Introduction

Carbonate mounds, also called coral build-ups, reefs, lithoherms or banks, are documented in the fossil record to have existed since the Proterozoic (Wilson 1975; Neumann et al. 1977; Bosence and Bridges 1995; Paull et al. 2000; Freiwald 2002). Their dynamics include the growth of living colonies and the build-up of in situ coral debris combined with the trapping of externally supplied sediment. This mixture of sediment and skeletal remains contributes to the formation of mound structures in up to 300-m-high and 10-km-long reefs which can grow in a wide range of water depths (between 50 and 1,300 m) and temperatures (between 4 and 12°C; Mullins et al. 1981; Frederiksen et al. 1992; Rogers 1999). The dominant cold-water coral species are Lophelia pertusa and Madrepora oculata; less common are species such as Desmophyllum cristagalli, Dendrophyllia cornigera, Enallopsammia rostrata and Solensmilia variabilis.

North Atlantic and Norwegian Sea carbonate mound areas have been intensively studied during the last two decades in the course of numerous marine expeditions (Hovland et al. 1994, 1998; Kenyon et al. 1998, 2003a; Henriet et al. 2001; Huvenne et al. 2003; van Weering et al. 2003; Kano et al. 2007; Wheeler et al. 2007). Various factors have been suggested to favour the initiation and development of these cold-water coral banks: (1) oceanographic conditions (mainly strong current activity, high nutrient supply and low sedimentation rates; Mullins et al. 1981; Kenyon et al. 1998; De Mol et al. 2002), (2) focused hydrocarbon seepage resulting in the fertilization of the seafloor (Hovland 1990; Hovland et al. 1994; Hovland and Thomsen 1997; Henriet et al. 1998, 2001; Hovland and Risk 2003) and (3) the availability of suitable substrates. Commonly, the settlement of coral colonies (mainly L. pertusa) takes place on hard substrata such as drop stones and rocky outcrops (e.g. Wilson 1975; Kenyon et al. 1998, 2003a) or eventually on methanogenic carbonate deposits which precipitate at seepage sites (Rogers 1999; León et al. 2007). Once initiated, the growth of the colony can become self-sufficient in terms of primary substrate; new corals grow on top of dead ones, the activity of the colony being largely controlled by ambient environmental conditions such as organic matter and sediment supply, current speed, pH, temperature and salinity. This fragile system can collapse when the equilibrium between sedimentation rate and biological growth is disturbed, and excessive sedimentation leads to a progressive burial of the coral colony (e.g. Kenyon et al. 2003a). However, if favourable conditions persist over sufficiently long periods of time, the carbonate build-ups can reach considerable dimensions. Conceptually, it seems probable that the growth of the mounds is further promoted by the presence of carbonate-cemented coral rubble mixed with trapped foraminifer-rich hemipelagic sediment. Such solidified deposits provide additional surfaces on which new coral colonies can develop. So far, this aspect has been examined only superficially.

The Logachev Mounds on the south-western Rockall Trough margin (Kenyon et al. 2003a; Mienis et al. 2007) represent a good example of an advanced stage of mound development. To date, the information on the internal structure of these and neighbouring carbonate mounds is limited (Expedition-Scientists-ODP-307 2005; Williams et al. 2006; Kano et al. 2007). High-resolution acoustic data obtained during surveys offshore Ireland revealed the presence of a strong reflector underneath the mounds (e.g. Kenyon et al. 2003a; van Weering et al. 2003). However, such profiles are often of little help because they usually show blank zones within these structures (e.g. Kenyon et al. 1998; Bailey et al. 2003; Huvenne et al. 2003). Thus, interpretations are overall limited to inferences based on shallow cores and visual observations.

Within this context, the oceanographic, geomorphological and biological conditions required for mound initiation and evolution have been investigated during several recent research cruises to the western Porcupine Bank offshore south-western Ireland (Fig. 1a). During the year 2000, the Irish National Seabed Survey (R/V Bligh) conducted by the Geological Survey of Ireland (GSI) completed multibeam and CTD surveys revealing, for the first time, the presence of a large cluster of inferred carbonate mounds on the western Porcupine Bank. In 2003, the area was once more visited during the Training Through Research 13 (TTR13) marine expedition of the R/V Professor Logachev, in the course of which detailed sampling, seafloor video recordings, and sidescan sonar and subbottom profiler surveys were performed. A preliminary report of this newly discovered deep-water coral mound province can be found in Kenyon et al. (2004).

Fig. 1
figure 1

a Locality map of the study area (red square), and location of the oceanographic transects (yellow band) in Fig. 5. b Multibeam bathymetric map of the survey area with sample stations (dots) and the outline of the sidescan sonar coverage (white dashed line). c Topographic profile (A–A′) along the edge of the canyon, showing the carbonate mound distribution. d Magnification of a part of the bathymetric map showing the locations of the CTD stations. e Nepheloid layers recorded in the study area (vertical profiles along B–B′ are from Dickson and McCave 1986), and acoustic single-beam water column imagery along profile C–C′ shown in b

Full size image

This paper provides a detailed presentation and interpretation of the sidescan sonar data from the western Porcupine Bank, supported by CTD and bathymetric data as well as selected seafloor video images. This is complemented by geochemical and petrographic analyses conducted on authigenic deposits associated with coral colonies sampled in the region during the TTR13 cruise. The aim of this paper is to evaluate the significance of the environmental controls in the growth of the coral colonies and the factors driving the formation of the authigenic deposits.

