Introduction

Cold-water coral (CWC) reefs are widespread along the continental margins of global ocean basins (Roberts et al. 2006). Since the discovery of spectacular CWC mounds in the Porcupine Seabight (Henriet et al. 1998), many more reef sites have been identified worldwide, predominantly in the North Atlantic. Indeed, the presence of Lophelia pertusa as a main frame-builder of CWC reefs is today known from, for example, the Norwegian margin (e.g. Freiwald 2002; Fosså et al. 2002; Hovland et al. 2012), the Porcupine, Rockall and Hatton banks (e.g. White et al. 2007; Sayago-Gil et al. 2010; Mazzini et al. 2012), the Celtic margin (e.g. Wheeler et al. 2007), the Armorican margin (e.g. Reveillaud et al. 2008), the Cantabrian Sea (e.g. Sánchez et al. 2014), the Iberian margin (e.g. Hernández-Molina et al. 2011; Somoza et al. 2014), the Mediterranean Sea (e.g. Freiwald et al. 2009; Palomino et al. 2011; Taviani et al. 2011; Fink et al. 2012; Gori et al. 2013; Savini et al. 2014), the Gulf of Cadiz (GoC; e.g. Van Rensbergen et al. 2005; Van Rooij et al. 2011; León et al. 2012), off northern Morocco (southern GoC; e.g. Foubert et al. 2008), off Mauritania (e.g. Eisele et al. 2011, 2014), the Angola margin (e.g. Le Guilloux et al. 2009), the Gulf of Mexico (e.g. Brooke and Schroeder 2007; Hübscher et al. 2010; Hebbeln et al. 2014), and around the Bahamas (e.g. Reed et al. 2006; Correa et al. 2012a, b). Many of these studies explored the impact of various hydrographic factors and nutrient availability (e.g. Hovland et al. 2012) on CWC reef growth. In addition, Martorelli et al. (2011) demonstrated the positive influence of topography-enhanced bottom current velocity on coral occurrences. Temperature and salinity are other key factors constraining the distribution of live corals (e.g. Davies et al. 2009; Davies and Guinotte 2011; Vierod et al. 2014), and indirectly ambient seawater density (e.g. Dullo et al. 2008; Flögel et al. 2014).

A new CWC province was discovered off western Morocco during research cruise 32 of RV Maria S. Merian in October 2013, which mainly targeted the morphology, sedimentary processes and geohazards of giant landslides of the Agadir Canyon system (Krastel et al. 2013). This study explores the recent oceanographic and bathymetric boundary conditions of these CWCs, based on a combination of seafloor bathymetric and acoustic data as well as conductivity-temperature-depth (CTD) data collected from the overlying water masses. One of the cruise aims was to test whether seawater density can be used as a predictive tool to detect living CWC occurrences. The new province is here named the Eugen Seibold coral mounds in honour of the pioneering marine geologist Eugen Seibold (1918–2013).

Physical setting

This study focuses on the Atlantic Moroccan continental slope off northwest Africa at 32°54′–31°08′N and 10°34′–10°49′W (Fig. 1). Three investigated sites exhibit the following water masses from the surface to the seafloor. The Central Water (CW) masses comprise Subtropical Underwater (SUW), South Atlantic Central Water (SACW) and North Atlantic Central Water (NACW; Mittelstaedt 1991; Van Camp et al. 1991; Morigi et al. 2001; Vandorpe et al. 2014). According to Mittelstaedt (1991) and Arístegui et al. (2009), SACW and predominantly NACW upwell seasonally. The presence of Mediterranean Outflow Water (MOW) below the NACW is associated with a pronounced increase in salinity favouring the formation of a pycnocline—for example, in the GoC (Fusco et al. 2008). Alves et al. (2011) described the MOW off Morocco at 34°N, while Pelegrí et al. (2005) reported this water mass extending as far south as 30.5°N. In the present study area, the deepest part of the water column is formed by the North Atlantic Deep Water (NADW).

