Introduction

The use of artificial structures for habitat restoration has increased due to their promising potential and successful initiatives in rehabilitating fish stocks (Caddy 1999; Fabi et al. 2011; Feary et al. 2011; Simon 2011). In addition, artificial reefs may also be used as scientific research tools because they are easily manipulated, allow for the control of structure variables (e.g., size and complexity) and can be arranged in a variety of spatial distributions (Jan et al. 2003; Jordan et al. 2005). Therefore, artificial reefs facilitate the validation of ecological theories and development of theoretical and predictive models, many of which represent the foundation of practical programs and can be applied to reefs in marine environments at a large scale (Polovina 1991; Fabi et al. 2011).

Artificial habitats are constructed from a variety of materials, such as concrete, rubber and oil platforms (Baine 2001; Fabi et al. 2011). Concrete is generally selected because of its durability and ability to form several types of reef structures (e.g., cubes, pyramids, cones and reefballs), areas and volumes (Baine 2001; Jan et al. 2003; Fabi et al. 2011; Hackradt et al. 2011; Gatts et al. 2014). The complexity of the structures can influence the associated fish communities, with holes and crevices providing shelter, breeding habitat and refuge against predators (Anderson et al. 1989; Bohnsack 1989; Hixon and Beets 1989; Seaman 1996; Charbonnel et al. 2002; Sherman et al. 2002; Brotto and Zalmon 2007; Hackradt et al. 2011).

The replication and spatial arrangement of reefs might alter the complexity of the environment and structure of the associated fish communities (Lindberg 1996; Jordan et al. 2005). Experimental reef designs with different sizes, numbers, volumes, areas and modular distances have been investigated as tools in the management and conservation of coastal resources because they can attract/aggregate target species (Schroeder 1987; Frazer and Lindberg 1994; Baine 2001; Charbonnel et al. 2002; Baine and Side 2003; Jordan et al. 2005; Simon 2011).

Organisms in sandy substrates adjacent to reefs are important in the diet of associated piscivorous and/or invertivorous fish, suggesting that reef communities are dependent on the biological production of the associated sediment (Lindquist et al. 1994; Relini et al. 2002; Leitão et al. 2007). The optimal foraging theory suggests that when less energy is expended during foraging, the energy gain is greater, and the risk of an organism being preyed upon is lower; therefore, fishes are expected to feed on preys closer to the reef (MacArthur and Pianka 1966; Milinski 1986; Stephens and Krebs 1986). This behavior of ichthyofauna can lead to a halo-like distribution of preys surrounding reef units, with an increase in the prey availability with increasing distance from these structures (Randall 1965; Ogden 1976; Bortone et al. 1998; Campbell et al. 2011). Lindberg (1996) and Jordan et al. (2005) observed increasing fish abundances with distance isolation (up to 25 m). Bortone et al. (1998), Campbell et al. (2011) and Machado et al. (2013) observed a decrease in prey density closer to artificial reefs.

On the other hand, natural and artificial reefs of smaller size or number that have a patchy distribution and complexity (i.e., spaced reef modules) may exhibit greater fish abundance and richness than larger and isolated reefs on sandy bottoms (Nanami and Nishihira 2002; Jan et al. 2003; Morton and Shima 2013; Yamamoto et al. 2014). Because of these contrasting scenarios, studies focused on these topics can provide information on how artificial structures affect the composition and structure of associated fish communities and can be used in marine resources management and conservation projects (Santos et al. 2010, 2011; Gatts et al. 2014). Therefore, this study evaluated the influence of the spatial distribution of artificial reef modules along the northern coast of Rio de Janeiro State on the composition and structure of the associated ichthyofauna. The hypothesis tested was whether an increase in distance from 0.5 to 15 m among artificial reef units will reduce the overlap of trophic halos, promoting a greater availability of prey and a higher richness, diversity, number of individuals and biomass of the ichthyofauna compared with modules that are located more closely together. Thus, the aim of this study was to suggest the best spatial configuration for patchy artificial reefs in coastal environments with homogeneous and less complex substrates that are seasonally influenced by the inflow of freshwater and sediment from large rivers, which represents the typical environmental conditions of the northern coast of Rio de Janeiro.

