The Inﬂuence of Manila Clam ( Ruditapes philippinarum ) on Macrobenthos Communities in a Korean Tidal Ecosystem

: We investigated the biological impact of extensive Manila clam ( Ruditapes philippinarum ) aquaculture on macrobenthic communities in a tidal ecosystem in Korea. We collected macrobenthos ( > 1 mm in length) samples seasonally in the intertidal zone in Geunsoman, Taean, Korea from April 2011 to December 2014. We identiﬁed 146 macrobenthos species, including 60 polychaetes, 53 crustaceans, and 16 mollusks. A biota–environment matching (BIO–ENV) analysis indicated that the benthic community was a ﬀ ected by mean sediment grain size (Mz), total organic carbon (TOC), and R. philippinarum biomass. We found no correlation between R. philippinarum and the main dominant species ( Heteromastus ﬁliformis , Ceratonereis erythraeensis, and Ampharete arctica ), which have a di ﬀ erent feeding strategy; thus, this may result in a lack of competition for food resources. In addition, we found that ﬂourishing R. philippinarum positively a ﬀ ects the macrobenthos density but negatively a ﬀ ects the biodiversity index. Moreover, competition between species does not occur clearly, and environmental variables (sediment, organic carbon) are important.


Introduction
The distributions of intertidal macrobenthic taxa are strongly influenced by environmental factors such as sediment type, temperature, salinity, organic carbon, etc. [1]. They play a critical role in the structure and functioning of marine ecosystems [2]. Benthos are consumed by fish and mammals, thereby providing food for higher trophic levels [3]. Macrobenthos are also important in organic matter cycling and nutrients and provide a link between the benthic and pelagic division of marine ecosystems [4]. They are used as indicators of coastal ecosystem health and environmental quality because this group is characterized by long-lived species with limited habitat ranges and high sensitivity to environmental change [5]. Macrobenthic animals are critical links between primary producers and high trophic level consumers in coastal food webs [6]. Therefore, macrobenthos are important to research targets in marine ecology and are essential to the structure and function of coastal ecosystems [7,8].
Studies of competition in marine benthic animals have been pivotal to our understanding of ecology systems overall; some of the earliest and most influential evidence of competition comes from studies of sessile marine benthic animals, and we continue to gain an understanding of community dynamics from this group [9]. Research on competition among marine invertebrates tends to focus on interference rather than exploitation [10]. Benthic organisms transport oxygen and organic matter from the surface to deeper layers, extending the habitat suitable for smaller fauna [11]. Competition, disturbance, and predation can also influence the spatial distribution of these small benthic animals [12]. Therefore, instead of multiple sites, one site was selected and sampled randomly. Samples were sieved through a 1 mm size mesh. Residue on the mesh was sorted and preserved in 10% formalin in seawater. A sample of surface sediments from the surface layer was collected to analyze sediment grain size (Mz) and total organic carbon (TOC) concentration, and these samples were frozen before analysis. In the laboratory, all organisms were sorted from the sediment and their wet weight was measured. Then, they were identified to the species level under a stereomicroscope.
analysis. In the laboratory, all organisms were sorted from the sediment and their wet weight was measured. Then, they were identified to the species level under a stereomicroscope.

Sediment Analysis
Sediment particle sizes were determined after treating samples with a solution of 10% hydrogen peroxide. Sediment samples were heated to >100 °C to evaporate the hydrogen peroxide and then washed at least three times with distilled water to remove organisms and salts. Washed samples were then passed through a 63-µm standard sieve. After drying, the sediments trapped by the sieve were weighed and subjected to automatic particle size analysis using a SediGraph 5120 device (Kunash Instruments, Mumbai, India) following the addition of sodium hexametaphosphate as a dispersing agent. We then calculated the average particle sizes and degrees of sorting. Sediments were categorized according to Folk's classification system [41]. The content of total organic carbon (TOC) in sediments was analyzed using a Shimadzu TOC analyzer, (SSM-5000A, Shimadzu, Japan).

Statistical Analyses
Density and biomass data are recalculated per square meter; statistical analyses were performed for all species. The Shannon-Wiener diversity index (H') was calculated using density data. Cluster and non-metric multi-dimensional scaling (nMDS) analyses on macrobenthic community data for each sampling period were analyzed using the Bray-Curtis similarity measure based on fourth-root transformed density data and group average linkage. A similarity profile (SIMPROF) permutation test was performed to determine the statistically significant clusters among the samples. A similarity percentage (SIMPER) analysis was used to determine the contribution of each species to similaritydissimilarity among groups. A biota-environment matching (BIO-ENV) analysis was conducted to determine the environmental factors that affect the spatial distribution of benthic animals [42]. Spearman's rank correlation analysis was used to determine relationships between biological and environmental variables. At the time of analysis, the density of R. philippinarum was considered an environmental variable and was not included in the calculation of Bray-Curtis similarity. The biomass of macrobenthos included all benthic animals, and only the biomass of R. philippinarum was calculated separately. All analyses were performed using PRIMER 6 software with the PERMANOVA add on package [43].

