Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
rspb.royalsocietypublishing.org
Research Cite this article: Bates AE, Stuart-Smith RD, Barrett NS, Edgar GJ. 2017 Biological interactions both facilitate and resist climaterelated functional change in temperate reef communities. Proc. R. Soc. B 284: 20170484. http://dx.doi.org/10.1098/rspb.2017.0484
Received: 7 March 2017 Accepted: 5 May 2017
Subject Category: Global change and conservation Subject Areas: ecology Keywords: climate change, community temperature index, functional traits, marine protected area, tropicalization
Author for correspondence: Amanda E. Bates e-mail: [email protected]
Electronic supplementary material is available online at https://dx.doi.org/10.6084/m9. figshare.c.3782237.
Biological interactions both facilitate and resist climate-related functional change in temperate reef communities Amanda E. Bates1, Rick D. Stuart-Smith2, Neville S. Barrett2 and Graham J. Edgar2 1 2
National Oceanography Centre, University of Southampton, Waterfront Campus, Southampton SO14 3ZH, UK Institute for Marine and Antarctic Studies, University of Tasmania, Hobart 7001, Tasmania, Australia AEB, 0000-0002-0198-4537 Shifts in the abundance and location of species are restructuring life on the Earth, presenting the need to build resilience into our natural systems. Here, we tested if protection from fishing promotes community resilience in temperate reef communities undergoing rapid warming in Tasmania. Regardless of protection status, we detected a signature of warming in the brown macroalgae, invertebrates and fishes, through increases in the local richness and abundance of warm-affinity species. Even so, responses in protected communities diverged from exploited communities. At the local scale, the number of cool-affinity fishes and canopy-forming algal species increased following protection, even though the observation window fell within a period of warming. At the same time, exploited communities gained turf algal and sessile invertebrate species. We further found that the recovery of predator populations following protection leads to marked declines in mobile invertebrates— this trend could be incorrectly attributed to warming without contextual data quantifying community change across trophic levels. By comparing long-term change in exploited and protected reefs, we empirically demonstrate the role of biological interactions in both facilitating and resisting climate-related biodiversity change. We further highlight the potential for trophic interactions to alter the progression of both range expansions and contractions.
1. Introduction Species’ distributions have been altered by contemporary climate change, creating shifting interaction strengths and novel communities. Large-scale biodiversity change across multiple taxonomic levels involve a combination of immigration of warm species, loss of cool species and changing abundances as temperatures approach or move away from species’ thermal optima [1–3]. Given the complexity of warming-related community restructuring and the importance of biological interactions in altering trajectories [4,5], an important objective is to identify conservation efforts with potential to improve biodiversity trajectories. Evidence is building that habitat protection offers a conservation tool to promote healthy biological systems through a variety of mechanisms. For example, parks may offer important migratory corridors for species to improve connectivity between present-day and likely future biotic and environmental habitats [6]. Intact trophic webs following reduced exploitation, particularly the recovery of top predators can also provide resilience to colonization by both range shifting and introduced species [7,8]. Thus, there is some evidence for increased warming-related resilience in fish communities and targeted species in response to protection from fishing. Habitat protection may also confer resilience by simply restricting human activities. For instance, spatial surveys of Mediterranean marine reserves linked high algal canopy biomass to well-enforced no-take
& 2017 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
(a) Dive surveys Reef communities were surveyed using standardized visual census methods in marine reserves (all reserves in this study can be regarded as effective no-take areas by global standards) and fished locations in Tasmania (electronic supplementary material, figure S1, table S1). Reserves were monitored annually from at least the time of implementation in 1992 (Bicheno, Maria Island, Ninepin and Tinderbox) and 2005 (Balthurst Harbour, Kent Group and Port Davey), and thus, all reserves have had sufficient time for protection effects to manifest at multiple trophic levels [14]. Each survey covered three taxonomic components, surveyed along the same transect line (at a 5 or 10 m depth contour). For the fishes, surveys involved recording the abundance and size of all species observed within 5 m of both sides of 200 m of the transect line (encompassing 2000 m2). Large mobile invertebrates (greater than 2.5 cm long) were counted in 1 m wide bands for 200 m of one side of the transect side (encompassing 200 m2:
(b) Community temperature index and thermal affinity We calculated an independent CTI for the fish, invertebrate and macroalgal communities to contrast warming responses across each taxonomic group. The CTI has been applied in diverse communities, including for fishes on temperate rocky reefs [17 – 19], and is usually calculated as a community-weighted mean with abundance weighting. Here, we did not apply abundance weighting in order to evaluate trends in range extensions and contractions (as in [8]), with abundance considered separately—to help separate these different mechanisms of community change. Thus, our CTI values represent an average thermal affinity of all species in a local community (note Centrostephanus was excluded from the invertebrate CTI as we included the urchin as a predictor). A species-specific thermal affinity was calculated as the median of the realized temperature distribution for each species (for invertebrates and fishes) or genus (for macroalgae, to minimize any inaccuracies arising during in situ identification of algal species by divers with different skill levels). The thermal distribution (electronic supplementary material, figure S2) was based on occurrence data returned from surveys across southern Australia for two decades, and included all sites, as summarized in electronic supplementary material, table S2 (rather than only the sites repeatedly monitored through time, electronic supplementary material, table S1). The thermal distribution is therefore calculated from the full geographical range of most species, across subtropical and temperate locations in western, southern and eastern Australia. However, because some species could potentially live in colder waters if habitat were available (Tasmania represents the edge of a continent), the median is the most robust and comparable metric of the thermal distribution (see also [8]). Our approach further allowed us to match occurrence records to sea surface temperature (SST) (in space and time) determined for the different taxa over the same geographical region and sampling period. Temperature data were at 18 resolution (latitudelongitude), averaged for each month of year (http:// www.esrl.noaagov/psd/data/gridded/data.noaa.oisst.v2.html). To assist in interpretation of species-level trends, we distinguished species based on whether their thermal affinity was relatively warm or cool in relation to others included in our study. Because the classification of warm or cool affinity is relative—and based on species’ ranks (upper or lower median distribution), this classification is somewhat independent of which parameter of the thermal distribution (as described in the preceding paragraph) is selected to make a distinction of warm versus cool. We simply used the pooled thermal affinity estimates for the animal species and macroalgal genera in the study, with the median thermal affinity (16.2588 C) as a cut-off, with upper distribution classed as ‘warm affinity’ and those falling in the lower distribution as ‘cool affinity’. Our results were robust to this
2
Proc. R. Soc. B 284: 20170484
2. Material and methods
divers brushed aside the algal canopy to closely search the rock surface). Sessile invertebrate and macroalgal cover were estimated by identifying taxa found under each of 50 intersections of wire in a 0.25 m2 quadrat, placed every 10 m along the line (thus 20 quadrats were surveyed per line and these data were pooled to provide estimates for 5 m2 in total). Full methods details, including size classes for fishes and invertebrates, and macroalgal counting technique have been detailed elsewhere [15,16]. A single 200 m transect was surveyed at any given time and site, with sites considered replicates in analyses. Given that the different taxa were measured using different techniques including abundance counts versus per cent cover, and sampling effort (survey area), we have focused our results on the standardized change in observed biological responses (i.e. anomaly) rather than absolute values, for direct comparisons of relative trends across the different taxa and functional groups.
rspb.royalsocietypublishing.org
marine reserves, in addition to human disturbance gradients [9]. It is presently unknown whether protection has the potential to increase ecological resilience to warming, and thus alter warming-related biodiversity trends across diverse species and multiple trophic levels. Here, we undertake a unique analysis of a dataset that includes (i) abundance data for all taxonomic and functional groups of plants and animals that can be visually surveyed on rocky reefs, in (ii) both protected and exploited coastal rocky reef habitats over a regional scale, and (iii) over a 21-year period that has undergone recent warming (sites ¼ 60, total # surveys ¼ 837) (electronic supplementary material, figure S1, table S1). Our dataset further captures the temporal and spatial window of a poleward range expansion of a habitat-altering sea urchin (Centrostephanus rodgersii, hereafter Centrostephanus). Centrostephanus drives cascading community change by forming extensive barrens lacking a macroalgal canopy, and therefore has the potential to reshape the seascape in Tasmania. Even so, Centrostephanus is vulnerable to predation by large lobsters, and thus protected areas where lobsters are allowed to reach large sizes act to limit its colonization [10,11]. Our overall objective is to quantify how protection, warming and invasion by the urchin interact to modify dynamics of major components of temperate reef communities across a region. First, we test for a difference in the responses of the macroalgal, invertebrate and fish communities to decadal temperature oscillations and a warming trend in protected and exploited areas using an ecological indicator of community-level warming, the community temperature index (CTI). We expect an increase in the CTI during periods of warming to be driven by increases in warm-affinity species [12], with synchronous decreases in cool-affinity species due to range contractions [13]. Second, we test whether protection results in change in the functional structure of communities (structural form for macroalgae and trophic level for animals) of communities in protected and areas open to fishing, to quantify the ecological consequences and drivers of functional community change. We predict increases in the abundance of higher trophic levels, with corresponding decreases in the abundance of lower trophic levels. We further expect increases in the macroalgal canopy in response to a release from grazers, including the urchin.
