WANG Zhenhua, CHEN Yong, ZHANG Shouyu, WANG Kai,ZHAO Jing, and XU Qiang



A Comparative Study of Fish Assemblages Near Aquaculture, Artificial and Natural Habitats

WANG Zhenhua1), 2), CHEN Yong2), ZHANG Shouyu1), *, WANG Kai1),ZHAO Jing1), and XU Qiang1)

1),,201306,2),,

Habitat plays a critical role in regulating fish community structure. Using the data collected from a monthly trammel net survey in Ma’an archipelago off the east coast of China, we evaluated impacts of five different habitats (artificial reefs, mussel farms, cage aquaculture, rocky reefs and soft bottom) on fish assemblages. This study suggests that artificial reefs (AR) have significantly higher species richness, abundance and diversity than mussel farms (MF) or soft bottom (SB) habitats during most seasons, and that fish taxa in the AR habitats are similar to those in the rocky reef (RR) habitats. Two different fish assemblage patterns were revealed in the study area using non-metric multidimensional scaling ordination: an assemblage dominated by reef fishes (especially by Scorpaenidae species) in AR, RR and cage aquaculture (CA) habitats and an assemblage dominated by Sciaenidae species in MF and SB habitats. We suggest that reef fishes play a key role in differentiating fish community structures in the study area. Although few differences in fish abundance and diversity were found between the CA and SB habitats, a more diverse age structure was observed in the CA habitats. A much more complex fish assemblage and enhanced population of local species were established as a result of the presence of both floating and fixed artificial structures, probably through improved survival rates.

artificial reef; species diversity; fish culture; mussel culture; aquaculture facility; fish assemblage

1 Introduction

The relationship between fish assemblage and habitat is an important research topic in ecology (Benaka, 1999; Fogarty, 1999;França., 2009; Simon., 2011; Burta., 2013). Most of these studies focus on coastal and estuarine habitats, such as shallow lagoons and straits (., Hajisamae and Chou, 2003; Martin., 2009), seagrass beds, mangroves, coral reefs and kelp forests (., Dorenbosch., 2006; James, 1988; Nagelkerken., 2007), rocky, sandy and soft bottom and algal-reefs (., Johan., 2007; Beyst., 2001). Understanding the fish assemblage-habitat relationships is particularly important because human activities such as fishing, aquaculture, oil and gas operations, offshore renewable energy developments and coastal development put considerable stress on natural habitats (Fraschetti., 2008). Such stress can lead to habitat degradation, destruction, fragmentation or loss (Sih., 2000; Gill, 2005; Hansson., 2007; Murdoch., 2007; Rybicki and Hanski, 2013). Approaches such as stock enhancement, sea ranching, and use of artificial reefs are often integrated into fishery management as an ecosystem approach to levitate habitat degradation (Leber., 2004).

The presence of artificial structures in aquatic systems may cause habitat fragmentation at different scales depending on the type of structures (Bulleri, 2005; Rybicki and Hanski, 2013). Structures such as artificial reefs and aquaculture facilities may increase habitat complexity by providing refuges and barriers that fragment the area, resulting in more heterogeneous assemblages (Sebens, 1991) and in turn changing fish community structure (Anderson., 2009;Fernández., 2009).

Marine structures, both natural and artificial, are known to aggregate fish (Pickering and Whitmarsh, 1997; Simon., 2011; Campbell., 2011). This has led to the development of fish aggregation devices (FADs) in fisheries using artificial structure (.., Stanley and Wilson, 2000; Machias., 2004; Valle., 2007; Pizzolon., 2008). Other studies claim that the presence of artificial structures can increase fish biomass (.., Bohnsack, 1989; Santos and Monteiro, 1997, 1998; Fabi., 2004). However, most of the studies only included a single artificial habitat. Few studies have compared fish assemblages among more than 3 artificial habitats within a relatively small area (Clynick., 2008a; Burta., 2013). Since these structures act as substitutes for natural habitat (., rocky shores), it is important to understand whether they support fish assemblages that are comparable to those found on natural substrata and if the environmental factors operating between and within artificial habitats and natural habitats are similar.

Mussel farming and fish cage aquaculture activities are widespread around coastal waters in many countries. Mussel farms can be found in New Zealand (Morrisey., 2006), Canada (Clynick., 2008b), Western Europe (Theodorou., 2010), China (Wang., 2010) and cage aquaculture in Western Europe (Machias., 2004; Valle., 2007; Sudirman., 2009), Southeast Asia (San Diego-McGlone., 2008; Sudirman, 2009) and China. Unique fish assemblages have been observed around deployed aquaculture facilities (Machias., 2004; Clynick., 2008b; Sudirman, 2009; Dempster., 2010), raising interesting questions about ecological roles of these structures in natural fishery ecosystems.

The objective of this study is to evaluate impacts of different habitats, in particular artificial structures and aquaculture facilities, on the dynamics of fish assemblages. Ma’an archipelago on the east coast of China provides an ideal experimental area for addressing this objective. This region features both aquaculture facilities set on soft bottom habitat and two artificial reef systems built in 2005-2006 and 2008 in the western marine reserve. All these artificial structures are situated near rocky shores. We test the following hypotheses: 1) the presence of additional structures on soft bottom habitat offers new living space for residents and transients, and thus increases both fish abundance and diversity on the whole, and 2) by attracting additional individuals to those artificial habitats, a more complex assemblage structure and age pattern can be formed within these areas.

The highlights in the present work are:

i) Aquaculture structures provide extra habitats for more local fishes and migrating fishes, and thus contribute to a small scale fisheries based on those species.

ii) Meta-habitats including both aquaculture and artificial habitats serve to establish a more complicated fish community structure. However, no significant differences can be found between them if reef fish are excluded.

iii) Age structure of fish community composition can be an effective tool in testing the population enhancement effects contributed by both aquaculture and artificial habitats.

