Advertisement

Ecological Research

, Volume 33, Issue 1, pp 271–281 | Cite as

Environmental factors affecting benthic invertebrate assemblages on sandy shores along the Japan Sea coast: implications for coastal biogeography

  • Yoshitake Takada
  • Naoto Kajihara
  • Hideki Sawada
  • Shota Mochizuki
  • Takuhiko Murakami
Original Article
  • 165 Downloads

Abstract

Assemblages of sandy shores are primarily structured by physical environment factors. This structuring provides a unique opportunity to evaluate biogeographic regions. In this study, the shallow subtidal zone (0.2–1.2 m) of 39 sites of sandy shores along the Japan Sea coast of Honshu were surveyed using a sledge net to identify invertebrate assemblages and to elucidate their relationships with environmental factors and geographic distributions. In total, 78 taxonomic units were obtained and six clusters of assemblage were recognized according to the Morisita-Horn dissimilarity index values between the assemblages of these sites. Indicator taxonomic units were identified for the clusters and a distance-based redundancy analysis (dbRDA) demonstrated that a set of five environmental factors (slope angle of the swash zone, sediment grain size, average wave fetch, spring sea surface temperature, and summer Chlorophyll-a concentration) significantly explained variations of the assemblages. Geographical distributions of two of the clusters were localized and mutually exclusive (one in the north and one in the south), while the other four clusters were scattered along the coast. On the dbRDA ordination, these two clusters were plotted on opposite ends of the 1st axis on which spring sea surface temperature and summer Chlorophyll-a concentration showed high contributions. The spatial gap between the two clusters was located in an area between the Noto Peninsula and the Sado Island, central Honshu, which can be proposed as a boundary of geographic regions of sandy shore organisms along the Japan Sea coast.

Keywords

Sandy shore Subtidal dbRDA Geographic distribution Benthic community 

Introduction

Understanding patterns of distribution of organisms is a fundamental subject in ecology and biogeography (Briggs 1995; Gaston 2003). The Japan Sea is a marginal sea of the northwestern Pacific between the Japanese islands and the Asian Continent (Fig. 1), and it has been focused on for biogeographic researches because of its unique geographic position (Nishimura 1965, 1981, 1992; Kafanov et al. 2000; Lutaenko and Noseworthy 2014). The Japan Sea has four narrow and shallow straits that connect neighboring seas. The Tsushima Current, a branch of the warm Kuroshio Current, flows into the Japan Sea through the Tsushima Strait and most of the Tsushima Current water flows out to the Pacific through the Tsugaru Strait. Under the effect of the Tsushima Current, the southwestern coastal areas of the Japan Sea are assigned as a part of warm temperate region (WTe in Fig. 1), while the northern coastal areas are included in a cold temperate region (MTe and CTe in Fig. 1) (Nishimura 1981, 1992; Briggs and Bowen 2012; Lutaenko and Noseworthy 2014). However, the boundary between the two regions remained to be verified, and subdivisional boundaries (such as between MTe and CTe in Fig. 1) have also been proposed (Nishimura 1965, 1981, 1992; Kafanov et al. 2000).
Fig. 1

Location of the 39 study sites on the Japan Sea coast of Honshu. Inset shows the biogeographic regions proposed by Nishimura (1992): A arctic, SA subarctic, CTe cold temperate, MTe mid temperate, WTe warm temperate, and STr subtropical

The lack of ability to differentiate the boundary is mainly due to a gradual declining effect of the Tsushima Current from the south to the north and the strong seasonality of the Asian monsoon (Nishimura 1969; Naganuma 2000). In the shallow coastal area, cold water species do not survive in summer because of the warm Tsushima current. Larvae or juveniles of warm water species transported to the Japan Sea may settle in summer, but they may not survive the next winter because of the seasonally reduced flow of the Tsushima Current and the strong cold winds from the Asian continent (Nishimura 1969; Naganuma 2000). The probability of successful settlement and survival of warm water species is considered to decrease gradually from the south to the north along the coast of the Tsushima Current, resulting in a gradual decline of warm water species in the coastal assemblage (Nishimura 1965; Motoh 2008). Thus, focusing on the warm water species, the lack of abrupt changes in assemblages makes it difficult to set a boundary between cold and warm regions on the coast of the Japan Sea. On the other hand, species with low dispersal ability have a possibility to show a better correspondence with the regional environment, resulting in a clear boundary of distribution. Peracarid crustaceans on sandy shores are one of the candidates that may show regionally characterized assemblages.

Sandy shore peracarids have two features enabling studies of environmental effects on assemblages: specificity to environmentally controlled habitats and a low dispersal ability. Assemblages on sandy shores are generally controlled by environmental factors, because of effects of harsh and unstable environments, including wave and wind, which override competitive interactions among sandy shore inhabitants. Furthermore the lack of sessile organisms precludes facilitative interactions by forming microhabitats for other organisms (McLachlan and Brown 2006). In addition, sandy shore peracarids potentially do not disperse over long distances, because females carry embryos in a brood pouch and juveniles develop directly without planktonic dispersal stages, and adults usually burrow into the sandy sediment. These characteristics provide a unique opportunity to examine relationships between assemblages and environmental factors (Lastra et al. 2006; Cisneros et al. 2011; Barboza et al. 2012; Schlacher and Thompson 2013; Rodil et al. 2014; Barboza and Defeo 2015; Takada et al. 2015b) and, also, to monitor effects of climate-induced environmental changes (Schoeman et al. 2014). On the Japan Sea coasts, peracarid crustaceans are one of the dominant components in sandy shore assemblages (Takada et al. 2016), and close relationships between environmental factors and densities of peracarids have been recognized in a regional scale (270 km) (Takada et al. 2015b). But these studies (Takada et al. 2015b, 2016) revealed that assemblages in intertidal sandy shores were less diverse and not suited to examine geographic variations. Studies on assemblages in subtidal sandy shores over a geographic scale (> 1000 km) may extend our insights into their geographical distribution and effects of environmental factors on Japan Sea coastal assemblages.

In this study, we examined assemblages of invertebrates, including peracarids, in the sandy shore subtidal zone in relation to environmental factors over an extensive geographic scale. Our specific aims were (1) to identify assemblages of sandy shore invertebrates along the coast of the Japan Sea, (2) to assess spatial patterns in assemblage, (3) to compare these patterns to previously described biogeographic regions, and (4) to identify relationships between assemblage and environmental characteristics of sandy shores, in order to provide insights on the coastal biogeography.

