Multiple occupancyabundance patterns in ... - Wiley Online Library

A Journal of Conservation Biogeography

Diversity and Distributions, (Diversity Distrib.) (2013) 19, 884–895

BIODIVERSITY RESEARCH

Multiple occupancy–abundance patterns in staghorn coral communities Zoe T. Richards1,2,3*, Craig Syms4, Carden C. Wallace5, Paul R. Muir5 and Bette L. Willis3,6

1

Department of Aquatic Zoology, Western Australian Museum, 49 Kew St, Welshpool, WA, 6106, Australia, 2Australian Museum, College Street, Sydney, NSW, 2010, Australia, 3School of Marine and Tropical Biology, James Cook University, Townsville, Qld, 4811, Australia, 4Australian Institute of Marine Science, Townsville, Qld, 4810, Australia, 5Museum of Tropical Queensland, Townsville, Qld, 4810, Australia, 6Australian Research Council Centre of Excellence for Coral Reef Studies, James Cook University, Townsville, Qld, 4811, Australia

ABSTRACT Aim Understanding patterns in a species’ occupancy and abundance across

multiple scales is important for management purposes, particularly for protecting threatened species. Here, we develop a new quantitative, multiscale model of occupancy and abundance that characterizes seven types of rarity and one of commonness in high-diversity communities of staghorn coral. We examine how rare species influence coral community structure and explore spatial variability in underlying patterns of community structure in the context of optimizing the data needed to protect biodiversity. Location North-west Pacific Ocean. Methods We present categorical abundance data for 87 species of staghorn

Diversity and Distributions

corals occurring within 100 sites across five reefs in the north-west Pacific Ocean. We develop a new model that combines measures of global distribution, local distribution and local abundance to describe eight mutually exclusive occupancy–abundance patterns, which can be used to prioritize regional species conservation. Traditional and new analytical approaches are compared to explore how rare species influence multidimensional space and community structure. Results We show that five types of occupancy–abundance relationships exist in

staghorn coral assemblages, including four patterns of rarity. The overwhelming majority of species (73%) are rare according to local abundance and/or distribution criteria. Occupancy–abundance patterns are spatially variable in staghorn coral communities, and no single underlying distribution fits all assemblages. Our findings suggest that 54 species are at risk of regional extinction, 30 of which are also classified as ‘vulnerable’ by the IUCN. Main conclusions Our model demonstrates that multiple occupancy–abundance

*Correspondence: Zoe T. Richards, Department of Aquatic Zoology, Western Australian Museum, 49 Kew St, Welshpool, WA 6106, Australia. E-mail: [email protected]

patterns exist in staghorn coral assemblages. We conclude that 66% of the pool of staghorn coral fauna in the north-west Pacific is at risk of regional extinction. At the locations and scales examined here, occupancy–abundance patterns and the underlying distributions of coral communities are spatially variable, suggesting that it may not be appropriate to apply unified ecological theory to communities with a large proportion of threatened species because this may jeopardize biodiversity conservation. Keywords Abundance and distribution, Acroporidae, community structure, extinction risk, rarity, scleractinia.

Against a backdrop of ecosystem decline and forecasts of escalating biodiversity loss, demographic research provides a

basis for evaluating options for conserving rare species (Gaston, 1994; Gaston et al., 2000; Brooks et al., 2006). In diverse communities, however, documenting the abundance of rare species is a logistical and analytical challenge. As a result,

DOI: 10.1111/ddi.12032 http://wileyonlinelibrary.com/journal/ddi

ª 2013 John Wiley & Sons Ltd

INTRODUCTION

884

Occupancy–abundance patterns in the coral genus Acropora prioritizing species conservation action in diverse coral reef ecosystems. Securing the health and resilience of coral reef ecosystems is imperative, not only because of their diversity and unique geological structure (Kleypas et al., 2001), but also because of their socio-economic value (Moberg & Folke, 1999; Access Economics, 2007). Over 600 species of extant scleractinian corals function in reef construction, primary production and nutrient recycling (Done et al., 1996; Veron, 2000) and provide microhabitat and food for a diverse range of fauna, flora and microbes (Paulay, 1997; Ainsworth et al., 2010; Stella et al., 2011). Despite their keystone role, corals are highly susceptible to cumulative anthropogenic and climate impacts, and today, one-third of shallow-water hard corals are threatened (Carpenter et al., 2008). Some of the most highly threatened species belong to the genus Acropora (staghorn corals). This diverse group (c. 120 extant species; Wallace et al., 2012) is highly susceptible to thermal bleaching, changes in water quality, disease and predation (Marshall & Baird, 2000; Willis et al., 2004; Fabricius et al., 2005); hence, 50% of the genus is included on the IUCN Red List of threatened species (www.iucnredlist.org). It is predicted that coral community structure will shift away from branching species of Acropora towards massive and encrusting corals such as Porites over the coming decades (Riegl & Purkis, 2009). To detect such shifts and safeguard this keystone group of corals, a greater understanding of the ways in which their current distributions and abundance

rare species are often excluded from ecological studies, either implicitly by the use of restricted sampling designs (i.e. inadequate searching, inappropriate search strategy and/or sampling within a limited subset of habitat types) or explicitly during analysis (Syms, 1998; Villard et al., 1999; Bai et al., 2004). Rare species have been described as contributing little other than ‘noise’ to a statistical solution (Gauch, 1982) and have been described as increasing computational time while lowering confidence scores (Marchant, 2002). These perceived difficulties in sampling and analysing rare species have hindered their representation in the numerical abundance and structure data upon which ecological theory is based (Gaston, 1994; Chapman, 1999). Part of the difficulty in documenting rare species is that rarity is an intuitive concept that is scale-dependent (Kunin & Gaston, 1997), thus it cannot be quantified easily in diverse and complex communities. The term ‘rare’ has been applied to a variety of patterns relating to a species’ abundance (e.g. relative or absolute), distribution (e.g. range size, endemism, extent of occurrence) and/or niche specialization (e.g. habitat/niche, space/diet) (Brown, 1984; Hanski & Gyllenberg, 1993; Gaston, 1994). To accommodate the multiple types of rarity in grasses, Rabinowitz (1981) proposed a model that identified seven forms of rarity and one of commonness (Fig. 1a). This approach has been successfully adapted for the study of mammals (Yu & Dobson, 2000; Harcourt et al., 2002), birds (Kattan, 1992) and trees (Pitman et al., 2001), and it holds promise as a tool for (a) 1.

