Habitat heterogeneity influences cold-seep macrofaunal communities within and among seeps along the Norwegian margin – Part 2: contribution of chemosynthesis and nutritional patterns

ORIGINAL ARTICLE

Habitat heterogeneity influences cold-seep macrofaunalcommunities within and among seeps along the Norwegianmargin – Part 2: contribution of chemosynthesis andnutritional patternsCarole Decker & Karine Olu

Departement Etude des Ecosystemes Profonds, IFREMER, Plouzane, France

Keywords

Cold-seep; macrofaunal nutrition; methane-

derived carbon; Norwegian margin; stable

isotope analysis.

Correspondence

Carole Decker, Departement Etude des

Ecosystemes Profonds, IFREMER, F-29280

Plouzane, France.

E-mail: [email protected]

Accepted: 31 July 2011

doi:10.1111/j.1439-0485.2011.00486.x

Abstract

The relative contribution of chemosynthesis in heterotrophic fauna at seeps is

known to be influenced by depth and by habitat. Using stable isotopes of car-

bon and nitrogen, we investigated macro- and megafaunal nutritional patterns

in Norwegian margin cold seeps by comparing food webs both among habitats

within a seep site and between different sites. The very active Hakon Mosby

mud volcano (HMMV) is characterized by geochemical gradients, microbial

activity and faunal zonation from the centre to the periphery. The Storegga

Slide (600–900 m depth) has pockmarks with patchy less active seeps, and also

shows concentric zonation of habitats but at much smaller spatial scale. The

dominant carbon source for macrofaunal nutrition in both areas was chemo-

synthetically fixed and the bulk of organic carbon was derived from sulphur-

oxidizing bacteria. In HMMV, food chains were clearly separated according to

habitats, with significantly lighter d13C signatures on microbial mats and adja-

cent sediment ()33.06 to )50.62&) than in siboglinid fields ()19.83 to

)35.03&). Mixing model outputs revealed that the contribution of methane-

derived carbon was small in siboglinid fields (0–17%) but significant (39–61%)

in the microbial mats. Moreover, the variability of macrofauna signatures

within this later habitat suggests the co-occurrence of two food chains, one

based on primary production via methanotrophy and the other via sulphide

oxidation. The length of the food chains also varied among habitats, with at

least one more trophic level in the siboglinid fields located at the periphery of

the volcano. Conversely, in Storrega pockmarks, faunal d13C signatures did not

vary among habitats but among species, although separate food chains seem to

co-occur. The small size of the seepage areas and their lower fluxes compared

to HMMV allow more background species to penetrate the seep area, increas-

ing the range of d15N and the trophic level number. Probably due to the higher

flux of photosynthetic particulate organic carbon, the overall chemosynthesis-

based carbon contribution in invertebrate nutrition was lower than that in

HMMV.

Marine Ecology. ISSN 0173-9565

Marine Ecology 33 (2012) 231–245 ª 2011 Blackwell Verlag GmbH 231

Introduction

Cold-seep ecosystems are found along the wide bathymet-

ric range of continental margins, in relation with release

of methane-enriched fluid or gas, often driven by tectonic

processes. In these ecosystems, methane and sulphide are

the two energy sources of the predominantly chemosyn-

thesis-based production. Nevertheless, cold-seep ecosys-

tems also receive variable input of organic matter from

surface waters, according to depth and dynamic context.

The importance of chemosynthetic nutritional pathways

versus photosynthesis has been related to depth by com-

paring seeps along the East Pacific (Levin et al. 2000;

Levin & Michener 2002) and along the Sakhalin margin

(Sahling et al. 2003), revealing an increase of chemosyn-

thesis contribution in invertebrate diets at the deepest

sites, where external inputs are lower than on continental

shelf and slopes. From stable isotopes analysis of animal

tissues and source signatures, these authors estimated a

methane-derived carbon contribution of 0–20% for the

shallowest California slope seeps (520 m depth) to 20–

50% in the deepest seeps along Alaska or Florida margins

(>3000 m). At 2000 m in the oligotrophic Eastern Medi-

terranean, invertebrates associated with methane seeps on

mud volcanoes are fuelled virtually exclusively by carbon

of chemosynthetic origin, with percentages of methane-

derived carbon varying among species from 0–33% (lower

estimates) to 21–100% (upper estimates) (Carlier et al.

2010). Finally, Thurber et al. (2010) reported a contribu-

tion 21–73% methane-derived carbon in the New Zealand

seep fauna diet. Taking into account the whole commu-

nity, chemosynthesis contribution is generally high in

seep communities dominated in biomass by chemoauto-

trophic symbiont-bearing invertebrates (siboglinid poly-

chaete tubeworms, mytilids, vesicomyid bivalves). Stable

isotope signatures of non-symbiotic endemic species or

opportunists also revealed a high use of chemosynthesis-

based carbon, relying directly or indirectly on free micro-

bial communities (methane-oxidizing archae or sulphur-

oxidizing filamenteous bacteria). These taxa include

dovilleid, polynoid, capitellid and ampharetid polychaetes,

several grazers (gastropods, alvinocarid shrimps) and

deposit feeders (sipunculids, echinoids, holothurids)

(Levin et al. 2000; Van Dover et al. 2003; MacAvoy et al.

2005; Olu et al. 2009; Carlier et al. 2010). Higher level

predators resident in seeps, such as a few fish, large crus-

taceans, sea-stars and buccinid gastropods, may rely

almost 100% on seep production (MacAvoy et al. 2002).

Large vagrant predators are abundant in bathyal depths

and can also export chemosynthesis-based material from

seeps (MacAvoy et al. 2002). A few fish also seem to

export chemosynthesis material at deeper sites, with up to

38% of seep material in their diet (Olu et al. 2009). But

there are also opportunist vagrant taxa from the back-

ground community at seeps that only complement their

diet with the seep production. The mean contribution of

seep production is therefore balanced between, on the

one hand, species highly dependent on chemosynthesis

and, on the other, opportunists and vagrant background

fauna benefitting more or less from local production.

The proportion of seep-dependent taxa may also vary

according to habitats and is generally related to the inten-

sity of the fluid flow. Biogenic structures created by large-

size symbiont-bearing species (Sibuet & Olu-Le Roy

2002), along with microbial mats, generate diverse habi-

tats at seeps (Levin 2005), where small heterotrophic

fauna live. Levin & Michener (2002) and Levin & Men-

doza (2007) evidenced differences in d13C signatures and

methane-derived carbon contribution among habitats

(microbial mats, vesicomyid fields, siboglinid fields)

regardless of the depth of the seeps.

The Hakon Mosby mud volcano (HMMV), a structure

2 km in diameter, is affected by high fluid, gas and mud

expulsion and characterized by concentric distribution of

habitats defined visually by the occurrence of microbial

mats, siboglinid fields, or bare sediment, and by geo-

chemical gradients, which also induce heterogeneity in

microbial communities and processes (De Beer et al.

2006; Niemann et al. 2006; Jerosch et al. 2007). This site,

located at mid-slope (1200 m) along the Norwegian

margin, is therefore a quite interesting one to test the

hypothesis of a decrease of the relative importance of che-

mosynthesis material in the nutrition of associated fauna

along the gradient of biogenic habitats. Also noticeable is

the presence of thick microbial mats of filamentous sul-

phur-oxidizing bacteria and the occurrence of methano-

trophic bacteria communities (Losekann et al. 2007).

Other cold-seeps have been recently discovered along the

Norwegian margin on the Storrega slide between 600 and

900 m depth (J.P. Foucher, C. Pierre, C. Decker, L.R.