2 Materials and Methods

2.1 Study Area and Site Selection

The western Ma’an archipelago off the Yangtze River estuary (Fig.1) is a traditional anchorage ground for international cargo ships and for offshore discharging. There are mainly two types of substrate in the subtidal area: rock-algal and clay. Less than 1% of this area is covered by rocky reefs. Most of the open waters share the same substrate, which is fine clay with particle diameter, ranging from 0.03 to 0.1mm. This soft bottom area is an important fishing ground for gill-netting and trawling. The area’s high productivity can be attributed to the runoff from the Qiantang and Yangtze Rivers interacting with the offshore Kuroshio current (Wang., 2010). It is an important habitat for the Chinese sturgeon () as well as other endangered species (Wang., 2010).

Fig.1 Spatial distribution of sampling sites and artificial habitats in western Ma’an Archipelago, East China Sea.

Many artificial structures were deployed in the Ma’an archipelago region in the last decade. About 1.8km2of mussel farms, divided into 6 regions, were developed during 2000–2008 in western Ma’an Archipelago. Nearly 140 fish cages were set in the south of western Lvhua Island between 2002 and 2004, covering an area of 1.6ha. In 2005 and early 2006, 554 individual concrete reef units of the same design (length×width×height=3m×3m×3m =27m3) were deployed between Dongku and Qiuzi Islands to form 13 clustered artificial reef groups, each consisting of 42–43 individual units with a reef area of approximately 330ha. In the vicinity of Sanheng Island, an additional 342 individual concrete reef units with two similar separate designs (length×width×height=3m×3m ×3m=27m3) were deployed in early 2008 to form 9 clustered artificial reef groups, each consisting of 37–39 units with a combined reef area of 200ha.

Eight sites were selected for our monthly fish sampling survey to examine the effects of artificial structures on fish assemblages (Fig.1). Sanheng (AR-1, average water depth=12.3±0.5m) and Dongku (AR-2, average water depth=15.3±1.3m) are two artificial reef (AR) sites located in the southeastern and eastern study areas. Two mussel farm (MF) sites, East Lvhua (MF-1, average water depth=10.1±0.3m) and West Lvhua (MF-2, average water depth=7.7±0.3m), are located in the northeast and northwest parts of the study area, respectively. Only one site, East Mandui (CA-1, average water depth=10.1±0.2m), was selected for cage aquaculture (CA) habitat due to limited spaces for trammel nets deployment. Three natural habitat sites, including 2 rocky reef (RR) sites (East Mantou (RR-1, average water depth=6.3±0.4m) and Tongqian reefs (RR-2, average water depth=6.2±0.1m)), and 1 soft bottom (SB) site (South Lvhua (SB-1, average water depth=9.7±0.7m)), were set as control sites because they are not artificially enhanced and share similar environmental characteristics (Carr and Hixon, 1997). All of the artificial structures were originally set in soft bottom habitats.

2.2 Field Sampling and Laboratory Analysis

We used multi-mesh trammel nets (MTNs) to sample our fish assemblage since most of the habitats are not easily trawled or seined. MTNs were set randomly each month in our selected 8 sites (each site is about 200m in length and 100m in width),.., no overlaps for the consecutive months. Multi-mesh gillnets and trammel nets were shown to be effective sampling tools for fish on various substrates (Acosta, 1997).

Four groups of experimental MTNs were deployed for approximately 24h on the same day at 4 sites. Thus it took 48h to sample all the 8 sites in each month. We sampled each site only once per month to avoid local overfishing (Olin., 2004) which can impact small-scale fish assemblages by severely reducing reef fish abundance over a small area. Monthly surveys at the sites were carried out over the same area (a transect of about 200m in length and 100m in width) within each habitat with the help of a GPS. All nets were set close to the bottom with anchors and placed in a direction perpendicular to the main tidal current (oriented northeast-southwest). Nets were set on top of natural reefs and soft bottom habitats and kept perpendicular to the main tidal current direction. They were deployed across the reef models in the AR habitats to make sure each part of the habitat was sampled effectively. In the CA habitat, nets were set around cages which were fixed to bottom structures. We set nets through gaps in the MF habitat and kept them perpendicular to tidal currents as well. Nets were checked twice a day in the summer to reduce mortality of captured fish.

Each group of trammel nets was composed of 2 nets (.., smaller-mesh net and larger-mesh net) that had 8 mesh sizes ranging from 2.5cm to 8.0cm. The panels in each net were ordered randomly. The size of smaller-mesh net was 1.5m×15m×4 panels with mesh sizes of 2.5cm, 3.4cm, 4.3cm and 5.8cm; and 27cm for covers. The size of larger-mesh net was 2.4m×30m×4 panels with mesh sizes of 5.0cm, 6.0cm, 7.0cm and 8.0cm; and 33cm for covers. Thus, the total length of our sampling transect was about 180m at each site. Due to high concentrations of juveniles and other small fish observed during previous sampling, we used only the 2.5cm mesh panel in the smaller-mesh net from January to August to ensure the efficiency of catching these size classes. Panels with mesh sizes of 3.4cm, 4.3cm and 5.8cm were included for smaller-mesh net from September to December. The larger-mesh net was kept the same throughout the year.

A total of 96 hauls were conducted in the 5 habitats from January to December in 2009. Fish collected in each survey were identified to species as well as measured, weighed and sexed. Stomach contents were collected for analyzing food selection preferences and otoliths were extracted for ageing. For each species, up to five individuals were aged in a 1cm-gap size group (if abundance ≤5, all individuals were aged).