Geological and oceanographic setting

Off the southwest coast of Ireland, the Porcupine Bank is framed by the amphitheatre-shaped Porcupine Seabight to the south, and the Porcupine Ridge and Slyne Ridge to the north. The Porcupine Basin is influenced by strong bottom currents which have formed channel systems and sediment waves (McDonnell and Shannon 2001). Three major mound provinces have been described in the bight, these being the Belgica Mounds, the Hovland and Magellan Mounds, and the Pelagia Mounds (Wheeler et al. 2007). The present study focuses on an area located in the westernmost part of the Porcupine Bank where a series of canyons with mound features are located (Fig. 1).

The region is known to be influenced mainly by the northward-flowing Eastern North Atlantic Water (ENAW) and Mediterranean Outflow Water (MOW; White 2007) but is also influenced by minor water masses derived from the western North Atlantic, which are carried northwards by the eastern boundary slope current (van Aken and Becker 1996; De Mol et al. 2005; White 2007).

Materials and methods

In the year 2000, a multibeam survey with the R/V Bligh investigated the study area, mapping it with a deep-water multibeam Kongsberg-EM120 (12 kHz, 1° × 1°) echosounder system and a 3.5-kHz subbottom profiler. Bathymetric and backscatter data were processed by means of the Caris HIPS 6.0 software. From this, bathymetric derivatives such as digital terrain models, and shaded relief and slope maps were produced. Multibeam backscatter mosaics were generated by applying angular and slope correction techniques. Conductivity–temperature–depth (CTD) measurements were conducted in the region during the months of July and August. Discussed here are the results of four stations located within the area outlined in Fig. 1d.

In the year 2003 (TTR13 cruise), a survey using a MAK deep-towed sidescan sonar operating at 30 kHz frequency with a built-in 5-kHz subbottom profiler was performed in the vicinity of mounds located within the area delimited in Fig. 1b. Survey lines MAKAT-99 to MAKAT-104 were run parallel to each other from north–northeast to south–southwest, obtaining a complete coverage of approx. 200–250 km2 of seafloor at water depths varying from 600 to 1,150 m. Based on the sidescan sonar images, 21 sites of particular interest were selected and then surveyed with an underwater video system (Fig. 1b, Table 1), samples being collected by means of a dredge, gravity corer, box-corer, and camera-guided grab sampler.

Table 1 Sampling stations on the western Porcupine Bank with geographic positions, water depths, and short descriptions of the geographical settings and sedimentary characteristics. Letters after the station numbers indicate the sampling method: D gravity core, B box-core, Gr remote-controlled TV grab (double coordinates refer to start and end of the TV survey). Note that stations TTR13AT502B and TTR13AT503G were collected outside the mound area and are not shown in Fig. 1
Full size table

Semi-quantitative bulk mineralogical compositions of the authigenic deposits were determined by X-ray diffraction analyses, using a Dron-3 instrument with CoKα radiation by step-scanning at 0.02 × 2 Å with fixed counting times (Shlikov and Kharitonov 2001). The petrography of the hard-rock samples was studied in thin sections following standard petrographic and fluorescence techniques. Scanning electron microscope (SEM and BSE) studies were performed by means of an ISI ABT-55 SEM.

Aliquots for isotopic analyses were collected from various rock types with a hand-held micro-drill. Bulk portions of cement where drilled at spots selected to minimise contamination from any bioclasts present in the matrix cement. Measurements were performed on CO2 liberated by the standard phosphoric acid technique (MaCrea 1950; Sharma and Clayton 1965) using a Carbonate Prep System linked to an AP2003 mass spectrometer. The δ13C and δ18O results are reported relative to the V-PDB standard (Craig 1957; Coplen 1994); the precision of the analyses is ±0.2‰.

Oceanographic data from 19,801 stations covering most of the western Irish Margin were retrieved from the World Ocean Database 2005 (Boyer et al. 2006). These were analysed using the Ocean Data View software developed by the Alfred Wegener Institute, Bremerhaven (Schlitzer 2007).

Results

Acoustic data

Multibeam and sidescan sonar data enabled the identification of a cluster of steep-sided mounds 50 to 220 m high, with base diameters varying from 300 to 2,000 m along the western edge and in the central part of the survey area. The seabed dips gently towards the southwest and, as a result, the mound tops are located at depths ranging from 550 to 900 m (Fig. 1b, c). The mound area is delimited to the east by a long escarpment, and to the north and northwest by the canyon head and a carbonate ridge (see Fig. 2 and profile a′–a). Between the escarpments and the edge of the canyon, the slope gradients across the intermound region range between 1–3°. To the northwest of the canyon head, at similar water depths, the gradients are less than 1° (Fig. 1). These mounds are characterised by high backscatter commonly associated with reflections from steep slopes (slope gradients between 25 to 35°; Figs. 2, 3). Some of the mounds have several peaks. The flanks of some mounds appear to have smooth surfaces, whereas others appear to be hummocky, presumably due to the presence of coral colonies. Transparent acoustic wedges found at the bases of some mounds probably represent aprons of coral debris (Fig. 4), as observed also on video images. Sidescan and multibeam records show patches of high and irregular backscatter at the base of some mounds, which are interpreted as small landslides, hardgrounds or coral patches covered by sediment (Fig. 4).