Materials and methods

Study sites and sampling

Field surveys were carried out at 39 sites of sandy shores along the Japan Sea coast of Honshu during summer (June and July) from 2010 to 2012 (Fig. 1, Table S1). The survey sites covered a total distance of 1145 km. The Japan Sea coast of this area is characterized by calm weather in summer and contrasting harsh wind-driven waves in winter (Naganuma 2000). The mean tidal range during spring tides is only about 0.3 m, except for the Hibiki-nada area (Sites 38 and 39, Fig. 1), southwestern part of the coast, where the range is 0.8 m. Because of this micro-tidal condition, the range of daily tidal fluctuations is smaller than the seasonal fluctuation of the mean sea level (Naganuma 2000). Nishimura (1992) updated his biogeographic regions of coastal animals around the Japanese islands by summarizing his earlier works (Nishimura 1969, 1981). On his schema, the present study spanned the mid temperate (MTe) and warm temperate (WTe) zones (Fig. 1).

At each site, two sampling lines were set perpendicular to the shoreline from a nearshore point of 0.2 m depth to an offshore point of 1.2 m depth, then the latitude and the longitude were recorded using a global positioning system (etrex-Vista-HCx, Garmin, Kansas). The length and depth of the sampling lines varied according to the bottom slope and the wave height at the sampling time: the offshore point was set close to the shoreline at the 15 m from the nearshore point when the bottom slope was gentle, or shorter if the waves were high (Table S1). Epibenthic and demersal animals along the lines were sampled by a sledge net. The sledge net had a rectangular opening (40 cm width and 25 cm height) with a mesh bag (0.76 mm mesh aperture) and a tickler-chain to stimulate an escape response in surface dwelling animals (Hirota et al. 1989). The sledge net was pulled by two persons about 5 m apart in order to avoid disturbing the sampling line. Any drifting algae and debris along the sampling line were removed manually before the sampling.

All invertebrates in the net were preserved in a plastic bag with 70% ethanol. In the laboratory, each sample was sorted out and identified to the lowest possible taxonomic level by using a binocular microscope (SZX7, Olympus, Tokyo). Amphipods were categorized to the family level, except for Dogielinotidae. After pooling the data of the two sampling lines, the identified species and the lowest possible taxonomic categories were used as taxonomic units for the following analyses.

Environmental variables

At each sampling site, seawater temperature and salinity in the swash zone were measured with salinometers (YSI-30 or YSI-63, YSI, Yellow Springs). Slope angle was measured in the swash zone with an inclinometer (Shinwa Rules, Sanjo). Sediment samples from up to 3 cm in depth were obtained from the swash zone to estimate the median grain size in phi-scale [ø = − log2 (grain size in mm)] and sorting coefficient [(ø16 − ø84)/2, where ø16 and ø84 are 16 and 84% percentile values of grain size in phi-scale, respectively] by using a laser diffraction particle size analyzer (SALD-3100, Shimazu, Kyoto). Among these, salinity (Sal), slope angle of the swash zone (SwSl), median of the sediment grain size in phi-scale (MdPh), and sorting coefficient (SoPh) were used as environmental variables in the following multivariate analysis. Latitude was not included in the analysis because of its high correlation to other variables.

Geographical and oceanographic variables were processed using a geographic information system (GIS, ArcView 9, ESRI). Seven variables representing geographical conditions of the shore were used as environmental variables. Within them, six variables, the presence/absence of artificial protective structures (APS) (such as breakwaters, groins, and seawalls), the length of the sandy shore (Len), the distance to the nearest river mouth from the sandy shore (DisR), the proportion of urbanized area within a 1-km buffer zone of the sandy shoreline (UrbA), the proportion of paddy fields within a 1-km buffer zone of the sandy shoreline (PadA), and the average wave fetch distance (AWF), were obtained from The Digital Map 25000 (Geospatial Information Authority of Japan) and the 5th Coastline alteration survey (Ministry of the Environment, Japan). AWF is a topographical index of wave exposure that was calculated as the average wave fetch distance of 16 angular sectors up to a distance of 200 km (Burrows et al. 2008). The AWF values were first obtained for 200-m grid sector of the coastline and then averaged for the corresponding sandy shores. Another geographical variable, the mean slope (degree) of the offshore seafloor (OfSl), was obtained within a 1-km buffer zone of the sandy shoreline using a 30-m grid of the depth data from the Bathymetric Charts (Basic Maps of Sea, Japan Hydrographic Association). Low value of OfSl represents an extensive shallow sea floor area off the shoreline.

In addition, Chlorophyll-a concentration (mg m−3) and sea surface temperature (SST) (°C) were obtained from Terra/MODIS (ECHO’s Reverb, https://reverb.echo.nasa.gov/reverb/). For each sandy shore containing the sampling sites, values of Chlorophyll-a and SST within a 5-km buffer zone of the shoreline were averaged for each 3-month period from March to August 2010 as a typical year. Because of a high collinearity between these variables, three variables, spring SST (SpT), summer SST (SmT), and summer Chlorophyll-a (SmCh), were used for the later analysis. Definitions and procedures of data processing of these environmental variables were the same as in the previous studies (Takada et al. 2015b, 2016).

Data analyses

For each site, richness (number of taxonomic units) and exponential Shannon diversity {expH′ = exp[− Σpiloge(pi)], where pi is the proportion of abundance for species i} were estimated. Assemblages of benthic invertebrates were analyzed using multivariate techniques. Abundance data of the taxonomical units were used to estimate similarity values of assemblages between all pairs of the sites by using the Morisita-Horn similarity index (Jost et al. 2011). The complement of the similarity index is the dissimilarity value of the Morisita-Horn index. Based on the site-by-site dissimilarity values, ordinations in reduced spaces were generated by nonmetric multidimensional scaling (nMDS) and the assemblages were clustered into several groups by K-medoids (Borcard et al. 2011). The number of clusters was decided by maximizing the average silhouette width, allowing one cluster with low silhouette values representing a cluster including outlier sites.

