Chemosynthesis influences food web and community structure in high-Arctic benthos

Abstract

Cold seeps are locations where seafloor communities are influenced by the seepage of methane and other reduced compounds from the seabed. We examined macro-infaunal benthos through community analysis and trophic structure using stable isotope analysis at 3 seep locations in the Barents Sea. These seeps were characterized by high densities of the chemosymbiotic polychaetes Siboglinidae, clade Frenulata (up to 32 120 ind. m−2), and thyasirid bivalves, Mendicula
cf. pygmaea (up to 4770 ind.m−2).We detected low δ13C signatures in chemosymbiotic polychaetes and in 3 species of omnivorous/ predatory polychaetes. These δ13C signatures indicate the input of
chemosynthesis-based carbon (CBC) into the food web. Applying a 2-source mixing model, we demonstrated that 28−41% of the nutrition of non-chemosymbiotic polychaetes originates from CBC. We also documented large community variations and small-scale variability within and among the investigated seeps, showing that the impact of seepage on faunal community structure transcends geographic boundaries within the Barents Sea. Moreover, aggregations of heterotrophicmacro- and megafauna associated with characteristic seep features (microbial mats, carbonate outcrops, and chemosymbiotic worm-tufts) add 3-dimensional structure and habitat complexity to the seafloor. Cold seeps contribute to the hydrocarbon-derived chemoautotrophy component of these ecosystems and to habitat complexity. These characteristics make the cold seeps of potential high ecological relevance in the functioning of the larger Arctic−Barents Sea ecosystem.

Full text

MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 629: 19–42, 2019
https://doi.org/10.3354/meps13101 Published October 24
1. INTRODUCTION
Offshore marine primary production originates
largely from phytoplankton in surface waters (Syvert-
sen 1991, Søreide et al. 2006). In the Arctic, primary
production is strongly linked to seasonal cycles and is
constrained by an intense growth period of short du-
ration during Arctic spring and summer (Wassmann
et al. 2006). In addition to phytoplankton production,
sea-ice algae can be an important marine photosyn-
© The authors 2019. Open Access under Creative Commons by
Attribution Licence. Use, distribution and reproduction are un -
restricted. Authors and original publication must be credited.
Publisher: Inter-Research · www.int-res.com
*Corresponding author: emmelie.k.astrom@uit.no
Chemosynthesis influences food web and
community structure in high-Arctic benthos
Emmelie K. L. Åström1,2,*, Michael L. Carroll1, 3, Arunima Sen1, 9,
Helge Niemann1,4,5, 6, William G. Ambrose Jr.3,7, Moritz F. Lehmann4, JoLynn Carroll3,8
1CAGE−Centre for Arctic Gas hydrate, Environment and Climate, Department of Geosciences,
UiT−The Arctic University of Norway, 9037 Tromsø, Norway
2Department of Arctic and Marine Biology, UiT−The Arctic University of Norway, 9037 Tromsø, Norway
3Akvaplan-niva, FRAM−High North Research Centre for Climate and the Environment, 9296 Tromsø, Norway
4Department of Environmental Sciences, University of Basel, Basel 4056, Switzerland
5NIOZ−Royal Netherlands Institute for Sea Research, Department of Marine Microbiology and Biogeochemistry,
and Utrecht University, 1790 AB Den Burg, Texel, The Netherlands
6Department of Earth Sciences, Faculty of Geosciences, Utrecht University, 3508 TC Utrecht, the Netherlands
7School of the Coastal Environment, Coastal Carolina University, Conway, South Carolina 29528, USA
8Department of Geosciences, UiT−The Arctic University of Norway, 9037 Tromsø, Norway
9Present address: Faculty of Bioscience and Aquaculture, Nord University, 8049 Bodø, Norway
ABSTRACT: Cold seeps are locations where seafloor communities are influenced by the seepage of
methane and other reduced compounds from the seabed. We examined macro-infaunal benthos
through community analysis and trophic structure using stable isotope analysis at 3 seep locations
in the Barents Sea. These seeps were characterized by high densities of the chemosymbiotic poly-
chaetes Siboglinidae, clade Frenulata (up to 32 120 ind. m−2), and thyasirid bivalves, Mendicula
cf. pygmaea (up to 4770 ind. m−2). We detected low δ13C signatures in chemosymbiotic polychaetes
and in 3 species of omnivorous/predatory polychaetes. These δ13C signatures indicate the input of
chemosynthesis-based carbon (CBC) into the food web. Applying a 2-source mixing model, we
demonstrated that 28−41% of the nutrition of non-chemosymbiotic polychaetes originates from
CBC. We also documented large community variations and small-scale variability within and
among the investigated seeps, showing that the impact of seepage on faunal community structure
transcends geographic boundaries within the Barents Sea. Moreover, aggregations of heterotro-
phic macro- and megafauna associated with characteristic seep features (microbial mats, carbonate
outcrops, and chemosymbiotic worm-tufts) add 3-dimensional structure and habitat complexity to
the seafloor. Cold seeps contribute to the hydrocarbon-derived chemoautotrophy component of
these ecosystems and to habitat complexity. These characteristics make the cold seeps of potential
high ecological relevance in the functioning of the larger Arctic−Barents Sea ecosystem.
KEY WORDS: Cold seeps · Benthos · Methane · Trophic structure · Stable isotopes · Barents Sea ·
Svalbard
O
PEN
PEN
A
CCESS
CCESS
This authors' personal copy may not be publicly or systematically copied or distributed, or posted on the Open Web,
except with written permission of the copyright holder(s). It may be distributed to interested individuals on request.

Mar Ecol Prog Ser 629: 19–42, 2019
thetic carbon source early in the spring (Syvertsen
1991, Hobson et al. 1995, Søreide et al. 2006). For ben-
thos and other organisms below the photic zone, these
communities are to a large extent reliant on the export
of surface-water primary production and partially de-
graded organic matter to the seabed i.e. pelagic−
benthic coupling (Graf 1989, Hobson et al. 1995, Re-
naud et al. 2008). In deep-sea systems, in situ dark
in orga nic carbon fixation via prokaryotes is also a sig-
nificant carbon source supporting benthic communi-
ties (Sweet man et al. 2019). In the Arctic, the large-
scale seasonal variations, sea-ice cover, and ice-algae
production in combination with sedimentation rates
and grazing pressure by zooplankton may create a
particularly strong cryo-pelagic− benthic coupling
that regulates the composition and function of seafloor
shelf-communities (Syvertsen 1991, Hobson et al.
1995, Wassmann et al. 2006). This conventional para-
digm of offshore marine productivity is based on car-
bon originating mainly from photosynthesis via bio -
synthesis in ocean surface waters. At cold seeps,
where hydrocarbons such as methane emanate from
the seafloor, chemosynthesis is an alternative mode of
carbon fixation (Brooks et al. 1987, Knittel & Boetius
2009, Zapata-Hernández et al. 2014).
At the base of cold seep-associated food webs are
specialized microbes that exploit methane and sul-
fide within anoxic sediments or within the tissues of
various organisms (Dubilier et al. 2008, Knittel &
Boetius 2009, Levin et al. 2016). Among the most
important are consortia of methanotrophic archaea
and sulfate-reducing bacteria that mediate the an -
aerobic oxidation of methane (AOM) coupled to
the reduction of sulfate. Free-living sulfur-oxidizing
(= thiotrophic) bacteria (SOB) can in turn utilize the
AOM end-product (sulfide) as an energy source for
carbon fixation, and mutualistic associations (i.e. sym -
bioses) occur between animals and aerobic methan-
otrophs and/or thiotrophic microbes. Energy is also
indirectly channeled into, and through, the seep
food web via trophic predator−prey interactions and
microbial grazing (Decker & Olu 2012, Niemann et
al. 2013, Zapata-Hernández et al. 2014). Such sym-
biotic relationships and trophic interactions can re -
sult in habitats consisting of specialized faunal com-
munities including typical chemosymbiotic cold seep
fauna (e.g. bathymodiolin mussels, vesicomyid bi -
valves, and siboglinid tubeworms) (Bergquist et al.
2005, Vanreusel et al. 2009, Levin et al. 2016).
The importance of chemosynthetic carbon (CBC) in
benthic food webs has been documented for cold
seeps globally (Levin 2005, Decker & Olu 2012, Zap-
ata-Hernández et al. 2014). Stable isotope analysis
has provided valuable insight into trophic interactions
and the utilization of different carbon sources (Brooks
et al. 1987, Ferrier-Pagès & Leal 2018). By combining
analyses of stable carbon (δ13C) and nitrogen (δ15N)
isotope ratios to assess energy sources and trophic
structure, it is possible to investigate community char-
acteristics and predator−prey interactions and to gen-
erate insights into resource utilization (Brooks et al.
1987, Hobson & Welch 1992, Vander Zanden et al.
1999). Chemosynthetically fixed carbon can have ei-
ther a 13C-enriched (δ13C of −9 to −16‰) or depleted
(δ13C < −35‰) carbon isotopic signature when com-
pared to photosynthetically fixed carbon (House et al.
2003, Ferrie-Pagès & Leal 2018). The lower (i.e. more
negative) δ13C values within this bimodal spectrum
for CBC can be attributed to the use of the enzyme
Rubisco I as catalyst in the carbon fixation step of the
Calvin Benson-Bassham cycle (CBB) (Robinson & Ca-
vanaugh 1995, Ferrier-Pagès & Leal 2018). The higher
(i.e. less negative) δ13C signatures can involve the Ru-
bisco II enzyme in the CBB cycle or the reductive tri-
carboxylic acid cycle (rTCA) for carbon fixation (Hü-
gler & Sievert 2011, Thiel et al. 2012). Methanotrophy
can result in even lower δ13C signatures than those as-
sociated with carbon fixation via CBB, and the degree
of the 13C depletion will depend on whether the source
of methane is either thermogenic (δ13C usually rang-
ing from −37 to −55‰) or microbial (δ13C≈−60 to
−80‰) (Brooks et al. 1987, Martens et al. 1991, Holler
et al. 2009). Finally, low δ13C-biomass isotopic signa-
tures may also arise from sequential fractionation, for
example, when autotrophic sulfur oxi dizers utilize
13C-de pleted CO2derived from me thane oxidation at
cold seeps (Lösekann et al. 2008). Such δ13C signatures
differ substantially from photosynthesized orga nic
car bon from phytoplankton (δ13C typically −20 to
−25‰) and ice algae (typically −15 to −20‰) (Hobson
et al. 1995, Søreide et al. 2006). Hence, due to differ-
ences in the carbon end-member isotope compositions
caused by different carbon fixation pathways, trophic
interactions, and carbon-source δ13C signatures, 13C/
12C ratios in biomass provide a valuable tracer of CBC
in marine food webs (Robinson & Cavan augh 1995,
Zapata-Hernández et al. 2014, Ferrier-Pagès & Leal
2018).
In addition to their importance in the context of
nutrient allocation for benthic communities, cold
seeps have other ecosystem functions and play an
im portant role in structuring faunal community com-
position. Seepage of fluids and gases in the sediment
creates strong geochemical gradients that influence
organisms on multiple spatial scales (Bergquist et al.
2005, Bowden et al. 2013, Levin et al. 2016) and mod-
20
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
ulates local seabed biodiversity and heterogeneity
(Vismann 1991, Sibuet & Olu 1998). At cold seeps,
chemosymbiotic energy sources can give rise to highly
specialized communities thriving in high abundances
and at high biomass (Sibuet & Olu 1998, Bergquist et
al. 2005, Vanreusel et al. 2009). Moreover, excess bi -
carbonate produced during AOM can precipitate and
form methane-derived authigenic carbonate pave-
ments (hereafter referred to as carbonate outcrops)
on the seafloor (Bohrmann et al. 1998). These carbon-
ate outcrops constitute important features, adding 3-
dimensional (3D) structure and enhancing faunal
habitat complexity with re spect to the surrounding
seafloor (Carney 1994, Levin et al. 2016, Åström et al.
2018).
Cold seeps occur in all the world’s oceans (Sibuet &
Olu 1998, Vanreusel et al. 2009, Levin et al. 2016),
including the Arctic (Decker et al. 2012, Paull et al.
2015, Savvichev et al. 2018). Natural methane seeps
have also been found in the Barents Sea, around the
Svalbard archipelago, where gas seepage likely re -
sults from the dissolution of hydrate reservoirs in the
sub-seabed caused by post-glacial isostatic re bound
from the last glacial maximum (LGM) (Portnov et al.
2016, Serov et al. 2017) and tectonic stress (Plaza-
Faverola & Keiding 2019). Several of these seeps
have been intensely studied with respect to geophys-
ical and geochemical aspects (e.g. Sahling et al.
2014, Andreassen et al. 2017, Serov et al. 2017), and
to a lesser extent, their ecological significance (Åström
et al. 2016, 2018, Sen et al. 2018a). However, the
trophic structure within these ecosystems is poorly
understood.
Studies on carbon cycling within Svalbard fjords
and the adjacent offshore shelves have demonstrated
the relative importance of carbon sources from terres-
trial origin versus phytoplankton, macroalgae, and
ice-algae contributions to the Arctic marine food web
(e.g. Søreide et al. 2006, Holding et al. 2017). The
contribution of CBC from cold seeps to high-Arctic
benthic food webs has scarcely been investigated
outside the area of the Håkon Mosby mud volcano
(HMMV) at the western border of the Barents Sea
slope (1200 m water depth) (e.g. Gebruk et al. 2003,
Decker & Olu 2012). To our knowledge, the impor-
tance of CBC in fueling benthic food webs has never
been assessed at cold seeps on the Barents Sea shelf.
Our goal was to examine to what extent CBC origi-
nating from methane sources at cold seeps in the Bar-
ents Sea is incorporated into benthic invertebrates in
these habitats. Furthermore, we investigated the
macro-infaunal community structure to compare the
communities at cold seep stations within the Barents
Sea with non-seep stations as references. Using com-
munity analysis of faunal composition as well as
stable isotope analyses, we evaluated the impact of
cold seeps and environmental variables on benthic
community structure and interactions at multiple
sites. The community analysis elucidated which dom-
inant faunal components drive observed community
structure patterns at the cold seeps, while the isotopic
measurements provided insights into the role of CBC
within the food web.