Figure 1 Rarity models. (a) ‘Seven forms of Rarity Model’ as developed by Rabinowitz (1981) to explain patterns of occupancy and abundance in grasses. The black panel represents commonness; the white panels represent the seven forms of rarity. Occupancy–abundance types are numbered to correspond with Fig. 1b. (b) Modified rarity model (adapted from Rabinowitz, 1981) depicting eight occupancy–abundance relationships (called ‘Occupancy types’) that describe patterns in the distribution and abundance of staghorn corals (Acropora and Isopora) in the north-west Pacific Ocean. Occupancy types 1–7 represent different types of rarity and type 8 represents commonness. Large circles represent species that are widespread or common, and small circles represent species that are restricted or rare at the designated level.

Small range Habitat specialist Small population size

2.

Large range Habitat specialist Small population size

5. Small range 6. Large range Habitat generalist Habitat generalist Small population size Small population size

(b) Occupancy Type

3.

4. Small range Small range Habitat specialist Habitat generalist Small population size Large population size

8. Large range 7. Large range Habitat specialist Habitat generalist Large population size Large population size

Global Local Local Distribution Distribution Abundance

Consequence

1

GLOBAL EXTINCTION

2

LOCAL EXTINCTION

3

PERSISTENCE

4

PERSISTENCE

5

PERSISTENCE

6

PERSISTENCE

7

PERSISTENCE

8

PERSISTENCE

Diversity and Distributions, 19, 884–895, ª 2013 John Wiley & Sons Ltd

via compensation via compensation

via compensation

885

Z. T. Richards et al. change across spatial scales, here referred to as occupancy– abundance patterns, is needed. Extensive global biogeographic records for Acropora indicate that the majority of species have a broad Indo-Pacific distribution (Veron & Wallace, 1984; Veron, 1993, 2000; Wallace, 1999; Wallace et al., 2012). However, few studies have focused specifically on patterns of local abundance and occupancy in this group, although more generally, reef-building corals are reported as sparse in abundance on local scales (e.g. DeVantier et al., 2006). If coral communities contain a large proportion of locally rare species, it would be expected that a long ‘tail’ would be evident in species abundance distributions (sensu Fisher et al., 1943); however, the expected ‘tail’ of rare species is typically missing in studies of coral community structure (Connolly et al., 2005). Thus, it is not clear how well rare coral species have been represented in current ecological theory. At a metacommunity level, published studies suggest there is little spatial variance in the patterns of abundance and occupancy among scleractinian corals (Karlson et al., 2011) or in the underlying species abundance distributions of coral communities (Connolly et al., 2005). These findings have led coral ecologists to propose that spatial variance and occupancy patterns in coral communities can be represented by a common, unified distribution for the purpose of developing theories about how coral communities are structured (Karlson et al., 2011). The implied benefit of unified ecological theory is that species abundance can be estimated at a fine scale from occupancy information obtained at coarser spatial scales and this saves both time and money (He & Gaston, 2000; Conlisk et al., 2009). However, if rare species have not been well represented in the numerical data upon which ecological theory is based, there is a risk that occupancy– abundance patterns and aspects of community structure will be oversimplified at the expense of species conservation outcomes, and this would be particularly problematic for rare and threatened species. Here, we apply a standard rapid ecological assessment method to maximize the detection of rare species. We develop a quantitative definition of occupancy–abundance relationships in corals that distinguishes seven different patterns of rarity and one of commonness and apply this model to five staghorn coral communities in the north-west Pacific Ocean. We then use the model to highlight which species have the highest regional extinction threat and compare this to current global assessments of threatened status using IUCN categories and criteria. Furthermore, we explore ß-diversity to examine how rare species influence community structure in multidimensional space and to investigate whether or not there is spatial variability in the underlying distribution patterns of staghorn coral communities. METHODS The study focussed on five reef locations in the north-west Pacific Ocean: four coral atolls in the Republic of the

886

Marshall Islands (RMI) and the fringing reefs of Kimbe Bay, New Britain, Papua New Guinea. Despite differences in geomorphic structure between atolls and fringing reefs, exposed sites in all five locations were generally similar – that is, all sites had steep reef slopes, with strong currents and the propensity for strong wave swell. ‘Protected’ sites in each location were lagoonal, featuring patch reefs generally protected from swell. All sites had very high water visibility (c. 20 m+). All species of staghorn coral (Genus Acropora Oken, 1815 and Genus Isopora Studer, 1878; Scleractinia: Acroporidae) occurring within 100 sites spread across the five locations were surveyed by the first author from July 2002 to September 2005 [Rongelap Atoll, Marshall Islands, 11.5833 ºN, 165.3333 ºE (n = 41); Mili Atoll, Marshall Islands, 6.1333 ºN, 171.9167 ºE (n = 20); Ailinginae Atoll, Marshall Islands, 11.1667 ºN, 166.3333 ºE (n = 7); Bikini Atoll, Marshall Islands, 11.5833 ºN, 165.3333 ºE (n = 19); Kimbe Bay, Papua New Guinea, 5.1667 ºS, 150.5000 ºE (n = 13)]. At each of the five locations, approximately equal numbers of sites were randomly selected within two wave exposure regimes (exposed and protected). At each site, a standard method of rapid ecological assessment (DeVantier et al., 1998) was undertaken on SCUBA, whereby the abundance of every species of staghorn coral was documented on a 60-min timed swim (covering c. 2500 m2). To maximize the detection of rare species, all available habitats (e.g. reef slope, crest and flat) were carefully searched from 1 m depth to a maximum of 30 m. To distinguish between different types of rarity, data on global distributions, local distributions and local abundance were compiled for all species of staghorn coral encountered. Global distribution was estimated from the maximum longitudinal and latitudinal limits in the WorldWide Acropora Database (founded upon 25,000 specimen-based records and over 30 years of collections; Wallace, 1999) and approximated as elliptical in shape. Local distribution was determined as the per cent of sites where a species was present. Local abundance was coded for each species at each site using a 5-point DAFOR scale (Jongman et al., 1995), which is broadly analogous to a logarithmic scale: 1 = rare (1–2 colonies); 2 = infrequent (3–5 colonies); 3 = frequent (6–20 colonies); 4 = common (21–50 colonies); and 5 = dominant (51 + colonies). If large colonies or stands of branching coral were encountered, every 1 m2 was classified as two colonies. A quantitative definition of rarity and occupancy– abundance patterns Rather than arbitrarily assigning an a priori cut-off value to represent rare species, we ranked global distribution, local distribution and the mean local abundance of each species from smallest to largest to look for discontinuities in the shapes of the graphs that might correspond to ‘groups’ of responses. After testing assumptions and observing distinct linear breaks in the distributions, we used segmented


