AQUATIC CONSERVATION: MARINE AND FRESHWATER ECOSYSTEMS
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014) Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/aqc.2495
Coral reef organisms as bioregion indicators off Halmahera, Moluccas, Indonesia ADRIAAN GITTENBERGERa,b,c,*, STEFANO G. A. DRAISMAd, UCU YANU ARBIe, VICTOR LANGENBERGf, PAUL L. A. ERFTEMEIJERf,g, YOSEPHINE TUTIh and BERT W. HOEKSEMAa a Department of Marine Zoology, Naturalis Biodiversity Center, Leiden, The Netherlands b Centre for Environmental Studies (CML) and Institute of Biology Leiden (IBL), Leiden University, Leiden, The Netherlands c GiMaRIS, Marine Research Inventory & Strategy Solutions, Leiden, The Netherlands d Institute of Ocean & Earth Sciences, University of Malaya, Kuala Lumpur, Malaysia e Research Centre for Oceanography (PPO), Indonesian Institute of Science (LIPI), North Sulawesi, Indonesia f DELTARES, Department of Water Quality and Ecology (WQE), Delft, The Netherlands g Sinclair Knight Merz (SKM), Perth, Australia h Research Centre for Oceanography (PPO), Indonesian Institute of Science (LIPI), Jakarta, Indonesia ABSTRACT 1. In the planning of marine protected areas for the conservation of coral reef systems, it is important to be able to distinguish between certain bioregions, i.e. regions with distinct species assemblages. This was done off western Halmahera (Moluccas, Indonesia), where three such bioregions were distinguished based on species inventories of 41 coral reef sites. 2. The relative value of species belonging to different trophic groups was examined with regard to their possible role as indicators for these three regions. The study focused on ascidians (Ascidiacea), macroalgae, mushroom corals (Fungiidae), and a selection of coral-associated gastropods (Epitoniidae and Coralliophilidae). 3. The best results for the detection of bioregions were obtained when datasets of all four trophic groups were pooled. When comparing the taxa and their indicator values, ascidians were the most suitable, followed by macroalgae, corals and gastropods with 98, 83, 71 and 66 % certainties, respectively. The occurrence of 17 species correlated strongly with the bioregions, which therefore were identified as potential indicator species consisting of 13 ascidians, three macroalgae, and one mushroom coral. These data suggest that ascidians have a significant value as indicators to evaluate bioregion boundaries. 4. Water quality measurements indicated that salinity and turbidity could be responsible for at least part of the differences between the species assemblages in the three bioregions. # 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Received 6 December 2013; Revised 11 June 2014; Accepted 4 July 2014 KEY WORDS:
coral; island; sublittoral; habitat mapping; marine protected area; conservation evaluation; algae; invertebrates; benthos
*Correspondence to: Adriaan Gittenberger, GiMaRIS, J.H. Oortweg 21, 2333 CH, Leiden, The Netherlands. E-mail:
[email protected]
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
A. GITTENBERGER ET AL.
INTRODUCTION Many coral reef species have an Indo-West Pacific distribution range and show maximum overlap in the Coral Triangle, the centre of maximum marine species diversity (Hoeksema, 2007). The exact ranges of many of these species have so far not been determined as a result of incomplete sampling (Dalleau et al., 2010). For an adequate sampling design to investigate the species richness of an area, it is necessary to recognize coral reef faunas with unique species compositions in so-called bioregions. If such bioregions are not taken into consideration, species with specific habitat requirements are likely to be overlooked. However, such species may be indicators of environmental conditions over time that otherwise are not easy to assess, such as current velocity and turbidity regimes (Scheffers et al., 2010). To support the designation of 36 000 km2 of New Caledonian coral reefs and lagoons as a UNESCO World Heritage Site, the environmental variability here had to be characterized first to demonstrate their natural value and diversity (Andréfouët and Wantiez, 2010). Similarly, one of the largest studies on the characterization of bioregions in coral reef systems was recently conducted for the conservation of the Great Barrier Reef as a World Heritage Area (Kerrigan et al., 2010). These bioregions were based on data on reef fishes, macroalgae, and hard and soft corals. Physical data were used as a surrogate for all other fauna and flora in the Great Barrier Reef because the range data of those species were found to be incomplete (Kerrigan et al., 2010). Dalleau et al. (2010) also advocated the use of habitats as surrogates for biodiversity in order to achieve coral reef conservation planning of Pacific Ocean islands, because little is known about the spatial distribution of coral reef species. With space-borne imagery for habitat mapping becoming increasingly available and affordable, such an approach is becoming ever more feasible and could appreciably facilitate and improve current coral reef conservation and enhance marine protected area (MPA) implementation. Surrogates for biodiversity and species are invaluable for assessing which habitats can be
linked to the presence of species and therefore which bioregions should be distinguished (Beger et al., 2007; Mumby et al., 2008; Dalleau et al., 2010; Kerrigan et al., 2010). In order to obtain reliable results and pinpoint the habitats and abiotic parameters that should be used as such surrogates, baseline studies must be done to link species to habitats; for which historical reference collections may be indispensable (Hoeksema and Koh, 2009; Rainbow, 2009; Van der Meij et al., 2009, 2010; Hoeksema et al., 2011). The present study is focused on distinguishing bioregions by the distribution of selected species of macroalgae, ascidians, mushroom corals and coral-associated gastropods during a coral reef survey off Halmahera, northern Moluccas in 2009. The main aim was to determine which coral reef bioregions can be distinguished on the basis of particular indicator species and/or species assemblages. In addition, it was investigated whether species from different trophic groups have equal value as indicators for these regions. Four model taxa were selected with different life-history strategies, dispersal capacities and reproduction strategies. Their selection also depended on the availability of taxonomic expertise in the research team. Although all these taxa are benthic, the bioregions that were distinguished may also be representative for pelagic species such as fish, as these often occur in close association with benthic taxa such as corals and macroalgae (Wilson et al., 2010). Potential correlations with basic water quality parameters were also investigated. MATERIAL AND METHODS Research localities A species inventory was carried out on 41 coral reef sites west off Halmahera from 23 October to 18 November 2009 (Figure 1, Supplementary Material Table 1S). West Halmahera was chosen because its marine biota are relatively under-explored (Tomascik et al., 1997) and its biogeographic position within the Coral Triangle was therefore unknown (Hoeksema and Van der Meij, 2010). Moreover, the area has recently become more easily accessible because of the establishment of a marine
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
CORAL REEF INDICATORS OF BIOREGIONS