2. MATERIALS AND METHODS
2.1. Study areas
The Barents Sea is a continental shelf sea located
between the mainland of northern Norway and the
Svalbard archipelago. To the west, it borders the
Norwegian Sea (North Atlantic Ocean), and to the
east, it is separated from the Kara Sea by the island of
Novaja Zemlja (Fig. 1). The Barents Sea shelf is influ-
enced by both warm and saline Atlantic water from
the southwest, and cold Arctic water masses from the
northeast, forming the oceanic boundary known as
the polar front (Loeng 1991). The average water
depth in the Barents Sea is approximately 230 m, and
21
Fig. 1. Sites used in this study (circles), located in the west-
ern and central Barents Sea region. GHM: gas hydrate
mounds ~380 m deep, SR: Storfjordrenna seep ~350 m; BR:
Bjørnøyrenna crater field ~335 m. SR, GHM, and reference
non-seep control station collectively are referred to as ‘Stor-
fjord’; BR stations including the reference non-seep control
site are referred to as the ‘Crater area.’ Bathymetry from
IBACO v. 3.0 (Jakobsson et al. 2012)
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
bathymetry is characterized by numerous troughs,
banks, and post-glacial features including many
wedges and iceberg plough marks (Patton et al.
2017).
Cold seeps in which seabed emissions have been
found to originate from sub-seabed dissociating gas
hydrates have been documented at numerous loca-
tions along the western Svalbard margin. Indeed,
several of the seeps occur near the predicted upper
depth limit of the gas hydrate stability zone, resulting
in a release of methane from these gas reservoirs
(Sahling et al. 2014, Portnov et al. 2016, Serov et al.
2017). In the western Barents Sea, 2 such locations
have been identified in the outer part of Storfjord -
renna, with clear evidence for sub-surface gas de -
posits and hydrocarbon seepage (Fig. 1). The south-
ernmost location is known as the Storfjordrenna seep
field (SR). This site is characterized by a predomi-
nantly soft-bottom plain, where small carbonate out-
crops, chemosymbiotic polychaete colonies and
microbial mats have been observed at a water depth
of 350 m (Åström et al. 2016). Approximately 30 km
northwest of SR, a cluster of methane-seeping gas
hydrate mounds (GHMs) was recently discovered
(Serov et al. 2017). These rounded domes are a few
hundred meters in diameter and rise approximately
8−10 m from the seabed at water depths of around
380 m. Several of the GHMs contain gas hydrates in
near-surface sediments, display elevated concentra-
tions of dissolved methane in the sediment pore
water, and are actively emitting methane gas into the
water column (Hong et al. 2017, Serov et al. 2017).
Further to the east into the central Barents Sea, an
area with verified seabed seepage activity, craters,
and crater–mound complexes has been documented
in the Bjørnøyrenna trough (BR) (Fig. 1) (Solheim &
Elverhøi 1993, Andreassen et al. 2017). A few of
these seeping craters and adjacent areas have been
investigated for their macrobenthic community struc-
ture (Åström et al. 2016).
The above-described sites were the target of our
study. These cold seeps were all once covered by the
Barents Sea Ice Sheet (BSIS) during the LGM. Mod-
eling and paleo-oceanographic records suggest that
the outer SR became ice-free around 19 000 cal. yr BP
(Rasmussen et al. 2007, Serov et al. 2017), and that
the central part of the Barents Sea, the BR, was ice-
free around 15 000 yr BP (Andreassen et al. 2017, Pat-
ton et al. 2017). Evidence of dynamic paleo-ice
stream activity at the seabed, such as mega-scale
glacial lineations, indicate that the large seabed
structures associated with cold seeps in this study
(mounds and craters) were formed after the deglacia-
tion of the BSIS from the region and can be consid-
ered relatively young formations (at least on a geo-
logical time-scale: <15 000 yr BP) (Serov et al. 2017,
Andreassen et al. 2017).
We designated 2 main sampling regions within the
study area: the ‘Storfjord’ and the ‘Crater area’,
which refer to the name of the trough and the charac-
teristic depressions, respectively (Fig. 1, Table 1).
Storfjord consists of the sites around the SR seep field
(SR 1, SR 15) and the GHMs (GHM 2, GHM 3), as
well as paired non-seep reference stations (SR 2C
and GHM C). The Crater area refers to Stns BR 3, BR
15, BR 16, and a non-seep reference site (BR C)
located in the BR trough (see Table 1 for details).
2.2. Benthic sampling
We collected faunal and sediment samples from 3
verified locations of methane seepage (BR, GHM, SR)
(Fig. 1). Additionally, we sampled at non-seepage (i.e.
control) locations. All sampling was conducted from
the R/V ‘Helmer Hanssen’ during the Arctic spring/
summer seasons in 2014, 2015, and 2016 (Table 1). Lo-
cations of active hydrocarbon seepage were selected
based on previous surveys in the region (Solheim &
Elverhøi 1993, Åström et al. 2016, Hong et al. 2017),
and where acoustically detected gas/ bubble streams
(i.e. acoustic flares) were re corded with a single beam
echo sounder (Simrad EK60; frequencies 18 and
38 kHz). Vertical CTD hydro casts (SBE 9 plus sensor)
were performed at each station prior to sampling on
the seafloor. Benthic sampling for faunal community
analyses in 2014 was conducted where acoustic re-
flections from bubble streams where observed in the
water column (see detailed description in Åström et
al. 2016), using a van Veen grab (0.1 m2). During the
2015 survey, samples were collected in areas where
characteristic seep features, such as flares, microbial
mats, and carbonate outcrops, were identified through
seafloor imagery using a towed camera-guided multi-
corer with CTD sensor (see details in Åström et al.
2018). In 2016, seafloor imaging was carried out using
a remotely operated vehicle (ROV; Sperre 30K, www.
ntnu. edu/aur-lab/ rov30k) operated by the Norwegian
University for Science and Technology. Video record-
ing and stereo camera imaging (resolution: 1360 ×
1024 pixels) allowed pinpointing areas of active gas
ebullition and detailed investigation of the seafloor
structures including carbonate outcrops and microbial
mats (Sen et al. 2018a). In addition to visual investiga-
tions of the sea floor using the towed camera and ROV
in 2015 and 2016, we collected sediment samples for
22
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
quantitative faunal analyses with van Veen
grabs (0.1 m2). For the same purpose, we also
included one sample taken with a blade core
(0.018 m2) from the ROV in a tuft of chemosym-
biotic polychaetes in the vicinity (within meters)
of a bubble stream of seeping gas in order to
obtain a detailed overview of the faunal com -
position in this habitat (Table 1). In total, we
sampled 10 van Veen grab stations from desig-
nated seeps and non-seep control sites in
order to characterize macrofaunal communities
(≥500 µm) (Table 1). Five replicate grabs were
taken at each station for quantitative faunal
analysis, ex cept for at one station (SR 15) where
only 4 replicates were collected. We took an ad-
ditional grab sample at each station for the de-
termination of sediment characteristics includ-
ing poro sity, grain size, total organic carbon
(TOC), and benthic pigments (chlorophyll a
[chl a] and phaeo pigments [PhP]).
In addition to the above samples used for
community quantification, a number of faunal
samples were collected from different types of
sampling gear and devices such as grabs, sed-
iment cores, triangle dredges, and through
‘manual’ ROV-sampling, i.e. scooping with a
benthic ‘butterfly’ net. These samples were
used for bulk stable isotope analyses (δ13C
and δ15N) of benthic organism tissue (Hobson
& Welch 1992, Hobson et al. 1995, Søreide et
al. 2006) to assess food web and trophic level
(TL) interactions. Animals for isotopic analysis
were sorted and identified immediately after
collection, and for some taxa (primarily for
mud-dwelling echinoderms), we in cluded a
1−2 d de puration period in a dark cold-room
onboard the vessel in order to clear the guts
of ingested sediment. All organisms collected
for isotope measurements were stored frozen
(−20°C) prior to further processing and labo-
ratory analysis.
2.2.1. Macrofaunal community
characterization
Benthic samples for macrofaunal community
structure analysis were sieved immediately on -
board the vessel using a 500 µm mesh. Material
retained on the sieve was fixed in formaldehyde
(4%), mixed with rose-bengal for staining live
tissue, and the solution was buffered with borax
(sodium tetra-borate deca hydrate). In the labo-
23
Area Region Station Replicate Equipment Characteristics Date Latitude Longitude Salinity T Depth
sampled (°N) (° E) (psu) (°C) (m)
Bjørnøyrenna crater field Crater area BR 3 5 Grab 0.1 m2 Seep 15.07.2014 74° 54.09’ 27° 33.39’ 35.1 1.7 337
Bjørnøyrenna crater field Crater area BR 16 5 Grab 0.1 m2 Seep 30.06.2016 74° 54.11’ 27° 33.40’ 34.9 2.1 335
Bjørnøyrenna crater field Crater area BR 15 5 Grab 0.1 m2 Seep 26.05.2015 74° 54.07’ 27° 33.41’ 35.0 2.3 334
Storfjordrenna seep field Storfjord SR 1 5 Grab 0.1 m2 Seep 09.07.2014 75° 50.48’ 16°35.55’ 35.1 2.4 353
Storfjordrenna seep field Storfjord SR 15 4 Grab 0.1 m2 Seep 24.05.2015 75° 50.49’ 16°37.55’ 35.0 2.2 352
Storfjordrenna gas hydrate mounds Storfjord GHM 3 5 Grab 0.1 m2 Seep 24.06.2016 76° 06.41’ 15°57.88’ 34.9 2.0 383
Storfjordrenna gas hydrate mounds Storfjord GHM 2 5 Grab 0.1 m2 Seep 23.05.2015 76° 06.36’ 16°02.20’ 34.9 0.5 380
Bjørnøyrenna Crater area BR Ca 5 Grab 0.1 m2 Non-seep 18.07.2014 75° 08.97’ 28° 35.50’ 35.1 1.4 334
control for BR
Storfjordrenna Storfjord SR 2C 5 Grab 0.1 m2 Non-seep 10.07.2014 75°52.47’ 16°38.56’ 35.1 2.4 350
control for SR
Storfjordrenna Storfjord GHM C 5 Grab 0.1 m2 Non-seep 26.06.2016 76° 04.56’ 15°58.43’ 35.0 2.4 385
control for GHM
Storfjordrenna gas hydrate mounds Storfjord GHM 3b 1 Blade core Adjacent flares, 22.06.2016 76° 06.39’ 15° 58.15’ 34.9 2.1 381
0.018 m2 siboglinid tuft
aThis station is equal to PFT 16 in Åström et al. (2016)
bQuantitative sample for detailed community composition in a Siboglinidae worm tuft, not included in overall community analysis
Table 1. Sampling stations in the Barents Sea: site, station designation, sampling equipment, sampling date, location, and physical parameters (salinity, temperature [T],
and depth)
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
ratory, samples were sorted and identified to the low-
est possible taxon and stored in 80 % ethanol. This pro-
cedure followed theISO16665: 2014fieldwork protocol
to ensure consistency and quality control of benthic
faunal surveys. Organisms were first separated into
main phyletic groups: Crustacea, Echinodermata,
Mol lusca, Polychaeta, and ‘diverse’ (containing mem-
bers of the taxonomic groups Bra chio poda, Chordata,
Cnidaria, Hemichordtata, Nemer tea, Oligochaeta,
Platyhel min thes, Priapulida, and Sipuncula). Taxon-
specific counts were compiled after each individual
was identified to species or the lowest possible taxo-
nomic level. Individuals of each phyletic group were
weighed collectively to obtain aggregated wet weight
of the respective groups. True planktonic taxa were
excluded from the counts, as were Foraminifera and
Nematoda, which are not quantitatively retained on a
500 µm mesh. Colonial taxa such as Porifera, Bryo -
zoa, and Hydrozoa were excluded in the faunal-
abundance analysis, but their biomass was included
in the ‘diverse’ phyletic group.
2.2.2. Benthic pigment and sediment analysis
Sediment-bound chl aand PhP were quantified in
sediment samples as indicators of photosynthetically
derived organic material deposited on the seafloor.
Sediment chl aindicates relatively recently produced
material, whereas PhP represent the degradation
product of chl a. Surface-sediment pigment concen-
trations (upper 0−2 cm) from sediment samples were
analyzed by fluorometery (Holm-Hansen et al. 1965).
Sediments and filters for chl aand PhP samples were
extracted with acetone for 12−24 h, covered with alu-
minum foil and kept in a freezer. After extraction,
samples were centrifuged, decanted, and measured
for fluorescence using a Turner Design Model 10 AU
fluorometer before and after acidification with hydro -
chloric acid (1 M). The measured concentrations
were corrected for sediment porosity.
Porosity of sediment samples was determined
using a wet−dry method, where pre-weighed vials of
known volume were filled with sediment, weighed,
dried at 60°C until all water evaporated, and re-
weighed. Density of the sediment was calculated
using the wet weight and the total water-filled pore
space of the sample (Zaborska et al. 2008).
Sediment grain size (fraction of pelite <0.63 µm)
and TOC were determined by subsampling surface
(0−2 cm) sediments (minimum 50 g) from grab sam-
ples. Grain size was determined according to Bale &
Kenny (2005). For TOC, samples were dried and
homogenized and first, total carbon (TC) was deter-
mined. Subsequently, acid (HCl) was added to
another sample aliquot and total inorganic carbon
(TIC) was quantitatively re moved. After acidification,
samples were dried at 70°C. Analysis of both the TC
and TOC content were performed on a Shimadzu
SSM 5000 and Elementar Vario Cube, where the
samples were combusted at 950°C and the resulting
CO2was quantified using a flame ionization detector.
2.3. δ13C and δ15N analysis
For stable isotope analyses, smaller organisms
were processed as whole units, whereas specific tis-
sues were sampled from larger individuals (e.g. mus-
cle, tail, intestine, foot) depending on taxa. Samples
were freeze-dried for 24 h and the dried tissue was
homogenized and weighed into tin capsules, ~1.5 mg
tissue sample−1. For sediments, ~25 mg of freeze-
dried material was weighed into tin capsules. Partic-
ulate organic matter (POM) samples were derived
from seawater samples (2.5−3 l) filtered onto pre-
combusted membrane filters (0.22 µm, 4.7 cm; Merck
Millipore) and stored frozen prior to freeze-drying.
δ15N and δ13C composition of bulk material (tissue
and sediments) was measured using an elemental
ana lyzer coupled to an isotope ratio mass spectro -
meter (EA-IRMS; INTEGRA2; Sercon). The combus-
tion and reduction reactors of the instrument were
operated at 1050 and 600°C, respectively. Gas chro-
matographic separation of the combustion and reduc-
tion products (N2and CO2) was achieved on a packed
column (stainless steel, 50 cm, 1/4” o.d., Carbosieve
G, 60/80 mesh; Analytical Columns) at a temperature
of 55°C, with He as a carrier gas (55 ml min−1). The raw
nitrogen and carbon isotopic data were blank-, linear-
ity, and drift-corrected by means of 2-point calibra-
tions based on EDTA and IAEA-N-2 or EDTA and
IAEA-CH-6 standards, respectively. Iso topic compo-
sitions are reported in the conventional δ-notation, as
δ15N and δ13C in per mille relative to air and VPDB
(Vienna Pee Dee Belemnite), respectively. Repro-
ducibility based on duplicate analyses of samples, as
well as internal and external standards was better
than ±0.25‰ for δ15N and better than 0.1‰ for δ13C.