Methodology

Study site

The study site (21°29′S, 41°00′W) was located on the continental shelf at the north of Rio de Janeiro State (Southeastern Brazil), and it is adjacent to the mouths of the Itabapoana and Paraíba do Sul rivers (Fig. 1). The northern coast of Rio de Janeiro is naturally depleted of rock substratum or other hard substrates, and it is covered by extensive sandy beaches with variable amounts of mud and calcareous nodules, such as rhodolites (Zalmon et al. 2002). Pluviometric precipitation in the Paraíba do Sul River drainage basin exhibits two distinct periods: a dry season from May to September and rainy season from October to April (Carvalho et al. 2002; see Gatts et al. 2014 for the monthly mean flow of the Paraíba do Sul River since 1995). Data on the average monthly flow of the Paraiba do Sul River in the region were obtained from the Laboratory of Environmental Science, University of North of Rio de Janeiro State.

Fig. 1
figure 1

Installation site and schematic drawing configuration of the artificial reefs (AR) along the northern coast of the Rio de Janeiro State

Full size image

Experimental design

The study was performed using concrete reefball modules (~1 m3). These modules contain orifices of 20 cm diameter and have an approximate weight of 500 kg. In studies evaluating the spatial configurations of artificial reefs, Santos et al. (2010) suggested that the distance between reef modules of the same treatment should not exceed 50 m because beyond this distance, the abundance and species richness associated with the reef declines. The reef complex covers an area of 50,000 m2, consisting of modular units and reef configurations randomly separated by >50 m to reinforce the independence among them. Moreover, the design was intended to allow for evaluations of spatial scale issues related to habitats of commercial interest to local artisanal fisheries. Therefore, to test how the degree of isolation of the artificial reefs affected the ichthyofauna, the modules were arranged at 9 m depth in equilateral triangular configurations of different distances (0.5, 5 and 15 m) (Fig. 1). This small-scale spatial design was defined based on the artificial reef spatial design investigated on FL, USA by Jordan et al. (2005). The precise location of the modules in the planned configurations was obtained using a geographic positioning system (GPS).

Sampling

The artificial reefs were assessed in four sampling periods according to the local seasonality (end of the dry season, September 2009 and October 2010, and end of the rainy season, April 2010 and April 2011) to reduce the influence of the transitional periods between seasons on the samples. The ichthyofauna were collected with gillnets (N = 24) of 25 m length and 3 m height with 30 mm of mesh between adjacent nodes. The nets were submerged for 24 h to capture diurnal and nocturnal species. The specimens were frozen on ice and taken to the laboratory for further analysis. The exclusive use of gillnets is justified because this fishing gear is the same as that used by artisanal fishermen in the region and the strong fluvial influence (Itabapoana River and, mainly, Paraíba do Sul River) restricts the subaquatic visibility (<0.5 m; Godoy et al. 2002) and prevents employing the visual census method. There was one gillnet per triangular artificial reef configuration per sampling period resulting in 12 gillnets for each period/reef configuration; see Gatts et al. (2014) for the arrangement of gillnets above the reef configurations.

Samples of the first 15 cm of sediment 1 m away from the units for each experimental distance (N = 3 replicates per treatment) were collected by divers using a stainless steel corer (0.018 m2) during the same ichthyofauna-sampling periods to identify potential fish predators and their respective reef-associated benthic prey. Infauna were fixed in 10 % formaldehyde, sieved through a 500-μm mesh, sorted and identified to the lowest taxonomic level. The macrofauna identification followed the methods by Rios (1994), Amaral and Nonato (1996) and Melo (1996).

Data treatment and analysis

Fish were identified, weighed and separated into groups, and their stomachs were removed. According to Morton and Shima (2013), numerical descriptors per standardized distance unit can reflect different responses of the fish assemblages associated with the reefs. Therefore, absolute and relative [i.e., standardized per distance unit (m)] values of species richness, total number of individuals (n) and biomass (g) were evaluated to detect variations in the structure of the ichthyofauna associated with each experimental distance (0.5, 5, and 15 m) and season (dry and rainy). Absolute values of Shannon’s diversity (H′), Simpson’s dominance (D), fish size (TL mm) and importance percentage (IP) index, which considers the occurrence frequency, number of individuals and percentage of biomass of each species (Zar 1999), were also assessed. Generalized linear models (GLM) were applied to test for differences in these descriptors [log10 (x + 1)-transformed data, except for diversity and IP indexes] among different reef distance arrangement and sampling periods. GLM were computed with the statistical package SPSS 15 (SPSS 2006). The effects of reef distance on species composition were assessed with canonical correspondence analysis (CCA). This technique constrains the ordination of species composition to be a linear function of the environmental variables, thus using the environmental and fish composition matrices simultaneously in a single analysis (Lepš and Šmilauer 2003). Canonical correspondence analysis was performed with CANOCO 4.5 (Lepš and Šmilauer 2003), down-weighting rare species. Seasonal changes in the composition and structure of the ichthyofauna associated with the different reef distances were assessed with a non-metric multidimensional scaling ordination analysis, using Bray–Curtis distance as a measure of similarity (Clarke and Warwick 2001).

