Biogenic habitat structure of seaweeds change along a latitudinal gradient in ocean temperature

1 Running head: Algal canopies across latitudes Biogenic habitat structure provided by temperate macroalgae change along a latitudinal gradient in ocean climate Thomas Wernberg1,2,3,*, Mads S. Thomsen3,4, Fernando Tuya3,5 and Gary A. Kendrick1 1School of Plant Biology, Oceans Institute (M096), University of Western Australia, Crawley WA 6009, Australia 2Australian Institute of Marine Science, Oceans Institute (M096), University of Western Australia, Crawley WA 6009, Australia 3Centre for Marine Ecosystems Research, Edith Cowan University, Joondalup WA 6027, Australia 4Marine Department, National Environmental Research Institute, University of Aarhus, P.O. Box 4000, Roskilde, Denmark 5BIOGES, Marine Sciences Faculty, University of Las Palmas de G.C., 35017 Las Palmas de G.C., Canary Islands, Spain * Email thom[email protected], Phone + 61 8 6369 4047 Manuscript type: New data Key words: Global warming, Ocean climate, Ocean temperature; Subtidal rocky reefs; Macroalgae; Patches and canopy structure; Western Australia 2 Abstract Global warming is affecting all major ecosystems, including temperate reefs where canopy-forming seaweeds provide biogenic habitat. In contrast to the rapidly growing recognition of how climate affects the performance and distribution of individuals and populations, relatively little is known about possible links between climate and biogenic habitat structure. We examined the relationship between several ocean temperature characteristics, expressed on time-scales of days, months and years, on habitat patch characteristics on 24 subtidal temperate reefs along a latitudinal gradient (Western Australia; ca 34 to 27º S). Significant climate related variation in habitat structure was observed, even though the landscape cover of kelp and fucalean canopies did not change across the climate gradient: monospecific patches of kelp became increasingly dominant in warmer climates, at the expense of mixed kelp-fucalean canopies. The decline in mixed canopies was associated with an increase in the abundance of Sargassum spp., replacing a more diverse canopy assemblage of Scytothalia doryocarpa and several other large fucoids. There were no observed differences in the proportion of open gaps or gap characteristics. These habitat changes were closely related to patterns in minimum temperatures and temperature thresholds (days > 20 °C), presumably because temperate algae require cool periods for successful reproduction and recruitment (even if the adults can survive warmer temperatures). Although the observed habitat variation may appear subtle, similar structural differences have been linked to a range of effects on canopy-associated organisms through the provision of habitat and ecosystem engineering. Consequently, our study suggests that the magnitude of projected temperature increase is likely to cause changes in habitat structure and thereby indirectly affect numerous habitat-dependent plants and animals. 1. Introduction Global warming has already affected all major ecosystems on earth (Rosenzweig et al. 2008) including temperate nearshore habitats where temperature increases of >0.5-1 oC have been measured over the past decades (Hawkins et al. 2003, Pearce & Feng 2007, Ridgway 2007). While the current understanding and documentation of the physical changes that are taking place is fairly good, the understanding of how this physical 3 forcing translates into effects in the biological world is comparatively poor, particularly in the sea (Harley et al. 2006, Poloczanska et al. 2007). Yet, this information is critical to understanding the range of consequences of global warming, and how they might be mitigated. Foundation species are critical to determining overall community structure in the habitats where they are found (Dayton 1972). The mechanisms behind their influence on community structure are multifaceted, but key ecological functions include modifying local environmental conditions (e.g., Wernberg et al. 2005, Irving & Connell 