2.3 Data Analysis and Statistical Methods

Fish abundance and biomass data were standardized for sampling (soak) time since the soak time was not exactly 24h for each net. Abundance per unit effort (APUE, indh−1) and biomass per unit effort (BPUE, gh−1) were used to represent the relative abundance of fishes in the targeted habitats.

Margalef’s species richness indexand Whilm’s species diversity indexwere used to compare the diversity levels among 5 habitats. We used the Whilm’s index because, based on our sampling experience, some species change habitat utilization strategies over their life history stages. For example, juvenile reef fishes such astend to be associated with algae and other structures while adult individuals choose mostly rocky reefs or hard bottom. The abundance-based diversity index cannot discern such differences if juvenile and adult fishes have similar abundance.

The number of species, APUE, BPUE and species diversity were analyzed seasonally (January–March,winter; April–June, spring; July–September, summer; October–December, autumn; the classification was based on the average seawater temperature recorded monthly in 2009) in order to better understand temporal variability of fish assemblage. Analyses of variance (ANOVA) were used for spatial, seasonal comparisons of those four indices among 5 habitats and 4 seasons. Posteriori comparison of means (relative abundance, relative biomass, species number and species diversity) was performed using the Turkey test.

Seasonal differences in fish assemblages among 5 habitats were examined using non-metric multidimensional scaling (NMDS) based on a Bray-Curtis similarity matrix. We used biomass matrix for ordination plots because some species might change favorable habitats in their life history stages, and the differences could be easily examined using individual weight rather than abundance data. Analysis of similarity (ANOSIM) was used for evaluating differences between groups (defined by two types of substrate samples and two types of habitats,, artificial and natural habitats). Two thousand permutations were run in the analysis. Rare species were excluded from the analyses (Araújo., 1999) and a square root transformation was performed on the biomass data before the analysis. All multivariate analyses were performed using PRIMER software package v5 (Clarke and Gorley, 2001). A significance level of 5% was used in all statistical analyses.

3 Results

3.1 Fish Abundance and Diversity

In total 66 fish species were identified belonging to 38 families (Table 1). Fish taxa, by either family or genus, were more abundant in the purpose-built AR habitats (27 families and 36 genera at AR-1; 23families and 32 genera at AR-2) compared with those in other non purpose-built artificial habitats and natural habitats (16 families and 23 genera at MF-1; 16 families and 21 genera at MF-2; 20 families and 27 genera in the CA habitat; about 20 families and 28 genera for the 2 RR habitats; and 15 families and 20 genera in the SB habitat). Fewer taxa were found in the RR and CA habitats. The lowest number of species in taxa was observed in the SB and MF habitats. Species diversity and fish abundance were highest in the AR habitat (Table 1). Site RR-1 had the lowest abundance (only 290 individuals were collected in 2009), but its total biomass was higher than biomass found in more than half of the study sites. Habitats with concrete structures on the seabed such as AR and CA had much higher concentration and occurrence (see Appendix A) of species likeand, than natural RR, SB and MF habitats. Those two species were observed in the CA habitat only. Fishes such as,,andwere most attracted to sites with artificial reefs (Appendix A).was also a common species, found most often in the MF habitat. Some seasonal species likeandwere also common at AR sites (Appendix A).

Species richness, relative abundance (APUE&BPUE) and species diversity varied considerably among the four seasons (Table 2). Significant seasonal differences were found within different habitats for species richness, Whilm’s diversity and APUE. However, differences in BPUE among the study sites were smaller (Table 2).

Table 1 The total number of species, abundance and biomass during the whole survey period

Table 2 Two-fixed-factor ANOVA

Notes: Species number, species diversity, abundance per unit effort and biomass per unit effort serve as functions of habitats (artificial reef, mussel farm, cage aquaculture, rocky reef and soft bottom) and seasons includes spring, summer, autumn and winter. NS,>0.05;*<0.05;**<0.01.

Many more species were found in the AR habitats than in the SB habitat in spring and summer (Fig.2a). AR-1 and AR-2 sites were found to be significantly higher in Margalef’s species richness than MF-2, MF-1 and CA-1 in autumn, and higher than RR-2 in winter. No other significant differences were found among those habitats during other seasons. Species collected in MF were similar to those from the SB habitat in each season; it also displayed low species richness. The habitat exhibiting the highest abundance was RR-2 in summer, which was significantly higher than the SB and MF habitats (Fig.2a).

The average APUE was higher at AR sites than that at the other sites in most of the seasons except for autumn (Fig.2b). Significantly higher APUE was found at AR-2 in both spring and summer, which was higher than those at MF-2 (=0.036) and SB-1 (=0.027) in spring and RR-1 (=0.026) in summer. AR-1 had a much higher APUE in winter than most of the sites (Fig.2b). Spatial variability of BPUE over the 8 sites in each season followed similar trends to that of APUE. All of the sites with concrete reefs (AR-1, AR-2&CA-1) had a significantly higher BPUE than one of the mussel farm sites in spring (Fig.2c). The highest BPUE was found at AR-2 in summer, much higher than those for the sites of RR-1 and SB-1.No significant differences were found in autumn among the 8 sites. Though some sites like RR-1 and SB-1 had a higher BPUE than the other sites on average, only the sites with artificial reefs were found to have a significantly higher BPUE than MF-1 and CA-1 in winter (Fig.2c).

Fig.2 Seasonal variations of fish abundance and diversity in 2009. (a), number of species; (b), abundance per unit effort (APUE); (c), biomass per unit effort (BPUE); (d), Whlim’s species diversity. Results are expressed as means±SD. The error bars represent the standard error. ‘*’ and combined lines indicate the significance level between the compared groups.*P<0.05;**P<0.01. Post hoc Turkey-HSD test was used.

Significantly higher Whilm’s species diversity was observed in all seasons for the 2 AR sites (Fig.2d). Only CA-1 had a higher diversity than MF-2 in summer for other artificial habitats. Sites without natural rocks or concrete reefs (.., MF-2, MF-1 and SB-1) were found to be much lower in fish diversity in most seasons.