obtaining a maximum variation in environmental parameters: exposure to waves, distance offshore, volcanic versus limestone substrate, depth of surrounding sea floor (compare Cleary et al., 2005; Waheed and Hoeksema, 2013, 2014). One dive was made per site during which four divers focused on scoring respectively the ascidians, macroalgae, coral, and coral-associated gastropod species present. The sites were classified by several parameters and characteristics, which were hypothesized to have influenced the present species assemblages. The four site-defining features used were (1) Region, i.e. TTH Channel (TTH = Ternate, Tidore and Halmahera), Halmahera bay and Guraici island, (2) Area, i.e. the islands Hiri, Ternate, Tidore, Maitara and Halmahera, (3) Bottom orientation, i.e. slope (more or less horizontal) shallow area and drop-off, and (4) Habitat, i.e. sandy with some coral, coral with sand patches, rubble, and muddy bay with coral (Supplementary Material Table 2S). Water quality parameters The salinity, pH and turbidity of replicate water samples, taken at the surface and taken near the bottom were measured at each site using an HI 9828 portable multimeter and an HI 93414 turbidity photometer (Hanna Instruments, Ann Arbor, USA). Species inventory Figure 1. Research area west of Ternate, indicating three bioregions: Ternate–Tidore–Halmahera (TTH) channel (green), Halmahera Bay (purple) and Guraici island (blue). The arrow on the map of Indonesia in the bottom right corner points at the research area.
field station with boat and diving facilities on the nearby volcanic island Ternate, which is operated by the Indonesian Institute of Sciences (Hoeksema and Van der Meij, 2010). The roving diving technique was used (Munro, 2005), as previously specifically applied to mushroom corals (Hoeksema and Koh, 2009). The reef sites were selected a priori using nautical charts 385 and 389 of the Indonesian Navy (Tentara Nasional Indonesia, Angkatan Laut, Jawatan Hidro-Oseanografi) depicting the central and southern part of West Halmahera with scale 1:200 000. Site selection was based on
A species inventory was made of four taxa: (1) macroalgae (Verheij and Prud’homme van Reine, 1993), (2) sea squirts (Tunicata: Ascidiacea) (Monniot and Monniot, 2001), (3) mushroom corals (Scleractinia: Fungiidae) (Hoeksema, 1989, 2014; Gittenberger et al., 2011), and (4) ecto- and endoparasitic snails (Gastropoda: Epitoniidae, Coralliophilidae) that live in association with mushroom corals and cup corals (Scleractinia: Dendrophylliidae) (Gittenberger et al., 2000; Gittenberger and Gittenberger, 2005, 2011; Gittenberger and Hoeksema, 2013). It was impossible to identify all sea squirts to species level, as many species have not been described and also because immature specimens of some species did not show diagnostic characters. If a specimen could not be identified, it was assigned to a morphotype. Whenever possible, specimens were photographed
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
A. GITTENBERGER ET AL.
underwater revealing diagnostic characters. Sampled specimens were photographed in the laboratory and then preserved as reference material. The mushroom coral species Cycloseris explanulata (Van der Horst, 1922) and C. wellsi (Veron and Pichon, 1980) were not included in the survey because they were not yet considered fungiids at that time (Benzoni et al., 2007, 2012). Voucher specimens were deposited in the zoological and herbarium collections of Naturalis Biodiversity Center, Leiden. The preservation methods were taxon-specific in order to preserve diagnostic morphological characters (Templado et al., 2010). Vouchers of algae were preserved either in 5% formalin diluted with seawater or in 96% ethanol. Subsamples were preserved in 96% ethanol for future DNA extractions. Statistical analyses The data were analysed with Primer-E v. 6.1.13 & PERMANOVA+ v. 1.0.3 software (Clarke and Gorley, 2006). All analyses were performed on the dataset of all species and also on each of the four datasets of the target taxa. To estimate the species richness of the area and to evaluate whether the number of survey sites was sufficient to yield a reliable overview of the species present, Michaelis–Menten species accumulation curves were calculated (Colwell and Coddington, 1994; Brose and Martinez, 2004). Using the recorded species absence/presence data, Sørensen distances (Sorensen, 1957) were calculated between all sites and then used to make MDS (multi-dimensional scaling) plots in order to analyse which site characters correlate best with the recorded coral reef communities. In addition, the Permanova test was conducted to check whether patterns discerned in MDS plots were significant. The Permanova analyses were done with the default settings in Primer-E v. 6.1.13: ‘main test’, ‘Permutation of residuals under a reduced model’, 999 permutations, Sums of Squares ‘Type III (partial)’. Two Permanova designs were used: one in which the two factors ‘Region’ and ‘Area’ were tested separately and one in which the factor ‘Area’ was tested nested within the factor ‘Region’. The latter was done only for the dataset including all species. A more
complete description of the Permanova method is given in Anderson (2001) and McArdle and Anderson (2001). The characteristics that were considered likely to correlate significantly with species assemblages were used in a canonical analysis of principal coordinates, i.e. a CAP (Anderson and Willis, 2003). This canonical analysis on principal coordinates is based on a symmetric distance matrix, including a test by permutation. The CAP analysis was done with the default settings of PERMANOVA+ v. 1.0.3, which includes the setting that the analysis determines the optimum number of canonical axes. The CAP analysis is discussed in detail in combination with the Permanova method by Irvine et al. (2011). To evaluate the performance of the CAP analyses, the ‘leave-one-out’ test was performed to examine to what degree the model resulting from the CAP analysis could successfully be used to identify a bioregion on the basis of its species community. The overlay vector function in Primer-E was used to illustrate the species occurrences with correlations >0.6 with the CAP graph to investigate which species serve as good indicators of the groups of sites separated in the CAP analyses. This function was also used to illustrate the correlations between the CAP graph and (1) species diversity, (2) salinity at the bottom and at the surface, (3) turbidity (in NTU) at the bottom and at the surface, and (4) pH at the bottom and at the surface. Finally, to ascertain the variation within the species assemblages that had been linked to a specific character, dissimilarity values were calculated with a Similarity Percentage analysis (SIMPER). This analysis Clarke (1993) calculates the average Sørensen dissimilarity between the selected communities. RESULTS Analyses of the complete dataset Presence/absence was recorded for 338 taxa. The Michaelis–Menten species accumulation curve (Figure 2) indicates that 365 taxa (Smax) would be expected in an infinite number of samples. In the MDS plot based on Sørensen’s distances, no distinct patterns were discernible, indicating that
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
CORAL REEF INDICATORS OF BIOREGIONS
Figure 2. Michaelis–Menten species accumulation curve (Smax = 365) based on the pooled dataset.