Rufine, S. Ker, G. Westbrook, K. Olu-Le Roy, J.P. Donval

and J.L. Charlou, unpublished observations). Patchy seep-

age areas are characterized by moderate fluid flow but

also revealed habitat zonation (microbial mats, siboglinid

fields) around the fluid source, at the decimetre to metre

scale, colonized by fauna taxonomically related to HMMV

fauna (Decker et al. 2011). Comparison of the spatially

structured mud volcano with the communities associated

with small-size seeps isolated within pockmarks of low

activity, likely receiving higher detrital particulate fluxes,

but characterized by habitat zonation similar to that

observed on the mud volcano but is quite interesting in

this context. In this paper we compare isotopic signatures

(d13C and d15N) of invertebrates sampled by ROV in

several habitats of the HMMV and of pockmarks of the

Storrega slide (including Nyegga area) to assess the role

Nutritional patterns at Norwegian cold seeps Decker & Olu

232 Marine Ecology 33 (2012) 231–245 ª 2011 Blackwell Verlag GmbH

of chemosynthetic versus photosynthetic food sources in

the nutrition of benthic communities (d13C) and to unra-

vel trophic relationships among their faunal components

(d13C and d15N). This multi-scale study aims to compare

these patterns both among habitats within a seep site and

between seeps, and could help to assess habitat versus

depth, seep size or activity influence on macro- and me-

gafaunal nutrition at seeps.

The following questions were addressed:l Do habitats influence macro ⁄ megafaunal nutritional

patterns, i.e. number and length of food chains and origin

of carbon, along the geochemical gradient of the Hakon

Mosby mud volcano?l Do habitats also influence these characteristics at smal-

ler and shallower seepage areas in pockmarks along the

Norwegian margin?

Study area

Hakon Mosby mud volcano

The Hakon Mosby mud volcano (HMMV 72� N,

14�44¢ E) is located at 1250 m depth along the Norwegian

margin (Vogt et al. 1997). The circular mud volcano

comprises: (i) a flat central and southern part interpreted

as an area of recent mud flows – this flat area was called

a crater by Milkov et al. (2004); (ii) a surrounding area

characterized by a hummocky sea floor interpreted to be

composed of deformed old mud flows; and (iii) a

pronounced moat at the periphery of the mud volcano

(Foucher et al. 2010). The mud volcano is characterized

by exceptionally high activity of fluid and gas ejections

from the sea bed with thermal and fluid flow gradients

from the centre to the periphery (Milkov et al. 2004;

Sauter et al. 2006; Feseker et al. 2008), and zonation of

microbial activity and sediment geochemistry (De Beer

et al. 2006; Niemann et al. 2006). Previous studies on

HMMV revealed that chemosynthetic habitats are distrib-

uted according to these activity gradients, with a central

zone apparently devoid of fauna, an intermediate zone

covered by microbial mats, and a peripheral zone colo-

nized by two species of symbiont-bearing siboglinid poly-

chaetes Sclerolinum contortum Smirnov, 2000 (Monilifera)

and Oligobrachia haakonmosbiensis Smirnov, 2000 (Frenu-

lata) (Gebruk et al. 2003; Jerosch et al. 2007). The meio-

faunal and macrofaunal communities associated with

these habitats are clearly structured by habitat, with a few

taxa in the centre of the volcano and a more diversified

fauna in the peripheral siboglinid fields (Gebruk et al.

2003; Soltwedel et al. 2005; Van Gaever et al. 2006;

Decker et al. 2011).

Storegga Slide

The Storegga Slide is a giant submarine landslide located

in the southern part of the Norwegian margin (64�45¢ N,

04�59¢ E) where some pockmarks have been discovered

between 600 and 900 m depth (Hustoft et al. 2007; Paull

et al. 2008; Foucher et al. 2009) (Fig. 1). Few of these are

active at present (Foucher et al. 2009) and seepage is

dominated by advection of methane in solution in

porewater, which could be derived from dissolution of

hydrate in the chimney (J.P. Foucher, C. Pierre, C.

Decker, L.R. Rufine, S. Ker, G. Westbrook, K. Olu-Le

Fig. 1. Location of the three sites sampled:

Hakon Mosby mud volcano and the Storegga

Storegga and Nyegga (G11) pockmarks.

Decker & Olu Nutritional patterns at Norwegian cold seeps


Roy, J.P. Donval and J.L. Charlou, unpublished observa-

tions). Chemosynthetic communities, similar in composi-

tion to those colonizing HMMV, have developed in

small areas, sparsely distributed around fluid sources

(Decker et al. 2011), but megafauna and particularly sus-

pension feeders are abundant along the slope, both in

and out of the pockmarks, likely related to a high partic-

ulate organic matter input (J.P. Foucher, C. Pierre, C.

Decker, L.R. Rufine, S. Ker, G. Westbrook, K. Olu-Le

Roy, J.P. Donval and J.L. Charlou, unpublished observa-

tions).

Nyegga area

On the northeastern flank of the Storegga Slide, in an area

called Nyegga, one pockmark called G11 was discovered at

750 m depth, and surveyed by ROV (Hovland et al. 2005).

Similar to the pockmarks found on the Storegga Slide,

small, scattered seep areas have been observed in the

Nyegga area with chemosynthetic communities distributed

around fluid sources (Decker et al. 2011). As in HMMV,

the habitat nearest to the fluid source consists of microbial

filaments, surrounded by siboglinid fields of Sclerolinum

cf. contortum and Oligobrachia cf. haakonmosbiensis. A ring

of small Rissoidae gastropods was observed at each seep

between the microbial mats and the siboglinids.

Inside the G11 pockmark, small mounds called

‘pingoes’, initially thought to be formed by gas hydrates

(Hovland & Svensen 2006), were also observed. These

mounds are colonized by siboglinid polychaetes S. cf. con-

tortum (Hovland & Svensen 2006; Vanreusel et al. 2009;

Decker et al. 2011). In contrast to HMMV, the Storegga

Slide and the Nyegga area are characterized by an abun-

dant background community and a high particulate fluid

flow, as observed during the ROV dive.

Material and Methods

Sampling strategy

In HMMV, macrofauna (‡250 lm) and megafauna

(fauna visible on video-images) were sampled during the

Vicking cruise (May–June 2006) aboard the R ⁄ V Pourquoi

pas? with the ROV Victor 6000 (chief scientist H. Nouze).

According to zonation, three different habitats were sam-

pled (Fig. 2) corresponding to sampling sites in Decker

et al. (2011): (i) white microbial mats (dominated by

white filamentous bacteria); (ii) sediment adjacent to

microbial mats; (iii) siboglinid polychaete fields domi-

nated by Sclerolinum contortum. In Storegga and Nyegga

pockmarks, two habitats found near the fluid source were

sampled: (i) microbial filaments surrounded by gastro-

pods and (ii) siboglinid fields composed of the two spe-

cies S. cf. contortum and Oligobrachia cf. haakonmosbiensis

(Fig. 3) as well as background megafauna.

The macrofauna and megafauna were sampled using a

suction sampler, manipulated by the ROV and the ROV

Fig. 2. Location of the three habitats

sampled on the Hakon Mosby mud volcano.



grab arm. In HMMV, three replicates in each habitat

were taken over a surface of 0.04 m2 delimited by four

laser beams. In Storegga and Nyegga pockmarks, two rep-

licates were taken in microbial filaments and three in

siboglinid fields and in background sediment. To carry

out isotope measurements, 1–20 individuals of the major

and larger taxa (‡1 mm) were isolated and fixed in liquid

nitrogen on board and then stored at )80 �C in the lab.

Sample processing

Samples for carbon and nitrogen isotope measurements

were obtained from 17 macro- or megafauna taxa identi-

fied to the family level or higher on board. Molluscs were

identified to species level by Anders Waren and polychae-

tes to family level by Marie Morineaux. One to 20 indi-

viduals per taxon were used for analysis. Pooled

individuals or alcohol-preserved individuals were used for

small taxa for which the amount of dry material of a sin-

gle specimen was insufficient. Tables 1 and 2 indicates the

number of samples and the number of individuals pooled

in each sample. Entire organisms were used for isotope

analysis, with the exception of comatulids, ophiurids (for

which analyses were performed on arms) and fishes (mus-

cle tissue), which were dissected on board. For bivalves

and tanaids preserved in alcohol, isotope signatures were

corrected based on the comparison of alcohol- and liquid

nitrogen-preserved individuals. To assess the correction,

individuals from five different taxa (siboglinid and capi-

tellid polychaetes, pycnogonids, amphipods and gastro-

pods) preserved in alcohol were analysed and compared

to frozen individuals of the same taxa. The relationship

between the isotope signatures of alcohol- and liquid

nitrogen-preserved individuals is linear with R2 > 0.86.

