Spatial and Temporal Variations in Sediment Accumulation in an Algal Turf and Their Impact on Associated Fauna

Marine Biology (2003) 142: 381–390 DOI 10.1007/s00227-002-0940-4 Anchana Prathep Æ R.H. Marrs Æ T.A. Norton Spatial and temporal variations in sediment accumulation in an algal turf and their impact on associated fauna Received: 18 March 2001 / Accepted: 16 August 2002 / Published online: 26 November 2002 Ó Springer-Verlag 2002 Abstract The relationships between the fauna inhabiting Introduction an intertidal algal turf, Osmundea pinnatifida, and the accumulated sediment were studied in the autumn and Extensive turfs of algae are a feature of the lower reaches summer. The investigation was carried out at two levels of rocky shores in many parts of the world (Myers and on sheltered, moderately exposed and very exposed Southgate 1980; Kain and Norton 1990). The crowded shores on the temperate rocky coast of the Isle of Man, fronds create a labyrinth that traps sediment (Scoffin British Isles. Twenty-four species of invertebrates were 1970; Stewart 1983; Airoldi et al. 1996) and is often in- found associated with the turfs, and their abundance habited by a rich and varied fauna (Nuemann 1970; and diversity varied with season, degree of wave expo- Myers and Southgate 1980; Chapman 1995). The accu- sure, shore level and the amount and particle size of the mulation of sediment in turfs varies in time and space sediment trapped within the turfs. Multivariate analysis horizontally and vertically on the shore (Whorff et al. indicated that most organisms were most strongly 1995; Airoldi et al. 1996), and seasonally, largely as a influenced by sediment accumulation and temporal consequence of wave action (Lewis 1964; Stephenson changes in the turf plants. Sediment provides a hetero- and Stephenson 1972; Stewart 1983). Terrestrial sources geneous habitat for psammophylic organisms, supplies of sediment washing down onto the shore as a result of materials for tube-building species, and is a food source deforestation and erosion may also be important in for detritivores. However, it also has adverse effects; its some regions (Airoldi et al. 1996). seasonal movement resulted in an unstable community. The effects of sediment on the intertidal and subtidal Sediment grain size was also important; certain grades zones have been well studied, but rarely with respect to appeared to be correlated with the abundance or spar- turf communities. Sediment is known to scour or oc- sity of some species. ‘‘Season’’ was, however, the over- clude the substratum, as well as smothering organisms riding factor influencing this microcommunity, since the or cutting off the light to juvenile attached plants frond complexity and the productivity of the turf plant, (Norton 1978). Thus, it interferes with the recruitment, as well as the supply and movement of sediment, vary growth and survival of algae (Neushul et al. 1976), as seasonally. well as zonation patterns and species diversity (Daly and Mathieson 1977; Little and Smith 1980; Taylor and Communicated by J.P. Thorpe, Port Erin Littler 1982; Littler et al. 1983). Sediment can also regulate the respiratory metabolism in some common A. Prathep (&) shore gastropods (Marshall and McQuaid 1989; Department of Biology, Faculty of Science, Chandrasekara and Frid 1998), clog the feeding mech- Prince of Songkla University, HatYai, Songkla, 90112, Thailand anism of suspension feeders (Eleftheriou and Basford E-mail: [email protected] 1989; Aller and Stupakoff 1996) and inhibit the settle- Tel.: +66-1-542-69-77 ment of a variety of marine invertebrates (Stewart Fax: +66-74-212917 1989). T.A. Norton In turf communities, studies have mostly focused on Port Erin Marine Laboratory, School of Biological Sciences, University of Liverpool, Port Erin, Isle of Man IM9 6JA the effect of sediment on the algae themselves rather than on the fauna harboured by the turfs (but see R.H. Marrs Applied Vegetation Dynamics Laboratory, Whorff et al. 1995). For example, the amount and type School of Biological Sciences, University of Liverpool, of sediment may structure both the species composition PO Box 147, Liverpool L69 3GS, UK and the diversity of the algal flora (Stewart 1983; 382 Kendrick 1991; Airoldi et al. 1995), but there is only one (count of species abundance, H’, J and total weight of sediment and report of the effects on fauna (Whorff et al. 1995). of each size fraction separately) were analysed using ANOVA for each season using a nested design with degree of exposure (shore) It is well known that algal turfs can trap sediment on as the main factor and shore level nested within the main factor. To shores where it would otherwise be absent, but little is test for temporal change, repeated measures were used; this was known about the relationships between turf communi- interpreted as a first approximation of seasonal effects. Seasonal ties, sediment accumulation and the