the species assemblages may be linked to the various areas, bottom orientations and habitats. Distinct clustering was observed in the MDS plot between the three bioregions. This was supported by the Permanova analysis, which found that the assemblages were significantly different in the three bioregions (df = 2, 159 pseudo-F = 2.32, P = 0.011; Supplementary Material Figure 1S): TTH channel, Halmahera Bay and the Guraici islands (Figure 1). SIMPER dissimilarities of these bioregions ranged from 59 to 61 (Table 1). The approximate contributions of the four taxa to these differences were Ascidiacea 40%, Macroalgae 35%, Scleractinia 20% and Gastropoda 5% (Table 1). Although the areas did not differ significantly in their species assemblages when all sites were included in the Permanova analyses, they did differ significantly from each other within the bioregions (Permanova, df = 4, pseudo-F =1.35, P = 0.049). The CAP analysis supported the Permanova results, indicating that the study sites could be classified into three distinct groups on the basis of their species assemblages (Figure 3(A)). Using the resulting CAP model and the ‘leave-one-out’ method it was found that the bioregions could be
identified with 100% certainty on the basis of their species assemblages. The occurrences of 17 species correlated > 60% with the CAP graph pattern (Figure 3(A)). Thirteen of these species were ascidians, three were macroalgae, and one was a mushroom coral. Correlations between the CAP model and species diversity, salinity, turbidity and pH are illustrated in Figure 3(B)–(E). The parameters turbidity and salinity were highly variable in surface water, while bottom water samples were much less variable (Supplementary Material Table 3S). On average, turbidity measured in water samples decreased with depth indicating greater water transparency and lower sedimentation at deeper sites (Table 3S). Salinity at virtually every locality was about 0.5 ppt less in surface water samples than in bottom water samples (Table 3S). Macroalgae In total, 98 macroalgal taxa (Supplementary Material Table 4S) were scored for presence/ absence at 38 sites (sites 16, 17 and 30 were not visited by the phycologist). These three sites were not included in analyses in which the macroalgal dataset was compared with the datasets of the other taxa. In addition to these macroalgal species, four seagrass species and one macroscopic golden alga (i.e. Chrysocystis fragilis C.S. Lobban, D. Honda and M. Chihara (Chrysophyceae)) were included in the macroalgal dataset in the further analyses as they concern the same trophic group (Table 4S). One taxon (the species-rich genus Halimeda) was found at all sites, whereas 23 other taxa were scored at only a single site. The Michaelis–Menten species (taxa) accumulation curve (Figure 4) indicated an expected number of 110 (Smax) species in an infinite number of samples. In total, 495 specimens of seaweed and seagrass taxa were collected as voucher material.
Table 1. SIMPER dissimilarities in species assemblages between the regions based on respectively all data, and on the macroalgae, Ascidiacea, Fungiidae and Gastropoda datasets separately. For these four taxa the relative contribution (%) to the dissimilarity value calculated for all data (first column), is provided in parentheses SIMPER Halmahera bay – TT channel Halmahera bay – Guraici islands Guraici islands - TT channel
Average Dissimilarity
Contribution Macroalgae
Contribution Ascidiacea
Contribution Fungiidae
Contribution Gastropoda
59.1 60.0 61.1
63.4 (37%) 68.8 (36%) 70.4 (33%)
78.7 (40%) 75.3 (40%) 74.3 (46%)
32.3 (19%) 40.7 (22%) 35.5 (18%)
88.4 (4%) 75.0 (2%) 93.7 (3%)
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
A. GITTENBERGER ET AL.
Figure 3. (A) CAP analysis (m = 27) based on the Sørensen distances between species assemblages of all taxa found per research site, illustrating the three bioregions TTH Channel (green), Halmahera Bay (purple) and Guraici (blue). The overlay vector function of Primer-E was used to illustrate the species whose occurrences have a correlation of more than 0.6 with the CAP graph. On the right correlations are illustrated between the CAP model and (B) the species diversity, (C) salinity, (D) turbidity and (E) pH. m = number of canonical axes.
Figure 4. Michaelis–Menten species accumulation curves based on the datasets including the macroalgae (black dots, Smax = 110), the Ascidiacea (crosses, Smax = 218), the mushroom corals (open squares, Smax = 37) and the parasitic gastropods (circles, Smax = 27).
In the MDS plot based on Sørensen distances, in which the symbols had been colour-coded according to area, bottom orientation and habitat, no distinct patterns were discernible. This was confirmed by Permanova tests, which showed no significant correlations between the species assemblages observed and the area, bottom orientation and/or habitat (P > 0.05). The macroalgal species assemblages were significantly linked to the bioregions (Permanova, df = 2,
pseudo-F = 2.40, P = 0.027). SIMPER dissimilarities between the three bioregions, based on the macroalgal data ranged from 63 to 70 (Table 1). However, when all taxa were included in the Permanova analyses, the areas within the bioregions differed significantly while an analysis solely on the macroalgal dataset did not show a significant difference (Permanova, df = 4, pseudo-F = 1.12, P = 0.271). The CAP analysis supported the Permanova results, which indicates that the study sites could be classified into three distinct groups on the basis of their species assemblages (Figure 5(A)). Using the resulting CAP model and the ‘leave-one-out’ method it was found that bioregions of sites could be identified with 82% certainty on the basis of their species assemblages. To be more precise, on the basis of their macroalgal species compositions 93% of the TTH Channel sites (25 out of 27) were categorized correctly, compared with only 33% of the Guraici sites (2 out of 6), and with 100% of the Halmahera Bay sites (5 out of 5). Four Guraici sites were miscategorized as TTH Channel sites by the CAP model. One TTH Channel sites was miscategorized as a Halmahera Bay site, while another was incorrectly categorized as a Guraici site.