Fig. 2
figure 2

Combination of bathymetry, slope gradients, and reflectivity data from the study area. a Bathymetry: 25-m grey-shaded relief. The topographic control of the mound province is clearly delimited. The line annotated FE indicates the faulted escarpment. Lines a–a′ and b–b′ indicate the locations of two profiles through the canyon and the carbonate mound field (see top and bottom of figure; m mound). The profile reproduced in Fig. 4 is marked by a thick dashed line. b Slope map in 25-m grey scale: light grey high slope (>1°), dark grey low slope (<1°). The N-S-oriented escarpment fault and the mound crests are clearly visible. Inset Rose diagram for mound crest orientation, white dashed line MAK-1 sidescan sonar coverage (cf. Fig. 1). c Reflectivity map (EM120 data) highlighting the patchy signature of the mound area, the very high reflectivity of the fault escapement, the low reflectivity of the intermound areas (loose fine sand, FS), the higher reflectivity of the iceberg-ploughed seabed and the more compacted sand (S) in the east, and the low reflectivity of mobile sands around mound clusters (e.g. arrow)

Full size image
Fig. 3
figure 3

a Bird’s eye view of the carbonate mound field (vertical exaggeration 1:4) based on multibeam echosounder and sidescan sonar data. It shows the pathways of bottom currents and sand transport, the locations of sediment drifts, the E-W-oriented carbonate ridge in the northern sector, and the two major carbonate mound clusters in the central sector. b Visualisation of the fault-related escarpment (vertical exaggeration 1:4) based on multibeam echosounder and sidescan sonar data. It shows the high degree of bedrock exposure (high acoustic backscatter) and small individual carbonate mounds growing along its edge. Also indicated are five sampling sites

Full size image
Fig. 4
figure 4

Slope-perched sediment drift: the area of low backscatter with seabed striations observed on the sonograph corresponds to the stratified, pinching-out units seen on the deep-tow subbottom profiler record. Sediment collapse on the mound slope can be observed on both the sonograph and the profiler record as thin, chaotic, wedging-out units (for profile location, see Figs. 2a and 3a)

Full size image

Most of the mounds are elongated in shape and have near-symmetrical slopes (Fig. 2b). Individual conical mounds were also recognised in the south of the study area along the edge of the canyon (Fig. 2). Towards the north, closer to the canyon head, mound distribution gets denser. Here the mounds have lower topography and form E-W-elongated ridge-like features.

The seafloor between these ridges is often characterised by sinuous bands of low backscatter, interpreted as pathways of bottom currents (Figs. 2, 3). This was corroborated by sediment samples collected during the cruise (Table 1, samples AT504G to 507 G) and by underwater video observations. The majority of the individual mounds and ridges are concentrated in two elongated clusters extending for 8 and 10 km in north–south and southwest–northeast directions. These clusters contain the largest mounds recorded in the area and are separated by one of the main sand transport pathways (Fig. 3a). A prominent E–W carbonate ridge rises up to 120 m above the surrounding seafloor and extends for more than 5 km along the western flank of the canyon head in the northwest of the study area (Fig. 3a). On the acoustic data the ridge is seen to have high backscatter levels suggesting a hard-bottom seabed. This high backscatter signature extends northwards of the ridge for hundreds of metres onto the flat seabed, becoming gradually irregular and patchier. Beyond this patchy zone, the seabed is characterised by mostly lower backscatter which is attributed to a sandy bottom. This ridge is interpreted to represent a carbonate build-up similar to the ones seen along the eastern flank of the canyon head. The distribution of the mounds, however, differs from that along the eastern flank as the ridge represents the only major topographic feature running predominantly along the edge of the canyon head. Small E-W-aligned clusters of low mounds (25–50 m high) can be found a few km north of the ridge (Figs. 1, 2). The eastern flank, by contrast, is characterised by the presence of several mound clusters oriented predominantly normal to the canyon edge (Fig. 3b).

Besides the described clusters, a system of escarpments was observed extending from south to north for up to 10 km in the north-eastern part of the study area at 600–700 m water depth (Fig. 1). These features are up to 60 m high and are separated by terraces characterised by moderate backscatter levels. The main lineament appears segmented into a number of curvilinear escarpments with high backscatter ascribed to their steepness and the presence of bedrock outcrops (Fig. 3b). In some places the high backscatter is disrupted by areas of low relief and patchy backscatter signatures. Multibeam data show that these discontinuities are associated with individual small carbonate build-ups growing along the edge of the escarpment. Trends of individual ridges within the main clusters (Fig. 2b) suggest that they also grow along a series of characteristic crescent-shaped escarpments. The crescent-shaped geometry indicates that, as in the case of the other escarpments, these are the seabed expression of a regional listric fault system.