2006b), and the provision of habitat, i.e. a 3-dimensional physical structure that can be inhabited by various organisms (e.g., Wernberg et al. 2004, Tuya et al. 2008, Thomsen et al. 2010). Canopy-forming macroalgae are the main foundation species on most temperate rocky reefs, where two groups, kelps (order Laminariales) and fucoids (order Fucales), are particularly abundant (Schiel & Foster 1986). Elevated temperature negatively affect most aspects of the biology of kelps and fucoids from temperate coasts (e.g., reproduction, recruitment and growth, resilience to disturbances, Kirkman 1984, Novaczek 1984, Wernberg et al. 2010, respectively) and the projected global warming has caused serious concern for permanent changes in habitat structure and loss of algal habitats (Kendall et al. 2004, Poloczanska et al. 2007) with indirect effects on other community components (Schiel et al. 2004) and ensuing loss of associated biodiversity (Graham 2004, Ling 2008) and cascading effects to higher trophic levels (Kendall et al. 2004, Ling et al. 2010). Indeed, habitat loss compounded by climate change, is probably the biggest threat to biodiversity and ecosystem function, in general (Pimm 2008). Despite concerns for the future persistence of algal habitats and their ecological functions, there have been few direct assessments of how algal habitats (as opposed to species or individuals) might respond to increasing water temperatures. Broad-scale oceanographic events causing elevated ocean temperatures (i.e., ENSO’s on the west coast of the America’s) have consistently been associated with loss of kelp canopies and substantial changes to biogenic habitat structure (Dayton et al. 1999, Martínez et al. 4 2003, Edwards & Estes 2006). The influence of temperature is, however, often ambiguous because of a strong negative relationship with water-column nutrient concentrations in many kelp-dominated systems (particularly in areas of upwelling, e.g., Edwards & Estes 2006). In a unique case study of the bay-wide effects of 10 years of induced heating by a thermal effluent from a power plant, Schiel et al. (2004) found substantial structural changes associated with a temperature increase of a few degrees Celsius; they reported large shifts in habitat formers where kelps, for example, decreased markedly in abundance. With the exception of such unique ‘opportunistic experiments’, it is impossible to manipulate temperature in subtidal habitats across spatial scales relevant to questions of landscape structure (> 1-10’s meters, Connell & Irving 2008). Questions about the influence of water temperature on habitat structure must, therefore, follow alternative lines of enquiry. One option is to study latitudinal gradients in ocean temperature at places where confounding by other co-varying factors is minimal (e.g., Wernberg et al. 2010). Understanding how latitudinal gradients in climate influence the distribution of key organisms may be particularly relevant in relation to predicting effects of climate change, because shifts in distribution patterns along these gradients have been one of the main detectable responses of both aquatic and terrestrial organisms (Walther et al. 2002, Parmesan & Yohe 2003, Poloczanska et al. 2007). Here, we took advantage of a latitudinal gradient in ocean temperature to explore the influence of ocean climate on biogenic habitat structure, defined as the relative abundance of key algal habitats and their patch size characteristics. In doing so, we tested the hypothesis that habitat structure would change gradually from latitude to latitude, and that the dominant kelp habitats would become increasingly fragmented at lower latitudes where ocean temperatures are higher, and presumably more stressful to temperate algae. 