field Pockmarks and methanogenic carbonates above the giant Troll gas ⁎

field Pockmarks and methanogenic carbonates above the giant Troll gas ⁎
Marine Geology 373 (2016) 26–38
Contents lists available at ScienceDirect
Marine Geology
journal homepage: www.elsevier.com/locate/margo
Pockmarks and methanogenic carbonates above the giant Troll gas field
in the Norwegian North Sea
A. Mazzini a,⁎, H.H. Svensen a, S. Planke a,b, C.F. Forsberg c, T.I. Tjelta d
a
CEED, University of Oslo, Oslo, Norway
VBPR, Oslo, Norway
NGI, Oslo, Norway
d
Statoil, Stavanger, Norway
b
c
a r t i c l e
i n f o
Article history:
Received 25 May 2015
Received in revised form 19 November 2015
Accepted 21 December 2015
Available online 28 December 2015
Keywords:
Norwegian North Sea
Troll
Methanogenic carbonates
Pockmarks
Gas hydrates
a b s t r a c t
Acoustic imaging has revealed more than 7000 pockmarks on the seafloor above the Troll East gas field in the
Norwegian North Sea. We present the first comprehensive study conducted on one of the World's largest
pockmark fields complementing the acoustic data with extensive sampling, geochemical and petrographical
studies. Specifically, we aimed at detecting possible active seepage still present over this vast area. The pockmarks
are present as isolated structures, on average ~35 m wide and up to 100 m in size. In addition, smaller satellite
pockmarks surround some of the pockmarks. In contrast to the muddy surroundings, parts of the investigated
pockmarks contain laterally extensive carbonate deposits or meter sized carbonate blocks. These blocks provide
shelter to abundant colonies of benthic megafauna. The carbonate blocks are comprised of micritic Mg-calcite
and calcite, micritic aragonite, and botryoidal aragonite. Framboidal pyrite is also commonly present. Carbon
isotopic values of the carbonates are 13C-depleted (δ13C as low as −59.7‰) and with δ18O up to 4.5‰, indicating
a methanogenic origin, possibly linked to gas hydrate dissociation. Pore water extracted from shallow cores from
the centre and the flanks of the pockmarks show similar Cl and SO4 profiles as the reference cores outside the
pockmarks, ruling out active methane seepage. This conclusion is also supported by seafloor video observations
that did not reveal any evidence of visual fluid seepage, and by the absence of microbial mats and by the fact that
the carbonate blocks are exposed on the seafloor and party oxidized on the surface. We conclude that methane
seepage formed this extensive gas field following to gas hydrate dissociation.
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
There are three main settings where pockmarks are commonly
present: 1) in offshore hydrocarbon provinces where fluids leaking
from reservoirs reach the seafloor; 2) in gas hydrate regions as the result
of ongoing or paleao dissociation of clathrates, and 3) in estuarine
and delta regions where the constantly deposited organic-rich
sediments or drowned wetlands trigger the production of shallow gas.
Comparative studies show that the fields with the highest density of
pockmarks are usually located at shallow depth associated with deltaic
and estuarine settings (e.g. Kelley et al., 1994; Rogers et al., 2006;
Brothers et al., 2012; Riboulot et al., 2013 and refs. therein). At these
localities the presence of microbial methane is mostly recorded as
opposed to thermogenic methane that is common at greater water
depths especially in hydrocarbon-rich provinces (e.g. Nickel et al.,
2013; Smith et al., 2014).
⁎ Corresponding author at: Centre for Earth Evolution and Dynamics (CEED) University
of Oslo, ZEB-bygningen Sem Saelandsvei 2A, Blindern 0371 Oslo Norway.
E-mail address: adriano.mazzini@geo.uio.no (A. Mazzini).
http://dx.doi.org/10.1016/j.margeo.2015.12.012
0025-3227/© 2015 Elsevier B.V. All rights reserved.
Former or ongoing methane seepage in pockmarks is commonly
coupled to anaerobic methane oxidation operated by microbial colonies
ions that bind with
of archea and bacteria. This reaction releases CO2−
3
Ca present in the seawater ultimately resulting in authigenic carbonate
precipitation (Valentine and Reeburgh, 2000; Boetius and Suess, 2004;
Boetius and Wenzhöfer, 2013). Methanogenic carbonates are indeed
common features at many pockmarks and seepage sites (Kocherla
et al., 2015; Magalhães et al., 2012; Hovland et al., 1987; Naehr et al.,
2000; Greinert et al., 2001; Gontharet et al., 2007; Akhmetzhanov
et al., 2008; Greinert et al., 2010; Haas et al., 2010).