Occupancy–abundance patterns in the coral genus Acropora regression models in R to delineate rare/restricted species (Vito & Muggeo, 2003, 2008). Mean local abundance per site was calculated according to the standard equation below, which does not include zero values (McArdle & Gaston, 1995; Gaston, 1996): mean local abundance ¼ minimum absolute abundance ðfrom DAFOR categoriesÞ ðnumber of sites occupiedÞ To define occupancy–abundance patterns, we combined measures of global distribution, local distribution and local abundance to describe eight mutually exclusive occupancy– abundance patterns (Fig. 1b) (here-in called ‘Occupancy types’). Each occupancy type leads to one of two general outcomes: persistence or extinction. We use the term compensation to represent the potential for species that have a restricted global distribution to offset the disadvantages of rarity by having either a large local distribution and/or large local abundance. Based on our quantitative definitions of rarity, the spectrum of occupancy–abundance types ranges from species that are rare across all three levels (Occupancy type 1) to species that are not rare according to any of the three measures (Occupancy type 8) (Fig. 1b).

Fitting statistical distributions To more closely examine the relationship between species abundance and occupancy, we examined the observed distribution of the numbers of species in different abundance categories with the number of sites occupied and fitted mixture distributions of lognormal functions. Unimodal, bimodal and trimodal distributions were fitted to explore which model provides the best fit (selected based on the lowest Schwartz’s Bayesian Information Criterion). Calculations were carried out with the ‘mix’ function in the ‘mixdist library’ in R (MacDonald & Du, 2010). Various binning strategies were examined, and the binning strategy presented is the most robust result. Rank–abundance plots were used to characterize the distribution shape of the different coral assemblages. Four classical models (Motomura, Broken Stick, Zipf and the Zipf–Mandelbrot; Frontier, 1985) were compared to the lognormal distribution using Akaike’s information criterion, and predicted curves from the best model were overlaid on the plot. Calculations were made out with the ‘radpart’ function in the ‘vegan library’ in R (Oksanen et al., 2010). The Kimbe Bay assemblage was not fitted well by any of the classical models, so an additional nonlinear four-parameter model was fitted for this location.

Multivariate assemblage structure Detecting rare species We used two approaches to ensure the sampling programme was adequate to detect rare species. Species accumulation curves were calculated using the ‘vegan::specaccum’ function in R with jackknifed standard errors (Oksanen et al., 2010). We also checked the ability of the sampling programme to detect rare species by using the Chao-2 estimator (Chao, 1987; Colwell & Coddington, 1994) in the ‘vegan::specpool’ function (Oksanen et al., 2010). This method estimates the total species pool at a site by:

Equation ¼

Number of species observed onceÞ2 ð2 Number of species observed twiceÞ

Diversity statistics Simpson’s and Shannon’s (natural log) diversity and evenness indices were calculated for each site and then averaged across each location. To enable comparisons across locations with different sampling efforts, we also used rarefaction methods (based on Hurlbert, 1971; with standard errors based on Heck et al., 1975) using the ‘vegan::rarefy’ function (Oksanen et al., 2010) to standardize the number of species expected according to smallest sample in the study (Ailinginae, n = 7). Additionally, b-diversity was calculated to determine how much diversity was contained within a single site, relative to the total species pool across sites at a given location. All diversity measures were calculated using the ‘vegan’ library in R (Oksanen et al., 2010).


Multivariate assemblage structure was analysed by non-metric multidimensional scaling (nMDS) in SAS (Statistic Analysis System v 9.1.3, SAS Institute Inc., Cary, NC, USA). Visual examination of species–frequency plots indicated that an x0.25 power transformation best balanced the influences of abundant, mid-range and rare species in the analysis. The Bray– Curtis distance measure was chosen because it weights cooccurrences of species in samples (Krebs, 1989; Legendre & Legendre, 1998). nMDS were run using a range of different numbers of axes to identify the dimensionality of the data, and stress values were used as a measure of goodness-of-fit. To improve the representation of rare species, we inverted the weighting of abundant and rare species by transforming data using their inverse value plus a constant: (x + 0.5)1. The value of 0.5 was chosen after examining a range of values because it gave the best ‘spread’ of values across species and did not over-weight 0’s relative to 1’s or other low values. To determine whether location, exposure or their interaction were statistically significant, multivariate analysis of variance tests were run on the sample scores of the set of nMDS axes that accounted for stress values of < 0.05 for each of the x0.25 and (x + 0.5)1 analyses. Similar approaches have been used elsewhere to deal with problems of too few samples for full analysis (e.g. Syms & Jones, 2000).

RESULTS A total of 82 staghorn coral species were recorded in surveys (80 species of Acropora/2 of Isopora). Species richness was

887

Z. T. Richards et al. highest at Kimbe Bay (69 species), followed by Rongelap Atoll (53 species), Bikini Atoll (41 species), Mili Atoll (38 species) and Ailinginae Atoll (32 species). Species accumulation curves approached asymptotes for all locations except Ailinginae, where diversity was under-represented due to the small sample size. However, the Chao method confirmed that the observed and predicted species number had a 1:1 relationship, indicating that all assemblages were sufficiently sampled to capture rare species. A quantitative definition of rarity From the results of the segmented regressions, we show that the most biologically meaningful cut-off points for delineating rare and restricted are scale-dependent. From the ranked distribution of global range size, there were three distinct groups (Fig. 2a). Fifty-seven per cent of species had range sizes > 65 million km2 and are considered widespread, 26%

Range size (millions of km sq.)

(a)

140 120 57% of species in top 44% of range size

100 80 60

Occupancy–abundance relationships

26% of species between 11% and 56% of range size

40 20

17% of species in bottom 11% of range size

0 0

10

20

30 40 50 Rank range size

60

70

80

(b) 100 Percentage of sites occupied (sq.root)

of species had moderate large range sizes (15 million km2 – 65 million km2) and 17% of species had restricted distributions. Measured at the local scale, these proportions changed, and two distinct groups were evident: 63% of species were locally restricted and 37% of species were locally widespread (Fig. 2b). When mean local abundance was analysed, three distinct, equally sized groups were present, with 33% of species being locally rare, 33% having intermediate abundance and the remaining being locally common (not shown). After categorizing species as rare or restricted according to each of the three factors of the coral occupancy–abundance model, we show that four types of rarity exist in staghorn corals in the north-west Pacific (Table 1). No species displayed occupancy types 3, 4 or 5. The greatest proportion of species (48%; n = 82 species) were globally widespread but had restricted local distributions and were locally rare (Occupancy type 2). Fifteen species were restricted on both global and local scales and were locally rare (Occupancy type 1). Thirteen species had widespread global and local distributions but were locally rare (Occupancy type 6). Three species had a widespread global distribution but were locally restricted, although locally common (Occupancy type 7). Twelve species were not rare according to any of the three measures. Overall, distribution and abundance patterns predict high extinction risk for 54 species (Occupancy 1 & 2) but persistence for 29 species (Occupancy types 3–8).