We calculated the TL of taxa to determine their
trophic position in the food web (i.e. first-order con-
sumer [grazer], second-order consumer [predator],
etc.). As a first step, we assumed photosynthetic POM
to be the sole carbon and nitrogen source in the tar-
geted areas. The POM δ13C and δ15N values deter-
mined in this study, integrated with published values
24
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
collected under similar conditions (Søreide et al.
2006, 2008), were used as a baseline isotopic signa-
ture for estimating the trophic position of organisms
in the food web (i.e. −25.1‰ for δ13C, with a range of
−23.5 to −27.3‰, and 4.0‰ for δ15N, with a range of
3.9−4.4‰). We assumed stepwise isotope fractiona-
tion values, with TLs of 3.4‰ for nitrogen and 0.6‰
for carbon (Hobson & Welch 1992, Søreide et al.
2006). Based on Søreide et al. (2006), the TL was then
calculated as:
TL = [(δ15NOrg. − δ15NPOM) / 3.4] + 1 (1)
where δ15NOrg. is the δ15N value of a given organism
and δ15NPOM is the TL baseline for the system. δ13C
values more negative than the baseline of −25.1‰
are indicative of the contribution from a 13C-depleted
carbon source (i.e. here considered a chemosyn-
thesis-based source). We used end-member calcula-
tions for the isotopic composition of potential food
sources to determine the relative contribution of
photo synthesis-based carbon vs. CBC at the targeted
seeps. The calculations were based on the results
from our stable isotope analyses and values from
Søreide et al. (2006, 2008). Hence, for organisms ex -
hibiting relatively low δ13C values, a 2-component
mixing equation was used to approximate the frac-
tion (Xfraction) of carbon that originated from the
assimilation of CBC originating from the seabed
seepage of hydrocarbons versus POM. The following
equation was applied:
Xfraction = (δ13COrg. − δ13CPOM) / (δ13CCBC − δ13CPOM) (2)
Here, analogous to Eq. (1), δ13COrg. is the δ13C value
of the biomass of a specific individual organism, and
δ13CPOM (−25.1‰) represents the baseline value for a
solely photosynthetic carbon source. δ13CCBC repre-
sents the δ13C value of chemosynthetic sources.
Few organisms can directly rely on methane as
an energy source i.e. AOM and microbial aerobic
methane-oxidation (MOx) mediating communities,
wherefore in sediment communities AOM is more
important (Reeburgh 2007, Knittel & Boetius 2009).
Given the carbon isotope fractionation (ε) that occurs
during AOM (ε≈12−38 ‰) (Holler et al. 2009), and
the δ13C of the methane itself, the AOM-biomass at
the investigated seeps (which can be incorporated in
the biomass of benthic organisms) will be markedly
depleted in 13C relative to photosynthetically derived
carbon sources. Methanotrophs, however, are cer-
tainly not the only microorganisms that can incorpo-
rate methane-derived carbon. Via AOM coupled to
sulfate reduction, hydrogen sulfide (H2S) and dis-
solved inorganic carbon (DIC), CBC becomes bio-
available in an ecosystem through SOB (Southward
et al. 1986, Brooks et al. 1987, Dubilier et al. 2008).
Hence, SOB can function as a carbon bridge between
chemical energy sources (in this study, seepage of
methane from the seeps) and fauna; and in this way,
various organisms inhabiting seeps can benefit from
chemosynthesis.
To estimate the contribution of CBC in this study,
we used different approaches: firstly, we used the
δ13C value of methane, δ13C = −47‰, sampled at the
GHM (Serov et al. 2017) as the key end-member. We
also provide calculations using the most negative
CBC end-member signal of −85‰, i.e. we assume
that the CBC is strongly fractionated and of AOM ori-
gin (CBC δ13C = −47‰, εup to −38‰; Holler et al.
2009). Finally, we made calculations based on the
least depleted CBC end-member, −35‰ (i.e. all car-
bon assumed to be derived from thio trophic symbi-
otic species, i.e. SOB). Here, we used a combination
of known values from the literature for SOB (South-
ward et al. 1986, Robinson & Cavanaugh 1995,
Decker & Olu 2012) (ranging be tween −27 and
−35‰). Furthermore, we integrated in our isotope
balance considerations the carbon isotopic composi-
tion of the siboglinids-reduced sediment, and bacter-
ial mat sediment samples from this study, where we
consider input of SOB. These abovementioned esti-
mates result in an average δ13C of SOB ~−35‰. For
our isotope balance model, we as sumed that carbon
fixation occurs primarily via the CBB-cycle (Robinson
& Cavanaugh 1995, Hügler & Sievert 2011). rTCA
has been shown to be a more common carbon fixa-
tion pathway at vents and is considered a minor con-
tributor to ecosystem carbon fixation in most seep
systems (Hügler & Sievert 2011, Thiel et al. 2012).
Estimates on the CBC (from both AOM and SOB) into
investigated taxa is presented in Table S1 in the Sup-
plement at www. int-res. com/ articles/ suppl/ m629 p019
_ supp. pdf.
In addition to assuming that POM derived from the
water column is the sole photosynthetic energy
source, we applied the same 2-component mixing
ap proach (Eq. 2) using a δ13C of −20.7 ‰ for sedimen-
tary organic matter (SOM), i.e. δ13CSOM instead of
δ13CPOM. This approach allows a refined estimate of
the extent to which different benthic organisms have
incorporated carbon from chemosynthetic sources
relative to SOM (which includes deposited, refrac-
tory organic material, and sediment-bound chloro-
phyll pigments) (Zapata-Hernández et al. 2014,
Alfaro-Lucas et al. 2018). If we consider that most of
the animals in this study were benthic (i.e. sediment
dwelling, deposit feeders or organisms preying on
25
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
deposit feeders), this approach examines the possible
contribution of CBC specifically to the benthos.
In summary, for the end-member calculations
(Eq. 2) of potential food sources in the system and to
determine the relative contribution of the chemical
energy carbon sources at the seeps, we considered
the following baseline values for carbon: CPOM
−25.1‰, CSOM −20.7‰, CCH4 −47‰, CSOB −35‰, and
Cfrac.AOM −85‰.
2.4. Statistical analysis
Diversity indices, including species richness (S),
Hurlbert rarefaction index (ES(100)), evenness (J’) and
Shannon-Wiener diversity (H’loge), were calculated
based on total faunal abundances (Hurlbert 1971). We
used faunal abundance to create a non-metric multi-
dimensional scaling plot (nMDS) and cluster ana lysis
based on the Bray-Curtis similarity matrix. Abundance
data were standardized and transformed (single
square-root) prior to analysis, in order to control for
the effect of large variability among replicates and to
balance highly abundant and rarer taxa. ANOSIM
was used to test the significance of the cluster results
(Clarke & Gorley 2006).
We calculated Pearson correlation coefficients (r) to
de termine individual pairwise relationships between
environmental variables and univariate faunal com-
munity data using Sigma Plot (v.12.5; Systat Soft-
ware). Moreover, we used multivariate analysis to de -
monstrate the influence of environmental variables
on overall faunal abundance at any given station, us-
ing the statistical computing program R. A canonical
correspondence analysis (CCA) with multi variate
constrained ordination was run to examine the ordi-
nation of station faunal abundance (we in cluded only
taxa that contributed to more than 1%
of the overall faunal abundance in the
entire data set) in relation to standard-
ized environmental variables (Ok sa -
nen et al. 2016). Environmental vari-
ables in cluded water depth, salinity,
bottom water temperature, and sedi-
ment characteristics such as grain size
(fraction <0.63 mm), TOC (%), and
ben thic pigments (total concentration
of chl aand PhP). In addition, we ad -
ded total faunal biomass and Sas ex-
planatory community parameters.
SIMPER analysis was used to test
similarities among taxon abundances
between clusters seen in the nMDS
plot. The clusters were grouped ac cording to overall
region, categorized either as Storfjord or Crater area,
with the exception of 5 samples that formed a sepa-
rate cluster. A refined control of the samples that
were grouped for respective regions in cluded sam-
ples both from non-seep controls as well as samples
collected at the seeps but without clear evidence of
strong impact from the seepage (hereafter classified
as ‘peripheral’ seep samples). The latter cluster in the
nMDS plot in cluded 5 separate samples that grouped
together regardless of region, and we categorized
them as ‘strong seep in fluence’. These samples were
all characterized by high abundance of chemosymbi-
otic siboglinids and displayed strong impact from the
hydrocarbon seepage (i.e. strongly reduced, black
and gassy sediments, and presence of carbonate
rocks). Furthermore, in the SIMPER analysis, we
tested the similarities on faunal abundances between
these strong seep-influenced samples and all re -
maining samples, categorized as ‘others’ (including
peripheral seep samples and non-seep controls).
These analyses were conducted with Primer v.6
(Clarke & Gorley 2006).
3. RESULTS
3.1. Community structure and environmental
influences
We identified 312 different taxa comprising nearly
40000 individuals from 10 grab stations and 49 repli-
cate samples in total. The overall average (±SE)
faunal abundance was 7936 ± 1172 ind. m−2, but was
highly variable among stations, ranging from 4575−
15 878 ind. m−2 (Fig. 2, Table 2). The 2 most numerous
taxa were siboglinid polychaetes (clade Frenu lata)
26
0
4000
8000
12 000
16 000
BR 3
BR 15
BR 16
BR C
GHM 2
GHM 3
GHM C
SR 1
SR 15
SR 2C
Abundance (m
–2
)
Station
Mendicula
Siboglinidae
Polychaeata (excl.Sibs.)
Mollusca (excl. Mend.)
Echinodermata
Diverse
Crustacea
Fig. 2. Faunal abundances aggregated by station. See Table 1 for station
information
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
and Mendicula cf. pygmaea bivalves (Fig. 3), con-
tributing 20.3 and 19.8%, respectively, to the total
faunal abundance. Their average faunal abundances
across all stations were 1607 ± 914 and 1569 ± 248 ind.
m−2, respectively. Siboglinids however, were absent
from 2 stations (SR 2C and BR C), whereas M. cf. pyg-
maea was present at all stations. At stations catego-
rized by strong seep-influenced replicates (GHM 3,
SR 1, and BR 3), the chemosymbiotic siboglinids and
M. cf. pygmaea bivalves encompassed 56.2% of total
faunal abundance (Table 3). At non-seepage control
stations (GHM C, SR 2C, and BR C), M. cf. pygmaea
was the most common species, comprising over 30%
of total abundance, followed by the polychaete Pri-
nospio cirrifera (Table 3). In one re plicate (van Veen
grab sample, 0.1 m2) from the GHM 3 seep, we
recorded extremely high abundances of siboglinid
polychaetes and M. cf. pygmaea (equivalent to 32120
and 4770 ind. m−2, respectively).
Benthic biomass was highly variable among sta-
tions (Table 2) (overall mean ± SE = 147.0 ± 30.4 g
wet weight [ww] m−2), with Polychaeta contributing
the most to the overall biomass (56.3%), followed by
Mollusca (22.7%). There were trends towards higher
faunal abundance and biomass at seeps compared to
non-seep control samples; however, these patterns
were not statistically significant owing to the high
variability of sampled individuals among replicates.
The nMDS and multivariate analysis based on total
faunal abundance at the replicate level showed 2 dif-
ferent clusters (R = 0.81, p < 0.001), separated by 36%
similarity (Fig. 4a). One cluster included 5 individual
replicates from Stns GHM 3, SR 1, and BR 3, while the
second cluster included all remaining replicates both
from non-seep paired controls and the peripheral seep
replicates, collected at the seeps but lacking obvious
signsof strongseepinfluence (as wede fined inSection
2.4). All replicates in the latter cluster were grouped
according to their station and overallsamplinglocation
(Fig. 4a). Furthermore, there was a distinct regional
separation (44% similarity), grouping replicates from
Storfjord and Crater area in addition to the separation
from the first cluster that in cluded the 5 strong seep-
influenced replicates from Stns GHM 3, SR 1, and BR 3
(Fig. 4a). These 3 clusters were significantly different
from each other (R = 0.87, p < 0.001). A refined hierar-
chical cluster analysis and dendrogram including only
the non-seep paired controls and the significantly dif-
ferent cluster of strong seep-influenced samples in
Fig. 4a further underscores that these 5 replicates
from GHM 3, SR 1, and BR 3 were distinctly different
with regards to their faunal composition compared to
other replicates (Fig. 4b).
In the CCA, the composite-station (across-stage
average) abundance, biomass, and Swere plotted
against the environmental variables TOC, grain size
fraction, temperature, benthic pigments, porosity,
salinity, and depth (Fig. 5). Three stations (GHM 3,
SR 1, and BR 3) were clearly separated from all other
stations along the first axis, which explained 48.7%
of the variability in the data (CCA I; Fig. 5). These
stations comprised similar environmental variables
and were also grouped by high abundances of sibo -
glinids, high S, and high biomass. The other stations,
including both reference non-seep controls and the
peripheral seeps, clustered together; however, these
stations were also separated by overall main region,
either belonging to Storfjord or Crater area along the
second axis (CCA II; Fig. 5), which ex plained 23.1%
of the ob served variability. The slightly shallower
27
Station Biomass Density Species Diversity Species ES(100) Pelite TOC Porosity Benthic Benthic
(g ww m−2) (ind. m−2) richness (%) evenness fraction (%) (0−2 cm) chl a phaeopigments
(no. station−1) J’ (%) (mg m−3) (mg m−3)
BR 3 130.3 12050 128 1.92 0.40 19.4 72.3 1.4 0.71a 2155.1a 8206.7a
BR 16 167.3 5714 123 3.19 0.66 33.9 57.1 2.0 0.67 1628.0 8746.4
BR 15 257.6 5264 126 3.33 0.69 35.6 90.4 2.0 0.71 2066.0 8497.9
SR 1 245.7 7684 125 3.31 0.69 32.2 88.5 1.6 0.66 1933.1 8056.5
SR 15 33.5 4584 99 3.36 0.73 33.2 84.8 1.6 0.66 782.5 5874.5
GHM 3 312.0 15878 145 2.55 0.51 25.2 76.1 1.8 0.62 2978.8 9744.5
GHM 2 77.7 7278 115 3.47 0.73 34.5 78.0 1.8 0.65 2582.0 10367.0
BR C 54.5 5040 120 3.03 0.63 32.3 86.0 1.9 0.62 2357.9 7115.1
SR 2C 64.8 5564 118 3.30 0.69 33.1 84.9 1.4 0.68 2141.6 8966.3
GHM C 126.6 10316 119 2.90 0.61 27.1 86.8 1.9 0.69 3146.9 10516.2
GHM 3bl 654.1 21833 45 2.64 0.69 − − 0.62 3012.7 11714.0
aUpdated and corrected porosity value from Åström et al. (2016)
Table 2. Faunal and environmental characteristics at each station; see Table 1 for station abbreviations. ES(100): Hurlbert index;
TOC: total organic carbon
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
28
Fig. 3. The 3 chemosymbiotic frenulates found in the study region: (A) possible Diplobrachia sp. (B) Oligobrachia sp. CPL-
clade (Sen et al. 2018b), and (C) possible Polybrachia sp. (note the tentacles with pinnules protruding out of its tube). (D)
Epibenthic foraminifera (Cibicidoides sp.) attached to an Oligobrachia sp. tube. (E) Mendicula cf. pygmaea bivalves occurred
at high abundances at all stations, regardless of seepage activity. These individuals were collected at the Bjørnøyrenna seeps.