The presence of pockmarks in the North and Norwegian Seas and all
the way north to the Barents Sea has been documented by several
authors (e.g. Hovland and Judd, 1988; Andreassen et al., 2000; Bouriak
et al., 2000; Bünz et al., 2003; Berndt et al., 2004; Hovland et al., 2005;
Mazzini et al., 2005, 2006; Forsberg et al., 2007; Paull et al., 2008;
Chand et al., 2012; Nickel et al., 2012). One of the largest pockmark
fields is located at Nyegga, in the southern part of the Vøring Plateau.
The first discovery of authigenic carbonates from this locality was reported by Mazzini et al. (2005). The first discovery of gas hydrates in
the area was reported more recently (Ivanov et al., 2007). Although
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
the features in the Nyegga region have always been described as
pockmarks, a large part of them are not morphological depressions
but rather positive structures containing large carbonate edifices in
the central area. In order to describe these “seeping positive structure”
Ivanov et al. (2010) coined the new term “seep mounds”.
Large pockmark fields represent a window into the subsurface
plumbing systems. They have been targeted for extensive geological
and geophysical explorations offshore Norway in order to investigate
the gas origin and if they are currently active or not. Active pockmarks
and mud volcanoes represent oasis of unique chemosynthetic life (Olu
et al., 1997; Sibuet and Olu, 1998; Menot et al., 2010; Ritt et al., 2011).
However, numerous questions remain unanswered in particular
for the pockmarks at northern latitudes. For example, is the seepage
entirely related to gas hydrate dissociation? Or could the leakage
from underlying gas and oil reservoirs represent the source of gas?
Was the gas release catastrophic or did the fluid migration occur at
27
slow mode? The climatic consequences of palaeo and modern seepgas release to the shallow ocean and atmosphere is still an open
question (Westbrook et al., 2009; Smith et al., 2014) and the
discovery of new large pockmark fields is of relevance for future
studies and global estimates.
In this paper we document the structure and nature of a giant
pockmark field in the Norwegian North Sea (Fig. 1) where large carbonate blocks have been identified and sampled. These results are part of a
large Statoil-funded project aiming at investigating gas migration
around the Troll A platform. The aim of the paper is to describe the
characteristics of this new discovery and to determine if active seepage
is occurring in the region based on a large sampling campaign inside
and outside the pockmarks. This issue has important implications for
the fluid flow and hazard aspects of the area. If the pockmarks are
actively leaking, there is a potential risk for formation of new pockmarks
with a possible influence on platform stability. The second important
Fig. 1. (A) Inset map of Norwegian channel and North Sea, framed the Troll study area. (B) Detail of the Troll East gas field (yellow) covered with an extensive multibeam survey (purple
line). (C) Troll field multibeam data. Note the large number of pockmarks (about 7500 in total). UTM Zone 31, WGS84 datum. Indicated the Troll A platform and the location of the three
main ROV pockmark study areas: in red (Septagram), green (Peanut), and blue (Arch). (For interpretation of the references to color in this figure legend, the reader is referred to the web
version of this article.)
28
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
aim is to find evidence of former seep activity and hydrocarbon flow
through the sea floor via pockmarks.
2. Troll A shallow gas migration project
The main objective of the Statoil “Troll A shallow gas migration
project” was to understand the cause of high pressures in the outer
annuli of the wells and in the foundation skirt compartments of the
Troll A platform (Tjelta et al., 2007). The Toll A platform is a large
concrete gravity base structure located in the Norwegian trench. High
pressures have to be bleed-off routinely since the installation of the
platform in 1995, typically once a month for the well annuli and more
irregularly for the foundation skirt compartments. This routine procedure raised a number of questions about the potential for natural gas
seepage in the Troll area, in particular if the numerous pockmarks in
the region could be sites for active focused fluid flow and thereby
representing a potential hazard. One of the main goals of the project
was therefore to search for evidence of active fluid seepage.
3. Geological setting
The giant Troll oil and gas field are located in the northern North
Sea, about 50 km west of Norway, in the 300 m deep Norwegian
Channel. The field extends over more than 700 km2, and is situated
on the Horda Platform between the northern Viking Graben to the
west and the Øygarden Fault Zone to the east. The reservoir consists
of faulted Jurassic sandstones of the Viking Group about 1.5 km
below the sea floor.
