The Gela Basin Pockmark Field in the Strait of Sicily (mediterranean Sea) Printer-friendly Version Interactive Discussion the Gela Basin Pockmark Field in the Strait of Sicily (mediterranean Sea): Chemosymbiotic Faunal and Carbonate Signatures of Postglacial to Modern Cold Seepage the Gela Basin Poc

Biogeosciences Discussions This discussion paper is/has been under review for the journal Biogeosciences (BG). Please refer to the corresponding final paper in BG if available. Abstract The geo-biological exploration of a pockmark field located at ca. −800 m in the Gela basin (Strait of Sicily, Central Mediterranean) provided a relatively diverse chemosym-biotic community and methane-imprinted carbonates. To date, this is the first occurrence of such type of specialized deep-water cold-seep communities recorded from 5 this key region, before documented in the Mediterranean as rather disjunct findings in its eastern and westernmost basins. The thiotrophic chemosymbiotic organisms recovered from this area include empty tubes of the vestimentiferan Lamellibrachia sp., loose and articulated shells of lucinids (Lucinoma kazani, Myrtea amorpha), vesi-comyids (Isorropodon perplexum), and gastropods (Taranis moerchi). A callianassid 10 decapod (Calliax sp.) was consistently found alive in large numbers in the pockmark mud. Their post-mortem calcified parts mixed with molluscs and subordinately miliolid foraminifers form a distinct type of skeletal assemblage (named DECAMOL). Carbon-ate concretions display δ 13 C values as low as −40 ‰ PDB suggesting the occurrence of light hydrocarbons in the seeping fluids. Since none of the truly chemosymbiotic 15 organisms was found alive, although their skeletal parts appear at times very fresh, some specimens have been AMS-14 C dated to shed light on the historical evolution of this site. Lamellibrachia and Lucinoma are two of the most significant chemosymbi-otic taxa reported from various Mediterranean cold seep sites (Alboran Sea and Eastern basin). E, −822 m) provided a sub-modern age of 484 ± 54 yr cal BP. These ages document that fluid seepage at this pockmark site has been episodically sustaining thiotrophic macrobenthic communities 25 since the end of the Younger Dryas stadial up to sub-recent times.

this key region, before documented in the Mediterranean as rather disjunct findings in its eastern and westernmost basins. The thiotrophic chemosymbiotic organisms recovered from this area include empty tubes of the vestimentiferan Lamellibrachia sp., loose and articulated shells of lucinids (Lucinoma kazani, Myrtea amorpha), vesicomyids (Isorropodon perplexum), and gastropods (Taranis moerchi). A callianassid 10 decapod (Calliax sp.) was consistently found alive in large numbers in the pockmark mud. Their post-mortem calcified parts mixed with molluscs and subordinately miliolid foraminifers form a distinct type of skeletal assemblage (named DECAMOL). Carbonate concretions display δ 13 C values as low as −40 ‰ PDB suggesting the occurrence of light hydrocarbons in the seeping fluids. Since none of the truly chemosymbiotic 15 organisms was found alive, although their skeletal parts appear at times very fresh, some specimens have been AMS-14 C dated to shed light on the historical evolution of this site. Lamellibrachia and Lucinoma are two of the most significant chemosymbiotic taxa reported from various Mediterranean cold seep sites (Alboran Sea and Eastern basin). Specimens from station MEDCOR78 (pockmark#1, Lat 36 • 46 10.18 N,