Organisms with a carbonate skeleton were acidified in

4.75 m HCl to remove carbonate and then dried at 60 �C

for 72 h. Other non-acidified organisms were freeze-dried

for 48 h. After homogenization, about 1.0 ± 0.5 mg of

material (precision 0.001 mg) was weighed in tin cap-

sules. Duplicates of each sample were prepared and sent

to the Scottish Crop Research Institute in Scotland for

analysis.

Samples were analysed using continuous-flow isotope-

ratio mass-spectrometry (CF-IRMS) using a Sercon 20–20

stable isotope analyser with an ANCA-GSL solid ⁄ liquid

preparation module (Sercon Ltd., Crewe, UK). Using the

dual isotope mode, both d15N and d13C were were mea-

sured on the same sample (precision 0.2&). The meth-

ods are fully described in Scrimgeour & Robinson (2003).

Carbon and nitrogen isotope ratios are expressed as d(delta) values, & deviation from standard reference

materials, which are PDB (Pee Dee Belemnite) for car-

bon, and atmospheric nitrogen for nitrogen (Peterson &

Fry 1987):

dX ¼ ½ðRsamples � RstandardÞ=Rstandard� � 1000

where X = 13C or 15N and R = 13C ⁄ 12C or 15N ⁄ 14N.

Data analysis

Mean values and standard deviations of d13C and d15N by

taxa living in a habitat were calculated. Differences in

d13C or d15N signatures among taxa or habitats were

tested using various non-parametric methods (Kruskal–

Wallis test, non-parametric multiple comparisons, npmc).

All analyses were performed using the free, open-source R

a b

Fig. 3. The two types of habitats sampled on Storegga (a) and Nyegga (b) pockmarks: (a) white microbial filaments and gastropods; (b) siboglinid

field.



Tab

le1.

Isoto

pe

signat

ure

s:d1

3C

and

d15N

(SD

)an

dm

ethan

eco

ntr

ibution:

F m/S

OB–F

m/P

OC

toth

edie

tof

maj

or

taxa

sam

ple

din

diffe

rent

hab

itat

son

HM

MV

.Se

ete

xtfo

rF m

/SO

Ban

dF m

/PO

C

estim

atio

ns.

taxo

n

white

mic

robia

lm

ats

adja

cent

sedim

ent

Scle

rolin

um

conto

rtum

fiel

d

N(n

)d

13C

(SD

)d

15N

(SD

)F m

/SO

B–F

m/P

OC

N(n

)d

13C

(SD

)d

15N

(SD

)F m

/SO

B–F

m/P

OC

N(n

)d

13C

(SD

)d

15N

(SD

)F m

/SO

B–F

m/P

OC

HM

MV

white

mic

robia

lm

ats

5)

35.3

(2.8

))

3.4

(3.1

)0.0

0–0

.35

Poly

chae

ta

Scle

rolin

um

conto

rtum

4(2

0)

)35.0

(0.2

))

1.7

(0.9

)0.0

0–0

.35

Am

phin

om

idae

3(9

))

34.9

(2.8

)7.1

(2.3

)0.0

0–0

.35

Cap

itel

lidae

2(1

5)

)47.8

(0.3

)2.9

(0.7

)0.5

2–0

.69

8(1

5)

)47.6

(7.1

)3.8

(0.8

)0.5

1–0

.69

Dorv

illei

dae

1(1

0)

)48.4

1.4

0.5

4–0

.71

Flab

ellig

erid

ae9

)19.8

(0.4

)8.8

(0.4

)0.0

0

Biv

alvi

a

Thya

siridae

5(2

0)

)27.2

(0.9

)5.4

(1.8

)0.0

0–0

.17

Gas

tropoda

Cry

pto

nat

ica

affinis

1(2

))

20.0

13.7

0.0

0

Alv

ania

sp.

1(2

0)

)46.6

7.2

0.4

7–0

.66

1(1

0)

)40.2

6.0

0.2

1–0

.49

Am

phip

oda

3(9

))

50.6

(0.2

)5.2

(0.5

)0.6

4–0

.77

Cap

relli

dae

3)

28.6

(2.7

)6.8

(0.5

)0.0

0–0

.19

Isopoda

4(8

))

26.3

(0.8

)9.4

(0.7

)0.0

0–0

.13

Tanai

dac

ea1

(3)

)29.3

7.3

0.0

0–0

.21

Panto

poda

1)

45.1

4.6

0.4

1–0

.62

8)

47.8

(4.1

)4.6

(1.0

)0.5

2–0

.69

6)

34.8

(1.9

)6.1

(2.2

)0.0

0–0

.35

Ophiu

roid

ea1

)22.2

9.3

0.0

0–0

.02

Zoar

cidae

4)

49.0

(4.2

)7.5

(0.9

)0.5

7–0

.72

Mea

nm

ethan

e-der

ived

carb

on

(SD

)

0.3

9(0

.22)–

0.6

1(0

.15)

0.4

9(0

.17)–

0.6

7(0

.11)

0.0

0–0

.17(0

.15)

N=

num

ber

of

indiv

idual

sor

poole

dsa

mple

s;n

=to

talnum

ber

of

indiv

idual

ssa

mple

dbef

ore

poolin

g.



Tab

le2.

Isoto

pe

signat

ure

s:d1

3C

and

d15N

(SD

)an

dm

ethan

eco

ntr

ibution:

F m/S

OB–F

m/P

OC

toth

edie

tof

maj

or

taxa

sam

ple

din

diffe

rent

hab

itat

son

Store

gga

and

Nye

gga

pock

mar

ks.

See

text

for

F m/S

OB,

F m/P

OC

estim

atio

ns.

white

mic

robia

lfila

men

tsan

dgas

tropods

Siboglin

idae

fiel

din

div

idual

sin

the

vici

nity

of

fluid

seep

age

N(n

)d

13C

(SD

)d

15N

(SD

)F m

/SO

B–F

m/P

OC

N(n

)d

13C

(SD

)d

15N

(SD

)F m

/SO

B–F

m/P

OC

N(n

)d

13C

(SD

)d

15N

(SD

)F m

/SO

B–F

m/P

OC

Store

gga

pock

mar

k

poly

chae

ta

Olig

obra

chia

cf.

haa

konm

osb

iensi

s

12

(10)

)62.4

(4.8

)3.7

(0.6

)1

Scle

rolin

um

cf.

conto

rtum

7(2

0)

)41.5

(5.1

)2.4

(0.4

)0–0

.54

Am

phin

om

idae

1)

45.8

10.8

0.2

7–0

.66

Lum

briner

idae

1)

41.9

8.6

0.0

3–0

.56

Gas

tropoda

Alv

ania

sp.

10

(10)

)42.1

(3.6

)7.4

(0.8

)0.0

4–0

.56

3(1

0)

)42.8

(2.8

)7.1

(0.3

)0.0

8–0

.58

Bucc

inum

sp.

2)

19.5

(2.1

)12.4

(0.4

)0.0

04

)18.8

(0.9

)12.6

(0.9

)0.0

0

Colu

ssa

bin

i3

)43.9

(1.8

)7.7

(0.6

)0.1

5–0

.61

Copep

oda

1)

22.9

6.2

0–0

.04

Car

idea

Car

idea

sp1

4)

20.5

(2.0

)11.2

(3.4

)0.0

04

)19.6

(2.0

)11.0

(2.3

)0.0

0

Car

idea

sp2

1)

29.7

12.4

0–0

.23

2)

24.2

(5.4

)11.6

(0.8

)0–0

.07

Car

idea

sp3

1)

43.4

11.8

0.1

2–0

.58

Holo

thuro

idea

2)

23.9

(5.9

)1.5

(1.8

)0–0

.07

Crinoid

ea18

)22.2

(1.5

)8.4

(1.6

)0–0

.02

Ophiu

roid

ea

Gorg

onoce

phal

us

sp.

