Isotopic composition of chemosymbiotic fauna
The overall carbon isotopic composition of known symbiont-bearing fauna from the CMSA (δ13C range = −34.6 to −28.6‰) indicates that the main energy source for these taxa was derived from thiotrophic symbionts (i.e. sulfide-oxidizing bacteria; Cavanaugh et al. 2006). Thyasirid bivalves are known to harbor endosymbiotic bacteria in their gills (Dando et al. 2004) and also to have the capacity for heterotrophic feeding (Dando & Spiro 1993). In this study, the isotopic value of T. methanophila, δ13C = −34.0‰ is clearly divergent from photosynthetic values, indicating that this species relies on chemoautotrophic food sources, a similar result found based on scanning electron microscopy by Oliver & Sellanes (2005). Lamellibrachia sp. was less 13C-depleted (δ13C = −28.6‰ ± 2.2) than other symbiont-bearing fauna from the CMSA, but more 13C-depleted than previously reported for this site and species (δ13C = −22.8‰; Sellanes et al. 2008). The δ13C values of vestimentiferans are influenced by dissolved inorganic carbon (DIC) sources and their proportions, growth rate, and morphology of individuals (Fisher et al. 1997; MacAvoy et al. 2005). Vestimentiferans potentially take up DIC from both the sediment across their roots as well as from the water column through their plumes, each with its own carbon isotopic signature, leading to different tissue δ13C values even if the same energy source was used (MacAvoy et al. 2005 and references therein). There is also a great deal of spatial variation in DIC leading to broad ranges in δ13C values among Lamellibrachia individuals in New Zealand seeps (Thurber et al. 2010). At the CMSA, frenulate siboglinids (pogonophorans) also had an isotopic composition typical of reliance on thiotrophic bacterial symbionts (δ13C = −32.1‰, δ15N = 4.3‰; Levin 2005). Similar carbon isotope values have been found in frenulates living at methane seeps in the Gulf of Alaska (δ13C = −38‰ and δ15N = 2.4‰; Levin & Michener 2002), Sea of Okhotsk (δ13C range = −38.0 to −30.5‰; Sahling et al. 2003), Hikurangi Margin off New Zealand (e.g. −38.6 to −33.5‰; Thurber et al. 2010), which are slightly more enriched than values of Oligobrachia haakonmosbiensis obtained from Haakon Mosby Mud Volcano (−66.7‰, Lösekann et al. 2008). These were the first collections of frenulates from this region, although recent surveys have found extended fields of this potentially novel species. It is interesting that they have similar isotopic values to those individuals collected on the other side of the Pacific Ocean, suggesting similar roles in the benthic ecosystems. A series of video and images captured by the ROV Kiel 6000 (GEOMAR) suggests that this species is potentially consumed by the shrimp Haliporoides diomedeae or that the shrimp feed on organic matter trapped in the tubes (Fig. 5A).
Figure 5. Sea-floor images of fauna living in the CMSA taken with the Ocean Floor Observation System (OFOS; see Treude et al. 2011) and ROV Kiel 6000 during ChiFlux cruise. (A) Unidentified frenulate field and shrimp (Haliporoides diomedeae) feeding on them (ROV#9, SC = 21). (B) Gastropod Neolepetopsidae sp. probably grazing on epibiotic bacteria of the vesicomyid Archivesica sp. (ROV#11, SC = 24). (C) Unidentified asteroid in the vicinity of Lamellibrachia sp. aggregations (ROV#7, SC = 19). (D) Echinoids Sterechinus cf. neumayeri grazing on the tubeworms Lamellibrachia sp. (OFOS#1). (E) Echinoid Dermechinus horridus apparently grazing on bacterial mats (arrows) (ROV#11, SC = 24). (F) Macrurids and hagfish swimming over a clam field (ROV#2, SC = 14). Scale bars: 2.5 cm (A,B), 10 cm (C,D), 3 cm (D) and 10 cm (E). SC, Station code on the map and Table 1.
Download figure to PowerPoint
Incorporation of chemosynthetic production by heterotrophic fauna
The wide variation of δ13C (42.9‰) in the heterotrophic fauna, indicate that a diversity of carbon sources fuel this habitat. These carbon sources come from a variety of types of chemoautotrophic production (e.g. methanotrophy and thiotrophy) and photosynthetic production and supporting a high trophic diversity among consumers. Greater than 87.5% of the species studied registered δ13C values > −24‰, indicative of incorporation of photosynthesis-derived organic matter. The remaining 12% registered highly negative δ13C values (<−24‰), indicating some chemosynthetic carbon incorporation (e.g. methanotrophy ≤−40‰ and thiotrophy = −35.2 to −24.3‰; MacAvoy et al. 2002; and references therein; Table 2).