On multibeam and sidescan sonar images, the south-eastern part of the study area is generally characterised by a smooth relief and homogeneous low backscatter levels (Fig. 2), implying a transitional thickening of the soft sedimentary drape towards the south. The 5-kHz subbottom profiler records show that the sedimentary cover often thickens at these locations. Here, the seabed has a convex morphology which onlaps onto the underlying sediments. Where the thickness is lower, erosional features are visible. Such sediment distributions and acoustic signatures indicate the presence of several small, fine-grained drifts formed by the northward-flowing bottom current (Fig. 4).

Oceanographic data

The regional, E-W-oriented temperature and salinity profiles across the area indicate the existence of a major interface between Eastern North Atlantic Water and Mediterranean water masses at about 1,100 m water depth (Fig. 5). CTD profiles taken from the southern (000730) and northern parts of the study area (000727) have a similar structure, showing a salinity peak associated with MOW between 900 and 1,150 m (Fig. 6). This corresponds to the lower limit of mound distribution. Furthermore, in the vicinity of the mound region, an irregular salinity pattern is observed over the depth range of the mounds (600–1,000 m). This could be caused by MOW mixing with the ENAW, but may also be induced by the locally irregular topography within the deep-water canyon system. Such irregular water mass patterns are not detected in the profiles of stations in deeper water to the southwest (000722) and west (000804) of the study area. Salinity plots of the latter two stations show a smoother transition between the ENAW and the MOW, suggesting that the MOW signature becomes more mixed in deeper water.

Fig. 5
figure 5

Oceanographic transects across the study area (see location in Fig. 1, yellow band) based on information extracted from the World Ocean Database 2005, GEBCO and GSI multibeam bathymetry data (cf. BODC 2003). Note the depth range of coral mound distribution

Full size image
Fig. 6
figure 6

Salinity profiles from four stations (see locations in Fig. 1d) collected during 2000 (last four digits of each station represent the month and day of sampling) showing the distribution of different water masses in the study region; the stations deployed in deeper waters to the southwest (000722) and west (000804) of the study area show an irregular water column structure in the depth range of mound occurrence (cf. main text for details)

Full size image

Profiles of oxygen, nitrate, phosphate, salinity and temperature show another interface at about 600 m water depth, which corresponds to the upper limit of mound occurrence (Fig. 5). The oxygen profile also reveals a marked oxygen minimum zone (OMZ) between the two interfaces.

Authigenic deposits: morphology and petrography

At many of the sampling stations (Table 1, Fig. 1), mostly rounded gravels and boulders (up to 0.5 m in diameter) of different lithologies were retrieved, indicating that the entire region has in the past been exposed to the discharge of ice-rafted material. More detailed sampling on selected carbonate mounds and along the north-eastern escarpment (Table 1) revealed the presence of two main rock types (carbonate, phosphate) associated with outcrops of authigenic deposits (Table 2). These rocks are all heavily bored, a feature ascribed to annelids and crustaceans coexisting with sponges and other encrusting organisms.

Table 2 Summary of XRD and isotope analyses on cementing groundmass samples. Alb Albite, Arag aragonite, Cal calcite, Chl chlorite, Dol dolomite, Flap fluorapatite, Goe goethite, Hal halite, Ill illite, MgCal magnesium calcite, Micr micrite, Plag plagioclase, Qz quartz
Full size table

Carbonate deposits

Carbonate deposits were retrieved in three distinct settings: the escarpment in the eastern part of the survey area, the upper parts of carbonate mounds, and the flanks of carbonate mounds.

The escarpment along the eastern side of the study area (Fig. 3b) was initially surveyed with an underwater video camera, which revealed irregularly shaped carbonate blocks with bored surfaces. Occasional small colonies of living corals were observed on these surfaces. These blocks consist of porous micritic calcite cementing foraminiferal (planktonic and benthic) and coccolith ooze. Some better lithified samples (e.g. station AT500D, Fig. 7a–d) show broadly distributed glauconite peloids and bryozoan remains; others samples (stations AT498D, AT499D, AT500D, Fig. 7e, f) are less lithified and lack any signs of glauconite.

Fig. 7
figure 7

Escarpment area. a Hand sample of lithified carbonate rock (station AT500D); the greenish colour denotes a high glauconite content. b Thin section plane light image showing glauconite aggregates often coated by organic matter in a foraminifer- and organic-rich micritic calcite groundmass. c SEM image showing some glauconite aggregates in micritic calcite groundmass. d Thin section plane light image showing bryozoans (arrow) in micritic, organic-rich groundmass with benthic and planktonic foraminifers and larger bioclasts. e Hand sample of less-lithified carbonate rock; note the greyish colour and the numerous borings. f SEM image of the internal part of the sample in e showing coccolith-rich groundmass devoid of glauconite. g Large block of dolomite (dredged). h Thin section plane light image showing fine-grained dolomite cement with microfossil ghost features (centre)

Full size image

Large reddish-pink elongated blocks of dolomite were also collected from the escarpment (Fig. 7g; e.g. station AT498D). Although borings have affected the outer surfaces of these samples, epibenthic biota were rarely observed. Thin sections and XRD analyses revealed the presence of a fine-grained dolomite groundmass with coarser-grained crystals occupying the voids. The observed microfossil ghost features are consistent with re-cementation of older hemipelagic deposits now exposed on the seafloor (Fig. 7h).