The effect of fish predation on the macroinvertebrate community was analyzed through the taxonomic composition of food items in the stomachs of the fish captured on each sampling distance (0.5, 5 and 15 m), period (dry and rainy) and year (1 and 2). The index of relative importance (IRI) was calculated for the main prey categories using the following equation: IRI i = (%W i  + %N i ) × %F i , where I is the prey item, %W is the percentage of weight, %N is the percentage of individuals and %F is the frequency of occurrence of prey type in ichthyofauna stomach contents (Zar 1999). Fish predation was evaluated by Pearson’s correlation coefficient between the number of captured fish at the 0.5, 5 and 15 m reef sets and total number of infauna at each distance, and between the ichthyofauna food items and the reef-associated infauna at each respective distance. Kolmogorov–Smirnoff tests were performed before the correlations to test for data normality (Zar 1999).

Results

Composition and structure of the ichthyofauna

A total of 24 species from 12 families were recorded during the four sampling campaigns, and the following five species accounted for 56.1 % of the associated ichthyofauna (Table 1) : Cynoscion virescens (16.3 %), Aspistor luniscutis (12.6 %), Paralonchurus brasiliensis (8.2 %), Larimus breviceps (8.1 %), Isopisthus parvipinnis (5.8 %) and Cynoscion jamaicensis (5.1 %). Fish species composition was similar among the distances tested, with two, three and four taxa exclusive to the 0.5-, 5-, and 15-m distance treatments, respectively (Table 1). C. virescens was the predominant species at the three distances (IP: 0.5 m = 13.5 %, 5 m = 18.9 % and 15 m = 16.9 %), followed by I. parvipinnis (12.6 %) and A. luniscutis (12.0 %) at 0.5 m; A. luniscutis (17.7 %) at 5 m; and L. breviceps (13.7 %) and P. brasiliensis (12.6 %) at 15 m (Table 1).

Table 1 Composition and importance percentage (IP) of the ichthyc species captured at each reef distance (0.5, 5 and 15 m) and in total
Full size table

The relationship between species composition and reef distance was summarized by the two first CCA axes, which explained 4.0 % (eigenvalue = 0.26) and 3.1 % (eigenvalue = 0.04) of the variation (total inertia = 6.1), respectively. The CCA biplot showed that the three reef treatments shared similar fish assemblages, as most species (N = 15; 62.5 %) were distributed near the central portion of the diagram (Fig. 2). Still, some species’ associations occurred exclusively in a specific treatment, such as C. jamaicensis and Menticirrhus americanus with reef modules spaced 0.5 m, Pellona harroweri, Sphyraena sp. and Pomatomus saltatrix with modules at 5 m apart and Odontognathus mucronatus, Centropomus parallelus, Haemulon steindachneri and Anisotremus surinamensis with reefs distanced 15 m apart. These results revealed a small increase in the number of exclusive species (i.e., two to four) and habitat-dependent fish (i.e., none to three) with increased distances among reef modules (i.e., 0.5–15 m).

Fig. 2
figure 2

Canonical correspondence analysis of fish composition (abundance) with reef distance treatment (0.5, 5 and 15 m). The species–environment correlations for the two axes were 0.57 and 0.43, respectively

Full size image

Seasonal effects of reef usage patterns

Species richness, total number of individuals, biomass of fish, size, Shannon’s diversity (H′) and dominance did not significantly differ among the three reef distances (0.5, 5 and 15 m) and seasons (dry and rainy) (Table 2).

Table 2 Generalized linear models (GLM) results for richness, abundance, biomass, Shannon’s diversity (H′), Simpson’s dominance (D) and size (TL mm) of the ichthyofauna captured at each reef distance (0.5, 5 and 15 m) and sampling period (dry and rainy)
Full size table

The non-metric multidimensional scaling indicated that there was a greater association between the fish community and sampling season than reef (Fig. 3).