Two major crustal extensional events, a Permo-Triassic and a late
Jurassic–early Cretaceous event, lead to the formation of the northern
North Sea rift system (Christiansson et al., 2000; Faleide et al., 2010;
Holgate et al., 2013). The rift basin was subsequently infilled with late
Cretaceous and Paleogene sediments during post-rift subsidence.
Southern Norway and the eastern flank of the North Sea were
subsequently uplifted and eroded, in particular from mid-Miocene
times. Repeated ice stream activity during the Pleistocene formed the
Norwegian Channel (Sejrup et al., 2003).
The stratigraphy of the Horda Platform is dominated by Mesozoic
and Cenozoic clastic sediments above a Paleozoic basement. Two
major angular unconformities are apparent in seismic reflection
data — the Base Cretaceous Unconformity (BCU) and the Upper Regional
Unconformity (URU) (Faleide et al., 2010). Sub-horizontal Quaternary
glacial and marine sediments, and a few tens of meters thick sequence
of post-glacial marine sediments, overlie the URU.
4. Methods
4.1. Marine expeditions and acoustic survey
After an initial multibeam survey to cover a large area around the
Troll field (~ 15,000 km2), four different main ROV cruises were
conducted in 2005 and 2006 to map in more detail the structure of six
complex pockmarks (Fig. 1). Initial surveying took place in March
2005 in the Septagram region using the vessel Edda Fonn. This survey
was followed up in July 2006 and October 2006 in the Septagram and
Peanut regions, and in December 2006 in the Arch region, using R/V
Normand Tonjer, Edda Fonn, and Normand Tonjer, respectively.
The ROV survey equipment for the two survey vessels were similar,
providing high-quality bathymetry (the DTM bin size is 0.5 by 0.5 m
for all except the high-resolution grid of Septagram which is 0.1 by
0.1 m), side-scan sonar (325 kHz), and sub-bottom profiler (4.5 to
10 kHz) data. Typically, a constant flight height of 10–15 m above
the regional sea floor was used. Video record and high resolution
image stills were collected during each ROV dive in order to document the presence of fluid activity, chemosyinthetic fauna, and possible authigenic carbonates.
4.2. Sampling
Sixteen ~ 6 m long cores and one 12 m core have been collected in
the Septagram region (i.e. from the Septagram pockmark and two
buried pockmarks) during two separate expeditions in 2005 and 2006.
In addition, 28 short (0.5 m) ROV cores were taken (19 in the Septagram
region, 4 in the Peanut region, 3 in Arch D, and 2 in Arch E). A collection
of biological samples and carbonate blocks was sampled using an ROV
grabber at the pockmark sites (Fig. A1, Tables A1–A2).
4.3. Petrography
Thin sections of the carbonate samples were studied using both optical and electronic microscopes, including UV light, secondary electron
and backscattered electron modes. A set of XRD analyses was conducted
on carbonate fractions in order to identify the carbonate phases.
A total of 42 individual mineral analyses were performed with
a CAMECA SX 100 electron microprobe (EMP) in order to determine the composition of the authigenic carbonate. The results
are presented in Table A3, both as weight percent oxides and the
calculated structural formulas.
3D CT scanning was conducted on one of the carbonate blocks in
order to determine the internal structure. Images were acquired using
a GE Lightspeed Ultra CT 99_OC0 spiral computed tomography scanner,
with an 8 sensor system (Majorstuen Røntgen Capio, Oslo, Norway).
The thickness of the scanned slices was 0.625 mm, and 247 scans
were made throughout the sample.
4.4. Geochemistry
Carbon, oxygen, and strontium stable isotopic analyses were done at
the Institute for Energy Technology (IFE, Norway) on both shell
fragments incorporated in the carbonate blocks and the carbonate
cements from five different blocks from four pockmarks. The carbonate
cements were ground and digested with a 0.1 ml 100% H3PO4 solution
for 2 h at 30.0 °C in a vacuumed environment. The released CO2 was
transferred to a Finnigan MAT DeltaXP isotope ratio mass spectrometer
(IRMS), for determination of δ13C and δ18O. The analyses were
controlled by in-house standards of calcite and aragonite. Results are
reported in ‰ relative to the PDB standard. The precision for δ13C is
± 0.1‰ and for δ18O ± 0.2‰. Three bulk crushed carbonate samples
(10–20 mg) were used for Sr isotope analysis. The isotopic composition
of Sr was determined by thermal ionization mass spectrometry (TIMS)
on a Finnigan MAT 261.