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
CORAL REEF INDICATORS OF BIOREGIONS
Figure 5. CAP analyses based on the Sørensen distances between species assemblages of (A) Macroalgae (m = 14), (B) Ascidiacea (m = 10), (C) mushroom corals (m = 5) and (D) parasitic gastropods (m = 6), illustrating the three bioregions TTH Channel (green), Halmahera Bay (purple) and Guraici (blue). m = number of canonical axes.
Ascidiacea (Tunicata)
Fungiidae (Scleractinia)
In total, 180 ascidian species and morphotypes of ascidians were distinguished (Supplementary Material Table 5S). The Michaelis–Menten species accumulation curve (Figure 4) is asymptotic, indicating an expected total of 218 (Smax). The following families were represented: Ascidiidae, Cionidae, Clavelinidae, Corellidae, Diazonidae, Didemnidae, Holozoidae, Perophoridae, Polycitoridae, Polyclinidae, Pycnoclavellidae, Pyuridae, and Styelidae. In the MDS plot based on Sørensen distances there was no discernible pattern in the colour-coded dots representing species assemblages in various areas, bottom orientations and habitats. Permanova tests (p > 0.05) supported this finding: they revealed no significant correlation with ascidian species assemblages, which themselves could be significantly linked (Permanova, df = 2, pseudo-F = 3.08, P = 0.002) to the bioregions. SIMPER dissimilarities between the three bioregions ranged from 74 to 79 (Table 1). When only the Ascidiacea dataset was analysed, there was no significant difference in occurrence between the areas within the bioregions (Permanova, df = 4, pseudo-F = 1.15, P = 0.191). The CAP analysis supported the Permanova results, indicating that the study sites could be classified into three distinct groups regarding their species assemblages (Figure 5(B)). When the ‘leave-one-out’ method was applied to the resulting CAP model, the bioregions of the sites were identified with 98% certainty on the basis of their species assemblages. Only one site was wrongly categorized: a Halmahera Bay site was misidentified as a Guraici site.
Thirty-six mushroom coral species were recorded (Supplementary Material Table 6S). The Michaelis–Menten species accumulation curve (Figure 4) indicated an expected total of 37 species (Smax) as asymptote. Four species appeared to be common, by being present at > 37 sites: Fungia fungites, Lithophyllon repanda, Pleuractis granulosa and P. paumotensis. Species encountered at only two localities occurred at greater depths in sheltered conditions: Cycloseris somervillei, Lithophyllon spinifer and Pleuractis taiwanensis. Halomitra clavator formed dense aggregations in one of its four sites as a result of asexual reproduction by fragmentation (Hoeksema and Gittenberger, 2010). In the MDS plot based on Sørensen distances there was no discernible pattern in the colour-coded dots representing mushroom coral assemblages in various areas, bottom orientations and habitats. This was confirmed by Permanova tests (P > 0.05), which found no significant correlation. Mushroom coral species assemblages could not be significantly linked (Permanova, df = 2, pseudo-F = 2.53, P = 0.112) to bioregions. SIMPER dissimilarities between the three bioregions, based on the Fungiidae data, ranged between 32 and 41 (Table 1). The CAP analysis supported the Permanova results, indicating that the study sites could not be classified into three distinct bioregions on the basis of their fungiid species assemblages (Figure 5(C)). When the ‘leave-one-out’ method was applied to the resulting CAP model, the bioregions of sites could be identified with only 66% certainty on the basis of their mushroom coral species assemblages. TTH Channel sites were best identified on the basis of the species present: 21 of the 27 sites
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
A. GITTENBERGER ET AL.
were correctly matched. Of the other six sites, four were misclassified as Guraici sites and two as Halmahera Bay sites. Most (five out of six) Guraici sites were wrongly categorized as TTH Channel sites. Only one Guraici site was correctly matched on the basis of the CAP model. Three of the five Halmahera Bay sites were correctly categorized, while one was wrongly categorized as a Guraici site and another as a TTH Channel site. Epitoniidae and Coralliophilidae (Gastropoda) Twenty coral-associated parasitic gastropods were recorded from their mushroom coral hosts: ten were ecto-parasites (Epitoniidae; Supplementary Material Table 7S) and ten were endo-parasites (Coralliophilidae; Supplementary Material Table 8S). The Michaelis–Menten species accumulation curve (Figure 4) indicated an expected total of 27 (Smax) species in an infinite number of samples. Two new associations were recorded, both involving epitoniid snails: Epifungium lochi with Cycloseris cyclolites, and Surrepifungium ingridae with Heliofungia actiniformis. In the MDS plot based on Sørensen distances, the colour-coded dots representing species assemblages of areas, bottom orientations and habitats did not reveal distinct patterns of correlation. Permanova tests (P > 0.05) confirmed this. The parasitic snail assemblages could not be significantly linked (Permanova, df = 2, pseudo-F = 1.22, P = 0.421) to bioregions. SIMPER dissimilarities between the three bioregions, based on the Gastropoda data, ranged from 75 to 94 (Table 1). The CAP analysis supported the Permanova results, indicating that the study sites could not be easily classified into three distinct bioregions (Figure 5(D)). When the ‘leave-one-out’ method was applied to the CAP model, the bioregions of the sites could be identified with 71% certainty. The TTH Channel sites could be best categorized on the basis of the gastropod species present: 15 of the 18 sites were correctly categorized. Of the other three sites, one was wrongly categorized as a Guraici site and two were misclassified as Halmahera Bay sites. The two Guraici sites where gastropods were recorded were both wrongly categorized as Halmahera Bay sites on the basis of the CAP model. Of the four Halmahera Bay sites, two were
correctly categorized, while the other two were wrongly categorized as Guraici sites.