3.2 Fish Assemblage Patterns and Age Structures

Two types of assemblage patterns, reef bottom and soft bottom, were identified based on seasonal fish community (left four panels in Fig.3). This classification was significant for all seasons (Table 3). However, there were no significant differences between reef habitats (AR, CA and RR) and soft bottom habitats (MF and SB) when reef fishes were excluded (right four panels in Fig.3, Table 3). Reef fishes, especially dominant ones such as,,,and, made major contributions to the reef bottom fish assemblage structures (Appendix A). Species such as,,,andwere much more abundant in SB and MF habitats than those in reef bottom habitats, and contributed to the formation of fish assemblage in soft substrate habitats (Appendix A). In spring, summer and autumn, fish assemblage patterns identified for those habitats containing artificial reefs were signify-cantly different from those in sites where there were only natural rocky reefs (left four panels in Fig.3, Table 3). When reef fish were excluded, only the summer assemblages showed significant difference between those two types of habitats (Fig.3, Table 3). The fish assemblages in MF were significantly different from those in SB habitats in all seasons. However, the data in summer and autumn showed no differences when reef fish were not included (Fig.3, Table 3).

Fig.3 NMDS plots on the assemblages of fish in 5 habitats at 8 sites (column a: reef fish included; column b: reef fish excluded). The biomass matrix was used in non-metric multidimensional scaling analysis. All original data was square-root transformed.

Table 3The list of R values from analysis of similarity (ANOSIM) between 3 groups

Notes: A=habitat with manmade reefs on bottom (AR&CA); B= habitat with natural reefs on bottom (RR); C=habitat with manmade or natural reefs on bottom (AR, CA&RR); D=soft bottom without any artificial structures (SB); E=soft bottom with mussel farming structures (MF); F=soft bottom with or without mussel farming structures (SB&MF). One Way Global Test for groups of C-F and One Way Pairwise Tests for groups of A-B and D-E. NS,>0.05;*<0.05;**<0.001.

In considering age structure, many more individuals older than 2 years were sampled from sites in the AR, CA and RR habitats than those from sites in the MF and SB habitats, especially at site CA-1 in the CA habitat (Fig.4). A greater variety of age groups and wider range of age structure were observed in the CA habitat, and most of those 3- or 4-year-old individuals were scorpion fish (). Age structure in the AR habitats was similar to that in the RR habitat with low abundance of fish in the 3-4 years old groups. Few individuals with ages older than 2 years were collected in the MF and SB habitats. Most individuals collected in our monthly sampling were under 1 year old in all sites (Fig.4).

Fig.4 Monthly variation of percentage of four age groups for all fishes at 8 sites during the study period. age1=0+, age2=1+, age3=2+, age4=3+. The relative percentage of each year group is shown by bars and the exact number of each age group is given in the table down below.

4 Discussion

4.1 The Differences in Species Composition, Fish Abundance and Diversity Among 5 Habitats

In our investigation three different types of artificial structures were found on natural SB habitat, each of which was showed to contribute differently to seasonal variations in species composition, fish abundance, and diversity. Masuda. (2010) compared fish richness and density among three types of ARs and found significant differences among them during most of the sampling years. Walker. (2002) also compared fish assemblages between concrete aggregates and quarry stone, and found significant seasonal variation of fish abundance and species richness for both types of ARs. However, the differences between these types of ARs were not significant. Although these studies may offer some evidence as to the functions of different types of ARs, few studies have compared fish assemblages in AR habitat with other artificial habitat, such as MF and CA. This makes it difficult to draw a comparable conclusion. Many studies suggest that marine structures can attract certain aquatic animals to colonize on them (Pickering and Whitmarsh, 1997), or just offer temporary uses for activities such as feeding, protection from predators and serving as spawning ground for migrating species (DeMartini., 1994).

In this study AR habitats had different fish taxa, abundance and species diversity compared with SB habitats in most seasons. However, indices for these characteristics did not differ greatly between the areas with aquaculture structures and natural habitats. Complex fish taxa composition and higher abundance and diversity of fishes were found mostly in the AR habitat, indicating that the AR tended to enhance local fish fauna complexity and diversity. This is consistent with many previous findings that artificial reef deployment has the potential to enhance fish assemblages and increase species diversity (., Santos., 1997; Rilov and Benayahu, 2000; Relini., 2007). However, a few positive changes have been noted in areas with improperly designed or sited artificial reefs (Baine, 2001; Polovina, 1994; Bortone., 1994). It has been reported that artificial reefs could provide additional carrying capacity and therefore abundance and biomass of reef biota, which explains why fish tended to be most abundant in the AR habitat in our study.

Although both MF and CA were established in SB areas, the composition of their dominant fish species was different. The MFs shared similar taxonomic structure with soft bottom, and onlywas found to be more abundant than in the natural SB habitat. However, species composition in the CA habitat was similar to that in the AR and RR habitats, especially for reef fish. The primary cause could be the difference in materials and spatial locations of the studied artificial structures. Residual food from feeding caged fish from April to November could be another reason why there were many wilder reef fishes sampled in this habitat. These wild fishes might not directly use food pellets dropped under the cages, but might prey on shrimp or crabs attracted by feeding events (Machias., 2004; Sudirman., 2009). By definition, fish species that respond directly to artificial or natural reefs and make them a permanent or seasonal habitat are referred to as reef-associated fishes (Bellwood, 1998). This might result in fish species composition in the CA and AR habitats being similar to that in the natural RR habitat.

