Saskia Van Gaever, Marine Biology Section, Department of Biology, Ghent University, Krijgslaan 281/S8, 9000 Ghent, Belgium. E-mail: email@example.com
Assessing the relative contribution of local diversity to regional biodiversity may be the key to understanding large-scale and even global patterns in species diversity. Here, the contribution of habitat heterogeneity of cold seeps at three spatial scales [micro-scale (ms), macro-scale (10 to 100s of ms), and mega-scale (10 to 100s of km)] to the total nematode biodiversity (genus level) along the Norwegian continental margin is evaluated. Due to the development of higher resolution bathymetry and increased bottom sampling in recent years, continental margins, once regarded as monotonous landscapes, are now acknowledged to have a high degree of habitat complexity and diversity. By calculating the additive partitioning of gamma diversity in alpha and beta fractions, we examined to what extent habitat diversity of seep sites significantly increases the nematode genus composition and diversity at different spatial scales. Siboglinidae patches and control sediments yielded comparably high levels of nematode genus richness. They exhibited low turnover rates within and across the different seep sites. In contrast, the bacterial mats at Håkon Mosby Mud Volcano (HMMV) and the reduced sediments at the Nyegga pockmarks harboured genus-poor nematode communities with an equally high dominance of one or two species, which were different for each seep. Different habitats, in particular at the HMMV, contributed significantly to the seep nematode richness. This study demonstrates that the presence of distinct habitat types within multiple seep sites contributes to the high diversity of nematode communities inhabiting the seeps in the Norwegian deep sea.
Historically, the relationship between spatial scale and patterns of species diversity has been a major topic in ecological research (Whittaker 1960, 1972; Legendre et al. 1997; Crawley & Harral 2001; Perry et al. 2002; Takashi 2004). Diversity is often supposed to vary with spatial scale, in the sense that an increase in scale could provide more resources to species and subsequently promote an increase in diversity (Crawley & Harral 2001). A large body of literature exists on the relationship between diversity patterns and multiple spatial scales in terrestrial systems (Chown & Gaston 2000; Godfray & Lawton 2001; Hill & Hamer 2004). Similar ecological studies in marine systems are more difficult to execute because these are open ecosystems that harbour numerous highly mobile species not constrained by the habitat boundaries typical for terrestrial systems. Nevertheless, some marine environments, for instance the rocky intertidal, coral reefs but also deep-sea environments such as hydrothermal vents and cold seeps, create a naturally occurring hierarchical patch structure for the associated benthos due to the high habitat heterogeneity. They are therefore very suitable to explore questions of diversity patterns related to scale in marine systems. However, the effect of spatial scale on nematode communities in the deep sea has, to date, been poorly studied, mainly because of sampling constraints. Fortunately, submersibles and ROVs are now providing easier access to certain hotspots on continental margins, and allow fine-scale to large-scale sampling.
As the exploitation of living and mineral resources is advancing much faster than ecological knowledge on continental slopes, a comprehensive analysis of species distribution, biodiversity patterns and processes on continental margins is urgently needed. In the present study, diversity patterns of the smaller-sized benthos (meiobenthos) inhabiting several cold-seep sites along the Norwegian continental margin were examined in relation to different spatial scales. The spatial scales included here range between meters (micro-scale), 10s to 100s of meters (macro-scale) and 10 to 100s of kilometres (mega-scale). Micro- and macro-scale comparisons were based on samples from different habitats collected on the Håkon Mosby Mud Volcano (HMMV). For the mega-scale comparison, similar habitats were identified and examined for three different seep sites studied along the Norwegian margin (HMMV, Storegga Slide, Nyegga area). Here we define a cold seep as the complex of all habitats affected at least to some degree by methane-rich fluid seepage, recognised by visual manifestations such as bacterial mats, siboglinid tubeworms or other chemosynthesis-dependent megafauna, reduced black sediments, carbonate crusts or the seepage of gas bubbles. The habitats that were compared in this study were the reduced black surface sediments (with or without bacterial mats) and the Siboglinidae fields. The mega-scale comparison was also applied to the control sediments outside any seep influence.