Fig. 1
figure 1

Top Map showing the locations of the study area off western Morocco (black box; green dot mound field), and the CTD stations (green crosses) of earlier cruises off Mauritania (MSM16/3) and in the Gulf of Cadiz (MSM1/3; cf. Table 1), with depth contours and water masses. North Atlantic Central Water: solid blue arrows after Vandorpe et al. (2014); dotted blue arrow its continuation, present study. Mediterranean Outflow Water: red area and arrows after Hernández-Molina et al. (2011), including preferred meddy pathways; orange arrows its continuation, CTD data of present study. North Atlantic Deep Water: solid dark grey arrows after Hernández-Molina et al. (2011); dark grey dotted arrow its continuation, present study. Bottom Green Multibeam bathymetry coverage. Note the location of CTD station MSM32-1-1 (cf. Fig. 3, Table 1)

Full size image

Unpublished analogue TOPAS (TOpographic PArametric Sonar) sub-bottom profiles kindly supplied by G. Ercilla of the Institute of Marine Sciences (CSIC), Barcelona had revealed the presence of acoustically transparent mounds at more than 750 m water depth ~40 nautical miles north of the upper Agadir Canyon (Fig. 1), one of the largest submarine canyons worldwide (Wynn et al. 2002) and characterised by tectonic morphologies and giant landslides (Talling et al. 2007; Frenz et al. 2009). Examination of those preliminary data showed that the dimensions and acoustic character of the features were similar to those of known carbonate mounds in the northeast Atlantic (De Mol et al. 2002; Rüggeberg et al. 2011). Subsequent cruises gathered hydrographic data including seawater density, reported to be a predictive tool for coral occurrence in other regions (Foubert et al. 2008; Somoza et al. 2014). Prospective sites were mapped by swath bathymetry and side scan sonar in order to, amongst others, define targets for box coring. Indeed, these confirmed the suspected presence of mostly dead and some living corals in an extensive province of >400 km2 (Fig. 2).

Fig. 2
figure 2

Top 3D perspective view of multibeam bathymetry (lateral pixel resolution 30 m) across the Eugen Seibold coral mounds, and locations of the seven box corer stations (cf. Table 2) with live or dead corals, as well as three areas selected for estimation of coral mound density (cf. Fig. 4). With the exception of CTD station MSM32-1-1, situated further north beyond the currently known boundaries of the coral mound site (cf. Fig. 1), the locations of the other ten CTD stations were essentially identical or very close to those of the box corer stations (cf. Table 1). Bottom Bathymetric profile 1 crosses the site from its shallowest to its deepest part, profile 2 incorporates a coral mound with living CWCs, and profile 3 a coral mound with dead corals (box corer stations MSM32-49-2 and 46-1 respectively). V.E. Vertical exaggeration

Full size image

Materials and methods

Hydrography

Water masses above the coral mounds were investigated by performing 11 CTD casts (Table 1) using a Seabird ‘SBE 911 plus’ underwater unit additionally equipped with two oxygen sensors and a turbidity sensor (including SBE data processing). CTD data were visualized using Ocean Data View (ODV) software, version 4.5.3 (Schlitzer 2013). Temperatures and salinities are reported as conservative temperature (Θ, °C) and absolute salinity (SA, g/kg; TEOS-10, McDougall and Barker 2011). T–S plots of temperature vs. salinity are commonly used to delineate water masses and their geographical distribution as well as dynamics (Tomczak 1999).

Table 1 CTD stations off Morocco during cruise MSM32 (present study), and off Mauritania (cruise MSM16/3) and in the Gulf of Cadiz (cruise MSM1/3 (for data sources, see main text). T, S, density and oxygen values are given for the deepest sampling locations
Full size table

Hydroacoustic data

Multibeam bathymetric data were collected by means of the shipboard Kongsberg Simrad EM122 multibeam echo sound system (MBES) operating with a nominal frequency of 12 kHz. Data were processed by QPS Fledermaus and MBE-system software (Caress and Chayes 1995), including application of local sound velocity profiles, the creation of a CUBE surface (Calder and Mayer 2003) with a pixel resolution of 30 m, and manual removal of spikes. Coral mound heights and extensions were evaluated from these data for each of three sites within the mound field (Fig. 2). Mound heights were measured along N–S oriented depth profiles run parallel to the slope, in each case crossing the highest elevation down to the first break in slope. It should be noted that these values represent conservative approximations.