Video images from the upper parts of carbonate mounds showed dense thickets of living white and pink corals abundantly colonized by various species of echinoderms and crustaceans up to an approx. water depth of 1,125 m (Fig. 8a). Thickets of L. pertusa, M. oculata and occasional D. cristagalli were observed to populate well-cemented carbonate blocks dredged at station AT509D (Fig. 8b).

Fig. 8
figure 8

a–c Top of coral mounds. a Dense thicket of corals (scale: ca. 1 m). b Hand sample consisting of micritic calcite cementing foraminiferal ooze; note the numerous borings and attached living corals. c Thin section plane light image showing micrite-cemented foraminiferal ooze with locally high concentrations of organic matter (dark colour). d–g Flank of coral mounds. d Extensive carbonate deposits rich in coral debris associated with attached thickets of living corals (scale: ca. 1 m). e Detail of a partly dissolved coral fragment extracted from carbonate deposits. f Hand sample of micritic calcite cementing hemipelagic sediments rich in coral and shell debris. g Thin section plane light image showing micritic calcite cementing hemipelagic sediments rich in microfossils (mostly planktonic) and bioclasts (e.g. shell fragment to right)

Full size image

Petrographic observations are similar to those made on rock samples from the escarpment region, micritic calcite being the main cement binding together coccolith and foraminiferal ooze but, in this case, with high contents of sponge spicules and partly dissolved coral branches. Organic matter and aggregates of pyrite framboids are concentrated in defined areas of the rock samples (Fig. 8c). A similar type of carbonate lithology was retrieved at station AT494G, representing the deepest site sampled in the canyon.

The video surveys run on the flanks of carbonate mounds revealed the presence of irregular blocks protruding from fine-grained sediment together with scattered bioclasts and ice-rafted debris. Dredge hauls from this area (stations AT509D, AT501D) retrieved rocks similar to those described from the tops of the mounds, except that the corals show evidence of much stronger dissolution (Fig. 8d–f). Noé et al. (2006) and Mienis et al. (2009) possibly describe similar authigenic deposits from carbonate mounds of the NE Atlantic, which contain carbonate-cemented layers with dissolved coral branches.

Phosphatic hardgrounds

The video record run along the escarpment shows widespread slabs populated by abundant fauna, in places including small colonies of living corals. The large slabs retrieved from dredge hauls have a yellowish-red colour, and are very well lithified and heavily bored (stations AT498D, AT499D, AT500D; Fig. 9a). Freshly broken surfaces reveal a greenish internal structure consisting of cm-sized, subrounded brownish intraclasts and large (>0.5 cm) bioclasts supported by a yellowish-green groundmass (Fig. 9b).

Fig. 9
figure 9

Escarpment area, phosphatic hardgrounds. a Large slab; note the intense boring and fauna colonizing the external surface. b Freshly broken surface of a slab showing yellowish-green groundmass and subrounded cm-sized light- and dark-coloured intraclasts. c Thin section plane light photo mosaic showing groundmass and intraclasts both consisting of collofane containing glauconite, bioclasts and clay

Full size image

Petrographic analyses (Table 2) show the cement of the microfossil-rich groundmass to consist of collophane binding together hemipelagic sediment, microfossils, apatite and sparsely distributed glauconite aggregates, the latter conferring the greenish colour to the samples (Fig. 9c). A similar composition is recorded for the lighter-coloured intraclasts (Mg-calcite 33.4%, fluorapatite 66.6%) as well as the darker ones (Mg-calcite 18.2%, fluorapatite 81.8%; Fig. 9c). The intraclasts typically contain higher amounts of fluorapatite, and occasionally also hemipelagic clay, hence the darker colour in comparison to the groundmass. Bioclasts (e.g. bivalves, corals, fish teeth, microfossils) are scattered throughout the sample. Microfossils are mostly disrupted within the intraclasts, in contrast to their good preservation in the groundmass. In addition, the contact intraclast–groundmass shows truncations of shells and glauconite, indicating prior transportation before their final deposition (Fig. 10a, b). In some instances the edges of irregularly shaped intraclasts reveal the presence of cross-cutting glauconite pellets on the intraclast margins (Fig. 10c, d), which is consistent with continued local in situ diagenesis. These intraclasts have sharp margins defined by a zoned rim of collophane cement (Fig. 10e, f).