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
(a)
(b) 0.6
(c)
DCTI (°C)
0.1
0.2 0
–0.1
–0.2
–0.2
–0.6 1995 2000 2005 2010 year
(d)
–0.3 1995 2000 2005 2010 year
(e)
1995 2000 2005 2010 year (f)
urchin SOI duration 0 0.05 0.10 coefficient
–0.15
–0.05 0.05 coefficient
0.15
–0.05
0
0.05 0.10 0.15 coefficient
Figure 1. Trends in CTI (relative to the overall mean) across southeast Australia between 1992 and 2012. (a– c) Change in CTI (8C), for the brown macroalgal, invertebrate and fish communities. (d – f ) Coefficients represent the effect of SST, the Southern Oscillation Index (SOI), Centrostephanus abundance and duration of protection (age of each marine reserve at the time of sampling). All fixed effects are scaled to facilitate direct comparisons of the coefficients. Grey circles indicate an effect with a 95% confidence interval overlapping zero, and black circles indicate a significant effect with intervals that exclude zero. Relationships (solid lines) and 95% confidence intervals (shading) in (a – c) are predicted using a generalized additive mixed effects model (GAMM) to account for the temporal and spatial structure present in the data, and coefficients in (d – f ) are reported in electronic supplementary material, table S4. threshold, e.g. while the response values differed slightly in magnitude, selecting the upper (warm) and lower (cool) third of species/genera for each taxon returned the same trends.
response data (log and square root), to ensure patterns were robust to our modelling approach.
(c) Functional traits
3. Results
Macroalgae were divided into the following functional groups expected to relate directly to habitat complexity: small foliose, including epiphytic species: less than 5 cm in length; canopyforming blade: greater than 5 cm in length and blade morphology; canopy-forming foliose: greater than 5 cm in length and foliose morphology; and surface canopy-forming: Macrocystis (electronic supplementary material, table S3a). Invertebrates and fishes were divided into trophic groups (electronic supplementary material, table S3b,c): herbivores, suspension feeders, omnivores or benthic invertivores for invertebrates and herbivores, planktivores, benthic invertivores or higher carnivores for fishes. Trait categories were assigned using information provided by Fishbase (http://www. fishbase.org/), Sealifebase (http://www.sealifebase.org/) and other local marine life texts [20].
A warming signature, independent of protection status, is evident across multiple taxonomic groups as an increase in CTI in the macroalgae, invertebrates and fishes (figure 1a–c, significant temperature coefficient estimate: electronic supplementary material, table S4a–c). The increase in CTI corresponds with regional warming in SST, detected as a 0.5–1.08C increase at surveyed sites over the study period. The rates of warming are reported for each 18 latitudinal bands from 38.58 to 43.58 S as slopes in electronic supplementary material, figure S3 and the model summary results table (electronic supplementary material, table S5). Decadal fluctuations in SST (evident in electronic supplementary material, ˜ o and La Nin ˜ a phases figure S3), in turn, relate to the El Nin in the Southern Hemisphere, as measured by the southern ˜ o-Southern Osciloscillation index (SOI is a metric of the El Nin lation climate mode; electronic supplementary material, figure S4). Indeed, CTI is positively related to SST, and in the case of the macroalgae negatively related to SOI, regardless of protection status (as evidenced by, respectively, positive and negative coefficient estimates in figure 1d–f). A warming signature, detected as an overall increase in CTI, is facilitated by Centrostephanus. After accounting for the spatial structure of the data (i.e. the random effects of site nested in location), CTI is higher in communities of macroalgae, invertebrates and fishes where Centrostephanus is most abundant (positive coefficient with 95% confidence limits that do not overlap zero: figure 1d–f). Moreover, Centrostephanus is an independent driver of CTI change, especially for the fishes, evidenced by comparing CTI change at locations where the
(d) Statistical modelling To test for differences in the CTI, richness and abundance through time, we used general additive and linear mixed effects models fitted using maximum likelihood, implemented, respectively, with the packages mgcv [21] and the function ‘gam’, and nlme [22] and ‘lme’ in R [23]. The random effects of survey site and location were included as nested factors to account for variation in the response variable due to the non-random spatial structure of the data (model structures, R code and data for figures 1 –3 are included as the electronic supplementary material). Year was included in all models to test for temporal trends, predicted as a response to warming. The model fit and residual structure were visually inspected to ensure that the test assumptions were met. We compared the results of models with different distribution families ( poisson/quasi-poisson and a log-link function) for count data, and transformation of the
Proc. R. Soc. B 284: 20170484
temp
–0.05
rspb.royalsocietypublishing.org
0.2
3
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
(a)
(b)
4
0.05
0.05
0.1
–0.05
–0.05
–0.1
–0.15
–0.15
–0.3 1995 2000 2005 2010
(d)
2
(e) 1.0
Dsqrt (abundance)
1995 2000 2005 2010
1
0.5
0.5
0
0
0
1995 2000 2005 2010
warm cool
Proc. R. Soc. B 284: 20170484
0.3
rspb.royalsocietypublishing.org
0.15
0.15 Dlocal richness
(c)
(f)
–1 –0.5
–2
–1.0 1995 2000 2005 2010 year
1995 2000 2005 2010 year
1995 2000 2005 2010 year
Figure 2. Richness and abundance trends contrasted for warm- and cool-affinity species across southeast Australia from 1992 to 2012. (a,c) Change in local richness and (d – f ) abundance in the brown macroagal, invertebrate and fish communities. Lines indicate trends in cool- and warm-affinity species relative to the mean, and shaded areas are the 95% CIs predicted using GAMMs. The sampling methods differed across the taxa, thus the trend rather than the absolute magnitude in richness and abundance change is of interest. See electronic supplementary material, table S7 for the model summary tables. A quasi-poisson distribution with a log link was selected to model richness trends, while the abundance data were square-root transformed. urchin’s abundance increases through time or remains low (shown in electronic supplementary material, figure S5). Specifically, there was a steeper increase in the CTI of the fish community in locations where urchins also increased the most in abundance through time (significant ‘YearUrchin’ interaction term; electronic supplementary material, table S6). We further find a dampening in the CTI increase in the macroalgae and invertebrate communities in reserves protected for longer durations (although this effect is borderline in the macroalgae), where fished sites were included as zero values with the age of each reserve at the time of sampling (figure 1d,e). Unexpectedly, protection leads to climate change resilience by supporting cooler algal and invertebrate communities (indicated by the negative coefficient for duration of the reserve: figure 1d, and as reported for the ‘Duration’ term in electronic supplementary material, table S4a,b). Even so, the same is not always the case for protected fish communities (as reported in [8]). Overall, the increase in CTI is due to an increasing frequency of occurrence of warmer-affinity species in surveys (i.e. increase in local richness: figure 2a–c, significant ‘Warm Genera’ term in electronic supplementary material, table S7a–c), in combination with increases in abundance (figure 2d–f, electronic supplementary material, table S7d–f ). While an overall decrease—compare the start and end of the monitoring period—in the relatively cooler-affinity macroalgae or fish species is not observed (figure 2b,e), there is a slight decrease in the cooler-affinity invertebrates (both richness and abundance) in comparison with observations from the early 1990s. Yet overall, biodiversity transition has been dominated by geographical expansion and population increases in warmer-affinity species with contrasting patterns in the cool and warm components of the community (e.g. figure 2c). Thus, the net community
richness and abundance responses to warming in this region have been positive (electronic supplementary material, figure S6, table S8a – c). Shifts in warm and cool species have led to changes in the functional structure of reef communities. In the macroalgal community, fluctuations in cover are present, such as for Macrocystis (the sole contributing taxon to the surface canopy-forming functional group in this region: figure 3a) that correspond with the SOI (electronic supplementary material, figure S4). In addition, linear trends are also evident. The dominant canopy-forming blade algae generally increase through time in reserves, including gains in genera with relatively cool thermal affinities, such as Durvillaea (electronic supplementary material, figure S7), but remain stable in fished areas (e.g. figure 3b; electronic supplementary material, table S9b). Conversely, small-foliose algal cover increase in sites open to fishing (figure 3d; electronic supplementary material, table S9d). Functional change across taxa also differs with protection status. Declines in the richness and abundances of mobile invertebrates (herbivores, omnivores and benthic invertivores) are greatest at sites protected from fishing (figure 3e,g,h, ‘Protected’ term is significant: electronic supplementary material, table S9e,g,h), while there was a lack of a significant decline in the fished seascape for the omnivores and benthic invertivores (figure 3g,h; electronic supplementary material, table S9g,h). By contrast, suspension-feeding invertebrates, including some sessile taxa (e.g. Herdmania grandis, a solitary sea squirt, electronic supplementary material, figure S8), increased dramatically with warming in locations open to fishing (figure 3f ). For the fishes, protected locations are notable by increases in higher carnivorous fishes (figure 3l, ‘Protected’ term is significant: electronic supplementary material, table S9l). By contrast, the herbivores, planktivores
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
5 (a)
surface canopy
(b)
blade canopy
(c)
canopy foliose
(d)
small foliose
Dabundance
0.1 0
0 –0.2
–1
–0.2
–1.0 –2
–0.6
–2.0 1995 2000 2005 2010 year
1995 2000 2005 2010 year
(e)
herbivore
0.6
(f)
suspension feeder
omnivore
0
0 –0.2
–4
(i)
herbivore
1995 2000 2005 2010 year
(j)
–0.5
–0.6
–6 1995 2000 2005 2010 year
invertivore
0.2
–2
–0.6
(h) 0.5
2
–0.2
(g)
1995 2000 2005 2010 year
0.6
4
0.2
–0.4 1995 2000 2005 2010 year
protected open
planktivore
1995 2000 2005 2010 year
(k)
benthic invertivore
1995 2000 2005 2010 year
(l)
higher carnivore
Dabundance
0.6 0.5
0.4
0.2
0
0.2
0
–0.5
0.4
0.5 –0.5
0 –0.2
–0.2
–1.5
–1.0 1995 2000 2005 2010 year
1995 2000 2005 2010 year
1995 2000 2005 2010 year
1995 2000 2005 2010 year
Figure 3. Abundance trends in functional groups in communities protected from fishing and those open to fishing across southeast Australia from 1992 to 2012. (a – d ) Macroalgae, (e– h) invertebrates and (i – l ) fishes. GAMMs (quasi-poisson distribution with a log link) were fitted independently for locations open to fishing (dotted line) and under protection (solid line). The 95% CIs are shaded for significant trends. The model results are summarized in electronic supplementary material, table S9. and benthic invertivores displayed similar trends in locations open to exploitation to those under protection, with general increases in the herbivores and benthic invertivores (figure 3i,j,k; electronic supplementary material, table S9i,j,k).