Fig. 3
figure 3

Non-metric multidimensional scaling of the ichthyic community in each reef distance (0.5, 5 and 15 m) and sampling season (dry and rainy)

Full size image

Temporal variations in the species association pattern were observed at all distances, which revealed a seasonal distribution of the main species in the region (Fig. 4a–d). A. luniscutis was caught exclusively during the rainy season, whereas C. microlepidotus and P. brasiliensis were observed only in the dry season (Fig. 4).

Fig. 4
figure 4

Relative distribution of the number of individuals (%) of the most representative species (importance percentage index >5 %) on the dry and rainy seasons: a all reef distances; b 0.5 m; c 5 m; d 15 m

Full size image

Predation

A total of 13 fish species that were captured at the artificial modules contained small fish and macrofaunal specimens in their stomachs, with the former showing the highest IRI values in the fish diets followed by crustaceans, mollusks and polychaetes (Table 3). The correlation of reef-associated infauna at the different distances with the fish stomach contents was not significant (r < 0.2; P > 0.05). Similarly, no significant relationship was found between the abundance of benthic species and fish captured near the modules at reef distances of 0.5, 5 and 15 m (P > 0.05).

Table 3 Index of relative importance (IRI) for the main prey categories of fish at distances of 0.5, 5 and 15 m per sampling period (dry 1: September 2009; dry 2: September 2010; rainy 1: April 2010; rainy 2: April 2011) (N = 3 replicates per treatment)
Full size table

Discussion

Limitations in the use of gillnets must be considered because the size of the mesh restricts the sampling of individuals according to their length (Acosta and Appeldoorn 1995; Acosta 1997). These nets select fishes with different habits, including transient, pelagic or demersal, from the unconsolidated substrate adjacent to reefs (Fabi and Fiorentini 1994; Brotto and Zalmon 2007). Gillnets have been traditionally used in the State of Rio de Janeiro by local artisanal fishermen (Garcez 2007), and strong fluvial influences (Itabapoana and Paraíba do Sul rivers) also explain the use of gillnets in scientific studies (Zalmon et al. 2002; Brotto and Zalmon 2007; Santos et al. 2010; Franco 2013).

The associated ichthyofauna did not clearly change with experimental patchy reef distances (0.5, 5 and 15 m). Our findings did not thus agree with the general hypothesis of increasing richness and abundance with distance of reef modules (Nanami and Nishihira 2002; Jan et al. 2003; Morton and Shima 2013; Yamamoto et al. 2014). It might be possible that differences among reef distance treatments were hidden by the non-mutually effects of the Paraíba do Sul and Itabapoana rivers on fish species composition and structure, especially during rainy season with higher outflows, and the relative low number of our treatment replicates. Therefore, further studies performed during the prevalence of clear and more saline waters, and with a greater sampling replication or effort or replication would be of great value to validate our findings for the fish assemblages associated with patchy artificial reefs on the northern coast of Rio de Janeiro.

Several authors (Randall 1963; Lindberg 1996; Nelson and Bortone 1996; Bortone et al. 1998; Jordan et al. 2005; Campbell et al. 2011) have stated that the greater availability of prey (fish and invertebrates) on reefs that are farther apart (25 m) compared with those closer (0.33 m) is a result of the lower overlap of trophic halos, which increased the richness of species and abundance of fishes in communities at greater distance. According to Randall (1963), the prey density is lower near the reef because benthivorous fishes forage in areas closer to the refuge to reduce the risk of being predated upon by transient fishes. Although certain reef predators directly reduce the infauna abundance near a reef through direct predation, transient predators increase such abundance by indirectly feeding on larger organisms, such as crab and fish (Lindberg 1996). The infauna in the sediment adjacent to the artificial reef at all tested distances is composed of annelids, mollusks and crustaceans. The presence of fish populations near the structures is not directly related to these organisms (Zalmon et al. 2014) but to larger benthic prey, such as decapod crustaceans, bivalve mollusks and small fishes, which likely use the interstices of the artificial structures; this result was confirmed by the diet analysis.

Morton and Shima (2013) argued that increased reef spacing (5 m) decreased fish recruitment between structures, whereas closer treatments (1 m) essentially concentrate larger number of fishes. In the present study, smaller individuals (<30 cm) of the most representative species, C. virescens, occurred on the reefs spaced 0.5 m between modules, and larger individuals were captured at the 5- and 15-m distance treatments, indicating that proximal structures concentrated recruits compared with reefs spaced at greater distances.

