Quantifying relationships between abundances of cold-water coral Lophelia pertusa and terrain features: A case study on the Norwegian margin

Continental Shelf Research 116 (2016) 13–26

Contents lists available at ScienceDirect

Continental Shelf Research journal homepage: www.elsevier.com/locate/csr

Research papers

Quantifying relationships between abundances of cold-water coral Lophelia pertusa and terrain features: A case study on the Norwegian margin Ruiju Tong a,b,n, Autun Purser c,d, Janine Guinan e, Vikram Unnithan c, Jinsongdi Yu b, Chengcheng Zhang f a

Department of Transportation, Fujian University of Technology, 350108 Fuzhou, China Spatial Information Research Center of Fujian Province, Fuzhou University, 350003 Fuzhou, China Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany d Alfred-Wegener Institut Helmholz-zentrum Für Polar-Und Meereschung, Deep Sea Ecology and Technology, Am Handelshafen 12, D-27570 Bremerhaven, Germany e INFOMAR, Marine and Geophysics Programme, Geological Survey of Ireland, Beggars Bush, Haddington Road, Dublin 4, Ireland f China Academy of Surveying and Mapping, Beijing 100830, China b c

art ic l e i nf o

a b s t r a c t

Article history: Received 1 April 2015 Received in revised form 5 December 2015 Accepted 19 January 2016 Available online 20 January 2016

An understanding of how terrain features influence abundance of a particular species greatly aids in the development of accurate predictive habitat suitability models. In this study, we investigated the observed seafloor coverage of cold-water coral Lophelia pertusa in relation to seabed topography at the Sotbakken and Røst Reefs on the Norwegian margin. The primary terrain features at the study sites are a SW-NE stretching mound at Sotbakken Reef and SW-NE running ridges at Røst Reef, located at depths of
300– 400 m and
250–320 m respectively. Ship-borne multibeam bathymetry data, JAGO dive video data and JAGO positioning data were used in this study. Terrain variables were calculated at scales of 30 m, 90 m and 170 m based on the bathymetry data. Additionally, we investigated the relationships between the terrain variables at multiple scales using the Unweighted Pair Group Method. The observed L. pertusa coverage at both reefs was found to be significantly correlated with most investigated terrain variables, with correlations increasing in strength with increase in analysis scale, suggesting that large scale terrain features likely play an important role in influencing L. pertusa distribution. Small scale terrain variations appear less important in determining the suitability of a region of seafloor for L. pertusa colonization. We conclude that bathymetric position index and curvature, as well as seabed aspect, most strongly correlate with coral coverage, indicating that local topographic highs, with an orientation into inflowing bottom currents, are most suitable for L. pertusa habitation. These results indicate that developing habitat suitability models for L. pertusa will benefit from inclusion of particular key terrain variables (e.g. aspect, plan curvature, mean curvature and slope) and that these should ideally be computed at multiple spatial scales with a greater gap in scales than we used in this study, to maximize the inclusion of the key variables in the model whilst minimizing redundancy. & 2016 Elsevier Ltd. All rights reserved.

Keywords: Cold-water coral Lophelia pertusa Terrain features Linear regression

1. Introduction The cold-water coral (CWC) Lophelia pertusa (azooxanthellate scleractinian) has a cosmopolitan distribution at depths
39– 3380 m, and is particularly abundant in North Atlantic waters (Davies and Guinotte, 2011; Fosså et al., 2002; Freiwald et al., 2004; Roberts et al., 2009a). L. pertusa is the primary reef

n Corresponding author at: Department of Transportation, Fujian University of Technology, 350108 Fuzhou, China. E-mail address: [email protected] (R. Tong).

http://dx.doi.org/10.1016/j.csr.2016.01.012 0278-4343/& 2016 Elsevier Ltd. All rights reserved.

framework builder in the North Atlantic (Roberts et al., 2009a, 2009b), forming large carbonate mounds, e.g. those on the Irish margin (Dorschel et al., 2007; Mienis et al., 2009; Wheeler et al., 2007), or extended reef complexes, at depths often associated with the continental shelf edge, particularly on the Norwegian margin (Freiwald et al., 2004; Mortensen, 2000; Purser et al., 2013). L. pertusa reef habitats support a high diversity of benthic species, with more than 1300 species identified at such reefs in the NE Atlantic (Buhl-Mortensen et al., 2010; Henry and Roberts, 2007; Kutti et al., 2014; Roberts et al., 2006). However, 30–50% of L. pertusa reef areas on the Norwegian margin are either damaged

14

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

or impacted by bottom trawling reported by Fosså et al. (2002). Lophelia pertusa distribution was previously thought to be dependent on local hydrocarbon seepage and chemoautotrophic production (hydraulic theory) (Hovland and Risk, 2003). Recent studies indicate that CWC ecosystems are often largely sustained by the passive settling delivery of phytodetritus, zooplankton, or particulate organic matter derived from near-surface primary productivity (Carlier et al., 2009; Davies et al., 2009; Dodds et al., 2009; Duineveld et al., 2007; Kiriakoulakis et al., 2007; van Oevelen et al., 2009; Wagner et al., 2011). Lateral advection of food particles may also play an important role in maintaining CWC communities (Thiem et al., 2006; Thomsen, 2002). Periodic downwelling of fresh, labile material from near-surface waters is an alternative primary food supply mechanism at some reefs (Davies et al., 2009; Duineveld et al., 2012; Wagner et al., 2011). In addition to the availability of a suitable food source, near-seabed water chemistry and hydrodynamics, seabed substrate and seabed topography are also controlling factors in determining CWC habitat suitability (Duineveld et al., 2012; Mienis et al., 2012; Purser, 2010; Purser et al., 2013; Rüggeberg et al., 2011; Somoza et al., 2014; Tittensor et al., 2009). Seabed topography influences CWC distribution by governing current regimes, therefore influencing the delivery of food particles (Mienis et al., 2007; Mortensen and Buhl-Mortensen, 2004; Thiem et al., 2006; Wagner et al., 2011), and also sediment distribution, which is important for initial coral settlement (Bryan and Metaxas, 2006). CWCs are often observed on topographic highs with accelerated currents, such as sills (Lavaleye et al., 2009; Wagner et al., 2011), ridges (Freiwald et al., 2004; Freiwald et al., 2002; Purser et al., 2013; Tong et al., 2013b), mound structures or seamounts (Lo Iacono et al., 2014; Mohn et al., 2014; Rowden et al., 2010; Vertino et al., 2010; White, 2007). The localized increased current velocity passing such topographic features promotes advection of food particles within the benthic boundary layer for utilization by CWCs (Kiriakoulakis et al., 2007; Thiem et al., 2006). High bottom current velocity also clears coral surfaces of deposited material, preventing living coral colonies from being buried by sediment (Mienis et al., 2007; White et al., 2005). On some areas of sloping topography, breaking internal waves resuspend organic matter or promote mixing to depth of surface waters, thereby further increasing food supply (Frederiksen et al., 1992). Hard substrate is commonly associated with steeper slopes or topographic highs. Therefore, the terrain variables slope, curvature (plan curvature, profile curvature and mean curvature) and bathymetric positioning index (BPI), which effectively capture the variation of sloping topography and topographic highs, may act as proxies for bottom current velocity and seabed substrate hardness, and may therefore be linked with the distribution of CWCs. Aspect provides information on the orientation of seabed terrain, and in regions with prevalent current direction conditions this parameter may be of particular relevance for suspension-feeding fauna (Guinan et al., 2009b; Wilson et al., 2007). Cold-water corals are often observed on areas with strong structural components, e.g. rocky outcrops, rather than flat areas (Fabri et al., 2014; Purser et al., 2013; Qurban et al., 2014; Roberts et al., 2005), which may be closely linked to the seabed variation captured by the terrain variability or complexity indices, rugosity and terrain ruggedness index (TRI) (Jenness, 2006; Riley et al., 1999). The terrain variability indices at a local scale distinguish complex habitats with strong structural features from flat terrain, whilst at broader scales capture variations related to seabed morphology characteristics (Wilson et al., 2007). In recent years, bathymetry-derived terrain variables such as slope, aspect, curvature (plan curvature, profile curvature, mean curvature), BPI, rugosity and TRI, are increasingly applied in studies of habitat classification and habitat suitability modeling

(Dolan et al., 2008; Giusti et al., 2014; Guinan et al., 2009a; Guinan et al., 2009b; Howell et al., 2011; Tong et al., 2013a, 2013b, 2012; Wilson et al., 2007). Terrain variables at multiple scales have been shown to have an ecological relevance in determining distribution of benthic fauna, with the terrain parameters often acting as proxies for bottom current velocity regimes (Guinan et al., 2009b; Rengstorf et al., 2012, 2013; Savini et al., 2014; Tong et al., 2013a, 2013b; Wilson et al., 2007). Wilson et al. (2007) comprehensively summarizes these variables to each belonging to one of four groups- slope, orientation (aspect), curvature and relative position (plan curvature, profile curvature, mean curvature and BPI), and terrain variability (rugosity and TRI). Investigating CWC distribution in relation to seabed topography is important for understanding the terrain habitat selection of these species, and for the development of predictive habitat models. Guinan et al. (2009b) report a strong correlation between L. pertusa cover and certain terrain variables (slope, aspect, BPI and rugosity) at scales of 90 m and 270 m within carbonate mound provinces on the Irish margin. Particularly strong correlations were observed between the L. pertusa cover and BPI, rugosity and slope at an analysis scale of 270 m (Guinan et al., 2009b). Accordingly, Guinan et al. (2009b) suggested that the relationship between the coral abundance and terrain variable is scale dependent. However, the degree to which this relationship applies to other CWC habitats, such as the Sotbakken Reef and Røst Reef provinces on the Norwegian margin, where reefs cover a more substantial area of seafloor with considerably more robust and vertically extensive structures, is not known. Additionally, the strength of the relationship between L. pertusa cover and other terrain variables often used in habitat modeling is also unclear. Similarly how closely terrain variables (extracted from multibeam bathymetry data) correlate with each other in these regions is also unknown. In this study, we investigate the video observations of L. pertusa cover in relation to multiscale terrain variables introduced above at the Røst and Sotbakken Reefs on the Norwegian margin. The aim of the investigation was to determine if any particular terrain variables (and at which analysis scales) significantly correlate with L. pertusa cover. Further, we investigate the interrelationship of these terrain variables and determine which should be adopted for use in developing habitat suitability models, to both include the key terrain information relating to coral distribution whilst simultaneously minimizing redundancy.

2. Methods 2.1. Study area Sotbakken Reef and Røst Reef are two CWC reef complexes located on the Norwegian margin (Purser, 2010). Both were investigated in 2007 during the International Polar Year (IPY) Polarstern ARK XXII/1a expedition (Klages and Thiede, 2011), with the acquired multibeam bathymetry data, video data, and submarine positioning data collected by the expedition used in this study. The Sotbakken Reef is located on a large mound at the NW extremity of a plateau in the northernmost cross-shelf trough Håkjerringdjupet (off Tromsø) on the Norwegian continental shelf (Ottesen et al., 2008) (Fig. 1). This mound is of
20 m height above the surrounding plateau, and stretches for 4 1000 m distance in a SW-NE direction at depths
250–320 m (Fig. 2). The Røst Reef complex was discovered in May 2002, and protected from bottom trawling from 2003 (Fosså et al., 2004). The reef complex is situated on the Norwegian shelf break, and composed of many large L. pertusa reefs at depths
300–400 m (Fosså et al., 2005) (Fig. 1). Numerous dissected ridges (tens of meters

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

15

Fig. 1. Overview of the study site – (A) Sotbakken Reef and (B) Røst Reef on the Norwegian margin. The map is projected in WGS 84 UTM Zone 34N.

height) run parallel to the continental shelf break with the L. pertusa reefs crowning these (Fosså et al., 2004; Purser et al., 2013). Both reefs are exposed to the relatively warm (temperature 45 °C) and salty (salinity 4 35%) Atlantic water within the density envelope of sigma-theta 27.35–27.65 kg m  3 (Dullo et al., 2008) and extensive L. pertusa growth has been reported at each (Klages and Thiede, 2011). The bottom current at Sotbakken Reef is likely to flow in a prevalently westerly direction, indicated by the facing direction of the concave surface of large Paragorgia arborea colonies (Tong et al., 2012). At the Røst Reef, food particles are cross-shelf transported from the productive shelf areas in a NW direction almost perpendicular to the reef community (Moje et al., 2011; Thiem et al., 2006; Tong et al., 2012). 2.2. Data acquisition and analysis 2.2.1. Multibeam bathymetry data and terrain variables at multiple scales Ship-borne multibeam bathymetric data were collected using the Hydrosweep DS-3 during the Polarstern ARK XXII/1a (2007) expedition (Klages and Thiede, 2011), and were further processed using the open source software MBSystem. The processed bathymetric data within the WGS 84 geographic coordinate system was projected into UTM Zone 34 N at Sotbakken Reef and UTM Zone 32 N at Røst Reef, then further gridded into raster data Digital Elevation Model (DEM) with 10 m cell size. In this study, the variables of slope, aspect, plan curvature, profile curvature, and mean curvature at the analysis scales of 30 m, 90 m and 170 m were calculated using Landserf v2.3 (Wood,

1996) based on the DEM provided by multibeam bathymetry data (10 m cell size), corresponding to moving window sizes of 3  3, 9  9 and 17  17 cells (Table 1). Similarly, the rectangular moving window was used to calculate the BPI at the three analysis scales using ArcMap 10 Raster Calculator. Additionally, the variables of TRI and rugosity at the analysis scale of 30 m were calculated using the Macro of ArcGIS 9.2 (Riley et al., 1999) and the ArcView 3.x. extension (Jenness, 2006), respectively. The terrain attribute of the cells through which JAGO dive transects passed (see Section 2.2.2) was extracted as the attribute value of the center of cells from the terrain variables. 2.2.2. JAGO submersible video data and positioning data Five survey dives were conducted at the reefs using the manned submersible JAGO (IFM-GEOMAR). Three surveys investigated the crest and southern side of the mound at Sotbakken Reef, with two passing over a number of ridges at the Røst Reef (Figs. 2a and 3a). The video data was recorded at high definition (HD) quality using the camera
1 m above the seafloor, with a vehicle speed
0.5 m s  1. A pair of parallel laser pointers (positioned 50 cm apart) were used to provide image scale for the majority of each dive. The JAGO dive positioning data was collected by a LinkQuest TrackLink 1500 HA USBL positioning system. The gross error points of the recorded dive tracks were identified as those too remotely distant from the previous positioning point to be accounted for by the maximum JAGO survey speed of 0.8 m s  1 and were therefore removed. The remaining positioning data was smoothed using Gaussian smoothing to decrease the random error in the data using Adelie GIS (IFREMER) software for underwater vehicle data

16

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

Fig. 2. Terrain variables at Sotbakken Reef. (a) Bathymetry data of the surveyed area, with the lines in different colors representing the three JAGO dive transects, (b) PlanCur30, (c) PlanCur90, (d) PlanCur170 with the observed % L. pertusa cover. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

post-processing. 2.2.3. Quantification of coral abundance Cover of L. pertusa was logged throughout each JAGO dive transect. From the collected video, L. pertusa passing through a1.5 m swathe in the center of the recorded video frames was

logged. The 1.5 m swathe was determined by the pair of lasers (50 cm spacing) recorded by the video, and equal length extensions to the left and right of these laser points. The L. pertusa cover was classified across five levels – 0, 1–25%, 25–50%, 50–75% and 75–100% live cover in 1 m distance intervals over the seabed, and then recorded as the mean values of 0, 12%, 37%, 62% and 87%

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

2.3. Statistics methods

Table 1 Terrain variables at multiple scales used in this study. Moving window size (m)

Slope Aspect Bathymetric position index Mean curvature Plan curvature Profile curvature Rugosity Terrain ruggedness index

17

Terrain variables at multiple analysis scales 30  30

90  90

170  170

Slope30 Aspect30 BPI30 MeanCur30 PlanCur30 ProfileCur30 Rugosity30 TRI30

Slope90 Aspect90 BPI90 MeanCur90 PlanCur90 ProfileCur90

Slope170 Aspect170 BPI170 MeanCur170 PlanCur170 ProfileCur170

cover (Fig. 4). These 1 m distance intervals were calculated from the positioning data. The % cover of L. pertusa in one grid cell was calculated as the mean of logged coral cover in this cell, and further converted to a value between 0 and 1 using the transformation (¼ 0.5log10(1 þ%cover)) to decrease the relative importance of high percentage cover (Aitchison, 1986; Guinan et al., 2009b).

2.3.1. Linear regression analysis The linear regression statistical method has been successfully applied to investigate the relationships between video observations of L. pertusa cover and seafloor features in the Rockall Trough, west of Ireland (Guinan et al., 2009b). Here we adopt a similar linear regression approach to quantify the relationships between the observed L. pertusa cover and terrain variables. Each correlation investigated was measured by the R-square (coefficient of determination) value, representing the variation in L. pertusa cover explained by the corresponding terrain variable as between 0 and 1. 2.3.2. Unweighted pair group method with arithmetic mean Unweighted Pair Group Method with Arithmetic Mean (UPGMA) is a commonly used cluster analysis method based on a similarity matrix (Sokal and Michener, 1958). In this study, the UPGMA algorithm was applied to hierarchically cluster the terrain variables at multiple scales within each reef area, using a Pearson Correlation Coefficient distance metric. At each step, UPGMA clustered the nearest two clusters of the terrain variables into a

Fig. 3. Terrain variables at Røst Reef. (a) Bathymetry data of the surveyed area, with the lines in different colors representing the two JAGO dive transects, (b) PlanCur30, (c) PlanCur90, (d) PlanCur170 with the observed % L. pertusa cover. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

18

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

Fig. 4. Sample of observed % L. pertusa cover. (a) 1–25%, (b) 25–50%, (c) 50–75%, (d) 75–100%.

higher-level cluster. The distance between the nearest two clusters was calculated as the average distance of all distances (Pearson

Correlation Coefficients) between pairs of objects “x” in A and “y” in B (Sokal and Michener, 1958).

Fig. 5. Correlations between the observed % L. pertusa cover and terrain variables at multiple scales. R2 value is the coefficient of determination. These terrain variables significantly correlated with the observed L. pertusa cover were highlighted in dark gray.

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

3. Results 3.1. Correlation between L. pertusa cover and terrain variables The linear regression analysis to quantify the relationship between the observed cover of L. pertusa and terrain variables measured by the R-square (coefficient of determination) yielded the results presented in Fig. 5. The observed L. pertusa cover was significantly correlated with terrain variables (shown in black in Fig. 5) with R2 40.06 at Røst Reef and R2 40.063 at Sotbakken Reef (P o0.05). Fig. 5 indicates that at both study sites L. pertusa cover was significantly correlated with seabed slope, aspect, BPI and curvature at certain analysis scales (Slope90, Slope170, Aspect170, BPI30, BPI90, BPI170, MeanCur30, MeanCur90, MeanCur170, PlanCur30, PlanCur90, PlanCur170, ProfileCur90 and ProfileCur170 at Sotbakken Reef, Slope170, Aspect90, BPI90, BPI170, MeanCur90, MeanCur170, PlanCur90, PlanCur170 and ProfileCur170 at Røst Reef). Additionally, L. pertusa cover was more significantly correlated with these variables at the largest tested analysis scale (170 m) at both study sites, with the exceptions of profile curvature at Sotbakken Reef and aspect at Røst Reef. 3.2. Terrain analysis and L. pertusa cover These multiple scale terrain variables captured the topographic properties of the study area at different spatial scales (e.g. the plan curvature at the three analysis scales shown in Figs. 2 and 3). PlanCur30 at Sotbakken Reef captured the local scale seabed topographic variations (Fig. 2b). PlanCur90 captured the crest of the mound feature, with local topographic highs shown in red and depressions in green (Fig. 2c). PlanCur170 was smoother across the region, highlighting the crest of the mound in red (Fig. 2d). The

19

highest Slope170 values (purple) (Fig. 6a) were recorded from the NW facing large slope, and the NW and SE facing flanks of the large mound, whilst lowest slope values (light blue) were characteristic of the crest of the mound. Aspect170 captured the orientation of the various mound sections – NW and SE facing flanks, and the crest with a distinct variable seabed orientation (Fig. 6b). Lophelia pertusa was most abundant on the crest of the mound feature, mainly associated with middle to low values of PlanCur170, PlanCur90 and low values of Slope170, with seabed orientations of Aspect170 [0, 150°) (Fig. 7). PlanCur30 at Røst Reef illustrated the specific local topographic variation, e.g. the uppermost crests of the ridges as topographic divergences (red) and the seabed depression areas between the ridges as local topographic convergences (green) (Fig. 3b). PlanCur90 smoothly captured the crest of the ridges in red and the depression areas between the ridges in green (Fig. 3c). PlanCur170 highlighted the crests of the large ridges. Where larger and smaller ridges were located in close proximity, PlanCur170 commonly smoothed these into one large ridge-like feature, highlighting the crest of the combined ‘ridge’ in red (Fig. 3d). MeanCur170 captured the crest and upper slope of the NW-facing flank of the large ridges and large ridge-like features in red, and the depression areas between these in blue (Fig. 8a). ProfileCur170 highlighted the upper slope of the NW-facing flank of the large ridges and the large ridge-like features in red (Fig. 8b). BPI170 highlighted the crest and the upper slope of the NW-facing flank of the ridges in red (Fig. 8c). Aspect90 captured the ridges running parallel to the edge of the continental margin as a spatially scalariform array orientated toward the NW in the central and southern area of the study site, as well as identifying a west facing large slide in the northerly surveyed area (Fig. 8d). The most abundant areas of L. pertusa at Røst Reef were predominantly observed on the crest of the large ridge-like feature

Fig. 6. The observed % L. pertusa cover with terrain variables (a) Slope170, (b) Aspect170 at Sotbakken Reef. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

20

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

Fig. 7. Relationships between the observed % cover of L. pertusa and (a) Slope170, (b) Aspect170, (c) PlanCur90, (d) PlanCur170 at Sotbakken Reef.

and the upper sections of the large slide in the north of the surveyed area, often associated with middle to low values of PlanCur170, middle to high values of MeanCur170, ProfileCur170 and BPI170 with seabed orientations of Aspect90 [0, 90°) and [270, 360°) (Fig. 9). 3.3. Relationships between terrain variables Figs. 10 and 11 show the clusters of terrain variables produced using UPGMA of Pearson's correlation coefficients for Sotbakken Reef and Røst Reef respectively. The strongest correlation between terrain variables (Pearson's correlation coefficient 40.5), was observed in two categories: (1) between terrain variables of different types at similar analysis scales, including between mean curvature, profile curvature and BPI, and between slope, TRI and rugosity and (2) between one terrain variable at multiple scales, particularly at analysis scales of 90 m and 170 m, e.g. between Slope30, Slope90 and Slope170 (Figs. 10 and 11). The terrain variables of BPI90/170, MeanCur90 and ProfileCur90 were strongly correlated with each other (illustrated in Fig. 12 at Sotbakken Reef and Fig. 13 at Røst Reef). The variables shown in Fig. 12 highlighted the local topographic highs on the crest of the mound and the NW-facing large slope at Sotbakken Reef. The terrain variables shown in Fig. 13 captured the topographic variations associated with the NE-SW running ridges, as well as the local topographic highs of the northerly slide feature.

4. Discussion 4.1. L. pertusa cover in relation to terrain variables In this study, the observed L. pertusa cover at both study areas was found to be significantly correlated with most of the terrain variables investigated, and particularly closely correlated with the variables at broader analysis scales, e.g. the Slope170 and PlanCur170, rather than the terrain variables at local scale (30 m) (Fig. 5). Bathymetric noise may account for a greater percentage of the observed variation of terrain variables at local scales compared

to those at larger scales (Albani et al., 2004). This result indicates that the L. pertusa cover is likely influenced by large scale terrain features, such as the mound feature at Sotbakken Reef or the large ridge structures at Røst Reef, with the progressively closer correlation with increase in analysis scale indicating broad area seafloor variability potentially playing a very significant role in determining L. pertusa distribution. Possibly, large features not captured within the scales investigated here, such as the shelf break at Røst Reef, may be of even greater relevance for L. pertusa distribution. This observation is supported by Mortensen and Buhl-Mortensen (2004), who suggested that such larger scale terrain features, (such as the shelf break and long running ridges of the Norwegian margin) play important role in controlling coral abundance by influencing current regimes. Further, Guinan et al. (2009b) reported that at a slightly coarser but comparable resolution (30 m grid) L. pertusa cover is more closely correlated with slope, aspect, BPI and rugosity at 270 m scale than at a scale of 90 m (Guinan et al., 2009b). Rengstorf et al. (2012) emphasize the need for high resolution multibeam bathymetry data (o 250 m grid) in habitat suitability modeling in order to identify the relatively small carbonate mounds that support a significant proportion of living L. pertusa on the Irish margin (Rengstorf et al., 2012). Previous investigation has shown live L. pertusa frameworks occupy a particular section of the dominant topographic features (the reefs) with a living coral ‘head’, such as at the Traena Reef on Norwegian margin (Buhl-Mortensen et al., 2010; Fosså et al., 2005), with the reefs of smaller size than at Røst Reef and Sotbakken Reef. Therefore, in habitat suitability modeling, the analysis scale of terrain variables should be chosen according to the scales of dominant topographic features in the particular study area. Tong et al. (2012) confirmed that the terrain variables BPI and curvature were most relevant to the observed distribution of two gorgonian coral species (Paragorgia arborea and Primnoa resedaeformis) at Sotbakken Reef, Røst Reef and Traena Reef. Similarly, in this study, L. pertusa cover was found to be most strongly correlated with the variables of curvature and BPI at the 170 m analysis scale (Fig. 5). This indicates a possible pattern in the importance of the relative position of focal localities to the surrounding terrain in influencing distribution of CWC species. The

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

21

Fig. 8. The observed % L. pertusa cover with terrain variables (a) MeanCur170, (b) ProfileCur170, (c) BPI170, (d) Aspect90 at Røst Reef. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article).

result of the linear regression shows that the high abundance of L. pertusa at Sotbakken Reef is mainly associated with middle to low values of PlanCur170 and PlanCur90, indicating a tendency for reef development in areas of topographic divergence (Wilson and Gallant, 2000; Wilson et al., 2007), i.e. on the crest of mound features (Figs. 2, 6 and 7). At Røst Reef, the high abundance of L. pertusa was mainly associated with middle to high values of MeanCur170, ProfileCur170 and BPI170, and middle to low values of PlanCur170, suggesting a tendency to occupy locations with slope increasing downhill on upper slopes and topographic divergences (Wilson and Gallant, 2000; Wilson et al., 2007), i.e. on the crests and upper section of the NW facing slopes of the large ridges and ridge-like features (Figs. 3, 8 and 9). High abundance of living CWCs has often been reported from areas with enhanced bottom current (Duineveld et al., 2012; Davies et al., 2009; Dorschel et al., 2007; Frederiksen et al., 1992; Mienis et al., 2007; Wagner et al., 2011). A recently developed model based on data describing benthic hydrodynamics and CWC occurrences illustrates that living CWCs are observed in areas with intensified near-bottom currents in contrast with areas with either a coral absence or a random background (Mohn et al., 2013).

Seabed topography features, such as the mound structure at Sobakken Reef, and the ridges and large ridge-like features at Røst Reef, may effectively enhance the inflowing bottom current velocity, thus enhancing nutrient supply, and therefore increasing particle encounter rates for the filter-feeding fauna such as L. pertusa (Davies et al., 2009; Dorschel et al., 2007; Mienis et al., 2009; Mienis et al., 2007; Wagner et al., 2011; White, 2007). L. pertusa framework may also play an important role in trapping the particulate organic matter (de Haas et al., 2009; Dorschel et al., 2007; Mienis et al., 2009; Mienis et al., 2012). Preliminary results of the Polarstern ARK XXII/1a (2007) expedition indicate that the reef community at Røst Reef biodeposits most of the labile phytodetritus within the bottom waters cross-shelf transported to the reef, with the chlorophyll a concentrations measured downstream diminished by passing through the reef (Moje et al., 2011). The enhanced bottom current velocity may resuspend a high percentage of trapped organic matter by promoting large amplitude local vertical mixing and organic matter fluxes, which may be in turn captured by feeding corals or again trapped within the coral structure or coral mucus (Mienis, 2008; Mienis et al., 2009; Mohn et al., 2013; Thiem et al., 2006). Additionally, the enhanced bottom

22

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

Fig. 9. Observed % L. pertusa cover in relation to (a) MeanCur170, (b) ProfileCur170, (c) BPI170, (d) Aspect90, (e) PlanCur170 at Røst Reef.

Fig. 10. Correlation of the multiscale terrain variables at Sotbakken Reef.

Fig. 11. Correlation of the multiscale terrain variables at Røst Reef.

current velocity promotes the periodic clearance of coral surfaces of deposited material, thus preventing living coral colonies from being buried by sediment (Dorschel et al., 2007; Duineveld et al., 2012; Mienis et al., 2007; White et al., 2005). Cold-water coral species rely on the food particles, i.e. zooplankton, suspended phytodetritus and particulate organic matter,

delivered by bottom currents (Carlier et al., 2009; Duineveld et al., 2004, 2007; Kiriakoulakis et al., 2007; van Oevelen et al., 2009). In this study, L. pertusa abundance was strongly correlated with Aspect170 at Sotbakken Reef, whilst strongly correlated with Aspect90 at Røst Reef. The bottom currents at Sotbakken Reef are likely to flow predominantly in a westerly direction, indicated by

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

23

Fig. 12. A example of the strongly correlated terrain variables at Sotbakken Reef – (a) BPI90, (b) BPI170, (c) MeanCur90, (d) ProfileCur90.

the predominant concave-side facing direction of the observed P. arborea colonies at the reef (Tong et al., 2012). L. pertusa was observed exclusively in areas with seabed orientation Aspect170 [0, 150°] (Figs. 6b and 7b), predominantly in areas with Aspect170 [80, 150°] facing into the inflowing prevalent bottom current to maximize the encounter rates with food particles. At Røst Reef, high L. pertusa abundance was observed exclusively in areas with seabed orientation Aspect90 [0, 90°] and [270, 360°], on the ridge area with a north to north-easterly facing direction [0, 90°] and on the upper section of the slide structure in a north to north-westerly facing direction [270, 360°] (Figs. 8d and 9d). However, the

bottom current flow at Røst Reef is prevalently in a NW direction, representing cross-shelf transport (Moje et al., 2011; Thiem et al., 2006). At the Røst Reef, the L. pertusa frameworks with dense live polyp cover were observed to be oriented predominantly SE, facing into the inflowing bottom current (Tong et al., 2012). This inconsistent result may be the result of the sampling bias related to Aspect90, as shown in Figs. 8d and 9d. The results of linear regression indicate that other factors are likely to play an important role in influencing L. pertusa abundance, e.g. the seabed substrate type (Howell et al., 2011; Mortensen and Buhl-Mortensen, 2004; Purser et al., 2013), with a hard

24

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

Fig. 13. A example of the strongly correlated terrain variables at Røst Reef – (a) BPI90, (b) BPI170, (c) MeanCur90, (d) ProfileCur90.

substrate such as bedrock, hardground or boulders being an essential requisite for coral settlement and growth. The widely distributed soft sediment observed around the coral reefs at Sotbakken Reef may play an important role in limiting the local cover of the CWCs, which is not the case at Røst Reef, with hardground and boulders widely distributed (Purser et al., 2013; Tong et al., 2012a). Additionally, anthropogenic impacts such as bottom trawling, directly influence the condition of CWC reefs (Fosså and Skjoldal, 2010; Murillo et al., 2011). Tong et al. (2012) reported that a large area of coral rubble was observed on the local topographic highs in the easterly mound crest region of the study area, whilst occasional rubble areas were observed at Røst Reef, with lost ropes recorded several times amongst the coral structure (Purser et al., 2013; Tong et al., 2012a). This damage and detritus are likely the result of heavy historic fishing pressure in these areas (Fosså and Skjoldal, 2010). 4.2. Relationships between terrain variables In a former study, terrain variables were summarized into four categories: (1) slope, (2) aspect, (3) plan curvature, mean curvature, profile curvature and BPI (curvature and relative position) and (4) rugosity and TRI (terrain variability) (Wilson et al., 2007).

However, according to the results of clusters of terrain variables produced using UPGMA in Figs. 10 and 11, the variables slope, TRI and rugosity at similar analysis scales are strongly correlated, and should for efficient analysis be classified into a single category. Similarly, mean curvature, profile curvature and BPI are also closely correlated. Therefore, terrain variables should most efficiently be grouped as follows: (1) slope, TRI and rugosity, (2) aspect, (3) plan curvature, (4) profile curvature, mean curvature and BPI. This result is supported by Evans (1979), who defined the five variables of elevation, slope, aspect, plan curvature and profile curvature for two dimensional continuous surface to have geomorphological meaning (Evans, 1979). The additional terrain variables investigated in this study-BPI and mean curvature, rugosity and TRI were closely correlated with these well defined terrain variables respectively. We found that the closely correlated terrain variables captured similar terrain features within both study areas. For example, BPI90/170, MeanCur90 and ProfileCur90 at Sotbakken Reef highlighted the local topographic highs and the NW-facing large slope (Fig. 12), whilst at Røst Reef the local topographic highs of the slide in the north and the NE-SW running ridges could be identified by these variables (Fig. 13). In former studies, these terrain variables were commonly used to model the habitat suitability for CWCs

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

(Dolan et al., 2008; Guinan et al., 2009a; Howell et al., 2011; Tong et al., 2013a, 2013b). According to our result, it is more efficient to select certain terrain variables for modeling, rather than including all variables into a model. Based on this study, in addition to aspect and plan curvature, the parameters of mean curvature (from the group BPI, mean curvature and profile curvature), and the parameter slope (from the slope, TRI30 and rugosity group), would appear to be usefully included in future habitat suitability modeling for L. pertusa. Each terrain variable investigated in this study was correlated across analysis scales, with a particularly strong correlation between 90 m and 170 m scales evident (Figs. 10 and 11). The strongly correlated variables captured similar terrain features, for example BPI90 and BPI170 highlighting the local topographic highs at Sotbakken Reef, and the ridges at Røst Reef (Figs. 12 and 13). This result suggests that in future, developing habitat suitability models with an increased spacing between the analysis window sizes (e.g. using window sizes of 3  3, 9  9 and 27  27) to compute the terrain variables. It would help to capture richer information on topographic variation over a greater range of scales, and may subsequently improve efforts to develop habitat suitability models.

5. Conclusions The observed L. pertusa cover at both Sotbakken Reef and Røst Reef was found to be significantly correlated with most of the investigated terrain variables, particularly strongly correlated with terrain variables at larger analysis scales, with increasingly strong correlations observed with analysis scale increase, indicating the importance of large scale terrain features in influencing L. pertusa distribution. A particularly strong correlation of L. pertusa with BPI and curvature, as well as with seabed aspect, indicate the tendency for the species to inhabit topographic highs exposed to the inflowing bottom current. The widely distributed soft sediments observed at Sotbakken Reef are likely to be an important factor in limiting L. pertusa expansion into the surrounding areas of seafloor. In addition to aspect and plan curvature, the strong correlation of the additional multiscale terrain variables suggest that the parameter mean curvature (from the group of BPI, mean curvature and profile curvature), and the parameter slope (from the group of slope, TRI3 and rugosity), are useful variables to consider in future development of habitat suitability models for L. pertusa. Additionally, the strong correlation between each terrain variable across the investigated analysis scales indicates that selecting a greater range in analysis (i.e. with a wider spread of scales than investigated here) in modeling efforts is to be recommended.

Acknowledgments This work is a contribution to the “Earth System Science Research School (ESSReS)”, an initiative of the Helmholtz Association of German research centers (HGF) at the Alfred Wegener Institute for Polar and Marine Research. Data contribution from the Polarstern ARK XXII/1a expedition (2007) is gratefully and duly acknowledged. The authors thank Dr. Margaret Dolan of Geological Survey of Norway for helpful comments on an earlier version of the manuscript. The authors also thank the anonymous reviewers, who spent valuable time and effort providing detailed comments which allowed for us to make significant improvements to the text. Ruiju Tong was funded by the School Scientific Research Fund of Fujian University of Technology (Grant no. GY-Z15123).

25

Additional funding to Autun Purser was provided through the European Community's Seventh Framework program (FP7/20072013) under the HERMIONE project (Grant agreement no. 226354). Jinsongdi Yu was funded by the National Natural Science Foundation of China (Grant agreement no. 41401454) and the China Civil Space Pre-research of Twelfth Five-year Plan (Grant agreement no. 01021405). Chengcheng Zhang was supported by the National Natural Science Foundation of China (Grant no. 41271451 and 41371425).

References Aitchison, J., 1986. The statistical analysis of compositional data. Chapman & Hall, London. Albani, M., Klinkenberg, B., Andison, D.W., Kimmins, J.P., 2004. The choice of window size in approximating topographic surfaces from Digital Elevation Models. Int. J. Geogr. Inf. Sci. 18, 577–593. Bryan, T.L., Metaxas, A., 2006. Distribution of deep-water corals along the North American continental margins: relationships with environmental factors. Deep-Sea Res. Pt I 53, 1865–1879. Buhl-Mortensen, L., Vanreusel, A., Gooday, A.J., Levin, L.A., Priede, I.G., Buhl-Mortensen, P., Gheerardyn, H., King, N.J., Raes, M., 2010. Biological structures as a source of habitat heterogeneity and biodiversity on the deep ocean margins. Mar. Ecol. 31, 21–50. Carlier, A., Le Guilloux, E., Olu, K., Sarrazin, J., Mastrototaro, F., Taviani, M., Clavier, J., 2009. Trophic relationships in a deep Mediterranean cold-water coral bank (Santa Maria di Leuca, Ionian Sea). Mar. Ecol. Prog. Ser. 397, 125–137. Davies, A.J., Duineveld, G.C.A., Lavaleye, M.S.S., Bergman, M.J.N., van Haren, H., Roberts, J.M., 2009. Downwelling and deep-water bottom currents as food supply mechanisms to the cold-water coral Lophelia pertusa (Scleractinia) at the Mingulay Reef Complex. Limnol. Oceanogr. 54, 620–629. Davies, A.J., Guinotte, J.M., 2011. Global habitat suitability for framework-forming cold-water corals. PLoS ONE 6, e18483. de Haas, H., Mienis, F., Frank, N., Richter, T., Steinacher, R., de Stigter, H., van der Land, C., van Weering, T., 2009. Morphology and sedimentology of (clustered) cold-water coral mounds at the south Rockall Trough margins, NE Atlantic Ocean. Facies 55, 1–26. Dodds, L.A., Black, K.D., Orr, H., Roberts, J.M., 2009. Lipid biomarkers reveal geographical differences in food supply to the cold-water coral Lophelia pertusa (Scleractinia). Mar. Ecol. Prog. Ser. 397, 113–124. Dolan, M.F.J., Grehan, A.J., Guinan, J.C., Brown, C., 2008. Modelling the local distribution of cold-water corals in relation to bathymetric variables: adding spatial context to deep-sea video data. Deep-Sea Res. Pt. I 55, 1564–1579. Dorschel, B., Hebbeln, D., Foubert, A., White, M., Wheeler, A.J., 2007. Hydrodynamics and cold-water coral facies distribution related to recent sedimentary processes at Galway Mound west of Ireland. Mar. Geol. 244, 184–195. Duineveld, G.C.A., Jeffreys, R.M., Lavaleye, M.S.S., Davies, A.J., Bergman, M.J.N., Watmough, T., Witbaard, R., 2012. Spatial and tidal variation in food supply to shallow cold-water coral reefs of the Mingulay Reef complex (Outer Hebrides, Scotland). Mar. Ecol. Prog. Ser. 444, 97–115. Duineveld, G.C.A., Lavaleye, M.S.S., Berghuis, E.M., 2004. Particle flux and food supply to a seamount cold-water coral community (Galicia Bank, NW Spain). Mar. Ecol. Prog. Ser. 277, 13–23. Duineveld, G.C.A., Lavaleye, M.S.S., Bergman, M.J.N., de Stigter, H., Mienis, F., 2007. Trophic structure of a cold-water coral mound community (Rockall Bank, NE Atlantic) in relation to the near-bottom particle supply and current regime. B. Mar. Sci. 81, 449–467. Dullo, W.-C., Flögel, S., Rüggeberg, A., 2008. Cold-water coral growth in relation to the hydrography of the Celtic and Nordic European continental margin. Mar. Ecol. Prog. Ser. 371, 165–176. Evans, I.S., 1979. An integrated system of terrain analysis and slope mapping, Final report on grant DA-ERO-591-73-G0040. University of Durham, England. Fabri, M.C., Pedel, L., Beuck, L., Galgani, F., Hebbeln, D., Freiwald, A., 2014. Megafauna of vulnerable marine ecosystems in French mediterranean submarine canyons: spatial distribution and anthropogenic impacts. Deep-Sea Res. Pt. II 104, 184–207. Fosså, J.H., Alvsvåg, J., Ottesen, D., Mortensen, P.B., 2004. Protection and management of deep-water coral reefs in Norway. In: Proceedings of the International Council for the Exploration of the Sea, CM AA, 08. Fosså, J.H., Lindberg, B., Christensen, O., Lundälv, T., Svellingen, I., Mortensen, P.B., Alvsvåg, J., 2005. Mapping of Lophelia reefs in Norway: experiences and survey methods. In: Freiwald, A., Roberts, J.M. (Eds.), Cold-water Corals and Ecosystems. Springer-Verlag Berlin Heidelberg, p. 22. Fosså, J.H., Mortensen, P.B., Furevik, D.M., 2002. The deep-water coral Lophelia pertusa in Norwegian waters: distribution and fishery impacts. Hydrobiologia 471, 1–12. Fosså, J.H., Skjoldal, H.R., 2010. Conservation of Cold-Water Coral Reefs in Norway. In: Grafton, R.Q., Hilborn, R., Squires, D., Tait, M., Williams, M.J. (Eds.), Handbook of Marine Fisheries Conservation and Management. Oxford University Press, New York, pp. 215–240. Frederiksen, R., Jensen, A., Westerberg, H., 1992. The distribution of the

26

R. Tong et al. / Continental Shelf Research 116 (2016) 13–26

scleractinian coral Lophelia pertusa around the Faeroe Islands and the relation to internal tidal mixing. Sarsia 77, 157–171. Freiwald, A., Fosså, J.H., Grehan, A., Koslow, T., Roberts, J.M., 2004. Cold-water coral reefs. UNEP-WCMC, Cambridge, UK, p. 84. Freiwald, A., Huehnerbach, V., Lindberg, B., Wilson, J.B., Campbell, J., 2002. The Sula Reef Complex, Norwegian Shelf. Facies 47, 179–200. Giusti, M., Innocenti, C., Canese, S., 2014. Predicting suitable habitat for the gold coral Savalia savaglia (Bertoloni, 1819) (Cnidaria, Zoantharia) in the South Tyrrhenian Sea. Cont. Shelf. Res. 81, 19–28. Guinan, J., Brown, C., Dolan, M.F.J., Grehan, A.J., 2009a. Ecological niche modelling of the distribution of cold-water coral habitat using underwater remote sensing data. Ecol. Inf. 4, 83–92. Guinan, J., Grehan, A.J., Dolan, M.F.J., Brown, C., 2009b. Quantifying relationships between video observations of cold-water coral cover and seafloor features in Rockall Trough, west of Ireland. Mar. Ecol. Prog. Ser. 375, 125–138. Henry, L.A., Roberts, J.M., 2007. Biodiversity and ecological composition of macrobenthos on cold-water coral mounds and adjacent off-mound habitat in the bathyal Porcupine Seabight, NE Atlantic. Deep-Sea Res. Pt. I 54, 19. Hovland, M., Risk, M., 2003. Do Norwegian deep-water coral reefs rely on seeping fluids? Mar. Geol. 198, 83–96. Howell, K.L., Holt, R., Endrino, I.P., Stewart, H., 2011. When the species is also a habitat: Comparing the predictively modelled distributions of Lophelia pertusa and the reef habitat it forms. Biol. Conserv. 144, 2656–2665. Jenness, J., 2006. Topographic Position Index (tpi_jen.avx) extension for ArcView 3. x, v. 1.3a. Jenness Enterprises, Available at 〈〈http://www.jennessent.com/arc view/tpi.htm〉〉. Kiriakoulakis, K., Freiwald, A., Fisher, E., Wolff, G., 2007. Organic matter quality and supply to deep-water coral/mound systems of the NW European Continental Margin. Int. J. Earth Sci. 96, 159–170. Klages, M., Thiede, J.E., 2011. The Expedition of the Research Vessel’“Polarstern” to the Arctic in 2007 (ARK-XXII/1a-c). Kutti, T., Bergstad, O.A., Fosså, J.H., Helle, K., 2014. Cold-water coral mounds and sponge-beds as habitats for demersal fish on the Norwegian shelf. Deep Sea Res. Part II: Topical Stud. Oceanogr. 99, 122–133. Lavaleye, M., Duineveld, G., Lundälv, T., White, M., Guihen, D., Kiriakoulakis, K., Wolff, G.A., 2009. Cold-Water corals on the Tisler Reef- preliminary observations on the dynamic reef environment. Oceanography 22, 76–84. Lo Iacono, C., Gracia, E., Ranero, C.R., Emelianov, M., Huvenne, V.A.I., Bartolome, R., Booth-Rea, G., Prades, J., 2014. The West Melilla cold water coral mounds, Eastern Alboran Sea: morphological characterization and environmental context. Deep Sea Res. Part II: Topical Stud. Oceanogr. 99, 316–326. Mienis, F., 2008. Environmental Constraints on Cold-Water Coral Growth and Carbonate Mound Formation. Vrije Universiteit, Amsterdam, p. 209. Mienis, F., de Stigter, H.C., de Haas, H., van Weering, T.C.E., 2009. Near-bed particle deposition and resuspension in a cold-water coral mound area at the Southwest Rockall Trough margin, NE Atlantic. Deep-Sea Res. Pt. I 56, 1026–1038. Mienis, F., de Stigter, H.C., White, M., Duineveld, G., de Haas, H., van Weering, T.C.E., 2007. Hydrodynamic controls on cold-water coral growth and carbonatemound development at the SW and SE Rockall Trough Margin, NE Atlantic Ocean. Deep-Sea Res. Pt. I 54, 1655–1674. Mienis, F., Duineveld, G.C.A., Davies, A.J., Ross, S.W., Seim, H., Bane, J., van Weering, T.C.E., 2012. The influence of near-bed hydrodynamic conditions on cold-water corals in the Viosca Knoll area, Gulf of Mexico. Deep-Sea Res. Pt. I 60, 32–45. Mohn, C., Rengstorf, A., White, M., Duineveld, G., Mienis, F., Soetaert, K., Grehan, A., 2013. Linking benthic hydrodynamics and cold-water coral occurrences: a highresolution model study at three cold-water coral provinces in the NE Atlantic. Prog. Oceanogr. 122, 92–104. Mohn, C., Rengstorf, A., White, M., Duineveld, G., Mienis, F., Soetaert, K., Grehan, A., 2014. Linking benthic hydrodynamics and cold-water coral occurrences: a highresolution model study at three cold-water coral provinces in the NE Atlantic. Prog. Oceanogr. 122, 92–104. Moje, A., Purser, A., Thomsen, L., Unnithan, V., Viergutz, T., 2011. Hydrodynamics and coral communities. in: Klages, M., Thiede, J. (Eds.) Proceedings of the The Expedition of the Research Vessel’“Polarstern” to the Arctic in 2007 (ARK-XXII/ 1a-c), pp. 32–36. Mortensen, P.B., 2000. Lophelia pertusa (Scleractinia) in Norwegian waters-Distribution, growth, and associated fauna. Department of Fisheries and Marine Biology, Department of Fisheries and Marine Biology. University of Bergen, Norway, pp. 1–42. Mortensen, P.B., Buhl-Mortensen, L., 2004. Distribution of deep-water gorgonian corals in relation to benthic habitat features in the Northeast Channel (Atlantic Canada). Mar. Biol. 144, 1223–1238. Murillo, F.J., Duran Munoz, P., Altuna, A., Serrano, A., 2011. Distribution of deepwater corals of the Flemish Cap, Flemish Pass, and the Grand Banks of Newfoundland (Northwest Atlantic Ocean): interaction with fishing activities. ICES J. Mar. Sci. 68, 319–332. Ottesen, D., Stokes, C.R., Rise, L., Olsen, L., 2008. Ice-sheet dynamics and ice streaming along the coastal parts of northern Norway. Q. Sci. Rev. 27, 922–940. Purser, A., 2010. Cold-Water Corals. Distribution of fauna and responses to environmental perturbation, School of Engineering and Science. Jacobs University Bremen, Bremen. Purser, A., Orejas, C., Gori, A., Tong, R., Unnithan, V., Thomsen, L., 2013. Local variation in the distribution of benthic megafauna species associated with coldwater coral reefs on the Norwegian margin. Cont. Shelf Res. 54, 37–51. Qurban, M.A., Krishnakumar, P.K., Joydas, T.V., Manikandan, K.P., Ashraf, T.T.M., Quadri, S.I., Wafar, M., Qasem, A., Cairns, S.D., 2014. In-situ observation of deep

water corals in the northern Red Sea waters of Saudi Arabia. Deep-Sea Res. Pt. I 89, 35–43. Rüggeberg, A., Flögel, S., Dullo, W.-C., Hissmann, K., Freiwald, A., 2011. Water mass characteristics and sill dynamics in a subpolar cold-water coral reef setting at Stjernsund, northern Norway. Mar. Geol. 282, 5–12. Rengstorf, A.M., Grehan, A., Yesson, C., Brown, C., 2012. Towards high-resolution habitat suitability modeling of vulnerable marine ecosystems in the Deep-Sea: resolving terrain attribute dependencies. Mar. Geod. 35, 343–361. Rengstorf, A.M., Yesson, C., Brown, C., Grehan, A.J., 2013. High-resolution habitat suitability modelling can improve conservation of vulnerable marine ecosystems in the deep sea. J. Biogeogr. 40, 1702–1714. Riley, S.J., DeGloria, S.D., Elliot, R., 1999. A terrain ruggedness index that quantifies topographic heterogeneity. Intermountain J. Sci. 5, 23–27. Roberts, J., Wheeler, A.J., Freiwald, A., 2006. Reefs of the deep: the biology and geology of cold-water coral ecosystems. Science 312, 543–547. Roberts, J.M., Brown, C.J., Long, D., Bates, C.R., 2005. Acoustic mapping using a multibeam echosounder reveals cold-water coral reefs and surrounding habitats. Coral Reefs 24, 16. Roberts, J.M., Wheeler, A., Freiwald, A., Cairns, S., 2009a. Cold-Water Corals: The Biology and Geology of Deep-Sea Coral Habitats. Cambridge University Press, Cambridge. Roberts, J.M., Wheeler, A., Freiwald, A., Cairns, S., 2009b. Cold-Water Corals: The Biology and Geology of Deep-Sea Coral Habitats. Cambridge University Press, 25–28. Rowden, A.A., Schlacher, T.A., Williams, A., Clark, M.R., Stewart, R., Althaus, F., Bowden, D.A., Consalvey, M., Robinson, W., Dowdney, J., 2010. A test of the seamount oasis hypothesis: seamounts support higher epibenthic megafaunal biomass than adjacent slopes. Mar. Ecol. 31, 95–106. Savini, A., Vertino, A., Marchese, F., Beuck, L., Freiwald, A., 2014. Mapping ColdWater Coral Habitats at Different Scales within the northern Ionian Sea (central Mediterranean): an assessment of coral coverage and associated vulnerability. PLoS ONE 9 (1), e87108. Sokal, R.R., Michener, C.D., 1958. A statistical method for evaluating systematic relationships 38. University of Kansas Science Bulletin, pp. 1409–1438. Somoza, L., Ercilla, G., Urgorri, V., León, R., Medialdea, T., Paredes, M., Gonzalez, F.J., Nombela, M.A., 2014. Detection and mapping of cold-water coral mounds and living Lophelia reefs in the Galicia Bank, Atlantic NW Iberia margin. Mar. Geol. 349, 73–90. Thiem, Ø., Ravagnan, E., Fosså, J.H., Berntsen, J., 2006. Food supply mechanisms for cold-water corals along a continental shelf edge. J. Marine Syst. 60, 207–219. Thomsen, L., 2002. The benthic boundary layer. In: Wefer, G., Billet, D., Hebbeln, D., Joergensen, B., Schlüter, M., van Weering, T.J. (Eds.), Ocean Margin Systems. Springer, Berlin Heidelberg, pp. 143–155. Tittensor, D.P., Baco, A.R., Brewin, P.E., Clark, M.R., Consalvey, M., Hall-Spencer, J., Rowden, A.A., Schlacher, T., Stocks, K.I., Rogers, A.D., 2009. Predicting global habitat suitability for stony corals on seamounts. J. Biogeogr. 36, 1111–1128. Tong, R., Purser, A., Guinan, J., Unnithan, V., 2013a. Modeling the habitat suitability for deep-water gorgonian corals based on terrain variables. Ecol. Inf. 13, 123–132. Tong, R., Purser, A., Unnithan, V., 2013b. Predicting habitat suitability for cold-water coral Lophelia pertusa using multiscale terrain variables. In: Lohmann, G., Grosfeld, K., Wolf-Gladrow, D., Unnithan, V., Notholt, J., Wegner, A. (Eds.), Earth System Science: Bridging the Gaps between Disciplines. Springer, Bremen, pp. 113–118. Tong, R., Purser, A., Unnithan, V., Guinan, J., 2012. Multivariate statistical analysis of distribution of deep-water Gorgonian corals in relation to seabed topography on the Norwegian margin. Plos One 7, e43534. van Oevelen, D., Duineveld, G., Lavaleye, M., Mienis, F., Soetaert, K., Heip, C.H.R., 2009. The cold-water coral community as a hot spot for carbon cycling on continental margins: A food-web analysis from Rockall Bank (northeast Atlantic). Limnol. Oceanogr. 54, 1829–1844. Vertino, A., Savini, A., Rosso, A., Di Geronimo, I., Mastrototaro, F., Sanfilippo, R., Gay, G., Etiope, G., 2010. Benthic habitat characterization and distribution from two representative sites of the deep-water SML Coral Province (Mediterranean). Deep Sea Res. Part II: Topical Stud. Oceanogr. 57, 380–396. Wagner, H., Purser, A., Thomsen, L., Jesus, C.C., Lundälv, T., 2011. Particulate organic matter fluxes and hydrodynamics at the Tisler cold-water coral reef. J. Mar. Syst. 85, 19–29. Wheeler, A., Beyer, A., Freiwald, A., de Haas, H., Huvenne, V.A.I., Kozachenko, M., Roy, K.O.L., Opderbecke, J., 2007. Morphology and environment of cold-water coral carbonate mounds on the NW European margin. Int. J. Earth Sci. 96, 20. White, M., 2007. Benthic dynamics at the carbonate mound regions of the Porcupine Sea Bight continental margin. Int. J. Earth Sci. 96, 1–9. White, M., Mohn, C., Stigter, H., Mottram, G., 2005. Deep-water coral development as a function of hydrodynamics and surface productivity around the submarine banks of the Rockall Trough, NE Atlantic. In: Freiwald, A., Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems.. Springer, Berlin, pp. 503–514. Wilson, J.P., Gallant, J.C., 2000. Digital Terrain Analysis. In: Wilson, J.P., Gallant, J.C. (Eds.), Terrain Analysis: Principles and Applications. John Wiley & Sons, Inc, New York, pp. 1–26. Wilson, M.F.J., O'Connell, B., Brown, C., Guinan, J.C., Grehan, A.J., 2007. Multiscale terrain analysis of multibeam bathymetry data for habitat mapping on the continental slope. Mar. Geod. 30, 3–35. Wood, J., 1996. The Geomorphological Characterisation of Digital Elevation Models. University of Leicester, Leicester, UK, Leicester, UK.