4. Discussion While there is limited empirical evidence that temperate marine reserves increase community resilience (although see [8]), we find that under warming, some protected macroalgal communities display a lower rate of tropicalization in comparison to fished locations. Moreover, the cover of the blade canopy increased through time in the protected locations only, suggesting protection is a mechanism with potential to promote macroalgal persistence. Clearly, the aggregation of the urchin (Centrostephanus), a species that grazes down the macroalgal canopy and maintains patches of bare rock, is a driver of macroalgal canopy loss. Yet, even in comparison to fished locations where the urchin is absent, greater gains in macroalgae are observed in the protected areas, including genera with relatively cool thermal affinities. Simultaneously, reserves generally resist establishment of small-foliose algal communities, typical of lower latitudes [24]. Thus, while reserves limit urchin populations through predation [10] (as discussed in more detail below), our work suggests additional mechanisms are at play. For instance, shifts in species interactions following protection include a decrease in other herbivorous invertebrates in marine reserves or changes in light levels with growth of the upper canopy. Greater biomass of canopy-forming macroalgae is also associated with reduced human-related disturbance in the Mediterranean [9],
suggesting that reducing human pressures, such as through protection, may offer a generally useful management tool for prolonging the persistence of this critical habitat type under climate change. While our previous work predicts that protection will buffer CTI increase in the fishes (Community Thermal Niche: [8]), we did not find a buffering response in all of the reserves. This may be because monitoring in the Maria Island Marine Reserve includes survey locations that can be directly compared with fished ‘reference’ sites that were carefully selected at the start of monitoring based on habitat and community similarity to the reserve, allowing for greater power to detect differences in community responses through time. In the Maria Island reserve, population recovery of larger bodied fishes, which also tend to have relatively lower thermal affinities, led to a dampening of the warming-related community thermal signal through time [8]. Inconsistencies in the biological responses in reserves may also be due to variability in environmental and habitat conditions among locations. The ecological effectiveness of different reserves can also vary markedly due to factors such as age and size [25,26], and thus, the responses of protected communities to warming may be variable, supporting the need for ongoing monitoring to inform decision-making. We further find significant declines in richness and abundance of invertebrates with protection, including the herbivores, omnivores and invertivores. At the same time, the richness and abundance of fishes are increasing—a greater diversity and abundance of fishes very likely translates to greater rates of predation. Increases in both carnivorous fishes and lobsters are also typical with recovery following historical exploitation and were observed here (figure 3l;
Proc. R. Soc. B 284: 20170484
Dabundance
0.2
0.2
0
rspb.royalsocietypublishing.org
0.4
0.2 1
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
Data accessibility. Data are provided as the electronic supplementary material.
Authors’ contributions. All authors contributed to the conceptual development of the manuscript. A.E.B. analysed the data and drafted the manuscript and figures, with significant input from all authors. N.S.B., G.J.E. and others collected the data.
Competing interests. We declare we have no competing interests. Funding. Salary support for R.D.S.-S. was provided by an ARC Linkage grant and through the Marine Biodiversity Hub, a collaborative partnership supported through the Australian Government’s National Environmental Science Programme. Acknowledgements. We thank the many field technicians who assisted with diving surveys over the 21-year monitoring period, P. Bosworth and A. Schaap for initiating MPA surveys and C. Buxton for operational support. Biological monitoring was supported by grants from the Australian and Tasmanian governments (Australian Research Council; Department of Primary Industries, Parks, Water and Environment; Fisheries Research and Development Corporation; National Resource Management and Natural Heritage Trust) with considerable logistic support from Tasmanian Parks and Wildlife Service. We appreciate comments from C. Waldock on a manuscript draft.
References 1.
Molinos JG et al. 2015 Climate velocity and the future global redistribution of marine biodiversity.
Nat. Clim. Change 6, 83– 88. (doi:10.1038/ nclimate2769)
2.
Poloczanska ES et al. 2013 Global imprint of climate change on marine life. Nat.
6
Proc. R. Soc. B 284: 20170484
transition of the community by facilitating warm-affinity fishes and barren formation, which may in turn influence colonization dynamics in other poleward shifting invertebrates. Patterns in the abundance of different functional groups further suggest the role of climate variability in mediating community patterns, over and above the previously documented effects of trophic cascades resulting from protection [14]. Marked fluctuations are evident over decadal timescales in multiple dimensions of diversity, including CTI, local richness and abundance, and also for abundance of the different functional groups, such as Macrocystis. These distinct fluctuations relate to climate signals (e.g. SOI) and thus change in numerous related physical variables. For example, warm years in this region are associated with lower nutrients when the oligotrophic East Australian Current strengthens in a polewards direction [10]. These conditions may impose physiological constraints on kelp performance [32] that decrease in warmer, low-nutrient conditions, therefore explaining change in cover of Macrocystis throughout the study period and recent decline. Interpreting the complexity inherent to the temporal biodiversity signal requires consideration of natural climate variability, including extreme events. Moreover, while our analysis focuses on warming, factors such as ocean acidification and associated ecological consequences may also influence biological responses, an important direction for future research [33,34]. Multi-decadal time-series records are integral for teasing apart both the drivers for change in community structure and the components of biodiversity that are responding, and reveal that net community outcomes are strongly affected by protection from exploitation. Our findings offer general support for the role of protection in facilitating greater cover of canopy-forming macroalgae and resisting colonization by barren-forming urchins. However, we also observe a marked decline in the invertebrates following the implementation of reserves, implicating the importance of biotic interactions in driving range shifts and community functional change, an important consideration for park managers.