Pore water extraction was done on thirteen cores and the main
cation and anion concentration was analyzed on a total of 140 pore
water samples. Isotope analyses were measured on 23 selected water
samples. The extracted pore waters were analyzed for Na, Cl, SO4, Ca,
Br and K at the Dept. of Geosciences at the University of Oslo. Cations
were analyzed on a Varian 300 flame atomic absorption spectroscope,
whereas the anions were analyzed on a Dionex QIC analyzer ion
chromatograph with a Ionpac AS4A-SC column. Cation and anion analyses were performed on highly diluted samples, thus the uncertainty
of the results is in the order of 5–10%. Oxygen, carbon, and hydrogen
isotope water analyses were performed at IFE. The water δ18O and
δ13C of the dissolved inorganic carbon (DIC) composition was measured
using a using a Finnigan DeltaXP isotope mass spectrometer, and is
reported as permil VSMOW. The δD composition was determined by a
Micromass Optima isotope mass spectrometer.
TOC and TIC analyses were done on dried bulk clay samples from
the cores in order to determine differences in the sedimentary
environments inside and outside the pockmarks. The analyses were
done at the Dept. of Geosciences, University of Oslo. Both acid treated
(HCl) and original samples were analyzed using a LECO Karbonanalyzer
CR-412 instrument. The samples were heated in a CO2 free atmosphere
to 1350 °C where the oxidized carbon (CO2) was measured with a CO2
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
IR detector. The inorganic carbon is calculated as the difference between
the acid treated sample (TOC) and the bulk sample.
5. Results
5.1. Multibeam and ROV surveys
An extensive multibeam survey in the area above the Troll field
revealed the presence of a huge pockmark field. More than 7000
pockmarks have been identified and most likely many more could be
mapped with a more extensive bathymetry study. More detailed ROV
multibeam bathymetry surveys (MBE) have provided excellent images
of the pockmarks morphology. The pockmarks targeted for more
comprehensive studies (Septagram, Peanut, Arch B–E) have similar
dimensions. The 0.5 m grid showed that the pockmarks are up to
100 m wide and 15 m deep and that the satellite pockmarks make the
complexes bigger (about 300 m in diameter) (Fig. 2).
The high resolution bathymetry allows mapping of the pockmarks in great detail. Commonly, the pockmark floors are flat with
29
occasional carbonate blocks in the central part. The western flanks
are typically steeper (up to 35°) and smoother, than the eastern
flanks. The latter often have small canyon-like structures. Small,
meter-sized so-called unit pockmarks are common both on the
flanks or may be scattered over a surrounding distance of at least
100 m from the center of the structure. Very small unit pockmarks
may be also clustered along trawling scars. The ROV sub-bottom
profiler data also revealed the presence of buried pockmarks
that are present just a few meters below the surface. ROV video observations did not show any evidence of fluids seepage, bubbling
nor microbial mats present at these sites.
5.2. Structure and mineralogy of the carbonate blocks
The carbonates from the Septagram and the Arch E pockmarks host a
variety of filter feeders and colonies of various fauna including
epibenthic fauna, large sponges, gorgonian corals, anemons, feather
stars, tube worms and bivalves (Fig. 3A–D) that typically populated
hard substratum environments. The carbonate blocks within the
Fig. 2. Example of multibeam lines acquired during the ROV dive over the (A) Arch, (B) Peanuts, and (C) Septagram. Note the distribution of the satellite smaller pockmarks around a
central “parent pockmark”. The location of three examples of cores sampled from centre, flank and outside the pockmark is marked in the Septagram image C.
30
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
Fig. 3. (A–D) Large carbonate blocks covered with epibenthic fauna, large sponges, gorgonian corals observed in the craters of the studied pockmarks. (E) Large carbonate blocks sampled
from the Septagram and Arch E pockmarks. (F) CT scans of Sample B67-4. The series of pictures represent different x-ray slices through the sample. Of particular interest is the distribution
of cavities within the sample (seen in black and dark gray), and the presence of small spherical cavities within the carbonate lobes. Similar features (small circular cavities) were identified
in the G11 Nyegga carbonates (see Mazzini et al., 2006).
pockmarks are generally large (meter size) and the retrieved samples
represent smaller portions of these. The carbonate samples share
many of the same macro and micro characteristics. Some consists of
carbonate-cemented shell debris, very similar to those from Nyegga
described by Mazzini et al. (2005), others have faint laminations within
the massive carbonate blocks suggesting distinct periods of carbonate
precipitation or possible different exhumation cycles (additional
research and further analyses on these carbonates are planned). We
focussed our detailed analyses on the larger blocks that were collected
(some reaching the size of ~ 40 cm). Macroscopic observations show
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
that the carbonates have a rusty and pitted external surface with abundant shell fragments, tube worms remains, and attached living
organisms (Fig. 1A–D). When split, the altered rim can easily be
observed as a 3–4 cm thick pitted and discolored zone. The discoloration
is due to oxidation of pyrite to iron hydroxides. The internal structure
of the samples (Fig. 3E) consists of a dendritic high porosity
31
carbonate framework with numerous clay-filled cavities. The
intricate carbonate-rich network of interconnected conduits and
voids cements detrital siliciclastic sediment with varying amounts
of carbonate content. No bivalves were observed in the internal
part of this specific carbonate type, suggesting that their formation
occurred entirely in the subsurface. Overall this type of carbonates
Fig. 4. A) Thin section scan showing the different carbonate growth stages. Note the rimming of the cavities filled by aragonite. Note the presence of clastic grains in the first two phases.
B–H: SEM backscatter images. B) The two stages defining the bulk of the sample, with a Mg-Calcite-dominated first zone and a aragonite-dominated second zone. Also indicated the last
stage of botryoidal aragonite. C) The transition between micritic (Phase 2) and botryoidal aragonite (Phase 3). D) The first phase of mineral growth contains abundant Mg-calcite
intergrown with aragonite. Both cavities and clastic silicates are common. E) Micritic aragonite with clastic grains (quartz). F) The matrix aragonite phase contains numerous
clay-filled cavities (dark). G) Small botryoidal aragonite along one of the cavities seen in F. H) Close-up of the last phase of botryoidal aragonite crystals.
32
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
strikingly resemble to the authigenic carbonates type G11-C described
by Mazzini et al. (2006) from the Nyegga region that is located to the
north of Troll.
Based on mineralogy and textures three distinct carbonate zones
can be identified in the samples (respectively from older to younger)
(Fig. 4):
• Zone 1: dark gray to very dark gray inner zone, sometimes containing
pelletoids and siliciclastic grains, with abundant micritic Mg calcite
(up to 5.3 wt.% MgO) and, in lesser amount, micritic aragonite.
Occasional calcite grains were also identified.
• Zone 2: gray to light gray matrix minerals, dominated by micritic aragonite and some Mg calcite. Mg-calcite is intergrown with aragonite
and minor calcite is also observed. Occasionally aragonite crystals
are larger than those present in zone 1, and the porosity is higher. In
both zone 1 and 2 siliciclastic grains (clay minerals, micas, quartz,
and feldspar) are present within the carbonate matrix.
• Zone 3: represents the last carbonate stage consisting of
fibrous/botryoidal white aragonite (as confirmed by XRD and
EMP analyses) commonly rimming cavities. Pyrite is also abundant.
No siliciclastic grains were observed. Interestingly, an outer rim
of fibrous/botryoidal white aragonite was also observed in the
Nyegga samples described by Mazzini et al. (2005) from the
Tobic seep mound.
Fig. 5. Carbon and oxygen isotope values for the Troll carbonates in comparison with
Nyegga carbonates: (compilation of data from: Mazzini et al., 2005, 2006; Ivanov et al.,
2010). All the Troll samples have 13C very depleted values, whereas there is a spread in
the δ18O values. Note that the shells analyzed show gradually decreasing 13C depletion,
but that some of them cannot be explained by a sea water carbon source. The trend
indicates mixing between carbonate with a seep origin and sea water origin.
5.4. Carbon content in the cores
The transition from zone 2 to zone 3 is sharp (Fig. 4) and Mg-calcite
is never present in this last stage of growth. Zones 1 and 2 occur in fairly
equal proportions in the samples, whereas the zone 3 generation is
volumetrically minor. Framboidal pyrite is commonly present in the
micritic and botryoidal carbonate. CT scan images of the selected
samples confirm that the internal structure of the carbonate block
consists of inter-fingered structures with numerous clay filled cavities.
5.3. Isotope geochemistry
The results of the carbon, oxygen, and strontium isotope analyses
are presented in Table 1. The shells gave δ13C values between 1.6
and − 8.1 ‰ VPDB and δ18O in the range 0–1.7 ‰ VPDB (Fig. 5).
The bulk carbonate samples have δ13C and δ18O values in the range
of −54.1 to −59.7 ‰ and 3.6 to 4.0 ‰, respectively.
Three samples have been analyzed for bulk 87Sr/86Sr. Although the
samples were taken from different carbonate blocks and pockmarks,
their ratio is similar. The B53-2 sample has a slightly higher ratio
(0.7093) than the B67-4 (0.7092) and the B67-6 (0.7092). The sea
water 87Sr/86Sr is for comparison 0.7091.