A TOBI (Towed Ocean Bottom Instrument) system, comprising a 30 kHz side scan sonar (Flewellen et al. 1993) and 7 kHz sub-bottom profiler, served to collect a ~13 km long side scan sonar line across the south-eastern sector of the coral province. Side scan sonar data were corrected for radiometric and geometric artefacts (Blondel 2009), and processed to a pixel size of ~3 m. The system was towed ~500 m above seafloor, producing a swath width of ~6 km (Murton et al. 1992).

Box cores

Based on hydroacoustic mapping, sites were selected for box coring (50×50 cm) to enable ground-truthing of seafloor sediments and potential CWC coverage. Seven box cores were taken (Table 2) distributed over the entire study area (Fig. 2).

Table 2 Box corer stations off Morocco during cruise MSM32
Full size table

Geochronology

Box core MSM32-3-1 was selected for a comparison of U-Th absolute ages of exposed surface branches of a living Lophelia reef structure with those of buried branch fragments from the underlying sediment. Measurements were conducted on a multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS, VG-Axiom), following Fietzke et al. (2005). The required mineralogical pre-investigation was based on X-ray diffractometry (XRD, Philips PW-1820). All analyses were performed using GEOMAR facilities.

Sample preparation, laboratory treatment and age determination followed Liebetrau et al. (2010). All uncertainties represent the 2 SE level, and the age data are rounded off to decades. Whole procedure blanks matched typical values of ca. 10 pg for U, 2 pg for 232Th and 0.1 fg for 230Th.

Results

Hydrography

In October 2013, assessments of regional oceanographic conditions characterising the deep-water MSM32-1-1 site to the north of the study area (4,224 m water depth; Figs. 1 and 3, Table 1) revealed an upper surface layer of ~30 m with a mean conservative temperature (Θ) of 23.06 °C, absolute salinity (SA) of 36.917 g/kg, and oxygen concentration of 4.53 ml/l. Below, the Central Water (CW) masses comprised Subtropical Under Water (SUW), South Atlantic Central Water (SACW) and North Atlantic Central Water (NACW). The SUW (31–64 m water depth) had Θ between 23.03 and 17.73 °C, SA of 36.914 to 36.526 g/kg, σΘ of 25.23 to 26.38 kg/m3, and oxygen values of 4.51 to 5.09 ml/l. The SACW (65–94 m) had Θ of 17.69 to 16.52 °C, SA of 36.535 to 36.483 g/kg, σΘ of 26.38 to 26.66 kg/m3, and oxygen values of 4.67 to 5.10 ml/l.

Fig. 3
figure 3

T–S plot (Θ cons. temperature, SA absolute salinity) for CTD stations MSM32-1-1 (see Fig. 1 for location) and MSM32-3-2 (location as for box corer station MSM32-3-1 in Fig. 2). Water masses comprise (from shallowest to deepest) an upper surface layer, Subtropical Under Water (SUW), South Atlantic Central Water (SACW), North Atlantic Central Water (NACW), Mediterranean Outflow Water (MOW), and North Atlantic Deep Water (NADW). Living CWCs, especially Lophelia pertusa, occur in the deeper part of the NACW roughly 150 m above the MOW characterised by still relatively elevated salinities and limited variation in temperature. Isopycnals are calculated with the reference pressure at 0 m (σΘ), i.e. sea level

Full size image

The top of the NACW (99–860 m) can be delimited by a strong decrease in Θ (from 16.71 to 10.75 °C) and in SA (from 36.512 to 36.072 g/kg). σΘ varies between 26.60 and 27.52 kg/m3 and oxygen values between 4.56 and 3.65 ml/l. The deeper part of the NACW can be constrained by a minimum SA range of 35.769 to 35.530 g/kg between 774 and 904 m water depth. Below the NACW, the Mediterranean Outflow Water (MOW, 863–1,252 m) can be identified in terms of small Θ variations in the range 10.77–9.71 °C, and elevated SA of 36.088–36.200 g/kg. σΘ ranges from 27.53 to 27.80 kg/m3, and oxygen values from 3.69 to 3.89 ml/l. Below the MOW at depths of 1,250–4,224 m, the North Atlantic Deep Water (NADW) had Θ=1.65 °C, SA=35.065 g/kg, σΘ=27.88 kg/m3 and [O2]=4.60 ml/l (Fig. 3).