Yellow, brownish color on the bivalve shell is encrusting material
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
Crater area stations differed from Storfjord stations
primarily in terms of environmental variables, such
as depth and sediment characteristics. The outcome
from statistical pairwise Pearson’s correlation coeffi-
cients between selected environmental variables
(TOC, grain size fraction, temperature, benthic pig-
ments, porosity, salinity, and depth) and faunal para -
meters (abundance, biomass, S, and J’) showed no
significant correlations.
A SIMPER analysis on replicate faunal abundances
between the clustered replicates (Fig. 4a) which sepa-
rated the 2 regions (Storfjord and Crater area) showed
a 56.3% dissimilarity in faunal commu-
nity composition. Three polychaete
species, Spio chae to pterus typicus,P.
cirrifera, and Galta tho wenia ocuelata,
explained most of this difference. This
pattern was also observed on the CCA
plot (Fig. 5), where stations in the Crater
area formed a cluster aligned with high
abundances of S. typicus, whereas Stor-
fjord stations clustered along high abun-
dances of P. cirrifera and G. ocuelata.A
similar SIMPER comparison of faunal
abundances be tween samples fromStns
GHM 3, SR 1, and BR 3 (which included
the strong seep influence samples) to
the remaining stations and samples in
the ‘others’ category showed a 64.5%
overall dissimilarity. Here, the sibogli-
nid polychaetes explained the largest
difference in faunal community compo-
sition between these 2 groups (10.2%).
In addition to the samples collected
for the regional analysis (above), the
ROV blade core at GHM 3 allowed for
detailed characterization of the faunal
composition inside one of the siboglinid
worm fields located in the immediate
vicinity of a methane gas flare at the
GHMs (Tables 2 & 3). The blade core
sample displayed by far the highest to-
tal faunal abundance (21 833 ind. m−2)
and biomass (654 g ww m−2) in this
study (Table 2). The blade core con-
tained high abundances of the chemo-
symbiotic siboglinids and M. cf. pyg-
maea bivalves, with densities of 7676
and 2500 ind. m−2, respectively. Over-
all, these 2 taxa alone contributed 35.1
and 11.5%, respectively, to the total
faunal abundance in the core (Table 3).
3.2. Stable isotope composition (δ13C and δ15N)
δ13C ratios among the studied taxa (Table 4) dis-
played a range of values between −16.2 and −47.1‰.
The highest (least negative) values were observed
for the starfish Solaster endeca (−16.2‰), whilst the
lowest values were observed for chemosymbiotic
sibo glinids (−47.1 ‰). The majority of the sampled
Barents Sea organisms in this study displayed δ13C-
values in the range between −24 and −16‰ (Table 4,
Fig. 6), suggesting a predominantly photosynthetic
29
Taxa No. ind. m−2 / Total faunal Cumulative
mean ind. m−2 (±SD) abundance (%) (%)
(A) Blade core GHM 3
Frenulata indet. 7667 35.1 35.1
Mendicula cf. pygmaea 2500 11.5 46.6
Cossura longocirrata 1389 6.4 53.0
Galathowenia oculata 1278 5.9 58.9
Yoldiella solidula 1111 5.1 64.0
Chaetozone sp. 1056 4.8 68.8
Aphelochaeta sp. 722 3.3 72.1
Polycirrus medusa 611 2.8 74.9
Maldane sarsi 556 2.5 77.4
Paradoneis lyra 500 2.3 79.7
Total: 79.7
(B) Seeps (GHM 3, SR 1 and BR 3)
Frenulata indet. 5187 (1062) 43.7 43.7
Mendicula cf. pygmaea 1485 (297) 12.5 56.2
Spiochaetopterus typicus 377 (182) 3.2 59.4
Yoldiella solidula 363 (166) 3.1 62.5
Galathowenia oculata 311 (101) 2.6 65.1
Prionospio cirrifera 277 (72) 2.3 67.4
Cossura longocirrata 269 (104) 2.3 69.7
Capitella capitata 229 (128) 1.9 71.6
Maldane sarsi 194 (45) 1.6 73.2
Levinsenia gracilis 170 (33) 1.4 74.6
Total: 74.6
(C) Non-seeps (GHM C, SR 2C, BR C)
Mendicula cf. pygmaea 2125 (343) 30.5 30.5
Prionospio cirrifera 408 (107) 5.9 36.4
Maldane sarsi 396 (33) 5.7 42.1
Yoldiella solidula 383 (128) 5.5 47.6
Nephasoma sp. 356 (206) 5.1 52.7
Galathowenia oculata 250 (73) 3.6 56.3
Spiochaetopterus typicus 195 (86) 2.8 59.1
Spiophanes kroyeri 177 (8) 2.5 61.6
Aphelochaeta sp. 155 (86) 2.2 63.8
Heteromastus filiformis 147 (21) 2.1 65.9
Total: 65.9
Table 3. Density of the top 10 most common taxa and their relative contribution
to total faunal abundance; see Table 1 for station abbreviations. (A) GHM 3
blade core from a worm tuft of chemosymbiotic frenulates, (B) ‘strong seep-in-
fluenced’ Stns GHM 3, SR 1, and BR 3, (C) non-seep control Stns GHM C, SR 2C
and BR C
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
source of carbon (Hobson & Welch 1992, Søreide et
al. 2006, 2008).
The food web baseline δ13C-values of photosyn-
thetically derived POM from the water column
ranged from −27.3 to −25.2‰, while SOM, at sites
without indication of microbial mats, reduced sedi-
ment patches, or free gas emissions, showed a base-
line between −21.3 and −20.5‰. A few sediment
samples showed different δ13C values than the SOM
baseline. For example, sediments collected with the
ROV blade core underneath a filamentous bacterial
mat exhibited a low δ13C signal (−25.3‰) compared
to sediment collected at a non-seep reference control
(−20.5‰) (Table 4). Despite the distinct (i.e. low)
δ13C-signature, sediment from the bacterial mat dis-
played δ15N values that were not markedly different
from other sediment samples (4.3 vs. 4.7‰, re -
spectively). The lowest (most nega-
tive) δ13C and δ15N signatures were re -
corded for dark-colored and re duced
sediment patches (δ13C = −34.5‰;
δ15N = 3.3 ‰). Sediment samples col-
lected from seeps without visual evi-
dence of bacterial mats or reduced
conditions displayed δ13C and δ15N
signals similar to the control non-seep
samples (δ13C = −20.6 to −20.9‰;
δ15N = 4.4−4.6‰).
The TL baseline, according to Eq. (1),
was represented by δ15N values from
POM sources of ~4‰ (Fig. 6, Table 4).
Moving up the food web, first-order
consumers of photosynthetically pro-
duced organic carbon (i.e. primarily
grazers at TL = 2.0) were Onisi mus sp.
amphipods (δ13C = −21.5‰, δ15N =
7.5‰; TL = 2.0), a gastropod, Hyalogy-
rina sp. (δ13C = −23.8‰, δ15N = 8.3‰;
TL = 2.3), and one unidentified species
of krill (Euphasidae) (δ13C = −23.2‰,
δ15N = 8.9‰; TL = 2.4) (Table 4). In ad-
dition, a few holo thurians (e.g. Molpa-
dia borealis) exhibited low δ15N values
that are typical for grazers, yet alto-
gether, we observed a relatively wide
range of δ15N values among individual
holothurians (δ15N = 5.8−11.3‰; mean
TL = 2.7). Most of the taxa in this survey
can be considered 2nd or 3rd order ben-
thic consumers at a TL ~3−4, i.e. typical
deposit feeders/ detrivores, predators,
and/ or scavengers. The starfish S. en-
deca (δ13C = −16.2‰, δ15N = 14.2‰;
TL = 4.0) and the amphipod Epimeria loricata (δ13C=
−19.8 ‰, δ15N = 14.4‰; TL = 4.1) wer e the top predators
we collected.
Siboglinid worms had distinctly low δ13C signa-
tures (δ13C = −38.2, to −47.1‰; Table 4, Fig. 6) in
accordance with a chemosynthesis-based lifestyle.
The siboglinid samples also exhibited the lowest δ15N
values among the secondary producers, ranging be -
tween −3.6 and 4.5‰.
Three different species of predatory/omnivorous
and non-chemosymbiotic polychaetes exhibited rela-
tively low δ13C values compared to the other hetero-
trophic organisms investigated in this survey (not in -
cluding the chemosymbiotic siboglinids). Individuals
of the carnivorous/omnivorous polychaetes Nephtys
sp. and Scoletoma fragilis, as well as the de posit
feeder Ophelina acuminata (Fauchald & Jumars
30
A
B
GHM C_1
GHM C_4
GHM C_ 2
GHM C_3
GHM C_5
SR2 C_2
SR2 C_5
SR2 C_4
SR2 C_3
SR2 C_1
SR1_2
SR1_5
SR1_3
SR1_4
SR1_1
GHM3_5
GHM3_2
GHM3_3
GHM3_4
GHM3_1
BR3_1
BR3_2
BR3_3
BR3_4
BR3_5
BR C_3
BR C_1
BR C_4
BR C_2
BR C_5
Smilarity
Replicates
25
50
75
100
Gas hydrate
mounds
Storfjord
seep-field
Crater area
SR
Control SR
Control G HM
Control BR
GHM 2
GHM 3
BR 3
Location
Similarity
36
44
2D stres s: 0.16
Standardise sam ples by total
Transform: square root
Resemblance Bray Curtis similarity
Standardise sam ples by total
Transform: square root
Resemblance Bray Curtis similarity
Fig. 4. Cluster analysis based on Bray-Curtis similarities of square-root trans-
formed faunal abundances. (A) Non-metric multidimensional scaling plot cal-
culated from all replicates. See Table 1 for detailed descriptions of locations.
(B) Dendrogram representing the cluster analysis as per replicate sample, at
Stns GHM 3, SR 1, and BR 3, and respective paired non-seep control replicates
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
1979, Cochrane et al. 2012) displayed δ13C values be -
tween −31.4 and −26.1‰, and δ15N values be tween
8.7 and 11.6‰. We also observed a large intra-
species variability in the δ13C signatures among
these taxa. Individual samples of nephtyids collected
within the same location (BR) exhibited a range of
almost 15‰ between extremes (δ13C = −31.8 vs.
−17.0‰). Similarly, nephtyids from SR as well as the
lumbrinerid S. fragilis collected from the GHM seeps
exhibited large intra-species differences of up to
~10‰ (e.g. δ13C = −25.5 vs. −16.8‰ for nephtyids;
δ13C = −29.1 vs. −18.9‰ for S. fragilis) between ex -
treme samples. Among the other taxa, no such large
variations were observed. Only the deposit feeding
starfish Ctenodiscus crispatus displayed a somewhat
elevated variability in its isotopic composition (δ13C =
−22.4 vs. −16.6 ‰).
The results from the 2-component mixing calcula-
tions using CPOM and CCBC as end-members for
photo synthetic vs. chemosynthetic (here
de fined as biosynthesis based on me -
thane) primary production sources at the
cold seeps indicated that tissues from
some invertebrates (not only chemosym-
biotic sibo glinids) reflected carbon from
chemical energy sources to varying de -
grees. We found individuals among the
nephtyids, lumbrinerids, and O. acumi-
nata with comparatively low δ13C signa-
tures. Further more, the calculated frac-
tion of incorporated CBC varied between
1.8 and 28.7% of their total carbon intake
(Table 4). Using the higher δ13C based on
SOM instead of POM (i.e. −20.7 vs.
−25.1‰) as a primary photosynthetic car-
bon source signature for the end-mem-
ber calculations yields an even larger
contribution of CBC for these poly-
chaetes (18.3−40.7%). Moreover, when
we applied δ13CSOM as the photosynthetic
end-member source, the carbon isotope
balance calculations indicated that sev-
eral other taxa, in addition to the sibogli-
nids and the abovementioned heterotro-
phic polychaetes, incorporate chemosyn-
thesized carbon to some extent (between
1.1 and 11.8%; Table 4). For the alterna-
tive ap proaches including SOB (δ13C =
−35‰) and the low carbon isotope signa-
ture for methane-derived carbon (δ13C =
−85‰) in the 2-component mixing model
respectively, we calculated a maximum
contribution of CBC ranging between
10.5 and 74.8% to the heterotrophic benthic commu-
nity (Table S1, Fig. S1).
4. DISCUSSION
We found high benthic biomass and high faunal
abundances at seep sites, despite large intra-station
variability. ROV-guided benthic sampling among
chemosymbiotic worm tufts within the seeps addi-
tionally revealed high population density of macro -
organisms, including high overall abundances of
chemosymbiotic siboglinid worms and Mendicula cf.
pygmaea bivalves. We documented a complex and
variable benthic seascape at seeps where habitat
heterogeneity was elevated across multiple spatial
scales by the presence of both biological and geolog-
ical seep-associated features (microbial mats, tufts of
chemosymbiotic tubeworms, and methane-derived
31
Fig. 5. Canonical correspondence analysis (CCA) based on station abun-
dances of taxa contributing to more than 1 % of the overall total faunal abun-
dance in the survey, ordinated with standardized environmental variables:
salinity, porosity, total benthic pigments, water depth, pelite fraction
(<0.63 µm), temperature, and total organic carbon in combination with ag-
gregated station species richness and biomass. See Table 1 for detailed de-
scriptions of station locations. Aphel: Aphelochaeta sp.; Cos_long: Cossura
longocirrata; Gal_fra: Gala tho wenia fragilis; Gal_oc: Galathowenia oculata;
Het_fil: Heteromastus fili formis; Leit_mam: Leitoscoloplos mammosus; Lev_
gra: Levinsenia gracilis; Lum_mix: Lumbrineris mixochaeta; Mal_sa: Mal-
dane sarsi; Mend: Mendicula cf. pygmaea; Neph: Nephasoma sp.; Pho_ass:
Pholoe assimilis; Sib:Siboglinidae (Frenu lata); Spi_kro: Spiophanes kroyeri;
Spi_ty: Spiochaetopterus typicus; Yol_sol: Yoldiella solidula
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
authigneic carbonates), which we suggest leads to
macro- and megafauna as well as smaller organisms
aggregating around the seeps (Figs. 3 & 7).