rspb.royalsocietypublishing.org
electronic supplementary material, figure S9), and presumably impose additional predation pressure. Our findings therefore implicate the importance of higher trophic levels and larger individuals [15,27] in biological interactions exacerbating warming-related processes of community change, such as geographical range contractions. Increases in abundance and richness across trophic levels with warming, including species with relatively cooler temperature affinities, in both protected and exploited locations suggest that warming can increase local-scale diversity. While populations of species living at the edges of their thermal tolerance may decline with warming [28,29], the cover of cooler-affinity macroalgae in Tasmania has not undergone a decrease since the start of monitoring and has increased during the past 10 years of warming. We find a similar pattern in the fishes, which may be underpinned by processes such as local adaptation. Thus, regardless of protection status, many species appear to have benefited and increased in abundance and occupancy, with recent warming. Indeed, richness increases with temperature are expected in temperate systems due to poleward geographical expansion of richer faunas from low latitudes [1,30]. The southeast Australian reef fish and invertebrate communities have a positive thermal bias [31] and, on average, are comprised of species with warmer affinity than the local environment (i.e. Tasmanian reef species also occur in much warmer regions). Thus, much of the fauna is poised to perform well with warming, at least with the magnitude of temperature change in this region to date. Even so, warming-related losses in diversity have been observed in other systems, such as in a study from Sweden documenting local extirpation in coldaffinity birds with warming and an increase in CTI [29]. Unlike the fauna, macroalgal diversity is higher in southern Australia than further north, so the flora is not predicted to increase in abundance and richness with warming, and may already be incurring climate debt. We confirm the importance of the urchin in facilitating a tropicalization of the reef communities across the region with quantitative data on the abundance of Centrostephanus as a predictor in a statistical modelling framework. In doing so, we offer the insight that the urchin may facilitate warm-affinity species and colonization by functional groups typical of lower latitudes. In particular, sessile suspension-feeding invertebrates (e.g. electronic supplementary material, figure S8) increase dramatically with warming, corresponding with increases in Centrostephanus, perhaps because bare habitat may offer colonization patches. As many of the suspensionfeeding species are typical of latitudes with warmer temperatures, the relative increase in this functional group explains the relatively large increase in the CTI of invertebrates relative to the macroalgae and fishes (respectively, approx. 0.88C versus 0.38C and 0.158C, as displayed in figure 1a–c). Another important driver of change likely relates to preferential settlement of warmer-affinity species to urchin barrens habitats rather than kelp-dominated reef. Our results therefore indicate that the poleward expansion of Centrostephanus is resulting in directional ecological change—i.e. the urchin is driving a
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
4.
5.
7.
8.
9.
10.
11.
12.
13.
15.
16.
17.
18.
19.
20. 21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
changes in herbivory and community phase shifts. Proc. R. Soc. B 281, 20140846. (doi:10.1098/rspb. 2014.0846) Edgar GJ et al. 2014 Global conservation outcomes depend on marine protected areas with five key features. Nature 506, 216 –220. (doi:10.1038/ nature13022) Edgar GJ, Barrett NS, Stuart-Smith RD. 2009 Exploited reefs protected from fishing transform over decades into conservation features otherwise absent from seascapes. Ecol. Appl. 19, 1967–1974. (doi:10.1890/09-0610.1) Sciherras M, Jenkins SR, Kaiser MJ, Hawkins SJ, Pullin AS. 2013 Evaluating the biological effectiveness of fully and partially protected marine areas. Environ. Evid. 2, 4. (doi:10.1186/2047-2382-2-4) Wernberg T et al. 2016 Climate-driven regime shift of a temperate marine ecosystem. Science 353, 169–172. (doi:10.1126/science.aad8745) Wernberg T, Russell BD, Thomsen MS, Gurgel CD, Bradshaw CJA, Poloczanska ES, Connell SD. 2011 Seaweed communities in retreat from ocean warming. Curr. Biol. 21, 1828–1832. (doi:10.1016/ j.cub.2011.09.028) Elahi R, O’Connor M, Byrnes JEK, Dunic J, Eriksson BK, Hensel MJS, Kearns PJ. 2015 Recent trends in local-scale marine biodiversity reflect community structure and human impacts. Curr. Biol. 25, 1938– 1943. (doi:10.1016/j.cub.2015.05.030) Stuart-Smith RD, Edgar GJ, Barrett NS, Kininmonth SJ, Bates AE. 2015 Thermal biases and vulnerability to warming in the world’s marine fauna. Nature 528, 88– 92. (doi:10.1038/nature16144) Dayton P, Tegner M, Edwards P, Riser K. 1999 Temporal and spatial scales of kelp demography: the role of oceanographic climate. Ecol. Monogr. 69, 219–250. (doi:10.1890/0012-9615(1999)069 [0219:TASSOK]2.0.CO;2) Connell SD, Kroeker KJ, Fabricius KE, Kline DI, Russell BD. 2013 The other ocean acidification problem: CO2 as a resource among competitors for ecosystem dominance. Phil. Trans. R. Soc. B. 368, 20120442. (doi:10.1098/rstb.2012.0442) Porzio L, Buia MC, Hall-Spencer JM. 2011 Effects of ocean acidification on macroalgal communities. J. Exp. Mar. Biol. Ecol. 400, 278–287. (doi:10.1016/ j.jembe.2011.02.011)
7
Proc. R. Soc. B 284: 20170484
6.
14.
Glob. Environ. Change 26, 27 –38. (doi:10.1016/j. gloenvcha.2014.03.009) Babcock RC, Shears NT, Alcala A, Barrett NS, Edgar GJ, Lafferty KD, McClanahan TR, Russ GR. 2010 Decadal trends in marine reserves reveal differential rates of change in direct and indirect effects. Proc. Natl Acad. Sci. USA 107, 18 251–18 255. (doi:10. 1073/pnas.0908012107) Barrett NS, Buxton CD, Edgar GJ. 2009 Changes in invertebrate and macroalgal populations in Tasmanian marine reserves in the decade following protection. J. Exp. Mar. Biol. Ecol. 370, 104–119. (doi:10.1016/j.jembe.2008.12.005) Stuart-Smith RD, Barrett NS, Stevenson DG, Edgar GJ. 2010 Stability in temperate reef communities over a decadal time scale despite concurrent ocean warming. Glob. Change Biol. 16, 122–134. (doi:10. 1111/j.1365-2486.2009.01955.x) Stevens JT, Safford HD, Harrison S, Latimer AM. 2015 Forest disturbance accelerates thermophilization of understory plant communities. J. Ecol. 103, 1253–1263. (doi:10.1111/1365-2745.12426) Devictor V, Julliard R, Couvet D, Jiguet F. 2008 Birds are tracking climate warming, but not fast enough. Proc. R. Soc. B 275, 2743 –2748. (doi:10.1098/rspb. 2008.0878) Stuart-Smith RD et al. 2017 Assessing national biodiversity trends for rocky and coral reefs through the integration of citizen science and scientific monitoring programs. Bioscience 67, 134– 146. (doi:10.1093/biosci/biw180) Edgar GJ. 2008 Australian marine life, revised edn. Melbourne, Australia: Reed Books. Wood SN. 2011 Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. R. Stat. Soc. B 73, 3 –36. (doi:10.1111/j.1467-9868.2010. 00749.x) Pinheiro J, Bates D, DebRoy S, Sarkar D, Team RC. 2015 nlme: linear and nonlinear mixed effects models. See http://CRAN.R-project.org/ package=nlme. R Core Team. 2014 R: a language and environment for statistical computing. See http://www.R-project. org/. Verge´s A et al. 2014 The tropicalization of temperate marine ecosystems: climate-mediated
rspb.royalsocietypublishing.org
3.
Clim. Change 3, 919 –925. (doi:10.1038/ nclimate1958) Chen IC, Hill JK, Ohlemueller R, Roy DB, Thomas CD. 2011 Rapid range shifts of species associated with high levels of climate warming. Science 333, 1024–1026. (doi:10.1126/science.1206432) Suttle KB, Thomsen MA, Power ME. 2008 Species interactions reverse grassland responses to changing climate. Science 315, 640–642. (doi:10.1126/ science.1136401) Cahill AE et al. 2012 How does climate change cause extinction? Proc. R. Soc. B 280, 20121890. (doi:10.1098/rspb.2012.1890) Thomas CD et al. 2012 Protected areas facilitate species’ range expansions. Proc. Natl Acad. Sci. USA 109, 14 063–14 068. (doi:10.1073/pnas. 1210251109) Teo DHL, Tan HTW, Corlett RT, Wong CM, Lum SKY. 2003 Continental rain forest fragments in Singapore resist invasion by exotic plants. J. Biogeogr. 30, 305–310. (doi:10.1046/j.1365-2699.2003.00813.x) Bates AE, Barrett NS, Stuart-Smith RD, Holbrook NJ, Thompson PA, Edgar GJ. 2014 Resilience and signatures of tropicalization in protected reef fish communities. Nat. Clim. Change 4, 62 –67. (doi:10. 1038/nclimate2062) Sala E et al. 2012 The structure of Mediterranean rocky reef ecosystems across environmental and human gradients, and conservation implications. PLoS ONE 7, e32742. (doi:10.1371/journal.pone. 0032742) Ling SD, Johnson CR, Frusher SD, Ridgway KR. 2009 Overfishing reduces resilience of kelp beds to climate-driven catastrophic phase shift. Proc. Natl Acad. Sci. USA 106, 22 341 –22 345. (doi:10.1073/ pnas.0907529106) Johnson CR et al. 2011 Climate change cascades: shifts in oceanography, species’ ranges and subtidal marine community dynamics in eastern Tasmania. J. Exp. Mar. Biol. Ecol. 400, 17 –32. (doi:10.1016/j. jembe.2011.02.032) Sunday JM et al. 2015 Species traits and climate velocity explain geographic range shifts in an ocean warming hotspot. Ecol. Lett. 18, 944–953. (doi:10. 1111/ele.12474) Bates AE et al. 2014 Defining and observing stages of climate-mediated range shifts in marine systems.
rspb.royalsocietypublishing.org
Research Cite this article: Bates AE, Stuart-Smith RD, Barrett NS, Edgar GJ. 2017 Biological interactions both facilitate and resist climaterelated functional change in temperate reef communities. Proc. R. Soc. B 284: 20170484. http://dx.doi.org/10.1098/rspb.2017.0484
Received: 7 March 2017 Accepted: 5 May 2017
Subject Category: Global change and conservation Subject Areas: ecology Keywords: climate change, community temperature index, functional traits, marine protected area, tropicalization
Author for correspondence: Amanda E. Bates e-mail: [email protected]
Electronic supplementary material is available online at https://dx.doi.org/10.6084/m9. figshare.c.3782237.