Material and methods
The study area was surveyed during cruises HERMES05 (July 2005), MEDCOR and DECORS December 2009 and August 2011 respectively of R/V Urania. Swathbathymetry data were acquired using a Kongsberg Simrad EM710 multibeam echosounder with a nominal sonar frequency of 70-100 kHz an angular coverage sector of 10 140 • , a beam width of 1 • × 1 • and 800 soundings per ping in dual mode. Chirp-sonar profiles were obtained using a hull-mounted sixteen-transducer source with a sweep modulated 2-7 kHz outgoing signal equivalent to a 3.5 kHz profile Bottom sampling were performed using a 1.2 ton gravity corer, large volume grab and epibenthic hauls (Table 1). 15 Water-column attributes were measured with a Conductivity/Temperature/Depth profiler (CTD) using a Seabird SBE 11 PLUS employing the SEASAVE V5.33 software.
Stable isotope analyses were performed at the Mass Spectrometry Laboratory of the CNR-IGG (Institute for Geosciences and Earth Resources) Pisa, using a Finnigan MAT252 mass-spectrometer. The carbonate samples were cleansed, powered and 20 then analysed following standard procedures by reacting carbonate aliquots with 100 % orthophosphoric acid under vacuum. Results are expressed as ‰ relative to VPDB (Vienna-Pee Dee Belemnite) standard by assigning a value of +1.95 ‰ ( 13 C) and a value of −2.2 ‰ ( 18 O) exactly to NBS 19 calcite; the analytical error is ±0.1δ (Table 2). 14 C-AMS dating was carried out at the Poznán Radiometric Laboratory, Poland (Ta-

The pockmark field of Gela
Pockmarks are common features at many continental margins in the Mediterranean basin (Stefanon, 1981;Stefanon et al., 1983;Curzi and Veggiani, 1985;Mazzotti et al., 1987;Hovland and Curzi, 1989;Trincardi et al., 2004;Geletti et al., 2008;Coleman et al., 2012;Curzi, 2012;Somoza et al., 2012). They have been recently documented 5 also in various locations of the Strait of Sicily, although most published studies regard relatively shallow-water (ca. 70-200 m) occurrences (Savini et al., 2009;Cangemi et al., 2010;Micallef et al., 2011). Pockmarks do occur in the Gela basin at bathyal depths (Minisini and Trincardi, 2009, this paper). Gela basin is the Plio-Quaternary foredeep of the Maghrebian fold-and-thrust belt 10 and is a site of widespread and repeated mass transport. Gela Slide, the largest masstransport complex in the area, has affected the northern margin of the basin mobilizing a volume of sediment in the order of 1000 km 3 in mid-Pleistocene times (Trincardi and Argnani, 1990). Following this major basin-wide event, several smaller-scales massfailure events impacted the area up to the Holocene (Minisini et al., 2007) and with 15 recurrence intervals in the order of few thousands of years between successive events (Minisini and Trincardi, 2009). The most recent mass-transport events involved, in various mixes, two basic distinctive kinds of slope sediments: the Pleistocene clinoforms that prograde with increasing slope angle and outbuilding the shelf seaward; and upper-Pleistocene to Holocene muddy contourite deposits characterized by water-saturated 20 and poorly consolidated sediment (Verdicchio and Trincardi, 2008). In all cases where stacked MTD deposits appear sandwiched within basin-wide draped units, the latter units appear disrupted by vertical features that have been ascribed to fluid-escape features from MTDs that became rapidly buried and perhaps pressurized (Minisini and Trincardi, 2009 data in 2005, the GBPF was then surveyed more in detail in winter 2009 and again in summer 2011 during Hermione cruise DECORS when high resolution multibeam data were acquired, and additional pockmark fields further west, identified. The resulting detailed morpho-bathymetric map documents the existence of discrete clusters of pockmarks in a restricted sector of the Gela basin (Figs. 1-4). Our data do 5 not show any clear preferential orientation of these features. All pockmarks in the studied field are clustered in seven seafloor patches for a total area of about 18 km 2 . The pockmarks presents generally a sub-circular shape, with the exception of the pockmark 1 (Fig. 2) which is elliptical-shape, and range in diameter from 40 to 310 m, although most of them are between 40 and 100 m and only 11 (Fig. 2) are between 200 m 10 and 310 m. The bottom of some of these is as much as 20 m below the surrounding seafloor; most however are less then 6 m deep. In cross section, extracted from the bathymetric profile, the pockmarks in most case appear to be U-shaped; but four of them are V-shaped and (Fig. 2) show a positive relief ranging from 4 m to 7 m along their rim. The side slope ranges from 7 • to 13 • reaching the maximum of 16 • . The 15 regional surrounding slope is about 2-3 • . Width to depth ratio were calculated for 11 pockmarks form bathymetric profiles crossed near the feature centres. The ratio fell between 11 : 1 and 40 : 1 with most near 18 : 1. There is not consistent relation between this ratio and relative size of the pockmark. Noticeably, acoustic signals relatable to the presence of gas entrapped in the sed-20 iment (such as BSR) or to active degassing (plumes) were not detected during both surveys (Fig. 5). Furthermore, CTD profiles show no T -S anomalies. The largest depression with an internal bulge, resulting from the coalescence of two pockmarks, was selected as the prime target for bottom sampling. The entire area is a fishing ground subjected for many years to a peculiar local fishing (mal)practice 25 known as cannizzi that litters the bottom by leaving behind scores of dead weights with long plastic threads attached (Taviani, 2010). Before proceeding with grab and core sampling, and CTD casting, it was therefore necessary to clear first the bottom by removing this litter. This goal was achieved by using a small modified haul that in fact successfully entangled some such cannizzi remains (threads and some weights), the former intensely overgrown by live cold water corals. A second trawl on the same trajectory intercepted the bottom collecting many dead shells of Lucinoma clams and tubeworms, living ghost shrimps and a few carbonate concretions amidst other benthic organisms and litter (Fig. 7a). Only at this stage, this large pockmark and some other 5 pockmarks in the area were tested with grab sampling and coring, integrating the exploration with CTD casts. Additional bottom samples were acquired in 2011 and on this occasion other pockmarks were also sampled.