Fig. 10
figure 10

SEM-BSE images of phosphate deposits. a Contact between nodule (top) and groundmass; sharp truncations of the glauconite and microfossils (arrowed) suggest fracturing and transportation of this nodule. b Groundmass and nodule (bottom right, dotted line); the truncated glauconite and foraminifers (arrowed), the fragmented microfossils within the nodule, and the better preserved microfossils in the groundmass suggest that the nodule was not fully lithified when transported to its final place of deposition. c, d Contact between nodule (left) and groundmass with well-preserved glauconite cross-cutting the margin between the two features. e Magnification of the area framed in d; note the lighter backscatter of the diagenetic rim framing the nodule. f Magnification of the area framed in e; zonations within the rim suggest variations in diagenetic conditions during rim formation

Full size image

Microfossil identification suggests a relatively recent age for these hardgrounds (Early Pliocene), as indicated by the presence of foraminifers such as orbulinids and numerous thin-walled Globigerinoides.

Carbon and oxygen stable isotopes

The majority of the analysed samples (Table 2) have small variations in stable isotope composition, with δ13C values between 0 and 2‰ and δ18O values between 3 and 4.6‰. The dolomite samples all have negative δ13C (−1.0‰) and δ18O (−4.1‰) values.

Discussion

Interrelationships between carbonate precipitation and oceanography

Sonar images and sediment samples taken around the mound area suggest that development of a small sediment drift is currently taking place. This is consistent with the observations of bottom current activity on the video footage of the seafloor surrounding the mounds. Furthermore, measurements conducted by a current-meter array a few kilometres farther south revealed average bottom current speeds of 30–50 cm/s (Dickson and McCave 1986), which are capable of transporting up to medium-grained sand. The interpretation of the acoustic data agrees with that of previous studies in other carbonate mound areas along the Irish margin (Kenyon et al. 1998, 2003a; Wheeler et al. 2005), where the low backscatter has been ascribed to the accumulation of fine- to medium-grained sand along the bottom current pathways (Kenyon et al. 2004). This interpretation indicates that the current from the south is responsible for the accumulations of lag sands and perhaps also for the transport of nutrients necessary for coral growth. It is suggested that the colonisation started from isolated symmetrical build-ups which then spread along the edge to form elongated complexes of merged mounds, the growth of which was facilitated by strong geostrophic flow along the escarpment and by tidal currents running in an E–W direction. Outcrops of basement rock, which occur along such escarpments, provide the necessary substratum for coral settlement (Wilson 1975). Mound growth is then facilitated by bottom currents and the hydrodynamic regime in the canyon where local current speed enhancement can be expected to occur due to the irregular topography around the edge of the canyon.

According to the sand transport pathways identified on the MAK-1 sonographs, prominent carbonate mounds also contribute to the topographic steering of the current. Abundant linear irregularities of high backscatter recognized on the sonographs are interpreted as half-buried escarpments located under the drift sequence along the southern extension of the previously described fault system (Fig. 2). The gradual northward variation in mound alignment from NW–SE to E–W, their shape evolution from conical to elongated ridge, and their increase in abundance emphasise the importance of the canyon head in enhancing the hydrodynamic conditions which then favour mound growth. Similar conditions within canyons funnelling nutrients and sediments to the benefit of coral growth have been reported by de Stigter et al. (2007) and Orejas et al. (2009). These observations, in combination with video surveys, also suggest that numerous coral colonies were initially emplaced along regional fault-related escarpments which provided a hard substratum suitable for later growth and preferential southward expansion into the current. A similar type of accretion was reported by Paull et al. (2000) from the Florida–Hatteras Slope.

Different mound distribution patterns along the western and eastern flanks of the canyon head may indicate the presence of cascading bottom flow (Shapiro and Hill 1997), which should be more intense on the eastern flank, thereby favouring mound growth. The western flank is less affected by the current, and the dominant process promoting coral growth there is probably the edge-related enhancement in turbulence.

The presence of several interfaces within the water column—e.g. between the MOW marked by the salinity anomaly, the OMZ and the ENAW—facilitates the widespread development of nepheloid layers in the depth interval between 300 and 1,000 m (Dickson and McCave 1986; Rice et al. 1991; De Mol et al. 1998). These nepheloid layers are an important source of detrital organic particles which increase the food availability for bottom fauna (Frederiksen et al. 1992). These layers spread out along the interfaces between the different water masses in the form of bottom-attached and intermediate nepheliod sheets (BNL and INL; Dickson and McCave 1986).

The development of nepheloid layers can be facilitated by a number of processes, including localised downwelling events following slack periods of tides (Davies et al. 2009, Mingulay Reef off NW Scotland), tidally pulsed Ekman drainage (driven by cascading of denser trapped winter water on the bank) which flushes the nutrient- and organic matter-rich ‘soup’ from shallow waters downslope into the bathyal realm (White et al. 2005), and others. To identify the process which is most active in the area lies beyond the scope of this paper, although prominent tidal influence can be inferred from analogy with other mound areas in the region (Dorschel et al. 2007; Mienis et al. 2007).

Both INL and BNL layers were initially described for the study area on the basis of CTD profiles (Dickson and McCave 1986; Fig. 1e). More recent single-beam surveys have provided acoustic images of three such layers in the north of the study area (Fig. 1e). These layers occur in the same depth range as the mounds (600–1,000 m; Fig. 1) and are therefore likely to be responsible for the enhanced coral growth.

