Advertisement

Coral Reefs

, Volume 36, Issue 2, pp 561–575 | Cite as

Relationships between structural complexity, coral traits, and reef fish assemblages

  • Emily S. Darling
  • Nicholas A. J. Graham
  • Fraser A. Januchowski-Hartley
  • Kirsty L. Nash
  • Morgan S. Pratchett
  • Shaun K. Wilson
Report

Abstract

With the ongoing loss of coral cover and the associated flattening of reef architecture, understanding the links between coral habitat and reef fishes is of critical importance. Here, we investigate whether considering coral traits and functional diversity provides new insights into the relationship between structural complexity and reef fish communities, and whether coral traits and community composition can predict structural complexity. Across 157 sites in Seychelles, Maldives, the Chagos Archipelago, and Australia’s Great Barrier Reef, we find that structural complexity and reef zone are the strongest and most consistent predictors of reef fish abundance, biomass, species richness, and trophic structure. However, coral traits, diversity, and life histories provided additional predictive power for models of reef fish assemblages, and were key drivers of structural complexity. Our findings highlight that reef complexity relies on living corals—with different traits and life histories—continuing to build carbonate skeletons, and that these nuanced relationships between coral assemblages and habitat complexity can affect the structure of reef fish assemblages. Seascape-level estimates of structural complexity are rapid and cost effective with important implications for the structure and function of fish assemblages, and should be incorporated into monitoring programs.

Keywords

Habitat diversity Species traits Functional ecology Reef architecture Scleractinian corals Coral reef fish 

Introduction

Scleractinian corals—the foundation species of tropical reef ecosystems—have long been recognized as providing essential habitat for reef-associated organisms (Luckhurst and Luckhurst 1978; Roberts and Ormond 1987; Stella et al. 2011). Similarly, structural complexity, defined as the physical three-dimensional configuration of a reef, can shape the abundance and diversity of reef fish assemblages across both large and small spatial scales (McCormick 1994; Nash et al. 2012; Ferrari et al. 2016). Several ecological hypotheses are proposed to underlie these relationships, notably that structural complexity and habitat diversity can mediate competition and predation, and facilitate cohabitation of an increased number of species (Hutchinson 1959; Hixon and Beets 1993; Beukers and Jones 1997).

Structural complexity is often positively associated with abundance and diversity of fishes in both temperate and tropical ecosystems (Friedlander et al. 2003; Graham and Nash 2013). Moreover, reductions in structural complexity and habitat diversity can result in reduced abundance, local extinctions, diversity loss (Graham et al. 2006; Holbrook et al. 2015; Newman et al. 2015), and declining fisheries productivity (Rogers et al. 2014), all outcomes with profound implications for reef biodiversity and associated ecosystem services.

Despite important relationships between structural complexity and reef fishes, coral reef monitoring programs typically focus on measuring total live cover and composition of reef-building corals. While many relationships between coral cover and reef fish assemblages have been weaker than structural complexity (Holbrook et al. 2008; Wilson et al. 2012), field experiments suggest that microhabitat (i.e., coral) diversity and some coral species can have a positive effect on fish diversity and community structure (Messmer et al. 2011; Holbrook et al. 2015). Recently, trait-based approaches have outlined important characteristics of coral assemblages beyond total coral cover or taxonomic composition (Darling et al. 2012; Madin et al. 2016a, b), which may provide insights for predicting the structure and diversity of reef fish communities. Traits that describe coral morphology, growth rate, and colony size may be expected to predict structural complexity, and thus the structure and composition of fish assemblages. As such, coral traits may provide a means of quantifying structural complexity and predicting fish assemblage properties from benthic surveys where direct measures of macrocomplexity were not taken, particularly in combination with the availability of open-access trait information (Madin et al. 2016a, b). For example, traits that describe branching, corymbose, or plating growth forms can provide keystone structures for reef fishes, which can preferentially select specific structural traits of corals for shelter (Noonan et al. 2012; Kerry and Bellwood 2012, 2015; Wilson et al. 2016). A range of other explicit links can be made for other coral traits (Table 1).
Table 1

Structural complexity and coral assemblage variables of total cover, richness and functional diversity, life histories, and community-weighted trait values considered as predictors of reef fish abundance, biomass, and species richness

Group

Variable

Description

VIF

Justification

Structural complexity

Habitat complexity

A visual score from 0 (flat) to 5 (highly complex)

1.60

More complex reefs offer greater structure for reef fishes

Coral cover

Hard coral cover,  %

Total  % cover of live hard corals

>5

Reefs with higher cover of hard corals provide more habitats for reef fishes. Coral cover models are presented in ESM 3 given high multicollinearity with life histories

Richness and diversity

Genera richness

Number of hard coral genera

4.69

Assemblages with greater hard coral richness and functional diversity are hypothesized to provide more functional resources that can be used by reef fishes. Greater habitat diversity typically supports a greater diversity of fishes because there are more available niches

Functional richness

Number of functional groups of hard corals

>5

Rao’s Functional diversity, Q

Abundance-weighted variance of the dissimilarities between all hard coral species pairs in multivariate functional space

1.67

Life historya

Competitive,  %

Absolute  % cover of hard corals in competitive life history

2.87

Coral life histories are trait-based groupings of species that are proposed to describe community successional patterns and responses to disturbances; we hypothesize that life histories associated with more disturbed environments can influence reef fish communities. For example, coral life history traits relate to disturbance regimes, which might influence the response of different fish functional groups based on their susceptibility to disturbance

Stress-tolerant,  %

Absolute  % cover of hard corals in stress-tolerant life history

4.49

Generalist,  %

Absolute  % cover of hard corals in generalist life history

2.10

Weedy,  %

Absolute  % cover of hard corals in weedy life history

2.42

Coral traitsa

Branching

Mean value of branching trait (0—not branching, 1—branching), weighted by abundance, across all species present at each site

3.01

Morphological traits can influence the size and shape of habitat refugia for fishes. For example, branching morphologies provide fine-scale structural complexity for small-bodied fishes, and for fishes specialized on live coral for habitat or diet (e.g., corallivores)

Growth rate

Mean value of growth rate trait (linear extension, mm year-1), weighted by abundance, across all species present at each site

>5

Coral assemblages with fast-growing corals might provide more structure and living coral for reef fishes

Maximum colony size

Mean value of maximum colony size trait (cm), weighted by abundance, across all species present at each site

3.36

Larger colonies can provide more structural complexity and shelter space for reef fish assemblages

Brooding

Mean value of brooding trait (0—not brooding, 1—brooding), weighted by abundance, across all species present at each site

1.86

Coral assemblages with brooding corals suggest a more disturbed or opportunistic community that provide less suitable habitat for reef fishes. Reproductive traits might also influence the availability of food from propagules or polyps

Fecundity

Mean value of fecundity trait, weighted by abundance, across all species present at each site

2.22

Coral assemblages with more fecund species may be able to persist or recover from disturbances to maintain live coral habitat for reef fishes. More fecund corals might also influence the availability of food from propagules or polyps

A description and justification is provided for each variable, as well as the variance inflation factor (VIF) used to assess independence of variables

aLife history classifications and trait values are provided in ESM Table S1

To the best of our knowledge, this study presents the first large-scale empirical test of trait-based relationships among coral communities, structural complexity, and reef fish assemblages. We use surveys across a large gradient of fish biomass in the Indian and Pacific oceans ranging from exploited sites to those within one of the world’s most pristine reef systems, the Chagos Archipelago (Graham and McClanahan 2013). Our objectives were to (1) evaluate which aspects of the physical and biological characteristics of benthic habitats best explain the structure of reef fish assemblages, in particular how well coral traits describe fish assemblages relative to physical measures of reef structure, and (2) investigate the relationships between hard coral cover, life histories and species traits, and structural complexity.