There is abundant evidence that habitat structure has important effects on the spatial distribution of benthic populations in seep environments, both for the macrofauna (Bergquist et al. 2003; Levin et al. 2003; Levin 2005; Olu-Le Roy et al. 2007) and for the meiofauna (Van Gaever et al. 2006; S. Van Gaever, J. Galéron, M. Sibuet, A. Vanreusel, unpublished observations). Macrofaunal species diversity at California methane seeps exhibited a habitat-related heterogeneity, consistent with sulphide inhibition of species richness and evenness (Levin et al. 2003). Meiofaunal communities, in particular the nematode assemblages, also experienced a significant structuring effect by different habitats and their characteristic geochemical conditions at a pockmark seep in the Gulf of Guinea (S. Van Gaever, J. Galéron, M. Sibuet, A. Vanreusel, unpublished observations).
The meiofaunal communities at the HMMV (Barents Sea, ∼1280 m) were previously described from distinct habitats (de Beer et al. 2006; S. Van Gaever, K. Olu, S. Derycke, A. Vanreusel, unpublished observations). The well-oxygenated surface sediments at the Siboglinidae fields (de Beer et al. 2006), the outer rim and the control sediments harboured the most diverse nematode communities, whereas the mud-flowing centre yielded only a few individuals. Unexpectedly high nematode densities (>11,000 ind. 10 cm−2) were observed in the sulphidic, anoxic sediments underneath the Beggiatoa mats, with more than 99% of the nematodes belonging to the species Halomonhystera disjuncta Bastian, 1865 (Van Gaever et al. 2006). This species also dominated the smaller, more isolated, grey bacterial mats at HMMV, although densities were 10-fold lower there (∼1000 ind. 10 cm−2). Another reduced, suboxic habitat was identified at the Nyegga seep off mid-Norway (∼740 m): blackish sediments without a sediment surface-covering bacterial mat. The nematode community identified in this black sediment was low in diversity, and was strongly dominated by Terschellingia longicaudata de Man, 1907 (S. Van Gaever, K. Olu, S. Derycke, A. Vanreusel, unpublished observations). Additionally, a Siboglinidae patch and control sediments at both the Nyegga seep and the nearby Storegga seep (∼730 m) were investigated for meiobenthos. Nematode diversity here was comparable to the corresponding Siboglinidae and control habitats at HMMV (S. Van Gaever, K. Olu, S. Derycke, A. Vanreusel, unpublished observations).
This paper focuses on the influence of different habitats and multiple spatial scales on genus diversity of the nematode communities associated with cold seeps. Sulphidic seep sediments are usually characterised by low meiofaunal diversity and/or increased dominance (Shirayama & Ohta 1990; Dando et al. 1991; Van Gaever et al. 2006; S. Van Gaever, K. Olu, S. Derycke, A. Vanreusel, unpublished observations), whereas siboglinid tubeworm fields may have a higher diversity, as has been found for macrofauna (Levin & Mendoza 2007). Therefore we expect habitat heterogeneity, rather than location, to be a key factor in structuring nematode diversity in such environments (Van Gaever et al. 2006; S. Van Gaever, K. Olu, S. Derycke, A. Vanreusel, unpublished observations). This will be tested by an additive partitioning analysis.
Material and Methods
The Håkon Mosby Mud Volcano (∼1280 m water depth) is located at the Norwegian-Barents-Spitsbergen continental margin at 72°N, an area characterised by major submarine slides and smaller sea-floor features (Vogt et al. 1997) (Fig. 1). This mud volcano is about 1.5 km in diameter and rises up to 10 m above the sea floor. It presents a concentric distribution of habitats: the bare sediments of the centre, the microbial mats, and the Siboglinidae fields (Figs 2 and 3A,B). These three habitats are surrounded by a habitat referred to as the outer rim (Pimenov et al. 2000; Gebruk et al. 2003). Observation with the ROV Victor 6000 revealed bubble and fluid discharge in the northern part of the mud volcano (‘active site’).