3)

18.5

(0.6

)13.5

(0.6

)0.0

0

Mea

nm

ethan

e-der

ived

carb

on

(SD

)

0.0

5(0

.07)–

0.2

9

(0.3

4)

0.3

2(0

.47)–

0.6

2

(0.3

2)

0.0

2(0

.04)–

0.1

5

(0.2

5)

Nye

gga

G11

Pock

mar

k

Poly

chae

ta

Scle

rolin

um

cf.

conto

rtum

7(2

0)

)36.0

(2.6

)2.0

(0.7

)0-0

.34

12

(20)

)34.8

(1.8

)0.5

(1.0

)0–0

.31

Cap

itel

lidae

5)

64.4

(20.5

)9.0

(0.9

)1

3)

28.6

(1.8

)6.9

(0.6

)0–0

.17

Unid

entified

Poly

chae

ta1

)45.7

8.1

0.3

5–0

.57

Biv

alvi

a

Thya

siridae

1)

29.9

9.7

0–0

.20

Gas

tropoda

Alv

ania

sp.

19

(10)

)37.8

(2.6

)6.0

(0.9

)0.0

6–0

.38

5(1

0)

)39.8

(1.9

)6.4

(0.9

)0.1

3–0

.43

1(1

0)

)37.2

6.5

0.0

4–0

.37

Am

phip

oda

Am

phip

oda

sp1

1)

21.3

7.6

0.0

02

)22.7

(0.3

)7.6

(1.4

)0–0

.03

Am

phip

oda

sp2

1)

21.2

11.4

0.0

0

Am

phip

oda

sp.

1)

20.0

10.4

0.0

0

Car

idea

Car

idea

sp1

5)

21.4

(0.8

)8.7

(0.1

)0.0

0

Car

idea

sp2

2)

37.2

(0.2

12.4

(0.1

)0.0

4–0

.37

3)

28.8

(13.1

)9.9

(2.1

)0–0

.18

Ast

eroid

ea1

)21.0

11.5

0

Mea

nm

ethan

e-der

ived

carb

on

(SD

)

0.2

9(0

.43)–

0.4

6

(0.3

7)

0.0

2(0

.05)–

0.1

6

(0.1

6)

0(0

.02)–

0.1

2

(0.1

7)

N=

num

ber

of

sam

ple

s;n

=num

ber

of

indiv

idual

spoole

dby

sam

ple

.V

alues

for

S.cf

.co

nto

rtum

sam

ple

don

the

‘pin

goe’

wer

eunder

lined

.



software (R Development Core Team 2008). The npmc

library was used for non-parametric multiple comparisons

(Helms & Munzel 2008).

Trophic relationships were discussed using both carbon

and nitrogen signatures, based on enrichments of £1& in

d13C and 3.4& in d15N, which are the generally accepted

values corresponding to the change in stable isotopes

between two successive trophic levels (Conway et al.

1994).

For each habitat and taxa, the percentage of methane-

derived carbon in the diet was estimated using a two-

source isotope mixing model (Fry & Sherr 1984). We

used the equation employed by Levin & Michener (2002):

Fm ¼ d13Ci � d13Csource2=d13Cmethane � d13Csource2

where Fm is the percentage of methane-derived carbon in

the diet of the individual, and d13Ci, the carbon isotopic

signature of the taxon. As assumed by Levin & Michener

(2002), no trophic shift was included as negligible (£1&

per trophic level) for d13C. To obtain upper and lower

estimates of methane-derived carbon infaunal tissues, two

estimates were made with two different ‘source 2’ end-

members: the lower estimate was obtained using the d13C

signature of sulphide-oxidizing bacteria (d13CSOB) (Fm/SOB)

and the upper estimate using the d13C signature of particu-

late organic carbon (d13CPOC) (Fm/POC). We used a

d13CPOC value of )22.11 ± 0.6& from surface sediment

sampled at a reference site in the volcano (Van Gaever et al.

2006), a d13Cmethane value of )59.2& in HMMV (Lein

et al. 1999), a d13Cmethane value of )57.9& in the Storegga

Slide pockmark (Paull et al. 2008) and )69.3& for the G11

pockmark in the Nyegga area (Hovland et al. 2005). For

d13CSOB, we used the signature of sulphur-oxidizing bacte-

ria from mats on HMMV. The mean signature of Scleroli-

num cf. contortum sampled in Storegga slide, which

probably contained the same sulphur-oxidizing symbionts

as S. contortum in HMMV (Losekann et al. 2008), were

used to determine the proportion of chemosynthesis-

derived carbon for Storegga Slide pockmark taxa. Indeed,

no estimation for d13C sulphur-oxidizing bacteria from

mats was available in Storegga and Nyegga pockmarks and

in HMMV, the d13C of S. contortum and microbial mats

were within 0.3 per mil.

Results

Stable isotope signatures

The d13C signatures of the 19 species of macro- and

megafauna sampled respectively on the microbial mats

(four species), adjacent sediment (five species) and sibo-

glinid fields (10 species) in HMMV ranged from )50.6 to

)20& (Table 1, Fig. 4). The most depleted taxa are those

associated with microbial mats (Dorvilleidae and Capitel-

lidae polychaetes, pygnogonids and Rissoidae gastropods)

or sampled a few metres apart in sediment (Amphipoda

and Zoarcidae, which have the lightest d13C). Macrofaunal

taxa associated with microbial mats and adjacent sediment

had very depleted but homogeneous d13C values,

all between )45 and )51&, although the microbial mat

signature had higher values ()35.3 ± 3.4&) (Table 1).

Capitellidae polychaetes and pycnogonids (Pantopoda)

living on mats and in adjacent sediment showed similar

carbon and nitrogen signatures in both habitats, whereas

Rissoidae gastropods Alvania sp. were less depleted in

carbon in adjacent sediment compared to microbial mats.

Signature variability among individuals tended to be

higher in adjacent sediment than on mats (e.g. Capitelli-

dae) but the number of individuals analysed from sedi-

ment was generally higher and likely represented a bias in

this comparison.

The three sampled habitats are clearly separated

according to macrofaunal d13C signatures (Kruskal-Wallis:

p = 0.0001), however, mats and adjacent sediment did

not significantly differ for macrofaunal d13C signatures

(non-parametric multiple comparisons: P = 0.970)

(Fig. 4). Indeed, the mean signature of macrofauna asso-

ciated with S. contortum fields had heavier d13C values

()27.0 ± 5.7&) than macrofauna on white microbial

mats or the adjacent sediment ()47.0 ± 1.46& and

)47.04 ± 4.0&, respectively).

There was a larger range of d13C signatures in sibogli-

nid fields than in other habitats, from)19.8 to )35.0&.

Together with S. contortum, Amphinomidae polychaetes

and pygnogonids showed the most depleted signatures,

tanaids, Caprellidae amphipods, Thyasiridae bivalves and

isopods had intermediate d13C (between 25 and 30&),

whereas the ophiurids, the gastropod Cryptonatica affinis

and Flabegilleridae polychaetes had higher d13C. Small

crustaceans (amphipods, isopods and tanaıds) showed

variable d13C signatures, suggesting heterogeneity of their

food sources. The carbon signature of pygnogonids, the

only taxa sampled in the three habitats, showed a clear

enrichment of at least 10& in the siboglinid fields.

In the mats and adjacent sediments, the d15N range

was relatively narrow, from the mat itself, with high sam-

ple variation ()3.4 ± 3.1&), to the Rissoidae gastropods

(7.2&), whereas Zoarcidae fish had the highest value

(7.5 ± 0.9&) in the adjacent sediment community. The

species sampled here have higher signatures in nitrogen

than those from mats (e.g. Dorvillieidae polychaete),

which may derive nutrition directly from microorgan-

isms.

A larger d15N range was obtained in taxa sampled in

siboglinid fields. Sclerolinum contortum had the lowest



d15N value ()1.7 ± 0.9&). Other polychaetes had different

isotopic signatures according to their family, revealing

different trophic levels. Nevertheless, d15N values were rela-

tively uniform (average +6.7&), except for isopods, which

were higher in the trophic chain (+9.4&). Gastropods

Cryptonatica affinis had the highest d15N value (+13.7&).