Materials and methods

Study sites

We surveyed 157 sites in the Seychelles, Maldives, Chagos Archipelago, and the Great Barrier Reef in Australia between 2010 and 2013 (Fig. 1). Sites were haphazardly sampled across three reef zones (reef crest, flat, and slope; also referred to as reef habitat) and included fished sites and sites within no-take marine reserves. Depth was recorded for each site and ranged from 1.5 m to 10 m. At each site, benthic and coral reef fish surveys were conducted using underwater visual census methods to evaluate coral communities, habitat complexity, and reef fish assemblages.
Fig. 1

Map of 157 coral reef study sites surveyed between 2010 and 2013 in the Seychelles, Maldives, Chagos Archipelago, and Australia’s Great Barrier Reef. Underwater visual censuses of both benthic and fish assemblages occurred at each site, and points are slightly offset to reveal sampling effort

Coral assemblages

Coral assemblages were surveyed using two methods: 4 × 50-m point intercept transects in Chagos, Maldives, and Australia, and 8 × 10-m line intercept transects in the Seychelles. On point intercept transects, the substrate directly below the transect tape was surveyed every 50 cm. On line intercept transects, the length of each substrate type was measured along the entire 10-m distance. For each method, we recorded major benthic categories (e.g., hard coral, soft coral, sand, rubble, pavement, algae, sponge) and identified hard corals to genus and growth form using standard morphological and genus names. These methods are typically comparable (Facon et al. 2016), and here, we directly compared these methods by conducting a sensitivity analysis where we reran our entire analysis without the line intercept transects from the Seychelles and reached similar overall conclusions (Electronic supplementary material, ESM 1).

We estimated structural complexity on each transect using a visual six-point scale, following Polunin and Roberts (1993). The scale ranges from 0 (no vertical relief) to 5 (exceptionally complex habitats with numerous caves and overhangs); for a full description of this scale and pictures, see ESM 2. This measure of structural complexity has been shown to correlate well with a range of other structural complexity measurement techniques (Wilson et al. 2007), to strongly predict coral and reef fish diversity (Polunin and Roberts 1993; Chong-Seng et al. 2012; Newman et al. 2015), and reef recovery following mass bleaching events (Graham et al. 2015). All visual observations of structural complexity were recorded by one observer. This method has also been shown to quickly provide a reliable and effective estimate of habitat complexity, with regular training of multiple observers limiting observer bias (Wilson et al. 2007).

We derived 17 metrics of coral communities based on benthic surveys, including total cover of scleractinian corals, genera richness, percent cover of life-history groups, and abundance-weighted trait means and functional diversity (Table 1). Total coral cover was estimated as the average percent of living hard coral at each site, and genera richness was estimated as the average number of coral genera on replicate transects. Richness was estimated from raw estimates and not adjusted by rarefaction after a post hoc sensitivity analysis of two underwater survey methods revealed similar conclusions (ESM 1). Life histories were evaluated based on the abundance (percent cover) of four life-history groups—competitive, stress-tolerant, weedy, or generalist corals—proposed by Darling et al. (2012), which have revealed community shifts in response to various disturbances on coral reefs (Darling et al. 2013; Graham et al. 2014; Sommer et al. 2014). The functional structure of coral assemblages was evaluated by five traits that define the life histories reported in Darling et al. (2012): colony growth form (branching, plating, or domed), maximum colony size, growth rate (measured as annual linear extension, mm yr−1), reproductive mode (brooding or broadcast spawning), and fecundity. These five coral traits were selected because they had tangible links to reef fish assemblages (see Table 1 for detailed justification). A full list of observed hard corals, growth forms, and their associated traits can be found in ESM Table S1; trait information is also freely available online at https://coraltraits.org (Madin et al. 2016b). To estimate genus-level trait values, we averaged available trait information for species that occurred in the appropriate Indo-Pacific faunal province, based on Keith et al. (2013) (Africa–India province: Chagos, Maldives, and Seychelles; Australian province: Great Barrier Reef). For categorical traits (colony growth form and reproductive mode), we calculated the proportion of species in each genus with the trait (ESM Table S1). Six genera (Heliopora, Millepora, Oulaphyllia, Polyphyllia, Tubastrea, Tubipora) were removed from life history and trait analyses because of limited trait information (and these genera had limited abundances within our surveyed coral communities).

To derive multivariate measures of functional diversity, we conducted a principal component analysis on the five coral traits using a Gower dissimilarity matrix. We then estimated two measures of functional diversity: functional richness and Rao’s quadratic entropy. Functional richness was derived from the number of distinct functional groups in a community based on a Ward’s clustering analysis, and Rao’s quadratic entropy was measured from the average distance of abundance-weighted maximum pairwise distances in functional space at each site (Mouillot et al. 2013). At each site, we calculated a community-weighted trait mean for each of the five traits by weighting the values of each trait by the relative abundance of corals (Mouillot et al. 2013). Overall, our approaches estimate three unique components of traits and functional diversity to consider (1) functional richness of trait combinations (FunctionalRichness), (2) functional entropy of a community (FunctionalRao), and (3) abundance-weighted trait values for five traits. We used the package “FD” in R for all functional analyses (Laliberté and Legendre 2010).

Reef fish assemblages

We estimated the density and individual sizes of all species of diurnally active, non-cryptic, reef-associated fish >8 cm total length at each site using two comparable methods. In the Seychelles, eight replicate 7-m-radius point counts were conducted at each site (154 m2 replicate−1); all other locations used four replicate 50 m × 5 m belt transects (250 m2 replicate−1) for large and mobile fishes, while smaller and site-attached, territorial species (e.g., pomacentrids) were recorded using four replicate 50 m x 2 m belt transects (100 m2 replicate−1). For each transect, we calculated the total numerical abundance, biomass, and species richness of reef fish assemblages and averaged these values at each site. All estimates were standardized to estimates 250 m−2 for comparison across sites. As described in the benthic analyses, we conducted a sensitivity analysis to account for these different methodologies by rerunning analyses with and without the Seychelles dataset, which revealed very similar findings to the full dataset analysis; while there was some influence of survey methods on the list of variables in the top models, overall our main findings and conclusions were supported (ESM 1). We also included fish method (point count vs. belt transect) as a variance–covariate in the model framework to account for uneven variation across methods (see “Data analysis”). Total reef fish abundance was strongly correlated with the numerical abundance of small fish (<20 cm; general linear model R 2 = 0.85, p < 0.001), and we modeled both total abundance and the abundance of small fish <20 cm in separate analyses. This allowed us to model the abundance of both the overall fish assemblage and juvenile or small-bodied fish that are often more associated with the reef habitat (Beets and Hixon 1989; Bergman et al. 2000; Graham et al. 2007; Wilson et al. 2010). Reef fish biomass was calculated using standard length–weight conversions, and diversity was estimated as the total number of species observed at each site. Fish species were assigned to the following trophic (feeding) groups, following Graham et al. (2006): corallivores, herbivories, invertivores, mixed-diet feeders, piscivores, and planktivores. The link between these feeding groups and specific characteristics of the coral assemblage may be expected to be stronger than for total biomass.

Data analysis

For all analyses, we used multi-model averaging of mixed-effects linear models with maximum likelihood estimation to evaluate relationships between coral assemblages, structural complexity, and reef fish communities. We evaluated all variables for multicollinearity and found, not surprisingly, that coral life histories were strongly correlated with coral cover. Therefore, we conducted two similar analyses: one with life histories and one with coral cover (ESM 3). Our final variable set consisted of independent predictors with variance inflation factors <5, following Zuur et al. (2010) (Table 1). Notably, growth rate was removed from the candidate set of traits due to multicollinearity (VIF > 5; Table 1). Using mixed-effects general linear models, we first evaluated the effects of coral diversity (genera richness), traits, life histories, and structural complexity on reef fish abundance, biomass, and diversity, followed by a second analysis that evaluated the effects of coral diversity, traits, and life histories on structural complexity. All models were repeated by replacing the four coral life histories with total coral cover to address issues of multicollinearity between life histories and coral cover. Total biomass and biomass for the six trophic groups of fishes were log-transformed in all models.

