Long term e¡ects of Ascophyllum nodosum canopy removal on mid shore community structure Jenkins, S.R.* P , Norton, T.A. O and Hawkins, S.J.* *Marine Biological Association, Citadel Hill, Plymouth, UK, PL1 2PB. O Port Erin Marine Laboratory (University of Liverpool), Port Erin, Isle of Man, British Isles, IM9 6JA. P Corresponding author: email: sjen@mba.ac.uk The long term e¡ects of macroalgal canopy removal on community composition were investigated over a 12 year period. Experimental removal of the dominant Ascophyllum nodosum canopy led to short term changes in community composition, the major features of which were still apparent 12 years later. Ascophyllum was slow to recover despite high recruitment, and experimental plots were dominated by Fucus species. After 12 years a mixed assemblage of Fucus serratus, Fucus vesiculosus and Ascophyllum had devel- oped. Canopy removal resulted in a change in the balance between grazing limpets and the cover of red algal turf in the understorey community.The cover of tur¢ng algae declined signi¢cantly allowing the area grazed by limpets to extend.This led to a 3^6 fold increase in the limpet population 12 years after canopy removal. INTRODUCTION Physical disturbance in the intertidal and shallow subtidal zones is frequently manifested as an extensive loss of macroalgal canopy cover (see Underwood 1998, for discussion), which can have severe knock-on e¡ects to the wider community. Canopy algae alter the physical envir- onment at the substratum through the lowering of light levels, amelioration of physical extremes, reduction of water movement and by the physical abrasion of sweeping algal fronds (see Jenkins et al., 1999). Experiments which aim to determine community level e¡ects of canopy loss typically extend over limited periods, rarely more than 3 years during which time recovery to the pre-disturbance state frequently fails to occur. Thus, it has not been possible to determine the long-term implications of canopy loss (but see Driskell et al., 2001 for recovery of Fucus gardneri over a 7 year period). Ascophyllum nodosum is a long lived canopy alga with an extensive distribution throughout the North Atlantic. At sheltered sites in north-western Europe, Ascophyllum forms a near monospeci¢c cover over much of the mid shore. Because of its commercial importance, a number of studies have investigated the potential of Ascophyllum for regrowth or recolonization in harvested or experimentally denuded areas (e.g. Printz, 1956; Keser & Larson, 1984) but exami- nation of the community level e¡ects of Ascophyllum loss has received less attention. Jenkins et al. (1999) described the important structuring role of the Ascophyllum canopy at the mid tide level of a sheltered site in the Isle of Man, and speculated on the long-term implications of canopy loss. We aimed to re-sample the original experiment 12 years after it was established to determine the long-term impact of canopy removal on this community. MATERIALS AND METHODS The study site selected, on the Langness peninsula, was typical of the sheltered rocky shores of the south of the Isle of Man with an almost complete cover of fucoid canopy algae (see Jenkins et al., 1999 for complete description of the shore). At mid tide level the shore was dominated by Ascophyllum nodosum with small patches of Fucus vesiculosus and Fucus serratus. The understorey community can be divided into two separate functional units: a multi- species, sediment trapping red algal turf, interspersed with patches of substratum, kept clear of erect algae by the grazing of Patella vulgata. These ‘bare’ patches consisted of a mosaic of bare rock and encrusting algae, predomi- nantly Phymatolithon lenormandii, and will subsequently be referred to as ‘bare substratum’. In November 1991 a two-way factorial experiment was established at mid shore level (3.3 m to 4.3 m above lowest astronomical tide) to investigate the e¡ect of the Ascophyllum canopy and the limpet, Patella vulgata, on the understorey community. Both factors had two levels, presence and absence, resulting in four orthogonal treat- ments with three replicates. Twelve plots were chosen, all positioned at least 5 metres apart in areas of smooth, gently sloping topography with a dense cover of Ascophyllum. At each plot an area 22 m square was measured and the four treatments were assigned at random to the twelve plots. For full details of the experi- mental set up seeJenkins et al., 1999. Plots were initially sampled at approximately 6 week intervals for a period of two years. Thereafter, sampling was undertaken at irregular intervals over the next four years until June 1997. A 0.50.5 m quadrat, subdivided into 25 equal squares was placed at four random positions within each plot. These four subsamples were used to calculate a mean value for each replicate. The percentage cover of dominant algal species, ‘bare substratum’ and number of grazers was estimated. In June 2003, almost 12 years after establishment of the experiment, and 6 years after the last sampling point, it was located and re-sampled. Owing to the lack of a signif- icant e¡ect of limpet grazing reported in Jenkins et al., 1999 the experiment was treated as a simple, single factor J. Mar. Biol. Ass. U.K. (2004), 84, 4496/1^3 Printed in the United Kingdom Journal of the Marine Biological Association of the United Kingdom (2004) Author: col. 2: not clear what the ‘two separate func- tional units’ are