The present findings are consistent with strong current activity facilitating the dissolution of aragonitic bioclasts and the precipitation of calcitic micrite. The observed aragonite dissolution (mainly from corals) operates as carbonate source for the precipitation of micritic calcite (e.g. Mazzullo 1980; Lasemi and Sandberg 1984). This process is favoured where coarser bioclasts are present and where a higher flux of water can be flushed through the pores to improve ion transport under oxic conditions. van der Land et al. (2010, 2011) have also demonstrated that carbonate precipitation is greatly enhanced by the presence of coccoliths which form the nucleation sites for calcite precipitation. In contrast to the previous sampling by Mienis et al. (2009), where carbonate-cemented intervals were cored in the subsurface, the present findings show that these rocks actually crop out on the seafloor. The model proposed herein is thus consistent with observations at the top and along the flanks of the carbonate mounds where coarse coral debris are present and where the currents are stronger, favouring better cementation and enabling the cemented blocks to remain exposed. The existence of such localized hard substrata would have an active role in promoting the continuous growth of the mounds.

The formation of carbonate deposits at the base of the mounds is unlikely to be due to high sedimentation rates or the supply of large amounts of coral debris from the upper part of the mounds. These conditions would not be favourable for carbonate dissolution and further precipitation, as confirmed by core AT495G (see log in Kenyon et al. 2004) recovered from the moat of a multi-peaked carbonate mound. The 5-m-long core penetrated a sequence rich in coral debris intercalated with clay or sand beds and occasional pebble layers. Similar cores from the bases of other mounds along the Irish margin are documented in Kenyon et al. (1998, 2003b). This implies that periods of fast growth (and higher rates of debris production) of the coral colony alternate with periods of dominant hemipelagic and terrigenous sedimentation (and perhaps erosional events). The presence of water-saturated coral debris layers suggests rapid burial, presumably occurring when the angle of repose of the mounds is exceeded, causing the collapse of the mounds accompanied by reduced coral growth. The absence of lithified hemipelagic sediment in the core strengthens the contention that carbonate precipitation occurs only at specific sites.

Interrelationships with phosphatic hardgrounds

As stated in the Introduction, the presence of phosphate deposits in carbonate mound regions of the Irish Margin has been poorly documented so far. Phosphatic hardgrounds were particularly common during the Cretaceous and are today cropping out in southern England (e.g. Kennedy and Garrison 1975a, 1975b), in Northern Ireland (Marshall-Neill and Ruffell 2004), and at various places along the Irish west coast (Wilson 1975). More recently, Haughton et al. (2005) documented the stratigraphy of three borehole sites on the western slope of the Porcupine High. At borehole 83/20-sb01, a thin phosphatised crust was cored between Maastrichtian chalky micrites and the overlying Miocene silty clays. However, it is currently unknown whether deposition of phosphate-rich sediments recurred in the recent past. The textural and petrographic characteristics of the hardgrounds in the present study area are very similar to those described by Haughton et al. (2005). By contrast, the Early Pliocene age estimated here indicates that they are much younger. It cannot be excluded that these laterally extensive hardgrounds correspond to the erosional unconformity observed at the base of mounds in the Porcupine Seabight and on the Rockall Bank (e.g. Stoker et al. 2005), but more data are required to verify this hypothesis.

Based on the present findings, the phosphatic hardgrounds could be the result of two different depositional processes. The rounded intraclasts, which contain more fragmented fossils and broken shells and minerals, represent the initial deposition on the upper part of the escarpment. These poorly lithified clasts were then locally moved along the seabed before they became fully cemented within the lighter-coloured groundmass. The well-preserved microfossils within the collophane groundmass suggest that cementation occurred at or close to the sediment–water interface.

In most cases, this type of hardground occurs as a consequence of the deposition of large amounts of organic material, this being frequently related to upwelling currents. The onshore (UK, Ireland) record of such deposits in the geological history indicates a recurrence of such events. Dinoflagellate blooms associated with major benthic and pelagic mortalities related to small-scale coastal upwelling events have indeed been documented along the southwest coast of Ireland (e.g. O’Boyle 2002; Silke et al. 2005), although of considerably smaller magnitude than those occurring, for example, off south-western Africa (e.g. Carr 2002; Weeks et al. 2004; Santana-Casiano et al. 2009). Such events occur where abrupt bathymetric changes take place (e.g. escarpments, continental shelf breaks). They generally induce high organic productivity and phytoplankton growth, resulting in the deposition of organic-rich sediments in the neighbouring escarpment area. Low sedimentation rates and bacterial decay of organic matter liberate phosphates which can then induce the phosphogenesis of hardgrounds containing glauconite. High-energy conditions are required to allow these reactions and to erode the hemipelagic capping sediment, or to prevent its deposition, in order to expose the hardgrounds at the seafloor.