Acropora nasuta

64 37% of species occupy > 17% of sites

36

16

There was a positive but noisy relationship between log10(total abundance) and maximum range size [log10 (total abundance) = 0.887 + 6.950* range in thousands of square km, R2 = 0.2469, Fig. 3a]. This indicates, somewhat intuitively, that more widespread species have larger population sizes. This relationship was particularly strong when the log10(total abundance) was regressed on log (number of sites occupied) [log10(total abundance) = 0.199 + 1.047* log10(number of sites occupied), R2 = 0.9562, Fig. 3b]. All coefficients were significant at P < 0.001. The slope of this relationship was effectively 1, indicating that the increase in population size with increased occupancy was not due to increased local abundance at some sites versus others. This finding was supported by the absence of a relationship between mean local abundance and either range size or number of sites occupied (Fig. 3c,d).

4 63% of species occupy < 17% of sites

Diversity statistics 0 0

10

20

30 40 50 Rank site occupancy

60

70

80

Figure 2 The scale-dependent nature of rarity. Comparison of the proportion of species classified as rare based on whether rarity is delineated according to global or local distributions. (a) Global distribution versus rank global distribution. (b) Percentage of local sites occupied versus rank percentage of local sites occupied. Note the square-root scale on the y-axis to linearize the relationship

888

Diversity and evenness indices were unstable (Table 2) and not particularly informative in our study, most likely because of the uneven sample sizes among locations. The rarefaction order mirrored the species accumulation curves, with the location diversity order being Kimbe > Rongelap > Bikini > Mili > Ailinginae. The b-diversity indices provided insights into within-location heterogeneity, and sites within Ailinginae, Kimbe and Mili appeared to be the most homogeneous, with each site containing a high proportion


Occupancy–abundance patterns in the coral genus Acropora Table 1 Summary of occupancy–abundance relationships exhibited by the 82 species of staghorn coral examined in this study, depicting the five occupancy types that were identified in the genus Acropora and the four patterns identified in the genus Isopora. No species display occupancy types 3, 4 or 5 Occupancy– abundance type

Global distribution

Type 1

Local distribution

Local abundance

No. species

Globally restricted

Locally restricted

Rare

n = 15

Type 2

Globally Widespread

Locally Restricted

Rare

n = 39

Type 6

Globally Widespread

Locally Widespread

Rare

n = 13

Type 7

Globally Widespread

Common

n=3

Type 8

Globally Widespread

Locally Restricted Locally Widespread

Common

n = 12

Predicted consequence

Species A. awi,* A. batunai,* A. chesterfieldensis, I. crateriformis,* A. jacquelineae,* A. kimbeensis,* A. loisetteae,* A. lokani,* A. pichoni, A. plumosa,* A. rongelapensis,† A. spathulata, A. walindii,* A. plana,† A. derawanensis* A. acuminata,* A. abrolhosensis,* A. anthocercis,* A. bushyensis, A. carduus, A. caroliniana,* A. clathrata, A. desalwii,* A. divaricata, A. donei,* A. echinata,* A. exquisita,† A. glauca, A. grandis, A. horrida,* A. insignis,† A. intermedia, A. kirstyae,* A. latistella,* A. listeri,* A. longicyathus, A. lovelli,* A. microclados,* A. microphthalma, A. paniculata,* A. prostrata,† A. sarmentosa, A. selago, A. solitaryensis,* A. speciosa,* A. spicifera,* A. subglabra. A. subulata, A. tenella,* A. tenuis, A. tortuosa, A. valenciennesi, A. vaughani,* I. cuneata* A. elseyi, A. humilis, A. hyacinthus, A. lutkeni, A. millepora, A. monticulosa, A. nana, A. robusta, A. samoensis, A. secale, A. striata, A. valida, A. verweyi A. abrotanoides, A. aspera, I. brueggemanni A. aculeus, A. austera, A. cerealis, A. cytherea, A. digitifera, A. florida, A. gemmifera, A. granulosa, A. loripes, A. muricata, A. nasuta, I. palifera

Global extinction

Local extinction

Persistence

Persistence Persistence

*Vulnerable as per IUCN. †Data Deficient as per IUCN.

(a) Log10 (Total Abundance)

2.5 2.0 1.5 1.0 0.5

0

(b)

3.0

0 20

40

60

80

100

120

2.5 2.0 1.5 1.0 0.5 0

140

0

0.5

Range (millions of sq. km)

1.0

1.5

2.0

Log10 (number of sites occupied)

(c)

(d)

3.5

3.5

Mean local abundance

Mean local abundance

Figure 3 Abundance–occupancy relationships for Acropora/Isopora species. Each point on the plots represents a species. In each plot, the thick line represents the linear regression, the dashed lines are the confidence intervals around the slope and the parallel solid lines are the 95% confidence intervals for the relationships between (a) log10 (total abundance) and maximum global distribution in millions of km2; (b) log10 (total abundance) and log10 (local distribution); (c) mean local abundance and maximum global distribution in millions of km2; and (d) mean local abundance and log10 (local distribution).

