Cold seeps of the deep Gulf of Mexico: Community structure and biogeographic comparisons to Atlantic equatorial belt seep communities

Cold seeps of the deep Gulf of Mexico: Community structure and biogeographic comparisons to Atlantic equatorial belt seep communities
of 17
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Deep-Sea Research I 54 (2007) 637–653 Cold seeps of the deep Gulf of Mexico: Community structureand biogeographic comparisons to Atlantic equatorialbelt seep communities Erik E. Cordes a, Ã , Susan L. Carney a,1 , Stephane Hourdez b , Robert S. Carney c ,James M. Brooks d , Charles R. Fisher a a Biology Department, Pennsylvania State University, 208 Mueller Lab, University Park, PA, USA b Equipe Ecophysiologie, CNRS-UPMC UMR 7127, Station Biologique, BP74, 29682 Roscoff, France c Coastal Ecology Institute, Louisiana State University, South Stadium Road, Baton Rouge, LA, USA d TDI-Brooks International, 1902 Pinon, College Station, TX, USA Received 4 May 2006; received in revised form 3 January 2007; accepted 5 January 2007Available online 25 January 2007 Abstract Quantitative collections of tubeworm- and mussel-associated communities were obtained from 3 cold seep sites in thedeep Gulf of Mexico: in Atwater Valley at 1890m depth, in Alaminos Canyon at 2200m depth, and from the FloridaEscarpment at 3300m depth. A total of 50 taxa of macro- and megafauna were collected including 2 species of siboglinidtubeworms and 3 species of bathymodiolin mussels. In general, the highest degree of similarity was between communitiescollected from the same site. Most of the dominant families at the well-characterized upper Louisiana slope seep sites of theGulf of Mexico were present at the deep sites as well; however, there was little overlap at the species level between theupper and lower slope communities. One major difference in community structure between the upper and lower slope seepswas the dominance of the ophiuroid Ophioctenella acies in the deeper communities. The transition between upper andlower slope communities appears to occur between 1300 and 1700m based on the number of shared species with theBarbados seeps at either end of this depth range. Seep communities of the deep Gulf of Mexico were more similar to theBarbados Accretionary Prism seep communities than they were to either the upper slope Gulf of Mexico or Blake Ridgecommunities based on numbers of shared species and Bray–Curtis similarity values among sites. The presence of sharedspecies among these sites suggests that there is ongoing or recent exchange among these areas. An analysis of bathymodioline mussel phylogeography that includes new collections from the west coast of Africa is presented. Thisanalysis also suggests recent exchange across the Atlantic equatorial belt from the Gulf of Mexico to the seeps of the WestNigerian margin. r 2007 Elsevier Ltd. All rights reserved. Keywords: Bathymodiolus ; Benthos; Biodiversity; Multivariate analysis; Phylogenetics; Vestimentiferans; Lamellibrachia ; Escarpia ; USA;Gulf of Mexico; Africa; Nigeria ARTICLE IN PRESS www.elsevier.com/locate/dsri0967-0637/$-see front matter r 2007 Elsevier Ltd. All rights reserved.doi:10.1016/j.dsr.2007.01.001 Ã Corresponding author. Current address: Organismic and Evolutionary Biology, Harvard University, 16 Divinity Ave., Cambridge,MA 01238, USA. E-mail address: ecordes@oeb.harvard.edu (E.E. Cordes). 1 Current address: Mote Marine Laboratories, 1600 Ken Thompson Parkway, Sarasota, FL, USA.  1. Introduction Cold seeps inhabited by dense macrofaunalcommunities are known from over 30 locations onactive and passive continental margins throughoutthe world’s oceans (Sibuet and Olu, 1998;Tyler et al., 2003). Additional studies continue to add tothe list of known seep sites with discoveries on theSunda Arc (Wiedicke et al., 2002) and Makranaccretionary prism (von Rad et al., 2000) in theIndian Ocean, as well as on the Gorda Escarpment(Stakes et al., 2002), in the Sea of Okhostk (Sahling et al., 2003), and near Lihir Island Papua NewGuinea (Schmidt et al., 2002;Southward et