Bulk clay samples were analyzed for TOC and TIC, and the results are
presented in Fig. A2.A and Table 2. All four analyzed cores show consistent results, with a slight decrease in TOC from 0.7 in the top to 0.55 wt.%
in the deepest levels. There is a spread in TIC within the top sections of
the cores (2.5–3 wt.%) and a decrease to 1.1 wt.% at 3–4 m depth. Below
5 m, there is an apparent increase in TIC which is not matched by the
TOC content. Authigenic carbonates have only been identified in the
central parts of the pockmarks and not in the analyzed cores. The TIC
content thus normally reflects the abundance of calcareous microfossils
and shell fragments. The similarities between TOC/TIC values inside and
outside the pockmarks rule out significant differences in sedimentation
and accumulation. This also means that there is no obvious indication of
an enhanced organic productivity within the pockmarks.
5.5. Water geochemistry
The results of the pore water analyses are presented in Tables 2 and
A4. The sulfate content (Fig. 6 upper part) gradually lowers from sea
water values in all the cores. Note that some of the irregularities in the
Table 1
Isotope analyses of Troll carbonate samples.
Sample
Type
δ13C (‰ V-PDB)
δ18O (‰ V-PDB)
87
Sr/86 Sr
B53-2
B67-4
B67-4
B67-4
B67-4
B67-4
B67-4
B67-4
B67-4 Shell
B67-6
B610-8 (1)
B610-8 (2)
B610-8 (3)
B610-8 (4)
B612-1 (1)
B612-1 (2)
B612-1 (3)
B612-1 (4)
Bulk sample
Bulk sample
Late stage aragonite
Late stage carbonate cement
Matrix carbonate. Centre
Shell A from surface
Shell B from surface
Tube from surface
Shell in carbonate cement
Bulk sample
Late stage aragonite
Aragonite from inside shell
Black carbonate from centre
Bulk gray carbonate from centre
Black carbonate
Brownish carbonate (altered)
Light gray carbonate
Late stage white aragonite
−56.4
−55.1
−59.7
−57.2
−55.9
1.6
−8.1
−0.7
−1.3
−54.1
−59.2
−57.0
−57.3
−56.8
−57.5
−59.4
−56.1
−52.7
4.0
4.5
2.7
2.7
3.7
0.0
1.7
1.7
0.3
3.6
3.5
3.6
3.6
4.0
3.9
3.8
3.9
1.7
0.709316
0.709210
0.709200
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
33
Table 2
Pore water geochemistry (A-series) and bulk sample carbon content.
Sample
201A-2E
201A-2F
201A-3B
201A-3F
201A-4B
201A-4D
201A-4F
201A-5F
201A-Ret
201A-ShoeA
201A-ShoeF
202A-2B
202A-2E
202A-3B
202A-3F
202A-4B
202A-4F
202A-5A
202A-5E
202A Shoe A
202A Shoe E
203A-2B
203A-2F
203A-3B
203A-3F
203A-4B
203A-4F
203A-5B
203A-5F
203A Shoe B
203A Shoe F
210A-2B
210A-2F
210A-3B
210A-3F
210A-4B
210A-4F
210A-5B
210A-5F
210A Shoe B
210A Shoe F
Troll1
Troll2
Position
Depth from
Depth to
Median
TC
TOC
TIC
Cl
Br
SO4
Na
Ca
K
δ18O
δD
δ13C TIC
m
m
m
wt.%
wt.%
wt.%
ppm
ppm
ppm
ppm
ppm
ppm
permil
SMOW
permil
permil
V-PDB
0.74
0.84
1.34
1.84
2.34
2.64
2.94
3.84
5.235
5.33
6.03
0.17
0.57
1.07
1.57
2.07
2.57
2.97
3.47
5.06
5.76
0.52
1.02
1.52
2.02
2.52
3.02
3.52
4.02
5.61
6.21
0.35
0.85
1.35
1.85
2.35
2.85
3.35
3.85
5.44
6.04
3.32
3.66
2.82
2.77
2.43
2.28
2.24
1.96
1.77
1.89
1.98
3.74
2.88
2.8
2.54
2.37
2.54
1.84
