surfaces with similar densities of urchins (ranging from 4-6 individualsjm2 ), and devoid of erect vegetation. The center of each experimental plot was marked with ametal stake drilled into the reef and labeled for subsequent identification. Two plots of each treatment (-U and +U) were established at 2depth strata-deep (16-18 m) versus shallow (8-9 m)-within each location (i.e., 8experimental plots per location). A11 plots were created at the beginning of March 2007, and were visited approximately every 4wk to maintain treatments and to quantify changes in assemblage structure, until the end ofthe experiment (March 2008). Because adult D. antillarum individuals show aclear "homing behavior" (Tuya et al. 2004c), colonization of urchin removal plots (~U) was mostly restricted to juveniles (personal observation) throughout the study, which hide in cryptic reef microhabitats such as cracks and crevices (Hernández 2006). Immigration of adult D. antillarum into urchin removal plots (-U), at any time during the 4-wk period, was, in all cases, less than or equal to 2 individuals per plot, and individuals were removed. Interactive Effects Betli'een Sea Urchin Presence andSedimentation: Experimental DesiKn The second experiment was conducted at the same locations using the same methods described for the previous experiment, but was only conducted at asingle depth stratum (l2~14 m). At each location, 2plots of each urchin treatment (-U and +U) were subjected, every 4wk, to an "enhanced" sedimentation treatment by adding sandy (0.28 mm mean grain diameter) sediments (40.44 gwet weight per experimental plot) as afine "rain" (Airaldi &Virgilio 1998, Connell 2005) uniformly distributed within each plot. The other 4plots were maintained at "natural" (i.e., unmanipulated) levels of sedimentation. To determine aproxy to the level of sedimentation in the study area, 4funnel-like sediment traps (63.58 cm2 of upper circular area) were established, 30 cm aboye the bottom at 1location, and the mean accumulation rate (measured in grams wet weight per square meter per'day) of sediments was quantified every day for 4successive days. Those plots assigned to the "enhanced" sedimentation treatment were subjected to a3-fold increase in sedimentation rates (0.67 gwet weightjm2jday) relative to natural sedimentation levels (0.22 gwet weightjm2jdayl). Such an increase was selected to represent arealistic scenario in the shifts of sediment loads within the study area (González Barbuzano 2003), which lacks major runoff sources (e.g., rivers, creeks) throughout the year. SamplinK For both experiments, and at each sampling time (i.e., every 4wk from March 2007 to March 2008), ascuba diver quantified in situ the percent cover of algae in four 50 X50-cm quadrats (0.25 m2)within each plot, following point-quadrat pracedures previously implemented for the study area (a grid of 121 points per quadrat was used in all cases [Tuya and Haroun 2006]). Quadrats, tens of centimeters apart, were haphazardly laid out ASSEMBLAGE PATTERNS ON ATLANTIC REEFS 949 within each plot. This is arapid, nondestructive technique to assess assemblage structure and dominance of sessile biota (Fowler-Walker &Connell 2002, McClanahan et al. 2003). Final values for each taxon were expressed as percentages. Taxa presented in less than a 4% cover were omitted. Responses of algae can be identified, without loss of significant information, through taxonomic groups instead of species (Terlizzi et al. 2003). Macroalgae were thus categorized into 3morphological groups, by taking into account the algal form groups reported in the literature (Steneck &Dethier 1994, Fowler-Walker & Connell 2002, McClanahan et al. 2003; Vaselli et al. 2008), and previously implemented in the study area (Tuya and Haroun 2006). Turf algae (hereafter referred to as TA) consist of small cushion-shaped and filamentous species, usually less than 5cm in height (e.g., Colpomenia sinuosa, Dasycladus vermicularis, Jania spp., Ceramium spp.). Bushlike algae (hereafter referred to as BA) are erect, coarsely branched algae (e.g., Asparagopsis spp., Corallina elongata, Dyctiota spp., Padina pavonica, Stypocaulon scoparium, Stypopodium zonale, Zonaria tournefortii), from 1-15 cm in height, which