2. Methods 2.1. Study area and experimental design. This study took place along a ~1,000 km stretch of the southwest coast of Western Australia, covering the latitudes 34 ºS to 27 ºS (Fig. 1A). This coastline runs north-south 5 and its near-shore environment is characterised by limestone and sandstone reefs. The ocean climate is dictated by the Leeuwin Current, a warm surface current which, in contrast to other eastern boundary currents in the southern hemisphere, runs southward and therefore suppresses up-welling (Pearce 1991). The result is a latitudinal temperature gradient of 2-3 ºC where nutrient concentrations remain consistently low (Pearce 1991, Lourey et al. 2006, Smale & Wernberg 2009). This temperature gradient captures the range of current scenarios for global warming of Australia’s temperate waters (25-year projection of 1-2 ºC increase and 50-year projection of 2-4 ºC, Poloczanska et al. 2007). Grazing pressure from fish and invertebrates is also low throughout this region (Wernberg et al. 2008, Vanderklift et al. 2009). The reefs along the coastline host a diverse algal flora dominated by a canopy of the only kelp in Western Australia, Ecklonia radiata (a species similar to northern kelps such as Eisenia arborea and Laminaria digitata) and a rich assemblage of fucalean algae mainly from the genera Cystophora, Sargassum and Scytothalia (Wernberg et al. 2003, Smale et al. 2010). Similar reef habitats and canopies characterise rocky reefs throughout temperate Australasia (Wernberg et al. 2003, Connell & Irving 2008). All sampling was completed within one month (November 2005) and followed a hierarchical design with four regions (each representing a different temperature regime) evenly separated by ~2º latitude (Fig. 1A). Six independent reefs, 8-12 m deep and separated by a minimum of 1 km, were nested within each region. The depth (transect start and end) and structural complexity (count of intersecting topographic features of 12 m and >2 m vertical relief) was measured at 25 transects (see below) for each reef. There were no differences in depth or vertical relief among regions (MSdepth = 5.61, F(3,20) = 1.0, P = 0.428; MSLN(1-2m+1) = 0.56, F(3,20) = 0.96, P = 0.430; MSLN(>2m+1) = 0.24, F(3,20) = 0.58, P = 0.634). Also, differences in light among regions due to latitude (510%) were masked by differences among reefs within regions due to differences in depths among them (~25%) (see calculations in Stæhr & Wernberg 2009). Here, each reef was considered a biologically independent replicate of the ocean climate represented by each region. This is justified because subtidal temperatures vary little 6 (<5%) between reefs at similar depth ranges (Smale & Wernberg 2009), and because most propagules from canopy-forming macroalgae have short dispersal distances (metres or less, e.g. Sargassum spinuligerum: Kendrick & Walker 1991) resulting in low genetic connectivity between reefs (e.g. Ecklonia radiata: Coleman et al. 2009). 2.2. Ocean climate. The unique ocean climate, and the broad latitudinal patterns of temperature change along the southwest coast, has long been established (Pearce 1991, Fig. 1A). Recently, Smale & Wernberg (2009) showed strong correlations between temperatures at 8-10 m depth and satellite-derived sea surface temperatures (SST) in this area. Consequently, historic satellite derived SST’s obtained by the Advanced Very High Resolution Radiometer (AVHRR) aboard the National Oceanic and Atmospheric Administration (NOAA) series of satellites could here be used to describe the ocean climate of each region prior to measurement of habitat structure. SST’s have been archived by the Australian Bureau of Meteorology. Daily SST maps of Australian waters were used to generate temperature profiles for each of the four study regions (Australian Bureau of Meteorology, www.bom.gov.au) (see Smale & Wernberg 2009 for details). Two data sets were generated; one covering seven years (1999-2005, assessed every 15th of January, April, July and October) and one covering 1 year (November 2004 – October 2005, assessed every 1st, 10th and 20th of each month). Ambient water temperature was also measured at the time of sampling (Onset Tidbit pendant loggers, logging every 5 minutes ~5 cm above the bottom at all six reefs within each region). The combination of these measures integrates long-term temperature exposure across >1 full generation of Ecklonia radiata (life span 1-4 years, Wernberg 2005), and more acute effects due to the particular year, season and time of sampling. The mean temperature, maximum and minimum temperatures, and number of ‘hot’ days > 20 oC, were used as proxies for chronic thermal exposure, extreme thermal exposure, and ‘adverse’ thermal exposure, respectively; 20 oC corresponds to the approximate threshold where temperature starts to negatively affect the productivity and growth of E. radiata (Kirkman 1984, Hatcher et al. 1987). 