Multi-scale spatial variability in intertidal benthic assemblages: Differences between sand-free and sand-covered rocky habitats

Download Multi-scale spatial variability in intertidal benthic assemblages: Differences between sand-free and sand-covered rocky habitats

Post on 18-Dec-2016

215 views

Category:

Documents

3 download

Embed Size (px)

TRANSCRIPT

<ul><li><p>al-c</p><p>apatPO</p><p>Accepted 10 August 2013Available online xxx</p><p>sedimentsspatial variationsintertidal environment</p><p>ofg to</p><p>diversity (heterogeneity). Here we examine the effects of sand deposition on rocky assemblages by</p><p>sand-covered and sand-free reefs, supporting the idea of that sedimentation is one of the major physical</p><p>proteenic im</p><p>processes and factors that structure assemblages (Chapman, 1999;</p><p>increased recently to provide a more in-depth knowledge ofassemblage heterogeneity for a broad variety of habitats(Benedetti-Cecchi, 2001; Prez-Ruzafa et al., 2007; Smale et al.,2010). In particular, marine benthic assemblages have been found</p><p>and physical factors in structuring intertidal assemblages along this, 1971; Schonbeck83; McCook andtrast, along-shorete the fact that iting on the spatial1). In recent years,ntied variabilityledge of the pro-</p><p>cesses that regulate species distribution from ne (patchiness) tobroad spatial scales (Menconi et al., 1999; Johnson et al., 2001;Coleman, 2003; Denny et al., 2004).</p><p>The physical factors traditionally explored as sources of varia-tion in intertidal assemblages include seawater temperature,salinity, wave exposure and intertidal height (Stephenson andStephenson, 1949; Underwood, 1978; Druehl and Green, 1982;McQuaid and Branch, 1984; Josselyn and West, 1985). However,despite the large role that sedimentation plays in modifying coastal</p><p>* Corresponding author.</p><p>Contents lists availab</p><p>Estuarine, Coastal a</p><p>.e</p><p>Estuarine, Coastal and Shelf Science xxx (2013) 1e12E-mail address: pdiaz@udc.es (P. Daz-Tapia).Anderson et al., 2005; Terlizzi et al., 2007). Causal relationshipsbetween inuential factors, ecological processes and subsequentdistribution of species need to be explored by experimentalresearch (Underwood et al., 2000), but quantifying the range ofnatural variation of assemblages may help to identify whichphysical and biological factors are most relevant to be explored rstunder an experimental approach (Underwood and Chapman, 1996;Menconi et al., 1999; Coleman, 2002). Thus, research efforts have</p><p>gradient of stress has long been studied (Daytonand Norton, 1980; Moreno and Jaramillo, 19Chapman, 1993; Jenkins et al., 1999). By convariation has received far less attention despimight be greater than vertical variation, dependscale (Benedetti-Cecchi, 2001; Valdivia et al., 201studies on horizontal heterogeneity have quaacross multiple scales of space, increasing knowtion of altered ecosystems rely largely on understanding the tical gradient imposed by tidal cycles. The role of biotic interactionsassemblage structurerocky shores</p><p>1. Introduction</p><p>The successful management andversity, the assessment of anthropog0272-7714/$ e see front matter 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.ecss.2013.08.019</p><p>Please cite this article in press as: Daz-Tapsand-free and sand-covered rocky habitats,factors inuencing the structure of benthic assemblages. More surprisingly, our ndings suggest thatsand-covered assemblages have greater spatial variation in terms of multivariate dispersion at smallspatial scales (from metres to 100s of metres) than sand-free assemblages. No differences were detectedbetween the two habitats in average species richness and Shannon diversity, whereas sand-coveredassemblages were found to be taxonomically more diverse. Thus, the effects of sedimentation on thediversity of assemblages from rocky shores remain unclear and further investigation is needed to clarifyits structuring role in combination with other environmental factors.