The morphology of the continental margin off North-west Norway at 64°N is marked by the Storegga Slide scar, produced by a series of giant Holocene slope failures. The most recent failure occurred about 8200 years ago and was the last megaslide (between 64° and 70°N) in this region (Bryn et al. 2005; Paull et al. 2008) (Fig. 1). Sea-floor pockmarks and subsurface chimney structures are common on the Norwegian continental margin north of the Storegga Slide scar. The Nyegga area, located on the northeastern flank of this slide, is characterised by a cluster of complex pockmarks at 740 m water depth. These pockmarks are circular in plan view and possess ridges of wide carbonate rocks up to 190 m long.
Two habitat types occurred at all sampling sites: Siboglinidae patches and control sediments. Reduced conditions at the sediment surface were only observed and sampled at HMMV and Nyegga. However, at the HMMV these reduced sediments were covered with bacterial mats (white and grey, Fig. 3A), whereas there was no evidence found of giant sulphide-oxidising bacteria on top of the black sediments at the Nyegga site (Fig. 3C). Because of the reduced surface conditions, these sediments were grouped as one type of habitat in the further analysis, contrasting with the more oxygenated surface sediments of the siboglinid fields. Although both Beggiatoa bacteria and siboglinid tubeworms inhabit reducing sediments with variable sulphide levels, a significant difference is found in the sulphide depth profiles of both habitats (de Beer et al. 2006). The sulphide-oxidising Beggiatoa bacteria require a significant sulphide flux up to the surficial sediment layers. In contrast, Siboglinidae fields can establish in sediments where sulphide is only present in the deeper sediment layers because the worms use the subsurface part of their tube (root) to take up sulphide from the subsurficial depths. On the other hand, oxygen is pumped through the anterior tube end reaching into the water column, and subsequently a steady source of sulphide and oxygen is transported via the highly adapted blood circulation system of the worm towards the symbiotic chemoautotrophic bacteria (Schulze & Halanych 2003). Several authors have proposed that the worm roots are not only important in sulphide uptake, but also in geochemical engineering of the sediments in their immediate environment (Julian et al. 1999; Bergquist et al. 2002), which could have a significant impact on the flourishing nematode assemblages there. The Siboglinidae fields along the Norwegian margin (HMMV, Storegga and Nyegga) consist of two species of tubeworms: Sclerolinum contortum and Oligobrachia haakonmosbiensis. The former species is the most abundant, whereas the latter forms only small patches and is associated with more reduced sediments. In this study, Siboglinidae samples were all retrieved from fields composed of S. contortum.
Micro- (ms) and macro-scale (10s to 100s of ms) sampling at HMMV
Sediment samples were collected at the HMMV (micro-scale) during the ARKTIS XIX/3b cruise aboard the RV Polarstern in June–July 2003. The ROV Victor 6000 collected a total of 12 cores of 21.2 cm2 along four distinct meter-scale transects sampled during four different dives (Fig. 2, Table 1). In the zone with bacterial mats, sampling of three cores occurred in a small area (maximum 4 m2), selecting the larger patches completely covered by Beggiatoa (dive transect 219, Fig. 3A). In the Siboglinidae area (dive transect 220, Fig. 3B) and the northern ‘active site’ with patchy appearance (dive transects 222, 224), a gradient was sampled each time by means of two to four cores from inside a dense Siboglinidae patch up to a less covered area with mixed patches of Beggiatoa and Siboglinidae. Cores taken on a single dive transect were located 1–5 m from each other (micro-scale). Distances between different dive sites ranged between 6 and 700 m (macro-scale). For micro-scale analysis, patterns were only considered within a single dive.