2.06
2.13
2.38
3.14
2.61
2.44
2.19
2.08
1.87
1.71
1.89
2.46
2.42
3.24
2.45
2.32
2.2
1.85
1.65
1.92
2.45
2.36
2.36
0.73
0.68
0.69
0.72
0.76
0.66
0.63
0.66
0.64
0.7
0.64
0.79
0.69
0.69
0.76
0.66
0.76
0.6
0.59
0.55
0.55
0.65
0.69
0.72
0.61
0.6
0.6
0.64
0.6
0.57
0.58
0.75
0.71
0.59
0.71
0.59
0.59
0.64
0.61
0.6
0.58
2.59
2.98
2.13
2.05
1.67
1.62
1.61
1.3
1.13
1.19
1.34
2.95
2.19
2.11
1.78
1.71
1.78
1.24
1.47
1.58
1.83
2.49
1.92
1.72
1.58
1.48
1.27
1.07
1.29
1.89
1.84
2.49
1.74
1.73
1.49
1.26
1.06
1.28
1.84
1.76
1.78
20,686
19,524
20,742
22,747
21,899
19,570
19,290
19,574
23,653
18,860
19,774
19,953
20,481
19,967
20,181
23,315
20,367
20,584
19,469
18,786
20,913
20,251
19,868
19,988
20,482
21,065
20,103
18,927
20,986
18,560
66
80
67
84
70
62
62
63
71
55
79
64
72
71
64
73
69
72
58
61
66
64
72
61
57
70
55
62
59
61
3164
3096
3360
3835
4175
3207
3125
3183
2190
1456
1472
2861
3234
3368
3399
4093
3203
2941
2326
1655
1820
3049
3406
3528
3263
3467
2865
2306
2223
1867
420
420
470
3.7
−12.5
0.9
1.4
−9.7
0.6
−0.7
−8.3
0.3
2.0
−12.6
1.1
2.6
−11.4
1.0
2.2
−11.1
0.8
0.8
2.3
0.7
−9.1
−9.5
0.5
−0.8
−10.7
60
55
62
60
70
67
54
54
61
60
60
77
3288
3314
3548
3152
3369
3158
2439
2235
1587
1584
2754
2660
660
620
620
660
690
590
520
540
660
450
450
590
610
580
590
640
560
530
510
470
490
590
550
610
550
570
510
490
530
460
470
640
540
530
510
560
590
550
490
450
440
320
320
0.9
21,347
19,548
19,473
18,981
20,554
22,856
20,166
18,703
19,155
18,780
18,685
19,208
11,500
10,600
10,400
12,300
11,100
10,300
10,000
10,300
12,700
9900
9400
10,500
10,700
10,000
10,300
12,700
10,600
10,800
10,300
9900
10,600
10,500
10,000
10,600
10,500
11,300
10,300
9800
10,200
9900
9800
11,000
10,400
10,300
10,000
10,100
11,900
10,400
9800
9300
9700
10,200
10,600
1.4
0.8
−11.4
0.8
1.4
−7.2
0.5
0.8
0.9
2.1
−7.9
−11.5
0.1
−1.4
−11.7
Centre
0.69
Centre
0.79
Centre
1.29
Centre
1.79
Centre
2.29
Centre
2.59
Centre
2.89
Centre
3.79
Centre
5.19
Centre
5.28
Centre
5.98
Centre
0.12
Centre
0.52
Centre
1.02
Centre
1.52
Centre
2.02
Centre
2.52
Centre
2.92
Centre
3.42
Centre
5.01
Centre
5.71
Flank
0.47
Flank
0.97
Flank
1.47
Flank
1.97
Flank
2.47
Flank
2.97
Flank
3.47
Flank
3.97
Flank
5.56
Flank
6.16
Reference
0.30
Reference
0.80
Reference
1.30
Reference
1.80
Reference
2.30
Reference
2.80
Reference
3.30
Reference
3.80
Reference
5.39
Reference
5.99
Sea water sample Peanut
Sea water sample Septagram
0.79
0.89
1.39
1.89
2.39
2.69
2.99
3.89
5.28
5.38
6.08
0.22
0.62
1.12
1.62
2.12
2.62
3.02
3.52
5.11
5.81
0.57
1.07
1.57
2.07
2.57
3.07
3.57
4.07
5.66
6.26
0.40
0.90
1.40
1.90
2.40
2.90
3.40
3.90
5.49
6.09
curves, e.g., the increase at 1.8–2.3 m depth for the central sites, are
possibly caused by post-sampling alteration of the samples. Cl profiles also reveal similarities between center, flank, and outside the
pockmarks with values fairly constant fluctuating around 1900 ppm
(Fig. 6 lower part). The interpretation of these profiles is however
more ambiguous due to the presence of peaks of higher chlorinity.