DISCUSSION The main aim of the present study was to use particular taxa and/or species assemblages to identify coral reef bioregions off Halmahera. It was also investigated whether species from different trophic groups have equal value as indicators for these regions. The large disparity between the total number of species predicted by the Michaelis–Menten species accumulation curve based on all data (Figure 2) and the number of species actually observed (365 minus 338 is 27 species) is striking. This difference appears to be primarily attributable to undersampling of the ascidians and parasitic gastropods. Examination of the curves of the separate taxa (Figure 4) reveals that ascidians were undersampled by 18% and the parasitic gastropods by 27%. According to the Smax resulting from the Michaelis–Menten analysis, only 6% of the macroalgal taxa were missed (Figure 4). This may be an underestimate, as several of the macroalgal taxa recorded certainly represent multiple species (e.g. Dictyota, Halimeda, Padina) that the algae expert could not distinguish in the field. The actual number of species collected could be >120. Some taxa (notably turf algae and crustose algae) were not scored because they are too inconspicuous. It is unlikely that a larger sampling effort would have yielded any additional mushroom coral species, as 36 species were found and 37 were expected (Figure 4). Addition of more sites would probably not have yielded much better data with which to achieve the research aims. Missing species were probably rare and therefore it can be assumed that they would not be very valuable as potential bioregion indicators. None of the datasets analysed indicated a significant correlation between species assemblages recorded and bottom orientation and/or habitat type as defined in the present study. Such relationships might have become clearer if the bottom orientation and/or habitat categories had been described in more detail (Scheffers et al., 2010).
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
CORAL REEF INDICATORS OF BIOREGIONS
Hoeksema and Moka (1989) compared similarity in mushroom coral fauna among 30 sites in the Flores Sea using the Sørensen similarity index and found bottom inclination and currents to be important factors determining similarity in species composition. They distinguished three major reef habitats: (1) reefs in open sea; (2) reefs along sheltered coastlines facing open sea; and (3) reefs in inland bays. Other surveys in the Coral Triangle or its close proximity have indicated that the following spatial components are important determinants of the species composition of mushroom coral assemblages: (1) distance offshore from the coastline with river run-off, (2) position around reefs in relation to predominant wind direction with major wave action, and (3) depth of the sea floor (Hoeksema, 1991, 2012a, b; Cleary et al., 2005, 2006, 2008; Waheed and Hoeksema, 2013, 2014). These findings do not disagree with the present results. Significant differences were found between the three major bioregions when the macroalgal and the ascidians datasets were analysed separately, but not when the mushroom coral and parasitic gastropod datasets were analysed independently. The most significant result was obtained when all four taxa sets were pooled and sites could be identified by their species composition with 100% certainty according to the ‘leave-one-out’ test conducted on the CAP-model, and areas within bioregions were found to vary significantly regarding species composition. Thus, although the mushroom coral and parasitic gastropod datasets were not very powerful for distinguishing bioregions in separate analyses, they added value when combined with the datasets of other taxa. Although ascidians have rarely been included in similar studies (the exceptions being Scheffers et al., 2010; Wilson et al., 2010) and little is known about the diversity of ascidians in the tropical regions worldwide (Lambert, 2003; Lee et al., 2013), they were by far the best bioregion indicators in the present survey. On the basis of the ascidian species dataset alone, the bioregion of a site could be identified with 98% certainty, which is higher than the 83, 66 and 71% for macroalgae, mushroom corals and parasitic gastropods, respectively. The macroalgal dataset explained as
much as 83% of the variation despite the fact that many taxa had been identified only to the genus level. In addition to species compositions that were diagnostic for the bioregions, potential indicator species were identified. To do this, in Primer-E, the correlation was calculated between the presence of species and the CAP model distinguishing between the bioregions (Figure 3(A)) with the assumption that species with a presence that correlates for more than r = 0.6 with this CAP model, are most suitable as indicator species for the bioregions distinguished in this model. By far the most (13) of the 17 species whose occurrences correlated with more than r = 0.6 with the CAP-model based on all taxa (Figure 3(A)), were ascidians; the others were three macroalgae and one mushroom coral species. Each bioregion appears to have its own typical ascidian species (Figure 3(A)). Three macroalgal species qualified as indicator species, one of which appears to be typical of the TTH Channel bioregion, while the other two appear to be typical of the Halmahera Bay bioregion. One mushroom coral species, Lithophyllon spinifer, appears to be a unique indicator species for Halmahera Bay (Figure 3(A)). The same species also appeared to be indicative for sheltered bays off eastern Sabah (Waheed and Hoeksema, 2013). The environmental parameters in the three bioregions could not be studied in detail. Some basic water quality parameters were, however, measured systematically at all sites. They appear to explain some of the general differences in species assemblages found in the three bioregions. The salinity measurements (Table 3S; Figure 3(C)) show that the highest surface water salinities were recorded in the offshore Guraici bioregion (Figure 1), which is remote from fluvial discharge. The largest differences between surface and bottom salinities were found in Halmahera Bay (Table 3S; Figure 3(C)); these are attributable to freshwater inflow from the coastal area of Halmahera Bay. Ascidians, macroalgae and corals are all known to be sensitive to either low or high salinities and to differences in salinity over time (Kirst, 1989; Gittenberger and Hoeksema, 2006; Epelbaum et al., 2009; Monniot, 2009). These sensitivities, some of which are species-specific, may explain at least some