Table 1. Coordinates, sampling device, water depth (m) and habitat type for all samples. ROV = push core; MUC = multicorer.
grey bacterial mats
grey bacterial mats
For the macro-scale analysis, additional samples were taken at HMMV during the same cruise ARKTIS XIX/3b, using a video-guided multiple corer (90 mm core diameter) (Table 1). At both habitats (microbial mats, Siboglinidae field), three replicate tubes were taken from at least two different deployments. During the VICKING cruise aboard the RV Pourquoi-Pas? in May–June 2006, two sediment samples were also collected at the grey bacterial mats using push cores of the ROV Victor 6000 (surface area of 21.2 cm2). Distance between samples ranged between 19 and 680 m (macro-scale). Replicate cores in the microbial mats and Siboglinidae fields were selected based on the presence of bacteria (Beggiatoa) or Siboglinidae. Detailed results on the composition of the meiofauna and nematode communities at the different habitats have been published in Van Gaever et al. (2006; S. Van Gaever, K. Olu, S. Derycke, A. Vanreusel, unpublished observations).
Sampling at mega-scale (10s to 100s of km)
During the cruise ARKTIS XIX/3b, two control sites outside HMMV (2.6 and 56.1 km) were sampled. During the VICKING cruise, sediment samples were collected at the Storegga seep and the Nyegga seep, located 13 km from each other and 900 km from HMMV (Table 1). Samples were taken at a Siboglinidae field (Fig. 3D) and at a control site at Storegga and at Nyegga. Meiofauna samples were also collected at a patch with blackish, anoxic sediments at Nyegga (Fig. 3C). Two replicate cores were taken from each habitat at each of the sampling sites. Samples at control sites were sampled with a multiple corer (one deployment per control, tube surface area of 30.2 cm2), while the other habitats were sampled with push cores of the ROV Victor 6000 (surface area of 21.2 cm2). Detailed data on the composition of the meiofauna and nematode communities are reported in S. Van Gaever, K. Olu, S. Derycke & A. Vanreusel (unpublished observations).
All samples were fixed with 4% buffered formaldehyde at sea prior to further processing. After sieving over a 32-μm-mesh sieve, meiofauna extraction was done by density gradient centrifugation, using Ludox (a colloidal silica polymer; specific gravity 1.18) as a flotation medium (Heip et al. 1985; Vincx 1996). The meiofauna was stained with Rose Bengal, sorted, counted and identified at higher taxon level under a stereomicroscope. Nematodes (600 per sample) were picked out randomly, mounted onto glycerine slides using the formalin–ethanol–glycerol technique of Seinhorst (1959) and Vincx (1996) to prevent dehydration, and identified up to genus level using Lorenzen (1994) and Warwick et al. (1998).
Over the different spatial scales, biodiversity was measured at the three levels introduced by Whittaker (1972): alpha, beta, and gamma diversity. Alpha diversity refers to the diversity within a particular habitat, and is usually expressed by the number of species or taxa (i.e. species richness) in that ecosystem. The degree of change in taxa diversity within and between habitats is reflected in beta diversity. Gamma diversity is a measure of the overall diversity for the different ecosystems within a region. Here, we additively partitioned the gamma diversity in one alpha and two beta fractions (within and between habitats) to investigate the contribution of different habitat types to the overall nematode biodiversity at the micro- and macro-scales at HMMV. Using the additive partitioning approach (PARTITION software; Veech & Crist 2007), total nematode diversity (γ diversity) at HMMV was calculated as the sum of alpha diversity (α) within each core sample, beta diversity (β1) within each habitat, and beta diversity (β2) between different habitats:
Alpha diversity was calculated at different spatial scales by means of different indices. Genus richness (N0) was used to describe the alpha diversity within the different habitats. Rarefaction curves were constructed from values of the expected number of genera for a theoretical sample of 600 individuals (Hurlbert 1971) to compare the nematode diversity among the different habitats at the three spatial scales. Other indices used to describe alpha diversity include EG(600), the Shannon–Wiener index H ’ (log-based 2), and Pielou’s evenness J ’. These indices were calculated using PRIMERv5 software (Plymouth Marine Laboratory, Clarke & Gorley 2001).
Beta diversity was calculated as the Bray–Curtis similarity (βS) using PRIMERv5 software. Beta diversity was also reflected by the turnover of genera (βT) for each diversity level, based on the formula of Simpson (1943):
where the minimum (min) number of genera restricted to one of a pair of samples is divided by itself plus the number of shared genera (a) between samples b and c.