constitute either large algal cushions or thin sheets. Finally, coralline algae (hereafter referred to as CA) consisted of algal crusts (e.g., the genera Lithothamnion, Lithophyllum, Neogoniolithon, Titanoderma), and were counted when not overgrown by other algae. Statistical Analyses For each experiment, differences between categories (and combinations) ofpredictive factors were evaluated using multiand univariate analysis of variance (ANOVA). Permutational analysis of variance (PERMANOVA [Anderson 2001]) was used to partition both multivariate and univariate variability, because this approach allows multivariate testing for interactions, and uses permutations to calculate Pvalues. The latter was preferable because the data were from unknown distributions and were overdispersed. The test statistic (pseudo F) is a multivariate analogue of the univariate Fisher's Fratio, and in the univariate context the two are identical when using Euclidean distance as the dissimilarity measure (Anderson 2001). Because data collected from the same experimental units through time are often autocorrelated, and so assumptions of analyses may not be met, we analyzed data at the end of the experimental period according to amixed effects, 4-factor ANOV Amodel, incorporating (l) location (random factor with 3levels), (2) urchins (fixed factor with 2levels; -U vs. +U and orthogonal to the previous factor), (3) depth (deep vs. shallow strata) or sedimentation (ambient vs. enhanced) (fixed factor with 2levels, and orthogonalto the previous factors), and (4) plots (random factor with 2levels, nested within the 2nd order interaction between location, urchin, and depth or sedimentation). Raw data were square root transformed to down-weight the most abundant taxa, and multivariate (the entire macroalgal assemblage) and univariate (percent coverage of TA, BA, and CA) analyses were based on Bray-Curtis dissimilarities and Euclidean distances, respectively. Pvalues were calculated from 4,999 unrestricted permutations ofthe raw data. When appropriate, pairwise aposterioricomparisonswere executed using 4,999 permutations to test for differences among levels of fixed factors. To visualize multivariate patterns, nonmetric multidimensional scaling ordination was carried out on the square root transformed data. Differences in variability (dispersion) in assemblage structure between treatments were tested through pairwise comparisons with the PERMDISP routine (Anderson 2004) via 4,999 permutations of the raw data. All analyses were performed using the PRIMER 6.0 (PRIMER-E Ltd, UK) statistical package. RESULTS Effects of Urchin Removal on Algal Assemblages Between Shallow and Deep Strata Differences in algal assemblage structure between urchin removal and untouched (control) plots at the end of the experiment shifted from the shallow to the deep stratum (Table 1, PERMANOVA, Ur XDe, P<0.05); this interactive effect was consistent among locations (Table 1, PERMANOV A, Lo X Ur XDe, P>0.1), although the effects of urchins and depth shifted among locations (Table 1, PERMANOVA, Lo X Ur and Lo XDe, P<0.01). In general, the structure of the algal assemblages at the shallow stratum was more variable (i.e., dispersed) relative to those algal assemblages at the deep strata TABLE 1. Results of multivariate and univariate ANOVA testing the effects of location (random factor), urchins (fixed factor with 2levels and orthogonal to the previous factor), depth (fixed factor with 2levels and orthogonal to the previous factors), and plots (random factor with 2levels, nested within the 2nd order interaction between location, urchins, and depth) on the structure of algal assemblages and the percent coverage of each algal morphological group at the end of the experimento AIgal Assemblage Turf AIgae Bush AIgae Crustose AIgae Source df MS FP MS FP MS FP MS FP Lo (location) 217,479 39.37 0.0002 0.41 2.25 0.1475 66.21 10.11 0.0027 4.35 12.28 0.0013 Ur (urchins) 125,678.2 5.77 0.0462 39.47 37.99 0.0253 799.26 24.49 0.039 0.09 0.86 0.4513 De (depth) 159,751 4.75 0.0762 52.01 45.95 0.0211 1186.76 11.87 0.0478 18.4 4.08 0.1808 P10t (Lo X Ur XDe) 12 443.94 2.58 0.0004 0.18 3.97 0.0001 65.521 5.60 0.0000 0.35 1.69 0.0866 Lo X Ur 24,448.1 10.02 0.0002 1.03 5.71 