Furthermore, we demonstrated that carbon de -
rived from chemical energy sources originating from
seeping methane in the Barents Sea is incorporated
into the macrofaunal food web. We found 3 different
species of heterotrophic worms from active cold
seeps that possess low δ13C signatures in comparison
to other individuals within the same taxa from this
study (Table 4). The low δ13C signatures indicate that
chemosynthetic carbon contributes up to 40.7% to
32
Group Sample ‰ δ13C ±SD ‰ δ15N ± SD n TL FCBC /POM FCBC/SOM
(%) (%)
POM POM GHM C −27.3 N/A 3.9 N/A 1 0.9 N/A N/A
POM GHM3 −25.5 0.44 3.9 0.44 2 1.0 N/A N/A
Sediment H2S sediment −34.5 N/A 3.3 N/A 1 0.8 N/A N/A
BR sediment −20.9 0.56 4.4 0.32 2 1.1 N/A N/A
GHM sediment −20.6 0.12 4.6 0.44 4 1.2 N/A N/A
BR bacterial mat −25.3 N/A 4.3 N/A 1 1.1 N/A N/A
Annelida
Polychaetea Nephtys sp. BR −31.4 N/A 8.7 N/A 1 2.4 28.7 40.7
Nephtys sp. SR −25.5 N/A 11.6 N/A 1 3.2 1.8 18.3
Nephtys sp. GHM −23.2 N/A 11.0 N/A 1 3.1 − 9.5
Nephtys sp. −17.7 0.64 12.1 0.65 13 3.4 − −
Ophelina acuminata −26.1 N/A 9.1 N/A 1 2.5 4.5 20.5
Pherusa plumosa −20.0 N/A 8.8 N/A 1 2.4 − −
Polynoid −21.8 N/A 11.8 N/A 1 3.3 − 4.2
Scoletoma fragilis −19.1 0.58 12.1 0.36 3 3.4 − −
Scoletoma fragilis GHM −29.1 N/A 11.0 N/A 1 3.0 18.2 31.9
Siboglinid tissue BRa −38.3 N/A −3.7 N/A 3 − 60.3 66.9
Siboglinid tubes BR a −38.1 N/A −1.7 N/A 3 − 59.3 66.2
Siboglinid + tube SR −47.1 N/A 4.5 N/A 1 1.1 100.5 100.4
Mollusca
Bivalve Astarte crenata −21.2 0.75 10.2 0.55 6 2.0 − 1.9
Astarte elliptica −21.0 0.86 10.7 0.50 15 3.0 − 1.1
Bathyarca glacialis −20.5 0.80 9.9 0.72 24 2.7 − −
Chlamys islandica −21.3 N/A 9.8 N/A 1 2.7 − 2.3
Gastropoda Buccinum sp. −21.2 N/A 12.1 N/A 1 3.4 − 1.5
Hyalogyrina sp. −23.8 N/A 8.6 N/A 1 2.3 − 11.8
Sipunculid Sipunculida indet. –18.0 N/A 10.6 N/A 1 2.9 − −
Arthropoda
Crustacea Amhipoda (scav) −21.4 N/A 12.7 N/A 1 3.6 − 2.7
Onisimus sp. −21.5 0.49 7.5 0.70 5 2.0 − 3.0
Epimeira loricata −19.8 N/A 14.4 N/A 1 4.1 − −
Euphasidae −23.2 N/A 8.9 N/A 1 2.4 − 9.5
Lebbeus polaris −19.6 N/A 13.7 N/A 1 3.8 − −
Pandalus borealis −18.7 0.54 12.5 0.71 14 3.5 − −
Sabinea septemcarinata −18.1 0.39 13.3 0.12 2 3.6 − −
Nemertea Nemertea indet. −17.5 1.41 13.3 0.79 7 3.7 − −
Echinodermata
Asteroidea Ctenodiscus crispatus −20.3 1.64 10.2 0.41 13 2.8 − −
Solaster endeca −16.2 N/A 14.2 N/A 1 4.0 − −
Holothuridea Molpadia borealis −21.8 1.01 9.6 2.52 5 2.7 − 4.2
Chordata
Pisces Myoxcephalus scorpius −18.2 0.8 12.4 0.8 1b 3.5 − −
aSeparated samples, tissue from 3 individuals and their tubes; bOne individual sub-sampled from fin, intestines and muscle
Table 4. Mean (±SD) δ13C and δ15N of benthic organisms at the study sites (see Table 1 for location abbreviations in the ‘Sample’
column). n: number of individual samples; TL: calculated trophic level; F: potential contribution of chemosynthetically derived
carbon in relation to photosynthetic carbon sources from particulate organic matter (POM) or sedimentary organic matter
(SOM) using the methane end-member (CCH4 −47‰). N/A: not applicable
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
their diet. These results highlight the in corporation of
CBC in heterotrophic, non-chemosymbiotic taxa,
complementing the photosynthetic carbon sources
generally used by these benthic orga nisms (e.g.
Decker & Olu 2012, Zapata- Hernández et al. 2014).
4.1. Cold seep community structure and
regional differences
Our results show that chemosymbiotic siboglinids
represent key organisms and are highly abundant at
cold seeps in the Barents Sea together with the thya -
sirid bivalve M. cf. pygmaea. Siboglinids dominated
the species abundance at the methane seep sites.
The mean density of GHM worm tufts was ~7000 m−2
and where they mass-occurred forming tuft-aggre-
gations, they contributed the most to the total poly-
chaete biomass. Overall, we noted a tendency to -
wards higher faunal abundance (up to ~1.5−2.5×
higher) and biomass (~2.5−3.5× higher) at seep sta-
tions compared to the control non-seep stations
(Table 2). Yet the relationship was not statistically
significant because of the high variability among the
replicates and stations.
We observed a considerably higher total benthic
abundance (mean = 7936 vs. 4340 ind. m−2) and bio-
mass (147.0 vs. 65.5 g ww m−2) in this study com-
pared to nearby Barents Sea stations at similar
depths in areas without seafloor methane seepage, as
investigated by Carroll et al. (2008) using identical
sampling procedures. At Stn GHM 3, a particularly
high faunal abundance and biomass was observed,
not only with respect to individual stations without
seepage in the Barents Sea (Carroll et al. 2008, Coch -
rane et al. 2012), but also in comparison to much shal-
lower stations (including seep sites) on the western
Svalbard shelf (Åström et al. 2016). The single blade
core replicate at GHM 3bl (Tables 1 & 2) from one of
the siboglinid worm tufts displayed by far the highest
faunal abundance (21 833 ind. m−2) and biomass
(654 g ww m−2). This sample would have been diffi-
cult to obtain without the video-guided ROV system
used here, which provided a unique view into the
macroinfaunal community in the tuft where the top 3
most abundant species account for more than 50% of
the total abundance (Table 3). Similarly, at 3 grab sta-
tions (GHM 3, SR 1, and BR 3), siboglinids and M. cf.
pygmaea bivalves dominated the faunal abundance
(43.7 and 12.5%, respectively) and these 2 species
contributed to over 50% of the total faunal abun-
dances at these stations (Table 3). Moreover, M. cf.
pygmaea bivalves were also prominent at the control
non-seepage stations, with average density of 2125
ind. m−2, contributing to 30.5% of the total faunal
abundance at these reference sites.
The small (mm-sized) thyasirid bivalve M. cf. pyg-
maea was the most abundant taxa across all stations.
It was abundant both at cold seeps and non-seep
control stations. In general, thyasirids are known to
be abundant in highly reduced habitats, such as seep
habitats and organic-rich sediments (Taylor & Glover
2010); however, the high abundances of M. cf. pyg-
maea across all stations in this study is not readily
explained. Low δ13C signature in bivalve tissue has
been recognized for many thyasirid species (Dando &
Spiro 1993, Dufour 2005), suggesting uptake of CBC.
Yet this does not necessarily imply a fully chemosym-
biotic lifestyle for all species, as Family Thyasiridae
includes species with a wide range of different
dietary adaptations, from microbial syntrophy and
chemosymbiosis to mixotrophy and heterotrophy
(Dando & Spiro 1993, Dufour 2005, Taylor & Glover
2010). To date, there are no studies that provide
direct evidence that M. cf. pygmaea is associated
with chemosymbionts for nutritional dependence
(Oliver & Killeen 2002, Dufour 2005, Taylor & Glover
2010). We were unfortunately unable to analyze M.
cf. pygmaea tissues specifically for δ13C and δ15N,
despite their high abundances. The samples in which
M. cf. pygmaea were found were contaminated with
33
Fig. 6. Simplified bi-plot of δ13C and δ14N values of selected
organisms as well as sediment and particulate organic mat-
ter (SOM and POM) in the Barents Sea. Vertical and hori-
zontal error bars: ± SD. Vertical axis to the right indicates
calculated trophic levels from Table 4
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
formaldehyde and buffering reagents as they were
assigned for community analysis before species iden-
tification was undertaken. Thus, we can only specu-
late as to whether these thyasirids were able to incor-
porate chemosynthesized carbon and use CBC as an
energy source (Oliver & Killeen 2002, Dufour 2005).
The marked association of high abundances of M. cf.
pygmaea with the seeps suggests, however, that they
may also benefit directly or indirectly from the re -
duced seep habitat, similar to many other thyasirid
species.
The nMDS plot and cluster analysis (Fig. 4a,b) re -
vealed high variability in faunal community structure
at the seeps over relatively small spatial scales (i.e.
grab samples). Moreover, variability in community
structure and environmental conditions were ob -
served between stations over larger scales (region-
ally) in the CCA (Fig. 5). Our observations support
the pattern that cold seeps are commonly character-
ized by high macrofaunal abundances and biomass
(Gebruk et al. 2003, Bowden et al. 2013, Sen et al.
2018a) and are inhabited by distinct faunal commu-
nities (Sibuet & Olu 1998, Bergquist et al. 2005, Levin
2005). In the Bray-Curtis similarity analysis and the
nMDS plot (Fig. 4a), 5 individual replicates were sig-
nificantly different with respect to community struc-
ture when compared to the rest of the samples in this
study, regardless of where they were sampled (SR,
BR, or GHM). The other peripheral seep samples (not
including the abovementioned 5 replicates) clustered
with the paired controls (non-seep stations and repli-
cates), and were clearly separated (44% similarity)
by region (i.e. Storfjord or Crater area) (Fig. 4a).
These regional differences in overall faunal commu-
nity structure were driven mostly by variations in
abundance among the taxa Spiochaetopterus typicus
(Crater area) and Prionospio cirrifera and Gala tho -
wenia oculata (Storfjord), as tested by SIMPER.
These results demonstrate that the typical faunal
characteristics that appear to be intrinsic to stations
with a strong seep influence override the regionally
constrained faunal characteristics within the Barents
Sea. The seep-specific faunal characteristics, how-
ever, are not pervasive enough to override larger-
scale oceanographic regional differences in the fau-
nal community between the West Svalbard shelf and
the Barents Sea, as demonstrated in Åström et al.
(2016). Moreover, the CCA analysis supports these
findings and observed patterns (Fig. 5). It highlights
3 stations (GHM 3, BR 3, and SR 1) that are distinct
from the rest along the x-axis (CCA I), regardless of
region. The rest of the stations clustered together
based on region (as either Storfjord or Crater area),
and diverged along the y-axis (CCA II). Storfjord sta-
tions are influenced to a larger extent by relatively
high concentrations of sediment-bound chlorophyll
pigments and water column depth (Fig. 5), a some-
what counterintuitive result since sediment pigment
concentrations are expected to de crease with depth
(Gage & Tyler 1991, Renaud et al. 2008). We explain
this pattern by the location of the 2 main regions and
the relatively small differences in depth (Storfjord:
~350−380 m; Crater area: ~330 m) (Loeng 1991). Sta-
tions at the Crater area, located in the central Barents
Sea, are less influenced by the North Atlantic Cur-
rent (NAC) regime, and are more distant from coast-
line and nutrient fluvial transport from land which
could result in less pronounced surface productivity,
offsetting a systematic depth-to-chl aconcentration
relationship between these stations.
The separation of the 3 disconnected stations
(GHM 3, BR 3, and SR 1) along the x-axis was mainly
driven by differences in S, biomass, abundance of the
siboglinids, and, to a lesser extent, temperature. Even
though this analysis was based on composite stations
reflecting a broader spatial scale, it further confirms
that the ecological impacts of gas seepage seem to
override regional difference.
In summary, our comparative analysis, both with
regards to individual samples (Fig. 4a,b) and among
stations (Fig. 5), underlines that the characteristic
benthic community shaped by the influence from
seafloor seepage overrides faunal and biogeographi-
cal aspects within the Barents Sea. This implies that
strong seep-influenced samples collected at cold
seeps more than 300 km away from each other are
more similar with regards to community characteris-
tics than seep samples categorized as peripheral, col-
lected just a few meters away. Such similarities, how-
ever, are not shared between Western Svalbard and
the Barents Sea where, as documented in Åström et
al. (2016), regional large-scale oceanographic differ-
ences override benthic community patterns at the
cold seeps.
4.2. Seep-associated food webs: carbon sources
and trophic structure
The lowest δ13C signatures in this study (−38.3 and
−47.1‰) were recorded for chemosymbiotic worms
(frenulate siboglinids). These values are between 20
and 30‰ lighter than the δ13C composition of other
organisms tested in this study, which likely rely
mainly on POM/SOM carbon sources. Siboglinids
represent a family of polychaetes commonly associ-
34
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
ated with hydrothermal vents, cold seeps, organic
falls, or other reduced habitats. Adult siboglinids
(Frenu lata) lack a digestive tract and obtain most of
their nutrition from thiotrophy or methanotrophy
through endosymbiotic bacteria housed in a special-
ized organ called the trophosome (Southward et al.
1986, Schmaljohann et al. 1990, Rodrigues et al.