STATISTICAv6 software was used to carry out parametric (t-test) and non-parametric (Mann–Whitney U-test or Kruskal–Wallis ANOVA by ranks) univariate analyses of variance to test the significance of differences in alpha nematode diversity, turnover rates and beta similarities among the different habitats.
Differences in nematode assemblage structure between the different samples were assessed by non-metric multidimensional scaling (nMDS) ordination based on Bray–Curtis similarity measures, with log-transformed abundance data. The stress value gives a measure for goodness-of-fit of the MDS ordination: a stress value below 0.2 gives a potentially useful two-dimensional picture (Clarke & Warwick 2001). A significance level for the observed differences between the dive transects at HMMV was calculated with the ANOSIM permutation test (PRIMERv5 software, Clarke & Green 1988).
Micro-scale (ms) diversity patterns at HMMV
The number of nematode genera identified per ROV core ranged from 1 to 27. On average, a more genus-rich community was observed in the cores from Siboglinidae patches (19 ± 6 genera) compared to those from Beggiatoa patches (6 ± 5 genera). The alpha diversity [EG(600), H ’, J ’] for pooled cores in the Siboglinidae patches was much higher than in the Beggiatoa patches (Table 2).
Table 2. Diversity measures at the different spatial scales. Alpha diversity is expressed in genus richness (N0) per habitat, expected number of genera for 600 individuals, Shannon–Wiener index (H’ on base 2), and Pielou’s evenness (J’). Beta diversity is expressed in Bray–Curtis similarity (βS) between samples and beta turnover rate of genera (βT) between samples.
βS mean (SD)
βT mean (SD)
within reduced sediments
between reduced sediments and Siboglinidae
within reduced sediments
between reduced sediments and Siboglinidae
within reduced sediments
There was no replacement of genera between the cores in the Beggiatoa habitat of dive 219 (βS = 97.3 ± 2.4). The Bray–Curtis similarity within the Siboglinidae habitat was much lower (βS = 40.7 ± 14.4), resulting in a significant difference in genus turnover rate between the cores within the Siboglinidae habitat and the cores within the Beggiatoa habitat (t5 = 40.55, P = 0.001).
The nMDS graph depicted in Fig. 4 displays the (dis)similarities between the micro-scale samples (genus level) of the different dive transects at HMMV. Cores collected in the large Beggiatoa zone (dive transect 219) were clearly separated from all other samples (ANOSIM between dives: R = 0.998; P = 0.005). The Beggiatoa cores yielded two completely different, species-poor nematode communities. The first community, at transect 219, was dominated by Halomonhystera disjuncta (96–100% and 1–4 genera). The second community (in Beggiatoa cores from the other dives) was dominated by Microlaimus (53–70% dominance and 6–18 genera in total), and also revealed high relative abundances of the genera Metalinhomoeus, Aponema and Dichromadora, but no H. disjuncta. The Siboglinidae cores harboured high relative abundances of Sabatieria, Molgolaimus, Metalinhomoeus and Dichromadora.
Additive partitioning of genus richness (α, β, γ) revealed that the beta fraction related to the difference between cores within a single habitat (from the same dive) was the most important contributor (49%) to the total (gamma) diversity at the micro-scale at HMMV (Fig. 5). In contrast, when additive partitioning was based on the Shannon–Wiener index (H’), the alpha diversity (47%) was the most important fraction.
Macro-scale diversity patterns at HMMV
The Beggiatoa habitat yielded significantly lower alpha diversity levels compared to the Siboglinidae and control habitats (N0: t5 = −4.68, P = 0.005; H’: t5 = −6.09, P = 0.002; J’: t5 = −4.84, P = 0.005).