0.0181 32.00 4.98 0.0266 0.10 0.31 0.7413 Lo XDe 212,581 28.33 0.0002 1.14 6.22 0.014 95.57 14.59 0.0006 4.52 12.78 0.0011 Ur X De 15,863.4 11.30 0.0268 2.09 7.94 0.1062 341.26 11.74 0.0757 1.01 1.85 0.3072 Lo X Ur X De 2518.8 1.16 0.3696 0.26 1.45 0.2723 29.07 4.44 0.0361 0.54 1.55 0.2529 Residual 72 172.27 0.04 1.17 0.20 950 ORTEGA-BORGES ET AL. (Table 2, PERMDISP, most pairwise comparisons, P<0.01), which, indeed, tended to cluster together in the ordination space (Fig. 1). For TA, the direction and magnitude of differences between urchin treatments (-U vs. +U) was consistent between depth strata(Tablel,Ur X DeandLo X Ur XDe, P>O.lO,Fig. 2). We detected differences in the percent coverage between urchin treatments (-U vs. +U) and depth strata (Table 1, urchin and depth, P<0.05, Fig. 2). Those plots where urchins were removed (-U) contained 3.1 times more algal cover than untouched (control) plots (+U), whereas plots at the shallow reef stratum contained 1.6 times more algal cover than the deep reef stratum (Fig. 2). These differences between urchin treatments and depth strata, however, differed slightly in magnitude amonglocations (Table 1, Lo XUr, Lo XDe, P<O.OS, Fig. 2). In the case of BA, differences between urchin treatments varied from the deep to the shallow reefstratum (Table 1, Ur X De, P=0.07, Fig. 2), as aresult of alarger magnitude of response (i.e., larger percent of coverage) in the shallow relative to the deep stratum (Fig. 2), which even varied among locations (Table 1, Lo X Ur XDe, P<0.05, Fig. 2). We detected differences in the percent coverage between urchin treatments (-U vs. +U) and depth strata (Table 1, urchin and depth, P< 0.05, Fig. 2). Those plots where urchins were removed (-U) had 3.4 times more cover of algae than plots where urchins were not eliminated (+U), whereas those plots at the shallow reefstratum had 5.7 times more cover of algae than the deep reef stratum (Fig. 2). These differences between urchin treatments and depth strata differed in magnitude among locations (Table 1, Lo X Ur, Lo XDe, P<0.05, Fig. 2). Finally, and for CA, location no. 1contained alarger percent coverage than the other locations at the end of the experiment (Table 1, location, P<0.01, Fig. 2), particularly at the shallow stratum, which resulted in an interactive effect between these 2factors (Table 1, Lo XDe, P<O.OI, Fig. 2). No other significant effects were detected. Effects o.f Urchin Removal on Algal Assemblages Between High and Low Leve/s oI Sedimentation 2D Stress: 0,04 • • •• o o O •+UShallow •-UShallow A+UDeep O-UDeep Figure 1. Two-dimensional nonmetric multidimensional scaling plot showing centroids of algal assemblages that were subjected to sea urchin removals (circles) or preserved at natural densities of sea urchins (untouched controls, triangles) at shallow (open symbols) and deep (filled symbols) strata at the end of the experimental periodo Replicated plots have been averaged within each treatment for each of the 3locations. treatments (all pairwise comparisons, P>0.05, PERMDISP, Table 4, except between -U -sediments and +U +sediments). For both TA and BA, we observed asignificantly larger cover, at the end of the experiment within those plots where urchins were removed (Table 3, urchins, P<0.05, 2.1 times for TA and 2.5 times for BA, Fig. 4), irrespective of levels of sedimentation (Table 3, Ur XSe, P> 0.05, Fig. 4). Differences between levels of sedimentation did not cause asignificant change in the cover of both algal groups (Table 3, sedimentation, P> 0.05, Fig. 4). Finally, location no. 1contained more CA than the other locations at the end of the experiment (Table 3, location, P<0.01, Fig. 4), which even resulted in an interactive effect with urchin treatments (Table 3, Ur XLo, P<0.05, Fig. 4). No other significant effects were detected. DISCUSSION TABLE 2. Results of pairwise comparisons of multivariate dispersion between sea urchin treatments (removals [-U] vs. untouched controls [+U)) from shallow to deep strata. The effect of urchin manipulations (-U vs. +U) on algal assemblage structure was irrespective of sedimentation levels at the end of the experiment (Table 3, PERMANOVA, Ur X Se, P> 0.05) (Fig. 3), although it varied among locations (Table 3, PERMANOVA, Lo