Fauna
The living component in the benthic macro-and megafauna inside the pockmarks at 10 the time of sampling was rather scarce and mostly represented by sparse bivalves (Ennucula corbuloides, Cardiomya costellata, Kelliella miliaris), caprellid amphipods and ghost shrimps. Numerous specimens of a callianassid ghost shrimp (Calliax sp.: Fig. 7f, Table 4) were found in pockmark # 6. This is likely a still undescribed species morphologically close to an elusive callianassid taxon reported from shallower settings in the Western Mediterranean (de Gaillande and Lagardère, 1966). This species is consistently associated with such type of environments since a few specimens of this same shrimp have been also found in active pockmarks in the central Adriatic (Taviani et al., 2010). Members of family Callianassidae are known to inhabit deep sea reducing habitats worldwide (summarized by Komai and Fujiwara, 2012 (Fig. 7b), the sole species recorded that far from the recent 5 Mediterranean basin. Such bivalves have been recorded from deep sea reducing habitats in the Mediterranean basin (Corselli and Basso, 1996;Cosel and Salas, 2001;Olu-Le Roy et al., 2004;Werne et al., 2004;Duperron et al., 2007;Duperron, 2010;Ivanov et al., 2010a, b;Ritt et al., 2010;Brissac et al., 2011;Shank et al., 2011;Taviani, 2011;Rodrigues et al., 2012).
Another remarkable megabenthic organism found in the GBPF is a siboglinid tubeworm of which only empty tubes have been collected by us (Fig. 7a). The morphology of the tube resembles Lamellibrachia, a genus recorded from deep water cold seep sites (Olu-Le Roy et al., 2004;Werne et al., 2004;Duperron et al., 2009;Ritt et al., 2010;Hilário et al., 2011;Taviani, 2011), sunken wrecks (Hughes andCrawford, 2006;15 Gambi et al., 2011), andhydrothermal vents (Lott andZimmerman, 2012) in the basin. Tubes might belong to Lamellibrachia anaximandri Southward, Andersen and Hourdez (2011), described from the Anaximander mud volcano in the eastern Mediterranean (Southward et al., 2011), and possibly much widespread in the entire Mediterranean basin (references in Taviani, 2013). 20 In addition to these well known deep-water chemosymbiotic metazoans, we can further list few other taxa preferentially coping with reducing environments i.e. thyasirid (Thyasira, Mendicula, Axinulus, Axinus) and lucinid (Myrtea spinifera) bivalves (Dufour, 2005;Löffler et al., 2005;Taylor et al., 2007;Taylor and Glover, 2010;Brissac et al., 2011;Rodrigues and Duperron, 2011), and the turrid gastropod Taranis moerchi 25 ( Fig. 7d) that has been observed to be particularly common at such seep sites (Olu-Le Roy et al., 2004, M. Taviani, unpublished data). Only some pockmarks were found to contain spoils of truly seep communities that are instead totally absent from the seabottom outside the pockmarks (Appendix A). Furthermore, the more emblematic cold seep BGD 10,2013 The Gela Basin pockmark field in the strait of Sicily (Mediterranean Sea)  Roy et al., 2007;Lessard-Pilon et al., 2010). As it is often the case, the sustainability of such chemosymbiotic bivalves and siboglinids relies upon high sulphide concentrations since their metabolic pathway is thiotrophic. Thus, they cannot be considered as direct fingerprints of hydrocarbon availability in the seeping fluids at the GBPF.