In the vicinity of two living CWC sites discovered during the cruise—MSM32-3-2 and MSM32-51-2 at 704 and 678 m water depth respectively (see below)—local bottom water masses comprise the deeper part of the NACW. The deeper site at 704 m had Θ=9.78 °C, SA=35.632 g/kg, oxygen=3.14 ml/l, whereas the shallow site at 678 m had Θ=9.94 °C, SA=35.632 g/kg, oxygen=3.15 ml/l (Table 1). Conservative temperature and absolute salinity are lower than those of the deep MSM32-1-1 station in 700 m water depth, exhibiting Θ=10.79 °C, SA=35.807 g/kg (Fig. 3).

Hydroacoustic data

Seafloor bathymetric data (Fig. 2) reveal that the mapped mound field extends up to 40 km alongslope (north to south) and up to 13 km downslope (east to west), covering an area of ~410 km2. Cold-water coral mounds occur between 678 and 863 m water depth, immediately downslope of a relatively flat ‘plateau’ on the upper continental slope (Fig. 2). Individual mounds have circular to elliptical shapes in planform, and reach maximum heights of up to 12 m in the shallower part of the mound field. However, heights between 4 and 8 m are dominant (Fig. 2). Lengths of mound axes vary between 80 and 240 m, the longer axis being E–W aligned parallel to the direction of slope inclination.

Maximum slope angle (12°) and maximum height (12 m) of individual mounds tend to decrease with increasing water depth (Fig. 4), as does the density of mounds. Densities of mounds with slope inclinations larger than approx. 3° were assessed in three areas, each 4 km2 in size (Fig. 2). Mound density varied from 12 mounds/km2 (area 1) at ~700 m water depth, to 8 mounds/km2 (area 2) at ~790 m water depth, and 2 mounds/km2 (area 3) at ~880 m water depth (Figs. 4, 5).

Fig. 4
figure 4

Shaded relief (top row) and slope map (bottom row) across selected coral mound areas at three different water depths based on multibeam bathymetry data. Abundance, elevation and slope of coral mounds decrease as water depths increase from area 1 to area 3 (700, 780 and 890 m respectively). See Fig. 2 for locations of selected areas

Full size image
Fig. 5
figure 5

Number of mounds with respect to height for the three selected areas (each 4 km2) shown in Fig. 4. Mound height decreases from shallower depths of area 1, where living corals have been sampled, to deeper depths of area 3. Mean elevation and standard deviation for area 1 = 6.9±2.2 m, area 2 = 6.0±1.4 m, and area 3 = 3.6±1.0 m. Refer to Fig. 2 for area locations

Full size image

The individual mound-like structures appear as areas of higher backscatter (bright) in TOBI side scan sonar data compared to the lower backscatter of the surrounding seafloor (Fig. 6). Acoustic shadows (black) are commonly observed on the flanks of the steeper mounds. Shallower settings with more elevated mounds (~700 m water depth) display higher backscatter intensities compared to deeper settings (>750 m) with less elevated mounds (Fig. 6).

Fig. 6
figure 6

Left TOBI side scan sonar (30 kHz) imagery across the carbonate mound province (swath width 6 km). Individual coral mounds are recognized by high acoustic backscatter on flanks inclined towards the side scan sonar track, and low backscatter shadows on the lee sides. Right Higher-resolution imagery in the vicinity of two box corer stations (red stars): top right MSM32-49-2, with living corals; bottom right MSM32-50-2, with dead corals. For corresponding locations in study area, see Fig. 2

Full size image

Box cores

All seven box cores (see Fig. 2 and Table 2) contained fragments of dead CWCs. The three box cores from the deeper sites (MSM32-46-1, 801 m; MSM32-47-1, 835 m; MSM32-48-1, 863 m) recovered 18–40 cm of a sandy to silty, light and pale brown mud containing dead Lophelia pertusa bioclasts of up to 2 cm. Most of the Lophelia branches are fragmented down to the size of one polyp. Two shallower box cores (MSM32-45-1, 701 m; MSM32-50-2, 788 m) sampled abundant dead Lophelia thickets up to 8 cm long. These are commonly accompanied by bivalve shells embedded in sandy to silty brown mud, with a maximum recovery of 15 cm. The abundance of bioclasts was more than twice as high in box core MSM32-50-2 and well exposed at the surface, in contrast to all other cores with dead CWC material.