The pair-wise comparison of all samples within the same habitat resulted in a high (although variable) similarity value for the nematode assemblages in the Beggiatoa mats (βS = 36.0 ± 45.5) at HMMV. In contrast, the nematodes communities from the Beggiatoa habitat and the Siboglinidae habitat were extremely dissimilar from each other (βS = 4.6 ± 7.1). The average turnover rate (βT) for the comparison between Beggiatoa and Siboglinidae (βT = 0.5 ± 0.3) equaled the turnover value for the comparison of samples within the Beggiatoa habitat (βT = 0.5 ± 0.4). In contrast, the genus turnover within the Siboglinidae habitat was much lower (βT = 0.3 ± 0.1).
Additive partitioning of genus richness (α, β, γ) at the macro-scale at HMMV revealed that the beta fraction related to the difference between habitats contributed the largest proportion (58%) to the total diversity (Fig. 5).
Mega-scale diversity patterns
Several diversity indices were calculated for the nematode communities from the same habitat at the mega-scale: (i) reduced sediments; (ii) Siboglinidae patches and (iii) control (Table 2). For reasons of comparison, the micro-scale ROV samples from HMMV were excluded. Differences in turnover of nematode genus richness within the same habitat were significantly different at the mega-scale, with the highest difference for the turnover rate between the samples within reduced sediments and those within Siboglinidae (Kruskal–Wallis: P < 0.0001, Fig. 6), and the highest difference for the beta similarity between the samples within reduced sediments and those within control sediments (Kruskal–Wallis: P < 0.0001, Table 2).
The MDS plot of all samples (Fig. 7) depicts the high variability in the nematode communities inhabiting reduced sediments, which was already visible at the micro-scale. In contrast, the control sediments that were located at different water depths and sampled in different sampling years, show a remarkable similarity at the mega-scale.
The rarefaction curves for the total community (pooled data) associated with the reduced sediments, Siboglinidae, and control sediments (HMMV, Storegga and Nyegga) show that the expected number of genera for reduced sediments falls far below that of the Siboglinidae and the control sediments (Fig. 8). The largest increase in EG is noticed within the Siboglinidae by expanding the scale from micro- to macro-scale.
Micro-scale diversity patterns at HMMV
The different dive transects at HMMV showed a very high patchiness in habitat type and associated nematode assemblages, even on a meter scale. Furthermore, nematode communities were strongly influenced by the neighbouring habitat patches. In particular the Beggiatoa mats revealed a high variation in nematode community composition. One community was dominated by Halomonhystera disjuncta (see also Van Gaever et al. 2006), while another more diverse but less dense community dominated by Microlaimus was also defined in the present micro-scale study. The latter community was found in the small-sized (∼25 cm2) Beggiatoa patches which were surrounded by Siboglinidae patches within the northern active site and the larger Siboglinidae field. The close proximity of the Siboglinidae patches possibly generates a more intensive migration rate of nematodes towards the Beggiatoa patches. As Hodda (1990) stated, species most easily available to colonise a certain patch are those from the surrounding patches. In contrast, microbial mats within the extensive Beggiatoa-covered area (38,244 m2, Jerosch et al. 2007) were probably colonised less by nematodes from the more southern Siboglinidae fields. Moreover, de Beer et al. (2006) measured a strong depletion of oxygen in the large Beggiatoa area, combined with sulphide levels toxic to most metazoan life, resulting in the absolute dominance of a single nematode species (H. disjuncta). It could be expected that the sediments underneath the small Beggiatoa patches yield lower sulphide concentrations compared to the large Beggiatoa area. Nevertheless, the presence of these bacteria suggests a sulphide flux reaching up to the sediment surface.
Small-scale laboratory experiments with defaunated sediments demonstrated that the colonisation of new niches via lateral interstitial migration is a species-specific process (Schratzberger et al. 2004). Microlaimus was probably the most successful migrant of the small, isolated bacterial patches at HMMV. This opportunistic nematode genus has been recognised as a good coloniser in diverse marine habitats such as disturbed Antarctic sediments (Vanhove et al. 2000; Lee et al. 2001) and canyon sediments exposed to frequent turbidity events (Congo canyon; S. Van Gaever, J. Galéron, M. Sibuet, A. Vanreusel, unpublished observations). Ullberg & Olafsson (2003) showed that the settling choice of nematodes appeared to be species-specific; Microlaimus actively selected and colonised the benthic algae treatment in their laboratory experiment with high numbers.
