Environ Biol Fish (2015) 98:1645–1657

DOI 10.1007/s10641-015-0390-8

Hierarchizing biological, physical and anthropogenic factors

influencing the structure of fish assemblages along tropical
rocky shores in Brazil
Tatiana Pires Teixeira-Neves & Leonardo Mitrano Neves &
Francisco Gerson Araújo

Received: 15 September 2014 / Accepted: 26 January 2015 / Published online: 7 February 2015
# Springer Science+Business Media Dordrecht 2015

Abstract Understanding the distribution patterns of increased with distance from the coast, and this relation-
reef fish and the relationships between those patterns ship is likely linked to the presence of large top preda-
and habitat and anthropogenic factors is important for tors and herbivores. Moreover, fish richness and density
the development of conservation policies by environ- increased with the physical complexity indicated by the
mental managers. Fish assemblage structure was studied physical structure index, suggesting that the presence of
over 22 rocky shores with different physical complexity a variety of refuges enhances the availability of shelter.
and benthic cover in Ilha Grande Bay, Southeastern We recommend that areas farthest from the urban cen-
Brazil. We aimed to test the relative influences on rocky tres and with higher physical complexity should be
reef fish assemblage descriptors (richness, density and prioritised in conservation policies.
biomass) of three categories of predictors: 1) biological
features as dominant benthic cover (in percentage), i.e., Keywords Fish assemblage . Habitat structure . Shallow
fleshy algae, turf algae and soft coral; 2) physical fac- waters . Anthropogenic impacts . Conservation
tors, i.e., depth and a physical structure index; and 3)
anthropogenic factors, i.e., distance from the coast, pop-
ulation of the nearest city, and influence of a marine
protection area. The main explanatory variables deter- Introduction
mining fish assemblage structure according to the dis-
tance based linear model (DistTLM) were depth Responses of marine assemblages to habitat compo-
(explaining 16.7 % of the variation) and distance from nents can be used to identify the specific physical or
the coast (14.0 %), followed by population of the nearest ecological attributes influencing distribution patterns
city (3.7 %) and turf algae (2.9 %). Similarly, fish (Moore et al. 2010). Factors known to influence the
species richness was positively associated with deeper spatial patterns of reef fish assemblages include the
areas and greater distance from the coast, thus being less biological, such as benthic cover (Chittaro et al. 2005;
accessible to human influence. Fish density and biomass McClanahan and Karnauskas 2011), physical, such as
depth (Srinivasan 2003) and wave exposure (Floeter
et al. 2007), and anthropogenic, such as sewage pollu-
tion (Azzurro et al. 2010), fishing pressure (Floeter et al.
2006) or proxies of human influences such as distance
T. P. Teixeira-Neves : L. M. Neves : F. G. Araújo (*) from a specific source (e.g., discharge of sewage or
Laboratório de Ecologia de Peixes, Universidade Federal
thermal effluent) and demographic metrics (Nguyen
Rural do Rio de Janeiro, BR 465, Km 7,
23870-000 Seropédica, RJ, Brazil and Phan 2008; Vincent et al. 2011; Gibran and Moura
e-mail: [email protected] 2012). Hierarchizing the influences of these various
1646 Environ Biol Fish (2015) 98:1645–1657

factors on the responses of species and assemblage and trophic structure. Therefore, variables such as dis-
structures to major environmental gradients requires tance from the coast (Williams 1982; Green et al. 1987;
the use of spatial models that incorporate the combined Adjeroud et al. 1998; Lecchini et al. 2003) and popula-
influence of those factors. Identifying the most impor- tion of the nearest city (Vincent et al. 2011) can be used
tant factors that determine spatial patterns of marine as a proxy for changes in spatial structure and distribu-
organisms is an essential step in helping environmental tion of fish assemblages. Concerns have been raised
managers design effective conservation measures. regarding the increasing degradation of coastal areas.
Habitat structure (biological and physical) is known Marine protected areas (MPAs) are an alternative for
to be among the main factors structuring fish communi- managing and conserving littoral resources (Lester et al.
ties (McGehee 1994; Lara and Gonzalez 1998; Wantiez 2009; Halpern et al. 2010). One objective of unfished
and Chauvet 2003). Further, benthic composition is an zones in MPAs is to help maintain viable fisheries in
important factor structuring habitat complexity adjacent areas by increasing the density and size of fish,
(Bouchon-Navarro and Bouchon 1989; Jones et al. providing centres for dispersal of individuals and larvae,
1991; Munday 2002; Floeter et al. 2007), adding niches and augmenting local fishery yields through biomass
to a diverse fauna of mobile and sessile invertebrates. exportation from the protected area (García-Charton
Benthic organisms may also contribute to structural et al. 2008). Although this region has an MPA, it is
complexity provided by boring and arborescent organ- poorly managed and is affected by factors including
isms on a scale of centimetres and can provide meso- coastal development (e.g., port activity, thermal pollu-
scale habitat modification for small benthic fishes tion, sewage discharge) and fishery activities (bottom
(Stephen et al. 2006). Physical structure (Luckhurst trawling, recreational fisheries targeting reef fishes).
and Luckhurst 1978; Roberts and Ormond 1987; Consequently, quality assessment and monitoring of
Ohman and Rajassuriya 1998; Teixeira et al. 2012) is marine ecosystems has become increasingly important
closely related to fish species richness and can reduce to ensure their sustainability (Spatharis and Tsirtsis
predation and competition by providing more refuges, 2010; Borja et al. 2012).
contributing to decreased encounter rates between In Brazil, the vast majority of the literature on rocky
predators and prey prey, and increasing the availability shores fish assemblage is focused on describing fish
of resources such as food, shelter, and spawning composition and its relationship with habitat structure
areas (Murdoch and Oaten 1975; Letourneur 1996; (Ferreira et al. 2001; Floeter et al. 2007; Krajewski and
Friedlander and Parrish 1998; Almany 2004; Grober- Floeter 2011). However, few studies have been conduct-
Dunsmore et al. 2008). Relationships of physical char- ed to assess the influences of anthropogenic factors
acteristics of the habitat such as depth (Williams 1991; (Floeter et al. 2006) and both biotic and abiotic vari-
Srinivasan 2003) and complexity with fish community ables, exploring their eventual interactions, and hierar-
structure have been reported elsewhere (Carpenter et al. chizing their influence on fish assemblages. In this
1981; Mcmanus et al. 1981; Grigg 1994; Chabanet et al. study, we aimed to test the relative influences of these
1997; Lewis 1997; Adjeroud et al. 1998). three categories of predictors (biological, physical and
The process of human development, with the growth anthropogenic factors) on descriptors (richness, density
of urban centres and demand for natural resources, is and biomass) of fish assemblages.
increasingly affecting coastal environments. Fish as-
semblages may not respond directly to human popula-
tion density but may respond instead to a variety of Material and methods
direct and indirect factors, such as urban development,
tourism, fishing, sedimentation, and habitat degradation, Study area
that are related to human population density (Benjamin
et al. 2012). Fish assemblages may respond to distur- Ilha Grande Bay (IGB) is located in the southern tip of
bances such as sewage discharges (Russo 1982; Smith Rio de Janeiro state (23°06 S, 44°42 W), SE Brazil,
et al. 1999; Guidetti et al. 2002; Islam and Tanaka 2004) and covers an area of approximately 600 km2, contain-
and fishing (Grigg 1994; Chabanet et al. 1995; Otway ing approximately 350 islands surrounded by shallow
et al. 1996) that influence the diversity, dominance, water (typically no more than 8 m in depth) (Ignacio
biomass, abundance, reproductive inhibition or failure et al. 2010). The bay is an oligotrophic system that is
Environ Biol Fish (2015) 98:1645–1657 1647