Contrary to HMMV, on the Storegga pockmark, the

values of d13C did not differ clearly among habitats

(Kruskal–Wallis: P = 0.10) but did vary with taxa, regard-

less of the habitat from which they were sampled

(Table 2, Fig. 5). The different species of gastropods

(Alvania sp., Buccinum sp. and Colus sabini) and shrimps

differed from each other in their signature, whatever their

habitat, but these differences were not statistically signifi-

cant. First, the two siboglinid species were distinct, with

lower d13C values for Oligobrachia cf. haakonmosbiensis

()62.4 ± 4.8&) than for S. cf. contortum ()41.5 ± 2.4&)

(Kruskal–Wallis, P < 0.001). Alvania sp. and Colus sabini

gastropods, the polychaetes and one Caridae species (sp3)

had depleted d13C values ranging from )46 to )42&,

whereas the Buccinum sp. gastropods, the echinoderms

(holothurids, crinoids and ophiurids) and the other

shrimps (species sp1 and sp2) had values ranging from

)30 to )18&. Regarding nitrogen signatures, siboglinids

and holothurids had the lowest values; the other taxa dif-

fered, with values ranging from 6 to 13&. As observed

for d13C values, d15N values fluctuated among taxa not

among habitats. The highest d15N values were observed in

the ophiurids Gorgonocephalus sp. (13.5&), Buccinum sp.

(12.6&) and shrimps (average 11.5 ± 0.4&).

As in Storegga pockmarks, in the G11 pockmark

in Nyegga, habitats did not have distinct d13C values

(Kruskal–Wallis: P = 0.19). However, taxa living far from

seeps appeared to have higher d13C values compared to

other taxa, although these differences were not statistically

significant (non-parametric multiple comparisons: white

microbial mats versus non-seep habitat: P = 0.37 and

siboglinid field versus non-seep habitat: P = 0.25). Capi-

tellid polychaetes that were abundant on microbial mats

differed greatly from other taxa and had the lowest d13C

values, despite large variation ()64.4 ± 20.5&). For d15N,

siboglinids showed the lowest values, whereas the

other taxa varied from 6 and 12&. Whatever the habitat,

Alvania sp. gastropods had a narrow range of signatures:

d13C ranged from )37.2 to )39.8& and d15N from 6 to

6.5&. However, this taxon was probably composed of

two species (A. Waren, personal communication), which

were not identified before analyses.

Methane-derived carbon contribution

According to the two-source isotope mixing model,

macrofauna living on HMMV microbial mats or adjacent

sediment depended mainly on chemosynthesis, whatever

the carbon source used in the model (Table 1). The mean

estimated contribution of methane-derived carbon was

39–61% for mats (48–67%, excluding the signature of the

mat itself) and 49–67% for taxa in the adjacent sediment.

Capitellids, zoarcids, amphipods and pycnogonids sam-

pled on or close to the microbial mats were the taxa that

depended the most on methane-derived carbon.

In siboglinid fields, the mean contribution of methane-

derived carbon was 0–17% and was highest (36%) for

amphinomid polychaetes and pycnogonids. Most of the

associated taxa (Flabelligerid polychaetes, ophiurids, gas-

tropods) seemed to rely mainly on photosynthesis-based

carbon, with d13C values of >)30& and estimates of phy-

toplankton-derived carbon (1–Fm/POC) of 89–100%.

In the Storegga pockmark, the mean methane-derived

carbon contribution was higher in siboglinid fields

Fig. 4. Isotope signatures of different taxa on Hakon Mosby mud volcano [ : macrofauna on white microbial mats (WhM), : macrofauna on

white mats in adjacent sediment (S), : macrofauna in siboglinid fields (Sc)]. Bars indicated standard deviation.



(32–62%) than on microbial mats (5–29%) (Table 2).

Whatever the habitat, some macrofauna taxa (Buccinum

sp., Caridae sp1) had slightly depleted d13C signatures

(>)30&), with percentages of methane-derived carbon

as low as 0%. Other taxa, including siboglinid, amphino-

mid and lumbrineid polychaetes, Alvania sp. and Colus

sabini gastropods, relied mainly on chemosynthesis-

derived carbon, as shown by low d13C signatures and

high FPOC values. Except for Oligobrachia cf. haakon-

mosbiensis, these taxa depended mainly on sulphide-oxi-

dizing-derived carbon, as shown by the low methane

contribution (Fm/SOB < 50%) but high chemosynthesis

contribution (Fm/POC > 50%) (Table 2). Most taxa

sampled in the vicinity of fluid seepage relied mainly on

photosynthesis, with less than 8% methane-derived

carbon, except Caridean shrimp sp3 and Alvania sp.,

whose signatures did not vary between on- and off-seep

habitats. The three species of shrimps (sp1, sp2 and sp3)

differed greatly in their carbon signatures and their

dependence on methane-derived carbon, with no rela-

tionship with the habitat where they were sampled. The

most depleted value of d13C was observed off-seep (sp3)

and the least depleted value off-seep (sp1). The large

echinoderms that were abundant around the seepage

areas (holothurids, crinoids and ophiurids) showed

only a limited consumption of methane-based carbon

(0–8%).

Similarly, on the G11 pockmark in the Nyegga area,

some of the taxa living on siboglinid fields and off-seep

depended on photosynthesis, with a contribution rang-

ing from 79 to 100% of photosynthesis-derived carbon

(1-Fm/POC). However, Alvania sp. gastropods showed

significant contributions of methane-based material in

their diet whatever the habitat, on- and off-seeps

(Fm = 4–43%). As in the Storegga pockmark, the

shrimp signatures in the Nyegga differed among the

two species. Other vagrant species, such as amphipods,

did not rely, or relied very little, on methane, even

those sampled on the microbial mats. Finally, the most

depleted taxa were capitellida polychaetes, with an esti-

mated 100% methane-derived carbon in the microbial

filament habitat.

a

b

Fig. 5. Isotope signatures of different taxa on the pockmarks: (a) Storegga pockmark and (b) Nyegga pockmark [ : macrofauna on white micro-

bial mats and gastropods (WhM ⁄ G), : macrofauna in siboglinid fields (Sc), : macrofauna in the vicinity of fluid seepage (Vic)]. Bars indicated

standard deviation.



Discussion

HMMV: a food web structured by habitat

Stable isotope values of macrofauna and megafauna along

the three study sites and estimation of the contribution of

chemosynthesis to their nutrition showed that their diet

is mainly based on chemosynthetic production. However,

some taxa living in the active seep area or in the vicinity

of fluid seepage rely partially or entirely on phytoplank-

ton-derived carbon.

In HMMV, the first stable isotope analyses performed

on a few macrofaunal taxa revealed high variability in

carbon isotope signatures, with a high contribution of

methane-derived carbon (Gebruk et al. 2003). In a preli-

minary study, we observed that carbon isotope signature

varies among habitats (Decker & Olu 2010). On micro-

bial mats and in adjacent sediment, the contribution of

methane-derived carbon was high (from 44 ± 19% to

63 ± 12%). Nematodes ()41.6 ± 0.4&), which dominate

the meiofaunal compartment on microbial mats, are

assumed – based on their d13C signature – to depend

on primary production by the sulphur-oxidizing Beggia-

toa sp. mat-forming bacteria ()42.7 ± 0.2&) (Van

Gaever et al. 2006). In our study, mats showed a higher

d13C value ()35.25 ± 2.84 &), and this value was very

different from that of most macrofauna living on mats

or in adjacent sediment (average )46.9 8 ± 1.46& and

)47.04 ± 4.01&, respectively). The macrofauna living

on microbial mats (capitellid and dorvilleid polychaetes,

rissoid gastropods and pycnogonids) seemed to depend

on methane-derived carbon (41–71%), which may be

produced by aerobic methanotrophic bacteria observed

in the top centimetre of sediment of this habitat or to

a single species of anaerobic methanotrophic archaea

(Losekann et al. 2007). Fauna sampled on sediment

adjacent to mats (amphipods, pygnogonids and zoarcid

fishes) also rely highly on methanotrophically fixed car-

bon. Zoarcids were frequently seen making incursions

in the mats during dives, where the microbial commu-

nity may contribute to their diet. Two food chains may

thus co-occur in this habitat, one based on sulphur-oxi-

dation and the other on aerobic or anaerobic methano-

trophy. The negative correlation observed between

meiofauna and macrofauna densities in HMMV (Van

Gaever et al. 2009; Decker et al. 2011), with a domi-

nance of nematodes on microbial mats, suggests that

these two groups compete for food and space. The

abundance of filamentous Beggiatoa mats compared to

methanotrophic bacteria or archaea may favour nema-

todes over macrofaunal taxa in this habitat. The zoarcid

fish could be predators of pygnogonids and caprellids

living in sediment adjacent to mats, and of capitellids

living in the mats themselves according to the nitrogen

and carbon signatures. Zoarcidae have been observed to

make frequent inclusions in the mats but lie on the

bottom in the bare adjacent sediment. They thus may

contribute to limit Capitellidae population and favour

the dominance of Nematoda.