2011). Our stable isotope measurements clearly re -
flect this symbiotic lifestyle (Table 4, Fig. 6) (South-
ward et al. 1986).
With respect to Arctic siboglinids, published δ13C
signatures from Oligobrachia frenulates suggest the
involvement of methanotrophic symbionts because of
the characteristically low δ13C signatures observed
among them (~−50 to −62‰; Gebruk et al. 2003,
Decker & Olu 2012, Paull et al. 2015). Furthermore,
Savvichev et al. (2018) observed methanotrophs in
Oligobrachia frenualtes from the Laptev Sea in trans-
mission electron microscope (TEM) images. Other
studies targeting the symbionts of these animals
were, however, more equivocal. Unlike Savvichev et
al. (2018), Lösekann et al. (2008) and Sen et al. (2018b)
only observed SOB in TEM images of Oligo brachia
haakonmosbiensis from HMMV, and in Oligo brachia
sp. CPL-clade from GHM and the Crater area, respec-
tively. Additionally, they were un able to amplify se-
quences for enzymes associated with methane oxida-
tion, although they were able to amplify sequences
associated with sulfur oxidation. In both these studies
and species, bacterial 16S rRNA sequencing from
trophosome tissue revealed the presence of a group of
bacteria whose relationship to sulfur- and methane-
oxidizing symbionts could not be determined but
whose closest cultivated relative is a facultative sulfur
oxidizer (Lösekann et al. 2008, Sen et al. 2018b). The
reason for such an apparent discrepancy between in-
direct evidence by isotopic signatures and observable
endosymbionts could be related to the worms’ uptake
of DIC that is ultimately derived from AOM. The DIC
at these seeps display a 13C-depleted signature and its
uptake could account for the lower δ13C signature of
worm tissue than would be expected from a sulfide-
based lifestyle (Löse kann et al. 2008). Indeed, Sen et
al. (2019) found that O. haakonmosbiensis worms
from a North Atlantic seep-site (Lofoten) displayed
δ13C signatures that were similar to those of carbonate
outcrops around them, suggesting that worms and
carbonates alike build in 13C-depleted inorganic car-
bon from the same sedimentary carbon pool. Frenu-
lates are also known to supplement their symbiont-
derived nutrition with uptake of sediment organic
molecules (Dando et al. 2008). If organic carbon in the
sediment at these seep sites is already 13C-depleted,
their up take could serve as an additional potential ex-
planation for the low δ13C values observed among
Oligo brachia frenulates (Lösekann et al. 2008).
Interestingly, the bulk δ13C signatures we observed
in this study were higher than all previously pub-
lished values for frenulate siboglinids (Oligobrachia
spp.) from high-latitude seeps: we measured −38.3 ‰
at BR and −47.1‰ at SR, in comparison to −51.1 to
−56.1‰ measured at HMMV (Gebruk et al. 2003),
−62.1‰ from the Storegga seeps (Decker & Olu
2012), −52.1‰ at the North Atlantic Lofoten seep site
(Sen et al. 2019), and −55.0‰ in the Beaufort Sea
(Paull et al. 2015). The carbon isotopic differences
may be species-related or may be due to different en-
vironmental conditions and sedimentological re -
gimes, or both. O. haakonmosbiensis has been identi-
fied as the frenulate species being present at both
HMMV and the Lofoten seep-site (Lösekann et al.
2008, Sen et al. 2019). From frenulates sampled in the
Beaufort Sea, Paull et al. (2015) reported 97 % similar-
ity with O. haakonmosbiensis based on cyto chrome c
oxidase subunit I (COI) mitochondrial DNA se-
quences, and Decker & Olu (2012) described the
frenulates at Storegga seeps as O. cf. haakon mos -
biensis. Sen et al. (2018b) described a cryptic Oligo -
brachia species (CPL-clade) occurring in the Barents
Sea and the Laptev Sea. We could not identify sibo-
glinids (frenulates) collected in this study to species
level; however, Oligobrachia was also ob served
(Fig. 3B), likely belonging to the Oligo brachia sp.
CPL-clade (Sen et al. 2018b). Therefore, the differ-
ences between values we obtained in this study and
those published in other studies could be due to dif-
ferent species being targeted. This itself suggests that
Oligobrachia species might differ in their symbiotic
partnerships and endobacterial populations or con-
sortia: differing extents of reliance on one type of re-
duced chemical versus another could lead to varying
carbon isotopic signatures across different species
(Rodrigues et al. 2011, Sen et al. 2018b). Alternatively,
if sediment DIC and organic matter uptake strongly
influences the carbon isotopic signature of Arctic
seep siboglinids, as discussed above, another expla-
nation for the marked difference between our and
previous isotopic results could be that different envi-
ronmental/sedimentological conditions exists, and
that different inorganic carbon sources are prevalent
at these high-latitude seeps. Such variability would
almost certainly manifest itself in diverse carbon iso-
tope compositions in resident Oligobrachia worms.
Indeed, the differences we observed be tween the BR
and SR regions, which supposedly host the same
Oligobrachia species (Sen et al. 2018b), suggests that
35
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
local environmental conditions might constrain the
carbon isotopic signatures of Arctic seep Oligobrachia
worms. Taken together, these results and observa-
tions highlight the paucity of studies and lack of
knowledge with regards to the species composition
and ecology of high-latitude siboglinids.
While hosting microbial endosymbionts is a strat-
egy that makes direct use of reduced substances in
seep-associated sediments, another strategy em -
ployed by seep organisms relies on benefitting indi-
rectly from chemosynthetic food production through
microbial grazing. This feeding strategy has been ob -
served at hydrothermal vents, cold seeps, and whale
falls, and is typically performed by organisms such as
gastropods, crustaceans, and polychaetes (e.g. Van
Dover & Fry 1989, Braby et al. 2007, Niemann et al.
2013). In a ROV core sample from the Crater area, a
number of mm-sized gastropods identified as Hyalo -
gy rina sp. (A. Warén pers. comm.) were recovered
from microbial mats (Fig. 7G). The Hyalogyrina
genus has been observed at vents, cold seeps, and
other reducing habitats where they are usually asso-
ciated with microbial mats (Braby et al. 2007, Guillon
et al. 2017). Their high densities among microbial
mats suggest that Hyalogyrina gastropods are micro-
bial grazers, which has been reported from other
seeps (Guillon et al. 2017). The Hyalogyrina from our
analysis displayed signatures of δ13C = −23.8‰ and
δ15N = 8.6‰. This δ13C value is relatively high in
comparison with other suggested microbe-grazing
gastropods, for example rissoids from the HMMV
which displayed δ13C values as low as −46.6‰ in
microbial mats and −40.2‰ in the adjacent sediment
(Decker & Olu 2012). Our reported δ13C and δ15N val-
ues for Hyalogyrina, in relation to the signatures
from the sediment where it was found (δ13C =
−25.3‰ and δ15N = 4.3‰; Table 4) and its calculated
TL of 2.3, however, suggest that Hyalogyrina is a
first-order consumer that is likely grazing the sedi-
ments around the microbial mats. Furthermore, the
output from the end-member mixing calculation
(Eq. 2) revealed partial input of CBC to Hyalogyrina
gastropods (4.8−21.7%; Tables 4 & S1). These results,
along with our ROV observations of Hyalogyrina
aggregations in the reduced sediment and at the
microbial mats, suggest that the gastropods partly
carry out microbial grazing.
A few predatory polychaetes exhibited large intra-
species variability with regards to the putative assim-
ilated carbon sources (Fig. 6, Table 4). The observed
range of δ13C values for Nephtys sp. was −16.8 to
−31.4‰, and that of Scoletoma fragilis similarly var-
ied between −18.6 and −29.1‰. These stable isotope
signatures suggest a large niche width and/or an
opportunistic diet. Furthermore, the signatures sug-
gest that the polychaetes have a wide range of pre-
ferred prey with variable δ13C signatures, feeding
partly upon organisms that directly assimilate CBC.
Alternatively, the predatory polychaetes prey on
organisms that use carbon that has been originally
produced chemolithotrophically and shunted be -
tween organisms within the benthic food web. Our
results from the end-member mixing calculation
(Eq. 2), assuming input sources of POM (−25.1 ‰)
and CBC (based on methane δ13C = −47.0‰), suggest
that some polychaetes (i.e. polychaetes with the low-
est δ13C) derive up to nearly 30% of their carbon diet
from chemical energy sources (total range between
1.8 and 28.7%). Considering that none of these poly-
chaetes are ‘seep-endemic’ (Bergquist et al. 2005,
Levin et al. 2016) or known to form symbiotic associ-
ations with microbes that can utilize methane or sul-
fide, such an input of CBC is somewhat surprising,
although still realistic considering the lifestyle of
these worms (i.e. motile predators and omnivorous
feeders). Further studies are necessary to assess the
modes by which they incorporate CBC in their diets:
whether they prey on animals that are directly asso-
ciated with microbial symbionts, consume prey that
grazes on mats of methanotrophic/sulfide-oxidizing
microbes, or whether their prey acquires a low δ13C
isotope signature indirectly by incorporating chemo -
synthesized carbon that has been recycled within
the microbial loop. Our estimates of the contribution
of chemosynthesized carbon originating from cold
seeps vs. photosynthetically derived carbon from the
water column (i.e. POM) must be considered conser-
vative. SOM (with a higher δ13C than POM) is proba-
bly the most utilized photosynthetic carbon source
for shelf benthos. Using SOM as the photosynthetic
end-member in the model yields contributions of
18.3− 40.7% of CBC into these non-chemosymbiotic
polychaetes (Table 4).
In a first attempt to more quantitatively assess the
incorporation of seep-derived carbon (i.e. CBC) into
the benthic community, we used the isotopic signa-
ture of methane from the GHM site (δ13C = −47.0 ‰)
as the primary end-member for δ13C balance calcula-
tions (Table 4). By alternatively using end-member
δ13C values for SOB (−35‰) and fractionated AOM
sources (−85‰), we take into account the whole poten-
tial range of different carbon substrates and acknowl-
edge the uncertainty of the true end- member δ13C,
leading to a broader range of estimates on the incor-
poration of CBC to the heterotrophic taxa be tween
10.5 and 74.8% for the most extreme samples
36
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
(Table S1). Our analysis revealed that taxa with low
δ13C values partially incorporate CBC regardless of
the modeled scenarios using different end-member
carbon sources (Tables 4 & S1).
Our stable isotope analyses overlapped only par-
tially with all taxa included in the community struc-
ture analysis. We did not sample smaller macro -
organisms or meiofauna and we examined the stable
37
Fig. 7. Representative photographs of seep habitats at cold seeps in the Barents Sea. (A) Storfjordrenna seep field: carbonate
outcrops with hard bottom anemones and a red rockfish (Sebastus sp.). (B) Dense mat of chemosymbiotic frenulate worms col-
onized by filamentous bacteria. (C) Gas hydrate mound (GHM) seeps: pom-pom anemones (Liponema sp.) and shrimps (Pan-
dalus borealis) in a frenulate worm tuft. (D) A spotted wolf fish Anarhichas minor next to carbonate rocks colonized by
anemones, solitary corals, and hydrozoans. (E) Close-up of a frenulate worm tuft at the GHM seeps (white arrow); lighter col-
ored sediment to the right (circle) indicates patches of microbial mats. (F) Crater area: rock slabs and hard surfaces on a slope
into one of the craters colonized by various epifauna (sponges, starfishes, and anemones). (G) Remotely operated vehicle sam-
pling in a patch of a microbial mat surrounded by a field of frenulate worms partially covered with filamentous bacteria.
Hyalogyrina sp. snails were recovered from the mat. Scale bar: 20 cm
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
isotopic composition only for a few first-order con-
sumers, top predators, and vertebrates. Potential
inter mediate food chain links between the obligate
chemosymbiotic siboglinids, chemoautotrophic con-
sumers, and associated organisms occupying the
worm tufts may therefore have been overlooked in
the constructed food web (Figs. 6 & S1, Tables 4
& S1). Nevertheless, all our results support the con-
clusion that CBC originating from seeps contributes
to the nutrition of the Barents Sea cold seep fauna
over a whole range of taxonomic groups, and well
beyond the chemo sym bi otic siboglinids.
4.3. Cold seep heterogeneity:
habitat variability for benthic fauna
The cold seeps we investigated displayed high
abundance and biomass compared to habitats in the
Barents Sea where cold seeps are absent (Carroll et
al. 2008, Cochrane et al. 2012). Furthermore, the
seeps in this study (SR, GHM, and BR) displayed a
high degree of within-station seafloor heterogeneity
(Fig. 7), both in terms of biological and geological/
structural features. Seafloor methane seepage is
typically associated with strong biogeochemical
gradients (Gebruk et al. 2003, Bergquist et al. 2005,
Sen et al. 2018a) in the sediment which leads to
small-scale habitat heterogeneity, for example by
creating carbonate precipitate layers, and support-
ing microbial mats and aggregations of chemosym-
biotic fauna. ROV and photo transects conducted
over Storfjord and Crater area seeps revealed
mosaics with distinct patches of microbial mats and
tufts of siboglinids (Fig. 7). Such localized zonation
patterns have also been recognized at other cold
seeps (Bergquist et al. 2005, Decker & Olu 2012,
Bowden et al. 2013) and once more underscore the
importance of ‘visual’ sampling (i.e. ROV or camera-
guided grabs and cores). Using such combinations
of equipment and visual tools at the seeps will cer-
tainly lead to a better ecological understanding of
the communities at these multifaceted habitats (Jør-
gensen et al. 2011, Bicknell et al. 2016). Our obser-
vations of worm tufts forming distinct zones on the
seabed is likely controlled by the siboglinids’ pref-
erence for a habitat with access to optimal sulfide
concentrations and/ or sulfide fluxes (Sahling et al.
2003, Dando et al. 2008, Sen et al. 2018a). Sibogli-
nids (Frenulata) are thin (<1 mm in diameter), al -
though they can reach considerable length (50−
60 cm) and build chitinous tubes that ex tend several
10s of cm into the sediment. These tubes protrude a
few cm from the seabed, creating mats or tuft-like
structures on the seafloor (Fig. 7B,E), and in this
way, the worm tufts add small-scale 3D heterogene-
ity to the seep habitat. Within tuft samples, we found
many small benthic organisms using the tubes as
substrate (e.g. caprellid amphipods, mollusks, and
polychaetes) and ROV video images re vealed ag -
gregations of shrimps and amphi pods within the
worm tufts. Moreover, several siboglinid tubes were
covered with epibenthic fora minifera (e.g. Cibici-
doides) and overgrown by filamentous bacteria, pat-
terns that are known from other seeps (Levin 2005,
Sen et al. 2018b) (Figs. 3 & 7B,G). Even though sibo-
glinids within the clade Frenulata are small in com-
parison to other siboglinids (Vestamentifera) inhab-
iting seeps, such as Lamellibrachia and Escarpia,
they function as important ecosystem engineers, pro-
viding substrate and re sources for micro- and meio-
benthos, as observed in this study. They are also
likely to influence the seabed geochemistry by en -
hancing the turnover rates of biogeochemical pro-
cesses through the up take and release of oxygen,
sulfide, and other compounds in the sediment and
overlying water (Freytag et al. 2001, Dando et al.