Increased habitat complexity is considered favorable for promoting diversity and abundance (Fernández., 2009). Ambrose and Swarbrick (1989) compared fish assemblages on artificial and natural reefs and found a greater number and density of benthic fish species in AR habitats compared to natural habitats, but species richness, diversity and total individuals were not different between artificial and natural reefs. Similar results were also found in other studies (Machias., 2004; Clynick., 2008b). Some other studies identified significantly higher abundance and diversity of wild fish around aquaculture structures (Valle., 2007; Sudirman., 2009). We believe that this might result from different sampling approaches in different studies. The three dimensional extension of aquaculture structures is quite different from the bottom set of ARs (Perkol-Finkel., 2008). Most of the facilities are suspended from the water surface for the MFs and CA habitats. Our bottom nets could barely sample those fishes near bottom and pelagic fishes, especially juvenile fish. Sudirman. (2009) reported that wild fishes are significantly more abundant in near-surface depths around the margins of the cages in the morning than at other times of the day. Similarly, we found species such asandin great numbers around MFs during our field investigation, but we could not effectively sample them. This was probably due to their sensory ability to detect obstacles (Duffy, 1987). Taking those pelagic individuals into consideration, there should be a much higher level of fish diversity in the MFs or CA habitats compared with the adjacent SB habitats.

4.2 Fish Assemblage Patterns and Age Structures

One of the key objectives of habitat restoration by deploying artificial structures is to enhance fish assemblages in a designated area and to maintain higher population densities of target species (Pickering and Whitmarsh, 1997). The deployed artificial reefs and established aquaculture structures around our study area are mostly acting as combined habitat enhancement systems and FADs, rather than habitat replacements. Thus new fish assemblage pattern could be easily found due to change of available food and living space. Masuda. (2010) demonstrated that deployment of ARs does not reduce the number of fish species or density of individuals in the adjacent areas. This also suggests that introduction of ARs in certain areas may have limited impacts on the adjacent natural habitat.

In our study two major fish assemblage patterns have been revealed, including the assemblage dominated by reef fishes (especially by Scorpaenidae species) in the AR, RR and CA habitats and the assemblage mainly dominated by Sciaenidae species such as Belenger’s croaker()and small yellow croaker () in the MF and SB habitats. There are significant differences between fish assemblages in artificial habitats and natural habitats for most seasons except winter. Ambrose and Swarbrick (1989) claimed that the fish assemblages on artificial reefs are generally similar to the assemblages on natural reefs, similar to our findings here. A recent study suggested that artificial reef can hardly retain reef fishes effectively, especially for groupers (Addis., 2013), indicating that they have an equal chance to choose habitat between natural and artificial reefs. If reef fish are excluded, the differences are found only in summer between the AR and RR habitats and in spring and winter between MF and SB habitats, which might result from seasonal changes in density of migratory species. This indicates that permanent residents or reef fishes are the driving factors in differentiating fish assemblages in the studied area. Early studies, such as the comparison of fish populations on artificial and natural reefs in the Florida Keys by Stone. (1979), suggested that ARs could be used to expand reef fish stocks. This explains why the AR and CA have similar fish assemblage pattern to that for the natural RR habitat. This study supports the hypothesis that deployment of artificial structures can enhance local fish assemblages and change community structure compared with natural habitats as a result of the newly formed reef fish aggregations.

Although our study did not identify any differences in abundance, biomass or diversity between CA habitat and RR habitats, an interesting result was revealed by examining the age structure of fish assemblages across the sampling sites. Many more individuals older than two years were sampled in CA habitat than any other artificial or natural habitats, indicating that permanent species inhabiting the CA habitat had higher survival rates than those in other habitats, and/or the CA habitat was more attractive to older fish than the other habitats. Most of the older individuals were species like scorpion fish () and Fat greenling (), which are the most important economic species in the region. Ogawa (1973) stated that properly constructed ARs or submarine forests could increase survival, growth, and feeding efficiency of certain juvenile fishes, thus increasing the total biomass of reef fishes (Stone., 1979). This conclusion can explain why the CA habitat could act as a reef fish stock enhancement system in our study. The higher food availability and lower disturbance from fishing activities in artificial habitats may explain the age structure observed in the artificial habitat in this study. No fishing around or within the CA area was allowed since the establishment of the cages except for our monthly sampling in 2009. However, gillnetting and fish trapping were often found to appear in the AR and MF habitats. It seems logical that population enhancement for specific species through deployment of artificial structures should at least involve protection measures as a prior consideration in habitat management.

5 Conclusions

The development of aquaculture structures and deployment of artificial reefs established a more complex and diversified habitat for fish species. This also restricts the use of mobile fishing gear and creates more undisturbed or minimally disturbed areas, providing extra habitats for local species (Burta., 2013) and enhancing populations of demersal fish in the area. A more complex age structure can be found in the CA habitat due to limited fishing. Migrating species also benefit from the existence of artificial structures, especially in the AR habitat.

Acknowledgements

We would like to thank Mrs. Zhiguo Zhang, Ming Xu, Wujun Tong, Mingchao Chen, Qingman Chen, and Linna Ye for their assistance in the fieldwork. This research was supported by the National High Technology Research and Development Program of China (863 Program, No. 2006AA100303) and the National Basic Research Program of China (No. 2011CB111608). Z. Wang and Y. Chen’s involvement in the project was partially supported by the Maine Sea Grant College Program at the University of Maine in the USA.

Acosta, A., 1997. Use of multi-mesh gillnets and trammel nets to estimate fish species composition in coral reef and mangroves in the southwest coast of Puerto Rico., 33: 45-57.

Addis, D. T., Patterson III, W. F., Dance, M. A., and Ingram Jr., G. W., 2013. Implications of reef fish movement from unreported artificial reef sites in the northern Gulf of Mexico., 147: 349-358.

Ambrose, R. F., and Swarbrick, S. L., 1989. Comparison of fish assemblages on artificial and natural reefs off the coast of southern California., 44: 718-733.