Biological interactions both facilitate and resist climate-related functional change in temperate reef communities Amanda E. Bates1, Rick D. Stuart-Smith2, Neville S. Barrett2 and Graham J. Edgar2 1 2
National Oceanography Centre, University of Southampton, Waterfront Campus, Southampton SO14 3ZH, UK Institute for Marine and Antarctic Studies, University of Tasmania, Hobart 7001, Tasmania, Australia AEB, 0000-0002-0198-4537 Shifts in the abundance and location of species are restructuring life on the Earth, presenting the need to build resilience into our natural systems. Here, we tested if protection from fishing promotes community resilience in temperate reef communities undergoing rapid warming in Tasmania. Regardless of protection status, we detected a signature of warming in the brown macroalgae, invertebrates and fishes, through increases in the local richness and abundance of warm-affinity species. Even so, responses in protected communities diverged from exploited communities. At the local scale, the number of cool-affinity fishes and canopy-forming algal species increased following protection, even though the observation window fell within a period of warming. At the same time, exploited communities gained turf algal and sessile invertebrate species. We further found that the recovery of predator populations following protection leads to marked declines in mobile invertebrates— this trend could be incorrectly attributed to warming without contextual data quantifying community change across trophic levels. By comparing long-term change in exploited and protected reefs, we empirically demonstrate the role of biological interactions in both facilitating and resisting climate-related biodiversity change. We further highlight the potential for trophic interactions to alter the progression of both range expansions and contractions.
1. Introduction Species’ distributions have been altered by contemporary climate change, creating shifting interaction strengths and novel communities. Large-scale biodiversity change across multiple taxonomic levels involve a combination of immigration of warm species, loss of cool species and changing abundances as temperatures approach or move away from species’ thermal optima [1–3]. Given the complexity of warming-related community restructuring and the importance of biological interactions in altering trajectories [4,5], an important objective is to identify conservation efforts with potential to improve biodiversity trajectories. Evidence is building that habitat protection offers a conservation tool to promote healthy biological systems through a variety of mechanisms. For example, parks may offer important migratory corridors for species to improve connectivity between present-day and likely future biotic and environmental habitats [6]. Intact trophic webs following reduced exploitation, particularly the recovery of top predators can also provide resilience to colonization by both range shifting and introduced species [7,8]. Thus, there is some evidence for increased warming-related resilience in fish communities and targeted species in response to protection from fishing. Habitat protection may also confer resilience by simply restricting human activities. For instance, spatial surveys of Mediterranean marine reserves linked high algal canopy biomass to well-enforced no-take
& 2017 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
(a) Dive surveys Reef communities were surveyed using standardized visual census methods in marine reserves (all reserves in this study can be regarded as effective no-take areas by global standards) and fished locations in Tasmania (electronic supplementary material, figure S1, table S1). Reserves were monitored annually from at least the time of implementation in 1992 (Bicheno, Maria Island, Ninepin and Tinderbox) and 2005 (Balthurst Harbour, Kent Group and Port Davey), and thus, all reserves have had sufficient time for protection effects to manifest at multiple trophic levels [14]. Each survey covered three taxonomic components, surveyed along the same transect line (at a 5 or 10 m depth contour). For the fishes, surveys involved recording the abundance and size of all species observed within 5 m of both sides of 200 m of the transect line (encompassing 2000 m2). Large mobile invertebrates (greater than 2.5 cm long) were counted in 1 m wide bands for 200 m of one side of the transect side (encompassing 200 m2:
(b) Community temperature index and thermal affinity We calculated an independent CTI for the fish, invertebrate and macroalgal communities to contrast warming responses across each taxonomic group. The CTI has been applied in diverse communities, including for fishes on temperate rocky reefs [17 – 19], and is usually calculated as a community-weighted mean with abundance weighting. Here, we did not apply abundance weighting in order to evaluate trends in range extensions and contractions (as in [8]), with abundance considered separately—to help separate these different mechanisms of community change. Thus, our CTI values represent an average thermal affinity of all species in a local community (note Centrostephanus was excluded from the invertebrate CTI as we included the urchin as a predictor). A species-specific thermal affinity was calculated as the median of the realized temperature distribution for each species (for invertebrates and fishes) or genus (for macroalgae, to minimize any inaccuracies arising during in situ identification of algal species by divers with different skill levels). The thermal distribution (electronic supplementary material, figure S2) was based on occurrence data returned from surveys across southern Australia for two decades, and included all sites, as summarized in electronic supplementary material, table S2 (rather than only the sites repeatedly monitored through time, electronic supplementary material, table S1). The thermal distribution is therefore calculated from the full geographical range of most species, across subtropical and temperate locations in western, southern and eastern Australia. However, because some species could potentially live in colder waters if habitat were available (Tasmania represents the edge of a continent), the median is the most robust and comparable metric of the thermal distribution (see also [8]). Our approach further allowed us to match occurrence records to sea surface temperature (SST) (in space and time) determined for the different taxa over the same geographical region and sampling period. Temperature data were at 18 resolution (latitudelongitude), averaged for each month of year (http:// www.esrl.noaagov/psd/data/gridded/data.noaa.oisst.v2.html). To assist in interpretation of species-level trends, we distinguished species based on whether their thermal affinity was relatively warm or cool in relation to others included in our study. Because the classification of warm or cool affinity is relative—and based on species’ ranks (upper or lower median distribution), this classification is somewhat independent of which parameter of the thermal distribution (as described in the preceding paragraph) is selected to make a distinction of warm versus cool. We simply used the pooled thermal affinity estimates for the animal species and macroalgal genera in the study, with the median thermal affinity (16.2588 C) as a cut-off, with upper distribution classed as ‘warm affinity’ and those falling in the lower distribution as ‘cool affinity’. Our results were robust to this
2
Proc. R. Soc. B 284: 20170484
2. Material and methods
divers brushed aside the algal canopy to closely search the rock surface). Sessile invertebrate and macroalgal cover were estimated by identifying taxa found under each of 50 intersections of wire in a 0.25 m2 quadrat, placed every 10 m along the line (thus 20 quadrats were surveyed per line and these data were pooled to provide estimates for 5 m2 in total). Full methods details, including size classes for fishes and invertebrates, and macroalgal counting technique have been detailed elsewhere [15,16]. A single 200 m transect was surveyed at any given time and site, with sites considered replicates in analyses. Given that the different taxa were measured using different techniques including abundance counts versus per cent cover, and sampling effort (survey area), we have focused our results on the standardized change in observed biological responses (i.e. anomaly) rather than absolute values, for direct comparisons of relative trends across the different taxa and functional groups.
rspb.royalsocietypublishing.org
marine reserves, in addition to human disturbance gradients [9]. It is presently unknown whether protection has the potential to increase ecological resilience to warming, and thus alter warming-related biodiversity trends across diverse species and multiple trophic levels. Here, we undertake a unique analysis of a dataset that includes (i) abundance data for all taxonomic and functional groups of plants and animals that can be visually surveyed on rocky reefs, in (ii) both protected and exploited coastal rocky reef habitats over a regional scale, and (iii) over a 21-year period that has undergone recent warming (sites ¼ 60, total # surveys ¼ 837) (electronic supplementary material, figure S1, table S1). Our dataset further captures the temporal and spatial window of a poleward range expansion of a habitat-altering sea urchin (Centrostephanus rodgersii, hereafter Centrostephanus). Centrostephanus drives cascading community change by forming extensive barrens lacking a macroalgal canopy, and therefore has the potential to reshape the seascape in Tasmania. Even so, Centrostephanus is vulnerable to predation by large lobsters, and thus protected areas where lobsters are allowed to reach large sizes act to limit its colonization [10,11]. Our overall objective is to quantify how protection, warming and invasion by the urchin interact to modify dynamics of major components of temperate reef communities across a region. First, we test for a difference in the responses of the macroalgal, invertebrate and fish communities to decadal temperature oscillations and a warming trend in protected and exploited areas using an ecological indicator of community-level warming, the community temperature index (CTI). We expect an increase in the CTI during periods of warming to be driven by increases in warm-affinity species [12], with synchronous decreases in cool-affinity species due to range contractions [13]. Second, we test whether protection results in change in the functional structure of communities (structural form for macroalgae and trophic level for animals) of communities in protected and areas open to fishing, to quantify the ecological consequences and drivers of functional community change. We predict increases in the abundance of higher trophic levels, with corresponding decreases in the abundance of lower trophic levels. We further expect increases in the macroalgal canopy in response to a release from grazers, including the urchin.