Fig. 1 Map of Ilha Grande Bay

showing the study sites. Mainland
sites are represented by axes,
MPA sites by circles, islands by
triangles, and Ilha Grande Island
rocky shores by trapezoids. Large
circles in the mainland represent
most populous cities in the area

freely connected to the Atlantic Ocean (Fig. 1), receiv- sources such as river discharges, mangroves, sewage
ing no significant input of freshwater from rivers. The outfalls and thermal discharges; thus, confounding ef-
shoreline is highly jagged, and coastal mountains reach fects were avoided, and the influences of the selected
the coastline, leaving little space for coastal plain evo- environmental variables could be better assessed.
lution (Guerra and Soares 2009). Economic activities Underwater visual censuses were performed along tran-
conducted in IGB include tourism, power generation sects of 20×2 m (40 m2) following current practice
(nuclear power plants), shipyards, private marinas, oil along the Brazilian coast (Ferreira et al. 2001; Floeter
terminals and fisheries. Despite its economic and eco- et al. 2006, 2007; Krajewski and Floeter 2011), at
logical importance, IGB has not been the subject of Bshallow^ (the upper limit of the low tide) and
systematic biological studies. The main physiographic Binterface^ (situated on the consolidated substrate, at
structure of the study area is characterised by narrow least 2 m from the sand/rocky interface) zones at each
rocky shores covered by granite boulders, ending in a site. The depth of the Bshallow^ zone varied from 1.2 to
sand bottom (interface). The sampled rocky shores had 2.9 m, whereas that of the Binterface^ area varied from 3
average depths ranging from 2.1 (shallow) to 5.1 m to 7.2 m. At each site, six evenly spaced transects were
(interface). IGB has an MPA at the Ecological Station performed (three transects at each depth zone), totalling
of Tamoios that began to operate in 1990 and comprises 114 transects in the entire study area. Each transect was
twenty-nine islets, islands, rocks and slabs. sampled twice for fish counts. In the first pass, the diver
Progressively increasing threats to the bay include sew- swam along the transect and recorded all mobile fishes
age discharge from coastal development, industrial fish- in water column. In the second pass, the observer fo-
eries (including destructive bottom trawling) and recre- cused on searching beneath rocks and in crevices to
ational fishing targeting rocky reef fishes (harpoon and observe the more cryptic species (e.g., Blenniidae,
hook-and-line fishing). Gobiidae, Muraenidae). The sampling unit, number of
fish per 40 m2, was defined as the pooled number of
Sampling surveys mobile and cryptic fishes.

Fish assemblages and habitat characteristics of twenty- Physical structure

two rocky shores (sites) were selected: five in the main-
land, seven in the islands, six in the MPA and four The quantitative assignment of physical structure was
around Ilha Grande. Sampling was conducted in the based on depth and a physical structure index. The depth
winters of 2010 and 2011. Rocky shores were chosen was obtained through the average of three measure-
in areas sheltered from wave activity, far from point ments for each fish transect, evenly spaced along the
1648 Environ Biol Fish (2015) 98:1645–1657

length of the transect. Twenty photographs were taken at overlaying 20 random points on each image and identi-
each fish sampling and were used to quantify the vari- fying the substratum under each point. Benthic organ-
ables describing physical structure following Chapman isms, expressed as percentage of benthic cover, were
et al. (1995). A digital camera was mounted onto a grouped as follows: 1) fleshy algae, comprised mainly
0.36 m2 polyvinyl chloride (PVC) photo quadrat frame, of crustose carbonate algae, coenocytic thalli, and
and the following physical characteristics were quanti- Sargassum spp.; 2) turf algae, composed of a matrix of
fied posteriorly: (1) number of refuges (holes and crev- small macroalgae mainly belonging to the orders
ices); (2) percentage of hard substratum; (3) number of Corallinales, Ceramiales and other green and red fila-
rocks of different sizes. The fourth characteristic, (4) mentous algae (Thrush et al. 2011); and 3) soft coral,
substratum height, was estimated during the sampling. represented by the Zoanthids Palythoa caribaeorum and
Number of refuges and number of rocks were grouped Zoanthus sociatus. Other invertebrates, such as tuni-
in three size ranges (<30 cm; 30 cm-1 m; >1 m) follow- cates, cirripedia, sponges, hydrozoa, Echinodermata,
ing the method proposed by Aburto-Opereza and Balart bryozoans and hard coral (scleractinians such as
(2001). Holes, crevices and rocks of different sizes form Mussismilia hispida and the invasive corals of the genus
gaps between structures that can provide paths for fish to Tubastraea) occurred rarely in this area and were not
escape from predators. BHard substratum^ refers to the included in the analyses.
percentage of substratum that was not sand, rubble or
shell-sand patches. Substratum height was a subjective
visual estimate taken at an intermediate position be- Anthropogenic factors
tween the top of the rocks and the lowest point for each
transect. The physical structure index was adapted from The following variables related to anthropogenic influ-
Gratwicke and Speight (2005). The scores were 1 for ence were included in this study: 1) distance from the
lowest, 3 for intermediate and 5 for the highest physical coast; 2) population of the nearest city; and, 3) whether
complexity. For example, rocky shores with a single the site was inside or outside an MPA. The distance from
refuge size receive scores of 1, whereas those with refuges the coast (measured in kilometres) was included as a
of all different size categories receive scores of 5. A total proxy of anthropogenic influences because the rocky
score was calculated by adding the scores of each of the shores were comparatively less accessible to human
four physical characteristics to give an estimate of the influence. For the rocky shores closer to the coast, where
overall degree of complexity of the sites (Table 1). there are ramps, marinas and beaches, access to the
islands by small boats in the bay also means greater
influence of anthropogenic activities. The population of
Benthic cover the nearest city can characterise the influence of urban
disturbances on the marine environment in relation to
Photographs were also analysed to measure the percent- the intensity of exploitation of nearby coastal resources.
age of benthic cover using Coral Point Count with Excel The MPA was also characterised as a factor related to
Extensions-CPCe 3.4 (Kohler and Gill 2006) by anthropogenic influence. Although the enforcement of

Table 1 Scoring system for the

physical structure index and the Descriptors Physical structure score
recorded values for each physical
characteristic 1 3 5