Seep sediments with siboglinid tubeworms appeared to harbour much more diverse nematode assemblages compared to the Beggiatoa patches, even at the micro-scale. Almost twice as many genera were identified in the Siboglinidae cores compared to the Beggiatoa cores. Moreover, the evenness of the community of the pooled Siboglinidae samples was six times as high as that from the Beggiatoa community. As a result, the similarity between the nematode communities of the Siboglinidae and Beggiatoa patches was much lower than the similarity within the same habitat type. Nevertheless, the additive partitioning at the micro-scale defined the beta fraction due to the differences among cores within a single habitat as the most important contributor to the total genus diversity, pointing to a stronger turnover at micro-scale within a habitat than between habitats.
Macro-scale diversity patterns
Sampling the different seep habitats with multiple replicates at a macro-scale increased the number of nematode genera identified at the Siboglinidae habitat. The expected number of genera EG(600) for the Siboglinidae habitat increased 27% from micro- to macro-scale, whereas EG(600) did not change for the Beggiatoa habitat because of the very high dominance levels and low genus diversity in combination with very high numbers.
The genus richness generated by the different habitats at HMMV contributed the largest proportion (70%) to the overall genus diversity. On the other hand, the local (sample core) genus richness (alpha diversity) reached high percentages at HMMV when the Shannon–Wiener index was considered. This high alpha diversity (H’) can be explained by typical, common deep-sea genera (Vincx et al. 1994; Soetaert & Heip 1995; Vanaverbeke et al. 1997) regularly distributed across the different cores within a habitat, e.g. Acantholaimus, Halalaimus, Microlaimus, Metalinhomoeus, Thalassomonhystera and Tricoma, hence the high evenness. On the other hand, the high number of unique genera with restricted distributions might explain the significantly higher than expected beta diversity between habitats. Raes & Vanreusel (2006) reported significantly different nematode assemblages associated with different microhabitats (dead coral fragments, dead sponge skeletons, and underlying sediment) in a cold-water coral degradation zone in the Porcupine Seabight (NE Atlantic). In comparison with the HMMV, however, the partitioning of diversity for the Porcupine Seabight showed a much lower contribution of beta diversity attributed to the different microhabitats when working with genus richness (21%). However, a similar composition pattern of the gamma diversity between both studies was observed when nematode genus abundance data were considered, with the highest proportion for the alpha (or within core) diversity (M. Raes, unpublished observations).
The present investigation of nematode genus diversity at the Norwegian seeps showed a disparity between the importance of different habitats (beta diversity) at the micro-scale and the macro-scale. Beta similarity, in terms of nematode genus composition within a similar habitat, was much higher at the micro-scale than at the macro-scale. The degree of substitution of nematode genera within the Beggiatoa habitat was higher at the macro-scale than at the micro-scale. Interestingly, the beta turnover rate between the Beggiatoa and Siboglinidae habitat was also higher at the macro-scale. These patterns can be assigned to the patchy occurrence of nematode assemblages, even within a single habitat type. In particular the Beggiatoa habitat harbours two completely different nematode communities depending on the surrounding habitats and the geochemical sediment conditions. We conclude that nematode communities associated with different seep habitats were much better sampled and better defined at the macro-scale than at the micro-scale. Similar patterns were observed in other ecological studies covering a wide range of environments, from marine to terrestrial habitats (e.g.Lennon et al. 2001; He et al. 2002; Okuda et al. 2004; Zhang et al. 2006; Kallimanis et al. 2008). In general, these studies indicated that by using scales of observation that differ only by one order of magnitude, the spatial pattern of alpha diversity could be considerably different. Areas of lower species richness at micro-scale might appear high in species richness at the macro- or mega-scale. An increase of genera from 28 to 41 from micro- to macro-scale was recorded in the Beggiatoa habitat. Overall, the comparison between the micro- and macro-scale sampling of a seep demonstrated that small-scale sampling at a seep environment is insufficient to cover the total nematode diversity at the genus level in the different habitats, especially in genus-rich habitats.