7 2.3. Habitat structure. The overall biogenic habitat structure of each reef was quantified by running ten 25 m transects along haphazard compass bearings and recording the distances (to nearest 10 cm) of habitat change. This technique has previously been successfully applied to describe habitat structure in algal dominated assemblages in other subtidal habitats (Kennelly 1987, Wernberg 2006, Connell & Irving 2008). We distinguished between four major habitat types: (1) mono-specific kelp canopy (>90% Ecklonia radiata), (2) fucalean canopy (>50% fucalean algae), (3) mixed kelp and fucalean canopy (50-90% kelp and 10-50% fucalean algae) and (4) canopy-free areas (gaps of < 10% canopy algae), typically dominated by small foliose red algae (<25 cm) (e.g., Hennedya crispa) or articulated coralline algae (e.g., Amphiroa anceps). The distinction between these habitat types were based on previous studies documenting how they affect their immediate surroundings differently, and support different associated benthic, phytal and fish communities (Harman et al. 2003, Irving et al. 2004, Wernberg et al. 2005, Irving & Connell 2006a, Tuya et al. 2008). Habitat structure was sampled in November 2005 (early Austral summer), a time where the kelp canopy has recovered from the thinning and pruning of winter storms (Wernberg & Goldberg 2008). Overall, 24 descriptors of habitat structure were considered, six for each habitat type: habitat cover was calculated as the proportion of each transect occupied by each habitat type. Median, mean and maximum habitat patch sizes (1-dimensional linear distance) as well as the number of habitat patches for each habitat type were determined for each reef after pooling all transects. Patch mode was not considered because there were virtually no differences (~1 m for all habitat types). 2.4. Statistical analyses. Differences in mean ocean temperature characteristics between adjacent regions were tested by either ordinary or paired t-tests (Zar 1996). Overall patterns in ocean climate were assessed by principal components analysis (PCA, Clarke & Gorley 2006) based on all nine temperature characteristics (cf. Table 1), and the presence of sequential change from region to region tested by non-parametric serial correlation (RELATE, Clarke & Gorley 2006). 8 Differences in biogenic habitat structure among regions, taking all 24 habitat characteristics into consideration, were tested by multivariate analysis of variance by permutation (PERMANOVA, Anderson et al. 2008). Constrained ordination by canonical analysis of principal coordinates (CAP, Anderson et al. 2008) was used to characterise, and visualise, these differences, and to evaluate their correspondence with individual climate variables in the environmental matrix by canonical correlation (CCorA, Anderson et al. 2008). These analyses were based on normalised data for habitat characteristics, using euclidian distances and 999 permutations. Subsequently, the influence of ocean climate on biogenic habitat was tested by linear regression (Zar 1996) of each habitat characteristic against PC1, the first principal components axis from the PCA of patterns in ocean climate (a univariate proxy for multivariate variation in ocean temperature characteristics). Regression analyses were done for habitat cover, habitat patch density and habitat patch size for all four habitat types, and for total cover of four dominant canopy taxa. 3. Results 3.1. Ocean climate. There were clear differences in ocean climate along the coastline, and the overall pattern for almost all temperature characteristics was an increase