Indicator taxonomical units were selected for the clusters by using the indicator value (IndVal). The IndVal for each taxonomical unit was obtained as the product of fidelity (relative frequency in the identified assemblage) and specificity (concentration of abundance measured as mean abundance in the assemblage against the sum of mean abundance over all assemblages) (Dufrêne and Legendre 1997). Significance of a taxonomical unit as an indicator of an estimated cluster of assemblages was evaluated using a randomization procedure (n = 1000).

The relationship between the assemblage and the environmental variables was analyzed by a constrained ordination technique using distance-based redundancy analysis (dbRDA) (Legendre and Anderson 1999). The explanatory variables were the 14 environmental variables (Sal, SwSl, MdPh, SoPh, APS, Len, DisR, UrbA, PadA, AWF, OfSl, SpT, SmT, and SmCh) and the dependent data were the assemblages of the 39 sites. The Morisita-Horn dissimilarity values were used as the distances between the assemblages. The environmental variables were standardized (mean = 0, SD = 1) to facilitate assessment of their relative importance. Before the analysis, Len and SmCh were log(10) transformed, and DisR, UrbA, PadA, and OfSl were square-root transformed to minimize skewness of the distribution of the variables. Multicollinearity of these variables was examined by the variance inflation factor (VIF) values that were less than 3.2. After analyzing the full model (14 variables), reduced models were analyzed by removing variables one-by-one by their permutation P values. These analyses were performed using R 3.3.2 software and packages cluster (pam) (Maechler et al. 2017), labdsv (indval) (Roberts 2016), and vegan (dbrda, metaMDS, and ordistep) (Oksanen et al. 2017). A biplot of the dbRDA results was produced using the vegan’s default setting.

Results

Overall 39 sites of the present study, 28960 individuals of benthic invertebrates were sampled. Removing juveniles and damaged samples, 28626 individuals were analyzed and 78 taxonomic units were recognized, including 60 crustaceans and 13 molluscans taxonomic units. Among them, dogielinotid amphipod Haustorioides japonicus was the most abundant (Table 1), followed by mysids Archaeomysis vulgaris and Archaeomysis kokuboi. The ten most abundant taxonomic units were all peracarids, consisting units of 5 amphipods and 5 mysids. On the other hand, 11 taxonomic units consisted of only a single specimen, and 22 taxonomic units occurred in only a single site (Table S2).
Table 1

Ten most abundant taxonomic units over all the samples

Rank

Taxonomic unit

Number of individuals (%)

Number of sites

1

Haustorioides japonicus

8210 (28.7)

25

2

Archaeomysis vulgaris

5358 (18.7)

21

3

Archaeomysis kokuboi

3981 (13.9)

18

4

Pontogeneiidae

2505 (8.8)

27

5

Atylidae

1708 (6.0)

22

6

Phoxocephalidae

1601 (5.6)

12

7

Archaeomysis japonica

1056 (3.7)

21

8

Orientomysis tamurai

782 (2.7)

2

9

Nipponomysis imparis

553 (1.9)

9

10

Oedicerotidae

475 (1.7)

21

The richness (number of taxonomic units) at each site ranged from 3 (Site 9) to 20 (Site 4) and the diversity (expH′) ranged from 1.31 (Site 2) to 9.92 (Site 4). The richness and the diversity did not show any significant correlations with latitude (df = 37, P > 0.05). Result of the K-medoids analysis on the assemblage of 39 sites showed 6 clusters (Table 2). Five of these clusters had high silhouette values (Clusters 2–6) representing high resemblance of assemblages within each clusters, but one cluster of the low silhouette value (Cluster 1, − 0.014) showed that it included a variety of assemblages. The two-dimensional ordination plot of nMDS showed a fairly good representation (stress = 0.20) (Fig. 2). On the plot, all sites of the same clusters were placed close together except for the sites belonging to Cluster 1.
Table 2

Indicator taxonomic units for the 6 clusters

 

Number of sites

Average silhouette width

Indicator taxonomic units (IndVal) (permutation P < 0.05)

Cluster 1

5

− 0.014

 

Cluster 2

4

0.298

Haustorioides japonicus (0.96)

Cluster 3

3

0.450

Atylidae (0.88), Pycnogonida (0.62), Ischyroceridae (0.61), Anisogammaridae (0.51), Parhyalella sp. (0.51), Caprellidae (0.47)

Cluster 4

4

0.551

*

Cluster 5

14

0.615

Archaeomysis kokuboi (0.92)

Cluster 6

9

0.725

Archaeomysis vulgaris (0.60)

Total

39

0.508

 

*Pontogeneiidae (0.77) showed a marginal significance (P = 0.052)

Fig. 2

Ordination by nMDS of the assemblage of sandy shore invertebrates of 78 taxonomic units. Sites are identified with the number and the six clusters of assemblages are indicated

Geographically, sites with the assemblages of Cluster 5 occurred on the northern part of the Japan Sea coastline and those of Cluster 6 occurred on the southern part (Fig. 3). The geographic ranges of these two clusters did not overlap. But the other clusters (Clusters 1–4) were widespread over the study area and did not show specific geographic distributions. Analysis of IndVal revealed that four of the six clusters had significant (P < 0.05) indicator taxonomical units (Table 2). Nine taxonomic units were designated as indicators, and among them H. japonicus showed the highest indicator value (0.96, for Cluster 2). Two species of Archaeomysis were indicators of different clusters: A. kokuboi for Cluster 5 and A. vulgaris for Cluster 6. Pontogeneiidae showed a high indicator value (0.77) for Cluster 4 but it was marginally significant. These clusters are characterized by the occurrence of these indicators.
Fig. 3