al., 2002) in the Pacific. Seeps supporting siboglinidtubeworms with symbiotic sulfide-oxidizing bacteria(vestimentiferans and pogonophorans) range indepth from the shallow waters (80m) of KagoshimaBay (Miura et al., 2002) to hadal depths in theKurile Trench (9345m,Mironov, 2000). Lush coldseep communities were first discovered in the Gulf of Mexico in the 1980s (Paull et al., 1984;Kennicutt et al., 1985) and are perhaps the most intensivelystudied of any cold seeps in the world. New sitescontinue to be found even in this well-exploredregion, with recent discoveries in Atwater Valley inthe northern Gulf of Mexico (Milkov and Sassen,2003;MacDonald et al., 2003) and the Campeche Knolls in the southern Gulf of Mexico (MacDonaldet al., 2004).The vast majority of cold-seep communities aredominated by the same families of symbioticorganisms, primarily the siboglinid polychaetes,bathymodioline mussels, and vesicomyid bivalves(Sibuet and Olu, 1998). The similarity among thesecommunities extends to the organisms inhabitingthe physical structure produced by the tubeworms,mussels, and clams; including the polynoid poly-chaetes, trochid gastropods, alvinocarid shrimp,and galatheid crabs among others (Sibuet andOlu, 1998). One of the persistent questions in thebiogeography of chemosynthetic ecosystems is thedegree to which the species are capable of dispersingamong relatively isolated cold-seep and hydrother-mal vent sites. To begin to address this question inthe seep setting, an adequate description of seepcommunities in well-defined areas is required. TheAtlantic Equatorial Belt, extending from the Gulf of Mexico through the Caribbean Sea and EquatorialAtlantic to the Nigerian Margin has been suggestedas a one of the target areas for these studies (Tyleret al., 2003).The Gulf of Mexico includes one of the mostgeologically active of the passive continental mar-gins of the world’s oceans. The geology of thenorthern Gulf is dominated by an underlying saltlayer formed during the successive opening andclosing of the connection between the Gulf and theAtlantic Ocean in the Jurassic, and potentiallyduring the Paleocene and Eocene as the Caribbeanplate moved to the east across the region (Rosenfeldand Pindell, 2003). Differential loading fromterrigenous sediment deposition on the continentalshelf over the incompressible salt sheet has causedthe formation of vertically mobile salt pillars anddomes (Humphris, 1979). As these structures impactthe sedimentary overburden, they compress oil-bearing reservoir sands and cause faulting in over-lying shale layers (Brooks et al., 1987;Aharon et al., 1992), resulting in the upward migration of hydro-carbons and brines (Kennicutt et al., 1988). Wherethe faults reach the sediment surface, they oftenform fluid and gas expulsion features: the seeps andmud volcanoes of the Gulf of Mexico.Most of the previous investigations of cold seepcommunity ecology in the Gulf of Mexico havefocused on sites in less than 1000m water depth onthe upper Louisiana slope (i.e.Kennicutt et al.,1988;MacDonald et al., 1990;Carney, 1994; Bergquist et al., 2003;Cordes et al., 2005;Cordes et al., 2006). This depth coincides with themaximum operating depth of the commonly usedJohnson Sea-Link submersibles. In deeper waters,there are published records from 5 cruises with theDSV Alvin at 7 sites in the Gulf of Mexico, with atotal of 58 submersible dives made to depths below1000m. Four of these sites contained vestimentifer-an tubeworms and/or bathymodioline mussels:Mississippi Canyon (MC 853,MacDonald et al.,2003), Atwater Valley (AT 425,MacDonald et al.,2003;Milkov and Sassen, 2003), Alaminos Canyon (AC 645,Bryant et al., 1990;Brooks et al., 1990; Carney, 1994), and Florida Escarpment (VN 945,Paull et al., 1984;Cary et al., 1989;Van Dover et al., 2003;Turnipseed et al., 2004). Here, we present the first data from quantitativecollections of the communities associated withtubeworm aggregations from greater than 1000mdepth in the Gulf of Mexico and new data fromcollections of mussels and associated fauna fromthese depths. We compare