from Hamelin Bay in the south towards Kalbarri in the north (Fig. 1, Table 1). Ocean temperatures showed a clear seasonal pattern of ~6-8 ºC variation between summer and winter in all four regions. Inter-annual differences were generally small, particularly after 2000 (Fig. 1B). The four regions spanned a temperature gradient of ~2-3 ºC; Hamelin Bay was substantially cooler in summer (Dec-Feb) whereas Kalbarri remained substantially warmer in winter (Jun-Oct) (Fig. 1C). Conditions at Marmion and Jurien Bay were intermediate (Table 1). Considering all temperature characteristics simultaneously, PCA revealed a strong serial correlation in multivariate thermal climate from region to region (RELATE,  = 0.926, P = 0.081, n = 4), emphasising a gradual environmental change along the southwest 9 Australian coastline. The first principal component (PC1) accounted for 85.2% of the multivariate differences in thermal climate, and thus provided a strong univariate proxy for overall variation in thermal climate among regions, where larger values were associated with warmer (more stressful) ocean climates. 3.2. Habitat structure. Based on all 24 descriptors, there were significant differences in biogenic habitat structure between regions (PERMANOVA, pseudo-F = 2.36, P = 0.020). The constrained ordination grouped all four regions into distinct groups (CAP, trace statistic = 0.410, P = 0.030, Fig. 2) where Kalbarri, Marmion and Jurien Bay were separated along CAP1 and Hamelin Bay from these three groups along CAP2. Canonical correlation (CCorA) against the temperature matrix revealed a significant relationship between ocean climate and biogenic habitat structure (CAP, trace statistic = 1.062, P = 0.003), where minimum temperature and the number of hot days in the year preceding sampling (2004-2005) were the best correlates of CAP1 and temperature at the time of sampling, maximum temperature in the preceding year, and the number of hot days during the preceding 7 years (1999-2005), were the best correlates of CAP2 (Table 2). These temperature characteristics have, therefore, likely influenced the formation of the observed patterns of biogenic habitat structure. All four habitat types were present at all four regions, although only monospecific kelp canopies and open gaps were present at all 24 reefs (Fig. 3). Mono-specific kelp canopies covered 20-90% of the reefs at all sites, and was by far the dominant habitat type. In contrast, fucalean canopies covered <30% of the reefs and was generally the least common habitat type. With a median size of 2-5 m, kelp patches were also the largest, followed by mixed canopy patches (~2-3 m), fucoid canopies (~1-3 m) and open gaps (~1-1.5m). There was substantial variation in most biogenic habitat characteristics both among and within climates (as represented by regions) (Fig. 3). For example, the cover of monospecific kelp canopies ranged from ~10-70% on reefs within Hamelin Bay and mixed kelp and fucoid canopies from 0-30% in Kalbarri. Despite this variation, there was a 16 Coleman MA, Gillanders BM, Connell SD (2009) Dispersal and gene flow in the habitat-forming kelp, Ecklonia radiata: relative degrees of isolation across an east-west coastline. Marine and Freshwater Research 60:802-809 Connell SD, Irving AD (2008) Integrating ecology with biogeography using landscape characteristics: a case study of subtidal habitat across continental Australia. Journal of Biogeography 35:1608-1621 Dayton PK (1972) Towards an understanding of community resilience and the potential effects of enrichment to the benthos at McMurdo Sound, Antarctica. In: Parker BC (ed) Colloquium on conservation problems in Antarctica. Allen Press, Lawrence, Kansas, USA, p 81-95 Dayton PK, Tegner MJ, Edwards PB, Riser KL (1999) Temporal and spatial scales of kelp demography: The role of oceanographic climate. Ecological Monographs 69:219-250 Edwards MS, Estes JA (2006) Catastrophe, recovery and range limitation in NE Pacific kelp forests: a large-scale perspective. Marine Ecology Progress Series 320:79– 87 Goodsell PJ, Fowler-Walker MJ, Gillanders