In the Sclerolinum contortum fields, the contribution of

methane-derived carbon was lower than on microbial mats

and varied among taxa (0 to 17 ± 15%). Only a few taxa

depended chiefly on chemosynthesis, namely S. contortum

and amphinomid polychaetes and pygnogonids, but the

levels of methane-derived carbon were very low (0–35%).

Regarding the food web, the length of food chains

seems to vary among habitats and within the microbial

mats ⁄ adjacent sediment habitat. Indeed, in both these

habitats, the food chain based on sulphur-oxidation is

about 9.8& length, whereas that based on methanotrophy

is shorter (5.8& length). On siboglinid fields, S. contor-

tum had the lowest d15N values and the various associated

fauna corresponded to different levels of consumers (d15N

from 5.39 to 9.38 &). We can expect two (microbial

mats ⁄ sediment habitat) to four (siboglinid fields) trophic

levels according to habitats. The occurrence of vagile

megafauna in the first- or second-level consumers may

contribute to the export of seep production from the

volcano. Nevertheless, the species sampled (ophiurids,

Cryptonatica affinis) showed a low contribution of che-

mosynthesis-based material to their diet. Pycnogonid and

smaller size crustaceans such as amphipods, isopods or

tanaids seem to benefit more from the local production

but likely have a limited dispersion around the seep area

and likely do not contribute much to the export of local

production. Zoarcidae also occur in siboglinid fields but

have not been sampled here. However, they seemed very

static. Although not sampled, Rajidae are quite numerous

and may play a role in these exchanges with the back-

ground community.

In HMMV, the mean percentage of methane-derived

carbon was clearly higher on microbial mats and in adja-

cent sediment (49–67%) than in siboglinid fields (0–17%).

This contrast between habitats has been observed on the

Florida Escarpment, with methane-derived carbon con-

tributing 50% on microbial mats and 23% in siboglinid

fields (Levin & Mendoza 2007) (Fig. 6). Although the

mean carbon signature of the fauna associated with

HMMV microbial mats was not as low as in the Florida

mat communities (Fig. 6), the mean methane-derived car-

bon contribution was higher at HMMV due to a less

depleted methane carbon signature ()59.2 versus )85& in

the HMMV versus the Florida escarpment, respectively).

Other estimations of methane-derived carbon on mats

are even lower, ranging from 20 to 44% on the Oregon

Margin and from 0 to 5% on the California continental

slope; however, depth influences the contribution of meth-



ane-derived carbon at these sites (Levin & Michener

2002). In contrast, siboglinid fields in the Gulf of Alaska

have higher methane-derived carbon estimates (32–51%)

(Levin & Mendoza 2007) than those on HMMV (0–

17%). Nevertheless, frenulate fields at the Uruti ridge off

New Zealand have lower methane-derived carbon (0–

1%) than siboglinid fields on HMMV (Thurber et al.

2010). These differences may be related to depth and

the influence of surface photosynthetic production being

higher in the shallower site: HMMV is located at

1200 m depth, whereas the Alaska seeps are deeper than

3000 m; the Uruti ridge is located at 716 m. Another

hypothesis is that the relatively mild environmental con-

ditions in the siboglinid fields on HMMV (low sulphide

concentrations and high penetration depth for oxygen;

De Beer et al. 2006) allows vagrant species to penetrate

from the background environment.

In HMMV, the relative importance of methanotrophic

versus thiotrophic micro-organisms in the nutrition of

invertebrates also varied, mainly among habitats, and

most of the organic carbon was derived from sulphur-

oxidizing bacteria [(1–Fm/SOB) average 70%] except on

microbial mats. In contrast, at the Blake Ridge seep site,

where ‘methanotrophic’ mussels form one of the habitats,

most of the organic carbon derives from methane (60%),

which is similar to our mean estimation on HMMV

microbial mats (49%). However, as on HMMV mats,

free-living thiotrophic bacteria also contribute to the

nutrition of the invertebrate community associated with

mussel bed (Van Dover et al. 2003).

Storegga and Nyegga pockmarks: influence of external

inputs

The fluid seepage areas of the Storegga and Nyegga pock-

marks are much more limited in area compared to

HMMV and did not show such a clear contrast in d13C

among habitats (Fig. 3). Siboglinid species differed in

their carbon signatures (Sclerolinum cf. contortum:

)41.46&, Oligobrachia cf. haakonmosbiensis: )62.37&),

as previously shown by Van Gaever et al. (2006) and

Losekann et al. (2008) in HMMV. As in HMMV, S. cf.

contortum likely harbours sulphur-oxidizing symbionts,

whereas the lower d13C value (<50&) of O. cf. haakon-

mosbiensis indicates a methane-carbon origin. However,

no methanotrophy genes (but those from sulphur oxida-

tion) have been detected in O. haakonmosbiensis (Losek-

ann et al. 2008) and the chemoautotrophic process

occurring in these species is still unresolved.

On seep areas of Storegga and Nyegga pockmarks, in

contrast to HMMV, no distinct food chains could be elu-

cidated according to habitats (Fig. 5). Indeed, signatures

of vagrant fauna and their chemosynthetic-carbon contri-

butions did not vary among habitats, but did among spe-

cies (Table 2). Alvania sp. and Colus sabini gastropods

observed on or around microbial filaments probably graze

on them. The d13C signature of Alvania sp. and its low

variability suggests that it may be endemic to seeps.

Shrimps are scavengers (Segonzac et al. 1993; Iken et al.

2001) but the three shrimps species found in the pock-

marks seem to differ in their use (or dependence) on che-

Fig. 6. Isotope signatures and macrofauna associated with habitats in different seep areas around the world [ : Hakon Mosby mud volcano

()1250 m; this study); : Storegga and Nyegga pockmarks ()750 m; this study); : Florida escarpment ()3300 m); : Oregon margin (600–

800 m) and Eel River ()500 m); : Gulf of Alaska ()4400 m); Kodiak ()3300 m); Unimak ()4400 m)] (Levin & Michener 2002; Levin & Mendoza

2007). Bars indicated standard deviation.



mosynthetic carbon production, as shown by their d13C

signatures. Indeed, one species (sp1) seems to be totally

independent of seep production, with a signature close to

)20& whatever the habitat, contrasting with (sp3), which

signature indicates significant chemosynthesis contribu-

tion in its diet. In contrast, the diet of the third species

(sp2) found both in Storrega and Nyegga areas seems to

vary according to the habitat where it was sampled

(Table 2). In these two pockmarks, we sampled several

background species (e.g. Buccinum sp., amphipods, one

species of shrimp, holothurids, ophiurids and asteroids)

that relied mostly on detritus-based production (1–Fm/

POC >92%). Although this is not as clear in Nyegga pock-

marks, two food chains could co-occur in Storegga pock-

marks, one chain based on chemosynthetic production

(d13C around )40&) and the other on photosynthetic

production (d13C around )20&). These chains have dif-

ferent lengths, 9.2& to 12&, suggesting respectively three

to four trophic levels. The lack of a trophic level near the

seep source, also observed on HMMV, could be explained

by the absence of predators or scavengers adapted to the

high sulphide concentration in these environments. Zoar-

cid fishes are an exception on HMMV, grazing on micro-

bial mats but seemed to mainly lie on adjacent sediment

(C. Decker et al. 2011).