the tubeworm- andmussel-associated communities from 3 of the mostwell-developed seep sites known at water depthsgreater than 1000m in the Gulf: Atwater Valley, ARTICLE IN PRESS E.E. Cordes et al. / Deep-Sea Research I 54 (2007) 637–653 638  Alaminos Canyon, and Florida Escarpment. Thedistribution of the tubeworms and mussels formingthe foundation of this community along with thebiogeography of their associated communities arecompared among the cold seeps of the upper andlower slope of the Gulf of Mexico, Blake Ridge, andBarbados accretionary prism in the Atlantic Equa-torial Belt, a key area of interest in the explorationand documentation of the diversity and biogeogra-phy of chemosynthetic ecosystems (Tyler et al.,2003). To complement and augment these biogeo-graphic comparisons, an analysis of the phylogeo-graphy of bathymodioline mussels from collectionsin the Gulf of Mexico and from the Nigerian seepsoff the coast of Africa is presented. 2. Methods Intact vestimentiferan aggregations were ob-tained with the DSV Alvin using the BushmasterJr. collection device (Urcuyo et al., 2003;Bergquist et al., 2003) in October 2003. This device has anopen diameter of approximately 70cm and is linedwith a 63 m m mesh net. The contents of eachBushmaster collection were sieved through a 2mmmesh (the size fraction of macrofauna considered inthis study) and sorted to the lowest possibletaxonomic level on board the ship. Smaller sizefractions were retained for complementary investi-gations by collaborators and will be presentedelsewhere. Mussel bed samples were obtained withthe hydraulically actuated Harbor Branch Oceano-graphic Institution (HBOI) ‘‘clam shell’’ samplermanipulated by the Alvin. This sampler is identicalto the sampler used on the HBOI Johnson Sea Linksubmersible and similar to the ‘‘Pac Man’’ samplerused by the Canadian ROV ROPOS. It consists of two halves of a cylinder split lengthwise, enclosed inan aluminum frame. The two halves rotate to forma cylinder that encloses the sample. Each mussel bedwas sampled with 3–4 successive grabs of 342cm 2 surface area using this sampling device. Sampleswere processed similarly to tubeworm samples. Allassociated fauna were fixed in 10% bufferedformalin and preserved in 70% ethanol for trans-port back to The Pennsylvania State University andfinal determination of taxonomic status. E.E.C. andS.H. identified all polychaetes. Primary identifica-tion of other groups was carried out and specimenssent to experts for further identification or verifica-tion. Colonial or encrusting organisms contributedto species richness, but these were not enumeratedand, therefore, were not included in quantitativeanalyses.Tubeworms and mussels were counted andmeasured on board, time permitting. The remainderof each collection was preserved in formalin andtransported to the laboratory for processing. Tube-worm length was measured to a standardizedposterior outer tube diameter of 2mm. Thisdiameter was the common point where the tube-worms entered a dense, tangled mass at the baseof the aggregation. Surface area of tubewormswas calculated as for a cone frustrum (Bergquistet al., 2002). Mussel length and height weremeasured at their greatest points. A functionrelating length and height to surface area wasdetermined using a subsample of mussel shells.Eighteen shells were covered with a single layer of aluminum foil. The foil was weighed and the surfacearea of the foil was calculated based on the mass of a 1cm 2 piece of foil. This empirically determinedsurface area (sa) was related to the surface areacalculated as a cone including the base in aregression function:sa ¼ 1 : 053 ð p r 2 þ p r  ffiffiffiffiffiffiffiffiffiffiffiffiffiffi r 2 þ l  2 p  Þ þ 14 : 38,where r is one half of the height of the shell and l  isthe length (  p o 0.0001, r 2 ¼ 0.987). This functionwas used to determine the surface area of the rest of the mussels from their lengths and heights. Anumber of different functions relating height andlength to mussel volume (including treating them asrectangular and cylindrical) were tested, but