Geographical distribution of the clustered assemblages

Measurements of the 14 environmental variables showed that the study sites covered a wide range of environmental gradients (Table 3). Some pairs of the environmental variables showed significant correlation with each other (Table 3), but the VIF of the environmental variables was low. Correlations with latitude showed that in higher latitude SpT (spring sea surface temperature) and OfSl (slope angle of the offshore seafloor) decreased and SmCh (summer Chlorophyll-a concentration) increased. In the full model dbRDA with the 14 environmental variables, the constrained inertia (sum of squared distance) was 7.87 (57.1% of the total inertia). Eigenvalues of the first and second axes were larger than 1. Five environmental variables (SwSl, MdPh, AWF, SpT, and SmCh) were selected for the reduced model dbRDA (Table 4) and the constrained inertia was 6.04 (43.8% of the total inertia). Only the eigenvalue of the first axis was larger than 1 but a permutation test showed that up to the third axes was significant (P < 0.05). Coefficients of the environmental variables showed that contribution of SpT was the largest followed by SmCh in the 1st axis both for the full and the reduced models (Table 4). In the second axis, UrbA (proportion of urbanized area) showed the largest contribution in the full model, while MdPh was the largest in the reduced model. The constrained ordination constructed by the reduced model dbRDA (Fig. 4) showed a similar configuration pattern to the unconstrained ordination of nMDS (Fig. 2). On the dbRDA ordination (Fig. 4), sites of Cluster 5 occur on the positive end of the 1st axis indicating low spring SST and high summer Chlorophyll-a concentration, and sites of Cluster 6 are on the negative end. Other sites are close to the origin of the 1st axis. Sites of Cluster 4 are on the positive end of the 2nd axis indicating a high slope angle of the swash zone and large median grain size, and sites of Cluster 2 show the opposite characteristics except for Site 24. Although it is not clear on the two-dimensional plot, sites of Cluster 1 are characterized to be wave sheltered (low AWF) that showed the largest contribution on the 3rd axis (Fig. S1).
Table 3

Summary of the 14 environmental variables

Environmental variables

Mean ± SD (n = 39)

Min.

Max.

Significant correlation (df = 37, P < 0.05)

Correlation with latitude

Sal (psu)

28.64 ± 4.87

12.8

33.9

+: DisR

0.003

SwSl (degree)

8.14 ± 3.14

3.0

14.7

−: MdPh

− 0.108

MdPh (ø)

1.57 ± 0.58

− 0.14

2.43

−: SwSl

0.020

SoPh (ø)

0.41 ± 0.15

0.25

0.90

+: UrbA

−: AWF, APS

− 0.257

APS

0.31

  

−: SoPh

0.003

Len (km)

3.54 ± 3.87

0.28

13.96

+: PadA, DisR, AWF, SmCh

−: SpT, SwSl

0.428*

DisR (km)

2.48 ± 2.53

0.11

12.33

+: Len, Sal

0.222

UrbA (%)

9.44 ± 9.74

0.00

34.69

+: SoPh

−: AWF

− 0.374*

PadA (%)

1.74 ± 2.06

0.00

8.10

+: SmCh, Len

−: SpT, SmT

0.497*

AWF (km)

42.71 ± 19.94

1.76

72.09

+: Len

−: SoPh, UrbA

0.307

OfSl (degree)

0.83 ± 0.48

0.25

2.51

+: SpT

−: Len, SmCh

− 0.571*

SpT (°C)

10.45 ± 1.89

7.49

14.00

+: OfSl

−: SmCh, PadA, Len

− 0.927*

SmT (°C)

25.24 ± 1.50

22.00

27.51

−: PadA

− 0.407*

SmCh (mg/m3)

1.23 ± 1.51

0.13

6.38

+: PadA, AWF, Len

−: SpT, OfSl

0.571*

Sal Salinity, SwSl slope angle of the swash zone, MdPh median of the sediment grain size in phi-scale, SoPh sorting coefficient of the sediment grain in phi-scale, APS presence or absence of artificial protective structures, Len length of the sandy shore, DisR distance to the nearest river mouth, UrbA proportion of urbanized area, PadA proportion of paddy fields, AWF average wave fetch distance, OfSl slope angle of the offshore seafloor, SpT spring sea surface temperature, SmT summer sea surface temperature, SmCh summer Chlorophyll-a concentration

*P < 0.05

Proportion of the APS present that is given for the binary variables

Table 4

Coefficients of full model (14 variables) and reduced model (5 variables) dbRDA up to 3rd axes

Environmental variables*

Full model

Reduced model

1st axis

2nd axis

3rd axis

1st axis

2nd axis

3rd axis

Sal

− 0.0063

− 0.0344

0.0510

   

SwSl

0.0121

0.0327

0.1264

0.0001

− 0.1026

− 0.0141

MdPh

− 0.0100

− 0.0690

0.0136

− 0.0191

0.0832

0.0520

SoPh

0.0105

− 0.0439

0.0515

   

APS

− 0.0014

0.0394

0.0028

   

Len

0.0028

− 0.0356

− 0.0232

   

DisR

0.0183

0.0451

− 0.0038

   

UrbA

0.0170

0.0938

− 0.0320

   

PadA

− 0.0356

0.0248

− 0.0907

   

AWF

− 0.0127

− 0.0762

0.0923

− 0.0412

0.0752

− 0.1420

OfSl

0.0008

− 0.0130

0.0615

   

SpT

− 0.1238

− 0.0572

− 0.0143

− 0.1281

0.0317

− 0.0358

SmT

− 0.0320

0.0229

− 0.0569

   

SmCh

0.0715

− 0.0585

0.0750

0.0664

0.0109

− 0.0155

Eigenvalue

4.2556

1.0807

0.9886

4.0507

0.8910

0.6140

% explained

30.9

7.8

7.2

29.4

6.5

4.5

Eigenvalues and explained percentage of variations are also listed

*Abbreviations of environmental variables are shown in Table 3

Fig. 4

Ordination by dbRDA based on biplot projections of the five environmental variables: SpT spring sea surface temperature, SmCh summer Chlorophyll-a concentration, SwSl slope angle of the swash zone, MdPh median of the sediment grain size in phi-scale, and AWF average wave fetch distance. Sites are identified with the number and the six clusters of assemblages are indicated

Discussion

The present study recognized six clusters of invertebrate assemblages along the coast of the Japan Sea. One of the clusters (Cluster 1) was a mixture of a variety of assemblages, which was represented by a low silhouette value. By allowing one cluster as a group including outlier sites, the present study successfully detected the other five, well-united, clusters of assemblages by the K-medoids method that is known to be more robust for outliers than the K-means method (Berkhin 2006). Thus in the present study, all the clusters, except for Cluster 1, can be characterized by the ordination on the environmental gradients of the dbRDA and by their indicator taxonomic units. Significant indicators recognized in the present results were all peracarid crustaceans, except for Pycnogonida, indicating the suitability of peracarids as indicators of sandy shore assemblages compared to the other taxonomic groups, such as decapods or bivalves.