Sediment Analysis
Sediment particle sizes were determined after treating samples with a solution of 10% hydrogen peroxide. Sediment samples were heated to >100 • C to evaporate the hydrogen peroxide and then washed at least three times with distilled water to remove organisms and salts. Washed samples were then passed through a 63-µm standard sieve. After drying, the sediments trapped by the sieve were weighed and subjected to automatic particle size analysis using a SediGraph 5120 device (Kunash Instruments, Mumbai, India) following the addition of sodium hexametaphosphate as a dispersing agent. We then calculated the average particle sizes and degrees of sorting. Sediments were categorized according to Folk's classification system [41]. The content of total organic carbon (TOC) in sediments was analyzed using a Shimadzu TOC analyzer, (SSM-5000A, Shimadzu, Japan).

Statistical Analyses
Density and biomass data are recalculated per square meter; statistical analyses were performed for all species. The Shannon-Wiener diversity index (H') was calculated using density data. Cluster and non-metric multi-dimensional scaling (nMDS) analyses on macrobenthic community data for each sampling period were analyzed using the Bray-Curtis similarity measure based on fourth-root transformed density data and group average linkage. A similarity profile (SIMPROF) permutation test was performed to determine the statistically significant clusters among the samples. A similarity percentage (SIMPER) analysis was used to determine the contribution of each species to similarity-dissimilarity among groups. A biota-environment matching (BIO-ENV) analysis was conducted to determine the environmental factors that affect the spatial distribution of benthic animals [42]. Spearman's rank correlation analysis was used to determine relationships between biological and environmental variables. At the time of analysis, the density of R. philippinarum was considered an environmental variable and was not included in the calculation of Bray-Curtis similarity. The biomass of macrobenthos included all benthic animals, and only the biomass of R. philippinarum was calculated separately. All analyses were performed using PRIMER 6 software with the PERMANOVA add on package [43].
We identified a total of 145 macrobenthos species. In April, July, and October 2011, 35, 45, and 47 species appeared, respectively. In 2012, the number of species detected increased from 40 in January to 55 species in April and declined to 39 in July and 36 species in October. On average, 34.8 ± 3.7 species were recorded in 2013, which increased to 40.8 ± 9.2 species in 2014 ( Figure 2).
Throughout the study period, the average macrobenthos biomass including Ruditapes philippinarum was 5990 ± 3837 g/m 2 ( Figure 2). The biomass was highest in July 2012 (13,914 g/m 2 ) and lowest in April 2014 (793 g/m 2 ). By year, 2012 was the highest (9684 ± 4511 g/m 2 ) and 2014 was the lowest (3333 ± 2960 g/m 2 ). The average Ruditapes philippinarum biomass was 5837 ± 3811 g/m 2 . The highest biomass was 11,291 g/m 2 , and the lowest biomass was 729 g/m 2 . The R. philippinarum biomass was similar in shape to the macrobenthos biomass ( Cluster analysis indicated two main groups (SIMPROF test, p < 0.001) (Figure 3). Group B was related to the October 2012, January 2013, and April 2013 sampling periods. The remaining sampling periods were all related to Group A, whereas in April 2011, it was separate from other periods. The SIMPER test indicated a 49.85% dissimilarity between Groups A and B (Table 2). A total of 15 species appeared, and of these, 6 species were polychaetes, 5 mollusks, and 4 crustaceans. The species making the largest contribution to this difference was Ampharete arctica Malmgren, 1866 (Polychaete), with an average dissimilarity of 1.86. The second most influential species was Musculus senhousia (Benson in Cantor, 1842) (Mollusca), with an average dissimilarity of 1.54. The third species was Crangon affinis De haan, 1849 (Crustacean), with an average dissimilarity of 1.21.     The BIO-ENV analyses and nMDS bubble plot showed the density and biomass of Ruditapes philippinarum, and the TOC showed the highest correlations with a macrofaunal composition (Rho = 0.32, P < 0.05; Table 4). However, the mean grain size (Mz) had relatively little effect (Table 3, Figure  4).   The BIO-ENV analyses and nMDS bubble plot showed the density and biomass of Ruditapes philippinarum, and the TOC showed the highest correlations with a macrofaunal composition (Rho = 0.32, p < 0.05; Table 4). However, the mean grain size (Mz) had relatively little effect (Table 3, Figure 4). Table 3. Environmental and biological variables affecting the macrobenthos community as determined by a biota-environment matching (BIO-ENV) analysis (Mz = mean grain size, TOC = total organic carbon, RP = Ruditapes philippinarum).  Table 3. Environmental and biological variables affecting the macrobenthos community as determined by a biota-environment matching (BIO-ENV) analysis (Mz = mean grain size, TOC = total organic carbon, RP = Ruditapes philippinarum).  Correlation analysis showed that the number of macrobenthos species was positively correlated with the Shannon diversity index ( Table 4). The diversity index had a negative correlation with Ruditapes philippinarum density. Additionally, macrobenthos biomass including R. philippinarum was positively correlated with R. philippinarum biomass. Two dominant species (Heteromastus filiformis and Ceratonereis erythraeensis) were positively correlated with the number of macrobenthos species; these two species are deposit and detritus feeders polychaetes, respectively, whereas R. philippinarum is a suspension feeder. Correlation analysis showed that the number of macrobenthos species was positively correlated with the Shannon diversity index ( Table 4). The diversity index had a negative correlation with Ruditapes philippinarum density. Additionally, macrobenthos biomass including R. philippinarum was positively correlated with R. philippinarum biomass. Two dominant species (Heteromastus filiformis and Ceratonereis erythraeensis) were positively correlated with the number of macrobenthos species; these two species are deposit and detritus feeders polychaetes, respectively, whereas R. philippinarum is a suspension feeder. Table 4. Spearman rank correlations between biological and environmental variables (Mz, TOC, macrobenthos species richness, biomass, and density, Shannon diversity: H', Ruditapas philippinarum (RP) density and biomass) (HF = Heteromastus filiformis, CE = Ceratonereis erythraeensis, AA = Ampharete arctica) (* = p < 0.05; ** = p < 0.01, *** = p < 0.001).