All models included a random effect of country; sites were nested within country to account for biogeographic differences of latitude, large-scale environmental conditions, and differences in fishing pressure among the countries studied (see Eq. 1: site j in country i). We also applied residual variance structures to account for identifiable structure in the error residuals of our full model. Variance structures are variables used to model the structure of the residuals, without a penalty of adding more model parameters (Zuur et al. 2009). We used Akaike’s information criterion (AIC) to compare variance structures and identify which one performed better in all models. For all models, a varIdent variance structure was applied to account for uneven residual error within management groups (marine reserve vs. open access) and fish methods (belt transect vs. point count). For fish abundance, we also applied a varPower variance structure to habitat complexity that accounted for greater observed residual variation in more complex habitats; this error structure was not required for the biomass or diversity models, although we did apply it in models for the six trophic groups. We also tested a varIdent variance structure for coral method (point count vs. line intercept), which did not improve model fit in the fish models, but did improve fit in the analysis of structural complexity.

For the analysis of reef fish assemblages, we applied the following model structure to reef fish abundance, biomass, species richness, and trophic groups:
$$ \begin{aligned} {\text{Fish}}\,{\text{response}}_{ij} & = \beta_{0 } + \beta_{1 } \times {\text{Reef}}\,{\text{zone}} + \beta_{2 } \times {\text{Depth}} \\ & \quad + \beta_{3 } \times {\text{Management}} + \beta_{4 } \times {\text{Complexity}} + \beta_{5 } \\ & \quad \times {\text{Genera}}\,{\text{Richness}} + \beta_{6 } \times {\text{Functional}}\,{\text{Diversity}} \\ & \quad + \beta_{7 } \times {\text{Branching}} + \beta_{8 } \times {\text{MaxSize}} + \beta_{9 } \times {\text{Brooding}} \\ & \quad + \beta_{10 } \times {\text{Fecundity}} + \beta_{11 } \times {\text{Competitive}} + \beta_{12 } \times {\text{Generalist}} \\ & \quad + \beta_{13 } \times {\text{Stress}}\,{\text{tolerant}} + \beta_{14 } \times {\text{Weedy}} + \varepsilon_{ij} \\ & \quad \varepsilon_{ij} \sim\,N\left( {0,\sigma^{2}_{\text{Management}} } \right)\quad \varepsilon_{ij} \sim\,N\left( {0,\sigma^{2}_{{{\text{Fish }}\,{\text{method}}}} } \right)\quad \varepsilon_{ij} \sim\,N\left( {0,\sigma^{2} \times {\text{Complexity}}^{2\delta } } \right) \\ \end{aligned} $$
(1)

Prior to analysis, all input variables were centered to a mean of zero and standardized to a standard deviation of ±1, in order to allow for direct comparisons of variable effect sizes (in R, function “rescale” in package arm; Gelman 2008). For each fish response, we evaluated 3473 models by completing every combination of variables (up to five variables per model) and comparing models using AIC corrected for small sample size (AICc) (Akaike 1974). Within a top model set of 4 ∆AICc of the best model, we calculated model-averaged coefficients and 95% confidence intervals; significant coefficients were identified from confidence intervals that did not overlap zero. We also calculated the variable importance for each coefficient as the sum of Akaike weights across all models, which provided a rank for each variable, where a relative variable importance of 1.0 identifies the highest ranked input variable (Burnham and Anderson 2002). For model validation, we assessed the heterogeneity and normality of residuals across the top 95% set of candidate models. We fit all models in the package nlme (Pinheiro et al. 2015) with multi-model averaging using the package MuMin (Bartoń 2016); all analyses were conducted in R (R Core Team 2015).

We applied the same statistical approach to model structural complexity using characteristics of coral diversity, traits, and life histories (Eq. 2). Here, 2380 models were evaluated using every possible combination (up to five variables) with a random effect of country and a variance–covariate varIdent structure applied to management and coral method. The model structure was defined as:
$$ \begin{aligned} {\text{Structural}}\,{\text{complexity}}_{ij} & = \beta_{0 } + \beta_{1 } \times {\text{Reef}}\,{\text{zone }} + \beta_{2 } \\ & \quad \times {\text{Depth}} + \beta_{3 } \times {\text{Management}} + \beta_{4 } \\ & \quad \times {\text{Genera}}\,{\text{Richness}} + \beta_{5 } \times {\text{Functional}}\,{\text{Diversity}} \\ & \quad + \beta_{6 } \times {\text{Branching}} + \beta_{7 } \times {\text{MaxSize}} + \beta_{8 } \times {\text{Brooding}} \\ & \quad + \beta_{9 } \times {\text{Fecundity}} + \beta_{10 } \times {\text{Competitive}} + \beta_{11 } \times {\text{Generalist}} \\ & \quad + \beta_{12 } \times {\text{Stress}}\,{\text{tolerant}} + \beta_{13 } \times {\text{Weedy}} + \varepsilon_{ij} \\ & \quad \varepsilon_{ij} \sim\,N\left( {0,\sigma^{2}_{\text{Management}} } \right)\quad \varepsilon_{ij} \sim\,N\left( {0,\sigma^{2}_{{{\text{Coral}}\,{\text{method}}}} } \right) \\ \end{aligned} $$
(2)

As described above, all input variables were centered and standardized to a mean of zero ±1 standard deviation prior to analysis, and we conducted the same multi-model inferences using an AIC approach.

Results

Structural complexity and reef zone were the strongest predictors of reef fish abundance, biomass, diversity, and trophic structure across ~38,000 mixed-effects models. Structural complexity was a consistently top-ranked predictor in all models of reef fish assemblages with a maximum relative variable importance of 1.0 in all model comparisons (Table 2; ESM 3–6). Characteristics of reef zone and depth also revealed the importance of abundance and diversity, but not biomass; slope and crest zones were typically associated with more abundant and diverse reef fish communities than reef flats (Table 2; Fig. 2a–c).
Table 2

Top models describing reef fish assemblages: (a) abundance, (b) biomass, and (c) species richness. Checkmarks indicate the presence of variables in the 95% top model candidate set

Rank

Habitat, site characteristics

Diversity

Coral traits

Coral life histories

df

AICc

Δ AICc

w

Depth

Reef zone

Mgmt

Complex

Gen. richness

Rao

Branch

Brood

Fecun

Col. size

C

G

ST

W

(a) Abundance

1

 

 

      

 

12

1871.07

0.00

0.38

Relative variable importance

1.00

1.00

 

1.00

 

1.00

      

1.00

     

(b) Biomass

1

  

          

7

47.60

0.00

0.24

2

  

         

8

49.19

1.59

0.11

3

  

         

8

49.44

1.83

0.10

4

   

          

6

49.86

2.26

0.08

5

  

       

 

9

50.73

3.13

0.05

Relative variable importance

  

0.86

1.00

0.26

       

0.09

0.19

    

(c) Species richness

1

 

   

 

    

8

1090.17

0.00

0.08

2

 

 

   

 

   

8

1091.69

1.53

0.04

3

 

 

  

 

    

8

1091.87

1.70

0.04

4

 

   

 

    

8

1091.96

1.79

0.03

5

 

 

  

 

    

8

1092.04

1.87

0.03

6

 

 

 

 

 

    

8

1092.06

1.89

0.03

7

 

 

   

 

   

8

1092.42

2.25

0.03

8

 

 

   

    

8

1092.42

2.25

0.03

9

 

 

   

 

  

 

8

1092.51

2.35

0.03

10

 

 

  

  

   

8

1092.64

2.47

0.02

11

 

     

   

8

1092.69

2.52

0.02

12

 

 

   

 

 

  

8

1092.70

2.54

0.02

13

 

 

   

 

    

7

1092.97

2.81

0.02

14

 

 

   

    

8

1093.25

3.08

0.02

15

 

 

    

   

8

1093.37

3.20

0.02

16

 

  

  

    

8

1093.63

3.46

0.01

17

 

    

    

8

1093.74

3.57

0.01

18

 

     

   

8

1093.96

3.79

0.01

19

 

 

  

  

   

8

1094.02

3.85

0.01

20

 