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
A. GITTENBERGER ET AL.
of the correlations found between the occurrences of these species and the bioregions (Figure 3(A)). Turbidity was found to be the highest in Halmahera Bay (Table 3S; Figure 3(D)), which can be attributed to the terrigenous impact from Halmahera mainland but also to resuspension of bottom sediments. Turbidity is usually associated with a high organic particle load in the water, which is important for filter feeders such as ascidians, although not all of these are able to tolerate much sedimentation. This is certainly true for many coral species (Erftemeijer et al., 2012), and may therefore explain why the coral Lithophyllon spinifer is correlated with the Halmahera Bay bioregion. Unlike other mushroom coral species, this species occurs on sediment-rich reef bases (Hoeksema, 2012a, b). The higher pH values found in the Guraici bioregion (Table 3S; Figure 3(A), (E)) may be attributable to the limestone underground of the Guraici islands, whereas Halmahera and the islands along the TTH Channel are of volcanic origin. Although the occurrences of some ascidians and mushroom corals are often associated with either a higher or lower pH (Monniot, 2009), it is uncertain whether the pH explains much in the present study, because the average pH value of the water at the research sites in the Guraici bioregion is only about 0.1 higher than in the TTH Channel and Halmahera Bay bioregions. This difference is significant from an acidity point of view, as it represents a 30% difference in hydrogen ion concentrations. Whereas both Guraici and Halmahera Bay have either relatively high or low salinities, turbidities and pH values, in the TTH Channel bioregion these water variables are neither exceptionally high nor low (Table 3S, Figure 3(B)–(E)). As argued before, many species are sensitive to the more extreme environmental conditions at Guraici (most exposed to oceanic water) and Halmahera Bay (most sheltered and with terrigenous impact). This might explain why the greatest species diversity was found in the TTH Channel bioregion (Figure 3(A)). Understanding the links between species distribution and habitat characteristics at various spatial scales is often the first step in clarifying biodiversity distribution as a basis for spatially explicit conservation plans (Dalleau et al., 2010;
Jimenez et al., 2012). Although similar multi-taxa coral reef studies have been conducted specifically for that purpose, they have not always been clear in distinguishing bioregions. A recent study in the Baa Atoll Man and Biosphere UNESCO Reserve (Maldives) for example, had some success with macroalgal and coral species data to distinguish between stations (Jimenez et al., 2012). The fish, hydroid and other macro-invertebrate species, however, did not appear to be of any value. As in the present study, no correlations could be found between species compositions and coral reef habitat characteristics. The multi-taxon reef study in the Maldives therefore supports the results of the present one. It makes an even stronger case for combining macroalgae and corals with ascidians to distinguish between bioregions, while assemblages of gastropods, fish and hydroids are less informative. Although not all taxa are equally valuable in distinguishing between bioregions, such studies do support the application of a multi-taxa approach when one aims at unravelling the mechanisms that control biodiversity distribution (Cleary et al., 2005, 2008; Becking et al., 2006) with the purpose of prioritizing coral reef areas for conservation purposes. Macroalgae, corals and ascidians are a good selection of taxa herein, not only because of their proven value as indicator species, but also because they are represented in nearly all coral reef systems and habitats. They represent taxa with either relatively long (corals), intermediate (macroalgae) and short (ascidians) life histories, that are highly (macroalgae), intermediately (corals) and usually not (ascidians) autotrophic and dependent on sunlight penetrating the water, which regulates their depth distribution (Hoeksema, 2012a) and their susceptibility to sedimentation (Erftemeijer et al., 2012). Such a selection of taxa may help to assess the health of coral reef areas and provide a tool to predict what will happen in the short- and long-term future, which can be relevant when setting conservation priorities. The occurrences of macroalgae and ascidians will give the best assessment of changes in coral system health that happened in recent years while coral diversity will result from changes that took place over a longer time span. For
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
CORAL REEF INDICATORS OF BIOREGIONS
assessments of changes in ecosystem health during the last few weeks to months, the presence and variety of zooxanthellae in corals will probably provide the best data (Krediet et al., 2013). Regardless of coral system health, a combination of data on occurrences of highly diverse taxa that differ strongly from each other in their life-history strategies, will enhance the accuracy with which their statistical analyses will be able to detect both regions with a high species richness and regions with a relatively high number of rare or range-restricted (endemic) species, which are not necessarily spatially concordant (Roberts et al., 2002; Orme et al., 2005; Selig et al., 2014). As these regions are the specific focus of setting global priorities for marine biodiversity conservation (Selig et al., 2014) the multi-taxon approach that was used in the present study can be used as a more general tool in the conservation planning of coral reef systems.
ACKNOWLEDGEMENTS
The research was organized by the Research Centre of Oceanography of the Indonesian Institute of Sciences (PPO-LIPI) and the Naturalis Biodiversity Center. It was based at the LIPI field station of LIPI on Ternate. The research permit was issued by the Indonesian State Ministry of Research and Technology (RISTEK). The authors gratefully acknowledge the support and contributions of the Building with Nature project SI 1.3, funded through the Ecoshape programme (www.ecoshape.nl). Bastian Reijnen is thanked for providing the geographical coordinates and habitat characteristics of the research localities. J. Burrough is thanked for critically reading and editing of an early version of the manuscript. REFERENCES Anderson MJ. 2001. A new method for non–parametric multivariate analysis of variance. Austral Ecology 26: 32–46. Anderson MJ, Willis TJ. 2003. Canonical analysis of principal coordinates: a useful method of constrained ordination for ecology. Ecology 84: 511–525. Andréfouët S, Wantiez L. 2010. Characterizing the diversity of coral reef habitats and fish communities found in a UNESCO
World Heritage Site: the strategy developed for Lagoons of New Caledonia. Marine Pollution Bulletin 61: 612–620. Becking LE, Cleary DFR, de Voogd NJ, Renema W, de Beer M, van Soest RWM, Hoeksema BW. 2006. Beta-diversity of tropical marine assemblages in the Spermonde Archipelago, Indonesia. Marine Ecology 27: 76–88. Beger M, McKenna SA, Possingham HP. 2007. Effectiveness of surrogate taxa in the design of coral reef reserve systems in the Indo-Pacific. Conservation Biology 21: 1584–1593. Benzoni F, Stefani F, Stolarski J, Pichon M, Mitta G, Galli P. 2007. Debating phylogenetic relationships of the scleractinian Psammocora: molecular and morphological evidences. Contributions to Zoology 76: 35–54. Benzoni F, Arrigoni R, Stefani F, Reijnen BT, Montano S, Hoeksema BW. 2012. Phylogenetic position and taxonomy of Cycloseris explanulata and C. wellsi (Scleractinia: Fungiidae): lost mushroom corals find their way home. Contributions to Zoology 81: 125–146. Brose U, Martinez ND. 2004. Estimating the richness of species with variable mobility. Oikos 105: 292–300. Clarke KR. 1993. Non–parametric multivariate analysis of changes in community structure. Australian Journal of Ecology 18: 117–143. Clarke KR, Gorley RN. 2006. PRIMER v6: User Manual/ Tutorial. PRIMER–E, Plymouth Marine Laboratory: Plymouth. Cleary DFR, Becking LE, De Voogd NJ, Renema W, De Beer M, Van Soest RWM, Hoeksema BW. 2005. Variation in the diversity and composition of benthic taxa as a function of distance offshore, depth and exposure in the Spermonde Archipelago, Indonesia. Estuarine, Coastal and Shelf Science 65: 557–570. Cleary DFR, Suharsono, Hoeksema BW. 2006. Coral diversity across a disturbance gradient in the Pulau Seribu reef complex off Jakarta, Indonesia. Biodiversity and Conservation 15: 3653–3674. Cleary DFR, DeVantier L, Giyanto, Vail L, Manto P, De Voogd NJ, Rachello–Dolmen PG, Tuti Y, Budiyanto A, Wolstenholme J, et al. 2008. Relating variation in species composition to environmental variables: a multi–taxon study in an Indonesian coral reef complex. Aquatic Sciences 70: 419–431. Colwell RK, Coddington JA. 1994. Estimating terrestrial biodiversity through extrapolation. In Biodiversity: Measurement and Estimation, Hawksworth DL (ed.). Royal Society: London; 101–118. Dalleau M, Andréfouët S, Wabnitz CCC, Payri C, Wantiez L, Pichon M, Friedman K, Vigliola L, Benzoni F. 2010. Use of habitats as surrogates of biodiversity for efficient coral reef conservation planning in Pacific Ocean islands. Conservation Biology 24: 541–552. Epelbaum A, Herborg LM, Therriault TW, Pearce CM. 2009. Temperature and salinity effects on growth, survival, reproduction, and potential distribution of two nonindigenous botryllid ascidians in British Columbia. Journal of Experimental Marine Biology and Ecology 369: 43–52. Erftemeijer PLA, Riegl B, Hoeksema BW, Todd PA. 2012. Environmental impacts of dredging and other sediment disturbances on corals: a review. Marine Pollution Bulletin 64: 1737–1765. Gittenberger A, Gittenberger E. 2005. A hitherto unnoticed adaptive radiation: epitoniid species (Gastropoda: Epitoniidae)
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
A. GITTENBERGER ET AL.
associated with corals (Scleractinia). Contributions to Zoology 74: 125–204. Gittenberger A, Gittenberger E. 2011. A largely cryptic, adaptive radiation of parasitic snails: sibling species in Leptoconchus (Gastropoda: Caenogastropoda: Coralliophilidae), associated with specific coral hosts (Scleractinia: Fungiidae). Organisms, Diversity & Evolution 11: 21–41. Gittenberger A, Hoeksema BW. 2006. Phenotypic plasticity revealed by molecular studies on reef corals of Fungia (Cycloseris) spp. (Scleractinia: Fungiidae) near river outlets. Contributions to Zoology 75: 195–201. Gittenberger A, Hoeksema BW. 2013. Habitat preferences of coral-associated wentletrap snails (Gastropoda: Epitoniidae). Contributions to Zoology 82: 1–25. Gittenberger A, Goud J, Gittenberger E. 2000. Six Epitonium species (Prosobranchia, Epitoniidae), four of which new, associated with mushroom corals (Scleractinia, Fungiidae) of Sulawesi, Indonesia. Nautilus 114: 1–13. Gittenberger A, Reijnen BT, Hoeksema BW. 2011. A molecularly based phylogeny reconstruction of mushroom corals (Scleractinia: Fungiidae) with taxonomic consequences and evolutionary implications for life history traits. Contributions to Zoology 80: 107–132. Hoeksema BW. 1989. Taxonomy, phylogeny and biogeography of mushroom corals (Scleractinia: Fungiidae). Zoologische Verhandelingen 254: 1–295. Hoeksema BW. 1991. Control of bleaching in mushroom coral populations (Scleractinia: Fungiidae) in the Java Sea: stress tolerance and interference by life history strategy. Marine Ecology Progress Series 74: 225–237. Hoeksema BW. 2007. Delineation of the Indo–Malayan Centre of Maximum Marine Biodiversity: the Coral Triangle. In Biogeography, Time and Place: Distributions, Barriers and Islands, Renema R (ed.). Springer: Dordrecht; 117–178. Hoeksema BW. 2012a. Distribution patterns of mushroom corals (Scleractinia: Fungiidae) across the Spermonde Shelf, South Sulawesi. Raffles Bulletin of Zoology 60: 183–212. Hoeksema BW. 2012b. Evolutionary trends in onshore–offshore distribution patterns of mushroom coral species (Scleractinia: Fungiidae). Contributions to Zoology 81: 199–221. Hoeksema BW. 2014. The ‘Fungia patella group’ (Scleractinia, Fungiidae) revisited with a description of the mini mushroom coral Cycloseris boschmai sp. n. ZooKeys 371: 57–84. Hoeksema BW, Gittenberger A. 2010. High densities of mushroom coral fragments at West Halmahera, Indonesia. Coral Reefs 29: 691. Hoeksema BW Koh EGL. 2009. Depauperation of the mushroom coral fauna (Fungiidae) of Singapore (1860s–2006) in changing reef conditions. Raffles Bulletin of Zoology Supplement 22: 91–101. Hoeksema BW, Moka W. 1989. Species assemblages and ecomorph variation of mushroom corals (Scleractinia: Fungiidae) related to reef habitats in the Flores Sea. Netherlands Journal of Sea Research 23: 149–160. Hoeksema BW, Van der Meij SET. 2010. Crossing Marine Lines at Ternate: Capacity Building of Junior Scientist in Indonesia for Marine Biodiversity Assessments. Naturalis: Leiden. Hoeksema BW, Van der Land J, Van der Meij SET, Van Ofwegen LP, Reijnen BT, Van Soest RWM, De Voogd NJ. 2011. Unforeseen importance of historical collections as baselines to determine biotic change of coral reefs: the Saba Bank case. Marine Ecology 32: 135 141.