XUr, P<0.05). Moreover, differences in dispersion in assemblage structure were not observed among Groups +U shallow versus -U shallow +U shallow versus +U deep +U shallow versus -U deep -U shallow versus +U deep -U shallow versus -U deep +U deep versus -U deep 51.606 61.981 35.019 29.805 0.61471 22.442 p 0.001 0.001 0.009 0.012 0.724 0.10 Ej]'ects oI Urchin Remova/on Alga/ Assemb/ages Between Shallow and Deep Strata The key role that sea urchins play in determining algal assemblage structure has been experimentally demonstrated throughout temperate latitudes (e,g., Dayton et al. 1992, Andrew 1993, Benedetti-Cecchi et al. 1998, Alves et al. 2001, Bulleri et al. 2002, Graham 2004). Typically, an increase in the cover of erect algae, including both turf and bushlike algae, within plots where sea urchins were eliminated (or reduced) has been observed, which is consistent with our observations from subtidal rocky reefs at Gran Canaria. Similarly, differences in the composition and structure of erect macroalgae between bathymetric strata have been observed at temperate subtidal reefs (e.g., Garrabou et al. 2002, Balata and Piazzi 2008, Vergés et al. 2009). More important, our study demonstrated that the magnitude of differences in algal assemblage structure between plots subjected to urchin manipulations (removal vs. untouched controls) shifted from shallow to deep waters. These results highlight, therefore, the capacity for physical conditions (i.e., depth) to interact with biotic processes (i.e., urchin abundance) ASSEMBLAGE PATTERNS ON ATLANTIC REEFS 951 Location 3 Location 3 Location 2 Location 2 Location 1 Location 1 - Deepreefs .. Shallow reefs 20 80 ~---------------, ,-----------------, o 80 <1> 60 ca C) Cü .:= 40 tIJ ~ ttl ~ 20 o O 80 Figure 2. Mean percentcoverage of each morphological group subjected to sea urchin removal (-U) or preserved at natural densities of sea urchins (+U) at shallow (gray bars) and deep (black bars) strata at the end of the experimental periodo Error bars are ± SE of means (n =8). to create patterns in the organization of subtidal, benthic assemb1ages. Differences in depth not only maintain differences in the composition and structure of algal assemb1ages, but can a1so give rise to divergence during succession when the densities of major grazers are a1tered. In this context, bushlike a1gae increased more in abundance (i.e., cover) in the shallow than deep waterswhen urchins were removed; a pattern that was not observed for turf algae. Such aresult demonstrates the capacity for depth to impose astrong synergistic effect that facilitates the assemb1y and maintenance of bushlike a1gae in shal10w waters. Typical1y, large patches where major grazers are excluded are colonized by arange of a1ga1 species, main1y via arrival of propagules from adjacent populations (Airoldi 2000). Differences in their abilities to survive and grow subsequent1y affect patch colonization. Because an increase in depth is linked with a decrease in light availability, and then adecline (limitation) in algal growth, the results of our experiment suggest that growth is more light limited for bushy a1gae relative to turf algae. This perception is in agreement with their different morphologies and eco10gies (Steneck &Dethier 1994, McC1anahan et al. 2003), and could partially explain the comparatively 1arge increase in the cover of bushlike a1gae where sea urchins were removed, from the deep to the shallow stratum. In general, a fi1amentous (or acushion-shaped) morphology is more efficient than abushy one to capture photosynthetically active radiation (PAR) light (Littler et al. 1983b; for example, self-shading is more limited relative to bushlike algae (Littler &Arnold 1980). Moreover, it is a1so possib1e that proximity to the source of propagule donors could explain sorne of these observations. Because extensive stands of bushlike algae dominate the community within the shallowest zone (0-3 m) in the study area (Tuya & Haroun 2006), these zones could supply propagu1es to the immediately adjacent strata (i.e., the shallow stratum in our experiment), because the dispersion of a1ga1 propagules generally shows asharp, exponential decrease severa1 meters away from the algal donor (Santelices 