(1) Number of refuge size categories (holes and crevices 0–1 2–4 4–6
in the following size categories for each refuge type:
<30 cm; 30 cm-1 m; > 1 m)
(2) Hard substratum (%) 0–20 20–40 40–50
(3) Substratum height (visual estimate of height of physical 0–50 51–100 >100
structure-cm)
(4) Number of rocks of different sizes (in the three following 0–1 2 3
size categories: <30 cm; 30 cm-1 m;> 1 m)
Environ Biol Fish (2015) 98:1645–1657 1649

restrictive measures is inadequate, we expect that areas dbRDA axes were then examined to identify the dom-
inside the MPA may be better preserved than areas inant species driving the response of the fish assemblage
outside the MPA because they are legally protected. structure to the physical, biological and anthropogenic
predictors. The strength and direction (positive or neg-
Data analyses ative) of the associations between the univariate fish
assemblage descriptors (fish richness, density and bio-
The following response variables derived from the fish mass) and the selected variables according to DistLM
assemblage data were used: fish assemblage structure, fish analyses were identified using the multiple partial cor-
richness, density and biomass. Total numbers of species relations of the variables with the first dbRDA axis and
(richness) and individuals (density) were calculated based scatterplots.
on observations from each transect. Total fish biomass per
transect was estimated by length-weight relationship
curves and allometric conversions: W=a×Lb, where pa- Results
rameters ‘a’ and ‘b’ are constants for the allometric equa-
tion from FishBase (www.fishbase.org), Macieira and A total of 67 fish species, distributed over 34 families,
Joyeux (2008) and Camilato et al. (2010). When coeffi- were identified during the entire survey. The distance-
cient values were not available for the species, we used based multivariate linear model (DistLM) analysis indi-
coefficients for the most closely related species. For the cated significant relationships between fish assemblage
other parameters, the size range and nearest region were and six of the predictors: depth, distance from the coast,
chosen. Prior to analyses, fish assemblage, fish richness, population of the nearest city, percentage of turf algae,
density and biomass data were square-root transformed. percentage of fleshy algae and percentage of soft coral.
Bray-Curtis similarity matrices were calculated for multi- Together, these factors accounted for 39.2 % of the
variate data (fish assemblage), while Euclidean similarity variance in the fish assemblage. Depth was the most
matrices were used for univariate variables (fish richness, important environmental factor, explaining 16.7 % of
density and biomass). Pairwise correlation coefficients the total variation, followed by distance from the coast
were calculated between all predictors to detect eventual (14.0 %) (Table 2). The remaining variables contributed
co-linearity (r<0.7; Sleeman et al. 2005; Leathwick et al. a small but significant component of the variance. These
2006), but no significant associations were found among included the population of the nearest city (3.7 %) and
these variables. the percentages of turf algae (2.9 %), fleshy algae
The relationships between fish assemblage descrip- (1.6 %) and soft coral (0.3 %). Turf algae, fleshy algae
tors (fish assemblage structure, fish richness, density and soft coral occurred almost in all sampling sites. In
and biomass) and the explanatory variables were fact, turf and fleshy algae were found at 100 % of the
analysed using distance-based linear models (DistLM; sites surveyed and accounted for approximately 36.9
Legendre and Anderson 1999; McArdle and Anderson and 23.9 % of the benthic cover of the habitat, respec-
2001) in the software PRIMER-E+. DistLM analysis tively. Soft corals were also common, identified at
was used to identify which of the potential examined 94.7 % of the sites and accounting for 25.1 % of the
variables explained most of the variability in fish assem- benthic cover.
blage. The Bbest^ selection method, according to the The first dbRDA accounted for 16.7 % of the total
Akaike Information Criterion (AIC), was used to select variation in the fish assemblage and distinguished sites
the final model. Moreover, a distance-based redundancy dominated by depth (shallow and interface sites)
analysis (dbRDA, Legendre and Anderson 1999; (Fig. 2a). The first dbRDA axis was positively correlat-
McArdle and Anderson 2001) was used to detect pat- ed with the percentages of turf algae (r=0.57), the
terns between the selected variables and fish assem- percentage of soft coral (r=0.49), the population of the
blage. This routine performs a constrained ordination nearest city (r=0.35) and the percentage of fleshy algae
of the fish assemblage data using the most parsimonious (r=0.23) and was negatively correlated with depth (r=–
DistLM model. To validate the interpretation of the 0.49). The second dbRDA axis was positively correlated
dbRDA, a distance-based principal coordinate analysis with depth (r=0.62) and distance from the coast (r=
(PCO) was used following Anderson et al. (2008). 0.68). The first axis clearly represented the gradient of
Species raw Pearson correlations with the first two depth on the rocky shores studied, with the right side
1650 Environ Biol Fish (2015) 98:1645–1657

Table 2 Results of the distance-based multivariate linear model (DistLM) for the fish assemblage, showing the percentage of variation
explained by significant environmental variables (P<0.001)

Axis Percentage of variation explained by individual axes

% Explained variation (fitted model) % Explained variation (total variation)

Individual Cumulative Individual Cumulative

Depth 42.62 42.62 16.73 16.73

DistCoast 35.79 78.41 14.05 30.78
Population of nearest city 9.46 87.87 3.71 34.49
% Turf algae 7.38 95.25 2.90 37.39
% Fleshy algae 4.00 99.25 1.57 38.96
% Soft coral 0.75 100.00 0.29 39.26

DistCoast Distance from the coast

characterised by shallow sites with increasing depth association with species richness (Fig. 3). Depth was
towards the left (Fig. 2a). Two groups were defined a strongly correlated with the first dbRDA axis (multiple
posteriori – shallow areas and deeper areas (near the partial correlation=–0.91). Physical structure index and
interface). As an additional way method of validation, soft coral were correlated with the first dbRDA axis
the dbRDA, a constrained ordination that uses the model (multiple partial correlations=–0.40 and –0.01, respec-
results, was compared to an unconstrained Principal tively) (Table 4). Density showed significant correla-
Coordinates Analysis plot (Fig. 2b). The two plots ap- tions with the physical structure index and distance from
proximate each other, indicating that our model is captur- the coast. This model explained 30.2 % of the variance.
ing the main patterns of variation (Anderson et al. 2008). The highest fish density was found in sites with the
To determine which environmental variable or com- highest values of the physical structure index (–0.546)
bination of environmental variables were represented by and the highest distance from the coast (–0.838). The
the dbRDA axes, raw Pearson correlations of each var- biomass was strongly inversely correlated with the dis-
iable with each dbRDA axis were examined (Table 3). tance from the coast (–1.0), with this model explaining
The omnivores with shoal distribution, Diplodus 27.4 % of the variance (Fig. 3 and Table 4).
argenteus and Abudefduf saxatilis, along with the terri-
torial herbivore Stegastes fuscus, were positively corre-
lated (Pearson correlations of>0.3) with axis 1, associ- Discussion
ated with the percentages of turf algae, fleshy algae and
soft coral. The groupers with heavy fishing pressure, Our data indicate that depth, distance from the coast,
Mycteroperca bonaci and Mycteroperca acutirostris, population of the nearest city and the dominant benthic
along with the invertivores Anisotremus virginicus and cover are the best predictors of rocky reef coastal fish
Serranus flaviventris, were negatively correlated with assemblage structure in the rocky shores of IGB.
axis 1, indicating that they are associated with deeper Moreover, the physical structure index and distance
areas. Axis 2, associated with depth, distance from the from the coast explained most of the variation in fish
coast, percentages of turf algae and soft coral and pop- richness, density and biomass. There was a positive
ulation of the nearest city, was positively correlated with relationship between the physical structure index, rich-
17 species, whereas the puffer fish Sphoeroides greeleyi ness and density, whereas distance from the coast was
showed the opposite pattern (Table 3). positively correlated with density and biomass.
The DistLM showed that 53.3 % of the variation in We found higher fish richness in the deeper sites than
the richness was attributed to the three following vari- in the shallow sites. Deep sites are located in areas
ables: physical structure index, depth and soft coral. between the consolidated substrate and the sandy bot-
Physical structure index and depth were negatively cor- tom, a transition zone that favours increases of species
related with the first dbRDA axis, indicating a positive richness where species pass through for foraging and
Environ Biol Fish (2015) 98:1645–1657 1651

structure and depth for Mediterranean rocky shores,

and Gibran and Moura (2012) found a similar correla-
tion across a gradient from nearshore areas to coastal
islands in Southeastern Brazil. In these studies, the
direct effects of depth are associated with pressure,
low temperature and lowlight intensity, as reported by
Srinivasan (2003).
Because microhabitat structure often changes with
depth (Done 1982), distributions of fishes may be ex-
plained by the availability of preferred microhabitats
(Srinivasan et al. 2010). However, in shallow rocky
shores of IGB, the influence of depth on the fish assem-
blages may be more closely related to its indirect effects
on increasing availability of substrate (by increasing the
surface area of the shore). Availability of rocky surface
is a critical factor for reef fishes, and reef sites with
higher depths present greater hard surface area and more
diverse structures in which marine life can settle.
Increases in habitat area often create a mosaic of micro-
habitats that are separated by barriers such as corridors
and patchily distributed resources. These factors play an
important role in altering competition, predation and
dispersion rates in different species (Kareiva 1990) and
thus have consequences on overall diversity. This high-
lights the importance of depth in structuring the distri-
Fig. 2 Ordination diagrams of the first two axes of the Distance- bution of the fish species in this study.
Based Redundancy Analysis (dbRDA) (a) and Principal Coordi- The highest fish densities were also found in sites
nates Analysis (PCO) (b), showing samples from the shallow with the highest values of the physical structure index.
(empty circles) and interface (filled circles) sites. Axes describe
the percentage of variation of the selected model in terms of the
This index was designed in a robust manner to compare
total variation of the fish assemblage. Vectors in dbRDA represent a range of physical complexities such as the number of
the environmental variables comprising the best selected models refuges of various size categories (holes and crevices),
(from Akaike’s information criterion). Codes for variables: the percentage of hard substratum, and the height and
DCoast Distance from the coast, Pop population nearest city, Turf
% turf algae, Fleshy % Fleshy algae, Soft % Soft coral
number of rocks of different sizes. Possible explanations
for the increased number of fishes in rugose areas in-
clude increased refuge from predators or increased pri-
camouflage in the sand. Moreover, these sites are suit- mary productivity on the hard surfaces that can support
able for predators such as serranids to search for prey. more fishes (Gratwicke and Speight 2005). Physical
By contrast, shallow sites are characterised by consoli- complexity has been widely used in coral reefs world-
dated substrate only, thus offering less habitat diversity wide as a good predictor of fish diversity (Chabanet
for fishes. Mendonça-Neto et al. (2008) also found that et al. 1997). Increasing the variety of refuge sizes would
species richness increases from shallow to interface enhance species richness by increasing the availability
areas in a tropical region along the southeastern coast of refuges to species with differing body sizes
of Brazil. The vertical distribution of reef fish assem- (Luckhurst and Luckhurst 1978), thereby reducing pre-
blages is highly correlated with the vertical distribution dation and competition, contributing to decrease en-
of benthic organisms, and it appears to be determined by counter rates between predators and preys, and increas-
factors such as habits and feeding behaviours, refuge ing resource availability (Almany 2004). Physical fea-
from predation and social interactions (Ferreira et al. tures such as rocks and holes of various sizes provide
2001). García-Charton and Pérez-Ruzafa (1998) also shelter and reproductive grounds for fish species
found a strong correlation between fish assemblage (Aburto-Opereza and Balart 2001). Therefore, the
1652 Environ Biol Fish (2015) 98:1645–1657

Table 3 The first two axes for

best discriminating fish species Species RDA1 RDA2 Trophic category
(>0.3) of the fish assemblage,
based on with heavy fishing Mycteroperca bonaci −0.457 – Piscivore*
pressure (Floeter et al. 2006)* and Anisotremus virginicus −0.530 – Mobile invertebrate feeder*
trophic category classification
Mycteroperca acutirostris −0.561 – Pscivore*
(Ferreira et al. 2001, 2004)
Serranus flaviventris −0.591 – Mobile invertebrate feeder
Diplodus argenteus 0.512 – Mobile invertebrate feeder
Abudfduf saxatilis 0.433 – Omnivore
Stegastes fuscus 0.390 0.377 Territorial herbivore
Coryphopterus glaucofraenum – 0.370 Mobile invertebrate feeder
Chromis multilineata – 0.394 Planktivores
Cantherhines pullus – 0.394 Omnivore
Epinephelus morio – 0.407 Carnivore*
Epinephelus marginatus – 0.445 Carnivore*
Haemulon aurolineatum – 0.448 Mobile invertebrate feeder
Sparisoma axillare – 0.474 Roving herbivore*
Stegastes pictus – 0.475 Territorial herbivore
Serranus baldwini – 0.502 Mobile invertebrate feeder
Elacatinus figaro – 0.512 Mobile invertebrate feeder
Halichoeres poeyi – 0.514 Mobile invertebrate feeder
Odontocion dentex – 0.539 Carnivore
Acanthurus chirurgus – 0.549 Roving herbivore
Canthigaster figueiredoi – 0.562 Sessile invertebrate feeder
Sparisoma frondosum – 0.619 Roving herbivore*
Pomacanthus paru – 0.626 Omnivore
Sphoeroides greeleyi – −0.442 Mobile invertebrate feeder

physical structure index is an important predictor for with distance from the coast belonged to the most impor-
fish assemblage structure and density in IGB coastal tant families of the herbivorous and carnivorous fishes.
reefs. Species such as parrotfishes (S. frondosum and
Densities of individual species were notably higher at S. axillare) and groupers (E. marginatus, E. morio, M.
greater distances from the coast, likely representing re- bonaci and M. acutirostris) suffer from heavy fishing
duced levels of disturbance and impacts from various pressure all along the Brazilian coast (Floeter et al.
sources on the mainland (e.g., sedimentation, sewage 2006). The association of these species with distance
pollution, and discharge from the rivers). Coastal areas from the coast indicates the importance of human action
at the sites close to the mainland have been degraded, and in populations of reef fishes. This trend is consistent with
those waters were very turbid compared to the offshore the expectation that harvesting pressure is lower in areas
areas. These findings are consistent with previous studies, away from the coast, due to the increasing difficulty of
which demonstrated that high sedimentation and low human access (Floeter et al. 2007). Distance from the
transparency indirectly influence fish distribution coast of the mainland (source of impacts from the main-
(Letourneur et al. 1998) or directly reduce coral diversity land) is also important in structuring fish communities
and coverage (Loya 1976). Benjamin et al. (2012), study- (Nguyen and Phan 2008).
ing environmental factors affecting coral reef fish in the The fish biomass was also correlated with the dis-
Mariana Archipelago, found that proximity to areas of tance from the coast. The low abundance of large pred-
high human population is likely to affect the distribution ators and herbivores found in this study indicates that
of large-bodied fishes that have a negative association fishing is an important factor structuring fish popula-
with population density. Species correlated positively tions on IGB. Historically, this bay was well populated
Environ Biol Fish (2015) 98:1645–1657 1653

Mycteroperca spp.) and snappers (Lutjanus spp.)

(Begossi et al. 2011). Fisheries appear to be affecting
the population size and size structure of fish populations
in studies on other coastal reefs in Brazil (Ferreira and
Gonçalves 1999; Frédou 2004; Gasparini et al. 2005;
Floeter et al. 2006). Stuart-Smith et al. 2008, studying
fishing impacts and the relationship with distance from
the nearest boat launching ramp, found low numbers of
large fish and a greater number of smaller fish at the sites
closest to access points.
Fish biomass is likely the most important factor in
accessing the status of reefs (Sabater and Tofaeono
2007; Francini-Filho and Moura 2008); thus, this vari-
able has direct implications for management. Reef sites
with the highest number of human communities are
those that have the poorest water quality, the lowest
coral cover and the lowest fish biomass (Dinsdale
et al. 2008). In the Brazilian National Park at the
Abrolhos Reef, a system with predominant coral cover,
significant spatial variability was recorded, with the
highest values for fish biomass occurring in protected
areas (Leão et al. 1988; Floeter et al. 2001; Francini-
Filho and Moura 2008; Bruce et al 2012). This confirms
that MPAs help maintain viable fisheries by augmenting
local fishery yields through biomass exportation from
the protected area (García-Charton et al. 2008).
Biological factors such as benthic organisms may
interfere in reef relief by increasing the structural com-
plexity provided by boring and arborescent organisms,
which, on a scale of centimetres, can provide mesoscale
habitat modification for small benthic fishes (Stephen
et al. 2006). In IGB, benthic cover the percentages of
Fig. 3 Scatterplot diagram comparing fish response variables turf algae, fleshy algae and soft coral) represented a
(richness, density and biomass) and the first Distance-based Re-
dundancy Analysis (dbRDA) axis
small (<5.0 %) influence on fish assemblage structure.
Although benthic cover on rocky shores is very impor-
tant in determining the habitat structure, when analysed
with larger groupers (pers. obs.), all of which were with other variables, it appears to have only limited
prized as food fish. Local resources exploited by the influence on fish assemblages. Previous studies on fish
artisanal fisheries in the region include rocky reef spe- community structure have shown high levels of variance
cies such as groupers (Epinephelus spp. and explained by benthic cover (Chabanet et al. 1997;

Table 4 Results of the distance-based multivariate linear model (DistLM) for the univariate variables richness, abundance and biomass.
AIC, Akaike’s information criterion; RSS, Residual sum of squared errors

Descriptors AIC R2 RSS N° Groups Selections

Richness 59.619 0.5333 147.82 3 Physical structure index, depth and soft coral
Density 254.42 0.3022 26,235 2 Physical structure index and distance from the coast
Biomass 380.43 0.2743 761,910 1 Distance from the coast
1654 Environ Biol Fish (2015) 98:1645–1657

Ferreira et al. 2001; Gratwicke and Speight 2005; responses to the biological, physical or anthropogenic
Floeter et al. 2007). However, Kuffner et al. (2010) variables tested in this study. Perhaps the lack of over-
found a weak relationship between benthic cover and sight prejudices the efficiency of these islands as
fish assemblage variables. The lack of relationships is protected area. Most areas of the Tamoios MPA are
consistent with several mechanisms, including stochas- subject to insufficient enforcement measures, and some
tic larval dispersal, priority effects of early colonisers, of them must even be considered merely as Bpaper^
and human disturbance (fishing) (Kuffner et al. 2010). reserves. As is often the case in Brazilian protected
Moreover, studies on benthic cover influences deter- areas, this MPA was created through a top-down ap-
mined Ba priori^ as a factor (e.g., levels of a given proach, by a governmental decree that delimits its
benthic cover; incrusting versus arborescent algae/ boundaries but does not resolve pending resource con-
coral; flat vs. complex coral) are more likely to have flicts (Begossi et al. 2011). Although nominally under
confounding results. Furthermore, potential generalisa- federal management and protection, as are most such
tions may be obscured due to a number of factors. These areas created by governmental fiat, its legitimacy and
include the enormous variation in the magnitude of credibility among artisanal fishers are weak (Begossi
changes to habitats observed, restrictive assumptions et al. 2010). Local conflicts include rules for the use of
about the forms of relationships between fish and ben- fishing areas established by artisanal fishers inside the
thic variables, inadequate descriptions of habitat and MPA and the advent of protected areas that close access
fish community structure, a heavy reliance on ‘natural’, to some fishing areas used by artisanal fisheries
pseudoreplicated experiments and a plethora of different (Begossi et al. 2011).
numerical and analytical techniques (Syms and Jones We assigned the characteristic of Bin^ or Bout^ for
2000). the sites within and outside the MPA together with
In this study, we observed high similarity in the other habitat structure and anthropogenic variables
benthic cover among the sites, which likely reduces in the model. This approach is important to avoid
the variability of this factor. The high frequency in the overestimate the effects of the MPA on fish assem-
sampled sites of turf, fleshy algae and soft coral were blage compared with studies that consider only the
likely associated with the weak relationship between site location within or outside the MPA. Protected
benthic cover and fish assemblage. However, the sites and non-protected areas differ in accessibility, ex-
with the highest percentage of soft coral showed the ploitation intensity, and fisheries activities, as well
lowest richness. Rocky shores dominated by soft coral as in the effectiveness of enforcement measures
such as P. caribaeorum may reduce the amount of (García-Charton et al. 2004; Claudet et al. 2010).
shelter available to fishes by overgrowing reef crevices Côté et al. (2001) compared data for 19 marine
and by reducing benthic diversity, consequently limiting reserves worldwide and concluded that the effect
food resources (Mendonça-Neto et al. 2008). Zoanthids on fish abundance is highly variable, based on fac-
do not generate complexity, and by covering huge por- tors such as intensity of exploitation outside
tions of the reef substratum, they can actually flatten the protected areas, variation in enforcement efficiency,
available complexity (Mueller and Haywick 1995; and habitat characteristics. Therefore, caution is re-
Haywick and Mueller 1997). The absence or low abun- quired when designing and developing programs to
dance of P. caribaeorum provides more space in those test the efficiency of MPAs, considering the poten-
sites for algal growth, which potentially benefits herbi- tial influences of anthropogenic activities and habi-
vores. In addition to soft coral, small fouling organisms tat structure.
that do not form large structures, including algae mostly In the case study of IGB, we found that the distance
composed of turf, were common in IGB. Although turf from the coast was the predictor that best explained the
algae does not form complex structures, it provides response variables, showing the importance of preserv-
strong effects on the distribution of macrofaunal organ- ing coastal areas that are losing species diversity. The
isms that live in the turf matrix and is very important for areas near the coast showed low levels of richness,
invertivore fish species as well as herbivores density and biomass, whereas areas farther from the
(Giangrande 1988; Sarda 1991; Kelaher et al. 2001). coast or more difficult to reach showed the opposite
The Ecological Station of Tamoios MPA, despite pattern. These areas should be constantly monitored to
having been in place for 24 years, has not shown avoid human disturbance. Moreover, re-organisation
Environ Biol Fish (2015) 98:1645–1657 1655