Discussion
Ruditapes philippinarum accounted for 37.6% of the total macrobenthos density and had greater biomass than all other species. Therefore, this species occupied more space than the other species in the study area and led to reduced macrobenthos species richness. The use of univariate measures of diversity (Shannon-Weaver diversity; H') is used to assess the level of stress in macrobenthos communities [44,45]. The index (H') was highly related to macrobenthos species richness [46]. Somerfield et al. (2009) [47] suggested that local species diversity is determined by disturbance, predation, and competition, and also by environmental structure and regional processes. In this study, the prosperity of the clams made the diversity index low. This is because a large increase in one species affects the entire benthic ecosystem. The correlation between the clams and other dominant species was not significantly related. This is because they do not compete for food, but coexist and survive.
Spatial patterns of dominant species allow us to understand the structure of target populations [48,49]. Choi (2003) [50] found that benthic groups at Gwangyang reflected the degree of dominance and regional distribution of dominant species. The proportion of the community occupied by dominant species plays a significant role in the overall community structure and provides a lens through which to interpret environmental conditions [51,52]. Although Ruditapes philippinarum was affected by multiple environmental factors (i.e., sediment type, temperature, salinity, organic carbon, etc.), it was most influenced by sediment composition [53]. Sediment characters are the key drivers of macrobenthos communities [54]. The proportion of sand in the substrate is important for Ruditapes philippinarum, which prefers substrates with a 50%-80% sand content [55]. The sediment in our study area had a sand content of 54%, which is suitable for R. philippinarum. Previous research in four areas in Gyeonggi Bay that had high densities of R. philippinarum found an average sediment grain size of 3.8 ± 0.1 ø [56], which was similar to the average grain size in our study area (4.1 ± 1.4 ø). The grain size and composition of sediment affect the lifecycle of benthic animals [57]. In a study conducted in Seonjaedo, Korea, Kim (2005) [58] reported an Mz of 3.47 ± 0.45 ø in the eastern part of the study area and 3.60 ± 0.34 ø in the western part. Thus, our study area had a sediment environment suitable for the clams to thrive and also had appropriate substrate conditions for clam growth.
A biota-environment matching (BIO-ENV) analysis indicated that the total organic carbon (TOC) influenced benthos communities. Analyzing environmental variables such as TOC is important for evaluating coastal marine ecosystems [59]. TOC is highly related to benthic food sources and is therefore associated with macrobenthic fauna [60]. Because the amount of clams is overwhelming, their presence has influenced the benthic fauna community, and it is necessary to examine the relationship with the dominant species other than the clams. The second most dominant species, the polychaete Heteromastus filiformis, is also a dominant species in other areas, such as Asan Bay and Gyeonggi Bay [61].