</p><p> 2013 Elsevier Ltd. All rights reserved.</p><p>ction of biological di-pacts and the restora-</p><p>to be highly variable across different scales of time and space(Coleman, 2002; Fraschetti et al., 2005).</p><p>With regard to intertidal systems, ecologists have devotedspecial attention to examining patterns of variation along the ver-Keywords:biodiversitycontrasting the multivariate composition, spatial variation and alpha diversity between sand-free as-semblages and assemblages covered by sand. The assemblage composition differed markedly betweenMulti-scale spatial variability in intertidDifferences between sand-free and sand</p><p>Pilar Daz-Tapia a,*, Ignacio Brbara a, Isabel Dez b</p><p>aBioCost Research Group, Universidade da Corua, Facultade de Ciencias, Campus da ZbDepartment of Plant Biology and Ecology, University of the Basque Country UPV/EHU,</p><p>a r t i c l e i n f o</p><p>Article history:Received 19 March 2013</p><p>a b s t r a c t</p><p>The presence of high loadsliving on rocky reefs, leadin</p><p>journal homepage: wwwAll rights reserved.</p><p>ia, P., et al., Multi-scale spatiEstuarine, Coastal and Shelf Sbenthic assemblages:overed rocky habitats</p><p>eira s/n, 15071 A Corua, SpainBox 644, 48080 Bilbao, Spain</p><p>sediment is often thought to be negatively associated with sessile speciesassemblages with low alpha diversity (average species richness) and beta</p><p>le at ScienceDirect</p><p>nd Shelf Science</p><p>lsevier .com/locate/ecssal variability in intertidal benthic assemblages: Differences betweencience (2013), http://dx.doi.org/10.1016/j.ecss.2013.08.019</p></li><li><p>t ofnd s</p><p>stalenvironments, its potential inuence on the structure of rocky as-semblages has been rarely studied until recently (but see Daly andMathieson, 1977; Littler et al., 1983; Airoldi et al., 1995). In recentyears there have been some studies on the effects of the increase ofanthropogenic sediment loads in rocky coastal assemblages(Airoldi, 2003 and references therein), which has been reported as amajor threat to marine biodiversity on a global scale (UnitedNations Environment Programme, 1995). By contrast, the role ofnatural sedimentation in the structure of benthic assemblages hasreceived little attention (but see e.g. Daly and Mathieson, 1977;Taylor and Littler, 1982; Littler et al., 1983; Airoldi and Hawkins,2007; Anderson et al., 2008b). The two approaches to the studyof sedimentation in regard to benthic assemblages agree in thatsedimentation affects the composition and distribution of rockycoast organisms, but they present contrasting views regarding itseffects on diversity. The prevalent opinion is that high sedimentloads related to anthropogenic activities are detrimental to theoverall diversity of rocky coast organisms (Airoldi, 2003 and ref-erences therein), but some authors support the hypothesis that thenatural presence of sediments promotes species diversity (Littleret al., 1983; McQuaid and Dower, 1990). However, a multiple-scale approach comparing the relevant spatial-scales of variationin diversity between sand-covered and sand-free rocky assem-blages has never been attempted. Indeed, the issue of scale has</p><p>Fig. 1. Map of the study area showing the sampling locations on the northwestern coasRinlo and 8) Serantes. Unshaded squares and shaded circles correspond to sand-free a</p><p>P. Daz-Tapia et al. / Estuarine, Coa2rarely been addressed in research into the impacts of sedimenta-tion on rocky coast assemblages (Airoldi, 2003).</p><p>Descriptive studies of rocky intertidal assemblages from theAtlantic coast of the Iberian Peninsula have shown differences incomposition between sand-free and sand-covered rocky assem-blages (e.g. Miranda, 1931; Ardr, 1970; Prez-Cirera, 1976; Prez-Cirera and Maldonado, 1982; Brbara, 1994; Brbara et al., 1995;Daz-Tapia and Brbara, 2005). However, quantitative data onspatial patterns of distribution of organisms are scarce (Boaventuraet al., 2002; Cremades et al., 2004; Arajo et al., 2005; Dez et al.,2009). In addition, all these studies are focused on sand-freerocky habitats, while there are no previous quantitative studies ofsand-covered assemblages, even though they are commonlydistributed along the Atlantic coastline of the Iberian Peninsula (seeDaz Tapia et al., 2011).</p><p>This study assessed differences in spatial patterns of variabilityin multivariate structure and diversity between sand-free andsand-covered rocky assemblages. We use a hierarchical design toquantify the magnitude of variation attributable to each of severalspatial scales at each of the two habitats. The use of nested hier-archical sampling designs to examine both univariate and</p><p>Please cite this article in press as: Daz-Tapia, P., et al., Multi-scale spatisand-free and sand-covered rocky habitats, Estuarine, Coastal and Shelfmultivariate response variables at multiple spatial scales has led toa greater appreciation of the importance of scale in ecology(Underwood and Chapman, 1996; Fraschetti et al., 2005). Thesedesigns provide unbiased, independent, rigorous quantitativemeasures of variability at predetermined spatial scales with a viewto testing structured hypotheses (Underwood and Chapman, 1996;Terlizzi et al., 2007). Specically, we address 3 main questions: (1)do sand-free and sand-covered rocky assemblages differ in terms ofmultivariate structure and alpha diversity (species richness, Shan-non and Taxonomic Distinctness measures); (2) are patterns ofvariability in assemblage structure and diversity dependent onspatial scale, and (3) are sand-free rocky assemblages spatiallymore heterogeneous than sand-covered assemblages?</p><p>2. Methods</p><p>2.1. Study area</p><p>The study area extends for approximately 100 km along thenorthern Galician coast in northwestern Spain (43330 N to 43470</p><p>N and 006560 W to 007480 W; Fig. 1). This is an open coastexposed to a large fetch where swell comes mainly from NW (38%)andWNW (29%), with mean signicant heights (Hs) of 3 and 2.5 m,respectively (Puertos del Estado, 2013). It consists mostly of rocky</p><p>the Iberian Peninsula: 1) Picn, 2) Bares, 3) Burela, 4) Peinzs, 5) Llas, 6) Catedrales, 7)and-covered reefs, respectively.</p><p>and Shelf Science xxx (2013) 1e12substratum interrupted irregularly by the presence of rias andbeaches. Tides are semidiurnal with a spring tidal range of up to3.8 m, and the sea surface temperature ranges from 11 C to 18 C(Brbara et al., 2005). The ora falls within warm-temperate NEAtlantic subregion 1 (WNE1) according to the phytogeographicalscheme proposed by Hoek and Breeman (1990).</p><p>2.2. Collection of data and sampling design</p><p>Sampling was conducted from July to August 2010. Twodifferent habitats at the low intertidal level (between 0.4 m and1.2 m above MLWL) were selected for this study: sand-free andsand-covered rocky platforms. The former consists of rocky shoreslocated at least 100 m apart from sandy beaches. The latter are rockoutcrops on sandy beaches constantly covered by a sand layer(&gt;1 cm thick) which is trapped within algal turfs. Sampling loca-tions were randomly selected along a stretch of coastline about100 km long, from a set of locations with comparable environ-mental conditions: N-NE-facing pristine open coastal shoresexposed to strong wave action, with stable substrata (continuousbedrock and large blocks), smooth surfaces and slight to moderate</p><p>al variability in intertidal benthic assemblages: Differences betweenScience (2013), http://dx.doi.org/10.1016/j.ecss.2013.08.019</p></li><li><p>following the AlgaeBase classication (Guiry and Guiry, 2013) andinvertebrate species were compiled according to the MarBEF datasystem (MarBEF, 2004). Step lengths from species to phylum wereequally weighted over each taxonomic level and the greatesttaxonomic distance was set to 100.