2008, Sen et al. 2019).
In addition to the small benthic invertebrates utiliz-
ing the substrate of the worm tubes, we observed
larger megafauna such as pycnogonids, amphipods,
and gastropods, as well as various fishes aggregating
around these tufts. In ROV videos and tow-camera
images from the 3 cold seep locations (SR, GHM, and
BR), abundant megafauna appeared in close vicinity
to siboglinid worm-tufts, microbial mats, and/or
around the carbonate outcrops. Similar observations
were documented in Åström et al. (2018) and Sen et
al. (2018a), who reported increased biodiversity and
high abundance of epifaunal megafauna around car-
bonate outcrops at Arctic cold seeps compared to
adjacent areas without carbonate outcrops. The vari-
ability in the seafloor relief provided by carbonate
outcrops and other hard surfaces at the seeps
(Fig. 7A,C,D,F) could lead to megafaunal accu mu -
lations similar to those observed at reefs or other
complex seabed structures that can provide ample
substrate and protection against predators in pre-
dominantly soft-bottom habitats (e.g. Jensen et al.
1992, Meyer et al. 2014). In conclusion, the nature of
the multi-faceted and heterogeneous habitat at cold
seeps allows organisms to exploit a multitude of
niches within these habitats on various spatial scales,
where carbonate rock formations and worm tufts
provide 3D structure for meio-, macro-, and mega -
fauna (Figs. 3 & 7).
38
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
4.4. Arctic cold seep fauna
A large portion of the biomass at cold seeps is com-
posed of ‘background’ heterotrophic fauna (also
reported in Sen et al. 2018a). We did not observed the
dense fields of obligate chemosymbiotic organisms
and typical seep fauna (e.g. large-bodied mussels,
vesicomyid bivalves, and vestimentiferan worms)
that are known from other seep ecosystems mostly in
lower latitude regions (e.g. Sibuet & Olu 1998, Berg -
quist et al. 2005, Levin et al. 2016). Siboglinids were
the only chemosymbiotic taxa, together with a few
possible chemosymbiotic thyasirids (i.e. Thyasira
gouldi), that we were able to collect. This is in line
with previous observations of few living chemosym-
biotic cold seep taxa at other high-Arctic seeps (e.g.
Decker & Olu 2012, Paull et al. 2015, Åström et al.
2017). Similarly, there have been only a few observa-
tions of chemosymbiotic fauna at Arctic hydrother-
mal vents along the mid-Atlantic ridge (Schander et
al. 2010, Sweetman et al. 2013). Therefore, it appears
that at both Arctic seeps and vents, the majority of
the faunal community is composed of ‘background’
orga nisms, and true specialists (i.e. chemosymbiotic
species and heterotrophic taxa specialized for chemo -
syn thesis) are missing (Sweetman et al. 2013, Åström
et al. 2017, Sen et al. 2018a).
Despite the presence of few chemosymbiotic orga -
nisms at the investigated cold seeps, we re corded 3
different morphotypes of siboglinids (Frenu lata).
Possibly, these are Oligobrachia (likely CPL-clade
[Sen et al. 2018b], Diplobrachia sp., and Polybrachia
sp.; Fig. 3). The genus Oligobrachia has been ob -
served at all other Arctic and North Atlantic seep
sites studied to date (e.g. Gebruk et al. 2003, Savvi -
chev et al. 2018, Sen et al. 2018b) and is hypothesized
to be the most widespread siboglinid among seeps on
the Arctic shelf. The presence/absence of specific
extant frenulate species is likely influenced by envi-
ronmental characteristics such as substrate, depth,
and concentrations and fluxes of sulfide/ methane in
the sediment (Southward et al. 1986, Dando et al.
2008, Sen et al. 2018a). The scarcity of chemosymbi-
otic cold seep taxa at the Arctic shelf seeps could also
be limited by depth, cold bottom water temperatures,
and dispersal barriers to norther ly latitudes. For
example, it has been suggested that a low abundance
and diversity of symbiont-bearing species at shelf
seeps is linked to the shallow water depth of these
habitats (Schmaljohann et al. 1990, Sibuet & Olu
1998, Sahling et al. 2003). Species that are present at
shallower seeps are less likely to be seep-obligate
(‘seep-endemic’) (Berg quist et al. 2005, Dando 2010,
Levin et al. 2016) be cause a shallow depth usually
im plies a higher input of photosynthetic organic
material from surface primary production, thereby
mitigating the competitive advantage of relying on
chemosynthesis (Sahling et al. 2003, Tarasov et al.
2005). Moreover, it has also been proposed that high
levels of competition and predation at shallow depths
may largely preclude chemosynthetic fauna from
these habitats (Sahling et al. 2003). Another hypoth-
esis is linked to the dispersal by chemosymbiotic
fauna to the Arctic and limitations because of consis-
tent sub-zero bottom water temperatures, which has
been suggested to have shaped deep-sea chemosym-
biotic bivalve communities since the LGM and dur-
ing the Holo cene (Ambrose et al. 2015, Hansen et al.
2017, 2019). This is, however, less likely at the shal-
lower shelf seeps we studied because the bottom
water temperatures are comparatively warm due to
the inflow of relatively warm water to the SW Barents
Sea from the NAC (Loeng 1991). Our results revealed
that most of the organisms colonizing the Barents Sea
cold seeps are non-chemosymbiotic, heterotrophic
macrofauna forming a distinct faunal community
structure, characterized by only a few taxa. The cold
seeps we studied were located at comparatively shal-
low water depths (<400 m) in contrast to many other
well-studied cold seep systems worldwide (Sibuet &
Olu 1998, Vanreusel et al. 2009, Levin et al. 2016).
This shallow depth could possibly explain the high
subset of heterotrophic ‘background species’ coloniz-
ing these seeps, exploiting a wide variety of habitat
and food resources. Still, there is a gap in our knowl-
edge regarding the trophic interactions among fauna
inhabiting the Barents Sea seeps, and whether auto -
chthonous chemical energy sources are utilized by a
wider background benthic community remains un -
certain. Here, we have provided the first insights into
the interplay between different fauna and carbon
sources at these shelf cold seeps, and demonstrated
that at least for some selected heterotrophic taxa,
CBC plays a significant role in their diets.
5. SUMMARY
We have demonstrated that highly localized me -
thane seepage and environmental gradients at cold
seeps drive strong community-level effects that over-
ride regionally constrained faunal characteristics in
the Barents Sea. We reported overall high faunal
abundance and biomass at the seeps despite high
variability among individual samples. Chemosymbi-
otic siboglinid worms along with small thyasirids
39
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
(Mendicula cf. pygmaea) were the 2 most dominant
and characteristic taxa at the investigated seeps.
We documented low δ13C signatures in obligate
chemo symbiotic siboglinids as well as in 3 species of
heterotrophic polychaetes, which clearly indicates
the input of CBC in their diets. Our results further
demonstrate that other heterotrophic invertebrates at
these seeps may partially incorporate CBC indirectly.
Moreover, we observed aggregations of macro-
and megafauna around characteristic seep features
(chemosymbiotic worm tufts, microbial mats, and
carbonate outcrops), suggesting that the seep habitat
provides both substrate and autochthonous food re -
sources. In this way, cold seeps in the Barents Sea
function as biological hotspots where chemosyn-
thesis provides important supplementary carbon
sources to high-Arctic benthos.
Acknowledgements. This work was funded through the
Centre for Arctic Gas Hydrate, Environment and Climate
(CAGE) and the Research Council of Norway through its
Centers’ of Excellence funding scheme, #223259. We
acknowledge the captain and crew on board RV ‘Helmer
Hanssen’ and chief scientists and scientific teams from
CAGE 14_3, CAGE 15_2 and CAGE 16_5 cruises. Thanks to
Woods Hole Oceanographic Institute (WHOI) and MISO,
Daniel Fornari, for the collaboration developing the towed
camera system during the cruise in 2015, to the Norwegian
University of Science and Technology (NTNU), AMOS/
AUR-lab, Martin Ludvigsen and the ROV team for the col-
laboration during 2016 and to Antje Boetius and Frank Wenz -
höfer, MPI/AWI for the support with ROV blade cores.
Thanks to sorters and taxonomic specialists at the biological
laboratory at Akvaplan-niva, Tromsø, for processing sam-
ples and Matteus Lindgren at IG laboratory UiT. Thanks
also to Thomas Kuhn who conducted the isotope analyses at
the Department of Environmental Sciences at the University
of Basel. We are grateful for valuable discussions on taxon-
omy with Graham Oliver (National Museum of Wales),
Anders Warén (Swedish Museum of Natural History), Mag-
dalena Georgieva (Natural History Museum, London), and
Paul Dando (Marine Biological Association of the UK, Ply-
mouth), and to Friederike Gründger, Wei-Li Hong, and
Pavel Serov for valuable discussions and input in general.
Thanks also to Paul Renaud for discussions about trophic
interactions among Barents Sea fauna and to David Ham-
menstig for assistance with photo material in the article. The
seabed images are stored at the CAGE data repository and
more information is available by contacting the responsible
author or data manager at CAGE (https://cage.uit.no/).
J.C.’s contribution to this manuscript was funded by the
Research Council of Norway (RCN #228107). E.Å is cur-
rently post-doctoral scholar funded by VISTA — a basic re -
search program in collaboration between The Norwegian
Academy of Science and Letters and Equinor (#6172).
LITERATURE CITED
Alfaro-Lucas JM, Shimabukuro M, Ogata IV, Fujiwara Y,
Sumida PYG (2018) Trophic structure and chemosyn-
thesis contributions to heterotrophic fauna inhabiting an
abyssal whale carcass. Mar Ecol Prog Ser 596: 1−12
Ambrose WGJ, Panieri G, Schneider A, Plaza-Faverola A
and others (2015) Bivalve shell horizons in seafloor pock-
marks of the last glacial-interglacial transition: a thou-
sand years of methane emissions in the Arctic Ocean.
Geochem Geophys Geosyst 16: 4108−4129
Andreassen K, Hubbard A, Winsborrow M, Patton H and oth-
ers (2017) Massive blow-out craters formed by hydrate-
controlled methane expulsion from the Arctic seafloor.
Science 356: 948−953
Åström EKL, Carroll ML, Ambrose WGJ, Carroll J (2016)
Arctic cold seeps in marine methane hydrate environ-
ments: impacts on shelf macrobenthic community struc-
ture offshore Svalbard. Mar Ecol Prog Ser 552: 1−18
Åström EKL, Oliver PG, Carroll ML (2017) A new genus and
two new species of Thyasiridae associated with methane
seeps off Svalbard, Arctic Ocean. Mar Biol Res 13: 402−416
Åström EKL, Carroll ML, Ambrose WG, Sen A, Silyakova A,
Carroll J (2018) Methane cold seeps as biological oases in
the high-Arctic deep sea. Limnol Oceanogr 63: S209−S231
Bale A, Kenny A (2005) Sediment analysis and seabed char-
acterisation. In: Eleftheriou A, McIntyre A (eds) Methods
for the study of marine benthos. Blackwell Publishing,
Oxford, p 43−87
Bergquist DC, Fleckenstein C, Knisel J, Begley B, MacDon-
ald IR, Fisher CR (2005) Variations in seep mussel bed
communities along physical and chemical environmental
gradients. Mar Ecol Prog Ser 293: 99−108
Bicknell AWJ, Godley BJ, Sheehan EV, Votier SC, Witt MJ
(2016) Camera technology for monitoring marine biodi-
versity and human impact. Front Ecol Environ 14: 424−432
Bohrmann G, Greinert J, Suess E, Torres M (1998) Authi-
genic carbonates from the Cascadia subduction zone and
their relation to gas hydrate stability. Geology 26: 647−650
Bowden DA, Rowden AA, Thurber AR, Baco AR, Levin LA,
Smith CR (2013) Cold seep epifaunal communities on the
Hikurangi Margin, New Zealand: composition, succes-
sion, and vulnerability to human activities. PLOS ONE 8:
e76869
Braby CE, Rouse GW, Johnson SB, Jones WJ, Vrijenhoek RC
(2007) Bathymetric and temporal variation among Osedax
boneworms and associated megafauna on whale-falls in
Monterey Bay, California. Deep Sea Res I 54: 1773−1791
Brooks JM, Kennicutt MC, Fisher CR, Macko SA and others
(1987) Deep-sea hydrocarbon seep communities: evi-
dence for energy and nutritional carbon sources. Science
238: 1138−1142
Carney RS (1994) Consideration of the oasis analogy for
chemosynthetic communities at Gulf of Mexico hydro-
carbon vents. Geo-Mar Lett 14: 149−159
Carroll ML, Denisenko SG, Renaud PE, Ambrose WGJ
(2008) Benthic infauna of the seasonally ice-covered
western Barents Sea: patterns and relationships to envi-
ronmental forcing. Deep Sea Res II 55: 2340−2351
Clarke KR, Gorley RN (2006) PRIMER V6: user manual/tuto-
rial. PRIMER-E, Plymouth
Cochrane SKJ, Pearson TH, Greenacre M, Costelloe J,
Ellingsen IH, Dahle S, Gulliksen B (2012) Benthic fauna
and functional traits along a Polar Front transect in the
Barents Sea —advancing tools for ecosystem-scale assess-
ments. J Mar Syst 94: 204−217
Dando PR (2010) Biological communities at marine shallow-
water vent and seep sites. In: Kiel S (ed) The vent and
seep biota. Springer, Dordrecht, p 333−378
Dando PR, Spiro B (1993) Varying nutritional dependence of
the thyasirid bivalves Thyasira sarsi and T. equalis on
40
Author copy

Åström et al.: Seep food webs and high-Arctic benthos
chemoautotrophic symbiotic bacteria, demonstrated by
isotope ratios of tissue carbon and shell carbonate. Mar
Ecol Prog Ser 92: 151−158
Dando PR, Southward AJ, Southward EC, Lamont P, Harvey
R (2008) Interactions between sediment chemistry and
frenulate pogonophores (Annelida) in the north-east
Atlantic. Deep Sea Res I 55: 966−996
Decker C, Olu K (2012) Habitat heterogeneity influences
cold-seep macrofaunal communities within and among
seeps along the Norwegian margin — Part 2: Contribu-
tion of chemosynthesis and nutritional patterns. Mar Ecol
33: 231−245
Decker C, Morineaux M, Van Gaever S, Caprais JCC and
others (2012) Habitat heterogeneity influences cold-seep
macrofaunal communities within and among seeps along
the Norwegian margin. Part 1: Macrofaunal community
structure. Mar Ecol 33: 205−230
Dubilier N, Bergin C, Lott C (2008) Symbiotic diversity in
marine animals: the art of harnessing chemosynthesis.