Box cores MSM32-49-2 (678 m) and MSM32-3-1 (719 m) contained large specimens of mainly dead but also a few living Lophelia. In box core MSM32-3-1, living branches of Lophelia exhibited up to 15 consecutive living polyps. Moreover, this box core had a single, live Madrepora oculata and Desmophyllum sp. specimen (Fig. 7). The coral framework stood well above the sediment between 10 and 20 cm. Further downcore, coral fragments and bioclasts associated with brown silty to sandy mud were recorded to at least 25 cm (maximum recovery). These shallower core sites are located within area 1 where highest elevations of coral mounds occur (see above). The box core sites deeper than 800 m are located within area 3 where the mounds are less elevated and less abundant (Table 2, Figs. 2, 4, 5). These cores recovered only dead CWCs, except for box core MSM32-50-2 (788 m; see above) which also contained a single living polyp of L. pertusa, unlike box core MSM32-3-1 (719 m) earmarked by more profuse live corals. In all box cores L. pertusa is by far the dominant CWC species.

Fig. 7
figure 7

Photograph showing the contents of the box core recovered from station MSM32-3-1. The large colony of Lophelia pertusa (1) exhibits up to 15 consecutive living polyps in one branch. Madrepora oculata (2) settled on dead branches of L. pertusa, which evidently collapsed prior to settlement as indicated by the different orientation of the polyps of both corals. This is the only living M. oculata recovered during the cruise. Desmophyllum sp. (3) settled also on dead L. pertusa branches. In this box core, the numerous dead corals partly buried in the muddy, bioclast-rich sediment comprise only L. pertusa. Among the coral debris, a few bivalves occur. See Fig. 2 for location of box core station

Full size image

Geochronology

U-Th isotope systematics (Table 3) were determined for the youngest parts of growing (live) Lophelia branches recovered above the sediment surface, and for macroscopically altered material buried at 25 cm sediment depth (max. penetration depth of box core MSM32-3-1, 719 m water depth). All analysed samples had aragonite contents exceeding 98%.

Table 3 U-Th isotope systematics and age determination of Lophelia pertusa from box core MSM32-3-1-1 (719 m water depth), cruise MSM32 off Morocco (uncertainties at 2 SE level)
Full size table

Based on Liebetrau et al. (2010), and in addition to the common detrital Th correction, the data were normalised in terms of a potentially site-specific initial 230Th/234U activity ratio of 0.0005±0.0002, here deduced from three live juvenile branch tops. These integrate only the youngest 2 to 4 mm of sub-recent skeletal growth and are therefore assumed to provide zero age reference values. Their δ234U signature of 149±2.5‰ is in good agreement with the modern seawater reference of 146.8±0.4‰ (2 SE) proposed by Andersen et al. (2010).

The growth systematics of this reef structure spanning approx. 40 cm (Fig. 7) enabled the identification of a root-like segment at its base with a maximum age of 60±20 years before present (BP) and a modern seawater matching initial δ234U value of 147.3±2.1‰. U-Th data on two branch fragments from 25 cm sediment depth revealed ages of 1,810±30 years BP (initial δ234U of 144.2±2.2‰) and 2,680±30 years BP (initial δ234U of 141.9±2.4‰). Inherently, the different degree of morphologically visible alteration implies the former as more reliable than the latter. This is supported by the fact that only the former fragment displays a within-error modern seawater matching initial δ234U value.

Discussion

Hydrographic measurements above the Eugen Seibold CWC mounds (between 65 and 863 m water depth) revealed that the oceanic water column is controlled mainly by the interaction of SACW and NACW with underlying MOW, which enters the Atlantic Ocean via the Strait of Gibraltar (Figs. 1, 3). The subsurface SACW flows northwards and typically displays lower salinity, oxygen depletion and nutrient enrichment (Arístegui et al. 2009) relative to the deeper southward-flowing NACW (Mittelstaedt 1991; Pastor et al. 2008; Vandorpe et al. 2014). The MOW is characterised by a very small temperature gradient, but shows a marked increase in salinity and density between 863 and 904 m water depth (Fig. 3). These water mass signatures are similar to those documented by Mittelstaedt (1991) and Pelegrí et al. (2005) off northwest Africa (25–35°N, and Cape Ghir at 30–36°N).