Log10 (Total Abundance)

3.0

3.0 2.5 2.0 1.5 1.0 0.5 0

3.0 2.5 2.0 1.5 1.0 0.5 0

0

20

40

60

80

100

120

Range (millions of sq. km)


140

0

0.5

1.0

1.5

2.0

Log10 (number of sites occupied)

889

Z. T. Richards et al. Table 2 Diversity statistics for the five reef locations in the north-west Pacific Ocean. All statistics were calculated at the site level and the means (standard errors) presented. b-diversity was calculated from each site relative to individual species pools and represents the ratio of the number of species found across all sites combined compared to the mean number of species per site. Low values indicate that sites are homogenous and that most of the diversity is maintained within sites; high values indicate that diversity is driven by between-site differences in species numbers and composition Diversity

Location

Simpson

Shannon

Simpson

Shannon

Rarefaction (No. species per 100 individuals

Ailinginae Bikini Mili Rongelap Kimbe

0.889 0.772 0.854 0.770 0.880

2.397 1.913 2.278 2.102 2.813

0.782 0.537 0.612 0.382 0.579

0.945 0.839 0.950 0.901 0.798

7.509 6.007 6.745 6.085 7.736

(0.021) (0.024) (0.015) (0.017) (0.030)

Evenness

(0.134) (0.099) (0.074) (0.056) (0.160)

(0.037) (0.043) (0.040) (0.031) (0.039)

of the overall location diversity. Rongelap and Bikini were more heterogeneous, as each site contained limited subsets of the overall diversity, indicating some finer scale partitioning of diversity. Statistical distributions We find no single pattern fits the underlying distributions of all assemblages. At a metacommunity scale, two distinct distribution patterns were evident. Over 80% of species occupied fewer than 20% of sites in a log-series decreasing function (55% of species were found at fewer than 10 sites) and the remaining species were normally distributed across the remaining sites (Fig. 4a). Similarly, for local abundance, the relationship was not a simple unimodal function (Fig. 4b) because over 80% of species occurred fewer than 50 times, in a log-series decrease, and the remainder were normally distributed. To examine whether the bimodal patterns result from pooling data, mixture distributions were fitted via location. The results were not consistent because the bimodal pattern was evident in the Ailinginae and Rongelap assemblages (see Fig. 4c); however, at Bikini, Mili and Kimbe, the assemblages closely fitted a unimodal decreasing function (see Fig. 4d). The varying pattern of community structure between locations was also evident in plots of the log-rank abundance distributions (RAD) (Table 3). A standard lognormal distribution was not the best approximate fit for any location. Ailinginae, Bikini and Mili assemblages were best fitted by the Motomura model (in which it is assumed the first species subsumes a fraction K of the resource, with the next species subsuming the same fraction of the remaining resource and so on). Rongelap was best fitted by the McArthur Broken Stick model, which indicates a distinct break in community structure, while Kimbe Bay was best fitted by a four-parameter lognormal distribution (Fig. 4e). Multivariate community structure MANOVA confirmed that staghorn coral assemblages differed between sites but not between exposed and protected

890

(0.014) (0.022) (0.036) (0.022) (0.048)

(0.347) (0.306) (0.214) (0.185) (0.370)

a-diversity (No. of species)

b-diversity

33 41 38 53 68

1.538 3.015 1.879 2.574 1.693

habitats (Table 4); however, the dimensionality of the data was very high. In the 2-dimensional scaling solution, all locations were relatively distinct from each other (Fig. 5a). However, this result did not appear to be biologically meaningful, given that, for example, there were a large number of rare species recorded only at Kimbe Bay or only in lagoonal habitats. Hence, we re-ran the MANOVA by inverse transformation, and by doing this, a strong location by exposure interaction was detected (Table 5). In 2D, the Kimbe Bay community was clearly distinct from the tightly clustered (and biologically similar) Marshall Island locations (Fig. 5b). DISCUSSION Our study of rarity and diversity using scaled distribution and abundance data demonstrates that multiple patterns of occupancy and abundance exist in north-west Pacific staghorn coral assemblages. Almost three-quarters of the species examined displayed one of four types of rarity (types 1, 2, 6 and 7) (Table 1). Considering that most species occurred in low numbers and/or were sparsely distributed on local scales, extensive searching across habitats was required to document occupancy patterns in staghorn corals. Furthermore, the appropriate cut-off points to delineate rare species varied according to scale and whether abundance or distribution was examined. Combining all three metrics for occupancy and abundance into one model enabled us to predict which species face the greatest risk of extinction in the north-west Pacific region. Fifty-four species (66%) are identified to be on extinction pathways. Fifteen species face triple jeopardy (i.e. have the highest extinction risk; Munday, 2004) because they are restricted across all three occupancy–abundance levels (i.e. Occupancy type 1). Thirty-nine species have restricted local distributions and abundances (i.e. Occupancy type 2). Occupancy type 2 species are of special interest for regional conservation planning because, despite having a large global distribution, they may be vulnerable to regional extinction. Insidious losses of local populations can have devastating


Occupancy–abundance patterns in the coral genus Acropora 30

30

(a)

Number of species

Number of species

20 15 10

20 15 10 5

5

0

0 0

20 40 60 80 Percentage of sites occupied

100

0

(b)

Number of species

20 15 10 5

10 20 Abundance per 10 sites (d)

25

25 Number of species

Rongelap Atoll

25

25

30

(c)

30

Bikini Atoll

20 15 10 5 0

0 0

50

100 150 Abundance

0

5

10 15 20 25 Abundance per 10 sites

30

100

Log10 Abundance

(e)

200

10

Bikini Kimbe Mili

Ailinginae

Rongelap

1 10

20

50

40 30 Species rank

60

Figure 4 Frequency histograms, fitted mixture distributions and rank abundance distributions for 82 species of Acropora in the NW Pacific showing (a) metacommunity data for the local distribution of species; (b) metacommunity data for the local abundance of species; (c) local abundance of species at Rongelap Atoll; (d) local abundance of species at Bikini Atoll; and (e) patterns of community structure across locations in the north-west Pacific as revealed by Rank Abundance Distribution Curves.

Table 3 Akaike’s information criteria for competing rank abundance models. The smallest values (shaded) reflect better fit. None of the best competing classical models fitted the Kimbe rank abundance distribution at both lower and upper extremes of the distribution due to high evenness of the most abundant species, so a four-parameter model was used to fit the curve

Location

Motomura

Broken stick

Lognormal

Zipf

Mandelbrot

Four-parameter lognormal

Ailinginae Bikini Mili Rongelap Kimbe

108.746 145.329 160.965 255.716 297.933

111.364 146.154 173.306 247.685 309.302

110.491 150.139 179.659 261.058 327.365

118.146 173.395 229.795 363.216 436.892

111.533 147.157 161.073 256.434 299.313

169.369


891

Z. T. Richards et al. Table 4 Multivariate analysis of variance of the first 14 dimensions (stress 1 0.0481) from nm-MDS of the Bray–Curtis distance of x0.25 transformed data. This analysis tests differences between assemblages weighted by dominant species Source

Pillai’s Trace

F

d.f.