In FL, Jordan et al. (2005) observed that the richness and abundance of fish were highest in patchy reefs spaced 0.33 m apart. Along the northern coast of Rio de Janeiro State, artificial structures with a patchy configuration at distances of 0.5 m apart may concentrate more fish compared with larger distances because such configurations have a greater complexity related to the availability of refuges and crevices between the closer modules, which have the potential to concentrate a greater abundance of resident prey, including epifauna and piscivorous predators (Krohling et al. 2006). These species can also use the crevices as refuge from predation related to transient fishes attracted to the site (Major 1978; Hixon and Beets 1989, 1993; Jordan et al. 2005).

The species belonging to the families Sciaenidae, Carangidae and Ariidae were most abundant in the reefs spaced more closely together, and they were classified by Santos et al. (2010) as piscivores (C. microlepidotus, C. jamaicensis, C. virescens, I. parvipinnis and L. breviceps) and invertivores (M. americanus, P. brasiliensis, S. rastrifer, C. chrysurus, A. luniscutis, B. marinus and G. genidens); these species seem to be most likely attracted by smaller fish concentrated in the crevices and surroundings of the more clustered structures as well as by the associated invertebrates (Hixon and Beets 1989; Jordan et al. 2005; Brotto and Zalmon 2007).

The presence of reef modules in a patchy configuration creates a more complex environment by offering a larger quantity of shelter (Brotto et al. 2006) in addition to concentrating a larger density of potential prey both on the reef itself and in the surrounding area (Krohling et al. 2006). Artificial reefs tend to attract the adjacent substrate organisms that are important in the diet of piscivorous and invertivorous fish, suggesting that transient shoals of opportunistic fishes are directly affected by the biological productivity of the associated sediments (Lindquist et al. 1994; Relini et al. 2002; Zalmon et al. 2002; Leitão et al. 2007). Optimal foraging theory suggests that when less energy is expended during foraging, the predation risk is reduced because the organism remains exposed for less time (MacArthur and Pianka 1966; Milinski 1986; Stephens and Krebs 1986; Bortone et al. 1998). Transient, small-sized opportunistic fishes, such as P. harroweri and O. oglinum, or juveniles of species whose adults generally reach greater sizes, such as C. jamaicensis and A. luniscutis, likely feed on prey that is closer to the reef.

The higher densities and species richness of the infauna and fishes assemblages at the tested reefs with the shortest distance suggest that the haloes of large-bodied infauna (>5 mm: 500 μm mesh) around reefs are not consistent, particularly for reefs located on open coast sediment that is influenced by seasonal inflows of freshwater and sediments from large rivers (Zalmon et al. 2014). Herrera et al. (2002) emphasized that predation pressure is less evident at sites that do not have resident predator species, which is a consistent with characteristics of many species of the studied areas (Santos et al. 2010, 2011).

The ichthyofauna association pattern observed in the present study was related to sampling period, highlighting the potential effect of the Paraíba do Sul and Itabapoana rivers. This seasonal influence, combined with the differentiation in the temporal distribution pattern of the main species in the region, may be masking potential differences among the reef distance treatments. As noted by Brotto and Zalmon (2007), adverse environmental conditions (strong bottom currents, turbid waters and the presence of polyhaline plumes) are likely the key factors affecting the colonization patterns of fishes in artificial reefs along the northern coast of Rio de Janeiro State (Zalmon et al. 2002; Krohling and Zalmon 2008; Santos et al. 2010).

In general, the main species of Sciaenidae captured on artificial reefs in the northern coast of Rio de Janeiro are coastal species that live in proximity to the mouths of major rivers and are predominant in the region throughout the year (Gomes et al. 2003; Fulgêncio 2004; Souza and Chaves 2007; Militelli et al. 2013), with adults and juveniles using shallow and estuarine areas for growth and feeding habitat (Menezes and Figueiredo 1980; Godefroid et al. 2004). These species are found in beach environments during periods of reproduction and recruitment, which occurs from spring to autumn (Godefroid et al. 2004). Thus, the migratory behavior of Sciaenidae species could explain their dominance in artificial reefs during the dry season.

In summary, increasing distance between artificial patchy reefs in the range of 0.5–15 m led to few or none effects on fish richness, abundance, biomass and diversity of the associated ichthyofauna. As the shorter reef distance exhibited generally the same richness and abundance compared with larger and more distant reefs, it is suggested that 1-m3 artificial reefs placed individually and dispersed in a patchy configuration at <5 m apart should be preferentially used in management programs designed to increase fish biodiversity in physically homogeneous and structurally simple coastal environments that are seasonally influenced by freshwater outflows and sediments from large rivers.