Seep carbonates
10 -Sediments. The pockmark field affects a predominantly silicoclastic-muddy sector of the southern Sicilian margin. Sampling however reveals the occurrence of biogenic carbonate production in the form of skeletal components shed primarily by calcified benthic organisms and subordinately by pelagics. A distinct type of carbonate biogenic sediment is produced here through cold seep factories in the 15 form of medium to coarse skeletal assemblages enriched in shells of chemosymbiotic bivalves (mainly vesicomyids), and calcified parts of callianassid decapods. This remarkable sediment represents a novelty not contemplated in the current schemes of skeletal sediment classification (e.g., Hayton et al., 1995). While the geographic location of the study area would require placing related skeletal sedi-20 ment here among heterozoan carbonates, their genesis is in fact substantially unrelated to climate but to chemical carbonate factories. Therefore, we introduce the name DECAMOL (Fig. 7) to highlight the two major forming skeletal components in the sand to hash fraction, i.e. DECApods (callianassid: 30-50 %) and MOLluscs (20-40 %). Secondary components in order of abundance are miliolid foraminifers Introduction basin. Since deep-water cold seep habitats owing modern-type chemosymbiotic molluscs and seep-related decapods are geographically diffuse in the Ocean, Cainozoic to present, it is predictable that DECAMOL sediment would be identified at other sites beside the Strait of Sicily.

Age constraints
The absence of distinct signals documenting either active degassing or gas charged-10 sediment close to surface and the lack of live chemosymbiotic macro-and megafauna was taken as an indication that seepage backing the genesis of the GBPF is at present either dormant/episodic or ended. In order to unravel this fundamental aspect we examined the core record, and carbon-dated faunas and carbonate concretions. Superficial bottom samples and two sedimentary cores were taken from pockmark 1 and 6

15
(respectively MEDCOR81 and MEDCOR70) to reconstruct the recent history of fluid seepage at this place (see Fig. 11