According to Dullo et al. (2008) and Flögel et al. (2014), most of the pristine and profuse live CWC coral reefs along the European continental margin occur at water mass densities of σΘ=27.35–27.65 kg/m3. Living CWCs of the Eugen Seibold province at water depths between 678 and 719 m with density values of about σΘ=27.33 kg/m3 correspond to the deeper part of the NACW. Their occurrence is therefore just above the defined density window of Dullo et al. (2008), and ~150 m above the upper limits of the MOW (Figs. 3, 8) identified at ~863 m water depth.

Fig. 8
figure 8

Detailed T–S plot (cf. Fig. 3) of water masses bathing living and dead CWCs of the Eugen Seibold coral mounds north of Agadir Canyon (cruise MSM32 of the present study, colour), compared with dead CWC sites in the Gulf of Cadiz (cruise MSM1/3, light grey) and living CWC sites off Mauritania (cruise MSM16/3, dark grey). Dashed ellipses Temperature–salinity occurrence range of dead CWCs; yellow, white circles living CWCs, all category III of Flögel et al. (2014). Isopycnals are calculated with the reference pressure at 0 m (σΘ), i.e. sea level

Full size image

Flögel et al. (2014) differentiated between three types of living CWC assemblages. Category I is characterised by dense horizontal and, most importantly, vertical reef growth, category II by patchy growth with colonies of some square meters, and category III by smaller and commonly more widely scattered isolated colonies. Living CWCs in the present study area are attributed to category III. This category is further supported by lower oxygen values of 3.14–3.15 ml/l at the study sites, contrasting with a global mean of ~6 ml/l for ambient water masses around flourishing CWC reefs (Davies et al. 2008). The classification of three categories in combination with seawater density works well in the Atlantic, but can not be applied to the Mediterranean due to its higher density and higher alkalinity (cf. Flögel et al. 2014).

Although the information available to date from box cores represents only snapshots of the coral mounds, the general distinction between living mounds clustering around 700 m water depth and dead mounds in deeper waters (Figs. 2, 3, 8) is supported by the acoustic data (Fig. 4). Contrasting with presumably dead mounds, areas inhabited by living CWCs generally have more elevated (up to 12 m) and more steeply (up to 12°) inclined mounds (Figs. 4, 5) associated with higher backscatter intensities (Fig. 6). One might argue that high mounds may result from the higher baffling capacity of CWC fabrics. Less elevated mounds with lower slope inclinations may be a sign of partial burial of dead CWCs, which may explain the reduced occurrence of mounds in such areas. Another mechanism could be that the bathymetrically deeper mounds are already affected by erosion, but this awaits confirmation by additional optical ground-truthing and dating.

A similar situation has been described for CWC mounds off Mauritania—e.g. the Banda Mounds at 17–18°N, which were actively growing from 65 to 57 ka, 45 to 32 ka, and at about 14 ka (Eisele et al. 2011). The Banda Mounds are bathymetrically above the density envelope of Dullo et al. (2008; σΘ=27.5±0.15 kg/m3), and are also situated within the lower part of the NACW (RV Maria S. Merian cruise MSM16/3 in October 2010, Westphal et al. 2012; see Fig. 8). Today, these mounds show only a very thin cover of living Lophelia between 415 and 570 m water depth, corresponding to category III of Flögel et al. (2014).

The Renard Ridge (35°N) in the Gulf of Cadiz (GoC) is another prominent region of CWC mound occurrence; however, there are no living CWC colonies reported so far (e.g. Foubert et al. 2008; Wienberg et al. 2009, 2010). These dead CWCs are today found at relatively shallow water depths typified by densities not exceeding σΘ=27.5±0.15 kg/m3 (RV Maria S. Merian cruise MSM1/3 of January 2006, Pfannkuche and cruise participants 2006; see Fig. 8). There is evidence that, over the last 400 ka, these reefs also flourished during glacial times and up to the very early Holocene (Wienberg et al. 2010).