Anderson, T. J., Syms, C., Roberts, D. A., and Howard, D. F., 2009. Multi-scale fish-habitat associations and the use of habitat surrogates to predict the organization and abundance of deep-water fish assemblages., 379: 34-42.

Araújo, F. G., Bailey, R. G., and Williams, W. P., 1999. Spatial and temporal variations in fish populations in the upper Thames Estuary., 55: 836-853.

Baine, M., 2001. Artificial reefs: A review of their design, application, management and performance., 44: 241-259.

Bellwood, D. R., 1998. What are reef fishes?–Comment on the report by D. R. Robertson: Do coral-reef fish faunas have a distinctive taxonomic structure?, 17: 179-186.

Benaka, L., 1999.. American Fisheries Society, Bethesda, MD, USA, 480pp.

Beyst, B., Hostens, K., and Mees, J., 2001. Factors influencing fish and macrocrustacean communities in the surf zone of sandy beaches in Belgium: temporal variation.,46: 281-294.

Bohnsack, J. A., 1989. Are high densities of fishes at artificial reefs the result of habitat limitation or behavioural preference?, 44: 631-645.

Bortone, S. A., Martin, T., and Bundrick, C. M., 1994. Factors affecting fish assemblage development on a modular artificial reef in a northern Gulf of Mexico estuary., 55: 319-332.

Bulleri, F., 2005. Role of recruitment in causing differences between intertidal assemblages on seawalls and rocky shores., 287: 53-65.

Burta, J. A., Feary, D. A., Cavalcante, G., Baumand, A. G., and Usseglio, P., 2013. Urban breakwaters as reef fish habitat in the Persian Gulf., 72: 342-350.

Campbell, M. D., Rose, K., Boswell, K., and Cowan, J., 2011. Individual-based modeling of an artificial reef fish community: Effects of habitat quantity and degree of refuge., 222: 3895-3909.

Carr, M. H., and Hixon, M. A., 1997. Artificial reefs: The importance of comparisons with natural reefs.,22: 28-33.

Clarke, K. R., and Gorley, R. N., 2001.. PRIMER-E, Plymouth UK, 91pp.

Clynick, B. G., Chapman, M. G., and Underwood, A. J., 2008a. Fish assemblages associated with urban structures and natural reefs in Sydney, Australia., 33: 140-150.

Clynick, B. G., McKindsey, C. W., and Archambault, P., 2008b. Distribution and productivity of fish and macroinvertebrates in mussel aquaculture sites in the Magdalen islands (Québec, Canada)., 283: 203-210.

DeMartini, E. E., Barnett, A. M., Johnson, T. D., and Ambrose, R. F., 1994. Growth and production estimates for biomass dominant fishes on a southern California artificial reef., 55: 484-500.

Dempster, T., Sanchez-Jerez, P., Uglemd, I., and Bjørn, P.-A., 2010. Species-specific patterns of aggregation of wild fish around fish farms., 86: 271-275.

Dorenbosch, M., Grol, M. G. G., Nagelkerken, I., and Van der Velde, G., 2006. Seagrass beds and mangroves as potential nurseries for the threatened Indo-Pacific humphead wrasse,and Garibbean rainbow parrotfish,., 129: 277-282.

Duffy, J. M., 1987. A review of the San Diego Bay stripied mullet,, fishery., 56:1-10.

Fabi, G., Grati, F., Puletti, M., and Scarcella, G., 2004. Effects of fish community induced by installation of two gas platforms in the Adriatic Sea., 273: 187- 197.

Fernández, T. V., Anna, G. D., Badalamenti, F., and Pérez-Ruzafa, A., 2009. Effect of simulated macroalgae on the fish assemblage associated with a temperate reef system., 376: 7-16.

Fogarty, M. J., 1999. Essential habitat, marine reserves and fishery management., 14: 133-134.

França, S., Costa, M. J., and Cabral, H. N., 2009. Assessing habitat specific fish assemblages in estuaries along the Portuguese coast.,, 83: 1-12.

Fraschetti, S., Terlizze, A., and Boero, F., 2008. How many habitats are there in the sea (and where)?, 366: 109-115.

Gill, A. B., 2005. Offshore renewable energy: Ecological implications of generating electricity in the coastal zone., 42: 605-615.

Hajisamae, S., and Chou, L. M., 2003. Do shallow water habitats of an impacted coastal strait serve as nursery grounds for fish?,, 56: 281-290.

Hansson, S., Hjerne, O., Harvey, C., Kitchell, J. F., Cox, S. P., and Essington, T. E., 2007. Managing Baltic Sea fisheries under contrasting production and predation regimes: Ecosystem model analyses., 36: 265-271.

James, L. B., 1988. Effects of kelp forest removal on associated fish assemblages in central California., 117: 227-238.

Johan, S., Leif, P., and Håkan, W., 2007. Food utilization by coastal fish assemblages in rocky and soft bottoms on the Swedish west coast: Inference for identification of essential fish habitats.,, 71: 593-607.

Leber, K. M., Kitada, S., Blankenship, H. L., and Svasand, T., 2004.. Blackwell Publishing Ltd., UK, 576pp.

Machias, A., Karakassis, I., Labropoulou, M., Somarakis, S., Papadopoulou, K. N., and Papaconstantinou, C., 2004. Changes in wild fish assemblages after the establishment of a fish farming zone in an oligotrophic marine ecosystem.,, 60: 771-779.

Martin, S., Alfred, S., Antti, L., Johanna, M., Kajsa, R., and Lauri, U., 2009. Fish assemblages in coastal lagoons in land-uplift succession: The relative importance of local and regional environmental gradients.,, 81: 247-256.