Comparing the volcano and pockmarks, the range of

d15N and therefore the number of trophic level is higher

in the pockmarks than in HMMV. The diversity of high

level consumers is much higher in the pockmarks than in

the volcano, with several species of shrimps, large size

gastropods and echinoderms, whereas only ophiurids and

the gastropod Crytonatica affinis were sampled in the

HMMV S. contortum fields. Export of seep production

and methane-derived carbon is likely much higher in the

Storrega ⁄ Nyegga pockmarks than around the mud vol-

cano. Although some have limited contribution of meth-

ane-derived carbon in their diet, several vagile megafauna

taxa including shrimps, echinoderms and gastropods

could contribute to the export of seep material to the

background.

The Storegga Slide probably receives higher vertical or

lateral fluxes of photosynthesis-produced material com-

pared to HMMV. Visual observations during dives

indicated high densities of particles in the water column

and high densities of suspension feeders (sponges,

Gorgonocephalus sp., crinoids, etc.) in the whole area.

HMMV is located deeper, is more isolated and is less

likely to receive external organic matter inputs. Moreover,

the small size of scattered seepages (average 30 cm diame-

ter) in the pockmarks, and their low fluid activity

compared to the mud volcano probably encourage back-

ground fauna to make incursions in the seeps, even if

some species do not utilize the chemosynthetic primary

production. Similar observations have been made at the

California continental slope and shelf and at Pacific seeps,

where the nutrition of macrofauna of shallower sites is

influenced less by seeps (Levin et al. 2000; Levin & Miche-

ner 2002). As shown in Fig. 6, the mean signatures of

macrofauna on microbial mats can vary greatly, from the

Florida escarpment ()56&; 3300 m) to the Eel River seeps

()22&; 500 m), and seem to be influenced by depth

(Levin & Michener 2002). The mean macrofaunal signa-

ture on HMMV mats was located between the Florida

escarpment mats and Oregon signatures, whereas Storegga

and Nyegga mat fauna had signatures with higher values,

due to incursions of background species. Nevertheless, the

pockmarks had more depleted signatures than Eel

River mat fauna because the pockmarks are inhabited by a

higher proportion of chemosynthesis-dependent species

(endemic or opportunists) compared to the vagrant back-

ground fauna that dominate at Eel River. For siboglinid-

associated fauna, the sites with the most depleted values

are the deepest cold-seep sites studied: Gulf of Alaska

seeps ()46&, 4400 m) and Unimak seeps ()24&,

4400 m), suggesting that depth is not the main controlling

factor of macrofaunal nutrition in this habitat.

Conclusion

The isotopic signatures of the seep-associated fauna

revealed food chains that were more or less distinct

among habitats, according to the seep sites we explored.

The signatures reflect differences in spatial patterns

between seep sites, with very active mud volcano with a

clear activity gradient from the centre to the periphery,

and pockmarks characterized by a mosaic of isolated

fluid seepages of lower intensity and restricted in space,

scattered in an area receiving higher external inputs.

The mud volcano showed contrasted food chains with

highly specialized fauna in habitats influenced by seep

activity, whereas the nutritional patterns on pockmarks

were structured less according to habitat and more

according to species. Moreover, the less depleted macro-

faunal d13C signatures on Storegga pockmarks suggest

that the size of the seep and its spatial organization

influence the distribution of seep endemic versus back-

ground species in the different habitats, and emphasize

the relative importance of the seep production in food

chains.

Acknowledgements

We are grateful to the chief scientists and project leader

of the Vicking cruise (H. Nouze, J. P. Foucher), and to



the captain and the crew of the R ⁄ V Pourquoi pas? and

the ROV Victor 6000 team. We thank our colleagues for

their help on board: A. Andersen, J. C. Caprais, M.-C.

Fabri, P. Noel, L. Toffin, S. Van Gaever. The isotope mea-

surements were processed by the Scottish Crop Research

Institute. Polychaetes were identified by Marie Morine-

aux, whose work was supported by the French ANR pro-

gram DeepOases. We thank Anders Waren and Marie

Morineaux for gastropod and polychaete identifications,

respectively, and Olivier Gauthier for the isotope plot

R-function design. We are grateful to the two anynomous

reviewers and to Prof. Craig Young, co-editor of Marine

Ecology for their useful comments to improve the manu-

script. The English was professionally edited by Carolyn

Engel-Gautier. The Vicking cruise was part of the FP6

European program HERMES. This publication is a

contribution to the FP7 European program Hermione.

References

Carlier A., Ritt B., Rodrigues C., Sarrazin J., Olu K., Grall J.,

Clavier J. (2010) Heterogeneous energetic pathways and

carbon sources on deep eastern Mediterranean cold seep

communities. Marine Biology, 157, 2545–2565.

Conway N.M., Kennicutt M.C., Van Dover C.L. (1994) Stable

Isotopes in the Study of Marine Chemosynthetic-Based

Ecosystems. In: Lajtha K., Michener R.H. (Eds), Stable

isotopes in ecology and environmental science. Blackwell

Scientific Publications, London: 156–186.

De Beer D., Sauter E., Niemann H., Kaul N., Foucher J.P.,

Witte U., Schluter M., Boetius A. (2006) In situ fluxes and

zonation of microbial activity in surface sediments of the

Hakon Mosby mud volcano. Limnology and Oceanography,

51, 1315–1331.

Decker C., Morineaux M., Van Gaever S., Caprais J.-C.,

Lichtschlag A., Gauthier O., Andersen A.C. & Olu K. (2011)

Habitat heterogeneity influences cold-seep macrofaunal

communities within and among seeps along the Norwegian

margin. Part 1: Macrofaunal community structure. Marine

Ecology, doi:10.1111/j.1439-0485.2011.00503.x.

Decker C., Olu K. (2010) Does macrofauna nutrition vary

among habitats at Hakon Mosby Mud Volcano? Cahiers de

Biologie Marine, 51, 361–367.

Feseker T., Foucher J.P., Harmegnies F. (2008) Fluid flow or

mud eruptions? Sediment temperature distributions on

Hakon Mosby mud volcano, SW Barents Sea slope. Marine

Geology, 247, 194–207.

Foucher J.P., Westbrook G., Boetius A., Ceramicola S., Dupre

S., Mascle A., Mienert J., Pfannkuche O., Pierre C., Praeg D.

(2009) Structure and drivers of cold-seep ecosystems.

Oceanography, 22, 92–109.

Foucher J.P., Dupre S., Scalabrin C., Feseker T., Harmegnies

F., Nouze H. (2010) Changes in seabed morphology,

mud temperature and free gas venting at the Hakon

Mosby mud volcano, offshore northern Norway, over

the time period 2003–2006. Geo-Marine Letters, 30,

157–167.

Fry B., Sherr E.B. (1984) Delta 13C measurements as indicators

of carbon flow in marine and freshwater ecosystems. Contri-

butions in Marine Science, 27, 13–46.

Gebruk A., Krylova E.M., Lein A., Vinogradov G.M., Anderson

E., Pimenov N.V., Cherkashev G.A., Crane K. (2003) Meth-

ane seep community of the Hakon Mosby mud volcano

(Norwegian Sea): composition and trophic aspects. Sarsia,

88, 394–403.

Helms J., Munzel U. (2008) Nonparametric Multiple Compari-

sons. R package version 1.0.7.

Hovland M., Svensen S. (2006) Submarine pingoes: indicators

of shallow gas hydrates in a pockmark at Nyegga, Norwe-

gian Sea. Marine Geology, 228, 15–23.

Hovland M., Svensen H., Forsberg C.F., Johansen H., Fichler

C., Fossa J.H., Jonsson R., Rueslatten H. (2005) Complex

pockmarks with carbonate-ridges off mid-Norway: products

of sediment degassing. Marine Geology, 218, 191.

Hustoft S., Mienert J., Buenz S., Nouze H. (2007) High-resolu-

tion 3D-seismic data indicate focussed fluid migration path-

ways above polygonal fault systems of the mid-Norwegian

margin. Marine Geology, 245, 89–106.

Iken K., Brey T., Wand U., Voigt J., Junghans P. (2001) Food

web structure of the benthic community at Porcupine

Abyssal Plain (NE Atlantic): a stable isotope analysis.

Progress in Oceanography, 50, 383–405.