Gastropods at seeps and vent habitats have been shown to host symbionts, deposit-feed and prey upon other invertebrates. One relationship that has been enigmatic to understand is the relationship between neolepetopsid limpets and the variety of hosts that they live upon. Members of this family typically live on bivalves and vestimentiferan tubes (Sasaki et al. 2010), and the neolepetopsid limpets found at the CMSA occurred growing on the chemosymbiotic bivalve Archivesica sp. (Fig. 5B). In this study the δ13C and δ15N values of the neolepetopsids were even more 13C-depleted than the tissues of the clam Archivesica sp. (Table 2) and Fm values of these limpets suggest a considerable methane-derived carbon incorporation (Fm min, max = 28.7–52.9%), as would be expected if it grazed on a diverse food source including both free-living thiotrophic and methanotrophic bacteria or archaea (Cordes et al. 2010b). Some species belonging to this family, feed on small invertebrates (Warén & Bouchet 2001, 2009; Warén et al. 2006), yet this was unlikely, as the neolepetopsids had an isotopic composition indicative of a primary consumer. Neolepetopsids may graze on authigenic carbonates, feeding directly on microbes precipitating the rocks, which often have low δ13C values (Levin et al. unpublished). Aerobic methanotrophy could offer an alternative food source, similar to that observed in hydrothermal ecosystems off Costa Rica (Levin et al. 2012); although this metabolic pathway has yet to be characterized at the CMSA, free-living aerobic methanotrophs may be abundant in surface sediments (Lösekann et al. 2007).
Gastropods of the genus Provanna are among the most often observed taxa in vents and seeps (Warén & Bouchet 2001), and are considered deposit-feeders (Warén & Bouchet 1986; Warén & Pounder 1991). However, Provanna sp. from the CMSA had 13C-depleted values (δ13C = −35.2 ± 3.1‰) that were similar to those of symbiont-bearing fauna. Similar negative values have been reported for vent and seep congeners: Provanna sp. (Levin & Michener 2002), Provanna lomana (Smith & Baco 2003), Provanna sculpta (MacAvoy et al. 2005; Cordes et al. 2010b), Provanna variabilis (Bergquist et al. 2007), Provanna sp. (Olu et al. 2009), and Provanna laevis (Soto 2009), in all cases suggesting grazing on free-living bacteria (e.g. among mytilid beds, vestimentiferan tube-worms, and upon bacterial mats) or nutrition from endosymbionts (Smith & Baco 2003; MacAvoy et al. 2005; Bergquist et al. 2007). Nevertheless, this latter case is unlikely since the ctenidium of Provanna is not hypertrophied, unlike in Alvinoconcha and Ifremeria (Sasaki et al. 2010).
The amount of seep production incorporated by mobile predators (e.g. echinoderms, decapods, fishes) can be an indicator of the residence time of these species in the methane seeps (Carney 2010). At the CMSA, the asteroid Lophaster stellans and the echinoid Sterechinus cf. neumayeri clearly assimilated chemosynthetically derived carbon via sulfide oxidation (Table 3). Both species had isotopic values that reflected a mixture of seep and photosynthetic production. Predation on symbiont-bearing fauna (e.g. vestimentiferan polychaetes) could possibly occur in some asteroids of this area (Fig. 5C). However, in the case of Sterechinus cf. neumayeri they could be grazing on the Lamellibrachia sp. tubes (Fig. 5D) since these tubes often harbor free-living chemoautotrophic bacteria (Julian et al. 1999) that may be an important food source for this echinoderm taxa. Another option could be the consumption of bacterial filaments, as suggested by in situ images of the echinoid Dermechinus horridus within the area (Fig. 5E).
Trophic structure of benthic consumers
Although diverse macro- and megafauna are the primary consumers at the CMSA, a variety of species occupied higher trophic levels (Appendix 1). The use of δ15N to identify the trophic level has the underlying assumption that there is a constant δ15N at the base of the food web. This assumption at methane seeps is often not valid as a variety of sources, including some of the symbiont-bearing taxa identified here, had 15N-depleted values (e.g. Calyptogena gallardoi, δ15N = 3.5‰). These species could provide a potential food source for grazers and scavengers, therefore affecting their isotopic composition and lead to a consumer being assigned to a lower trophic level. An alternative explanation could be that the high percentage of microbial production, as well as the large physical size of certain bacteria in bathyal habitats (i.e. Gallardo 1977) leads to an expansion of the trophic niche in traditional carnivores to an omnivorous diet, where consumers can feed on both chemoautotrophic producers (e.g. bacteria that fix local nitrogen) and small invertebrates (Fanelli et al. 2009; Hoyoux et al. 2009). Further, it is worth noting that both deposit-feeding and suspension-feeding taxa had wide ranges of δ13C and δ15N, which could be associated with temporal and spatial variability in the quality and quantity of food (Iken et al. 2001) or the exploitation of organic matter in different stages of degradation (e.g. fresh detritus to highly refractory; Grémare et al. 1997). Moreover, anomalously low nitrogen isotopic composition likely indicates the incorporation of seep production, as seep microbes may use local nitrogen sources that result in low δ15N values (Levin & Mendoza 2007) or the fauna themselves may assimilate DIN or DON from the available N pool. On the other hand, bivalve species can consume particulate organic matter (POM) and also feed on dissolved organic matter; this last nutritional alternative also has been suggested for deposit-feeding species (Lopez & Levinton 1987) and numerous invertebrate phyla (e.g. Porifera, Annelida, Mollusca, Sipunculida; Sepers 1977). Therefore, the base of the food web at seeps is complex, where numerous food sources may be exploited by consumers and yet these are masked by shifting nitrogen isotopic values and divergent nitrogen isotopic ratios at the base of the food web.