and expansion of the MPA should be an important evaluated by means of water quality, microbial diversity,
benthic cover, and fish biomass data. PLoS ONE 7:1–13
management practice to protect fish species, especially
Camilato V, Simon T, Pinheiro HT, Pimentel CR, Joyeux JC
large predators and herbivores. (2010) Length-weight relationships for some cryptobenthic
reef fishes of Guarapari, southeastern Brazil. J Appl Ichthyol
26:463–464
Acknowledgments We thank Carolina Corrêa and Daniel Carpenter KE, Midat RI, Albaladejo VD, Corpuz VT (1981) The
Gouvêa for logistical and field assistance, and Juan C. Ortiz for influence of substratum structure on the local abundance and
his help with data analyses. We also thank Estação Ecológica de diversity of Philippine reef fishes. 4th International Coral
Tamoios for research permits. Financial support was provided by Reef Symposium 2:495–502
the Brazilian Funding Agencies CNPq and CAPES; Teixeira- Chabanet P, Dufour V, Galzin R (1995) Disturbance impact on reef
Neves TP and Neves LM received doctoral scholarship from CNPq fish communities in Reunion Island (Indian Ocean). J Exp
and ‘sandwich’ scholarship from CAPES. Finally, we gratefully Mar Biol Ecol 188:29–48
acknowledge the Marine Spatial Ecology Lab (The University of Chabanet P, Ralambondrainy H, Amanieu M, Faure G, Galzin R
Queensland) for the opportunity to ‘sandwich’ internship. This (1997) Relationships between coral reef substrata and fish.
research was conducted under SISBIO Collection of Species Permit Coral Reefs 16:93–102
number 10707 issued by ICMBio, Brazilian Environmental Chapman MG, Underwood AJ, Skilleter GA (1995)
Agency. Variability at different spatial scales between a subtidal
assemblage exposed to the discharge of sewage and
two control assemblages. J Exp Mar Biol Ecol 189:
103–122
Chittaro P, Usseglio P, Sale P (2005) Variation in fish density,
References assemblage composition and relative rates of predation
among mangrove, sea grass and coral reef habitats. Environ
Biol Fish 72:175–187
Aburto-Opereza O, Balart EF (2001) Community structure of reef Claudet J, García-Charton JA, Lenfant P (2010) Combined effects
fish in several habitats of a rocky reef in the Gulf of of levels of protection and environmental variables at differ-
California. Mar Ecol 22:283–305 ent spatial resolutions on fish assemblages in a marine
Adjeroud M, Letourneur Y, Porcher M, Salvat B (1998) Factors protected area. Conserv Biol 25:105–114
influencing spatial distribution of fish communities on a Côté IM, Mosqueira I, Reynolds JD (2001) Effects of marine re-
fringing reef at Mauritius, SW Indian Ocean. Environ Biol serve characteristics on the protection of fish populations: a
Fish 53:169–182 meta-analysis. J Fish Biol 59:178–189
Almany GR (2004) Differential effects of habitat complexity, Dinsdale EA, Pantos O, Smriga S, Edwards RA, Angly F, Wegley
predators and competitors on abundance of juvenile and adult L, Hatay M, Hall D, Brown E, Haynes M, Krause L, Sala E,
coral reef fishes. Oecologia 141:105–113 Sandin SA, Thurber RV, Willis BL, Azam F, Knowlton N,
Anderson M, Gorley R, Clarke K (2008) PERMANOVA+ for Rohwer F (2008) Microbial ecology of four Coral Atolls in
PRIMER. PRIMER-ELtd, Plymouth the Northern Line Islands. PLoS ONE 3:1–17
Azzurro E, Matiddi M, Fanelli E, Guidetti P, La Mesa G, Scarpato Done TJ (1982) Patterns in the distribution of coral communities
A, Axiak V (2010) Sewage pollution impact on across the central Great Barrier Reef. Coral Reefs 1:95–107
Mediterranean rocky-reef fish assemblages. Mar Environ Ferreira CEL, Gonçalves JEA (1999) The unique Abrolhos reef
Res 69:390–397 formation (Brazil): need for specific management strategies.
Begossi A, Lopes PFM, Oliveira L, Nakano H (2010) Ecologia de Coral Reefs 18:352
Pescadores artesanais da Ilha Grande. Editora Rima, Sao Ferreira CEL, Gonçalves JEA, Coutinho R (2001) Community
Carlos structure of fishes and habitat complexity on a tropical rocky
Begossi A, May PH, Lopes PF, Oliveira LEC, Vinha V, Silvano shore. Environ Biol Fish 61:353–369
RAM (2011) Compensation for environmental services from Ferreira CEL, Floeter SR, Gasparini JL, Ferreira BP, Joyeux JC
artisanal fisheries in SE Brazil: policy and technical strate- (2004) Trophic structure patterns of Brazilian reef fishes: a
gies. Ecol Econ 71:25–32 latitudinal comparison. J Biogeogr 31:1093–1106
Benjamin LR, Ivor DW, Oliver JV, Gareth JW (2012) Floeter SR, Guimarães RZP, Rocha LA, Ferreira CEL, Rangel CA,
Environmental factors affecting large-bodied coral reef fish Gasparini JL (2001) Geographic variation in reef-fish assem-
assemblages in the Mariana Archipelago. PLoS ONE 7:1–25 blages along the Brazilian Coast. Glob Ecol Biogeogr 10:
Borja A, Dauer DM, Grémare A (2012) The importance of setting 423–431
targets and reference conditions in assessing marine ecosys- Floeter SR, Halpern BS, Ferreira CEL (2006) Effects of fishing
tem quality. Ecol Indic 12:1–7 and protection on Brazilian reef fishes. Biol Conserv 128:
Bouchon-Navarro Y, Bouchon C (1989) Correlations between 391–402
chaetodontid fishes and coral communities of the Gulf of Floeter SR, Krohling W, Gasparini JL, Ferreira CEL, Zalmon IR
Aqaba (Red Sea). Environ Biol Fish 25:47–60 (2007) Reef fish community structure on coastal islands of
Bruce T, Meirelles PM, Garcia G, Paranhos R, Rezende CE, southeastern Brazil: the influence of exposure and benthic
Moura RL, Francini Filho R, Coni EOC, Vasconcelos AT, cover. Environ Biol Fish 78:147–160
Amado-Filho GM, Hatay M, Schmieder R, Edwards R, Francini-Filho RB, Moura RL (2008) Dynamics of fish assem-
Dinsdale E, Thompson FL (2012) Abrolhos bank reef health blages on coral reefs subjected to different management
1656 Environ Biol Fish (2015) 98:1645–1657