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
(a)
(b) 0.6
(c)
DCTI (°C)
0.1
0.2 0
–0.1
–0.2
–0.2
–0.6 1995 2000 2005 2010 year
(d)
–0.3 1995 2000 2005 2010 year
(e)
1995 2000 2005 2010 year (f)
urchin SOI duration 0 0.05 0.10 coefficient
–0.15
–0.05 0.05 coefficient
0.15
–0.05
0
0.05 0.10 0.15 coefficient
Figure 1. Trends in CTI (relative to the overall mean) across southeast Australia between 1992 and 2012. (a– c) Change in CTI (8C), for the brown macroalgal, invertebrate and fish communities. (d – f ) Coefficients represent the effect of SST, the Southern Oscillation Index (SOI), Centrostephanus abundance and duration of protection (age of each marine reserve at the time of sampling). All fixed effects are scaled to facilitate direct comparisons of the coefficients. Grey circles indicate an effect with a 95% confidence interval overlapping zero, and black circles indicate a significant effect with intervals that exclude zero. Relationships (solid lines) and 95% confidence intervals (shading) in (a – c) are predicted using a generalized additive mixed effects model (GAMM) to account for the temporal and spatial structure present in the data, and coefficients in (d – f ) are reported in electronic supplementary material, table S4. threshold, e.g. while the response values differed slightly in magnitude, selecting the upper (warm) and lower (cool) third of species/genera for each taxon returned the same trends.
response data (log and square root), to ensure patterns were robust to our modelling approach.
(c) Functional traits
3. Results
Macroalgae were divided into the following functional groups expected to relate directly to habitat complexity: small foliose, including epiphytic species: less than 5 cm in length; canopyforming blade: greater than 5 cm in length and blade morphology; canopy-forming foliose: greater than 5 cm in length and foliose morphology; and surface canopy-forming: Macrocystis (electronic supplementary material, table S3a). Invertebrates and fishes were divided into trophic groups (electronic supplementary material, table S3b,c): herbivores, suspension feeders, omnivores or benthic invertivores for invertebrates and herbivores, planktivores, benthic invertivores or higher carnivores for fishes. Trait categories were assigned using information provided by Fishbase (http://www. fishbase.org/), Sealifebase (http://www.sealifebase.org/) and other local marine life texts [20].
A warming signature, independent of protection status, is evident across multiple taxonomic groups as an increase in CTI in the macroalgae, invertebrates and fishes (figure 1a–c, significant temperature coefficient estimate: electronic supplementary material, table S4a–c). The increase in CTI corresponds with regional warming in SST, detected as a 0.5–1.08C increase at surveyed sites over the study period. The rates of warming are reported for each 18 latitudinal bands from 38.58 to 43.58 S as slopes in electronic supplementary material, figure S3 and the model summary results table (electronic supplementary material, table S5). Decadal fluctuations in SST (evident in electronic supplementary material, ˜ o and La Nin ˜ a phases figure S3), in turn, relate to the El Nin in the Southern Hemisphere, as measured by the southern ˜ o-Southern Osciloscillation index (SOI is a metric of the El Nin lation climate mode; electronic supplementary material, figure S4). Indeed, CTI is positively related to SST, and in the case of the macroalgae negatively related to SOI, regardless of protection status (as evidenced by, respectively, positive and negative coefficient estimates in figure 1d–f). A warming signature, detected as an overall increase in CTI, is facilitated by Centrostephanus. After accounting for the spatial structure of the data (i.e. the random effects of site nested in location), CTI is higher in communities of macroalgae, invertebrates and fishes where Centrostephanus is most abundant (positive coefficient with 95% confidence limits that do not overlap zero: figure 1d–f). Moreover, Centrostephanus is an independent driver of CTI change, especially for the fishes, evidenced by comparing CTI change at locations where the
(d) Statistical modelling To test for differences in the CTI, richness and abundance through time, we used general additive and linear mixed effects models fitted using maximum likelihood, implemented, respectively, with the packages mgcv [21] and the function ‘gam’, and nlme [22] and ‘lme’ in R [23]. The random effects of survey site and location were included as nested factors to account for variation in the response variable due to the non-random spatial structure of the data (model structures, R code and data for figures 1 –3 are included as the electronic supplementary material). Year was included in all models to test for temporal trends, predicted as a response to warming. The model fit and residual structure were visually inspected to ensure that the test assumptions were met. We compared the results of models with different distribution families ( poisson/quasi-poisson and a log-link function) for count data, and transformation of the
Proc. R. Soc. B 284: 20170484
temp
–0.05
rspb.royalsocietypublishing.org
0.2
3
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
(a)
(b)
4
0.05
0.05
0.1
–0.05
–0.05
–0.1
–0.15
–0.15
–0.3 1995 2000 2005 2010
(d)
2
(e) 1.0
Dsqrt (abundance)
1995 2000 2005 2010
1
0.5
0.5
0
0
0
1995 2000 2005 2010
warm cool
Proc. R. Soc. B 284: 20170484
0.3
rspb.royalsocietypublishing.org
0.15
0.15 Dlocal richness
(c)
(f)
–1 –0.5
–2
–1.0 1995 2000 2005 2010 year
1995 2000 2005 2010 year
1995 2000 2005 2010 year
Figure 2. Richness and abundance trends contrasted for warm- and cool-affinity species across southeast Australia from 1992 to 2012. (a,c) Change in local richness and (d – f ) abundance in the brown macroagal, invertebrate and fish communities. Lines indicate trends in cool- and warm-affinity species relative to the mean, and shaded areas are the 95% CIs predicted using GAMMs. The sampling methods differed across the taxa, thus the trend rather than the absolute magnitude in richness and abundance change is of interest. See electronic supplementary material, table S7 for the model summary tables. A quasi-poisson distribution with a log link was selected to model richness trends, while the abundance data were square-root transformed. urchin’s abundance increases through time or remains low (shown in electronic supplementary material, figure S5). Specifically, there was a steeper increase in the CTI of the fish community in locations where urchins also increased the most in abundance through time (significant ‘YearUrchin’ interaction term; electronic supplementary material, table S6). We further find a dampening in the CTI increase in the macroalgae and invertebrate communities in reserves protected for longer durations (although this effect is borderline in the macroalgae), where fished sites were included as zero values with the age of each reserve at the time of sampling (figure 1d,e). Unexpectedly, protection leads to climate change resilience by supporting cooler algal and invertebrate communities (indicated by the negative coefficient for duration of the reserve: figure 1d, and as reported for the ‘Duration’ term in electronic supplementary material, table S4a,b). Even so, the same is not always the case for protected fish communities (as reported in [8]). Overall, the increase in CTI is due to an increasing frequency of occurrence of warmer-affinity species in surveys (i.e. increase in local richness: figure 2a–c, significant ‘Warm Genera’ term in electronic supplementary material, table S7a–c), in combination with increases in abundance (figure 2d–f, electronic supplementary material, table S7d–f ). While an overall decrease—compare the start and end of the monitoring period—in the relatively cooler-affinity macroalgae or fish species is not observed (figure 2b,e), there is a slight decrease in the cooler-affinity invertebrates (both richness and abundance) in comparison with observations from the early 1990s. Yet overall, biodiversity transition has been dominated by geographical expansion and population increases in warmer-affinity species with contrasting patterns in the cool and warm components of the community (e.g. figure 2c). Thus, the net community
richness and abundance responses to warming in this region have been positive (electronic supplementary material, figure S6, table S8a – c). Shifts in warm and cool species have led to changes in the functional structure of reef communities. In the macroalgal community, fluctuations in cover are present, such as for Macrocystis (the sole contributing taxon to the surface canopy-forming functional group in this region: figure 3a) that correspond with the SOI (electronic supplementary material, figure S4). In addition, linear trends are also evident. The dominant canopy-forming blade algae generally increase through time in reserves, including gains in genera with relatively cool thermal affinities, such as Durvillaea (electronic supplementary material, figure S7), but remain stable in fished areas (e.g. figure 3b; electronic supplementary material, table S9b). Conversely, small-foliose algal cover increase in sites open to fishing (figure 3d; electronic supplementary material, table S9d). Functional change across taxa also differs with protection status. Declines in the richness and abundances of mobile invertebrates (herbivores, omnivores and benthic invertivores) are greatest at sites protected from fishing (figure 3e,g,h, ‘Protected’ term is significant: electronic supplementary material, table S9e,g,h), while there was a lack of a significant decline in the fished seascape for the omnivores and benthic invertivores (figure 3g,h; electronic supplementary material, table S9g,h). By contrast, suspension-feeding invertebrates, including some sessile taxa (e.g. Herdmania grandis, a solitary sea squirt, electronic supplementary material, figure S8), increased dramatically with warming in locations open to fishing (figure 3f ). For the fishes, protected locations are notable by increases in higher carnivorous fishes (figure 3l, ‘Protected’ term is significant: electronic supplementary material, table S9l). By contrast, the herbivores, planktivores
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
5 (a)
surface canopy
(b)
blade canopy
(c)
canopy foliose
(d)
small foliose
Dabundance
0.1 0
0 –0.2
–1
–0.2
–1.0 –2
–0.6
–2.0 1995 2000 2005 2010 year
1995 2000 2005 2010 year
(e)
herbivore
0.6
(f)
suspension feeder
omnivore
0
0 –0.2
–4
(i)
herbivore
1995 2000 2005 2010 year
(j)
–0.5
–0.6
–6 1995 2000 2005 2010 year
invertivore
0.2
–2
–0.6
(h) 0.5
2
–0.2
(g)
1995 2000 2005 2010 year
0.6
4
0.2
–0.4 1995 2000 2005 2010 year
protected open
planktivore
1995 2000 2005 2010 year
(k)
benthic invertivore
1995 2000 2005 2010 year
(l)
higher carnivore
Dabundance
0.6 0.5
0.4
0.2
0
0.2
0
–0.5
0.4
0.5 –0.5
0 –0.2
–0.2
–1.5
–1.0 1995 2000 2005 2010 year
1995 2000 2005 2010 year
1995 2000 2005 2010 year
1995 2000 2005 2010 year
Figure 3. Abundance trends in functional groups in communities protected from fishing and those open to fishing across southeast Australia from 1992 to 2012. (a – d ) Macroalgae, (e– h) invertebrates and (i – l ) fishes. GAMMs (quasi-poisson distribution with a log link) were fitted independently for locations open to fishing (dotted line) and under protection (solid line). The 95% CIs are shaded for significant trends. The model results are summarized in electronic supplementary material, table S9. and benthic invertivores displayed similar trends in locations open to exploitation to those under protection, with general increases in the herbivores and benthic invertivores (figure 3i,j,k; electronic supplementary material, table S9i,j,k).