theconical function provided the best fit to the data.Species abundances were standardized to tube-worm or mussel surface area to provide a measureof species density irrespective of collection size.Community diversity of each collection was as-sessed using the Shannon–Wiener diversity index( H  0 ): H  0 ¼ À X i   p i  ln ð  p i  Þ ,where p i  is the relative abundance (%) of the i  thspecies. Pairwise similarity in community structurebetween aggregations was examined using theBray–Curtis (BC) similarity index based on densityof fauna. BC similarity was determined using thefollowing function: Sjk  ¼ 100 1 À P  pi  ¼ 1 j  y ij  À y ik  j P  pi  ¼ 1 ð  y ij  þ y ik  Þ ! , ARTICLE IN PRESS E.E. Cordes et al. / Deep-Sea Research I 54 (2007) 637–653 639  where y ij  is the abundance of the i  th species in the  j  th sample and p is the total number of species.Tubeworms and mussels were not included in thisanalysis when they were the primary foundationspecies, although Bathymodiolus brooksi  is listedwhen it occurred within a tubeworm aggregation.Sub-samples of mussels from the Florida Escarp-ment were dissected to confirm the presence of thecommensal polychaete Branchipolynoe seepensis ,but because not all mussel individuals were dissectedin this study, B. seepensis is not included in thequantitative analyses. Pairwise similarity betweensites (upper Louisiana slope, deep Gulf of Mexico,Barbados accretionary prism, Blake Ridge) wasbased on presence/absence of species in the samplespresented here and as reported inOlu et al. (1996),Sibuet and Olu (1998),Bergquist et al. (2003),Van Dover et al. (2003),Turnipseed et al. (2004),Cordes et al. (2005), andCordes et al. (2006). Because the Barbados species inventories were primarily fromgrab samples and video and photographic records(Olu et al., 1996;Olu et al., 1997), only megafaunal taxa (those easily observed in video records or likelyto be retained by grabs) and mussel commensalswere included in these comparisons (Table 3).Potential shared taxa that were not identified tospecies level in previous studies were included in thisanalysis but are indicated by a question mark inTable 3. The number of shared species wasdetermined twice, once including these species andonce excluding them. Therefore, ranges in numberof shared species between pairs of sites are presentedbased on these differences in taxonomic resolution.In the statistical analyses, all potentially synon-ymous species were included; therefore, the resultsare not conservative, but rather are based on themaximum number of shared taxa between any twosites. Cluster analysis of pairwise BC similarityvalues was based on group average linkage inPRIMER software (Clarke and Warwick, 2001).Mussel tissue samples were obtained from theclam shell collections at the Florida Escarpment andAlaminos Canyon, and from Atwater Valley collec-tions in 2000 (dive 3633). Mussel mantle tissue wasdissected from the organism and frozen at À 80 1 Cuntil processing back in the laboratory. Samples of the ‘‘long’’ ( Bathymodiolus sp. A) and ‘‘short’’( Bathymodiolus sp. B) mussel types from theNigerian seeps were obtained in 50 Â 50-cm stainlesssteel box core samples (4 1 59 0 N, 4 1 08 0 E, 1700 and2200m depth). Mussels were frozen whole andshipped to the Pennsylvania State University foranalyses. DNA extractions were performed by astandard phenol-chloroform method (Ausubelet al., 1989). The mitochondrial cytochrome oxidasesubunit I (COI) gene was amplified using theprimers of Folmer et al. (1994)and sequenced bydye terminator cycle sequencing according to themanufacturer’s protocol (Beckman Coulter). Gen-Bank accession numbers for previously publishedsequences are as follows: B. brooksi   — AY649797; B. heckerae  — AY649794; Tamu fisheri   — AY649803; Idas macdonaldi   — AY649804; Bathymodiolus child-ressi   — AY649800; Benthomodiolus lignola  — AY275545. Sequences for the new Nigerian musselsequences in this study, as well as those for newlyidentified haplotypes of  B. brooksi  and B. childressi  ,have been submitted to GenBank (accession num-bersEF051241 – EF051246). Phylogenies were de- termined by the maximum parsimony and neighbor- joining methods of MEGA3 (Kumar et al., 2004)using Tajima-Nei distances. 