Significance

Location Exposure Location*Exposure

2.292 1.730 0.694

7.58 1.14 1.18

56,316 14,76 56,218

< 0.0001 0.3427 0.1793

Stress1 0.30

(a)

Ailinginae

Dimension 2

Rongelap

Bikini

Kimbe Mili

Dimension1 Stress1 0.18

(b)

Dimension 2

Marshall Is.

Kimbe

Dimension1

Figure 5 Two-dimensional nMDS of Acropora and Isopora assemblages using the Bray–Curtis distance measure. (a) x0.25 transformed data and (b) (x + 0.5)1 transformed data, which weighted rare species over common ones. Ailinginae Mili Rongelap Bikini Atoll Kimbe Bay. Open symbols are sheltered locations, shaded symbols are exposed locations. nMDS, non-metric multidimensional scaling. Table 5 Multivariate analysis of variance of the first 12 dimensions (stress 1 0.0491) from nm-MDS of the Bray–Curtis distance of (x + 0.5)1 transformed data. This analysis tests differences between assemblages weighted by rare species Source

Pillai’s Trace

F

d.f.

Significance

Location Exposure Location*Exposure

2.092 0.250 0.765

7.49 2.19 1.62

48,306 12,79 48,328

< 0.0001 0.0196 0.0087

892

effects on the global status of species (Ehrlich & Daily, 1993; Frankham, 1995; Hughes et al., 1997). In a global context, 10 species with type 1 and 20 species with type 2 occupancy relationships are recognized by the IUCN as ‘vulnerable’ (species marked with ‘*’ in Table 1; see Carpenter et al., 2008). Our study of regional demographics provides further validation that these 30 species face an elevated level of risk (at global and/or regional scales) and should be prioritized in future coral biodiversity monitoring and research programmes. Five additional species are listed by the IUCN as ‘data deficient’ (species marked with ‘†’ in Table 1), and our data suggest that these five species may also be vulnerable to regional extinction. The remaining nineteen species with a type 1 occupancy–abundance relationship are listed as either ‘near threatened’ or ‘least concern’ by the IUCN but may be at risk of regional extinction. We found no evidence to suggest that rare staghorn species of the north-west Pacific compensate for restricted global distributions by having large local distributions or abundance, that is, no species occurred in Occupancy types 3, 4 or 5. Hence, ecological compensation does not appear to offset geographical rarity, as has been reported in terrestrial fauna (Williams et al., 2006). Considering that no relationship was observed between mean local abundance and local or global distribution (Fig. 3c,d), we find no evidence that staghorn corals have aggregated distributions, as has been postulated for mobile taxa (Gaston et al., 2000; He & Gaston, 2000) and corals (Cornell et al., 2007; Karlson et al., 2007). Hence, while local abundance patterns may be predictable, local diversity is not. Fine-scale partitioning of diversity at Bikini and Rongelap Atolls indicates that species did not have homogeneous abundance–occupancy relationships (Table 2). At these locations, diversity is driven by between-site differences in species numbers and composition, meaning that each site contains a subset of the overall diversity. Contrary to this pattern, diversity is more homogeneous at Ailinginae, Mili and Kimbe Bay, indicating that each site contains a representation of the overall location-wide diversity. Our results suggest that patterns of occupancy and abundance in staghorn coral communities in the north-west Pacific Ocean are spatially variable. Variation was particularly evident in the modality of fitted mixture distributions and in RAD. Our results provide preliminary evidence that two general underlying distribution types are present in the staghorn coral assemblages studied, that is, the log-series and lognormal (Fig. 4e), and that not all communities are unimodal (see McGill et al., 2007). A log-series distribution (i.e. Motomura, 1932) was supported by AIC for Ailinginae, Bikini and Mili Atolls, while Rongelap Atoll and Kimbe Bay were best described by variants of the lognormal model (i.e. the broken stick, MacArthur, 1957, 1960 for Rongelap and a four-parameter lognormal model for Kimbe Bay). However, given that AIC is asymptotic and some of the values are very close (Table 3), we caution against further interpreting which model provides the best comparative fit to our data.


Occupancy–abundance patterns in the coral genus Acropora Staghorn corals present interesting contradictions to accepted principles regarding the underlying distributions of natural communities (see also Dornelas & Connolly, 2008). Evidence of multiple occupancy–abundance patterns and spatial variance in community structure has implications for biodiversity conservation and for managing the resilience of coral reef ecosystems. For example, if a community is fitted by an underlying lognormal curve, this indicates that the majority of species reach a relatively similar local abundance (i.e. high evenness), whereas a community fitted by a log-series curve has low evenness, suggesting that a small number of species dominate and many species are rare. If a species were lost from a community with low evenness, this may have lasting effects at the community/ecosystem level, whereas in a community with high evenness, the loss of a single species may have lesser functional impact because it is possible that other species could fill the empty space. Therefore, if the underlying distribution of a community reflects its resilience (assuming higher evenness equates to higher resilience), or responds to disturbance (as proposed by Gray, 1979), monitoring underlying distribution curves through time could enable detection of changes in community structure and provide a mechanism for evaluating the success of conservation actions (i.e. serve as an indicator of ecosystem health, Bakkes, 1994). Thus, with further conceptual development of the relationship between evenness and resilience, RAD’s may provide a way to ‘operationalize resilience theory’ (as called for by Nystr€ om et al., 2008). The coral occupancy–abundance model proposed here provides a framework for analysing biodiversity, not only in staghorn coral assemblages, but also in other diverse and/or threatened groups. Our study suggests that multiple and spatially variable patterns of occupancy and abundance is a common component of the north-west Pacific Ocean communities studied. We propose that the underlying distribution patterns of species within coral communities should be further investigated as a tool for conservation biology. We caution against applying unified ecological theories to groups such as staghorn corals that contain a large proportion of threatened species because, while this would reduce the time and money spent on data collection, it may jeopardize the success of biodiversity conservation efforts.

ACKNOWLEDGEMENTS Thank you to past and present members of the Department of Biological Sciences at James Cook University and the Museum of Tropical Queensland. Thanks to the Orpheus Island Research Station, the College of the Marshall Islands and the Mahonia Na Dari Research Station. Sincere thanks to Maria Beger for fieldwork collaboration. This project was funded by a Queensland Smart State Fellowship, International Society for Coral Reef Studies Fellowship and Chadwick Biodiversity Fellowship awarded to ZR.