1990, Kendrick &Walker 1991, Coleman 2003). Obviously, several other potential factors (or even combinations) that vary with depth (e.g., wave-induced turbulence) cou1d also affect each group differentially. Further experimental approaches are necessary to unravel the various proposed mechanisms behind the observed patterns. The lack of response observed for encrusting coralline algae may be exp1ained by the short duration of the experiment. It is likely that both turf and bushlike algae did not have sufficient time to monopolize the entire available space within plots, which prevented the competitive exclusion of encrusting coralline algae. In addition, encrusting coralline algae are able to reduce settlement of 952 ORTEGA-BORGES ET AL. TABLE 3. ResuIts of muItivariate and univariate ANO VA testing the effects of location (random factor), urchins (fixed factor with 2levels and orthogonal to the previous factor), sedimentation (fixed factor with 21evels and orthogonal to the previous factors), and plots (random factor with 2levels nested within the 2 nd order interaction between locations, urchins, and sedimentation) on the structure of algal assemblages and the percent coverage of each algal morphological group at the end of the experimento Algal Assemblage Turf Algae Bush Algae Crustose Algae So urce dI MS FP MS FP MS FP MS FP Lo (lacatian) 25,414 42.78 0.0002 1.09 4.73 0.0306 6.70 18.43 0.0002 483.29 119.90 0.0001 Uf (urchins) 15,649.7 7.21 0.0310 13.65 38.65 0.0249 27.33 28.82 0.0330 49.59 2.27 0.2711 Se (sedimentatian) 11.861.4 5.32 0.0924 0.05 0.14 0.7485 0.67 4.12 0.1794 12.76 1.08 0.4075 Plat (Lo x Uf XSe) 12 126.5 0.81 0.7261 0.23 4.59 0.0000 0.36 2.51 0.0081 4.03 1.16 0.3265 Lo X Uf 2784.1 6.19 0.0002 0.35 1.53 0.2564 0.94 2.60 0.1150 21.87 5.43 0.021 Lo XSe 2349.7 2.76 0.0033 0.38 1.68 0.2275 0.16 0.45 0.6481 11.79 2.93 0.0923 Uf XSe 11,303.9 1.63 0.2876 1.29 1.47 0.3489 4.35 3.03 0.2237 0.84 0.05 0.8487 Lo X Ur XSe 2800.93 6.32 0.0004 0.88 3.81 0.0524 1.13 3.13 0.0982 14.00 3.47 0.1055 Residual 72 156.7 0.05 0.14 3.46 poten ti al competitors, suggesting that they do not rely on grazing by herbivores (e.g., sea urchins) to prevent acomplete swamping by erect algal species (Bulleri et al. 2002). Finally, results from this experiment indicated that the structure of the algal assemblages at the shallow stratum was considerably more variable compared with algal assemblages at the deep stratum. These results support the notion that shallow environments are typically more variable, and more extreme, than deeper ones (Garrabou et al. 2002). Indeed, shallow habitats are notoriously affected by fiuctuations in asuite of environmental factors, such as seawater temperature, turbulence and exposure to hydrodynamic forces, salinity, light, and so forth, through annual cycles (Denny 1988). As aresu1t, substantial f1uctuations in such arange offactors induce ahigh degree of environmental, and thus ecological, heterogeneity. Effects o/ Urchill Removal 011 Algal Assemblages Betweell High alld Low Levels oI Sedimelltatioll An increase in the cover of both turf and bushlike algae was detected, again, where sea urchins were eliminated. However, patterns in the structure of these assemblages were irrespective of sedimentation levels. Filamentous algae may tolerate heavy sediment accumulations (Airoldi &Virgilio 1998, Airoldi 2003, Connell 2005). Similarly, encrusting coralline algae have been also experimentally observed to be unaffected by large sedimenta tion (Kendrick 1991, Vaselli et al. 2008), although negative effects of sediment are possible for some taxa (Airoldi TABLE 4. 