Mega-scale diversity patterns
An increase of 66% in EG(600) from micro- to mega-scale was observed for nematode communities associated with the Siboglinidae habitat (Fig. 8), which can be explained by a high degree of small-scale variation resulting from highly patchy occurrences of individuals (Hodda 1990). Additionally, the average Bray–Curtis similarity within a habitat still decreased from macro- to mega-scale for both the Beggiatoa mats and Siboglinidae fields. Given their small body size and low mobility, nematodes are expected to be very susceptible to within-habitat physical changes. This, coupled with a limitation to long-distance dispersal and likely restrictions in gene flow (Derycke et al. 2007), also resulted in a significant decrease in nematode community similarity with distance (i.e. within 0.1–23 km) at two offshore muddy habitats off the west coast of the UK (Schratzberger et al. 2008). Temporal variation, due to the different years in which sampling occurred, is not taken into account in our study. Nevertheless, the control sediments which were located at different water depths and sampled in different sampling years, show a remarkable similarity at the mega-scale.
The nematode community at a single seep is composed of a number of very patchily distributed, different nematode assemblages, ranging from extremely low diversity communities in reduced sediments to highly diverse communities associated with siboglinid tubeworm fields. The HMMV Siboglinidae fields [H ’ (log-based e) = 3.03] and control sediments [H ’ (log-based e) = 3.05] investigated here exhibited nematode diversities similar to those reported for deep-sea sediments at a comparable depth in the Hausgarten site (Greenland Sea) [H’ (log-based e) = 3.29 ± 0.02] (E. Hoste, unpublished observations), and slightly higher than those documented from the Central Arctic [H ’ (log-based e) = 2.08 ± 0.16 at 1000–3890 m, H ’ = 2.42 ± 0.16 at 1020–2150 m] (Renaud et al. 2006). The nematode communities inhabiting these siboglinid fields were composed of nematode genera similar to those reported from soft-sediment deep-sea benthos at comparable depths (Vanaverbeke et al. 1997; Soltwedel 2000; Vanreusel et al. 2000; Renaud et al. 2006).
Despite the many inherent difficulties in evaluating diversity patterns and species richness in deep-sea ecosystems, it has been proven worthwhile to investigate the nematode diversity patterns in extreme environments such as cold seeps. The cold-seep environments, consisting of different habitats along the Norwegian margin, maintain nematode communities with some broadly distributed common deep-sea genera and a high number of rare genera with restricted habitat range. Our study suggests a considerable effect of habitat heterogeneity on the spatial structure and diversity patterns of nematode communities at the macro-scale within a seep. Some of the observed nematode assemblages were strongly restricted to a specific seep habitat. To assure the long-term survival of diverse deep-sea nematode assemblages, a certain diversity of suitable habitats is needed. The cold seeps located along the Norwegian margin offer a sufficiently large number of different habitats for nematodes in the otherwise rather monotonous deep-sea environment.
The authors thank AWI (Germany) and IFREMER (France) for a successful collaboration and for providing sampling facilities on board of the RV Polarstern and RV Pourquoi-Pas?. We thank the Captains, the ship crews, the ROV crew, and the head scientists Michael Klages and Hervé Nouzé for their efforts during the two sampling campaigns (ARKTIS XIX/3b in 2003, VICKING in 2006), and all other scientists onboard who have assisted during sampling. Thanks to Dr Vikram Unnithan, Jeroen Ingels, and the sixth FP IP HERMES for providing a high-quality map of the sampling area. Special thanks go to Dr Marleen De Troch, Prof. Lisa Levin and three anonymous reviewers for critically reading the manuscript and for making constructive remarks. This research was supported by the Ghent University, the sixth FP IP HERMES and MARBEF (Network of Excellence). This publication is contribution number MPS-08056 of MarBEF (Marine Biodiversity and Ecosystem Functioning).