     

  

 

8

1094.04

3.88

0.01

21

 

    

    

8

1094.06

3.90

0.01

22

     

    

8

1094.14

3.97

0.01

Relative variable importance

0.36

1.00

0.08

1.00

0.16

0.09

0.11

0.73

0.07

0.96

0.21

0.04

0.07

0.07

    

Mgmt management, Complex habitat complexity, C competitive life history, G generalist life history, ST stress-tolerant life history, W weedy life history, Branch branching, Brood brooding, Fecun fecundity, Col. size colony size, df degrees of freedom, AICc AIC scores corrected for small sample sizes, w model weight

Fig. 2

Multi-model-averaged parameter coefficients and 95% confidence intervals for a reef fish abundance, b biomass, and c species richness. Positive coefficients indicate a positive relationship between coral and fish characteristics, and negative coefficients indicate a negative relationship. The relationships between structural complexity and reef fish abundance (d), biomass (e), and diversity (f). Blue lines are linear model fits, and red lines are LOESS locally weighted smoothers with standard error for each. Shaded points show significant predictors of reef zone and management

Management (marine reserve vs. open access), coral cover and genera richness, functional diversity, species traits of colony size and branching, and some life histories were also associated with different characteristics of reef fish assemblages, including the abundance of small fishes and trophic groups (Table 2; Figs. 2, 3; ESM 3–6). Total reef fish abundance decreased with depth and functional diversity, and increased with more complex habitats, cover of stress-tolerant corals and slope zones (Fig. 2a); these patterns were consistent for the abundance of smaller fish <20 cm, which also included weaker effects of life histories, coral traits, genera richness, and management variables in the top models (ESM 4), although complexity only appeared in two of the top 11 models for smaller fish abundance and had a lower relative importance than aforementioned variables. Reef fish biomass increased with structural complexity within no-take reserves and, across all management types, with greater cover of stress-tolerant corals and decreased cover of weedy corals (Fig. 2b). Species richness was greater at sites in slope and crest habitats, which had higher complexity and larger maximum colony sizes. In addition, a suite of life histories, traits, depth, and management had weaker influences, i.e., with 95% confidence intervals overlapping zero and lower relative variable importance (Table 2; Fig. 2c). Overall, total abundance and biomass were predicted by fewer variables (5–6 in top models) than the abundance of smaller fish and diversity, which included 15 variables in each top model set (ESM 4).
Fig. 3

Influence of structural complexity and coral traits on the functional structure of reef fish assemblages: a corallivores, b herbivores, c invertivores, d mixed-diet feeders, e piscivores, and f planktivores. Model-averaged coefficients are shown with 95% confidence intervals

The biomass of reef fish trophic groups also included structural complexity as a main driver in each analysis; sites with more complex reef structure were associated with greater biomass of each trophic group (Fig. 3; ESM 5). Each trophic group was predicted by a different suite of traits, life histories, and other variables; for example, the biomass of corallivores increased with the abundance of competitive branching and plating corals and with depth (Fig. 3a), while the biomass of piscivores increased with depth and management, in addition to a suite of weaker trait and life history variables (Fig. 3e).

We also separated out the effects of management in these relationships by conducting separate analyses of total reef fish biomass within no-take marine reserves versus open-access fished reefs (ESM 6). This revealed that the influence of complexity, life histories, traits, and habitat on total biomass within marine reserves was more complex (i.e., 7 top variables) than on fished reefs, where biomass was predicted by complexity and a weaker influence of coral genera richness (ESM 6).

Structural complexity was predicted by a large suite of benthic and habitat characteristics, including hard coral traits and life histories (Fig. 4). Habitat and the abundance of competitive, stress-tolerant and generalist life histories, genera richness, and traits of colony size, branching, and fecundity were the strongest predictors of structural complexity and had the highest relative variable importance in the top 95% model set after evaluating ~3000 mixed-effects models (Fig. 4; Table 3). Structural complexity was typically higher on the reef slope and crest than reef flat and increased with percentage cover of competitive and generalist corals, genera richness, and maximum colony size; structural complexity was lower on reefs with a high abundance of corals with branching traits and high fecundity (Fig. 4b). Coral cover was also a significant and positive predictor of structural complexity; sites with higher coral cover were associated with higher structural complexity (ESM 3).
Fig. 4

a Multi-model-averaged parameter estimates and 95% confidence intervals for predictors of structural complexity. b Relationships between significant coral characteristics and structural complexity; blue lines are linear model fits, and red lines are LOESS locally weighted smoothers with standard error. The fit in the bottom right plot between coral cover and structural complexity is from model results presented in ESM 3

Table 3

Top models describing structural complexity of 157 sites in Seychelles, Maldives, Chagos, and the Great Barrier Reef. Checkmarks indicate the presence of variables in the 95% top model candidate set

Rank

Habitat, site characteristics

Diversity

Coral life histories

Coral traits

df

AICc

Δ AICc

w

Depth

Reef zone

Mgmt

Gen. richness

Rao

C

G

ST

W

Branch

Brood

Fecun

Col. size

1

 

 

 

   

   

10

215.06

0.00

0.14

2

 

 

 

   

 

 

11

215.66

0.61

0.10

3

 

 

 

   

  

11

215.89

0.84

0.09

4

 

 

 

  

   

11

217.53

2.47

0.04

5

 

   

  

   

10

217.59

2.53

0.04

6

 

   

 

   

11

218.04

2.98

0.03

7

 

 

 

   

  

11

218.33

3.27

0.03

8

 

 

   

   

11

218.51

3.45

0.02

9

 

   

 

    

10

218.68

3.62

0.02

10

 

 

 

  

   

11

218.76

3.70

0.02

11

 

   

  

  

11

218.87

3.81

0.02

Relative variable importance

0.04

1.00

 

0.80

 

1.00

0.23

0.04

0.09

0.96

0.09

0.18

0.20

    

Mgmt management, Complex habitat complexity, C competitive life history, G generalist life history, ST stress-tolerant life history, W weedy life history, Branch branching, Brood brooding, Fecun fecundity, Col. size colony size, df degrees of freedom, AICc AIC scores corrected for small sample sizes, w model weight

Discussion

Understanding the associations between habitat structure and reef fishes and mobile invertebrates has been a key focus of coral reef ecology for decades (Luckhurst and Luckhurst 1978; Roberts and Ormond 1987; Stella et al. 2011). Recently, trait-based metrics for both corals (Darling et al. 2012; Madin et al. 2016a) and reef fishes (Mouillot et al. 2013) have been proposed to offer new insights into patterns of diversity and function on coral reefs to provide a more mechanistic understanding of associations between benthic characteristics and fish assemblages. Here, we show that the abundance, biomass, diversity, and trophic structure of diurnal, non-cryptic reef fishes were strongly and consistently predicted by structural complexity and reef zone while characteristics of hard coral traits, total cover, life histories, and functional diversity added subtle, if weaker, effects to these relationships (Table 2; Fig. 2). Importantly, hard coral characteristics of total cover, traits, and life histories strongly influenced structural complexity (Table 3; Fig. 3), which highlights the indirect relationships by which living hard corals can provide structural complexity and habitat diversity for reef fishes.

The consistent influence of structural complexity and reef zone emphasize the importance of both living and non-living components of the reef that can structure reef fish assemblages. For example, we show that the living “veneer” provided by live hard corals and their associated traits, diversity, and life histories can also influence reef fish assemblages and contributes critically to structural complexity. Interestingly, different characteristics of the coral assemblage contributed to different components of the reef fish assemblage. Coral genera richness and functional diversity appeared in the top model sets for total biomass and diversity, the abundance of small fish, and the biomass of herbivores, piscivores, and planktivores, yet were mainly weak variables compared to the influence of complexity and habitat. Similarly, coral life histories and traits influenced total abundance, biomass and diversity, the abundance of small fish, and the biomass of corallivores, mixed-diet feeders, piscivores and planktivores but were also typically weaker predictors (Figs. 2, 3; ESM 4, 5). Thus, we conclude that the mechanistic hypothesis that coral diversity, life histories, and traits support the structure of reef fish assemblages is only weakly supported (e.g., Table 1), compared to the stronger influences of overall structural complexity and reef zone. It should be noted, however, that only reef fish >8 cm in length were surveyed. It is possible that smaller fishes would show a stronger relationship with some coral growth forms, such as branching corals.