2003). For bushlike algae, the lack ofresponse to achange in the level of sedimentation is somehow unexpected, particularly when coupled with achange in the densities of major grazers (Valentine & Johnson 2005). Several reasons can explain this outcome. First, it is possible that the dose added was insufficient to detect asignificant change in these algal assemblages; it is also possible that the duration of the experimental period, and even the frequency of sediment additions, were not sufficient to detect aperceptible change in their macroscopic cover. The duration and intensity of sediment treatments greatly affect the outcome of these experiments (Airoldi &Cinelli 1997, Airoldi 2003, Vaselli et al. 2008). For example, the frequency of sediment additions might have been considerably long; it is possible that sediments were swept away by swells between any 2consecutive additions. Second, it is also possible that bushlike algae in the study area may tolerate sedimentation more efficiently than previously suspected. Third, we cannot rule out the possibility that the experiment was carried out at adepth too deep for BA. At the moment, it is difficult to estimate the relative importance ofthese explanations. For example, despite 2D Stress: 0,1 •O ••O O ... 1::.. 1::.. ... ResuIts of pairwise comparisons of muItivariate dispersion between sea urchin treatments (removals [-U vs. untouched controls [+U)) and levels of sedimentation. Groups +U -sediments vs.~U -sediments +U -sediments vs. +U +sediments +U -sediments vs. -U +sediments -U -sediments vs. +U +sediments -U -sediments vs. -U +sediments +U +sediments vs.-U +sediments 58.242 14.399 16.404 60.828 39.072 0.501 p 0.058 0.078 0.067 0.003 0.088 0.438 A+U-Sed •-U-Sed 1::.. +U+Sed O-U+Sed Figure 3. Two-dimensional nonmetric multidimensionalscaling plot showing centroids of algalassemblages that were subjected to seaurchin removal (circles) or preserved at natural densities of sea urchins (untouched controls, triangles) at enhanced (open symbols) or ambient (fiUed symbols) levels of sedimentation at the end of the experimental periodo Replicated plots have been averaged within each treatment for each of the 3locations. ASSEMBLAGE PATTERNS ON ATLANTIC REEFS 953 80 + Sediments Location 1Location 2 Location 3 ~ 60 -Sediments C) e; "- ... 40 :::J ~ ~ o20 o 80 Location 1Location 2 Location 3 Q) 60 ca C) e; J: 40 ti) :::J m ~ 20 o o 80 Location 2 Location 3 Q) ~ 60 e; Q) ti) 40 o - ti) :::J ... 20 o ~ oo -U +U -U +U -U +U Figure 4. Mean percent coverage oC each morphological group subjected to sea urchin removal (-U) or preserved at natural densities oC sea urchins (+U) at enhanced (+sediments) or ambient levels (- sediments) oC sedimentation at the end oC the experimental periodo Error bars are ± SE oC means (n =8). the fact that most studies dealing with the effects of sedimentation on macroalgae have experimentally added doses 1-2 orders of magnitude 1arger than ours (Airo1di &Virgilio 1998, Connell 2005), aslight reduction in the amount of sediments sinking over algal assemblages can affect the diversity and dominance of algae, particularly with regard to colonization of patches of bare rock (Airoldi &Cinelli 1997), similar to those plots where sea urchins were removed. Because species-specific responses to sediment-induced disturbances and the local characteristics of the regime of sedimentation considerably affect the response (Airoldi 2003), it is obvious that further experimental approaches are necessary to address the interactive potential of sedimentation with biotic mechanisms (i.e., grazing, competition) shaping the organization of subtidal benthic assemblages in the study area. ACKNOWLEDGMENTS The authors thank I. Bertocci, M. S. Thomsen, M. A. Vanderklift, and T. Wernberg for providing positive criticism and discussion in previous drafts of this manuscript. Financial support for this work was provided by the Cabildo de Gran Canaria through apostgraduate research fellow to L. 0.- B. LITERA TURE CITED Airoldi, L. 2000. Responses of algae with different life histories to temporal and spatial variability of disturbance in subtida1 reefs. Mar. Ecol. Prog. Ser. 195:81-92. Airo1di, L. 2003. The effects of sedimentation on rocky coast assemblages. Oceanogr. Mar. Biol. Annu. Rev. 41:161-236. Airoldi, L. &F. Cinelli. 1997. Effects of sedimentation on subtidal macroalga1 assemblages: an experimental study from aMediterranean rocky shore. J. Exp. M al'. Biol. Ecol. 215:269-288. Airoldi, L. &M. Virgilio. 1998. Responses of turf-forming algae to spatial variations in the deposition of sediments. Mar. Ecol. Prog. Ser. 165:271-282. Alves, F. M. A., L. M. Chicharo, E. Serrao &A. D. Abreu. 2001. A1ga1 cover and sea-urchin spatia1 distribution at Madeira Island (NE Atlantic). Sci. Mar. 65:383-392. Anderson, M. J. 2001. Anew method for non-parametric multivariate analysis of variance in ecology. Austral. Ecol. 26:32--46. 