REFERENCES Access Economics Pty Limited (2007) Measuring the economic and financial value of the great barrier reef marine park, 2005–2006. Great Barrier Reef Marine Park Authority. Townsville, Qld. Ainsworth, T.D., Thurber, R.V. & Gates, R.D. (2010) The future of coral reefs: a microbial perspective. Trends in Ecology & Evolution, 25, 233–240. Bai, Y., Han, X., Wu, J., Chen, Z. & Li, L. (2004) Ecosystem stability and compensatory effects in the inner Mongolia grassland. Nature, 431, 181–184. Bakkes, J.A. (1994) An overview of environmental indicators: state of the art and perspectives. UNEP/Earthprint, Bilthaven, the Netherlands. Brooks, T.M., Mittermeier, R.A., da Fonseca, G.A.B., Gerlach, J., Hoffman, M., Lamoreux, J.F., Mittermeier, C.G., Pilgrim, J.D. & Rodrigues, A.S.L. (2006) Global biodiversity conservation priorities. Science, 313, 58–61. Brown, J.H. (1984) On the relationship between abundance and distribution of species. The American Naturalist, 124, 255–279. Carpenter, K.E., Abrar, M., Aeby, G. et al. (2008) One-third of reef-building corals face elevated extinction risk from climate change and local impacts. Science, 321, 560–563. Chao, A. (1987) Estimating the population size for capturerecapture data with unequal catchability. Biometrics, 43, 783–791. Chapman, M.G. (1999) Are there adequate data to assess how well theories of rarity apply to marine invertebrates? Biodiversity and Conservation, 8, 1295–1318. Colwell, R.K. & Coddington, J.A. (1994) Estimating terrestrial biodiversity through extrapolation. Philosophical Transactions: Biological Sciences, 345, 101–118. Conlisk, E., Conlisk, J., Enquist, B., Thompson, J. & Harte, J. (2009) Improved abundance prediction from presenceabsence data. Global Ecology and Biogeography, 18, 1–10. Connolly, S.R., Hughes, T.P., Bellwood, D.R. & Karlson, R.H. (2005) Community structure of corals and reef fishes at multiple scales. Science, 309, 1363–1365. Cornell, H.V., Karlson, R.H. & Hughes, T.P. (2007) Scaledependent variation in coral community similarity across sites, islands, and island groups. Ecology, 88, 1707–1715. DeVantier, L.M., De’ath, G., Done, T.J. & Turak, E. (1998) Ecological assessment of a complex natural system: a case study from the Great Barrier Reef. Ecological Applications, 8, 480–496. DeVantier, L.M., De’ath, G., Turak, E., Done, T.J. & Fabricius, K.E. (2006) Species richness and community structure of reef-building corals on the nearshore Great Barrier Reef. Coral Reefs, 25, 329–340. Done, T.J., Ogden, J.C., Wiebe, W.J. & Rosen, B.R. (1996) Biodiversity and ecosystem function of coral reefs. Functional roles of biodiversity (ed. by H.A. Mooney, J.H. Cushman, E. Medina, O.E. Sala and E.D. Schulze), pp. 393–429. Chichester, John Wiley & Sons.

893

Z. T. Richards et al. Dornelas, M. & Connolly, S.R. (2008) Multiple modes in a coral species abundance distribution. Ecology Letters, 11, 1008–1016. Ehrlich, P.R. & Daily, G.C. (1993) Population extinction and saving biodiversity. Ambio, 22, 64–68. Fabricius, K., De’ath, G., McCook, L., Turak, E. & Williams, D.M. (2005) Changes in algal, coral and fish assemblages along water quality gradients on the inshore Great Barrier Reef. Marine Pollution Bulletin, 51, 384–398. Fisher, R.A., Corbet, A.S. & Williams, C.B. (1943) The Relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12, 42–58. Frankham, R. (1995) Effective population size/adult population size ratios in wildlife: a review. Genetic Resources, 66, 995–1007. Frontier, S. (1985) Diversity and structure in aquatic ecosystems. Oceanography and Marine Biology and Annual Review, 23, 253–312. Gaston, K.J. (1994) Rarity. Chapman and Hall, London. Gaston, K.J. (1996) The multiple forms of the interspecific abundance–distribution relationship. Oikos, 75, 211–220. Gaston, K.J., Blackburn, T.M., Greenwood, J.D., Gregor, R.D., Quinn, R.M. & Lawton, J.H. (2000) Abundance– occupancy relationships. Journal of Applied Ecology, 37, 39– 59. Gauch, H.G. (1982) Multivariate analysis in community ecology. Cambridge University Press, Cambridge, UK. Gray, J.S. (1979) Pollution-induced changes in populations. Journal of the Philosophical Transactions of the Royal Society of London B, 286, 545–561. Hanski, I. & Gyllenberg, M. (1993) Two general metapopulation models and the core-satellite species hypothesis. The American Naturalist, 142, 17–41. Harcourt, A.H., Coppeto, S.A. & Parks, S.A. (2002) Rarity, specialization and extinction in primates. Journal of Biogeography, 29, 445–456. He, F. & Gaston, K.J. (2000) Estimating species abundance from occurrence. American Naturalist, 156, 553–559. Heck, K.L., van Belle, G. & Simberloff, D. (1975) Explicit calculation of the rarefaction diversity measurement and the determination of sufficient sample size. Ecology, 56, 1459– 1461. Hughes, J.B., Daily, G.C. & Ehrlich, P.R. (1997) Population diversity: its extent and extinction. Science, 278, 689–692. Hurlbert, S.H. (1971) The nonconcept of species diversity: a critique and alternative parameters. Ecology, 52, 577–586. Jongman, R.H.G., Braak, C.J.F. & Van Tongeren, O.F.R. (1995) Data analysis in community and landscape ecology. Cambridge, Cambridge University Press. Karlson, R.H., Cornell, H.V. & Hughes, T.P. (2007) Aggregation influences coral species richness at multiple spatial scales. Ecology, 88, 170–177. Karlson, R.H., Connolly, S.R. & Hughes, T.P. (2011) Spatial variance in abundance and occupancy of corals across broad geographic scales. Ecology, 92, 1282–1291.