Although Davies and Guinotte (2011) predict a high suitability for CWC growth, especially of Lophelia pertusa, between the GoC and Mauritania, to date living CWCs have been found only off Mauritania (17–18°N) and now off western Morocco (31°N) as small isolated colonies or forming a thin and patchy cover. Vigorous and extensive CWC reef growth apparently occurred only during glacial times in this region, which would explain why the few modern living colonies reported to date do not match bathymetrically with the water mass density values known for CWC reefs belonging to category I of Flögel et al. (2014). The Eugen Seibold coral mounds exhibit a few living colonies (albeit based on limited sampling), and most sampled mound structures consist of a dead fabric.

Dating of a large living coral branch indicated a time span of 60±20 years BP from the base to its living topmost part. The large uncertainty of ±20 years mainly results from the correction for the initial 230Th/234U activity ratio. This implies a mean growth rate of 0.5 to 1 cm/year. An alternative age estimate is provided by counting the number of polyps in straight succession from the root segment to the juvenile top, which on average could correspond to the number of years according to Freiwald (2002). Resulting in 30±5 polyps (varying with the sub-branch to follow), this approach suggests a slightly younger age. Based on CaCO3 accumulation rates of 8.70±3.18×10–3% per day (Form and Riebesell 2012), the maximum mass of 465 g of the dated branch implies a similar value of 9±2×10–3% per day in terms of polyp counting age, and 5±2×10–3% per day in terms of U-Th age. Deducing in first approximation a local sediment accumulation rate from the most reliable age of the dead fragments (1,810±30 years BP) and the observed burial depth results in 14 cm/1,000 years. This is consistent with the background flux of 12.15±0.6 cm/1,000 years for the nearby Canary Basin reported by Hendericks et al. (2002).

In their overview of north-eastern Atlantic CWC reefs and climate change, Frank et al. (2011) argue that CWC reefs were growing south of 50°N predominantly during the last glacial periods and the deglacial on the Armorican margin (Reveillaud et al. 2008), the Iberian margin (Hernández-Molina et al. 2011), in the GoC (Wienberg et al. 2009, 2010) and the Banda Mounds (Eisele et al. 2011, 2014), including the Mediterranean (Savini et al. 2014), with the exception of sites in the Alboran Sea still growing until 5 ka (Fink et al. 2012). This would be in contrast with the CWC reef formation north of 50° being constrained to interglacials, including the late Holocene (e.g. Rüggeberg et al. 2007; Frank et al. 2009; Raddatz et al. 2013). The predominantly dead CWC assemblages of the newly discovered Eugen Seibold coral mounds off western Morocco may support the hypothesis of Frank et al. (2011). However, the initial dataset displays CWC growth since about 2 ka at least. Therefore, interglacial growth periods cannot be excluded in general south of 50°N, as demonstrated also by the discovery of live juvenile corals in the present case. Providing a more comprehensive record of emplacement time intervals for the Eugen Seibold coral mounds requires more detailed archives, notably a drill core through at least one of the major mound structures well identified in this study.

Conclusions

The newly discovered Eugen Seibold coral mounds off western Morocco extend over more than 410 km2, associated with known oceanographic settings of CWC occurrences along the north-eastern Atlantic margin. Mounds with living CWCs occur in a bathymetric range between 678 and 719 m characterised by Lophelia pertusa, at one site accompanied by a single Madrepora oculata excrescent and a single Desmophyllum sp. specimen. Most CWCs are represented by dead thickets which are more fragmented in the mounds at deeper water depths (>710 to 860 m). Shallower mounds are more elevated (12 m) and exhibit higher backscatter in contrast to mounds at deeper sites. Absolute ages of dead coral fragments indicate growth since the late Holocene at least, and imply a sediment accumulation rate of about 14 cm/1,000 years. These initial geochronological results and the patchy distribution of the cold-water corals suggest re-colonization of this area by Lophelia, rather than the continuous evolution of large flourishing reef build-ups. This interpretation for the Eugen Seibold coral mounds awaits confirmation based on additional evidence from optical ground-truthing and dating of longer downcore records.