Masuda, R., Shiba, M., Yamashita, Y., Ueno, M., Kai, Y., Nakanishi, A., Torikoshi, M., and Tanaka, M., 2010. Fish assemblages associated with three types of artificial reefs: Density of assemblages and possible impacts on adjacent fish abundance., 108: 162-173.

Morrisey, D. J., Cole, R. G., Davey, N. K., Handley, S. J., Bradley, A., Brown, S. N., and Madarasz, A. L., 2006. Abundance and diversity of fish on mussel farms in New Zealand., 252: 277-288.

Murdoch, T. J. T., Glasspool, A. F., Outerbridge, M., Ward, J., Manuel, S., Gray, J., Nash, A., Coates, K. A., Pitt, J., Fourqurean, J. W., Barnes, P. A., Vierros, M., Holzer, K., and Smith, S. R., 2007. Large-scale decline in offshore seagrass meadows in Bermuda., 339: 123-130.

Nagelkerken, I., van der Velde, G., Gorissen, M. W., Meijer, G. J., van’t Hof, T., den Johan, S., Leif, P., and Håkan, W., 2007. Food utilization by coastal fish assemblages in rocky and soft bottoms on the Swedish west coast: Inference for identification of essential fish habitats.,, 71: 593-607.

Ogawa, R., 1973. Various biological questions regarding artificial reefs., 3: 21-30.

Olin, M., Kurkilahti, M., Peitola, P., and Ruuhijärvi, J., 2004. The effects of fish accumulation on the catchability of multimesh gillnet., 68: 135-147.

Perkol-Finkel, S., Zilman, G., Sella, I., Miloh, T., and Benayahu, Y., 2008. Floating and fixed artificial habitats: Spatial and temporal patterns of benthic communities in a coral reef environment.,, 77: 491-500.

Pickering, H., and Whitmarsh, D., 1997. Artificial reefs and fisheries exploitation: A review of the ‘attraction versus production’ debate, the influence of design and its significance for policy., 31: 39-59.

Pizzolon, M., Cenci, E., and Mazzoldi, C., 2008. The onset of fish colonization in a coastal defence structure (Chioggia, Northern Adriatic Sea).,, 78: 166-178.

Polovina, J. J., 1994. Function of artificial reefs.,55: 1349.

Relini, G., Relini, M., Palandri, G., Merello, S., and Beccornia, E., 2007. History, ecology and trends for artificial reefs of the Ligurian Sea, Italy., 580: 193-217.

Rilov, G., and Benayahu, Y., 2000. Fish assemblages on natural versus vertical artificial reefs: The rehabilitation perspective., 136: 931-942.

Rybicki, J., and Hanski, I., 2013. Sepecies-area relationships and extinctions caused by habitat loss and fragmentation., 16: 27-38.

San Diego-McGlone, M. L., Azanza, R. V., Villanoy, C. L., and Jacinto, G. S., 2008. Eutrophic waters, algal bloom and fish kill in fish farming areas in Bolinao, Pangasinan, Philippines., 57: 295-301.

Santos, M. N., and Monteiro, C. C., 1997. The Olhão artificial reef system (south Portugal): Fish assemblages and fishing yield., 30: 33-41.

Santos, M. N., and Monteiro, C. C., 1998. Comparison of the catch and fishing yield from an artificial reef system and neighbouring areas off Faro (Algarve, south Portugal)., 39: 55-65.

Sebens, K. P., 1991. Habitat structure and community dynamics in marine benthic systems. In:. Bell, S. S.,., eds., Chapman and Hall, New York, 211pp.

Sih, A., Jonsson, B. G., and Luikart, G., 2000. Habitat loss: ecological, evolutionary and genetic consequences., 15: 132-134.

Simon, T., Pinheiro, H. T., and Joyeux, J. C., 2011. Target fishes on artificial reefs: Evidences of impacts over nearby natural environments., 409: 4579-4584.

Stanley, D. A., and Wilson, C. A., 2000. Variation in the density and species composition of fishes associated with three petroleum platforms using dual beam hydroacoustics., 47: 161-172.

Stone, R. B., Pratt, H. L., Parker, R. O., and Davis Jr., G. E., 1979. A comparison of fish populations on an artificial and natural reef in the Florida Keys., 41: 1-11.

Sudirman, H. H., Jompa, J., Iswahyudin, Z., and McKinnon, A. D., 2009. Wild fish associated with tropical sea cage aquaculture in South Sulawesi, Indonesia., 286: 233-239.

Theodorou, J. A., Sorgeloos, P., Adams, C. M., Viaene, J., and Tzovenis, I., 2010. Optimal farm sie for the production of the mediterranean mussel () in Greece. IIFET 2010 Montpellier Proceedings, 6pp.

Valle, C., Bayle-Sempere, J. T., Dempster, T., Sanchez-Jerez, P., and Gimenez-Casalduero, F., 2007. Temporal variability of wild fish assemblages associated with a sea-cage fish farm in the south-western Mediterranean Sea.,, 72: 299-307.

Walker, B. K., Henderson, B., and Spieler, R. E., 2002. Fish assemblages associated with artificial reefs of concrete aggregates or quarry stone offshore Miami Beach, Florida, USA., 15: 95-105.

Wang, Z. H., Zhang, S. Y., and Wang, K., 2010. Fish and macroinvertebrates community structure in artificial habitat around Sanheng Isle, Shengsi, China., 30: 2026-2035 (in Chinese).