894

Kattan, G.H. (1992) Rarity and vulnerability: the birds of the Cordillera Central of Colombia. Conservation Biology, 6, 64–70. Kleypas, J.A., Buddemeier, R.W. & Gattuso, J.P. (2001) The future of coral reef in an age of global change. International Journal of Earth Sciences, 90, 426–437. Krebs, C.J. (1989) Ecological methodology. Harper Collins, New York. Kunin, W.E. & Gaston, K.J. (1997) The biology of rarity: causes and consequences of rare-common differences. Chapman and Hall, London. Legendre, P. & Legendre, L. (1998) Numerical ecology, 2nd edn. Elsevier, Amsterdam. MacArthur, R.H. (1957) On the relative abundance of bird species. Proceedings of the National Academy of Sciences USA, 45, 293–295. MacArthur, R.H. (1960) On the relative abundance of species. American Naturalist, 94, 25–36. MacDonald, P. & Du, J. (2010) mixdist: finite mixture distribution models. R package version 0.5-2. Available at: http:// CRAN.R-project.org/package=mixdist (accessed 21 May 2011). Marchant, R. (2002) Do rare species have any place in multivariate analysis for bioassessment? Journal of the North American Benthological Society, 21, 311–313. Marshall, P.A. & Baird, A.H. (2000) Bleaching of corals on the Great Barrier Reef: differential susceptibilities among taxa. Coral Reefs, 19, 155–163. McArdle, B.H. & Gaston, K. (1995) The temporal variability of densities: back to basics. Oikos, 74, 165–171. McGill, B.J., Etienne, R.S., Gray, J.S., Alonso, D., Anderson, M.J., Benecha, H.K., Dornelas, B., Enquist, B.J., Green, J.L., He, F., Hurlbert, A.H., Magurran, A.E., Marquet, P.A., Maurer, B.A., Ostling, A., Soykan, C.U., Ugland, K.I. & White, E.P. (2007) Species abundance distributions: moving beyond single prediction theories to integration within an ecological framework. Ecology Letters, 10, 995–1015. Moberg, F. & Folke, C. (1999) Ecological goods and services of coral reef ecosystems. Ecological Economics, 29, 215–233. Motomura, I. (1932) A statistical treatment of association (in Japanese). Japan Journal of Zoology, 44, 379–383. Munday, P.L. (2004) Habitat loss, resource specialization, and extinction on coral reefs. Global Change Biology, 10, 1642–1647. Nystr€ om, M., Graham, N.A.J., Lokrantz, J. & Norstrom, A.V. (2008) Capturing the cornerstones of coral reef resilience: linking theory to practice. Coral Reefs, 27, 795–809. Oksanen, J., Guillaume, B.F., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H. & Wagner, H. (2010) Vegan: community ecology package. R package version 2.0-2. Available at: http://CRAN.R-project.org/package=vegan (accessed 13 April 2011). Paulay, G. (1997) Diversity and distribution of reef organisms. Life and death of coral reefs (ed. by C. Birkeland), pp. 298–353. Chapman and Hall, New York.


Occupancy–abundance patterns in the coral genus Acropora ~ez, P., Pitman, N.C.A., Terborgh, J.W., Silman, M.R., N un Neill, D.A., Cer on, C.E., Palacios, W.A. & Aulestia, M. (2001) Dominance and distribution of tree species in upper amazonian terra firme forests. Ecology, 82, 2101–2117. Rabinowitz, D. (1981) Seven forms of rarity. The biological aspects of rare plant conservation (ed. by H. Synge and J. Chichester), pp. John Wiley and Sons, New York. Riegl, B.M. & Purkis, S.J. (2009) Model of coral population response to accelerated bleaching and mass mortality in a changing environment. Ecological Modelling, 220, 192–208. Stella, J.S., Pratchett, M.S., Hutchings, P.A. & Jones, G.P. (2011) Coral-associated invertebrates: diversity, ecological importance and vulnerability to disturbance. Oceanography and Marine Biology Annual Review, 49, 43–104. Syms, C. (1998) Disturbance and the structure of coral reef fish communities on the reef slope. Journal of Experimental Marine Biology and Ecology, 230, 151–167. Syms, C. & Jones, G.P. (2000) Disturbance, habitat structure, and the dynamics of a coral reef fish community. Ecology, 81, 2714–2729. Veron, J.E.N. (1993) A biogeographic database of hermatypic corals. Australian Institute of Marine Science Monograph Series 10. pp. 433. Townsville, Australia. Veron, J.E.N. (2000) Corals of the World, Vol 1. Australian Institute of Marine Science, Townsville, Australia. Veron, J.E.N. & Wallace, C.C. (1984) Scleractinia of Eastern Australia, Part V. Family Acroporidae. AIMS Monograph Series 6. pp. 485. Townsville, Australia. Villard, M.A., Trzcinski, M.K. & Merriam, G. (1999) Fragmentation effects on forest birds: relative influence of woodland cover and configuration on landscape occupancy. Conservation Biology, 13, 774–783. Vito, M. & Muggeo, R. (2003) Estimating regression models with unknown break-points. Statistics in Medicine, 22, 3055–3071. Vito, M. & Muggeo, R. (2008) Segmented: an R package to fit regression models with broken–line relationships. R News, 8/ 1, 20-25. Available at: http://cran.r-project.org/doc/Rnews/ (accessed 6 May 2011).


Wallace, C.C. (1999) Staghorn corals of the world. CSIRO Publishing, Victoria, Australia. Wallace, C.C., Done, B.J. & Muir, P.R. (2012) Revision and catalogue of worldwide staghorn corals Acropora and Isopora (Scleractinia: Acroporidae) in the Museum of Tropical Queensland. Memoirs of the Queensland Museum, 57, 255. Williams, Y., Williams, M., Alford, R.A., Maycott, M. & Johnson, C.N. (2006) Niche breadth and geographical range: ecological compensation for geographical rarity in rainforest frogs. Biology Letters, 2, 532–535. Willis, B.L., Page, C.A. & Dinsdale, E.A. (2004) Coral disease on the Great Barrier Reef. Coral health and disease (ed. by E Rosenberg. and Y. Loya.), pp. 69–104. Springer-Verlag, Berlin. Yu, J. & Dobson, F.S. (2000) Seven forms of rarity in mammals. Journal of Biogeography, 27, 131–139.

BIOSKETCH Zoe Richards is a coral biodiversity expert whose research spans multiple biological disciplines including systematics, evolution, ecology and conservation. Richards has worked for government, industry and NGO’s on a broad variety of national and international coral monitoring, capacity building and research projects and is currently employed as a coral biodiversity research scientist at the Western Australian Museum. Author contributions: Z.R. conceived the research, secured funding, conducted fieldwork, analysed the data and wrote the manuscript. C.S. conducted segmented regressions, curve fitting and multivariate analyses and provided intellectual input. C.W. verified specimen identifications and enabled access to the W.W. Acropora database. P.M. calculated global range sizes. B.W. edited the manuscript and supervised Richards during her affiliated PhD.

Editor: Omar Defeo

895