3. Results and discussion 3.1. Site descriptions and communities Atwater Valley lies to the south of MississippiCanyon and represents a continuation of thisgeological structure extending from approximately1800m depth out to the abyssal plain (Fig. 1).Within the Atwater Valley (AT) 425 lease block(Table 1) is a large knoll representing a fluid releasesite on top of a shallow salt ridge (Milkov andSassen, 2003). Along the flanks of the knoll, atapproximately 1890m, are a series of ridges withextensive areas of seepage (MacDonald et al., 2003;Sassen et al., 2003). On the northwest slope, therewere mussels lining the channels between ridges(MacDonald et al., 2003). Approximately 1km tothe south, along the western slope, carbonateoutcrops occur on the crests of the ridges withsmall mounds similar to gas hydrate mounds andbubbling sediments indicating active seepage. Smalltubeworm aggregations and mussel beds werepresent in dark (presumably brine-stained) sedi-ments in depressions in this area.The community associated with one of the musselbeds was sampled at the AT 425 site. The bed wascomposed of  B. brooksi  , identified based onmorphological characters and mitochondrial COIgene sequences, which were identical to otherreported Gulf of Mexico B. brooksi  sequences(Fig. 2). This extends the known upper depth limit ARTICLE IN PRESS E.E. Cordes et al. / Deep-Sea Research I 54 (2007) 637–653 640  of  B. brooksi  from 2200 to 1890m. There were only2 species of macrofauna associated with this musselbed collection, the brittle star Ophioctenella acies and the shrimp Alvinocaris muricola (Table 2). Inaddition to mussel beds, a few Escarpia laminata aggregations were noted and an undescribed speciesof lamellibrachid tubeworm was collected duringthe same submersible dive (DSV Alvin dive 3918).This lamellibrachid may be the same Lamellibrachia sp. nov. reported from the Alaminos Canyon site(Nelson and Fisher, 2000), though additionalcollections coupled with morphological and geneticinvestigations are required to resolve the taxonomyof lamellibrachids at these sites.Alaminos Canyon cuts through the continentalslope off the coast of Texas between approximately1500 and 3000m, exposing alternating salt andcarbonate layers (Bryant et al., 1990). The seep sitesin Alaminos Canyon (AC 645;Table 1) are locatedat 2200m depth on the eastern side of the canyon inan area of relatively flat bathymetry (Fig. 1).Throughout the area is a thin authigenic carbonateveneer and scattered outcrops (Aharon et al., 1997)with isolated tubeworms and small to moderate-sized (less than 1m in diameter) aggregations. Themost well-developed area is characterized bymassive uplifted carbonate blocks (Roberts andAharon, 1994) associated with larger vestimentifer-an aggregations, mussel beds and areas covered indisarticulated mussel shells (Brooks et al., 1990).The undersides of some of the overhangs werecovered in a white crust (resembling methanehydrate) where methane bubble streams becametrapped. ARTICLE IN PRESS Fig. 1. Collection sites in Gulf of Mexico and locations of Blake Ridge and Barbados seeps. Contour lines are 1000, 3000, and 5000mbathymetry. Bushmaster and mussel bed samples were collected at Atwater Valley (1890m), Alaminos Canyon (2200m), and FloridaEscarpment (3300m). The other seeps of the western side of the Atlantic Equatorial Belt that have been sampled previously are alsoshown: Blake Ridge (2150m,Van Dover et al., 2003), and the seeps near Barbados including El Pilar (1300m,Olu et al., 1997), Orenoque A (1700m,Olu et al., 1997), Orenoque B (2000m,Olu et al., 1997), and the Barbados Trench sites (4850m,Olu et al., 1996). The upper slope sites of the Gulf of Mexico (ULS) lie just north of the Atwater Valley and Alaminos Canyon sites in water depths between 500 and800m. E.E. Cordes et al. / Deep-Sea Research I 54 (2007) 637–653 641
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks