Marine Biology

, Volume 144, Issue 6, pp 1223–1238

Distribution of deep-water gorgonian corals in relation to benthic habitat features in the Northeast Channel (Atlantic Canada)

Authors

    • Institute of Marine Research
  • Lene Buhl-Mortensen
    • Institute of Marine Research
Research Article

DOI: 10.1007/s00227-003-1280-8

Cite this article as:
Mortensen, P.B. & Buhl-Mortensen, L. Marine Biology (2004) 144: 1223. doi:10.1007/s00227-003-1280-8

Abstract

The distribution and abundance of deep-water gorgonian corals were investigated along 52 transects at 183–498 m depth in the Northeast Channel, between Georges Bank and Browns Bank in the northwest Atlantic, using a remotely operated vehicle and a towed video-camera system. Three species (Paragorgia arborea, Primnoa resedaeformis, and Acanthogorgia armata) were observed. Primnoa occurred on 35 transects below 196 m depth, with highest local abundance in stands of 104 colonies per 100 m2. Paragorgia was present on 21 transects deeper than 235 m, with highest local abundance of 49 colonies per 100 m2. Acanthogorgia was observed at only four transects between 231 m and 364 m, with a local maximum abundance of 199 colonies per 100 m2. The maximum abundance averaged for whole transects was 19.2 and 6.2 colonies per 100 m2 for Primnoa and Paragorgia, respectively. The corals were more common in the outer part of the channel along the shelf break and slope than on the shelf in the inner part. All three species showed a patchy distribution with no signs of competitive exclusion at any spatial scale. Transects with high abundance of corals were characterised by depths greater than 400 m, maximum temperatures less than 9.2°C, and a relatively high percentage coverage of cobble and boulder (more than 19% and 6%, respectively). High temperatures probably control the upper depth limit of the corals, and Primnoa seems to tolerate slightly higher temperatures than Paragorgia. Abundance of both species was negatively correlated with average temperature and positively with cobbles. Together, temperature, percentage cobble and salinity accounted for 38% of the variance of Primnoa. The comparable figure for Paragorgia was 15%. The observed distribution indicated that the abundance of coral is controlled by additional factors such as larger-scaled topographic features governing the current regimes and thus also the supply of food and larvae.

Introduction

It is well known that corals in the photic zone of temperate and tropical waters form complex habitats with a high diversity of associated invertebrates and fish. However, less is known about the biology and ecology of deep-water corals, especially deep-water gorgonians. Thus, the role of corals in deep-water ecosystems as habitat for other species, and their possible link to shallower shelf or pelagic ecosystems is poorly understood. At the same time, there is a growing concern that human activities (e.g. fishing and oil exploration) along the continental shelf edges and slopes may damage both the coral and their habitats in these environments. As fish stocks in shallower water have been overexploited, fisheries are moving into deeper water. Deep-water corals are thought to be particularly vulnerable to disturbance because of their arborescent morphology and assumed slow growth. Deep-water gorgonians are found in all of the world’s oceans and the Mediterranean Sea, commonly at depths between 200 m and 1,000 m (Broch 1935, 1957; Madsen 1944; Tendal 1992; Mistri and Ceccherelli 1994; Heifetz 2002), but have been recorded from depths as great as 4,000 m in the Northeast Atlantic (Grasshoff 1982). Ten gorgonian species have been recorded off Atlantic Canada, mainly below 200 m depth along the shelf edge, in submarine canyons and in deep channels between fishing banks (Verrill 1922; Deichman 1936; Breeze et al. 1997; MacIsaac et al. 2001). The three gorgonian species subject for this study (Acanthogorgia armata, Paragorgia arborea, and Primnoa resedaeformis) occur on both sides of the North Atlantic Ocean (Madsen 1944; Tendal 1992; Breeze et al. 1997). Paragorgia arborea and Primnoa resedaeformis are found co-occurring in the boreal Atlantic and Pacific (Broch 1935), whereas Acanthogorgia is confined to the Atlantic (Madsen 1944).

Information on distribution and abundance of deep-water gorgonians is scarce, but is needed to enable the assessment of human impacts. Quantitative information on sessile deep-water fauna is hard to obtain with traditional benthic sampling equipment (i.e. trawl and dredge) because of the poor spatial resolution and unknown capture efficiency. Fortunately, recent development of remotely operated vehicles (ROVs), submersibles and video-equipped towed gear has provided an opportunity for controlled sampling and detailed observation of specific deep-water habitats. This study is based on direct observations with an ROV and a towed platform equipped with video and still-photo camera, which have a minimal destructive effect on coral habitats.

The objectives for this study are as follows: (1) to describe the distribution and abundance of corals in the Northeast Channel; (2) to compare the abundance of corals in the study area with other areas; and (3) to relate the distribution patterns to a number of environmental factors (bottom type, depth, topography, and hydrography) in order to discern small-scale and broad-scale patterns of associations. Quantifying these patterns is an important first step towards understanding the environmental controls on deep-water gorgonians off Nova Scotia.

Materials and methods

Description of the study area

The Northeast Channel crosses the Scotian Shelf between Browns Bank and Georges Bank (Fig. 1A, B). It connects the Gulf of Maine and the North Atlantic Ocean, and has a sill depth of about 230 m. The seabed is composed of thick glacial till (Scotian Shelf Drift) deposited by grounded glaciers (Lawrence et al. 1985; Fader et al. 1988). Linear iceberg plough marks and circular iceberg pits are widespread. Both pits and plough marks are flanked by berms of boulders (Fader et al. 1988). Three shallower, ridged areas at the shelf break represent terminal moraines. The central ridge is known as Romeys Peak. Beyond the shelf break (>400 m), the seabed drops off steeply and the seascape changes into submarine canyons. Some areas have sand transported by strong currents.
Fig. 1 A

Location of the Northeast Channel. B The topography of the Northeast Channel, based on multi-beam bathymetry collected by the Canadian Hydrographic Service in 1999. C Map of the Northeast Channel showing the location of video transects as filled circles

Salinity and temperature at 300–400 m range from 34.87‰ to 35.05‰ and 5.74°C to 7.64°C, respectively (Petrie et al. 1996). Current measurements are not available close to the bottom at the investigated sites, but strong semidiurnal tidal currents are generally dominant in the Northeast Channel (Ramp et al. 1985; Smith and Schwing 1991). The maximum speed 16 m off the bottom at the sill is between 40 cm s-1 and 50 cm s-1 (Ramp et al. 1985). The inflow occurs on the northeastern side while the current is directed outward on the southwestern side (Ramp et al. 1985; Shore et al. 2000; Loder et al. 2001).

To aid the discussion of coral distribution, three geographical divisions were defined (Fig. 1C), each covering individual submarine valleys as follows: (1) the northern valley, covering the area between the flank of Browns Bank and Romeys Peak; (2) the middle valley, between Romeys Peak and the area known as the Rips by local fishers; and (3) the southern valley just west of the Rips.

Selection of study sites

The study was limited to sites <500 m, because the inspection gear could not reach to greater depths. In total, 48 sites (Fig. 1C) covering a depth range of 183–498 m were selected for deployment of the gear. A detailed bathymetric map (Fig. 1B) from recent mapping with a multi-beam echo sounder by the Canadian Hydrographic Service, was used as a guide. Twenty-three sites were selected within areas with seabed topography most likely to support corals (i.e. at noses and ridges along the shelf edge at depths in the range 300–500 m). Previous studies (e.g. Genin et al. 1986; Mortensen et al. 2001) have shown that rugged seafloor often supports deep-water coral communities. The rest of the sites were selected randomly in order to describe the geographical distribution and the upper depth limits for the corals.

Video recording of the seabed

Fifty-two transects were video recorded from the 48 sites with the ROV ROPOS (remotely operated platform for ocean science), or the video and still photo camera system Campod. Both ROPOS and Campod were equipped with an ultra-short-baseline navigation system (ORE Trackpoint II) providing detailed records of their tracks along the seabed. The geographical positions provided by the navigation instruments were quite noisy with an error in the order of ±5 m. However, after post-processing and filtering of the data the error was reduced to ±2 m. Transects varied between 301 m and 1,683 m in length. A total distance of 32.4 km along the seabed was video-recorded, covering an area of approximately 97,095 m2. The total time of the video records was 20 h and 5 min.

ROPOS is a specialised ROV operated by the Canadian Scientific Submersible Facility. It is designed for scientific purposes, equipped with specialised sampling equipment, depth sensor, compass, and two parallel laser beams. The laser beams provided a 10 cm scale for measuring the width of the video frame. ROPOS was operated in a shallow (<500 m depth) diving mode without the heavy docking/transport station, which is needed for work at greater depths. Seven transects (Fig. 1C, 636–642) were recorded with ROPOS in August 2001, during a cruise with C.C.G.S. Martha L. Black. The transects covered depths between 331 m and 498 m and varied from 301 m to 1,683 m in length (Table 1). The plan for the dives was to move along 1.5 km transects on selected ridges. However, strong currents restricted the ROV to staying within areas of approximately 400 m2 at each dive site.
Table 1

Brief description of visual transects where corals were observed in the Northeast Channel. Mean percentage cover is given for sand, pebble, cobble, and boulder. Number of colonies and mean abundance are given for Paragorgia arborea, Primnoa resedaeformis and Acanthogorgia armata

Geara

Transect number

Depth (m)

Length (m)

Mean cover (%)b

Paragorgiacd

Primnoacd

Acanthogorgiacd

Sa

Pe

Co

Bo

n

Colonies per 100 m2

Id

n

Colonies per 100 m2

Id

n

Colonies per 100 m2

Id

The northern valley

  Campod

2

225–265

1,163

10

46

37

8

0

0

13

0.4 (9)

15.0

0

0

  Campod

5

247–250

1,087

86

5

7

4

0

0

57

1.2 (10)

5.6

0

0

  Campod

7

246–251

796

71

16

8

5

0

0

17

0.61 (6.1)

5.3

0

0

  Campod

8

288–319

1,010

60

15

17

8

4

0.1 (2.5)

11.5

74

2.18 (26.1)

6.8

5

0.4 (12.8)

25.4

  Campod

12

357–385

594

58

24

12

7

0

0

1

0.07 (3.3)

0

0

  Campod

13

334–364

266

90

5

5

1

0

0

0

0

0

0

  Campod

14

355

96

94

3

2

1

0

0

0

0

0

0

  Campod

16

354–373

889

48

26

20

6

19

0.6 (3.1)

1.1 (ns)

155

4.7 (18.4)

2.1

2

0.03 (1.2)

36 (ns)

  Campod

17

361–366

784

78

14

6

3

10

0.7 (15.7)

14.7

20

1 (18.0)

10.1

2

0.1 (2.5)

39

  Campod

18

301–319

683

40

29

23

9

1

0.1 (2.5)

108

6.8 (80.4)

5.2

0

0

  Campod

102

266–369

633

96

3

0

0

0

0

0

0

0

0

  Campod

103

236–251

408

37

29

25

9

0

0

0

0

0

0

  Campod

104

212–226

269

47

28

23

2

0

0

0

0

0

0

  Campod

105

287–309

725

25

33

25

17

0

0

3

0.03 (0.5)

52.9 (ns)

0

0

  Campod

106

312–341

321

40

17

28

10

1

0.2 (3.3)

14

6.0 (31.4)

3.6

0

0

  Campod

107

373–407

759

71

21

8

1

0

0

0

0

0

0

  Campod

108

416–422

340

71

18

2

0

0

0

0

0

0

0

  Campod

121

246

405

61

14

20

6

0

0

0

0

0

0

  Campod

122

255–265

354

86

1

4

4

2

0.3 (7.6)

22.0 (ns)

1

0.3 (4.9)

0

0

  Campod

123

290–321

826

31

38

22

8

1

0.1 (6.9)

7

0.9 (25.6)

92.4 (ns)

0

0

  Campod

139

247–248

382

59

21

18

2

0

0

0

0

0

0

  Campod

140

240–245

407

95

4

1

0

0

0

0

0

0

0

  ROPOS

636

462–498

301

52

28

14

6

17

2.3 (11)

2.1

162

19.0 (74.2)

2.1

0

0

  ROPOS

637

464–486

1,683

49

15

26

10

89

4.0 (23.3)

1.7

582

19.2 (104)

2.1

0

0

  ROPOS

638

457–476

1,018

44

17

29

10

16

0.9 (6.2)

3.5

228

10.6 (99.6)

3.6

0

0

  ROPOS

642

458–485

494

43

13

42

2

40

3.3 (10.0)

1.6

466

37.6 (136.0)

1.7

0

0

The middle valley

  Campod

1

233–261

598

38

33

23

6

0

0

75

4.3 (44.4)

4.8

0

0

  Campod

3

303–327

493

10

55

29

6

0

0

3

0.3 (13.5)

49.0

0

0

  Campod

6

230–237

848

55

22

14

9

0

0

45

1.5 (27.0)

10.0

203

17.6 (199.1)

6.4

  Campod

9

354–387

1,076

25

44

19

11

1

0.02 (1.1)

56

1.5 (18.5)

6.7

0

0

  Campod

11

372–384

310

21

50

19

10

0

0

0

0

0

0

  Campod

109

383–458

1,590

42

26

26

6

2

0.1 (6.2)

59.2

27

0.7 (12.5)

9.2

0

0

  Campod

110

240–270

567

92

2

4

2

0

0

47

1.6 (14.5)

6.6

0

0

  Campod

115

181–188

360

78

7

10

4

0

0

0

0

0

0

  Campod

116

251–265

399

40

15

26

19

8

0.5 (6.5)

6.3

53

4.9 (84.8)

8.9

0

0

  Campod

117

304–323

530

20

30

32

18

21

1.0 (16.1)

7.6

105

8.1 (70.4)

4.3

0

0

  Campod

118

343–365

715

30

33

26

12

80

6.2 (49.0)

2.5

68

4.9 (50.5)

4.8

0

0

  Campod

119

284–312

628

63

10

15

13

2

0.3 (7.1)

19.9

88

14.6 (121.4)

4.8

0

0

  Campod

120

256–258

371

50

21

20

10

0

0

0

0

0

0

  Campod

135

244–260

387

57

19

22

2

0

0

0

0

0

0

  Campod

137

249–257

792

35

16

36

13

0

0

1

0.1 (2.4)

0

0

  Campod

138

254

362

61

19

17

2

0

0

0

0

0

0

  ROPOS

639

387–426

792

36

54

6

4

1

0.2 (1.8)

12

1.9 (9.8)

3.4

0

0

  ROPOS

640

423–452

466

56

14

16

14

5

1.1 (6.1)

2.4

36

3.6 (12.8)

2.0

0

0

The southern valley

  Campod

111

367–389

469

48

19

24

9

0

0

6

0.3 (6.3)

20.5

0

0

  Campod

113

196–211

674

39

27

26

8

0

0

67

2.8 (8.6)

1.7

0

0

  Campod

114

242–279

540

38

30

23

9

1

0.1 (2.0)

21

4.0 (91.4)

15.7

0

0

  Campod

132

201–223

536

49

13

25

12

0

0

5

0.2 (1.9)

6.4 (ns)

0

0

  Campod

133

204–213

583

59

14

17

10

0

0

2

0.1 (1.8)

6.5 (ns)

0

0

  Campod

134

183–189

354

69

11

14

6

0

0

0

0

0

0

  Campod

136

196–204

503

97

1

1

0

0

0

0

0

0

0

  ROPOS

641

330–383

729

24

20

42

16

1

0.1 (3.8)

38

3.5 (24.3)

3.66

0

0

Sum

32,365

322

2,663

212

aGear was either the video and still-photo camera system Campod, or ROPOS (remotely operated platform for ocean science)

bSa Sand, P pebble, C cobble, B boulder

cn Number of colonies, Id Morisita’s index, insufficient data

dMaximum values are given in parentheses

Campod is an observation platform without propulsion. It is equipped with a high-resolution video camera for viewing the seabed directly below and an oblique video camera providing an overview of the seabed ahead (Gordon et al. 2000). Campod was deployed on 45 sites within the study area during two cruises (HUD2000-020 and HUD2001-055) with C.C.G.S. Hudson in June 2000 (transects 1–18), and September/October 2001 (transects 102–140). Campod was deployed while the ship was slowly (<1 kn) drifting over the investigation site, and was kept close (1–2 m) to the seabed for at least 15 min on each tow. The Campod transects covered depths between 196 m and 458 m, and were between 321 m and 1,590 m long.

Video analysis

To capture fine-scale distribution of corals and habitat features, the video records of the video transects were divided into shorter intervals later referred to as video sequences. These sequences were mainly of 30 s duration but were made shorter when abrupt changes in the habitat occurred earlier. Geographical positions and depth were registered at the start and end of each sequence. In total, the 52 video-recorded transects were divided into 1,751 sequences for analysis. Each video sequence covered a distance of 5–25 m (the average being 14 m), estimated from the navigation data. Abundance of coral colonies was estimated by dividing the number of colonies within a video sequence with the approximate area of the sequence. There were two sources of errors in estimated area of video sequences: inaccurate geographical positions, and variable width of the visual field. The navigation error of ±2 m induced an error of 17% in the abundance of corals in video sequences. Unfortunately, there was no practical way to estimate the width of the visual field continuously, which varied with the height above bottom and the pitch and roll angles of the video-camera lens. However, most of the time the inspection platforms were kept ca 1.5 m above the bottom, with very little variation in the pitch and roll. This provided a visual field width of 2 m and 4 m respectively for the vertical and oblique cameras on Campod, and 4.5 m for the camera on ROPOS. The variable field width probably induced an error in coral abundance <10%. The errors in area estimates has a much smaller influence on the abundance estimates at a large scale (whole transects), than at a small scale (video sequences).

The percentage cover of substrate types (i.e. sand, pebble, cobble, and boulder) was estimated to a precision of 5% intervals (0, 5, 10, .... 100%) from the video sequences, following the size classes as defined by the Wentworth scale (Wentworth 1922). The presence of any bottom type with a coverage <5% was given the value 1%. In cases where the substratum composition or cover varied within a video sequence, average values of two or more estimates were used. This method is not necessarily as precise as a point-count method. Percentage cover is estimated with the point-count method as the ratio of points in a grid net overlying a certain taxon or seabed type to the total number of points in the grid net. However, this method would have been very time-consuming to apply to all transects. Estimates gained by the visual “subjective” method were compared with estimates made by counting points overlying the different substratum classes in a grid of 20 points. The comparison was based on parallel estimates applying the two techniques to 155 video frames from three transects with a total of 72 video sequences. The correlation between the two estimation techniques was good (r2=0.86, P<0.001) and justified using the simple and fast visual method.

The roughness of the seabed was quantified for each of the video transects, for correlation with the abundance of corals. An index of roughness was estimated as the three dimensional length of the bottom profile along the transect divided by the two dimensional length. The angle of the seabed inclination was measured as the average inclination along each transect.

Temperature and salinity

Data on temperature and salinity for the study area were extracted from the hydrographic database assembled at the Bedford Institute of Oceanography, and Fisheries and Oceans Canada (Petrie et al. 1996, also accessible at http://www.mar.dfo-mpo.gc.ca/science/ocean/home.html). The data from 1933 to the present represent 862 unevenly distributed records (single measurements or vertical profiles) from different times of the year. The data set did not have sufficient resolution in time and space to enable detailed maps of the distribution of near-bottom temperature and salinity. To gain information on these variables for the video transects, all temperature and salinity records from the study area were plotted against depth. The plot was used to determine minimum, average and maximum values for temperature and salinity for the depths of each of the video transects.

Statistical methods

The patchiness of the coral distribution along the video transects was determined using Morisita’s index of dispersion [Id (Morisita 1962)]. Random spatial distribution is indicated when the value of this index equals one, while the distribution tends to be continuous as the value becomes greater than one. The null hypothesis of randomness was tested by a χ2-test at a 5% significance level. Classification of sites was based on the Bray-Curtis measure (Bray and Curtis 1957) using group-average sorting. By this method, transect groups join at the average level of similarity between all members of one group and all members of the other group. Linear correlation and stepwise regression analysis were used to relate coral abundance to environmental variables, i.e. bottom type, depth, topography, and hydrography.

Results

Geographical distribution and abundance of corals

In total, 3,197 coral colonies were recorded from the 52 video transects, belonging to three species, Primnoa resedaeformis, Paragorgia arborea and A. armata. Primnoa was the most abundant and widely distributed gorgonian in the area (Fig. 2). It occurred on 67% of the transects and represented 83% of the total number of colonies observed (Table 1). Paragorgia and Acanthogorgia contributed 10% and 7% to the total number of colonies and were observed on 40% and 2% of the inspected transects, respectively. On average, for all transects where any gorgonian was observed, Primnoa occurred with 4.8 colonies per 100 m2, Paragorgia with 0.6 colonies per 100 m2 and Acanthogorgia with 0.5 colonies per 100 m2. The highest average abundance of Primnoa for whole transects was observed on transect 637 with 19.2 colonies per 100 m2 and for Paragorgia on transect 118, with 6.2 colonies per 100 m2 (Table 1). On transect 6, a total of 203 colonies of A. armata was observed with a maximum abundance of 199.1 colonies per 100 m2 and an average of 17.6 colonies per 100 m2.
Fig. 2A–C

Densities of coral colonies (colonies per 100 m2) in the Northeast Channel. A Paragorgia arborea. B Primnoa resedaeformis. C Acanthogorgia armata

Primnoa and Paragorgia were more common along the shelf break and slope (>250 m) than on the shelf (<250 m). Figure 3 shows the abundance of Paragorgia and Primnoa along the shelf break. Classification analysis of the transects based on colony abundance identified three areas with high abundance of Primnoa and Paragorgia in the three valleys (Fig. 4). In the southern valley, Primnoa was observed on 66% of the transects, whereas Paragorgia was present on only 22%. The outer parts of the middle valley had the widest distribution and highest local abundance of Primnoa and Paragorgia; which were present on 81% and 50% of the transects, respectively. The highest abundance of Primnoa and Paragorgia, estimated for video sequences, was 121 colonies per 100 m2 and 49 colonies per 100 m2, respectively (transects 119 and 118). In the northern valley, the highest abundance occurred on the northern side of Romeys Peak at transects 636–639 and 642 (Table 1). No corals were observed on the northern side of the northern valley.
Fig. 3A, B

Average abundance of coral colonies (colonies per 100 m2) along the shelf break going from south to north. Maximum values are indicated with vertical lines. A Primnoa resedaeformis. B Paragorgia arborea. Only transects closer than 5 km from the 300-m depth contour line were included in the graph

Fig. 4

Seven groups of video transects identified with cluster analysis based on environmental information. Encircled transects have high abundance of Paragorgia arborea and/or Primnoa resedaeformis

Small-scale distribution and co-occurrence of corals

All three gorgonians showed a patchy distribution along the transects. This was demonstrated by Morisita index values above one (Table 1). Patches were identified at different spatial scales. In video sequences where Primnoa occurred in high abundance, distinct patches were confined to single boulders. At a larger scale, these small patches were not evenly distributed, but occurred within distances of 10–160 m. Paragorgia was rarely observed with more than one colony per boulder. It occurred in patches between 10 m and 50 m (average of 18 m) along the transects. The corals were generally higher in dense patches than in sparse patches. There was a relatively strong correlation between the abundance of Primnoa and Paragorgia, with r values of 0.50 (P<0.001, n=1,751) and 0.51 (P<0.05, n=52) for video sequences and video transects respectively (Tables 2, 3). The abundance of Acanthogorgia was not correlated with the abundance of the two other gorgonians. Paragorgia was never observed on any transects without Primnoa and they frequently occurred together on the same boulder. Primnoa was observed alone on 13 transects, and it occurred alone with Acanthogorgia on one transect. All three species occurred together on three transects.
Table 2

Correlation (r values) for the relationship between abundance (colonies per 100 m2) of the three gorgonians (Primnoa resedaeformis, Paragorgia arborea and Acanthogorgia armata) and environmental variables estimated for video sequences (n=1,751)

Paragorgia

Primnoa

Acanthogorgia

Depth (m)

Sand

Pebble

Cobble

Boulder

Cobble and boulder

Paragorgia

1.00

Primnoa

0.50*

1.00

Acanthogorgia

−0.02

−0.02

1.00

Depth (m)

0.33*

0.32*

−0.08

1.00

Sand

−0.08

−0.13*

−0.03

−0.07

1.00

Pebble

−0.05

−0.11*

−0.02

0.02

−0.56*

1.00

Cobble

0.14*

0.24*

0.00

0.07

−0.65*

−0.05

1.00

Boulder

0.10

0.17*

0.11*

<0.01

−0.42*

−0.21*

0.26*

1.00

Cobble and boulder

0.16*

0.26*

0.06

0.05

−0.69*

−0.15*

0.87*

0.71*

1.00

*P< 0.001

Table 3

Correlation (r values) for the relationship between abundance (colonies per 100 m2) of the three gorgonians (Primnoa resedaeformis, Paragorgia arborea and Acanthogorgia armata) and environmental variables estimated for video transects (n=52)

Paragorgia

Primnoa

Acanthogorgia

Tmeana

Tmaxa

Tmina

Tdiffa

Smeana

Smaxa

Smina

Sdiffa

Depth (m)

Topographical index

Angle

Sand

Pebble

Cobble

Boulder

Cobble and boulder

Paragorgia

1.00

Primnoa

0.51*

1.00

Acanthogorgia

−0.05

−0.04

1.00

Tmean

−0.34*

−0.47*

0.09

1.00

Tmax

−0.32

−0.43*

0.14

0.95*

1.00

Tmin

0.25

0.25

−0.17

−0.85*

−0.86*

1.00

Tdiff

−0.30

−0.39*

0.17

0.89*

0.98*

−0.83*

1.00

Smean

0.03

0.02

0.07

−0.34*

−0.31

0.30

−0.28

1.00

Smax

−0.30

−0.36*

0.02

0.94*

0.95*

−0.83*

0.92*

−0.49*

1.00

Smin

0.23

0.29

−0.08

−0.87*

−0.83*

0.84*

−0.77*

0.69*

−0.88*

1.00

Sdiff

−0.27

−0.31

0.01

0.89*

0.89*

−0.79*

0.85*

−0.66*

0.98*

−0.93*

1.00

Depth

0.36*

0.52*

−0.13

−0.98*

−0.97*

0.82*

−0.93*

0.23

−0.92*

0.78*

−0.84*

1.00

Topographical index

<0.01

0.11

−0.13

−0.12

−0.06

0.12

−0.02

0.07

−0.14

0.14

−0.14

0.07

1.00

Angle

0.10

0.15

−0.15

−0.35*

−0.29

0.34*

−0.24

0.04

−0.31

0.31

−0.28

0.30

0.66*

1.00

Sand

−0.18

−0.17

0.01

0.22

0.17

−0.19

0.14

−0.16

0.20

−0.25

0.21

−0.19

0.17

0.06

1.00

Pebble

0.08

−0.04

0.02

−0.24

−0.21

0.25

−0.19

0.16

−0.24

0.26

−0.24

0.20

−0.20

−0.02

−0.84*

1.00

Cobble

0.24

0.38*

−0.06

−0.13

−0.07

0.02

−0.04

0.09

−0.09

0.12

−0.10

0.11

−0.05

−0.05

−0.83*

0.44*

1.00

Boulder

0.15

0.12

0.06

−0.07

−0.04

0.13

−0.02

0.12

−0.08

0.16

−0.09

0.05

−0.15

−0.11

−0.66*

0.30

0.62*

1.00

Cobble and boulder

0.23

0.33*

−0.02

−0.12

−0.07

0.06

−0.04

0.11

−0.09

0.15

−0.10

0.10

−0.09

−0.08

−0.85*

0.43*

0.96*

0.81*

1.00

aT Temperature, S salinity

*P< 0.05

Bathymetrical distribution

The mean depth and depth range for each species showed that Primnoa has a shallower distribution than Paragorgia in the Northeast Channel. The shallowest observation of Primnoa was at 196 m at transect 113, while Paragorgia occurred deeper than 235 m. Acanthogorgia was observed at depths between 231 m and 364 m. The maximum abundance of Acanthogorgia on top of Romeys Peak was not related to high abundance of the other two species. There was a significant (P<0.05) correlation between the average abundance of Paragorgia and Primnoa and average depth for transects (Table 4). This correlation was stronger for Primnoa (r=0.53) than for Paragorgia (r=0.32).
Table 4

Environmental factors and coral abundance (colonies per 100 m2) for transect groups (1–7) identified by classification of 52 video transects based on environmental factors, Bray-Curtis similarity measure and group-average clustering

Group 1

CP2

Group 2

Group 3

Group 4

Group 5

Group 6

Group 7

Average

Range

Average

Range

Average

Range

Average

Range

Average

Range

Average

Range

Average

Range

Paragorgia arborea

0.04

0–0.5

0.00

0.09

0–0.3

0.00

0.00

0.23

0–1.0

0.99

0–6.2

0.17

0–0.7

0.87

0–3.3

Primnoa resedaeformis

1.15

0–4.3

0.39

0.77

0–1.6

0.00

0.00

4.85

0.03–14.6

2.13

0–4.7

0.25

0–1

10.27

0–37.6

Acanthogorgia armata

1.10

0–17.6

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.05

0–0.4

0.02

0–0.06

0.00

0.00

Middle depth (m)

240

203–261

245

252

242–260

190

184–200

309

298–327

367

354–378

346

318–364

441

390–480

Topographical index

1.57

0.1–5.3

0.07

0.72

0–1.2

0.94

0.5–1.4

1.19

0.2–0.3

0.96

0.2–2.3

3.88

0.1–0.8

1.34

0.1–0.4

Bottom angle

1.7

0.4–3.9

2.0

1.5

0.2–3.0

1.2

0.9–1.5

2.3

1.5–4.9

2.3

1.2–4.2

4.6

0.4–9.2

2.7

0.8–6.5

Percentage sand

50

35–71

10

90

86–95

81

69–78

36

10–63

36

21–58

90

78–96

52

36–71

Percentage pebbles

21

13–39

46

3

1–5

6

1–11

28

10–55

31

19–50

6

3–14

23

13–54

Percentage cobbles

21

8–36

37

4

1–7

9

1–14

24

15–32

23

12–42

3

0–6

19

2–42

Percentage boulders

8

1.8–19

8

3

0–4

4

0–6

11

6–18

10

6–16

1

0–3

6

0–14

Temperature

  Average

7.3

7.1–7.7

7.1

7.2

7.2–7.3

8.2

7.9–8.6

6.4

6.1–6.5

6.1

6.0–6.2

6.2

5.6–8.2

5.5

5.3–5.9

  Max.

10.7

9.5–12.1

11.2

10.1

9.5–10.7

13.0

12.1–13.9

9.2

8.6–9.7

7.2

6.7–7.9

8.0

7.3–9.0

6.0

5.6–6.5

  Min.

3.7

3.4–4.3

3.6

3.9

3.4–4.3

3.2

2.9–3.4

4.9

4.8–5.1

5.2

4.8–5.7

5.2

5.1–5.4

5.1

4.5–5.6

  Diff.

3.4

2.3–4.7

4.0

2.9

2.3–3.4

4.8

4.2–5.2

2.8

2.3–3.2

1.0

0.6–1.8

1.8

1.1–2.7

0.5

0.3–1.1

Salinity

  Average

34.99

34.9–35.0

34.99

35.00

35.00

34.89

34.9

34.99

35.50

34.98

35.00

34.99

35.00

34.98

35.00

  Max.

35.22

35.2–35.3

35.21

35.19

35.20

35.40

35.3–35.5

35.16

35.1–35.2

35.06

35.10

35.09

35.10

35.03

35.00

  Min.

34.40

33.8–34.7

34.49

34.54

34.3–34.7

33.54

33.1–33.8

34.86

34.90

34.87

34.90

34.87

34.90

34.93

34.90

  Diff.

0.23

0.2–0.4

0.21

0.19

0.20

0.51

0.4–0.7

0.17

0.1–0.2

0.08

0.10

0.10

0.1–0.2

0.05

0.0–0.1

Based on abundance of colonies within 20-m depth intervals, Paragorgia and Primnoa occurred with two major peaks at different depths (Fig. 5). Highest abundance for both species was found between 410 m and 490 m depth. A shallower peak in abundance occurred between 300 m and 365 m depth. The maximum abundance of Paragorgia (within 20-m depth intervals) was 3.4 colonies per 100 m2 between 340 m and 360 m depth in the middle valley and 3.3 colonies per 100 m2 between 480 m and 500 m depth in the northern valley. Primnoa occurred with a maximum abundance of 26.5 colonies per 100 m2 between 440 m and 460 m depth in the northern valley. At shallower depths the abundance was always below 5 colonies per 100 m2. The maximum abundance for Acanthogorgia at transect 6 was at depths between 220 m and 237 m on the western side of Romeys Peak in the middle valley. However, on the northern side of Romeys peak it occurred down to around 360 m in low abundance. Paragorgia occurred at shallower depths in the inward extension of the local valleys than on the peaks between. The shallowest occurrence of Paragorgia on the peaks was 319 m on Romeys Peak and 383 m at the Rips. In the inner parts of the valleys, the shallowest occurrence of Paragorgia was at 248 m in the northern valley, 256 m in the middle valley, and 267 m in the southern valley. This topographically related difference was not observed for Primnoa, which occurred on all of the shallow transects on the peaks.
Fig. 5A, B

Average abundance of coral colonies (colonies per 100 m2) versus depth of transects in three divisions of the Northeast Channel. A Primnoa resedaeformis. B Paragorgia arborea. Error bars indicate standard deviations. The number of investigated video sequences for the each depth interval in the different divisions is given in Table 7

Table 7

Number of video sequences within 20-m depth intervals for Paragorgia arborea, Primnoa resedaeformis and Acanthogorgia armata

Depth interval (m)

Northern Valley

Middle Valley

Southern Valley

180–200

0

0

74

200–220

16

0

59

220–240

56

94

1

240–260

176

154

10

260–280

19

27

26

280–300

39

0

0

300–320

161

60

0

320–340

22

24

13

340–360

63

45

7

360–380

99

52

28

380–400

49

54

7

400–420

1

90

0

420–440

0

30

0

440–460

5

11

0

460–480

151

0

0

480–500

28

0

0

Sum

885

641

225

Coral abundance in relation to environmental factors

Transects with high abundance of corals were characterised by great depth (400–500 m), low maximum temperatures, and high percentage of cobbles and boulders (Table 5). Depth was strongly correlated with abundance of coral colonies for both Primnoa and Paragorgia (Tables 2, 3). This correlation was strongest for video transects as a sample unit. However, depth was also strongly correlated with temperature (R2=0.95) and cannot be considered to be an environmental factor per se.
Table 5

Environmental characteristics for transects, grouped according to the abundance (colonies per 100 m2) of Paragorgia arborea and Primnoa resedaeformis, using the same abundance intervals as in Fig. 2. Standard deviations are given in parentheses

Abundance

n

Salinity

Temperature

Depth

Topo ind

Angle

Bottom typesa

Mean

Max.

Min.

Mean

Max.

Min.

Sa

Pe

Co

Bo

Paragorgia

  0

32

34.97

35.27

33.95

7.35

10.16

5.28

278

1.47

2

59

19

16

6

(0.19)

(0.18)

(1.18)

(0.98)

(1.91)

(0.44)

(75)

(1.66)

(1.8)

(25)

(14)

(10)

(4)

  <0.15

7

34.86

35.14

34.15

6.67

8.39

5.21

332

1.53

2.5

37

29

25

10

(0.37)

(0.07)

(1.3)

(0.5)

(1.13)

(0.31)

(53)

(1.78)

(1.1)

(12)

(10)

(8)

(3)

  0.15–1

8

34.96

35.12

34.64

6.65

8.68

5.11

343

1.22

2

54

19

17

9

(0.05)

(0.1)

(0.55)

(0.9)

(2.29)

(0.15)

(72)

(1.08)

(1.4)

(19)

(16)

(10)

(6)

  >1

5

35.01

35.1

34.35

6.25

7.72

5.27

411

1.46

3.4

40

24

26

10

(0.02)

(0.04)

(0.83)

(0.45)

(0.64)

(0.18)

(74)

(1.3)

(1.9)

(15)

(9)

(12)

(6)

Primnoa

  0

17

34.99

35.27

33.73

7.27

10.12

5.26

276

2.19

2.4

68

16

12

3

(0.09)

(0.21)

(1.36)

(1.15)

(2.08)

(0.45)

(77)

(1.95)

(2.2)

(22)

(12)

(9)

(4)

  <1

12

35

35.22

34.22

7.19

9.43

5.14

293

0.98

1.8

44

25

22

9

(0.06)

(0.15)

(1.01)

(0.76)

(1.8)

(0.43)

(68)

(0.77)

(0.9)

(23)

(15)

(10)

(4)

  1–5

14

34.84

35.17

34.66

6.97

9.19

5.28

308

0.93

2.1

51

23

18

9

(0.34)

(0.14)

(0.56)

(0.96)

(2.28)

(0.31)

(73)

(1.42)

(1.3)

(21)

(14)

(10)

(5)

  >5

8

35.02

35.18

33.35

6.62

8.91

5.29

393

1.53

2.7

44

20

26

10

(0.03)

(0.08)

(1.23)

(0.64)

(1.18)

(0.18)

(87)

(1.13)

(2)

(12)

(8)

(9)

(5)

aSa Sand, P pebble, C cobble, B boulder

Cluster analysis of the video transects based on environmental variables (percentage cover of bottom types, topography, seabed angle, depth, temperature and salinity) identified seven habitat groups. These groups are indicated on the map in Fig. 4, and the environmental characteristics and coral abundance for the groups are given in Table 4. Three groups (group 2, 3 and 6) had no or little corals. The common feature for these groups was high percentage cover of sand. The cluster groups 4, 5, and 7 had the high abundance of Primnoa and Paragorgia. The transects in these groups were all situated at depth (>298 m), and in cold water (with an average maximum temperature<9.2°C). The bottom was steep, dominated by pebble, cobble and boulder. Group 1, which had the highest abundance of Acanthogorgia, had transects with relatively high percentage cover of pebble and cobble, and relatively warm water (average maximum temperature 10.7°C). Transect 2 was not classified with any other transects. It is situated on steep bottom at a depth of ca 245 m on top of Romeys Peak. The bottom is dominated by pebble, cobble and boulder. The water is relatively warm with a maximum temperature of 11.2°C. Primnoa occurred with intermediate abundance. This transect was unusual by having the combination of shallow depth and relative high percentage cover of cobble and boulder.

Multiple stepwise linear regression analysis of the relationship between abundance of colonies and environmental factors was performed for Primnoa and Paragorgia (Table 6). Together, temperature, percentage cobble, and salinity accounted for 38% of the variance of Primnoa. The comparable figure for Paragorgia was 15%. By including depth as a factor, additionally 3% and 8% of the variance is explained for Paragorgia and Primnoa, respectively. Figure 6 shows a three-dimensional plot of the relation between abundance of coral colonies, mean temperature, and percentage cover of cobble for video transects.
Table 6

Multiple stepwise regression analysis of the relationship between abundance of Primnoa resedaeformis and Paragorgia arborea and environmental variables. Only variables with significant correlation to abundance were used in the analysis

Environmental factor

Paragorgiaa

Primnoab

Correlation r

MR with depth

MR without depth

Correlation r

MR with depth

MR without depth

Mean depth

0.36**

0.36

0.52**

0.52**

Average temp.

−0.34**

0.01

0,34**

−0.47**

0.03

−0.47**

Percentage cobble

0.24*

0.05

0.05

0.38**

0.10*

0.1*

Max. salinity

−0.30*

0.00

0.01

−0.36**

0.08*

0.07*

Multiple r

0.43

0.39

0.68

0.62

Adjusted R2

0.18

0.15

0.46

0.38

aParagorgia abundance=3.26−(0.44×Tmean)

bPrimnoa abundance=−1,471−(8.86×Tmean)+(0.8×percentage cobble)+(43.5×Smax)

*P<0.1; **P<0.05

Fig. 6A, B

Three-dimensional plots of abundance of coral colonies (colonies per 100 m2) versus mean temperature, and percentage cover of cobble. A Primnoa resedaeformis. B Paragorgia arborea

Seabed substratum and topography

Cobble and boulder were patchy distributed along all transects except one (102, Table 1). The average percent cover of cobble and boulder for the whole study area was 18% and 7%, respectively. In general, the corals were restricted to areas with boulders and/or cobbles. However, the abundance of colonies was not strongly correlated with the percentage cover of stones in any of the three size classes (pebble, cobble and boulder) (Tables 2, 3). The strongest correlation for all transects was found between the abundance of Primnoa and percent cover of cobble (r=0.38, P< 0.05). Uncolonised cobbles and boulders were far more numerous than those colonised by gorgonians.

Large Paragorgia colonies were observed almost exclusively on boulders, whereas smaller colonies (<25 cm) were observed on cobbles as well. At one location a large Paragorgia colony was found attached to cobbles. This colony had its holdfast spread out over several cobbles. Primnoa occurred on both cobbles and boulders while Acanthogorgia was found only on cobbles. The abundance of Acanthogorgia, however, was not correlated with the percentage cover of cobbles. No corals were observed attached to pebbles.

Judging from the map, the distribution and abundance of corals was positively related to large-scale topographic features such as the shelf break and ridges. On the scale of video transects the abundance of coral was not significantly correlated with the angle of the seabed or the roughness of the seabed estimated from the profiles along the video transects. On an even finer scale, the roughness of the seabed is correlated with the percent cover of cobble and boulder.

Temperature and salinity

The abundance of Primnoa and Paragorgia was correlated with estimated near-bottom temperature and salinity for each site. This was done with hydrographic data pooled for the whole area, as well as for sub-areas (inner vs. outer, and inner and outer parts of the three valleys). The strongest correlation was found when applying data pooled for the whole study area. The abundance of Primnoa and Paragorgia was significantly and negatively correlated with both average and maximum bottom temperatures estimated from the general temperature curve for the investigation area (Table 3). The highest abundance of colonies for Primnoa and Paragorgia was found along transects with average temperatures between 5.3°C and 6.5°C (Fig. 7). Primnoa occurred over a wider temperature range than Paragorgia and was present on transects with a maximum temperature up to 12.1°C. The highest maximum temperature for transects where Paragorgia occurred was 9.7°C.
Fig. 7 A, B

Colony abundance (colonies per 100 m2) versus average and maximum temperature. A Primnoa resedaeformis. B Paragorgia arborea

Discussion

Abundance of deep-water gorgonians in the Northeast Channel compared with other areas

There are more studies of gorgonian distribution and abundance in temperate and tropical shallow waters than in deep-water. Hecker et al. (1980) provided an early quantitative study of the distribution of corals and megafauna, from photographs and submersible observations in three canyons off the northeast coast of USA. Averaged for 100-m depth intervals they estimated a highest abundance for Acanthogorgia of 3.2 colonies per 100 m2 and for Paragorgia 0.25 colonies per 100 m2 between 400 m and 800 m depth in Oceanographers Canyon. They do not give information about the abundance of Primnoa, but this species seem to be less common within their study area. Our estimated average abundance (estimated for 20-m depth intervals) (3.3 colonies per 100 m2 for Paragorgia and 12.1 colonies per 100 m2 for Acanthogorgia) are about a factor of ten higher than values calculated by Hecker et al. (1980). Methodological differences could account for the discrepancy, but the Northeast Channel may also be a unique high-abundance area for gorgonian corals in the northwestern Atlantic Ocean. Mortensen et al. (1995) studied the megafauna on deep-water coral reefs (Lophelia pertusa) off mid-Norway, and found Primnoa and Paragorgia with average abundance of 6.9 colonies per 100 m2 and 4.3 colonies per 100 m2, respectively.

The gorgonian habitat in the Northeast Channel has higher abundance of Primnoa resedaeformis and Paragorgia arborea than is observed in other areas. However, the abundance is clearly lower than for shallow-water gorgonian populations. Temperate and tropical shallow-water gorgonians generally occur with higher abundance than that indicated for deep-water gorgonians (Grigg 1977; Mistri 1995). This may be due to different environmental conditions and food supply. There also seems to be a general negative relationship between size of colonies and abundance, where smaller gorgonians in general, have higher abundance than larger species. This was also demonstrated in the Northeast Channel where the smaller species (Acanthogorgia) occurred with the highest local abundance, whereas the lowest local maximum was found for the largest species (Paragorgia).

Distribution and environmental controls

In general, the distribution of deep-water gorgonians is uneven and controlled by environmental factors. Of physical factors, suitable seabed substratum, water temperature and salinity are obvious requirements that must be met to enable the colonisation and growth of corals. We found that the distribution of deep-water gorgonians in the Northeast Channel was patchy even within areas with high percentage of cobble and boulder, and within the depth range where the highest abundance was found. This indicated that other factors varying over small spatial scales are also important. Competition for space does not seem to be an important factor for the distribution of either shallow-water tropical gorgonians (Yoshioka and Yoshioka 1989), or the deep-water gorgonians in this study. Uncolonised boulders were common in areas with corals while at the same time corals co-occurred on boulders. No clear signs of competition between the corals (such as absence of other gorgonians in the vicinity of a colony, or malformation of co-occurring colonies of different species) were observed in this study. Food supply is an important biological factor, which may control the distribution of sessile invertebrates at both large and small scales. Unfortunately, the food requirement of the deep-water gorgonians is generally unknown. The type and concentration of food available for the corals is in turn controlled by other factors such as current patterns and hydrography, which are discussed below.

Substratum and topography

With the exception of a few species having anchorage structures for soft sediments, deep-water gorgonians are found on hard substrata such as large stones or bedrock. Even though sandy sediment was the dominant bottom type, availability of a hard bottom substratum did not seem to be a limiting factor for the distribution of the gorgonians. Even a percentage cover as low as 9% of boulder and cobbles combined for whole transects was sufficient to support an abundance of Paragorgia higher than average (Table 1). In addition, the corals’ requirements with respect to the micro-environment for settlement is not known. These conditions probably explain the relatively poor correlation between coral abundance and percentage cover of hard bottom.

No clear relationship between coral abundance and topography at an intermediate scale (tens of metres) was observed in the Northeast Channel. However, the concentration of corals close to the shelf break and on sides of ridges, such as Romeys Peak, indicate that large scale topographic structures play an important role generating a favourable environment for corals. Topographic influence on current patterns and particle concentration may determine the distribution of deep-water corals locally (Genin et al. 1986; Mortensen et al. 2001). The occurrence of deep-water corals on topographic highs or close to edges seem to be a general feature. The most probable explanation for this is that topographic highs or seabed relief increase the encounter rate of food particles, due to increased current velocity. Currents may also be created around these features that concentrate food particles, owing to formation of eddies and consequently retention of particles. On wide peaks, higher abundance is found along the edge rather than in the centre of the top, while narrow peaks also support corals at their top (Genin et al. 1986). Genin et al. (1986) suggest that the acceleration of flow over the top of seamounts leads to upwelling of deeper water. More studies on the local current patterns in coral habitats are needed to explain the possible relationship between the small-scale hydrodynamics and the availability of food and larvae.

Depth and hydrography

Many factors that co-vary with depth can be expected to determine vertical distribution patterns. In the Northeast Atlantic different regional maximum depths of deep-water corals in offshore areas generally reflect different maximum depths of water masses with suitable temperatures (Frederiksen et al. 1992; Tendal 1992; Freiwald 1998; Mortensen et al. 2001).

Our observation of A. armata at 231 m is to our knowledge the shallowest occurrence of this species. It occurs at greater depths on the shelf off the USA (400–1,300 m) (Hecker et al. 1980) than those that have been recorded previously off Nova Scotia (250–600m) (Deichmann 1936; Breeze and Davis 1998). Primnoa has a wide depth distribution within the investigated area while Paragorgia mainly occurs in the deeper parts. Primnoa is reported to occur below 100 m depth off Nova Scotia (Breeze et al. 1997). Paragorgia has an upper depth limit around 300 m off the northeast coast of USA (Hecker et al. 1980), which is quite similar to our findings. Unfortunately, our investigation was limited to depths above 500 m, which prevents discussion of the local maximum depths of corals in the Northeast Channel. There are few attempts to study corals below this depth off Nova Scotia. These studies, using dredges or trawls, have found Paragorgia and Primnoa as deep as 1,097 m and 457 m, respectively (Verrill 1922; Deichman 1936; Breeze et al. 1997). On the northeast coast of the USA, Hecker et al. (1980) found Paragorgia down to 800 m depth, and Primnoa down to 560 m. This indicates that peak abundance for these corals may occur more deeply than indicated by our study.

Tendal (1992) describes the distribution of Paragorgia in the North Atlantic, and suggests that it is connected to the North Atlantic Current, characterised by temperatures generally between 4°C and 8°C, and stable salinity around 35‰. Madsen (1944) regards Paragorgia and Primnoa as being extremely stenotherm with temperature requirements between 5°C and 8°C. Similarly to the reports by Madsen (1944), we found a shallower upper limit for Primnoa (196 m) than for Paragorgia (235 m), suggesting that these species have different environmental requirements. We suggest that the maximum temperature for Primnoa is about 2°C higher than for Paragorgia.

Paragorgia occurred more shallowly in the inward extension of the local valleys than out on the peaks. This is best explained by different depth ranges of water masses following the general topography of the seabed in the study area. In the outer parts of the channel, the temperature ranges between 7.8°C and 10.3°C at depths between 200 m and 300 m (Petrie and Dean-Moore 1996). At the same depths in the inner parts, the temperature varies between 5.7°C and 8.4°C. Based on the suggestions by Madsen (1944) and Tendal (1992), this indicates that high temperatures may prevent the corals from establishing at shallow depths in the outer parts of the study area. The salinity is above 34.3‰ below 200 m depth all over the study area, and should thus not be a restricting factor.

The absence of corals on the northern side of the northern valley is partly explained by the hydrography and currents. The main current here is directed inwards (Loder et al. 2001), and the salinity, temperature and current velocity are variable and related to weather conditions. Both mean and maximum temperatures are higher in this area than on the southern side. At 200 m depth the maximum temperature is around 9.6°C on the northern side compared to 8.1°C on the southwestern side. The range is around 2.5°C on the northern side compared to 1.5°C on the southern side (Ramp et al. 1985). These differences may also relate to differences in food supply.

Acknowledgements

This study was funded by the Environmental Studies Research Fund (ESRF), and the Department of Fisheries and Oceans (DFO) of Canada. We thank Don Gordon and Anna Metaxas for organising the cruises. The ROPOS team, and the crew onboard C.C.G.S. Hudson and C.C.G.S. Martha Black were very helpful in arranging for the sampling and video recording at sea. Thanks to Dave M. McKeown for invaluable help in processing the navigation data for the Campod. We thank Ken Paul of the Canadian Hydrographic Service for providing the multi-beam bathymetric data of the Northeast Channel and Stan Johnson of the Oceans and Coastal Management Division for producing the general bathymetric map. Thanks to Don Gordon, Barry Hargrave and Shelley Armsworthy for help with the manuscript.

Copyright information

© Springer-Verlag 2004