BM, Connell SD (2004) Variations in the configuration of algae in subtidal forests: Implications for invertebrate assemblages. Austral Ecology 29:350-357 Graham MH (2004) Effects of Local Deforestation on the Diversity and Structure of Southern California Giant Kelp Forest Food Webs. Ecosystems 7:341-357 Harley CDG, Randall Hughes A, Hultgren KM, Miner BG, Sorte CJB, Thornber CS, Rodriguez LF, Tomanek L, Williams SL (2006) The impacts of climate change in coastal marine systems. Ecology Letters 9:228-241 Harman N, Harvey ES, Kendrick GA (2003) Differences in fish assemblages from different reef habitats at Hamelin Bay, south-western Australia. Marine and Freshwater Research 54:177-184 Hatcher BG, Kirkman H, Wood WF (1987) Growth of the kelp Ecklonia radiata near the northern limit of its range in Western Australia. Marine Biology 95:63-72 Hawkins SJ, Southward AJ, Genner MJ (2003) Detection of environmental change in a marine ecosystem –evidence from the western English Channel. The Science of the Total Environment 310:245–256 Huisman JM (2000) Marine plants of Australia, Vol. University of Western Australia Press, Perth Irving AD, Connell SD (2006a) Physical disturbance by kelp abrades erect algae from the understorey. Marine Ecology Progress Series 324:127-137 Irving AD, Connell SD (2006b) Predicting understorey structure from the presence and composition of canopies: an assembly rule for marine algae. Oecologia 148:491502 Irving AD, Connell SD, Gillanders BM (2004) Local complexity in patterns of canopybenthos associations produces regional patterns across temperate Australasia. Mar Biol 144:361-368 Kendall MA, Burrows MT, Southward AJ, Hawkins SJ (2004) Predicting the effects of marine climate change on the invertebrate prey of the birds of rocky shores. Ibis 146:40-47 Kendrick GA, Lavery PS, Philips JC (1999) Influence of Ecklonia radiata kelp canopy structure on macro-algal assemblages in Marmion Lagoon, Western Australia. Hydrobiol 399:275-283 17 Kendrick GA, Walker DI (1991) Dispersal distances for propagules of Sargassum spinuligerum Sargassaceae Phaeophyta measured directly by vital staining and venturi suction sampling. Marine Ecology Progress Series 79:133-138 Kennelly SJ (1987) Physical disturbances in an australian kelp community. I. Temporal effects. Mar Ecol Prog Ser 40:145-153 Kirkman H (1984) Standing stock and production of Ecklonia radiata (C.Ag.) J. Agardh. J Exp Mar Biol Ecol 76:119-130 Ling S (2008) Range expansion of a habitat-modifying species leads to loss of taxonomic diversity: a new and impoverished reef state. Oecologia 156:883-894 Ling SD, Johnson CR, Frusher SD, Ridgway KR (2010) Overfishing reduces resilience of kelp beds to climate-driven catastrophic phase shift. Proceedings of the National Academy of Sciences of the United States of America in press Lourey MJ, Dunn JR, Waring J (2006) A mixed-layer nutrient climatology of Leeuwin Current and Western Australian shelf waters: Seasonal nutrient dynamics and biomass. Journal of Marine Systems 59:25-51 Martinez EA, Cardenas L, Pinto R (2003) Recovery and genetic diversity of the intertidal kelp Lessonia nigrescens (Phaeophyceae) 20 years after El Nino 1982/83. Journal of Phycology 39:504-508 Matson P, Edwards M (2007) Effects of ocean temperature on the southern range limits of two understory kelps, Pterygophora californica and Eisenia arborea , at multiple life-stages. Marine Biology 151:1941-1949 Novaczek I (1984) Response of gametophytes of Ecklonia radiata (Laminariales) to temperature in saturating light. Marine Biology 82:241-245 Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37-42 Pearce A, Feng M (2007) Observations of warming on the Western Australian continental shelf. Marine and Freshwater Research 58:914-920 Pearce AF (1991) Eastern boundary currents of the southern hemisphere. Journal of the Royal Society of Western Australia 74:35-45 Phillips J (2001) Marine macroalgal biodiversity hotspots: why is there high species richness and endemism in southern Australian marine benthic flora? Biodiversity and Conservation 10:1555-1577 Pimm SL (2008) Biodiversity: Climate Change or Habitat Loss — Which Will Kill More Species? Current Biology 18:117-119 Poloczanska ES, Babcock RC, Butler A, Hobday AJ, Hoegh-Guldberg O, Kunz TJ, Matear R, Milton DA, Okey TA, Richardson AJ (2007) Climate change and Australian Marine life. Oceanogr Mar Biol Ann Rev 45:407-478 Ridgway KR (2007) Long-term trend and decadal variability of the southward penetration of the East Australian Current. Geophys Res Lett 34 Rosenzweig C, Karoly DJ, Vicarelli M, Neofotis P, Wu Q, Casassa G, Menzel A, Root TL, Estrella N, Seguin B, Tryjanowski P, Liu C, Rawlins S, Imeson A (2008) Attributing physical and biological impacts to anthropogenic climate change. Nature 453:353-357 Schiel DR, Foster MS (1986) The structure of subtidal algal stands in temperate waters. Oceanogr Mar Biol Ann Rev 24:265-307 Schiel DR, Steinbeck JR, Foster MS (2004) Ten years of induced ocean warming causes comprehensive changes in marine benthic communities. Ecology 85:1833-1839 18 Schils T, Wilson SC (2006) Temperature threshold as a biogeographic barrier in northern Indian Ocean macroalgae. Journal of Phycology 42:749-756 Smale DA, Kendrick GA, Wernberg T (2010) Assemblage turnover and taxonomic sufficiency of subtidal macroalgae at multiple spatial scales. J Exp Mar Biol Ecol 384:76-86 Smale DA, Wernberg T (2009) Satellite-derived SST data as a proxy for water temperature in near-shore benthic ecology. Marine Ecology Progress Series 387:27-37 Stæhr PA, Wernberg T (2009) Physiological responses of Ecklonia radiata (Laminariales) to a latitudinal gradient in ocean temperature. Journal of Phycology 45:91-99 Thomsen MS, Wernberg T, Altieri A, Tuya F, Gulbransen D, McGlathery K, Holmer M, Silliman BR (2010) Habitat cascades: the conceptual context and global relevance of facilitation cascades via habitat formation and modification. Integrative and Comparative Biology 50:158-175 Toohey BD, Kendrick GA, Harvey ES (2007) Disturbance and reef topography maintain high local diversity in Ecklonia radiata kelp forests. Oikos 116:16181630 Tuya F, Wernberg T, Thomsen M (2009) Habitat structure affect abundances of labrid fishes across temperate reefs in south-western Australia. Environmental Biology of Fishes 86:311-319 Tuya F, Wernberg T, Thomsen MS (2008) The spatial arrangement of reefs alters the ecological patterns of fauna between interspersed algal habitats. Estuarine, Coastal and Shelf Science 78:774–782 Vanderklift MA, Lavery PS, Waddington KI (2009) Intensity of herbivory on kelp by fish and sea urchins differs between inshore and offshore reefs. Marine Ecology Progress Series 376:203–211 Walther G-R, Post E, Coney P, Menzel A, Parmesan C, Beebee TJC, Fromentin J-M, Hoegh-Guldberg O, Bairlein F (2002) Ecological responses to recent climate change. Nature 416:389-395 Wernberg T (2005) Holdfast aggregation in relation to morphology, age, attachment and drag for the kelp Ecklonia radiata. Aquat Bot 82:168-180 Wernberg T (2006) Scale of impact determines early post-disturbance assemblage structure in subtidal Fucus beds in the Baltic Sea (Bornholm, Denmark). European Journal of Phycology 41:105-113 Wernberg T, Connell SD (2008) Physical disturbance and subtidal habitat structure on open rocky coasts: effects of wave exposure, extent and intensity. Journal of Sea Research 59:237–248 Wernberg T, Goldberg N (2008) Short-term temporal dynamics of algal species in a subtidal kelp bed in relation to changes in environmental conditions and canopy biomass. Estuarine, Coastal and Shelf Science 76:265-272 Wernberg T, Kendrick GA, Phillips JC (2003) Regional differences in kelp-associated algal assemblages on temperate limestone reefs in south-western Australia. Diversity and Distributions 9:427-441 Wernberg T, Kendrick GA, Toohey BD (2005) Modification of the physical environment by an Ecklonia radiata (Laminariales) canopy and implications for associated foliose algae. Aquat Ecol 39:419-430 19 Wernberg T, Thomsen MS, Staehr PA, Pedersen MF (2004) Epibiota communities of the introduced and indigenous macroalgal relatives Sargassum muticum and Halidrys siliquosa in Limfjorden (Denmark). Helgoland Mar Res 58:154-161 Wernberg T, Thomsen MS, Tuya F, Kendrick GA, Staehr PA, Toohey BD (2010) Decreasing resilience of kelp beds along a latitudinal temperature gradient: potential implications for a warmer future. Ecology Letters 13:685-694 Wernberg T, White M, Vanderklift MA (2008) Population structure of turbinid gastropods on wave-exposed subtidal reefs: effects of density, body size and algae on grazing behaviour. Marine Ecology Progress Series 362:169-179 Zar JH (1996) Biostatistical analysis, Vol. Prentice Hall, Upper Saddle River 20 Table 1. Ocean temperature characteristics of the regions studied. Temperature data for time of sampling were measured in situ where 1 habitat characteristics were sampled. Temperature data for 2004-2005 and 1999-2005 were obtained by remote sensing (see methods). 2 Temperature characteristics Hamelin Bay Marmion Jurien Bay Kalbarri Pattern Time of sampling (Tidbit data logger) Mean (oC, ± SE) 17.95 ± 0.03 19.34 ± 0.11 19.97 ± 0.13 21.21 ± 0.11 H < M < J < K (P < 0.001)a 2004-05 (SST, AVHRR) Mean (oC, ± SE) 19.82 ± 0.27 20.93 ± 0.43 21.08 ± 0.48 22.94 ± 0.34 H < M = J < K (P < 0.001) b Minimum (oC) 17.28 17.28 16.35 19.14 H = M = J < K Maximum (oC) 23.79 24.72 25.65 25.65 H < M < J = K Hot days >20 oC (%) 55.6 61.1 61.1 94.4 H < M = J < K 1999-2005 (SST, AVHRR) Mean (oC, ± SE) 19.48 ± 0.50 20.13 ± 0.51 21.42 ± 0.52 22.10 ± 0.53 H < M < J < K (P < 0.010) b Minimum (oC) 15.18 15.18 16.28 17.22 H = M < J < K Maximum (oC) 27.16 27.16 28.24 29.32 H = M < J < K Days >20 oC (%) 46.4 60.7 71.4 78.6 H < M < J < K aOrdinary t-tests 3 bPaired t-tests, paired by sampling dates 4 21 Table 2. Canonical eigenvectors (correlation coefficients) for individual ocean 5 temperature characteristics in relation to the constrained ordination of biogenic habitat 6 structure on reefs within regions. The five best correlated variables are highlighted in 7 bold. 8 9 Temperature Variable CAP1 (  12 = 0.54) CAP2 (22 = 0.45) Temperature, time of sampling -0.152 -0.486 Mean temperature 2004-2005 -0.305 -0.399 Minimum temperature 2004-2005 -0.636 0.060 Maximum temperature 2004-2005 0.083 -0.474 Percent hot days >20 ºC, 2004-2005 -0.489 -0.038 Mean temperature 1999-2005 -0.139 -0.107 Minimum temperature 1999-2005 -0.314 0.273 Maximum temperature 1999-2005 -0.336 0.268 Percent hot days >20 ºC, 1999-2005 -0.052 -0.465 10 11 22 Fig. 1. A: Position of the four regions included in this study on the southwest coast of 12 Western Australia. Summer (short dash) and winter (long dash) isotherms after Pearce 13 (1991). B: Sea Surface Temperature (SST) measured by remote sensing (AVHRR). The 14 dotted line indicate 20 oC, a threshold for when the growth and productivity of Ecklonia 15 radiata starts to decline (Kirkman 1984, Hatcher et al. 1987). 16 17 18 19 20 21 23 Fig. 2. Constrained ordination of principal coordinates for habitat structure at six reefs 22 within Hamelin Bay (HAM), Marmion (MAR), Jurien Bay (JUR) and Kalbarri (KAL). 23 24 25 26 27 28 Normalise Rese mbl a nce : 01 Euclidean d istance O .• L oe Á HAM • MAR ... DJUR O KAL 0.2 O N ... ... O "- [!\. " ... O '-' o • O • O O • O • 00 • • O O ·0.2 I I I I I ·0.4 ·0.2 o 0.2 O .• CAP1 24 Fig. 3. Relationship between ocean climate at four regions (from left to right: Hamelin 29 Bay, Marmion, Jurien Bay and Kalbarri) and characteristics of the biogenic habitat at 6 30 independent sites within each region. Kelp canopy refer to mono-specific patches of E. 31 radiata. The ocean climate incorporates temperature variation at multiple temporal 32 scales (cf. Table 1). Solid lines indicate significant relationships (P < 0.05) and dashed 33 lines near-significant relationships (0.05<P < 0.10). 34 35 36 37 38 25 Fig. 4. Relationship between ocean climate and abundance of dominant canopy-forming 39 taxa. Note, E. radiata also includes kelps found in mixed canopies. Increments in ocean 40 climate corresponds to (from left to right) the studied regions (Hamelin Bay, Marmion, 41 Jurien Bay and Kalbarri). Solid lines indicate significant relationships (P < 0.05). 42 43 44 45 46 47