We also found that some components of reef fish assemblages were influenced by more variables than others: notably species richness, the abundance of smaller individuals (8–20 cm), and the biomass of piscivores and planktivores, with >10 variables in the top model set (Figs. 2, 3; ESM 4). This suggests that different types of living corals and their traits influence different components of reef fish assemblages, as well as provide important habitat for the abundance of all fishes and smaller juvenile fish recruits even when adults of their species may have less affinity for live corals (Jones et al. 2004). Clearly, there are many complex and indirect interdependencies between reef corals and reef fishes that require further investigation (Pratchett et al. 2015), and a combination of structural complexity and living corals is likely required for optimal reef fish productivity and diversity.

Overall, the strong influence of seascape-level structural complexity on reef fish assemblages may be because this index integrates both living hard corals and non-living features of the reef matrix such as caves, crevices, or overhangs that can reflect reef growth and erosion processes accumulated over decades, centuries, or even millennia (Kleypas et al. 2001; ESM 2). Thus, the reef matrix can provide some resilience to disturbances that directly kill living corals (e.g., bleaching and outbreaks of crown-of-thorns starfish), which may also explain why fish diversity can be maintained following extensive coral mortality on inherently complex reefs (Wilson et al. 2009; Graham et al. 2015). However, we have shown that various aspects of the living corals are critical to providing structural complexity, and likely maintaining it long term. However, the seascape metric of structural complexity we used does not capture small-scale complexity, such as the presence of small holes <10 cm (Wilson et al. 2007), and our surveys focused on fish >8 cm, which limits our ability to make inferences about the how smaller fish use finer-scale structural complexity (e.g., smaller spaces within branching coral colonies).

No-take marine reserves were also, unsurprisingly, a key factor in predicting total reef fish biomass (e.g., MacNeil et al. 2015). This suggests that strategically locating marine reserves or fisheries management (e.g., gear restrictions) in areas with structurally complex reefs may provide greater returns on investment for biomass recovery than if focused on low-complexity reefs (McClanahan et al. 2011; MacNeil et al. 2015). Piscivore biomass was greater in reserves, and there were more complex associations between habitat and coral traits in reserves compared to fished reefs. This supports well-described evidence that overexploitation of fish can weaken predator–prey processes, and decouple biophysical relationships (Madin et al. 2010; Houk and Musburger 2013; Williams et al. 2015).

Here, our analysis of reef fish assemblages focused on site-level variables of habitat zonation and exposure and not broad biogeographic gradients of wave exposure or other environmental variables (e.g., currents, bleaching, or cyclone disturbances). For example, wave-induced water motion and flow velocity has been shown to structure the distribution and functional structure of reef fishes on Lizard Island (Fulton et al. 2005). Similarly, our correlative study did not consider how reef fish behavior—e.g., swimming ability and the use of flow or habitat refuges—might affect these patterns (Johansen et al. 2008). However, our findings agree with experimental studies on how reef fish use benthic habitat space, for example the finding that Acropora table corals can provide keystone structure for large reef fishes (Kerry and Bellwood 2012, 2015) and field observations that show tight links between specific coral and fish species (Coker et al. 2014). Clearly, controlled experimental studies will be necessary to understand the relationships between coral traits, life histories, and diversity that we have investigated here using field-based surveys and correlative analyses.

Hard coral assemblages strongly influenced structural complexity; sites with higher coral cover, greater genera richness, more abundant competitive and generalist life histories, and traits of larger colony sizes and higher growth rates were associated with more complex reefs (Table 3; Fig. 4; ESM 3). Surprisingly, the abundance of corals with branching traits was negatively associated with structural complexity (Fig. 3). This contrasts with studies that have found a positive association between the abundance of branching corals and fish assemblages via structural complexity (Chabanet et al. 1997; Messmer et al. 2011; Graham and Nash 2013). This may be because the refuge spaces associated with many fine branching corals are small (<10 cm), and this scale of complexity is not well captured by the visual assessments of structural complexity used here (Wilson et al. 2007). In addition, the dominance of corals with branching traits may also create a homogeneous habitat that might reduce the overall structural complexity of a reef. Alternatively, coral identities (specific genera) or life histories may be a better predictor of structural complexity than the individual coral traits investigated in this analysis. For example, the cover of competitive life histories (typically branching and plating acroporids; Darling et al. 2012) was positively associated with structural complexity, as was the number of coral genera. Previous studies have also found strong relationships between specific genera with structural complexity; for example, large, massive Montastrea (now Orbicella) colonies are associated with greater structural complexity (Alvarez-Filip et al. 2013) while the cover of branching Pocillopora was negatively related to reef complexity (Perry et al. 2015). However, as the new systematics of Scleractinia continue to inform and revise morphological classifications (e.g., Kitahara et al. 2016), trait selection and inference must carefully consider how traits are matched to genera observed on monitoring surveys.

While structural complexity and reef zonation are known to be major drivers of reef fish assemblages (Wilson et al. 2009; Graham and Nash 2013), here we show that coral traits and diversity can also have subtle effects on these relationships, as well as being key variables in determining structural complexity. This suggests that the long-term maintenance and production of future structural complexity on carbonate reefs will rely on living hard corals continuing to build the carbonate framework of reefs since dead corals inevitably break down and erode. Furthermore, this emphasizes the importance of past disturbance histories and recovery trajectories, particularly as disturbed coral assemblages shift in composition toward smaller weedier or opportunistic species that may provide less complexity overall (Darling et al. 2013; Alvarez-Filip et al. 2013).

As trait-based approaches move to advance coral reef science (Madin et al. 2016a, b), our study highlights two important caveats and considerations for future trait-based investigations. First, coral traits may have weak associations with reef fish assemblages, even if these traits are mechanistically associated with ecosystem processes of interest (e.g., growth forms, colony size, or growth rates). Here, the five traits used in this analysis had general relevance for fish community structure, although considering different traits, such as inter-colony volume, actual colony size, or volume under colonies, may have resulted in stronger relationships and different conclusions. Traits might also be explicitly tied to reef fish processes through specific links to energy production or provision of shelter, but empirical data on these processes are typically limited and may require additional experimental (e.g., Kerry and Bellwood 2012, 2015) or modeling approaches (e.g., biomass spectrum models or size-based theory; Trebilco et al. 2013). Second, most species-level coral trait information is largely available only as global averages, which does not reflect regional or intraspecific variation in coral traits. Given that intraspecific variation in coral morphology influences associations between fishes and host corals (e.g., Noonan et al. 2012), this will limit the usefulness of trait information when averaged at the level of genus (the common taxonomic resolution of most underwater coral surveys). Here, we used genus-level trait averages estimated from regionally appropriate species lists (the Indo-Pacific faunal provinces of Keith et al. 2013), yet, we still found relatively weak associations between traits and coral reef fish assemblages. Evaluating the coral traits of specific colonies (vs. species or generic averages) may yield stronger relationships, although this requires significantly more field effort which is logistically challenging for regional or global comparisons (Coker et al. 2014). However, we also found that simple life history groups based on species traits (from Darling et al. 2012) are useful predictors of structural complexity, which suggests that even coarse groupings of life history traits may have some utility in describing reef functions, such as structural complexity and the trophic structure of reef fish assemblages. Overall, the availability of new, open-access databases can aggregate and provide trait information online to test these hypotheses in the future (Madin et al. 2016b)

An obvious symptom of reef degradation is the transition toward net erosion, due to declines in reef accretion caused by declines in abundance and/or growth of calcifying organisms, and especially scleractinian corals (Kleypas et al. 2001; Alvarez-Filip et al. 2009; Perry et al. 2015). Maintaining key demographic processes of coral growth and carbonate accretion will be increasingly important to maintain structural complexity and habitat diversity for reef fishes and invertebrates. However, this is a challenging recommendation for managers and decision-makers in the context of increasing local stressors, ocean warming and acidification, and sea-level rise (Hoegh-Guldberg et al. 2007; Manzello et al. 2008; DeCarlo et al. 2015). Identifying relationships and thresholds between key demographic processes and specific management actions (like water clarity or the biomass of herbivorous fishes) remains a key focus for future research. Furthermore, ongoing coral community shifts toward stress-tolerant, weedy, and low-complexity species may superficially maintain coral cover while silently masking declines in complexity, coral accretion, and ecosystem services (Alvarez-Filip et al. 2013; Rogers et al. 2014).

Monitoring reef condition requires indicators that capture key processes and early signs of decline (Hughes et al. 2010; McClanahan et al. 2011). Here, we show that seascape visual estimates of structural complexity can be easily incorporated into monitoring and management programs (ESM 2). Interestingly, this method could also be calculated retrospectively from underwater images that provide a panoramic view about transects (Collins et al. 2016), or with machine-learning methods applied to satellite images or digital reef terrain maps, which would allow this methodology to scale up beyond SCUBA-based surveys to larger temporal and spatial scales (e.g., Pittman et al. 2009). For example, including a panoramic assessment at the start and end of the transect line that includes geomorphological reef features like caves or overhangs can provide important assessments of structural complexity relevant to fishes (e.g., Wilson et al. 2009). Further, while our study was based on observations, our results are supported by experimental studies that have manipulated structural complexity and specific habitats (e.g., the space below tabulate corals) to reveal the importance of complexity for reef fishes (Beets and Hixon 1989; Syms and Jones 2000; Gratwicke and Speight 2005; Kerry and Bellwood 2012, 2015). Overall, incorporating a standard measure of structural complexity into long-term monitoring programs may provide useful information on the loss of reef architecture and knock-on effects for coral reef ecosystems (e.g., Graham et al. 2015).

Continued declines in the three-dimensional complexity and coral diversity of tropical reefs will have consequences for reef fishes, fisheries, and the human societies that depend on these ecosystem services (Alvarez-Filip et al. 2013; Hicks and Cinner 2014; Rogers et al. 2014). Here, we highlight the importance of structural complexity as well as the more nuanced and subtle influences of coral traits and functional diversity on reef fishes. Maintaining structural complexity and coral diversity should be a key focus for managers, because it can help reefs recover to a coral-dominated state after mass bleaching (McClanahan et al. 2011; Graham et al. 2015) and supports fish diversity and fisheries productivity (Rogers et al. 2014; Holbrook et al. 2015). Managing for structural complexity may include prioritizing conservation and no-take reserves to reefs with high existing complexity and abundant live corals, such as including reef slopes with high complexity, or reefs where the underlying complexity is not so reliant on the accretion of reef by living corals (e.g., granitic reefs) (Harris et al. 2014). Management should also reduce activities that directly damage reef structural complexity or coral functional diversity, such as destructive fishing gears (dynamite, drag nets, some traps), physical damage by tourists, and boat anchoring. Here, we have highlighted the utility of a simple and standardized methodology for assessing large-scale patterns reef complexity, as well as the need to understand the subtler effects of hard coral traits and diversity, to meet the future challenges of monitoring a changing ecosystem.

Notes

Acknowledgements

We thank Sally Keith for providing species lists for Indo-Pacific provinces. ESD was supported by a David H. Smith Conservation Research Fellowship from the Cedar Tree Foundation, a Banting Fellowship from the Natural Sciences and Engineering Research Council of Canada, and the John D. and Catherine T. MacArthur Foundation. NAJG was supported by the Australian Research Council and a Royal Society University Research Fellowship. Field work in the Seychelles was supported by the Seychelles Fishing Authority, Seychelles National Parks Authority, Nature Seychelles, and Global Vision International. We thank the reviewers and Editor for suggestions that substantially improved earlier versions of this manuscript.

Supplementary material

338_2017_1539_MOESM1_ESM.docx (4.6 mb)
Supplementary material 1 (DOCX 4742 kb)

References

  1. Akaike H (1974) A new look at the statistical model identification. IEEE Trans Automat Contr 19:716–723CrossRefGoogle Scholar
  2. Alvarez-Filip L, Carricart-Ganivet JP, Horta-Puga G, Iglesias-Prieto R (2013) Shifts in coral-assemblage composition do not ensure persistence of reef functionality. Sci Rep 3:3486CrossRefPubMedPubMedCentralGoogle Scholar
  3. Alvarez-Filip L, Dulvy NK, Gill JA, Cote IM, Watkinson AR (2009) Flattening of Caribbean coral reefs: region-wide declines in architectural complexity. Proc R Soc Lond B Biol Sci 276:3019–3025CrossRefGoogle Scholar
  4. Bartoń K (2016) MuMIn: multi-model inference. R package version 1(15):6Google Scholar
  5. Beets J, Hixon M (1989) Shelter characteristics and Caribbean fish assemblages: experiments with artificial reefs. Bull Mar Sci 44:666–680Google Scholar
  6. Bergman K, Ohman M, Svensson S (2000) Influence of habitat structure on Pomacentrus sulfureus, a western Indian Ocean reef fish. Environ Biol Fishes 59:243–252CrossRefGoogle Scholar
  7. Beukers JS, Jones GP (1997) Habitat complexity modifies the impact of piscivores on a coral reef fish population. Oecologia 114:50–59CrossRefGoogle Scholar
  8. Burnham KP, Anderson D (2002) Model selection and multimodel inference: a practical information-theoric approach. Springer, New York, NY, p 488Google Scholar
  9. Chabanet P, Ralambondrainy H, Amanieu M, Faure G, Galzin R (1997) Relationships between coral reef substrata and fish. Coral Reefs 16:93–102CrossRefGoogle Scholar
  10. Chong-Seng KM, Mannering TD, Pratchett MS, Bellwood DR, Graham NAJ (2012) The influence of coral reef benthic condition on associated fish assemblages. PLoS One 7:e42167CrossRefPubMedPubMedCentralGoogle Scholar
  11. Coker DJ, Wilson SK, Pratchett MS (2014) Importance of live coral habitat for reef fishes. Rev Fish Biol Fish 24:89–126CrossRefGoogle Scholar
  12. Collins DL, Langlois TJ, Bond T, Holmes TH, Harvey ES, Fisher R, McLean DL (2016) A novel stereo‐video method to investigate fish Holmes TH, Harvey ES, Fisher R, McLean DL (2016) A nove [DOI:  10.1111/2041-210X.12650]
  13. Darling ES, Alvarez-Filip L, Oliver TA, Mcclanahan TR, Côté IM (2012) Evaluating life-history strategies of reef corals from species traits. Ecol Lett 15:1378–1386CrossRefPubMedGoogle Scholar
  14. Darling ES, McClanahan TR, Côté IM (2013) Life histories predict coral community disassembly under multiple stressors. Glob Chang Biol 19:1930–1940CrossRefPubMedGoogle Scholar
  15. DeCarlo TM, Cohen AL, Barkley HC, Cobban Q, Young C, Shamberger KE, Brainard RE, Golbuu Y (2015) Coral macrobioerosion is accelerated by ocean acidification and nutrients. Geology 43:7–10CrossRefGoogle Scholar
  16. Facon M, Pinault M, Obura D, Pioch S, Pothin K, Bigot L, Garnier R, Quod J (2016) A comparative study of the accuracy and effectiveness of line and point intercept transect methods for coral reef monitoring in the southwestern Indian Ocean islands. Ecol Indic 60:1045–1055CrossRefGoogle Scholar
  17. Ferrari R, McKinnon D, He H, Smith R, Corke P, González-Rivero M, Mumby P, Upcroft B (2016) Quantifying multiscale habitat structural complexity: a cost-effective framework for underwater 3D modelling. Remote Sens (Basel) 8:113CrossRefGoogle Scholar
  18. Friedlander AM, Brown EK, Jokiel PL, Smith WR, Rodgers KS (2003) Effects of habitat, wave exposure, and marine protected area status on coral reef fish assemblages in the Hawaiian archipelago. Coral Reefs 22:291–305CrossRefGoogle Scholar
  19. Fulton CJ, Bellwood DR, Wainwright PC (2005) Wave energy and swimming performance shape coral reef fish assemblages. Proc R Soc Lond B Biol Sci 272:827–832CrossRefGoogle Scholar
  20. Gelman A (2008) Scaling regression inputs by dividing by two standard deviations. Stat Med 27:2865–2873CrossRefPubMedGoogle Scholar
  21. Graham NAJ, McClanahan TR (2013) The last call for marine wilderness? Bioscience 63:397–402CrossRefGoogle Scholar
  22. Graham NAJ, Nash KL (2013) The importance of structural complexity in coral reef ecosystems. Coral Reefs 32:315–326CrossRefGoogle Scholar
  23. Graham NAJ, Chong-Seng KM, Huchery C, Januchowski-Hartley FA, Nash KL (2014) Coral reef community composition in the context of disturbance history on the Great Barrier Reef. Australia. PLoS One 9:e101204CrossRefPubMedGoogle Scholar
  24. Graham NAJ, Jennings S, MacNeil MA, Mouillot D, Wilson SK (2015) Predicting climate-driven regime shifts versus rebound potential in coral reefs. Nature 518:94–97CrossRefPubMedGoogle Scholar
  25. Graham NAJ, Wilson SK, Jennings S, Polunin NVC, Bijoux JP, Robinson J (2006) Dynamic fragility of oceanic coral reef ecosystems. Proc Natl Acad Sci U S A 103:8425–8429CrossRefPubMedPubMedCentralGoogle Scholar
  26. Graham NAJ, Wilson SK, Jennings S, Polunin NVC, Robinson J, Bijoux JP, Daw TM (2007) Lag effects in the impacts of mass coral bleaching on coral reef fish, fisheries, and ecosystems. Conserv Biol 21:1291–1300CrossRefPubMedGoogle Scholar
  27. Gratwicke B, Speight MR (2005) The relationship between fish species richness, abundance and habitat complexity in a range of shallow tropical marine habitats. J Fish Biol 66:650–667CrossRefGoogle Scholar
  28. Harris A, Wilson S, Graham N, Sheppard C (2014) Scleractinian coral communities of the inner Seychelles 10 years after the 1998 mortality event. Aquat Conserv Mar Freshw Ecosyst 24:667–679CrossRefGoogle Scholar
  29. Hicks CC, Cinner JE (2014) Social, institutional, and knowledge mechanisms mediate diverse ecosystem service benefits from coral reefs. Proc Natl Acad Sci U S A 111:17791–17796CrossRefPubMedPubMedCentralGoogle Scholar
  30. Hixon M, Beets J (1993) Predation, prey refuges, and the structure of coral reef fish assemblages. Ecol Monogr 63:77–101CrossRefGoogle Scholar
  31. Hoegh-Guldberg O, Hooten AJ, Steneck RS, Greenfield P, Gomez E, Harvell CD, Sale PF, Edwards AJ, Caldeira K, Knowlton N, Eakin CM, Iglesias-Prieto R, Muthiga N, Bradbury RH, Dubi A, Hatziolos ME (2007) Coral reefs under rapid climate change and ocean acidification. Science 318:1737–1742CrossRefPubMedGoogle Scholar
  32. Holbrook SJ, Schmitt RJ, Brooks AJ (2008) Resistance and resilience of a coral reef fish community to changes in coral cover. Mar Ecol Prog Ser 371:263–271CrossRefGoogle Scholar
  33. Holbrook SJ, Schmitt RJ, Messmer V, Brooks AJ, Srinivasan M, Munday PL, Jones GP (2015) Reef fishes in biodiversity hotspots are at greatest risk from loss of coral species. PLoS One 10:e0124054CrossRefPubMedPubMedCentralGoogle Scholar
  34. Houk P, Musburger C (2013) Trophic interactions and ecological stability across coral reefs in the Marshall Islands. Mar Ecol Prog Ser 488:23–34CrossRefGoogle Scholar
  35. Hughes TP, Graham NAJ, Jackson JBC, Mumby PJ, Steneck RS (2010) Rising to the challenge of sustaining coral reef resilience. Trends Ecol Evol 25:633–642CrossRefPubMedGoogle Scholar
  36. Hutchinson GE (1959) Homage to Santa Rosalia or why are there so many kinds of animals? Am Nat 93:145–159CrossRefGoogle Scholar
  37. Johansen JL, Bellwood DR, Fulton CJ (2008) Coral reef fishes exploit flow refuges in high-flow habitats. Mar Ecol Prog Ser 60:219–226CrossRefGoogle Scholar
  38. Jones GP, McCormick MI, Srinivasan M, Eagle JV (2004) Coral decline threatens fish biodiversity in marine reserves. Proc Natl Acad Sci U S A 101:8251–8253CrossRefPubMedPubMedCentralGoogle Scholar
  39. Keith SA, Baird AH, Hughes TP, Madin JS, Connolly SR (2013) Faunal breaks and species composition of Indo-Pacific corals: the role of plate tectonics, environment and habitat distribution. Proc R Soc Lond B Biol Sci 280:20130818CrossRefGoogle Scholar
  40. Kerry JT, Bellwood DR (2012) The effect of coral morphology on shelter selection by coral reef fishes. Coral Reefs 31:415–424CrossRefGoogle Scholar
  41. Kerry JT, Bellwood DR (2015) Do tabular corals constitute keystone structures for fishes on coral reefs? Coral Reefs 34:41–50CrossRefGoogle Scholar
  42. Kitahara MV, Fukami H, Benzoni F, Huang D (2016) The new systematics of Scleractinia: integrating molecular and morphological evidence. In: Goffredo S, Dubinsky Z (eds) The Cnidaria, Past, Present and Future. Springer, Netherlands, pp 41–59CrossRefGoogle Scholar
  43. Kleypas J, Buddemeier RW, Gattuso J-P (2001) The future of coral reefs in an age of global change. Int J Earth Sci 90:426–437CrossRefGoogle Scholar
  44. Laliberté E, Legendre P (2010) A distance-based framework for measuring functional diversity from multiple traits. Ecology 91:299–305CrossRefPubMedGoogle Scholar
  45. Luckhurst E, Luckhurst K (1978) Analysis of the influence of substrate variables on coral reef fish communities. Mar Biol 323:317–323CrossRefGoogle Scholar
  46. MacNeil MA, Graham NAJ, Cinner JE, Wilson SK, Williams ID, Maina J, Newman S, Friedlander AM, Jupiter S, Polunin NVC, McClanahan TR (2015) Recovery potential of the world’s coral reef fishes. Nature 520:341–344CrossRefPubMedGoogle Scholar
  47. Madin EMP, Gaines SD, Madin JS, Warner RR (2010) Fishing indirectly structures macroalgal assemblages by altering herbivore behavior. Am Nat 176:785–801CrossRefPubMedGoogle Scholar
  48. Madin JS, Hoogenboom MO, Connolly SR, Darling ES, Falster D, Huang D, Keith SA, Mizerek T, Pandolfi JM, Putnam HM, Baird AH (2016a) A trait-based approach to advance coral reef science. Trends Ecol Evol 31:419–428CrossRefPubMedGoogle Scholar
  49. Madin JS, Anderson K, Andreason M, Bridge TCL, Cairns S, Connolly SR, Darling ES, Diaz M, Falster D, Franklin EC, Gates RD, Hoogenboom MO, Huang D, Keith SA, Kosnik M, Kuo C-Y, Lough JM, Lovelock CE, Luiz O, Martinelli J, Mizerek T, Pandolfi JM, Pochon X, Putnam HM, Pratchett MS, Roberts E, Stat M, Wallace CC, Widman E, Baird AH (2016b) The Coral Trait Database, a curated database of trait information for coral species from the global oceans. Sci Data 3:160012CrossRefGoogle Scholar
  50. Manzello DP, Kleypas JA, Budd DA, Eakin CM, Glynn PW, Langdon C (2008) Poorly cemented coral reefs of the eastern tropical Pacific: possible insights into reef development in a high-CO2 world. Proc Natl Acad Sci U S A 105:10450–10455CrossRefPubMedPubMedCentralGoogle Scholar
  51. McClanahan TR, Graham NAJ, MacNeil MA, Muthiga NA, Cinner JE, Bruggemann JH, Wilson SK (2011) Critical thresholds and tangible targets for ecosystem-based management of coral reef fisheries. Proc Natl Acad Sci U S A 108:17230–17233CrossRefPubMedPubMedCentralGoogle Scholar
  52. McCormick M (1994) Comparison of field methods for measuring surface topography and their associations with a tropical reef fish assemblage. Mar Ecol Prog Ser 112:87–96CrossRefGoogle Scholar
  53. Messmer V, Jones GP, Munday PL, Holbrook SJ, Schmitt RJ, Brooks AJ (2011) Habitat biodiversity as a determinant of fish community structure on coral reefs. Ecology 92:2285–2298CrossRefPubMedGoogle Scholar
  54. Mouillot D, Graham NAJ, Villéger S, Mason NWH, Bellwood DR (2013) A functional approach reveals community responses to disturbances. Trends Ecol Evol 28:167–177CrossRefPubMedGoogle Scholar
  55. Nash KL, Graham NAJ, Wilson SK, Bellwood DR (2012) Cross-scale habitat structure drives fish body size distributions on coral reefs. Ecosystems 16:478–490CrossRefGoogle Scholar
  56. Newman SP, Meesters EH, Dryden CS, Williams SM, Sanchez C, Mumby PJ, Polunin NVC (2015) Reef flattening effects on total richness and species responses in the Caribbean. J Anim Ecol 84:1678–1689CrossRefPubMedGoogle Scholar
  57. Noonan SHC, Jones GP, Pratchett MS (2012) Coral size, health and structural complexity: effects on the ecology of a coral reef damselfish. Mar Ecol Prog Ser 456:127–137CrossRefGoogle Scholar
  58. Perry CT, Murphy GN, Graham NAJ, Wilson SK, Januchowski-Hartley FA, East HK (2015) Remote coral reefs can sustain high growth potential and may match future sea-level trends. Sci Rep 5:18289CrossRefPubMedPubMedCentralGoogle Scholar
  59. Pinheiro J, Bates D, DebRoy S, Sarkar D (2015) nlme: linear and nonlinear mixed effects models. R package version 3.1-120Google Scholar
  60. Pittman SJ, Costa BM, Battista TA (2009) Using lidar bathymetry and boosted regression trees to predict the diversity and abundance of fish and corals. J Coast Res 10053:27–38CrossRefGoogle Scholar
  61. Polunin NVC, Roberts CM (1993) Greater biomass and value of target coral-reef fishes in two small Caribbean marine reserves. Mar Ecol Prog Ser 100:167–176CrossRefGoogle Scholar
  62. Pratchett M, Anderson K, Hoogenboom M, Widman E, Baird A, Pandolfi J, Edmunds P, Lough J (2015) Spatial, temporal and taxonomic variation in coral growth—implications for the structure and function of coral reef ecosystems. Oceanogr Mar Biol Annu Rev 53:215–296CrossRefGoogle Scholar
  63. Core Team R (2015) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, AustriaGoogle Scholar
  64. Roberts CM, Ormond RFG (1987) Habitat complexity and coral reef fish diversity and abundance on Red Sea fringing reefs. Mar Ecol Prog Ser 41:1–8CrossRefGoogle Scholar
  65. Rogers A, Blanchard JL, Mumby PJ (2014) Vulnerability of coral reef fisheries to a loss of structural complexity. Curr Biol 24:1000–1005CrossRefPubMedGoogle Scholar
  66. Sommer B, Harrison PL, Beger M, Pandolfi JM (2014) Trait-mediated environmental filtering drives assembly at biogeographic transition zones. Ecology 95:1000–1009CrossRefPubMedGoogle Scholar
  67. Stella JS, Pratchett MS, Hutchings PA, Jones GP (2011) Coral-associated invertebrates: diversity, ecological importance and vulnerability to disturbance. Oceanogr Mar Biol Annu Rev 49:43–104Google Scholar
  68. Syms C, Jones GP (2000) Disturbance, habitat structure, and the dynamics of a coral-reef fish community. Ecology 81:2714–2729CrossRefGoogle Scholar
  69. Trebilco R, Baum JK, Salomon AK, Dulvy NK (2013) Ecosystem ecology: size-based constraints on the pyramids of life. Trends Ecol Evol 28:423–431CrossRefPubMedGoogle Scholar
  70. Williams GJ, Gove JM, Eynaud Y, Zgliczynski BJ, Sandin SA (2015) Local human impacts decouple natural biophysical relationships on Pacific coral reefs. Ecography 38:751–761CrossRefGoogle Scholar
  71. Wilson SK, Graham NAJ, Polunin NVC (2007) Appraisal of visual assessments of habitat complexity and benthic composition on coral reefs. Mar Biol 151:1069–1076CrossRefGoogle Scholar
  72. Wilson SK, Depczynski M, Fulton CJ, Holmes TH, Radford BT, Tinkler P (2016) Influence of nursery microhabitats on the future abundance of a coral reef fish. Proc R Soc Lond B Biol Sci 283:20160903CrossRefGoogle Scholar
  73. Wilson SK, Babcock RC, Fisher R, Holmes TH, Moore JY, Thomson DP (2012) Relative and combined effects of habitat and fishing on reef fish communities across a limited fishing gradient at Ningaloo. Mar Environ Res 81:1–11CrossRefPubMedGoogle Scholar
  74. Wilson SK, Dolman AM, Cheal AJ, Emslie M, Pratchett MS, Sweatman HPA (2009) Maintenance of fish diversity on disturbed coral reefs. Coral Reefs 28:3–14CrossRefGoogle Scholar
  75. Wilson SK, Fisher R, Pratchett MS, Graham NAJ, Dulvy NK, Turner RA, Cakacaka A, Polunin NVC (2010) Habitat degradation and fishing effects on the size structure of coral reef fish communities. Ecol Appl 20:442–451CrossRefPubMedGoogle Scholar
  76. Zuur AF, Ieno EN, Elphick CS (2010) A protocol for data exploration to avoid common statistical problems. Methods Ecol Evol 1:3–14CrossRefGoogle Scholar
  77. Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed-effects models and extensions in ecology with R. Springer, New York, NYCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2017

Authors and Affiliations

  • Emily S. Darling
    • 1
    • 2
  • Nicholas A. J. Graham
    • 3
    • 4
  • Fraser A. Januchowski-Hartley
    • 5
  • Kirsty L. Nash
    • 6
    • 7
  • Morgan S. Pratchett
    • 4
  • Shaun K. Wilson
    • 8
    • 9
  1. 1.Marine Program, Wildlife Conservation SocietyBronxUSA
  2. 2.Department of Ecology and Evolutionary BiologyUniversity of TorontoTorontoCanada
  3. 3.Lancaster Environment CentreLancaster UniversityLancasterUK
  4. 4.ARC Centre for Excellence for Coral Reef StudiesJames Cook UniversityTownsvilleAustralia
  5. 5.IRD UMR 9190 MARBEC, IRD-CNRS-IFREMER-UMUniversité de MontpellierMontpellierFrance
  6. 6.Centre for Marine SocioecologyHobartAustralia
  7. 7.Institute of Marine and Antarctic StudiesUniversity of TasmaniaHobartAustralia
  8. 8.Marine Science Program, Department of Parks and WildlifeKensingtonAustralia
  9. 9.Oceans InstituteUniversity of Western AustraliaCrawleyAustralia

Personalised recommendations