Irvine KM, Dinger EC, Sarr D. 2011. A power analysis for multivariate tests of temporal trend in species composition. Ecology 92: 1879–1886. Jimenez H, Bigot L, Bourmaud C, Chabanet P, Gravier– Bonnet N, Hamel MA, Payri C, Mattio L, Menou JL, Naeem S, et al. 2012. Multi–taxa coral reef community structure in relation to habitats in the Baa Atoll Man and Biosphere UNESCO Reserve (Maldives), and implication for its conservation. Journal of Sea Research 72: 77–86. Kerrigan B, Breen D, De’ath G, Day J, Fernandes L, Tobin R, Dobbs K. 2010. Classifying the biodiversity of the Great Barrier Reef World Heritage Area. Australian Government. Great Barrier Reef Marine Park Authority Research Publication Series 104: 1–55. Kirst GO. 1989. Salinity tolerance of eukaryotic marine algae. Annual Review of Plant Physiology and Plant Molecular Biology 40: 21–53. Krediet CJ, Ritchie KB, Paul VJ, Teplitski M. 2013. Coralassociated microorganisms and their roles in promoting coral health and thwarting diseases. Proceedings of the Royal Society B 280: 2012–2328. Lambert G. 2003. Marine biodiversity of Guam: the Ascidiacea. Micronesia 35–36: 584–593. Lee SS, Teo SLM, Lambert G. 2013. New records of solitary ascidians on structures in Singapore waters. Marine Biodiversity Records 6: e93. McArdle BH, Anderson MJ. 2001. Fitting multivariate models to community data: a comment on distance-based redundancy analysis. Ecology 82: 290–297. Monniot F. 2009. Some ascidians from Indonesia marine lakes (Raja Ampat Islands, West Papua). Zootaxa 2106: 13–40. Monniot F, Monniot C. 2001. Ascidians from the tropical western Pacific. Zoosystema 23: 201–383. Mumby PJ, Broad K, Brumbaugh DR, Dahlgren CP, Harborne AR, Hastings A, Holmes KE, Kappel CV, Micheli F, Sanchirico JN. 2008. Coral reef habitats as surrogates of species, ecological functions, and ecosystem services. Conservation Biology 22: 941–951. Munro C. 2005. Diving systems. In Methods for the Study of Marine Benthos, 3rd edn. Eleftheriou A, McIntyre AD (eds). Blackwell Science: Oxford; 112–119. Orme CDL, Davies RG, Burgess M, Eigenbrod F, Pickup N, Olson VA, Webster AJ, Ding TS, Rasmussen PC, Ridgely RS, et al. 2005. Global hotspots of species richness are not congruent with endemism or threat. Nature 436: 1016–1019. Rainbow PS. 2009. Marine biological collections in the 21st century. Zoologica Scripta 38(S1): 33–40. Roberts CM, McClean CJ, Veron JEN, Hawkins JP, Allen GR, McAllister DE, Mittermeier CG, Schueler FW, Spalding M, Wells F, et al. 2002. Marine conservation hotspots and conservation priorities for tropical reefs. Science 295: 1280–1284. Scheffers SR, Van Soest RWM, Nieuwland G, Bak RPM. 2010. Coral reef framework cavities: is functional similarity reflected in composition of the cryptic macrofaunal community? Atoll Research Bulletin 583: 1–26. Selig ER, Turner WR, Troëng S, Wallace BP, Halpern BS, Kaschner K, Lascelles BG, Carpenter KE, Mittermeier RA. 2014. Global priorities for marine biodiversity conservation. PLoS ONE 9: e82898. Sorensen T. 1957. A method of establishing groups of equal amplitude in plant sociology based on similarity of species
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
CORAL REEF INDICATORS OF BIOREGIONS
and its application to analyses of the vegetation on Danish commons. Biologiske skrifter kongelige Danske videnskabernes selskab 5: 1–34. Templado J, Paulay G, Gittenberger A, Meyer C. 2010. Sampling the marine realm. ABC Taxa 8: 273–307. Tomascik T, Mah AJ, Nontji A, Moosa MK. 1997. The Ecology of the Indonesian Seas. Periplus: Singapore. Van der Meij SET, Moolenbeek RG, Hoeksema BW. 2009. Decline of the Jakarta Bay molluscan fauna linked to human impact. Marine Pollution Bulletin 59: 101–107. Van der Meij SET, Suharsono, Hoeksema BW. 2010. Long-term changes in coral assemblages under natural and anthropogenic stress in Jakarta Bay (1920–2005). Marine Pollution Bulletin 60: 1442–1454. Verheij E, Prud’homme van Reine WF. 1993. Seaweeds of the Spermonde Archipelago, SW Sulawesi, Indonesia. Blumea 37: 385–510.
Waheed Z, Hoeksema BW. 2013. A tale of two winds: species richness patterns of reef corals around the Semporna peninsula, Malaysia. Marine Biodiversity 43: 37–51. Waheed Z, Hoeksema BW. 2014. Diversity patterns of scleractinian corals at Kota Kinabalu, Malaysia, in relation to exposure and depth. Raffles Bulletin Zoology 62: 66–82. Wilson SK, Depczynski M, Fisher R, Holmes TH, O’Leary RA, Tinkler P. 2010. Habitat associations of juvenile fish at Ningaloo reef, Western Australia: the importance of coral and algae. PLoS ONE 5(12): e15185.
SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article at the publisher’s web site.
# 2014 The Authors. Aquatic Conservation: Marine and Freshwater Ecosystems published by John Wiley & Sons, Ltd.
View publication stats
Aquatic Conserv: Mar. Freshw. Ecosyst. (2014)