954 ORTEGA-BORGES ET AL. Anderson, M. J. 2004. PERMDISP: a FORTRAN computer program for permutational analysis of multivariate dispersions (for any two-factor ANOVA design) using permutation tests. Auckland, New Zealand: Department of Statistics, University of Auckland. 11 pp. Andrew, N. L. 1993. Spatial heterogeneity, sea-urchin grazing, and habitat structure on reefs in temperate Australia. Eco!ogy 74:292302. Andrew, N. L. &AJ. Underwood. 1993. Density-dependent foraging in the sea urchin Centrostephanus rodgersii on shallow subtidal reefs in New South Wales, Australia. Mar. Eco!. Prog. Ser. 99:89-98. Balata, D. &L. Piazzi. 2008. Patterns of diversity in rocky subtidal macroalgal assemblages in relation to depth. Bot. Mar. 51:464-471. Balch, T. &R. E. Scheibling. 2000. Temporal and spatial variability in settlement and recruitment of echinoderms in kelp beds and barrens in Nova Scotia. Mar. Eco/. Prog. Ser. 205:139-154. Benedetti-Cecchi, L., F. Bulleri & F. Cinelli. 1998. Density dependent foraging of sea urchins in shallow subtidal reefs on the west coast of Italy (western Mediterranean). Mar. Eco/. Prog. Ser. 163:203-211. Bulleri, F., 1. Bertocci &F. Micheli. 2002. Interplay of encrusting coralline algae and sea urchins in maintaining alternative habitats. Mar. Eco!. Prog. Ser. 243:101-109. Coleman, M. A. 2003. The role ofrecruitment in structuring patterns of small-scale spatial variability in intertidal and subtidal algal turfs. 1. Exp. Mar. Bio!. Eco!. 291:131-145. Connell, S. D. 2005. Assembly and maintenance of subtidal habitat heterogeneity: synergistic etTects of light penetration and sedimentation. Mar. Eco!. Prog. Ser. 289:53-61. Dayton, P. K., M. J. Tegner, P. E. Parnell &P. B. Edwards. 1992. Temporal and spatial patterns of disturbance and recovery in akelp forest community. Eco/. Monogr. 62:421-445. Denny, M. W. 1988. Biology and the mechanisms of wave-swept environment. New York: Princeton University Press. 329 pp. Ebert, T. A. 1983. Recruitment in echinoderms. In: M. Jangoux &J. Lawrence, editors. Echinoderm studies. Rotterdam: Balkema. pp. 169-203. Fowler-Walker, M. J. & S. D. Connell. 2002. Opposing states of subtidal habitat across temperate Australia: consistency and predictability in kelp canopy-benthic associations. Mar. Eco!. Prog. Ser. 240:49-56. Garrabou, J., E. Ballesteros & M. Zabala. 2002. Structureand dynamics of North-western Mediterranean rocky benthic communities along a depth gradient. Estuar. Coast. She/! Sci. 55:493-508. González Barbuzano, J. R. 2003. Las Canteras y Bahía de El Confital: libro blanco. Las Palmas de Gran Canaria: Cabildo de Gran Canaria, Consejería de Obras Públicas e Infraestructuras. 617 pp. Graham, M. H. 2004. Effects of local deforestation of the diversity and structure of southern California giant kelp forest food webs. Ecosystems (N. Y., Print) 7:341-357. Hernández, J. C. 2006. Estrategia reproductiva de la población canaria del erizo Diadema aff. antillarum, Phillipi 1845: maduración gonadal, asentamiento larvario y reclutamiento. PhD dissertation, U niversidad de La Laguna. 242 pp. Irving, A. D. & S. D. Connell. 2002. Sedimentation and light penetration interact to maintain heterogeneity of subtidal habitats: algal vs. invertebrate dominated assemblages. Mar. Eco!. Prog. Ser. 245:8391. Kendrick, G. A. 1991. Recruitment of coralline crusts and filamentous turf algae in the Galapagos archipelago: effect of simulated scour, erosion and accretion. 1. Exp. Mar. Bio/. Eco!. 147:47-63. Kendrick, G. A. &D. 1. Walker. 1991. Dispersal distances for propagules of Sargassum spinu/igerum (Sargassaceae, Phaeophyta) measured directly by vital staining and Venturi suction sampling. Mar. Eco/. Prog. Ser. 79: 133-138. Lawrence, J. M. 2001. Edible sea urchins: biology and ecology. Amsterdam: EIsevier Science. 432 pp. Littler, M. M. & K. E. Arnold. 1980. Sources of variability in macroalgal primary productivity: sampling and interpretive problems. Aquat. Bot. 3: 141-156. Littler, M. M., D. S. Littler &P. R. Taylor. 1983a. Evolutionary strategies in a tropical barrier reef system: functional form groups of marine macroalgae. J. Phyco!. 19:223-235. Littler, M. M., D. R. Martz & D. S. Littler. 