4. Discussion While there is limited empirical evidence that temperate marine reserves increase community resilience (although see [8]), we find that under warming, some protected macroalgal communities display a lower rate of tropicalization in comparison to fished locations. Moreover, the cover of the blade canopy increased through time in the protected locations only, suggesting protection is a mechanism with potential to promote macroalgal persistence. Clearly, the aggregation of the urchin (Centrostephanus), a species that grazes down the macroalgal canopy and maintains patches of bare rock, is a driver of macroalgal canopy loss. Yet, even in comparison to fished locations where the urchin is absent, greater gains in macroalgae are observed in the protected areas, including genera with relatively cool thermal affinities. Simultaneously, reserves generally resist establishment of small-foliose algal communities, typical of lower latitudes [24]. Thus, while reserves limit urchin populations through predation [10] (as discussed in more detail below), our work suggests additional mechanisms are at play. For instance, shifts in species interactions following protection include a decrease in other herbivorous invertebrates in marine reserves or changes in light levels with growth of the upper canopy. Greater biomass of canopy-forming macroalgae is also associated with reduced human-related disturbance in the Mediterranean [9],
suggesting that reducing human pressures, such as through protection, may offer a generally useful management tool for prolonging the persistence of this critical habitat type under climate change. While our previous work predicts that protection will buffer CTI increase in the fishes (Community Thermal Niche: [8]), we did not find a buffering response in all of the reserves. This may be because monitoring in the Maria Island Marine Reserve includes survey locations that can be directly compared with fished ‘reference’ sites that were carefully selected at the start of monitoring based on habitat and community similarity to the reserve, allowing for greater power to detect differences in community responses through time. In the Maria Island reserve, population recovery of larger bodied fishes, which also tend to have relatively lower thermal affinities, led to a dampening of the warming-related community thermal signal through time [8]. Inconsistencies in the biological responses in reserves may also be due to variability in environmental and habitat conditions among locations. The ecological effectiveness of different reserves can also vary markedly due to factors such as age and size [25,26], and thus, the responses of protected communities to warming may be variable, supporting the need for ongoing monitoring to inform decision-making. We further find significant declines in richness and abundance of invertebrates with protection, including the herbivores, omnivores and invertivores. At the same time, the richness and abundance of fishes are increasing—a greater diversity and abundance of fishes very likely translates to greater rates of predation. Increases in both carnivorous fishes and lobsters are also typical with recovery following historical exploitation and were observed here (figure 3l;
Proc. R. Soc. B 284: 20170484
Dabundance
0.2
0.2
0
rspb.royalsocietypublishing.org
0.4
0.2 1
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
Data accessibility. Data are provided as the electronic supplementary material.
Authors’ contributions. All authors contributed to the conceptual development of the manuscript. A.E.B. analysed the data and drafted the manuscript and figures, with significant input from all authors. N.S.B., G.J.E. and others collected the data.
Competing interests. We declare we have no competing interests. Funding. Salary support for R.D.S.-S. was provided by an ARC Linkage grant and through the Marine Biodiversity Hub, a collaborative partnership supported through the Australian Government’s National Environmental Science Programme. Acknowledgements. We thank the many field technicians who assisted with diving surveys over the 21-year monitoring period, P. Bosworth and A. Schaap for initiating MPA surveys and C. Buxton for operational support. Biological monitoring was supported by grants from the Australian and Tasmanian governments (Australian Research Council; Department of Primary Industries, Parks, Water and Environment; Fisheries Research and Development Corporation; National Resource Management and Natural Heritage Trust) with considerable logistic support from Tasmanian Parks and Wildlife Service. We appreciate comments from C. Waldock on a manuscript draft.
References 1.
Molinos JG et al. 2015 Climate velocity and the future global redistribution of marine biodiversity.
Nat. Clim. Change 6, 83– 88. (doi:10.1038/ nclimate2769)
2.
Poloczanska ES et al. 2013 Global imprint of climate change on marine life. Nat.
6
Proc. R. Soc. B 284: 20170484
transition of the community by facilitating warm-affinity fishes and barren formation, which may in turn influence colonization dynamics in other poleward shifting invertebrates. Patterns in the abundance of different functional groups further suggest the role of climate variability in mediating community patterns, over and above the previously documented effects of trophic cascades resulting from protection [14]. Marked fluctuations are evident over decadal timescales in multiple dimensions of diversity, including CTI, local richness and abundance, and also for abundance of the different functional groups, such as Macrocystis. These distinct fluctuations relate to climate signals (e.g. SOI) and thus change in numerous related physical variables. For example, warm years in this region are associated with lower nutrients when the oligotrophic East Australian Current strengthens in a polewards direction [10]. These conditions may impose physiological constraints on kelp performance [32] that decrease in warmer, low-nutrient conditions, therefore explaining change in cover of Macrocystis throughout the study period and recent decline. Interpreting the complexity inherent to the temporal biodiversity signal requires consideration of natural climate variability, including extreme events. Moreover, while our analysis focuses on warming, factors such as ocean acidification and associated ecological consequences may also influence biological responses, an important direction for future research [33,34]. Multi-decadal time-series records are integral for teasing apart both the drivers for change in community structure and the components of biodiversity that are responding, and reveal that net community outcomes are strongly affected by protection from exploitation. Our findings offer general support for the role of protection in facilitating greater cover of canopy-forming macroalgae and resisting colonization by barren-forming urchins. However, we also observe a marked decline in the invertebrates following the implementation of reserves, implicating the importance of biotic interactions in driving range shifts and community functional change, an important consideration for park managers.