Jerosch K., Ludtke A., Schluter M., Ioannidis G.T. (2007)

Automatic content-based analysis of georeferenced image

data: detection of Beggiatota mats in seafloor video mosaics

from the Haakon Mosby Mud Volcano. Computer and

Geosciences, 33, 202–218.

Lein A., Vogt P., Crane K., Egorov A., Ivanov M. (1999)

Chemical and isotopic evidence for the nature of the fluid

in CH4-containing sediments of the Hakon Mosby Mud

Volcano. Geo-Marine Letters, 19, 76–83.

Levin L.A. (2005) Ecology of cold seep sediments: interactions

of fauna with flow, chemistry and microbes. Oceanography

and Marine Biology Annual Review, 43, 1–46.

Levin L.A., Mendoza G.F. (2007) Community structure

and nutrition of deep methane-seep macrobenthos from

the North Pacific (Aleutian) Margin and the Gulf of

Mexico (Florida Escarpment). Marine Ecology, 28, 131–

151.

Levin L.A., Michener H.M. (2002) Isotopic evidence for

chemosynthesis-based nutrition of macrobenthos: the

ligntness of being at Pacific methane seeps. Limnology and

Oceanography, 47, 1336–1345.

Levin L.A., James D.W., Martin C.M., Rathburn A.,

Harris L., Michener R. (2000) Do methane seeps

support distinct infaunal assemblages? Observations on

community structure and nutrition from the northern



California slope and shelf. Marine Ecology Progress Series,

208, 21–39.

Losekann T., Knittel K., Nadalig T., Fuchs B., Niemann H.,

Boetius A., Amann R.L. (2007) Diversity and abundance of

aerobic and anaerobic methane oxidizers at the Haakon

Mosby Mud Volcano, Barents Sea. Applied and Environmen-

tal Microbiology, 73, 3348–3362.

Losekann T., Robador A., Niemann H., Knittel K., Boetius A.,

Dubilier N. (2008) Endosymbioses between bacteria and

deep-sea siboglinid tubeworms from an Arctic Cold Seep

(Haakon Mosby Mud Volcano, Barents Sea). Environmental

Microbiology, 10, 3237–3254.

MacAvoy S.E., Carney R.S., Fisher C.R., Macko. S.A. (2002)

Use of chemosynthetic biomass by large, mobile, benthic

predators in the Gulf of Mexico. Marine Ecology Progress

Series, 225, 65–78.

MacAvoy S.E., Fisher C.R., Carney R.S., Macko S.A. (2005)

Nutritional associations among fauna at hydrocarbon seep

communities in the Gulf of Mexico. Marine Ecology Progress

Series, 292, 51–60.

Milkov A., Vogt P., Crane K., Lein A., Sassen R., Cherkashev

G. (2004) Geological, geochemical, and microbial processes

at the hydrate-bearing Hakon Mosby Mud Volcano: a

review. Chemical Geology, 205, 347–366.

Niemann H., Losekann T., de Beer D., Elvert M., Nadalig T.,

Knittel K., Amann R., Sauter E.J., Schluter M., Klages M.,

Foucher J.P., Boetius A. (2006) Novel microbial communi-

ties of the Haakon Mosby mud volcano and their role as a

methane sink. Nature, 443, 854–858.

Olu K., Caprais J.-C., Galeron J., Causse R., Cosel R.v., Bud-

zinsky H., Le Menach K., Le Roux C., Levache A., Khripou-

noff A., Sibuet M. (2009) Influence of seep emission on the

non-symbiont-bearing fauna and vagrant species at an active

giant pockmark in the Gulf of Guinea (Congo–Angola

margin). Deep-Sea Research II, 56, 2380–2393.

Paull C.K., Ussler W., Holbrook W., Hill T.M., Keaten R.,

Mienert J., Haflidason H., Johnson J.E., Winters W.J.,

Lorenson T.D. (2008) Origin of pockmarks and chimney

structures on the flanks of the Storegga Slide, offshore

Norway. Geo-Marine Letters, 28, 43–51.

Peterson B.J., Fry B. (1987) Stable isotopes in ecosystem studies.

Annual Review of Ecology and Systematics, 18, 293–320.

R Development Core Team. (2008) R: A language and environ-

ment for statistical computing. R Foundation for Statistical

Computing, Vienna, Austria.

Sahling H., Galkin S.V., Salyuk A., Greinert J., Foerstel H.,

Piepenburg D., Suess E. (2003) Depth-related structure and

ecological significance of cold-seep communities – a case

study from the Sea of Okhotsk. Deep-Sea Research I, 50,

1391–1409.

Sauter E.J., Muyakshin S.I., Charlou J.-L., Schluter M., Boe-

tius A., Jerosch K., Damm E., Foucher J.-P., Klages M.

(2006) Methane discharge from a deep-sea submarine mud

volcano into the upper water column by gas hydrate-coated

methane bubbles. Earth and Planetary Science Letters, 243,

354–365.

Scrimgeour C.M., Robinson D. (2003) Stable isotope analysis

and applications. In: Smith K.A., Cresser M.S. (Eds), Soil

and Environmental Analysis: Modern Instrumental

Techniques. Marcel Dekker Inc., New York: 381–431.

Segonzac M., de Saint-Laurent M., Casacova B. (1993)

L’enigme du comportement trophique des crevettes

Alvinocarididae des sites hydrothermaux de la dorsale

medio-atlantique. Cahiers de Biologie Marine 34, 535–571.

Sibuet M., Olu-Le Roy K. (2002) Cold seep communities on

continental margins: structure and quantitative distribution

relative to geological and fluid venting patterns. In: Wefer

D.B.G., Hebbeln D., Jorgensen B.B., Van Weering T.

(Eds), Ocean Margin Systems. Springer Verlag, Berlin:

235–251.

Soltwedel T., Portnova D., Kolar I., Mokievsky V., Schewe I.

(2005) The small-sized benthic biota of the Hakon Mosby

Mud Volcano (SW Barents Sea slope). Journal of Marine

Systems, 55, 271.

Thurber A.R., Kroger K., Neira C., Wiklund H., Levin L.A.

(2010) Stable isotope signatures and methane use by

New Zealand cold seep benthos. Marine Geology, 272,

260–269.

Van Dover C., Aharon P., Bernhard J.M., Caylor E., Doerries

M., Flickinger W., Gilhooly W., Goffredi S.K., Knick K.,

Macko S.A., Rapoport S., Raulfs E.C., Ruppel C., Salerno J.,

Seitz R.D., Sen Gupta B.K., Shank T., Turneipseed M.,

Vrijenhoek R.C. (2003) Blake Ridge methane seep: charac-

terization of a soft-sediment, chemosynthetically based

ecosystem. Deep-Sea Research I, 50, 281–300.

Van Gaever S., Moodley L., de Beer D., Vanreusel A. (2006)

Meiobenthos at the Arctic Hakon Mosby Mud Volcano,

with a parental-caring nematode thriving in sulphide-rich

sediments. Marine Ecology Progress Series, 321, 143–155.

Van Gaever S., Olu K., Derycke S., Vanreusel A. (2009)

Metazoan meiofaunal communities at cold seeps along the

Norwegian margin: influence of habitat heterogeneity and

evidence for connection with shallow-water habitats.

Deep-Sea Research I, 56, 772–785.

Vanreusel A., Andersen A.C., Boetius A., Connelly D., Cunha

M.R., Decker C., Heeschen K., Hilario A., Kormas K.A.,

Maignien L., Olu K., Pachiadaki M., Ritt B., Rodrigues C.,

Sarrazin J., Tyler P.A., Van Gaever S., Vanneste H. (2009)

Biodiversity of cold seep ecosystems along the European

margins. Oceanography, 22, 118–135.

Vogt P.R., Cherkashev G., Ginsburg G., Ivanov G., Milkov A.,

Crane K., Lein A., Sundvor E., Pimenov N., Egorov A.

(1997) Haakon Mosby mud volcano provides unusual

example of venting. Eos, 78, 549–557.



Documents

Habitat heterogeneity influences cold-seep macrofaunal communities within and among seeps along the Norwegian margin – Part 2: contribution of chemosynthesis and nutritional patterns