</p><p>Univariate PERMANOVA analyses based on Euclidean distancewere performed to test the null hypothesis that there were nodifferences in SR, H0, D*, and the mean percentage cover of the mostabundant species using the same multi-factor experimental designapplied in multivariate analyses. Prior to the analyses, Cochrans C-test was performed using the GMAV5 package (University of Syd-ney) to check for homogeneity of variance, and data were properlytransformed when necessary. The most stringent criterion ofa &lt; 0.01 was used to reject the null hypothesis when varianceswere heterogeneous and analyses were conducted on untrans-formed data.</p><p>3. Results</p><p>3.1. Assemblage composition</p><p>Overall, 127 taxa were identied in this study. The most widelyrepresented phylum was Rhodophyta with a total of 90 taxa, fol-lowed by Ochrophyta with 18 and Chlorophyta with 14 (Appendix1). Sessile invertebrates were scarcely represented, with just twotaxa belonging to Mollusca and two to Arthropoda. PERMANOVAanalyses on multivariate data detected signicant differences be-</p><p>Fig. 2. nMDS ordinations of assemblages based on BrayeCurtis dissimilarity measureson square root transformed cover data (a) and Jaccard coefcients on presence/absencedata (b). Unshaded and shaded symbols correspond to sand-free and sand-coveredreefs, respectively (see Fig. 1).</p><p>stalslope (0e45). Data were collected following a full nested hierar-chical sampling design taking into account four spatial scales:Location (four coastal stretches about 1 km long at least 10 kmapart for each habitat) (Fig. 1), Site (three stretches 100 m long atleast 100 m apart), Patch (three stretches 10 m long at least 10 mapart), and Quadrat (three 20 20 cm quadrats at least 1 m apart).The percentage cover of seaweeds and sessile invertebratesexceeding 0.5 cm in length was visually estimated in each quadratas described by Dethier et al. (1993). Representative specimens ofthose taxa that could not be identied in situ were collected foridentication at the laboratory and deposited in the SANT-Algaeherbarium (University of Santiago de Compostela).</p><p>2.3. Statistical analyses</p><p>Differences in the multivariate structure of assemblages be-tween the two habitats studied were assessed with permutationalmultivariate analysis of variance (PERMANOVA) using the PRIMERV. 6. PERMANOVA package (Clarke and Gorley, 2006; Andersonet al., 2008a). The factors considered were: Habitat (xed, 2levels), Location (random, 4 levels, nested in habitat), Site (random,3 levels, nested in locations), and Patch (random, 3 levels, nested insites) with n 3. The resemblance matrices comparing pairs ofsamples were calculated using the BrayeCurtis index on square-root transformed data and the Jaccard coefcient on presence/absence data. The tests used 9999 permutations under a reducedmodel, and signicance was accepted at p &lt; 0.05. Given that therewere too few possible permutations (</p></li><li><p>signicantly greater in sand-covered habitats than in sand-freerocky habitats at the scale of metres (quadrat), tens of metres(patch), and hundreds of metres (site), while no differences weredetected at the km scale (location) (Fig. 4A). With respect tocomposition, sand-free rocky assemblages were also more homo-geneous than sand-covered assemblages at the smallest spatialscales (i.e. quadrat &amp; patch), but PERMDISP showed that rocky as-semblages were marginally (p 0.05) more heterogeneous at thescale of kilometres (Fig. 4B).</p><p>SIMPER analysis on BrayeCurtis measures showed that ninetaxa accounted for most of the differentiation between sand-freeand sand-covered rocky assemblages (Table 2). Each taxon ac-counts for more than 2% of the average dissimilarity between thetwo habitats (84.8%), and their cumulative contribution is 51.82%.Their contribution to the average dissimilarity between habitats isconsistent across pairs...</p></li></ul>

Recommended

View more >