interaction between effects were assessed using repeated measures. This analysis was physical factors, sediment supply and fauna. In an at- done using untransformed and transformed data (species numbers Öx; sediment amounts ln x) with PROC ANOVA (SAS 1985). tempt to investigate these relationships, we report the Statistical results are presented based on the analyses of trans- results of a study of turfs of Osmundea (Laurencia) formed data, but, for clarity, graphical output is based on the pinnatifida (Hudson) Lamouroux on rocky shores of the untransformed means. Isle of Man. O. pinnatifida was chosen as a model system The potential relationships between species abundance and site factors, season and sediment were then assessed using stepwise because it forms discrete patches of short turf, with ir- multiple regression (PROC GLM; SAS 1985). Here degree of ex- regular branches that trap sediment. These patches posure, shore elevation and season were treated as categorical represent small ‘‘islands’’ of favourable habitat in an variables and sediment values as an ordinal variable. The aim was otherwise adverse terrain of open rock and within them not to develop predictive models, but rather to select those com- binations of variables that were associated with the abundance of there is a greater animal diversity than on the adjacent individual species, to give information on the types of factors that rock. might be controlling their distribution. The univariate approach We measured the amount and types of sediment in was used to generate hypotheses for individual species which could patches of O. pinnatifida in two contrasting seasons, on be tested experimentally in future work. Transformed data were shores with different degrees of wave exposure and at used throughout. Multivariate analysis was also carried out using CANOCO for different levels on the same shore. The fauna associated WINDOWS (ter Braak and Smilauer 1998). The data (24 species, with the turfs was examined at the same time. 60 samples) were analysed using a sequence of techniques; in all We hypothesised that (1) the Osmundea turf harbours multivariate analyses the species data were transformed (to ln a fauna whose composition, distribution and diversity x+1) and the downweighting option for rare species was not used. Initially the species x plot data were analysed with DEC- will vary between sites and seasons and (2) the nature of ORANA (DCA) to measure the gradient lengths. The gradient the associated fauna will be determined chiefly by the length on the first axis was 2.9, suggesting that the linear model amount and type of sediment that accumulates in the should be used. Thereafter, two constrained ordinations were run turf. using environmental variables in which wave exposure, shore elevation and season were considered as using categorical information, and the different sediment fractions were treated as ordinal data. The first constrained ordination was a redundancy Materials and methods analysis (RDA), in which the Forward Selection procedure was used with a Monte Carlo test with 499 permutations, to determine Study site and sampling which environmental variables were important in explaining the variation in the species in each data set. The following variables The study sites were all in the south of the Isle of Man, in the Irish were selected (at P<0.05) in order: degree of exposure, shore Sea. Three shores were selected with different degrees of wave ex- elevation, season, and amount of sediment in two fractions posure: sheltered (S) Port Erin (54°06¢N, 4°46¢W), moderately ex- (63–125 lm, 500 lm–1 mm). In the final RDA the eigenvalues for posed (M) Port St. Mary Ledges (54°0¢N, 4°44¢W), and exposed (E) the four axes were 0.141, 0.125, 0.029 and 0.366, and the Monte Scarlet Point (54°04¢N; 4°39¢W). Exposure was assessed using the Carlo tests with 999 permutations were significant, P=0.001 (first biological exposure scale of Lewis (1964). Each shore was sampled canonical axis, F=9.207, P<0.01; overall model, F=7.835, at two shore levels within the turf zone: upper limit (h) 3.5 m above P<0.001). lowest astronomical tide (LAT) and lower limit (l) 2.0 m above LAT. Samples were taken in two seasons: late autumn (November 1998) and summer (August 1999). Thirty patches of turf were Results chosen randomly at each shore level in each season. Areas of turf measuring 10 cm·10 cm were removed using a paint scraper. This was done carefully to prevent the loss of associated organisms and Sediment trapped in algal turfs sediment. Samples were immediately placed in plastic bags and transferred to the laboratory for processing. The total amount of sediment trapped varied greatly, Samples were washed with filtered sea water to remove sedi- between 1,126 and 4,628 g/m2, and the amount present ment and organisms. The sediment was then filtered using distilled water and gentle suction, and dried at 60°C to constant weight.

Load more