rspb.royalsocietypublishing.org
electronic supplementary material, figure S9), and presumably impose additional predation pressure. Our findings therefore implicate the importance of higher trophic levels and larger individuals [15,27] in biological interactions exacerbating warming-related processes of community change, such as geographical range contractions. Increases in abundance and richness across trophic levels with warming, including species with relatively cooler temperature affinities, in both protected and exploited locations suggest that warming can increase local-scale diversity. While populations of species living at the edges of their thermal tolerance may decline with warming [28,29], the cover of cooler-affinity macroalgae in Tasmania has not undergone a decrease since the start of monitoring and has increased during the past 10 years of warming. We find a similar pattern in the fishes, which may be underpinned by processes such as local adaptation. Thus, regardless of protection status, many species appear to have benefited and increased in abundance and occupancy, with recent warming. Indeed, richness increases with temperature are expected in temperate systems due to poleward geographical expansion of richer faunas from low latitudes [1,30]. The southeast Australian reef fish and invertebrate communities have a positive thermal bias [31] and, on average, are comprised of species with warmer affinity than the local environment (i.e. Tasmanian reef species also occur in much warmer regions). Thus, much of the fauna is poised to perform well with warming, at least with the magnitude of temperature change in this region to date. Even so, warming-related losses in diversity have been observed in other systems, such as in a study from Sweden documenting local extirpation in coldaffinity birds with warming and an increase in CTI [29]. Unlike the fauna, macroalgal diversity is higher in southern Australia than further north, so the flora is not predicted to increase in abundance and richness with warming, and may already be incurring climate debt. We confirm the importance of the urchin in facilitating a tropicalization of the reef communities across the region with quantitative data on the abundance of Centrostephanus as a predictor in a statistical modelling framework. In doing so, we offer the insight that the urchin may facilitate warm-affinity species and colonization by functional groups typical of lower latitudes. In particular, sessile suspension-feeding invertebrates (e.g. electronic supplementary material, figure S8) increase dramatically with warming, corresponding with increases in Centrostephanus, perhaps because bare habitat may offer colonization patches. As many of the suspensionfeeding species are typical of latitudes with warmer temperatures, the relative increase in this functional group explains the relatively large increase in the CTI of invertebrates relative to the macroalgae and fishes (respectively, approx. 0.88C versus 0.38C and 0.158C, as displayed in figure 1a–c). Another important driver of change likely relates to preferential settlement of warmer-affinity species to urchin barrens habitats rather than kelp-dominated reef. Our results therefore indicate that the poleward expansion of Centrostephanus is resulting in directional ecological change—i.e. the urchin is driving a
Downloaded from http://rspb.royalsocietypublishing.org/ on June 14, 2017
4.
5.
7.
8.
9.
10.
11.
12.
13.
15.
16.
17.
18.
19.
20. 21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
changes in herbivory and community phase shifts. Proc. R. Soc. B 281, 20140846. (doi:10.1098/rspb. 2014.0846) Edgar GJ et al. 2014 Global conservation outcomes depend on marine protected areas with five key features. Nature 506, 216 –220. (doi:10.1038/ nature13022) Edgar GJ, Barrett NS, Stuart-Smith RD. 2009 Exploited reefs protected from fishing transform over decades into conservation features otherwise absent from seascapes. Ecol. Appl. 19, 1967–1974. (doi:10.1890/09-0610.1) Sciherras M, Jenkins SR, Kaiser MJ, Hawkins SJ, Pullin AS. 2013 Evaluating the biological effectiveness of fully and partially protected marine areas. Environ. Evid. 2, 4. (doi:10.1186/2047-2382-2-4) Wernberg T et al. 2016 Climate-driven regime shift of a temperate marine ecosystem. Science 353, 169–172. (doi:10.1126/science.aad8745) Wernberg T, Russell BD, Thomsen MS, Gurgel CD, Bradshaw CJA, Poloczanska ES, Connell SD. 2011 Seaweed communities in retreat from ocean warming. Curr. Biol. 21, 1828–1832. (doi:10.1016/ j.cub.2011.09.028) Elahi R, O’Connor M, Byrnes JEK, Dunic J, Eriksson BK, Hensel MJS, Kearns PJ. 2015 Recent trends in local-scale marine biodiversity reflect community structure and human impacts. Curr. Biol. 25, 1938– 1943. (doi:10.1016/j.cub.2015.05.030) Stuart-Smith RD, Edgar GJ, Barrett NS, Kininmonth SJ, Bates AE. 2015 Thermal biases and vulnerability to warming in the world’s marine fauna. Nature 528, 88– 92. (doi:10.1038/nature16144) Dayton P, Tegner M, Edwards P, Riser K. 1999 Temporal and spatial scales of kelp demography: the role of oceanographic climate. Ecol. Monogr. 69, 219–250. (doi:10.1890/0012-9615(1999)069 [0219:TASSOK]2.0.CO;2) Connell SD, Kroeker KJ, Fabricius KE, Kline DI, Russell BD. 2013 The other ocean acidification problem: CO2 as a resource among competitors for ecosystem dominance. Phil. Trans. R. Soc. B. 368, 20120442. (doi:10.1098/rstb.2012.0442) Porzio L, Buia MC, Hall-Spencer JM. 2011 Effects of ocean acidification on macroalgal communities. J. Exp. Mar. Biol. Ecol. 400, 278–287. (doi:10.1016/ j.jembe.2011.02.011)
7
Proc. R. Soc. B 284: 20170484
6.
14.
Glob. Environ. Change 26, 27 –38. (doi:10.1016/j. gloenvcha.2014.03.009) Babcock RC, Shears NT, Alcala A, Barrett NS, Edgar GJ, Lafferty KD, McClanahan TR, Russ GR. 2010 Decadal trends in marine reserves reveal differential rates of change in direct and indirect effects. Proc. Natl Acad. Sci. USA 107, 18 251–18 255. (doi:10. 1073/pnas.0908012107) Barrett NS, Buxton CD, Edgar GJ. 2009 Changes in invertebrate and macroalgal populations in Tasmanian marine reserves in the decade following protection. J. Exp. Mar. Biol. Ecol. 370, 104–119. (doi:10.1016/j.jembe.2008.12.005) Stuart-Smith RD, Barrett NS, Stevenson DG, Edgar GJ. 2010 Stability in temperate reef communities over a decadal time scale despite concurrent ocean warming. Glob. Change Biol. 16, 122–134. (doi:10. 1111/j.1365-2486.2009.01955.x) Stevens JT, Safford HD, Harrison S, Latimer AM. 2015 Forest disturbance accelerates thermophilization of understory plant communities. J. Ecol. 103, 1253–1263. (doi:10.1111/1365-2745.12426) Devictor V, Julliard R, Couvet D, Jiguet F. 2008 Birds are tracking climate warming, but not fast enough. Proc. R. Soc. B 275, 2743 –2748. (doi:10.1098/rspb. 2008.0878) Stuart-Smith RD et al. 2017 Assessing national biodiversity trends for rocky and coral reefs through the integration of citizen science and scientific monitoring programs. Bioscience 67, 134– 146. (doi:10.1093/biosci/biw180) Edgar GJ. 2008 Australian marine life, revised edn. Melbourne, Australia: Reed Books. Wood SN. 2011 Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. R. Stat. Soc. B 73, 3 –36. (doi:10.1111/j.1467-9868.2010. 00749.x) Pinheiro J, Bates D, DebRoy S, Sarkar D, Team RC. 2015 nlme: linear and nonlinear mixed effects models. See http://CRAN.R-project.org/ package=nlme. R Core Team. 2014 R: a language and environment for statistical computing. See http://www.R-project. org/. Verge´s A et al. 2014 The tropicalization of temperate marine ecosystems: climate-mediated
rspb.royalsocietypublishing.org
3.
Clim. Change 3, 919 –925. (doi:10.1038/ nclimate1958) Chen IC, Hill JK, Ohlemueller R, Roy DB, Thomas CD. 2011 Rapid range shifts of species associated with high levels of climate warming. Science 333, 1024–1026. (doi:10.1126/science.1206432) Suttle KB, Thomsen MA, Power ME. 2008 Species interactions reverse grassland responses to changing climate. Science 315, 640–642. (doi:10.1126/ science.1136401) Cahill AE et al. 2012 How does climate change cause extinction? Proc. R. Soc. B 280, 20121890. (doi:10.1098/rspb.2012.1890) Thomas CD et al. 2012 Protected areas facilitate species’ range expansions. Proc. Natl Acad. Sci. USA 109, 14 063–14 068. (doi:10.1073/pnas. 1210251109) Teo DHL, Tan HTW, Corlett RT, Wong CM, Lum SKY. 2003 Continental rain forest fragments in Singapore resist invasion by exotic plants. J. Biogeogr. 30, 305–310. (doi:10.1046/j.1365-2699.2003.00813.x) Bates AE, Barrett NS, Stuart-Smith RD, Holbrook NJ, Thompson PA, Edgar GJ. 2014 Resilience and signatures of tropicalization in protected reef fish communities. Nat. Clim. Change 4, 62 –67. (doi:10. 1038/nclimate2062) Sala E et al. 2012 The structure of Mediterranean rocky reef ecosystems across environmental and human gradients, and conservation implications. PLoS ONE 7, e32742. (doi:10.1371/journal.pone. 0032742) Ling SD, Johnson CR, Frusher SD, Ridgway KR. 2009 Overfishing reduces resilience of kelp beds to climate-driven catastrophic phase shift. Proc. Natl Acad. Sci. USA 106, 22 341 –22 345. (doi:10.1073/ pnas.0907529106) Johnson CR et al. 2011 Climate change cascades: shifts in oceanography, species’ ranges and subtidal marine community dynamics in eastern Tasmania. J. Exp. Mar. Biol. Ecol. 400, 17 –32. (doi:10.1016/j. jembe.2011.02.032) Sunday JM et al. 2015 Species traits and climate velocity explain geographic range shifts in an ocean warming hotspot. Ecol. Lett. 18, 944–953. (doi:10. 1111/ele.12474) Bates AE et al. 2014 Defining and observing stages of climate-mediated range shifts in marine systems.