Characteristics of the clustered assemblages

Cluster 2 was characterized by the occurrence of Haustorioides japonicus. Although this species showed the highest value of IndVal, it was the most prevalent species observed at 25 sites (64% of all the sites). The previous study on the intertidal zones (Takada et al. 2016) demonstrated that H. japonicus occurred over the entire geographic range of the Japan Sea coast of Honshu and its main vertical distribution range is in the swash zone (between the limits of the run-up and the run-down of waves). The present study, which focused on the subtidal zone of sandy shores, indicated that H. japonicus extended its vertical distribution range to the subtidal zone in the sites of Cluster 2. Two major mechanisms can be proposed: (1) some individuals spilled over into the subtidal zone because of the high density in the swash zone, and (2) some individuals were dislocated to the subtidal zone because of strong waves and/or sediment reworking. In both cases, spilled over or dislocated H. japonicus were more abundant than the other invertebrates selected as indicators. On the dbRDA plot, Cluster 2 showed an association with finer grain size of the sediment substrate (higher value of MdPh), indicating that the range extension of the H. japonicus to the subtidal zone tends to occur on shores with finer sand.

Indicators of Cluster 3 included Ischyroceridae, Anisogammaridae, and Caprellidae. Some species of these amphipods are known to associate with drifting algae and debris, including terrigenous leaf litter (Sano et al. 2003; Sakurai and Yanai 2006). So possibly, Cluster 3 can be characterized as an associated assemblage with drifting algae and/or debris. Stranded algae and debris are well known to affect invertebrate assemblages of sandy shores as food source and refuges from predation and environmental stresses (Colombini and Chelazzi 2003; Kochi and Yanai 2006). Because these drifting algae and debris are variable in space and time floating on the shoreline back and forth before they were stranded, associated assemblages would also be transient.

Pontogeneiidae was the indicator of Cluster 4, although it was marginally significant. Because only four sites were assigned to Cluster 4, increased number of survey sites may improve statistical power to clarify the probability of significance. Cluster 4 showed the association with steeper slope of the swash zone and larger median grain size. This result is concordant with observations that pontogeneiid amphipods occur on sandy substrates with medium to coarse grain size (0.25–1.0 mm) (Jewett et al. 2010). A species of Pontogeneiidae (Pontogeneia rostrata) is known to be the dominant amphipods on a sandy shore of southern Korea (Yu et al. 2002), indicating that Cluster 4 may also occur on the coasts of Korean Peninsula.

Clusters 5 and 6 showed contrasting characteristics, represented by the 1st axis of the dbRDA ordination. Cluster 5 occurred only on the northern sites exposed to the lower spring SST and the higher summer Chlorophyll-a concentration, while Cluster 6 showed the opposite trend. There is a spatial gap between the geographic distributions of Clusters 5 and 6 at around Noto Peninsula and Sado Island (Fig. 3). This spatial division corresponds to the geographic distribution of their indicators: A. kokuboi for Cluster 5 and A. vulgaris for Cluster 6. It is known that A. kokuboi occurs in the north and A. vulgaris in the south of the Japan Sea coasts with little overlaps in this area (Hanamura 1997; Takada et al. 2016).

Geographical distribution of the clusters

Along the Japan Sea coast, only two of the six clusters (Clusters 5 and 6) showed geographically localized distributions, while the other four were widely scattered. This result shows that biogeographic subdivisions along the Japan Sea coast do not have a clear boundary, as discussed by the previous studies (Nishimura 1981, 1992; Lutaenko and Noseworthy 2014). Focusing on the scattered distributing clusters (Clusters 1–4), it might be possible to consider that the Japan Sea coasts is a single biogeographic region as also proposed by the Marine Ecoregions of the World (MEOW) (Spalding et al. 2007). However, the geographic boundaries between the six clustered assemblages were not perfectly concordant. The present results highlight that the area between the Noto Peninsula and Sado Island was the biogeographical boundary between the northern Cluster 5 and the southern Cluster 6. Thus, the Japan Sea coast can be subdivided at this area according to the invertebrate assemblages of subtidal sandy shores. Takada et al. (2016), which examined the intertidal zonation of sandy shores along the Japan Sea coast, demonstrated that the southern limit of the assemblage characterized by A. kokuboi was Mase (Site 16 in the present study), the same as the southern limit of Cluster 5 in the present result. Kafanov’s schemata of biogeographic regions in the Japan Sea based on coastal bivalves (Kafanov and Volvenko 1997) and fish (Kafanov et al. 2000) also recognized the boundary around the Noto Peninsula, but additional boundary around the Oga Peninsula was proposed. Lutaenko and Noseworthy (2014) extended the transitional boundary from the Noto Peninsula to the Tsugaru Strait by combining the two boundaries at the Noto and Oga peninsulas. The present results further emphasized the importance of the area around the Noto Peninsula as a biogeographical boundary. Contrary, this boundary is different from the biogeographic regions proposed by Nishimura (1992) (Fig. 1) and Briggs and Bowen (2012) that set a boundary in Shimane Prefecture (around Site 32 and Site 35, respectively).

Presence of physical barriers, such as land-bridges or ocean currents, produces a sympatric range limit of coastal species (Briggs 1995; Gaston 2003), which is recognized as a clear boundary of biogeographic regions. Along the Japan Sea coastline, declining effects of the Tsushima current and monsoonal changes of climate do not produce a sympatric range limit of coastal species (Nishimura 1969). Differences in positions of boundaries of these biogeographic schemata may be explained by several reasons. First, various definitions of the biogeographic regions are used in different studies. MEOW schema (Spalding et al. 2007) is a hierarchical classification system of coastal regions based on a combination of various information including geographic range of organisms, geomorphological features, currents, and climate regimes. Briggs and Bowen (2012) focused on the proportion of endemic species, while similarity of faunal compositions (presence/absence) was focused on in some other studies (Nishimura 1992; Kafanov et al. 2000; Lutaenko and Noseworthy 2014). The present study also used the similarity of faunal compositions but employed the abundance data of assemblages. Presence/absence data tend to bias the importance of rare species in assemblage analyses (Jost et al. 2011) and may emphasize transient occurrence of tropical species transported by the Tsushima Current, regardless of their abundance. So, it enables recognition of southwestern coastal areas of the Japan Sea as the warm temperate region (Nishimura 1992; Kafanov et al. 2000).