Nat Rev Microbiol 6: 725−740
Dufour SC (2005) Gill anatomy and the evolution of symbio-
sis in the bivalve family Thyasiridae. Biol Bull (Woods
Hole) 208: 200−212
Fauchald K, Jumars PA (1979) The diet of worms: a study of
polychaete feeding guilds. Oceanogr Mar Biol Annu Rev
17: 193−284
Ferrier-Pagès C, Leal MC (2018) Stable isotopes as tracers of
trophic interactions in marine mutualistic symbioses.
Ecol Evol 9: 723−740
Freytag JK, Girguis PR, Bergquist DC, Andras JP, Childress
JJ, Fisher CR (2001) A paradox resolved: Sulfide acquisi-
tion by roots of seep tubeworms sustains net chemoauto-
trophy. Proc Natl Acad Sci USA 98: 13408−13413
Gage JD, Tyler PA (1991) Deep-sea biology: a natural his-
tory of organisms at the deep-sea floor. Part I. Cambridge
University Press, Cambridge
Gebruk AV, Krylova EM, Lein AY, Vinogradov GM and oth-
ers (2003) Methane seep community of the Håkon Mosby
mud volcano (the Norwegian Sea): composition and
trophic aspects. Sarsia 88: 394−403
Graf G (1989) Benthic-pelagic coupling in a deep-sea ben-
thic community. Nature 341: 437−439
Guillon E, Menot L, Decker C, Krylova E, Olu K (2017) The
vesicomyid bivalve habitat at cold seeps supports hetero-
geneous and dynamic macrofaunal assemblages. Deep
Sea Res I 120: 1−13
Hansen J, Hoff U, Sztybor K, Rasmussen TL (2017) Taxon-
omy and palaeoecology of two late Pleistocene species of
vesicomyid bivalves from cold methane seeps at Sval-
bard (79°N). J Molluscan Stud 83: 270−279
Hansen J, Ezat MM, Åström EKL, Rasmussen TL (2019) New
late Pleistocene species of Acharax from Arctic methane
seeps off Svalbard. J Syst Palaeontology, doi: 10. 1080/
14772019.2019.1594420
Hobson KA, Welch HE (1992) Determination of trophic rela-
tionships within a high Arctic marine food web using
δ13C and δ15N analysis. Mar Ecol Prog Ser 84: 9−18
Hobson KA, Ambrose WG Jr, Renaud PE (1995) Sources of
primary production, benthic-pelagic coupling, and trophic
relationships within the Northeast Water Polynya: insights
from δ13C and δ15N analysis. Mar Ecol Prog Ser 128: 1−10
Holding JM, Duarte CM, Delgado-Huertas A, Soetaert K
and others (2017) Autochthonous and allochthonous con-
tributions of organic carbon to microbial food webs in
Svalbard fjords. Limnol Oceanogr 62: 1307−1323
Holler T, Wegener G, Knittel K, Boetius A, Brunner B,
Kuypers MMM, Widdel F (2009) Substantial 13C/ 12C and
D/H fractionation during anaerobic oxidation of methane
by marine consortia enriched in vitro. Environ Microbiol
Rep 1: 370−376
Holm-Hansen O, Lorenzon C, Holmes R, Strickland J (1965)
Fluorometric determination of chlorophyll. ICES J Mar
Sci 30: 3−15
Hong WL, Torres ME, Carroll J, Crémière A, Panieri G, Yao
H, Serov P (2017) Seepage from an Arctic shallow marine
gas hydrate reservoir is insensitive to momentary ocean
warming. Nat Commun 8: 15745
House CH, Schopf JW, Stetter KO (2003) Carbon isotopic
fractionation by Archaeans and other thermophilic
prokaryotes. Org Geochem 34: 345−356
Hügler M, Sievert SM (2011) Beyond the Calvin Cycle: auto-
trophic carbon fixation in the ocean. Annu Rev Mar Sci 3:
261−289
Hurlbert SH (1971) The nonconcept of species diversity:
a critique and alternative parameters. Ecology 52:
577–586
Jakobsson M, Mayer L, Coakley B, Dowdeswell JA and oth-
ers (2012) The international bathymetric chart of the Arc-
tic Ocean (IBCAO) version 3.0. Geophys Res Lett 39:
L12609
Jensen P, Aagaard I, Burke RA Jr, Dando PR and others (1992)
‘Bubbling reefs’ in the Kattegat: submarine landscapes of
carbonate-cemented rocks support a diverse ecosystem at
methane seeps. Mar Ecol Prog Ser 83: 103−112
Jørgensen LL, Renaud PE, Cochrane SKJ, Lindal L, Renaud
PE, Cochrane SKJ (2011) Improving benthic monitoring
by combining trawl and grab surveys. Mar Pollut Bull 62:
1183−1190
Knittel K, Boetius A (2009) Anaerobic oxidation of methane:
progress with an unknown process. Annu Rev Microbiol
63: 311−334
Levin LA (2005) Ecology of cold seep sediments: interactions
of fauna with flow, chemistry and microbes. Oceanogr
Mar Biol Annu Rev 43: 1−46
Levin LA, Baco AR, Bowden D, Colaço A and others (2016)
Hydrothermal vents and methane seeps: rethinking the
sphere of influence. Front Mar Sci 3: 72
Loeng H (1991) Features of the physical oceanographic con-
ditions of the Barents Sea. Polar Res 10: 5−18
Lösekann 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). Environ
Microbiol 10: 3237−3254
Martens CS, Chanton JP, Paull CK (1991) Biogenic methane
from abyssal brine seeps at the base of the Florida
escarpment. Geology 19: 851−854
Meyer KS, Soltwedel T, Bergmann M (2014) High biodiver-
sity on a deep-water reef in the eastern Fram Strait.
PLOS ONE 9: e105424
Niemann H, Linke P, Knittel K, MacPherson E and others
(2013) Methane-carbon flow into the benthic food web at
cold seeps —a case study from the Costa Rica subduction
zone. PLOS ONE 8: e74894
Oksanen J, Blanchet F, Kindt R, Legendre P, O’Hara R (2016)
vegan: community ecology package. R Package version
2.3-3. https://CRAN.R-project.org/package= vegan
Oliver PG, Killeen IJ (2002) The Thyasiridae (Mollusca:
Bivalvia) of the British continental shelf and North Sea oil
fields: an identification manual. National Museums &
Galleries of Wales, Cardiff
Patton H, Hubbard A, Andreassen K, Auriac A and others
(2017) Deglaciation of the Eurasian ice sheet complex.
Quat Sci Rev 169: 148−172
41
Author copy

Mar Ecol Prog Ser 629: 19–42, 2019
Paull CK, Dallimore SR, Caress DW, Gwiazda R and others
(2015) Active mud volcanoes on the continental slope of
the Canadian Beaufort Sea. Geochem Geophys Geosyst
16: 3160−3181
Plaza-Faverola A, Keiding M (2019) Correlation between
tectonic stress regimes and methane seepage on the
western Svalbard margin. Solid Earth 10: 79−94
Portnov A, Vadakkepuliyambatta S, Mienert J, Hubbard A
(2016) Ice-sheet-driven methane storage and release in
the Arctic. Nat Commun 7: 10314
Rasmussen TL, Thomsen E, lubowska MA, Jessen S, Sol-
heim A, Koç N (2007) Paleoceanographic evolution of the
SW Svalbard margin (76° N) since 20000 14C yr BP. Quat
Res 67: 100−114
Reeburgh WS (2007) Oceanic methane biogeochemistry.
Chem Rev 107: 486−513
Renaud PE, Morata N, Carroll ML, Denisenko SG, Reigstad
M (2008) Pelagic−benthic coupling in the western Bar-
ents Sea: processes and time scales. Deep Sea Res II 55:
2372−2380
Robinson JJ, Cavanaugh CM (1995) Expression of form I
and form II Rubisco in chemoautotrophic symbioses:
implications for the interpretation of stable carbon iso-
tope values. Limnol Oceanogr 40: 1496−1502
Rodrigues CF, Hilário A, Cunha MR, Weightman AJ, Web-
ster G (2011) Microbial diversity in Frenulata (Siboglin-
idae, Polychaeta) species from mud volcanoes in the Gulf
of Cadiz (NE Atlantic). Antonie van Leeuwenhoek 100:
83−98
Sahling H, Galkin SV, 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 Res I 50:
1391−1409
Sahling H, Römer M, Pape T, Bergès B and others (2014) Gas
emissions at the continental margin west of Svalbard:
mapping, sampling, and quantification. Biogeosciences
11: 6029−6046
Savvichev AS, Kadnikov VV, Kravchishina MD, Galkin SV
and others (2018) Methane as an organic matter source
and the trophic basis of a Laptev Sea cold seep microbial
community. Geomicrobiol J 35: 411−423
Schander C, Rapp HT, Kongsrud JA, Bakken T and others
(2010) The fauna of hydrothermal vents on the Mohn
Ridge (North Atlantic). Mar Biol Res 6: 155−171
Schmaljohann R, Faber E, Whiticar MJ, Dando PR (1990)
Co-existence of methane- and sulphur-based endosym-
bioses between bacteria and invertebrates at a site in the
Skagerrak. Mar Ecol Prog Ser 61: 119−124
Sen A, Åström EKL, Hong WL, Portnov A and others (2018a)
Geophysical and geochemical controls on the megafau-
nal community of a high Arctic cold seep. Biogeosciences
15: 4533−4559
Sen A, Duperron S, Hourdez S, Piquet B and others (2018b)
Cryptic frenulates are the dominant chemosym-
biotrophic fauna at Arctic and high latitude Atlantic cold
seeps. PLOS ONE 13: e0209273
Sen A, Himmler T, Hong WL, Chitkara C and others (2019)
Atypical biological features of a new cold seep site on the
Lofoten-Vesterålen continental margin (northern Nor-
way). Sci Rep 9: 1762
Serov P, Vadakkepuliyambatta S, Mienert J, Patton H and
others (2017) Postglacial response of Arctic Ocean gas
hydrates to climatic amelioration. Proc Natl Acad Sci
USA 114: 6215−6220
Sibuet M, Olu K (1998) Biogeography, biodiversity and fluid
dependence of deep-sea cold-seep communities at active
and passive margins. Deep Sea Res II 45: 517−567
Solheim A, Elverhøi A (1993) Gas-related sea floor craters in
the Barents Sea. Geo-Mar Lett 13: 235−243
Søreide JE, Hop H, Carroll ML, Falk-Petersen S, Hegseth
EN (2006) Seasonal food web structures and sympagic−
pelagic coupling in the European Arctic revealed by sta-
ble isotopes and a two-source food web model. Prog
Oceanogr 71: 59−87
Søreide JE, Falk-Petersen S, Hegseth EN, Hop H, Carroll
ML, Hobson KA, Blachowiak-Samolyk K (2008) Seasonal
feeding strategies of Calanus in the high-Arctic Svalbard
region. Deep Sea Res II 55: 2225−2244
Southward AJ, Southward EC, Dando PR, Barrett RL, Ling R
(1986) Chemoautotrophic function of bacterial symbionts
in small pogonophora. J Mar Biol Assoc UK 66: 415
Sweetman A, Levin L, Rapp H, Schander C (2013) Faunal
trophic structure at hydrothermal vents on the southern
Mohn’s Ridge, Arctic Ocean. Mar Ecol Prog Ser 473:
115−131
Sweetman AK, Smith CR, Shulse CN, Maillot B and others
(2019) Key role of bacteria in the short-term cycling of
carbon at the abyssal seafloor in a low particulate
organic carbon flux region of the eastern Pacific Ocean.
Limnol Oceanogr 64: 694−713
Syvertsen EE (1991) Ice algae in the Barents Sea: types of
assemblages, origin, fate and role in the ice-edge phyto-
plankton bloom. Polar Res 10: 277−288
Tarasov VG, Gebruk AV, Mironov AN, Moskalev LI (2005)
Deep-sea and shallow-water hydrothermal vent commu-
nities: Two different phenomena? Chem Geol 224: 5−39
Taylor JD, Glover EA (2010) Chemosymbiotic bivalves. In:
Kiel S (ed) The vent and seep biota. Springer, Dordrecht,
p 107−135
Thiel V, Hügler M, Blümel M, Baumann HI and others (2012)
Widespread occurrence of two carbon fixation pathways
in tubeworm endosymbionts: lessons from hydrothermal
vent associated tubeworms from the Mediterranean Sea.
Front Microbiol 3: 423
Vander Zanden MJ, Shuter BJ, Lester N, Rasmussen JB
(1999) Patterns of food chain length in lakes: a stable iso-
tope study. Am Nat 154: 406−416
Van Dover CL, Fry B (1989) Stable isotopic compositions of
hydrothermal vent organisms. Mar Biol 102: 257−263
Vanreusel A, Andersen A, Boetius A, Connelly D and others
(2009) Biodiversity of cold seep ecosystems along the
European margins. Oceanography (Wash DC) 22: 110−127
Vismann B (1991) Sulfide tolerance: physiological mecha-
nisms and ecological implications. Ophelia 34: 1−27
Wassmann P, Reigstad M, Haug T, Rudels B and others
(2006) Food webs and carbon flux in the Barents Sea.
Prog Oceanogr 71: 232−287
Zaborska A, Carroll J, Papucci C, Torricelli L, Carroll ML,
Walkusz-Miotk J, Pempkowiak J (2008) Recent sediment
accumulation rates for the western margin of the Barents
Sea. Deep Sea Res II 55: 2352−2360
Zapata-Hernández G, Sellanes J, Thurber AR, Levin LA,
Chazalon F, Linke P (2014) New insights on the trophic
ecology of bathyal communities from the methane seep
area off Concepción, Chile (~36° S). Mar Ecol 35: 1−21
42
Editorial responsibility: James McClintock,
Birmingham, Alabama, USA
Submitted: March 25, 2019; Accepted: August 9, 2019
Proofs received from author(s): October 19, 2019
Author copy