1983b. Effects ofrecurrent sand deposition on rocky intertidal organisms: importance of substrate heterogeneity in afluctuating environment. Mar. Eco!. Prog. Ser. 11:129-139. Lüning, K. 1990. Seaweeds: their environment, biogeography, and ecophysiology. New York: Wiley. 572 pp. McClanahan, T. R., E. Sala, P. A. Stickels, B. A. Cokos, AC. Baker, C. J. Starger & S. H. Jones. 2003. Interaction between nutrients and herbivory in controlling algal communities and coral condition on Glover's Reef, Belize. Mar. Eco/. Prog. Ser. 261:135-147. Roberts, D. E., S. P. Cummins, AR. Davis & M. G. Chapman. 2006. Structure and dynamics of sponge-dominated assemblages on exposed and sheltered temperate reefs. Mar. Eco!. Prog. Ser. 321: 19-30. Sala, E., C. F. Boudouresque & M. L. Harmelin-Vivien. 1998. Fishing, trophic cascades, and the structure of algal assemblages: evaluation of an old but untested paradigm. Oikos 83:425-439. Santelices, B. 1990. Patterns of reproduction, dispersal and recruitment in seaweeds. Oceanogr. Mar. Bio!. Annu. Rev. 28: 177-276. Schiel, D. R. & M. S. Foster. 1986. The structure of subtidal algal stands in tempera te waters. Oceanogr. Mar. Bio!. Annu. Rev. 24:265307. Schils, T. &E. Coppejans. 2003. Spatial variation in subtidal plant communities around the Socorra Archipelago and their biogeographic affinities within the Indian Ocean. Mar. Eco!. Prog. Ser. 251: 103-114. Shears, N. T. &R. C. Babcock. 2003. Continuing trophic cascade effects after 25 years of no-take marine reserve protection. Mal'. Eco!. Prog. Ser. 246:1-16 Shears, N. T., R. C. Babcock &A. K. Salomon. 2008. Contextdependent effects of fishing: variation in trophic cascades across environmental gradients. Eco!. App/. 18:1860-1873. Steneck, R. S. & M. N. Dethier. 1994. A functional group approach to the structure of algal-dominated communities. Oikos 69:476498. Terlizzi, A, S. Bevilacqua, S. Fraschetti & F. Boero. 2003. Taxonomic sufficiency and the increasing insufficiency of taxonomic expertise. Mar. Po!hil. Bull. 46:544-560. Terlizzi, A., M. J. Anderson, S. Fraschetti &L. Benedetti-Cecchi. 2007. Scales of spatial variation in Mediterranean subtidal sessile assemblages at different depths. Mar. Eco!. Prog. Ser. 332:25-39. Tuya, F., A. Boyra, P. Sánchez-Jerez, R. J. Haroun &C. Barberá. 2004a. Can one species determine the structure of a rocky benthic community on a tempera te rocky reef: the case ofthe long-spined sea urchin Diadema antillarum (Echinodermata: Echinoidea) in the eastern Atlantic. Hydrobio!ogia 519:211-214. Tuya, F., A. Boyra, P. Sánchez-Jerez, R. J. Haroun &C. Barberá. 2004b. Relationships between rocky-reef fish assemblages, the sea urchin Diadema antillarum and macroalgae throughout the Canarian Archipelago. Mar. Eco!. Prog. Ser. 278:157-169. Tuya, F., J. Cisneros-Aguirre, L. Ortega-Borges &R. J. Haroun. 2007. Bathymetric segregation of sea urchins on reefs of the Canarian Archipelago: role of flow-induced forces. Estuar. Coast. She/f Sci. 73:481--488. Tuya, F. &R. J. Haroun. 2006. Spatial patterns and response to wave exposure of photophilic algal assemblages across the Canarian Archipelago: a multiscaled approach. M al'. Eco!. Prog. Ser. 311: 15-28. Tuya, F., J. A. Martin &A. Luque. 2004c. Patterns of nocturnal movement of the sea urchin Diadema antillarum (Philippi) in Gran ASSEMBLAGE PATTERNS ON ATLANTIC REEFS 955 Canaria (Canary Islands, central east Atlantic Ocean). Helgol. Mar. Res. 58:26-3l. Valentine, J. &C. R. Johnson. 2005. Persistence of sea urchin (Heliocidaris erythrogramma) barrens on the east coast of Tasmania: inhibition of macroalgal recovery in the absence of high densities of sea urchins. Bol. Mar. 48:106-115. Vaselli, S.,!. Bertocci, E. Maggi &L. Benedetti-Cecchi. 2008. Effects of mean intensity and temporal variance of sediment scouring events on assemblages of rocky shores. Mar. Ecol. Prog. Ser. 364:57-66. Vergés, A., T. Alcoverro &E. Ballesteros. 2009. Role of fish herbivory in structuring the vertical distribution of canopy algae Cystoseira spp. in the Mediterranean Sea. Mar. Ecol. Prog. Ser. 375:1-1l. Witman, J. D. &P. Dayton. 2001. Rocky subtidal communities. In: M. D. Bertness, S. D. Gaines &M. E. Hay, editors. Marine community ecology. New York: Sinauer Associates. pp. 339--366.