Second, phylogenetic and habitat biases may influence the recognition of biogeographic boundaries. Sandy shores are not distributed uniformly along the coast of the Japan Sea, although they are widespread (Koike 1977). Because the Noto Peninsula is one of the largest areas dominated by rocky shores along the Japan Sea coast, distribution gaps of sandy shore invertebrates, which are dominated by poor dispersal peracarids, tend to occur at such habitat gaps rather than in areas of continuous habitats. On the contrary, coastal fishes are thought to be more sensitive to seawater flows (Nishimura 1969; Kafanov et al. 2000) because of their dispersal ability along the coastal waters.

Third, taxonomic resolution is known to affect the results of multivariate analyses of assemblages (Anderson et al. 2005; Hirst 2006). In the present study, amphipods were classified up to family level except for Dogielinotidae. It may be possible that some species of the same family show geographically segregated distributions, as in the case of Archaeomysis observed in the present result. Overall, results of this study provided insights for biogeography of the Japan Sea coast by focusing on sandy shore invertebrates. Appropriate use of biogeographic information increases model prediction in sandy shore biodiversity (Barboza and Defeo 2015). Various methods have their merits in developing new research concepts and practical strategies for coastal management.

Environmental factors on the sandy shore assemblages

Due to the ongoing increase of SST associated with global climate change, some marine species are extening or shifting their range of distribution to higher latitudes (Perry et al. 2005; Poloczanska et al. 2016). The Japan Sea coastline is one of the areas facing rapid environmental changes (Belkin 2009; Japan Meteorological Agency 2016). Although species distribution may not respond simply to climate change, the present results of the dbRDA indicate some insights for the effects of environmental factors.

On the 1st axis of the dbRDA, increase of the spring SST and decrease of summer Chlorophyll-a concentration indicates a negative effect on Cluster 5 and a positive effect on Cluster 6. Various studies demonstrated that sandy shore assemblages are affected by SST (Barboza et al. 2012; Schlacher and Thompson 2013; Rodil et al. 2014; Barboza and Defeo 2015) and Chlorophyll-a concentration (Lastra et al. 2006; Cisneros et al. 2011; Rodil et al. 2014). In contrast, a previous study on three peracarids species on sandy shores around Niigata (corresponding to sites from 12 to 18 of the present study), the central area of the Japan Sea coast (Takada et al. 2015b) demonstrated that the SST had marginally significant or insignificant effects on their density and that the Chlorophyll-a concentration was not significant. This discrepancy is probably due to differences in the spatial scale of the studies. Takada et al. (2015b) surveyed on an area of 270 km of coastline, while the present study surveyed wider geographic area (1145 km), enabling to cover enough range of variations in the SST and Chlorophyll-a concentration. On the 2nd axis of the dbRDA, increase of sediment grain size and slope angle indicates a negative effect on Cluster 2 and a positive effect of Cluster 4. Sediment grain size is also known as a primary factor affecting sandy shore assemblages (Lastra et al. 2006; Barboza et al. 2012; Schlacher and Thompson 2013; Rodil et al. 2014; Barboza and Defeo 2015). On the Japan Sea coast, Takada et al. (2015b) also showed that density of H. japonicus increased with decreasing sediment size (increasing phi scale), corresponding to the present result that showed the association of fine sediment and Cluster 2 which has an indicator of H. japonicus. Slope angle is also recognized as an important factor such the abundance of bivalve Donax semigranosus increased in decreasing slope angle in a previous study on the Japan Sea coasts (Takada et al. 2015a).

Thus, the results of the dbRDA indicate that these environmental factors have a high explanatory power to estimate the present distribution of the sandy shore assemblages. However, little is known about whether sandy shore animals will be able to change habitat with the increasing SST and shift their geographic distribution to higher latitude (Schoeman et al. 2014). Because of the poor dispersal ability of most peracarids, it is unlikely that they readily colonize to new favorable shores in higher latitudes as compared to other taxonomic groups with planktonic dispersal stages. The geographic distribution of the sandy shore assemblages shown in the present study is useful as baseline information to evaluate effects of the ongoing climate change. Recent increases of anthropogenic impacts on the global climate and ecosystem functioning, including over-exploitation of natural resources and introduction of alien species, affect the distribution of organisms (Perry et al. 2005; Cheung et al. 2013; Poloczanska et al. 2016). To evaluate these impacts and to develop better management practices, efforts for monitoring distribution of organisms are required with an aid of accumulated knowledge of biogeography (Spalding et al. 2007; Belanger et al. 2012; Barboza and Defeo 2015).

In summary, the present study recognized the six clusters of invertebrate assemblage on sandy shores of the Japan Sea coast. These clusters were characterized by indicator taxonomic units. Geographical distribution patterns of the clusters showed that the area between the Noto Peninsula and Sado Island is a possible boundary of biogeographic regions. Spring sea surface temperature and some other environmental factors showed significant effects on the variation of the assemblages.

Notes

Acknowledgements

We thank S. Abe, T. Iseki, T. Fujii and Y. Yagi for their help with the fieldwork, and H. Saito for his constructive comments on the manuscript. S. Ishimaru kindly assisted in the identification of some amphipod specimens.

Funding

Part of this study was supported by “A project for development of assessment methods for the coastal environment” (FY2010–2012, Fisheries Agency) to YT, NK, and HS and JSPS KAKENHI Grant Nos. 22510252 and 25340114 to YT and TM, and 15H02265 to YT.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary material

11284_2017_1553_MOESM1_ESM.pdf (225 kb)
Supplementary material 1 (PDF 225 kb)