Discussion
The pockmark area investigated in this study is located at the base of the northern slope of Gela basin where Chirp-sonar profile detect the top of the last basin-wide MTD at a depth of few tens of meters under a succession of basin-floor muddy turbidites. The pockmarks area are not randomly distributed but corresponds to a bulge 5 at the base of slope with a positive relief of few km 2 (Figs. 1 and 3) and the overlying succession, characterized by draped and onlapping plain-parallel reflectors (muddy turbidites?) thins toward the slope and are breached by the pockmarks (Fig. 6). The perfect matching of the pock mark field with the slide-related base-of-slope bulge is suggestive of a possible causal relationship. In this view the fluids emanating through 10 the pockmarks would be fed from the MTD, acoustically-transparent and possibly gascharged deposit. However, in the lack of deeper-penetration MCS profiles we cannot rule out the possibility that fluids come, at least in part, from beneath the MTD rather than from it. The genesis of the GBPF under study is thus not fully understood yet but a potential 15 link between defluidization (pockmarks) and one of the main motif of this part of the continental margin (mass sediment failure) merits to be considered. In fact, this cluster of pockmarks in the Gela basin are just located distally in front of a complex made up by stacked multiple submarine mass-transport deposits of late Pleistocene age (Minisini et al., 2007). These composite units are by large buried and only a few are exposed 20 at present on the seabottom, the best examples of which are the "twin slides" whose multi-phase emplacement occurred since the last Post-Glacial (Minisini et al., 2007;Minisini and Trincardi, 2009). Pore fluid gradients, resulting in fluid migration and eventual expulsion onto the seabottom, are suggested to be a potential co-driving factor, although not necessarily the dominant, in triggering episodic mass-transport events that 25 shape this region (Minisini et al., 2007;Minisini and Trincardi, 2009 two phenomena (twin-slide-slumping and pockmark-fluid expulsion) does exist. A series of still-unexplored pockmarks has been, however, identified at the toe of the "twin slides" (Minisini and Trincardi, 2009). It is predictable that they may in principle contain similar faunas and carbonates as their counterparts. The multiswath topographic map suggests the subsurface existence of now-draped lobate bodies northwards of the 5 pockmarks, that in all likeness represent buried mass-transport deposits comparable to those described by Minisini et al. (2007) from this general area. The mise-en-place process of the slumping units may in theory cause squeezing out of entrapped fluids or, otherwise, the underpressuring of fluids related to sliding off sediment that may be conducive to gas release pockmarking the bottom. Release of hydrocarbons has 10 been invoked as a likely mechanism to trigger submarine slumps (Carpenter, 1981 and many others). On-land, similar causative scenarios have been suggested to account for localized Miocene chemosymbiotic faunal assemblages and/or authigenic carbonates associated with slumped bodies in the Apennine chain (Berti el al., 1994;Conti and Fontana, 2002;Lucente and Taviani, 2005). 15 The GBPF shows a certain degree of homogeneity regarding the chemosymbiotic assemblages inhabiting individual pockmarks. This is especially true for small vesicomyids (Isorropodon), gastropods (Taranis) and callianassid decapods (Calliax) that are almost ubiquitous in our samples. Such homogeneity is not at all surprising once we consider the overall reduced variability of cold seepage at the GBPF when compared 20 to larger and heterogenous seep sites as those observable elsewhere in the ocean (Olu et al., 1997;Sibuet and Olu-Le Roy, 2002;Olu-Le Roy et al., 2007;Cordes et al., 2010), or even in the Mediterranean itself (Olu-Le Roy et al., 2004). Such variability often reflects primarily variations in the fluid regimes (Teichert et al., 2003). Nevertheless, minor faunal differences are equally observable at the GBPF such as for example the 25 comparable rarity of the large chemosymbiotic clam L. kazani or of the large vestimentiferan Lamellibrachia which are not recorded thus far from more than two pockmarks.
32 Fig. 6. Chirp sonar profiles across the pockmark-hosting morphologic bulge at the base of the slope showing the top of the last basin-wide mass-transport event (Gela slide?) and the thinning of plane-parallel reflectors (basin-plain fine-grained turbidites and hemipelagites) toward the bulge. Where thinner, this unit is likely easier to breach by the ascending fluids than in other areas.      Stable isotope compositions of bivalve shells (core MEDCOR 81), callianassid claw and authigenic carbonates all from st. MEDCOR70 and MEDCOR78 expressed in the conventional δ notation relative to the PDB reference standard.