regimes in the Abrolhos Bank, eastern Brazil. Aquat Conserv Islam MS, Tanaka M (2004) Impacts of pollution on coastal and
18:1166–1179 marine ecosystems including coastal and marine fisheries and
Frédou T (2004) The fishing activity on coral reefs and adjacent approach for management: a review and synthesis. Mar
ecosystems: a case study of the Northeast of Brazil. Cybium Pollut Bull 48:624–649
28:274 Jones GP, Ferrell DJ, Sale PF (1991) Fish predation and its impact
Friedlander AM, Parrish JD (1998) Habitat characteristics affect- on the invertebrates of coral reefs and adjacent sediments. In:
ing fish assemblages on a Hawaiian coral reef. J Exp Mar Sale PF (ed) The ecology of fishes on coral reefs. Academic,
Biol Ecol 224:1–30 New York, pp 156–178
García-Charton JA, Pérez-Ruzafa A (1998) Correlation between Kareiva P (1990) Population dynamics in spatially complex environ-
habitat structure and a rocky reef fish assemblage in the ments: theory and data. Philos Trans R Soc Lond 330:175–190
southwestern Mediterranean. PSZN I: Mar Ecol 19:111–28 Kelaher BP, Chapman MG, Underwood AJ (2001) Spatial patterns
García-Charton JA, Pérez-Ruzafa A, Sánchez-Jerez P, Bayle- of diverse macrofaunal assemblages in coralline turf and their
Sempere JT, Reñones O, Moreno D (2004) Multi-scale spa- associations with environmental variables. J Mar Biol Assoc
tial heterogeneity, habitat structure, and the effect of marine UK 81:917–930
reserves on Western Mediterranean rocky reef fish assem- Kohler KE, Gill SM (2006) Coral Point Count with Excel exten-
blages. Mar Biol 144:161–182 sions (CPCe): a visual basic program for the determination of
García-Charton JA, Pérez-Ruzafa A, Marcos C, Claudet J, coral and substrate coverage using random point count meth-
Badalamenti F, Benedetti-Cecchi L, Falcón JM, Milazzo M, odology. Comput Geosci 32:1259–1269
Schembri PJ, Stobart B, Vandeperre F, Brito A, Chemello R, Krajewski JP, Floeter SR (2011) Reef fish community structure of
Dimech M, Domenici P, Guala I, Le Diréach L, Maggi E, the Fernando de Noronha Archipelago (Equatorial Western
Planes S (2008) Effectiveness of European Atlanto- Atlantic): the influence of exposure and benthic composition.
Mediterranean MPAs: do they accomplish the expected ef- Environ Biol Fish 92:25–40
fects on populations, communities and ecosystems? J Nat Kuffner IB, Grober-Dunsmore R, Brock JC, Hickey TD (2010)
Conserv 16:193–221 Biological community structure on patch reefs in Biscayne
Gasparini JL, Floeter SR, Ferreira CEL, Sazima I (2005) Marine National Park, FL, USA. Environ Monit Assess 164:513–531
ornamental trade in Brazil. Biodivers Conserv 14:2883–2899 Lara EN, Gonzalez EA (1998) The relationship between reef fish
Giangrande A (1988) Polychaete zonation and its relation to algal community structure and environmental variables in the
distribution down a vertical cliff in the western southern Mexican Caribbean. J Fish Biol 53:209–221
Mediterranean (Italy): a structural analysis. J Exp Mar Biol Leão ZMAN, Araújo TMF, Nolasco MC (1988) The coral reefs
Ecol 120:263–276 off the coast of eastern Brazil. Proceedings of the 6th
Gibran FZ, Moura RL (2012) The structure of rocky reef fish International Coral Reef Symposium 1:339–347
assemblages across a nearshore to coastal islands gradient Leathwick JR, Elith J, Hastie T (2006) Comparative performance
in Southeastern Brazil. Neotropical Ichthyol 10:369–382 of generalised additive models and multivariate adaptive
Gratwicke B, Speight MR (2005) The relationship between fish regression splines for statistical modelling of species distri-
species richness, abundance and habitat complexity in a range butions. Ecol Model 199:188–196
of shallow tropical marine habitats. J Fish Biol 66:650–667 Lecchini D, Adjeroud M, Pratchett MS, Cadoret L (2003) Spatial
Green DG, Bradbury RH, Reichelt RE (1987) Patterns of predict- structure of coral reef fish communities in the Ryukyu
ability in coral reef community structure. Coral Reefs 6:27– Islands, southern Japan. Oceanol Acta 26:537–547
34 Legendre P, Anderson MJ (1999) Distance-based redundancy
Grigg RW (1994) Effects of sewage discharge, fishing pressure analysis: testing multispecies responses in multifactorial eco-
and habitat complexity on coral ecosystems and reef fishes in logical experiments. Ecol Monogr 69:1–24
Hawaii. Mar Ecol Prog Ser 103:25–34 Lester SE, Halpern BS, Grorud-Colvert K, Lubchenco J,
Grober-Dunsmore R, Frazer TK, Beets JP, Lindberg WJ, Zwick P, Ruttenberg BI, Gaines SD, Airamé S, Warner RR (2009)
Funicelli NA (2008) Influence of landscape structure on reef Biological effects within no-take marine reserves: a global
fish assemblages. Landsc Ecol 23:37–53 synthesis. Mar Ecol Prog Ser 384:33–46
Guerra JV, Soares FLM (2009) Circulation and flux of suspended Letourneur Y (1996) Dynamics of fish communities on reunion
particulate matter in Ilha Grande Bay, SE Brazil. J Coastal fringing reefs, Indian Ocean: 1. Patterns of spatial distribu-
Res 56:1350–1354 tion. J Exp Mar Biol Ecol 195:1–30
Guidetti P, Fanelli G, Frashetti S, Terlizzi A, Boero F (2002) Letourneur Y, Kulbicki M, Labrosse P (1998) Spatial structure of
Coastal fish indicate human-induced changes in the commercial reef fish communities along a terrestrial runoff
Mediterranean littoral. Mar Environ Res 53:77–94 gradient in the northern lagoon of New Caledonia. Environ
Halpern BS, Lester SE, McLeod KL (2010) Placing marine Biol Fish 51:141–159
protected areas onto the ecosystem-based management sea- Lewis AR (1997) Effects of experimental coral disturbance on the
scape. Proc Natl Acad Sci U S A 107:18312–18317 structure of fish communities on large patch reefs. Mar Ecol
Haywick DW, Mueller EM (1997) Sediment retention in Prog Ser 161:37–50
encrusting Palythoa spp. a biological twist to a geological Loya Y (1976) Effects of water turbidity and sedimentation on the
process. Coral Reefs 16:39–46 community structure of Puerto Rican corals. Bull Mar Sci 26:
Ignacio BL, Julio LM, Junqueira AOR, Ferreira-Silva MAG 450–466
(2010) Bioinvasion in a Brazilian Bay: filling gaps in the Luckhurst BE, Luckhurst K (1978) Analysis of the influence of
knowledge of southwestern Atlantic Biota. PLoS ONE substrate variables on coral reef fish communities. Mar Biol
5:1–9 49:317–323
Environ Biol Fish (2015) 98:1645–1657 1657

Macieira RM, Joyeux JC (2008) Length–weight relationships for Islands (western Mediterranean). PSZNI: Mar Ecol 12:341–
rockpool fishes in Brazil. J Appl Ichthyol 25:358–359 360
McArdle BH, Anderson MJ (2001) Fitting multivariate models to Sleeman JC, Kendrick GA, Boggs GS, Hegge BJ (2005)
community data: a comment on distance-based redundancy Measuring fragmentation of seagrass landscapes: which in-
analysis. Ecology 82:290–297 dices are most appropriate for detecting change? Mar Freshw
McClanahan T, Karnauskas M (2011) Relationships between ben- Res 56:851–864
thic cover, current strength, herbivory, and a fisheries closure Smith AK, Ajani PA, Robert DE (1999) Spatial and temporal
in Glovers Reef Atoll, Belize. Coral Reefs 30:9–19 variation in fish assemblages exposed to sewage and impli-
McGehee MA (1994) Correspondence between assemblages of cations for management. Mar Environ Res 47:241–260
coral-reef fishes and gradients of water motion, depth, and Spatharis S, Tsirtsis G (2010) Ecological quality scales based on
substrate size off Puerto-Rico. Mar Ecol Prog Ser 105:243– phytoplankton for the implementation of water framework
255 directive in the Eastern Mediterranean. Ecol Indic 10:840–
Mcmanus JW, Miclat RI, Palaganas VP (1981) Coral and fish reef 847
community structure of Sombrero Island, Batangas, Srinivasan M (2003) Depth distributions of coral reef fishes: the
Philippines. In: Proceedings of the Forth International Coral influence of microhabitat structure, settlement, and post-
Reef Symposium, Manila, p 271–280 settlement processes. Oecologia 137:76–84
Mendonça-Neto JP, Ferreira CEL, Chaves LCT, Pereira RC (2008) Srinivasan UT, Cheung WWL, Watson R, Sumaila UR (2010)
Influence of Palythoa caribaeorum (Anthozoa, Cnidaria) Food security implications of global marine catch losses due
zonation on site-attached reef fishes. An Acad Bras Cienc to overfishing. J Bioecon 12:183–200
80:495–513 Stephen JD, Larson RJ, Pondella DJ (2006) Rocky reefs and kelp
Moore CH, Harvey ES, Van Niel K (2010) The application of beds. In: Allen LG, Pondella DJ, Horn MH (eds) The ecology
predicted habitat models to investigate the spatial ecology of of marine fishes: California and adjacent waters. University
demersal fish assemblages. Mar Biol 157:2717–2729 of California Press, Berkeley, pp 227–252
Mueller E, Haywick DW (1995) Sediment assimilation and calci- Syms C, Jones GP (2000) Disturbance, habitat structure, and the
fication by the Western Atlantic reef zoanthid, Palythoa dynamics of a coral-reef fish community. Ecology 81:2714–
caribaeorum. Bull Inst Océanogr 14:89–100 2729
Munday PL (2002) Does variability determine geographical-scale Stuart-Smith RD, Barrett NS, Crawford CM, Frusher SD,
abundances of coral-dwelling fishes? Coral Reefs 21:105– Stevenson DG, Edgar GJ (2008) Spatial patterns in impacts
116 of fishing on temperate rocky reefs: are fish abundance and
Murdoch WW, Oaten A (1975) Predation and population stability. mean size related to proximity to fisher access points? J Exp
Adv Ecol Res 9:125–132 Mar Biol Ecol 365:116–125
Nguyen LV Phan HK (2008) Distribution and factors influencing Teixeira TP, Neves LM, Araújo FG (2012) Thermal impact of a
on structure of reef fish communities in Nha Trang Bay nuclear power plant in a coastal area in Southeastern Brazil:
Marine Protected Area, South-Central Vietnam. Environ effects of heating and physical structure on benthic cover and
Biol Fish 82:309–324 fish communities. Hydrobiologia 84:161–175
Ohman MC, Rajassuriya A (1998) Relationships between habitat Thrush SF, Chiantore M, Asnaghi V, Hewitt J, Fiorentino D,
complexity and fish communities on coral and sandstone Cattaneo-Vietti R (2011) Habitat–diversity relationships in
reefs. Environ Biol Fish 55:19–31 rocky shore algal turf infaunal communities. Mar Ecol Prog
Otway NM, Gray CA, Craig JR, McVea TA, Ling JE (1996) Ser 424:119–132
Assessing the impacts of deep-water sewage outfall on spa- Vincent V, Hincksman CM, Tibbetts IR, Harris A (2011) Biomass
tially and temporally variable marine communities. Mar and abundance of herbivorous fishes on coral reefs off
Environ Res 41:45–71 Andavadoaka, Western Madagascar. W Indian Ocean J Mar
Roberts CM, Ormond RFG (1987) Habitat complexity and coral Sci 10:83–99
reef fish diversity and abundance on Red Sea fringing reefs. Wantiez L, Chauvet C (2003) First data on community structure
Mar Ecol Prog Ser 41:1–8 and trophic networks of uvea coral reef fish assemblages
Russo AR (1982) Temporal changes in fish community structure (Wallis and Futuna, South Pacific Ocean). Cybium 27:83–
near a sewage ocean outfall, Mokapu, Oahu, Hawaii. Mar 100
Environ Res 6:83–98 Williams DMB (1982) Patterns in the distribution of fish commu-
Sabater MG, Tofaeono SP (2007) Scale and benthic composition nities across the central Great Barrier Reef. Coral Reefs 1:35–
effects on biomass and trophic group distribution of reef 43
fishes in American Samoa. Pac Sci 61:503–520 Williams DMCB (1991) Patterns and processes in the distribution
Sarda R (1991) Polychaete communities related to plant covering of coral reef fishes. In: Sale PF (ed) The ecology of fishes on
in the mediolittoral and infralittoral zones of the Balearic coral reefs. Academic, San Diego, pp 437–474