References

  1. Anderson MJ, Connell SD, Gillanders BM, Diebel CE, Blom WM, Saunders JE, Landers TJ (2005) Relationship between taxonomic resolution and spatial scales of multivariate variation. J Anim Ecol 74:636–646.  https://doi.org/10.1111/j.1365-2656.2005.00959.x CrossRefGoogle Scholar
  2. Barboza FR, Defeo O (2015) Global diversity patterns in sandy beach macrofauna: a biogeographic analysis. Sci Rep-UK 5:14515.  https://doi.org/10.1038/srep14515 CrossRefGoogle Scholar
  3. Barboza FR, Gómez J, Lercari D, Defeo O (2012) Disentangling diversity patterns in sandy beaches along environmental gradients. PLoS One 7:e40468.  https://doi.org/10.1371/journal.pone.0040468 CrossRefPubMedPubMedCentralGoogle Scholar
  4. Belanger CL, Jablonski D, Roy K, Berke SK, Krug AZ, Valentine JW (2012) Global environmental predictors of benthic marine biogeographic structure. PNAS 109:14046–14051.  https://doi.org/10.1073/pnas.1212381109 CrossRefPubMedPubMedCentralGoogle Scholar
  5. Belkin IM (2009) Rapid warming of large marine ecosystems. Prog Oceanogr 81:207–213.  https://doi.org/10.1016/j.pocean.2009.04.011 CrossRefGoogle Scholar
  6. Berkhin P (2006) Survey of clustering data mining techniques. In: Kogan J, Nicholas C, Teboulle M (eds) Grouping multidimensional data: recent advances in clustering. Springer, Berlin, pp 25–71CrossRefGoogle Scholar
  7. Borcard D, Gillet F, Legendre P (2011) Numerical ecology with R. Springer, New YorkCrossRefGoogle Scholar
  8. Briggs JC (1995) Global biogeography. Elsevier, AmsterdamGoogle Scholar
  9. Briggs JC, Bowen BW (2012) A realignment of marine biogeographic provinces with particular reference to fish distribution. J Biogeogr 39:12–30.  https://doi.org/10.1111/j.1365-2699.2011.02613.x CrossRefGoogle Scholar
  10. Burrows MT, Harvey R, Robb L (2008) Wave exposure indices from digital coastlines and the prediction of rocky shore community structure. Mar Ecol Prog Ser 353:1–12.  https://doi.org/10.3354/meps07284 CrossRefGoogle Scholar
  11. Cheung WWL, Watson R, Pauly D (2013) Signature of ocean warming in global fisheries catch. Nature 497:365–368.  https://doi.org/10.1038/nature12156 CrossRefPubMedGoogle Scholar
  12. Cisneros KO, Smit AJ, Laudien J, Schoeman DS (2011) Complex, dynamic combination of physical, chemical and nutritional variables controls spatio–temporal variation of sandy beach community structure. PLoS One 6:e23724.  https://doi.org/10.1371/journal.pone.0023724 CrossRefGoogle Scholar
  13. Colombini I, Chelazzi L (2003) Influence of marine allochthonous input on sandy beach communities. Oceanogr Mar Biol Ann Rev 41:115–159Google Scholar
  14. Dufrêne M, Legendre P (1997) Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecol Monogr 67:345–366. https://doi.org/10.1890/0012-9615(1997)067[0345:SAAIST]2.0.CO;2Google Scholar
  15. Gaston KJ (2003) The structure and dynamics of geographic ranges. Oxford University Press, OxfordGoogle Scholar
  16. Hanamura Y (1997) Review of the taxonomy and biogeography of shallow-water mysids of the genus Archaeomysis (Crustacea: Mysidacea) in the North Pacific Ocean. J Nat Hist 31:669–711.  https://doi.org/10.1080/00222939700770341 CrossRefGoogle Scholar
  17. Hirota Y, Tominaga O, Kamiharako T, Kodama K, Sadakata T, Tanaka M, Furuta S, Kojima K, Koshiishi Y (1989) Geographical distribution of mysids in coastal waters in the Japan Sea. Contr Fish Res Jpn Sea 15:43–57 (in Japanese) Google Scholar
  18. Hirst AJ (2006) Influence of taxonomic resolution on multivariate analysis of arthropod and macroalgal reef assemblages. Mar Ecol Prog Ser 324:83–93.  https://doi.org/10.3354/meps324083 CrossRefGoogle Scholar
  19. Japan Meteorological Agency (2016) Climate change monitoring report 2015. http://www.jma.go.jp/jma/en/NMHS/ccmr/ccmr2015_high.pdf. Accessed 8 May 2017
  20. Jewett SC, Bodkin JL, Chenelot H, Esslinger GG, Hoberg MK (2010) The nearshore benthic community of Kasatochi Island, one year after the 2008 volcanic eruption. Arct Antarct Alpine Res 42:315–324.  https://doi.org/10.1657/1938-4246-42.3.315 CrossRefGoogle Scholar
  21. Jost L, Chao A, Chazdon RL (2011) Compositional similarity and β (beta) diversity. In: Magurran AE, McGill BJ (eds) Biological diversity: frontiers in measurement and assessment. Oxford University Press, Oxford, pp 66–84Google Scholar
  22. Kafanov AI, Volvenko IV (1997) Bivalve molluscs and Cenozoic paleoclimatic events in the northwestern Pacific Ocean. Palaeogeogr Palaeoclimatol Palaeoecol 129:119–153.  https://doi.org/10.1016/S0031-0182(96)00057-0 CrossRefGoogle Scholar
  23. Kafanov AI, Volvenko IV, Fedorov VV, Pitruk DL (2000) Ichthyofaunistic biogeography of the Japan (East) Sea. J Biogeogr 27:915–933.  https://doi.org/10.1046/j.1365-2699.2000.00438.x CrossRefGoogle Scholar
  24. Kochi K, Yanai S (2006) Decomposition of leaves in coastal brackish water and their use by the macroinvertebrate Anisogammarus pugettensis (Gammaridea). Mar Freshw Res 57:545–551.  https://doi.org/10.1071/MF05140 CrossRefGoogle Scholar
  25. Koike K (1977) The recent change of sandy shorelines in Japan. Komazawa Geogr Komazawa Univ Tokyo 13:1–16Google Scholar
  26. Lastra M, de La Huz R, Sánchez-Mata AG, Rodil IF, Aerts K, Beloso S, López J (2006) Ecology of exposed sandy beaches in northern Spain: environmental factors controlling macrofauna communities. J Sea Res 55:128–140.  https://doi.org/10.1016/j.seares.2005.09.001 CrossRefGoogle Scholar