Although applying isotopic trophic approaches to seep systems entails many challenges and assumptions, in most instances the results are consistent with the known natural history of the taxa considered in the analysis. Deep-sea chitons typically feed on sponges, hydroids and corals (Gracia et al. 2005); this agrees with the trophic positions estimated for all polyplacophoran taxa analysed in this study (i.e. Leptochiton sp., Leptochiton americanus, Stenosemus sp., and Polyplacophora sp.) (TP = 3.8 to 5.1; Appendix 1). The high trophic position of the polyplacophoran Placiphorella sp. in this area (TP = 5.1) may be associated with a peculiar morphological adaptation that it possesses to trap prey; this genus has an enlarged anterior girdle that is employed to trap mobile food sources, including other animals (McLean 1962; Eernisse et al. 2007).
The Patagonian toothfish Dissostichus eleginoides and the dusky cat shark Bythaelurus canescens were the top predators of the methane seep food web at the CMSA (Fig. 3). Both species are vagrants at methane seeps (sensu Carney 1994) and are likely attracted by the diversity of prey available in these environments (Sellanes et al. 2008). Indeed, D. eleginoides was significantly more abundant at the CMSA than at the surrounding area (Sellanes et al. 2012). Moreover, D. eleginoides was identified as a generalist consumer and omnivore based on its wider isotopic trophic niche (indicated by its larger SEAB area). This agrees well with observations of Flores & Rojas (1987), who concluded that this species is a non-selective carnivore that consumes a diversity of prey species (Garcia de la Rosa et al. 1997) with a seasonally variable diet (Murillo et al. 2008). In contrast, a more specialized feeding behavior was detected in the combtooth dogfish Centroscyllium nigrum, based on the estimation of the SEAB area. The trophic position estimated for C. nigrum (TP = 3.9; Fig. 2) agrees with the known trophic position of its congener Centroscyllium fabricii, TP = 3.8 (Cortés 1999). The blue hake Antimora rostrata occupied a lower trophic level (Appendix 1), probably reflecting a diet based on pelagic/benthopelagic and benthic prey, as has been observed in individuals inhabiting the Atlantic Ocean (Houston & Haedrich 1986). The consumption of carrion has been widely observed in this species (Collins et al. 1999; Yau et al. 2002; Kemp et al. 2006) and abyssal macrourids (Drazen et al. 2008; Jeffreys et al. 2010) linking benthic and pelagic trophic pathways (Iken et al. 2001). On the other hand, despite the absence of isotope values that reflect incorporation of chemosynthetic production in the fish species, the presence of numerous macrourids and hagfishes on the chemoautotrophic assemblages (i.e. clam field) suggests a potential ecological relationship between these faunal groups (e.g. trophic, refuge, nursery; Fig. 5E) as reported by Treude et al. (2011).
Food web length (FWL) is a measure of trophic links from primary producers to top predators in an ecosystem, and has been considered a key aspect of food webs, since it regulates a wide range of ecological processes (Kondoh & Ninomiya 2009), plays a role in regulating biogeochemical fluxes, and influences fisheries productivity (Vander Zanden & Fetzer 2007 and references therein). Although examples of FWL in benthic deep-sea environments are scarce, three trophic levels have been estimated in the Algerian Basin (650–780 m; Fanelli et al. 2009), Porcupine Abyssal Plain (~4840 m; Iken et al. 2001), and Håkon Mosby Mud Volcano (~1250 m; Gebruk et al. 2003), agreeing with the average FWL estimated for other marine ecosystems (3.97 ± 0.47; Vander Zanden & Fetzer 2007). However, for the bathyal food web at the CMSA, four trophic levels were estimated, as in the Arctic Canada Basin (625–3900 m; Iken et al. 2005) and the Blake Ridge methane seep (~2155 m; Van Dover et al. 2003). At present, it is recognized that heterogeneity generated by carbonate reefs and a number of associated bioengineer species (e.g. tubeworms, corals, sponges, bivalves, Levin & Dayton 2009; Cordes et al. 2010a; Govenar 2010), increase the richness and diversity in these environments, attracting a large number of non-seep taxa of different trophic levels (e.g. echinoderms, fishes, gastropods, shrimps). These taxa exploit chemosynthetic production and increased habitat diversity for the benefit of their feeding, in turn generating more complex trophic interactions at these sites over time.