This scenario seems to apply to the Porcupine Bank Canyon escarpment region where the presence of strong currents was confirmed by sidescan sonar data and the medium- to coarse-grained sand with foraminifers and shell debris retrieved at sampling station AT496G (Mazzini et al. 2004). Moreover, the water depth in this area coincides with the limits of the previously described OMZ. The regional oceanography also provides complementary arguments to explain the presence of phosphate deposits. Glenn et al. (1994) describe phosphate precipitation as a widespread phenomenon in marine sediments, but emphasise that the accumulation of such deposits requires the maintenance of the phosphogenic system for extended periods of time coupled with prolonged concentration of phosphatic particles by current winnowing and transport. In the Porcupine Bank Canyon area, the water column stratification and the circulation pattern can provide the mechanism for efficient trapping and concentration of organic matter within narrow zones along the continental margin. The bottom nepheloid layers described by Dickson and McCave (1986) and Rice et al. (1991) can evidently supply large volumes of suspended organic matter, and the long-term persistence of these conditions along the Irish continental margin would thus promote the formation of phosphate crusts which, in turn, provide the hard substrata required for the establishment of deep-water coral bioherms. These conditions persist to the present day, as indicated by the thriving coral colonies and by the glauconite aggregates observed within the carbonate deposits of the escarpment (e.g. station AT500Gr). This implies constant slow accumulation rates, which is a precondition for the in situ formation of glauconite at the water–sediment interface.

The association of phosphatic hardgrounds with corals has also been widely documented in the literature. For example, Jarvis (2006) described deep-water colonial coral thickets (Diblasus arborescens) in conjunction with phosphate-rich hardgrounds in the Santonian–Campanian phosphatic chalks of England and France. Phosphate ‘nodules’ and intraclasts were drilled close to the base of the Magellan carbonate mound (Expedition-Scientists-ODP-307 2005). Furthermore, glauconite has been observed to occur in the upper bathyal coral zone of the Rockall Bank (Scoffin et al. 1980), in the hardgrounds near Connacht Mound (Noé et al. 2006), in bathyal gorgonian skeletons from the Bay of Biscay margin (Noé et al. 2007), on the Sula Reef of the mid-Norwegian shelf (Freiwald et al. 2002), and in Gonicorella thickets of the Chatham Rise, New Zealand (Dawson 1984; Kudrass and von Rad 1984).

The data available to this study are insufficient to determine to what extent the localised upwelling events known to occur in the area have contributed to the development and maintenance of the deep-water coral ecosystem. The distribution of the mounds along the escarpments along the canyon head suggests that the role of upwelling is substantial. However, the increased current-induced turbulence at the escarpment edges is probably also sufficient to create the enhanced nutrient fluxes needed for coral growth. A combination of both factors would, of course, also be plausible. Whichever the case, the formation of phosphate crusts greatly enhances the probability of widespread coral settlement. The high backscatter recorded in the escarpment area together with the phosphate deposits suggest that it would not be unrealistic to expect much more extensive deposits of this kind in the Porcupine Bank region.

Insights from isotopic signatures

The isotopic values show that seawater was the initial source of the carbon in the cements and bioclasts of the authigenic deposits in the study area. With δ18O = 0.6‰SMOW as the present-day value for the Porcupine region bottom waters (F. Mienis, personal communication), and applying the equations of, for example, Friedman and O’Neil (1977) and Grossman and Ku (1986), the oxygen isotopes of the analyzed samples indicate that these carbonates precipitated in equilibrium with the actual seafloor conditions (~7°C). Moreover, the δ13C values show no significant depletion which could be related to seepage of hydrocarbon-rich fluids. This is consistent with the findings of Noé et al. (2006) who obtained similar δ18O and δ13C values for carbonate hardgrounds from the Propeller Mound in the Porcupine Seabight. The association of coral reefs, carbonate deposits and methane seepages therefore remains a speculative issue at this stage.

Conclusions

The western margin of the Porcupine Bank is characterised by pronounced bathymetric variations and a high variety of depositional environments. In the course of this multidisciplinary study, a large province of carbonate coral mounds was examined together with the environmental controls and depositional processes involved in their formation. The mounds are located in a distinctive geomorphological and oceanographic setting when compared to other large mound provinces in Irish waters, and thus represent an important case study for research on mound development.

The data presented in this study show that particular oceanographic processes are the predominant factor controlling mound distribution and growth. Water column properties indicate mixing of water masses across the mound depth interval. In addition, a distinct salinity peak at about 1,000 m associated with the MOW coincides with the lower limit of the mounds. The available geophysical evidence indicates that the distribution of the mounds follows the pre-existing seabed topography, being located on the almost flat seabed along the edges of the canyon system framed by fault escarpments in the easternmost part. Furthermore, mound density increases towards the canyon head where the specific hydrodynamic regime effectively influences the shape and orientation of mound morphology.

Different types of authigenic deposits (carbonate and phosphatic) have been identified. On the top and along the flanks of the mounds, calcite and coral-rubble deposits are present. The model proposed in this study implies that strong currents facilitate the dissolution of coral debris and the precipitation of micritic calcite. Phosphatic and carbonate deposits have been sampled from the eastern escarpment. These phosphatic hardgrounds may extend well beyond the eastern part of the survey area. The bottom currents in this area and localised upwelling favour the formation of these phosphate-rich deposits, at the same time preventing their burial.

Sampling and video observations show that all the described authigenic deposits represent a hard substratum enhancing the growth of the mounds and promoting the establishment of new coral colonies along the escarpment. The available geochemical data do not support any relationship between methane seepage and coral colony emplacement.