(Edited by Qiu Yantao)

Appendix

A List of taxa recorded and their total individual numbers and frequency of occurrence (% in 12 months) in associated habitats during the study period

Category /speciesFamilySites in artificial habitatsSites in natural habitats SHDKELHWLHEMDEMTTQRSLH Permanent residents Sebastiscus marmoratus Scorpaenidae138/100144/10020/41.727/58.3194/10068/10097/91.71/8.3 Hexagrammos agrammus Hexagrammidae1/8.33/252/8.39/41.74/33.34/25 Hexagrammos otakii Hexagrammidae30/83.338/91.735/7516/58.34/25 Platycephalus indicus Platycephalidae35/5061/66.75/16.76/16.72/8.35/2513/58.3 Paralichthys olivaceusBothidae 5/33.319/58.31/8.32/16.711/33.36/25 Zebrias japonicaCynoglossidae1/8.32/16.71/8.32/16.7 Cynoglossus abbreviatesCynoglossidae7/16.71/8.31/8.32/8.3 Cynoglossus joyneri Cynoglossidae1/8.35/8.32/16.71/8.33/16.7 Odontamblyopus rubicundus Taenioididae1/8.31/8.3 Amblychaeturichthys hexanema Gobiidae 1/8.33/8.36/8.37/16.7 Nibea albiflora Sciaenidae35/58.336/66.722/506/41.739/58.324/58.323/66.739/41.7 Johnius belangeriiSciaenidae19/58.325/33.3186/100116/91.754/58.320/33.319/41.762/75 Miichthys miiuySciaenidae13/2549/41.717/2563/8.341/16.75/33.37/16.7 Muraenesox cinereus Muraenesocidae1/8.33/253/16.73/16.71/8.39/41.7 Raja porosaRajidae2/8.3 Acanthopagrus schlegeli Sparidae1/8.31/8.38/41.713/41.7 Acanthopagrus latusSparidae1/8.3 Rhabdosargus sarbaSparidae2/16.72/8.34/16.75/25

Category /speciesFamilySites in artificial habitatsSites in natural habitats SHDKELHWLHEMDEMTTQRSLH Pagrus majorSparidae2/16.79/253/8.31/8.31/8.3 Lateolabrax maculatesSerranidae4/252/16.72/16.713/33.34/33.3 Microcanthus strigatusKyphosidae5/8.32/16.7 Oplegnathus fasciatus Oplegnathidae1/8.32/8.32/8.31/8.3 Parapristipoma trilineatum Pomadasyidae1/8.31/8.31/8.33/25 Harpadon nehereusHarpodontidae2/16.72/16.71/8.31/8.33/251/8.3 Seasonal residents Apogon semilineatus Apogonidae6/16.73/254/251/8.31/8.3 Decapterus maruadsi Carangidae60/16.714/257/16.75/16.76/16.711/16.724/16.727/8.3 Psenopsis anomala Centrolophidae1/8.31/8.3 Chaetodon modestus Chaetodontidae3/8.31/8.31/8.3 Goniistius zonatusCheilodactylidae3/8.32/8.36/8.31/8.3 Konosirus punctatus Clupeidae14/16.7 Harengula thrissina Clupeidae1/8.31/8.311/16.71/8.5 Thryssa kammalensis Engraulididae501/41.7693/33.3201/25254/25106/33.373/33.3129/33.3199/33.3 Coilia ectenes Engraulididae3/16.72/8.31/8.3 Setipinna tay Engraulididae2/8.31/8.31/8.3 Stolephorus commersonii Engraulididae46/16.7100/257/16.72/16.71/8.34/16.72/8.3 Engraulis aponicas Engraulididae22/25119/258/16.76/8.31/8.31/8.312/16.71/8.3 Thrissa mystax Engraulididae23/16.74/252/8.32/16.73/16.71/8.32/8.311/16.7 Sphyraena japonicaHemiramphidae1/8.3 Lophius litulon Lophiidae2/8.36/16.7 Stephanolepis cirrhifer Monacanthidae2/16.72/16.76/254/8.37/25 Plotosus anguillaris Plotosidae6/2566/2510/251/8.3 Eleutheronema tetradactylum Polynemidae2/8.35/16.71/8.31/8.32/8.3 Collichthys lucidus Sciaenidae1/8.3 Argyrosomus argentatusSciaenidae3/8.31/8.3 Larimichthys croceaSciaenidae4/252/16.71/8.3 Larimichthys polyactisSciaenidae92/58.395/41.744/7536/66.717/8.34/16.715/33.317/33.3 Nibea japonica Sciaenidae39/8.31/8.311/8.3 Scomberomorus niphonius Scombridae5/8.32/16.7 Scomber japonicas Scombridae10/16.7 Pampus argenteus Stromateidae3/16.75/16.72/8.3 Takifugu oblongusTetrodontidae1/8.33/8.3 Pelates quadrilineatus Theraponidae2/8.31/8.32/8.3 Trichiurus haumela Trichiuridae1/8.35/16.72/8.3 Chelidonichthys kumuTriglidae3/8.32/8.3 Casual/rare species Navodon modestus Aluteridae1/8.31/8.3 Goniistius quadricornis Cheilodactylidae1/8.3 Ilisha elongataClupeidae1/8.31/8.3 Cynoglossus gracilisCynoglossidae1/8.3 Elops saurus Elopidae1/8.3 Oplegnathus punctatus Oplegnathidae1/8.3 Hapalogenys mucronatus Pomadasyidae1/8.3 Epinephelus akaara Serranidae1/8.3 Sillago sihama Sillaginidae1/8.3 Sillago japonica Sillaginidae1/8.3 Hyporhamphus sajori Sphyraenidae1/8.3 Takifugu niphobles Tetrodontidae1/8.3 67 species39 families10901529566576552290421411

DOI 10.1007/s11802-015-2455-x

ISSN 1672-5182, 2015 14 (1): 149-160

© Ocean University of China, Science Press and Spring-Verlag Berlin Heidelberg 2015

(August 18, 2013; revised November 18, 2013; accepted August 29, 2014)

* Corresponding author. Tel: 0086-21-61900336 E-mail: syzhang@shou.edu.cn