  27. Legendre P, Anderson MJ (1999) Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecol Monogr 69:1–24. https://doi.org/10.1890/0012-9615(1999)069[0001:DBRATM]2.0.CO;2Google Scholar
  28. Lutaenko KA, Noseworthy RG (2014) Biodiversity and biogeography patterns of bivalve mollusks in the Sea of Japan. In: Song S, Adrianov AV, Lutaenko KA, Xiao-Xia S (eds) Marine biodiversity and ecosystem dynamics of the Northwest Pacific Ocean. Science Press, Beijing, pp 160–188Google Scholar
  29. Maechler M, Rousseeuw P, Struyf A, Hubert M, Hornik K (2017) Cluster: cluster analysis basics and extensions. R package version 2.0.6Google Scholar
  30. McLachlan A, Brown AC (2006) The ecology of sandy shores, 2nd edn. Academic Press, AmsterdamGoogle Scholar
  31. Motoh H (2008) Macruran decapods from the Sea of Japan—I. A check list based on literature. Rep Noto Mar Cent 14:13–27 (in Japanese with English summary) Google Scholar
  32. Naganuma K (2000) The sea of Japan as the natural environment of marine organisms. Bull Jpn Sea Nat Fish Res Inst 50:1–42 (in Japanese with English summary) Google Scholar
  33. Nishimura S (1965) The zoogeographical aspects of the Japan Sea part I. Publ Seto Mar Biol Lab 13:35–79CrossRefGoogle Scholar
  34. Nishimura S (1969) The zoogeographical aspects of the Japan Sea part V. Publ Seto Mar Biol Lab 17:67–142CrossRefGoogle Scholar
  35. Nishimura S (1981) Sea and life on the earth: an introduction to marine biogeography. Kaimei-sha, Tokyo (In Japanese) Google Scholar
  36. Nishimura S (1992) Guide to seashore animals of Japan with color pictures and keys. I. Hoikusha, Osaka (in Japanese) Google Scholar
  37. Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, O’Hara RB, Simpson GL, Solymos P, Henry M, Stevens H, Sozoecs E, Wagner H (2017) Vegan: community ecology package. R package version 2.0.6Google Scholar
  38. Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate change and distribution shifts in marine fishes. Science 308:1912–1915.  https://doi.org/10.1126/science.1111322 CrossRefPubMedGoogle Scholar
  39. Poloczanska ES, Burrows MT, Brown CJ, Molinos JG, Halpern BS, Hoegh-Guldberg O, Kappel CV, Moore PJ, Richardson AJ, Schoeman DS, Sydeman WJ (2016) Responses of marine organisms to climate change across oceans. Front Mar Sci 3:62.  https://doi.org/10.3389/fmars.2016.00062 CrossRefGoogle Scholar
  40. Roberts DW (2016) Labdsv: ordination and multivariate analysis for ecology. R package version 1.8-0Google Scholar
  41. Rodil IF, Compton TJ, Lastra M (2014) Geographic variation in sandy beach macrofauna community and functional traits. Estuar Coast Shelf Sci 150:102–110.  https://doi.org/10.1016/j.ecss.2013.06.019 CrossRefGoogle Scholar
  42. Sakurai I, Yanai S (2006) Ecological significance of leaf litter that accumulates in a river mouth as a feeding spot for young cresthead flounder (Pleuronectes schrenki). Bull Jpn Soc Fish Oceanogr 70:105–113Google Scholar
  43. Sano M, Omori M, Taniguchi K (2003) Predator–prey systems of drifting seaweed communities off the Tohoku coast, northern Japan, as determined by feeding habit analysis of phythal animals. Fish Sci 69:260–268.  https://doi.org/10.1046/j.1444-2906.2003.00616.x CrossRefGoogle Scholar
  44. Schlacher TA, Thompson L (2013) Environmental control of community organization on ocean-exposed sandy beaches. Mar Freshw Res 64:119–129.  https://doi.org/10.1071/MF12172 CrossRefGoogle Scholar
  45. Schoeman DS, Schlacher TA, Defeo O (2014) Climate-change impacts on sandy-beach biota: crossing a line in the sand. Glob Chang Biol 20:2383–2392.  https://doi.org/10.1111/gcb.12505 CrossRefPubMedGoogle Scholar
  46. Spalding MD, Fox HE, Allen GR, Davidson N, Ferdaña ZA, Finlayson M, Halpern BS, Jorge MA, Lombana A, Lourie SA, Martin KD, McManus E, Molnar J, Recchia CA, Robertson J (2007) Marine ecoregions of the world: a bioregionalization of coastal and shelf areas. Bioscience 57:573–583.  https://doi.org/10.1641/B570707 CrossRefGoogle Scholar
  47. Takada Y, Kajihara N, Abe S, Iseki T, Yagi Y, Sawada H, Saito H, Mochizuki S, Murakami T (2015a) Distribution of Donax semigranosus and other bivalves on sandy shore swash zones along the Japan Sea coast of Honshu. Venus 73:51–64.  https://doi.org/10.18941/venus.73.1-2_51 Google Scholar
  48. Takada Y, Kajihara N, Mochizuki S, Murakami T (2015b) Effects of environmental factors on the density of three species of peracarid crustaceans in micro-tidal sandy shores in Japan. Ecol Res 30:101–109.  https://doi.org/10.1007/s11284-014-1215-5 CrossRefGoogle Scholar
  49. Takada Y, Kajihara N, Iseki T, Yagi Y, Abe S (2016) Zonation of macrofaunal assemblages on microtidal sandy beaches along the Japan Sea coast of Honshu. Plankton Benthos Res 11:17–28.  https://doi.org/10.3800/pbr.11.17 CrossRefGoogle Scholar
  50. Yu OH, Soh HY, Suh HL (2002) Seasonal zonation patterns of benthic amphipods in a sandy shore surf zone of Korea. J Crust Biol 22:459–466. https://doi.org/10.1651/0278-0372(2002)022[0459:SZPOBA]2.0.CO;2Google Scholar

Copyright information

© The Ecological Society of Japan 2017

Authors and Affiliations

  1. 1.Japan Sea National Fisheries Research Institute, Fisheries Research and Education AgencyChuoJapan
  2. 2.National Research Institute of Fisheries and Environment of Inland Sea, Fisheries Research and Education AgencyHatsukaichiJapan
  3. 3.Maizuru Fisheries Research StationKyoto University, NagahamaMaizuruJapan
  4. 4.Graduate School of Science and TechnologyNiigata UniversityNishiJapan

Personalised recommendations