The origin of the high Cl waters could be due to post-sampling
effects (such as evaporation/drying). Similarly to the sulfate content,
the chlorinity trends show that the samples from the centre of the
pockmarks versus reference cores have similar profiles.
The pore water isotope data (Fig. A3) show a slight decrease in δ18O
with depth. There is no difference between centre, flank, and reference
cores. The values overall fluctuate between − 9‰ and − 12‰ with
some peaks reaching values as high as − 7‰. This implies that the
pore waters essentially represent sea water with a slight diagenetic
overprint that increases with depth. The δD values are more scattered,
but there is a tendency towards lower δD values (e.g. shifting
respectively from 2.6‰ to −0.8‰) with depth. The δ13C of the dissolved
inorganic carbon in the pore fluids shows enrichment in 13C from
shallow depths (δ13C from ~− 12‰ to ~− 8‰) between 1 and 4 m.
The lowermost waters from the 5–6 m interval have comparable values
to the upper meter waters.
610
520
470
480
410
300
270
410
410
470
450
600
460
430
360
300
320
390
470
460
490
500
440
360
330
320
270
410
480
520
460
470
450
360
330
250
280
402
406
6. Discussion
6.1. Methane seepage and carbonate precipitation
From the carbon isotope data, we may ultimately address the
possible fluid and carbon sources. The overall low δ13C values (as low
as − 59.7 ‰) support the hypothesis that the carbonates formed at
shallow subsurface from hydrocarbon-rich fluid seepage at the sea
floor. Compared to the Nyegga authigenic carbonates (Mazzini et al.,
2005, 2006; Ivanov et al., 2010), the Troll samples have similar δ13C
values although the Nyegga carbonates are more enriched in 18O
(Fig. 5). The figure also shows the mixing trend between the marine
carbon and oxygen isotope signatures, in the top left part of the plot,
towards seep signatures in the lower part. The substantial spread in
δ18O values reflects both the temperature of precipitation (higher δ18O
equals colder precipitation water temperatures) and the composition
of the source fluids. Very 18O enriched fluids are usually interpreted to
have a gas hydrate or clay dehydration origin (Savin and Epstein,
1970; Kastner et al., 1993; Aloisi et al., 2000; Dählmann and de Lange,
2003). The colder temperature (i.e. N δ18O) signature of the Nyegga
samples can be attributed to deeper and colder waters at that locality
compared to the Troll site. The late stage aragonite precipitation is
34
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
Fig. 6. Sediment pore water profiles for sulfate (upper three) and chlorine (lower three).
systematically more depleted in 13C compared to the bulk samples, with
δ13C values as low as −59.7 ‰ VPDB. This likely represents a combination of fluid isotopic composition variations during precipitation and
fractionation effects between calcite and aragonite. The Sr isotope
results are consistent with a seawater source for the strontium, and
thus the major source of cations such as Na and Ca. Interestingly, isotopic data from shells may show a significant 13C depletion suggesting
that methane-rich fluids were also metabolized by the clams within
the pockmarks.
Given that Troll and Nyegga carbonates share similar δ13C values,
we may suggest a similar origin for the hydrocarbons fueling both
systems. Based on isotopic studies and fluid inclusion analyses
Mazzini et al. (2005) suggested that the gas leaking through the
pockmarks could be a mix of thermogenic (e.g. gas originating from
the deep reserves confined to Tertiary domes) and more recent microbial (shallow production) methane. Similar conclusions were suggested
also by Ivanov et al. (2010) combining the results of a large carbonate
collection coupled with gas hydrate geochemistry and gas headspace
analyses from cores. The available data thus suggest that the signature
of more recent microbial gas (i.e. generated by microbial alteration
of organic matter and gas in shallow sediments) is coupled to the
migration of ancient thermogenic gas. Similar conclusions were later
confirmed by Vaular et al. (2010) that sampled another pockmark in
the Nyegga region. The less 13C depleted (older) carbonate phases and
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
the more 13C depleted (latest) aragonite coating further supports this
scenario. The hydrocarbons trapped in the Troll reservoir provide
important information regarding the gas composition migrating from
the Viking Graben (Horstad and Larter, 1997). We therefore suggest
that a thermogenic methane with composition similar to the one
present in the Troll reservoir (i.e. δ13C = − 44.5‰ Thomas et al.,
1985) migrated in the whole region of the Norwegian Channel and
was mixed with shallower microbial methane. This would explain the
depletion values recorded in the Troll carbonates.
The presence of thousands of pockmarks over a large area, suggests
formation during a major gas release event. Therefore, given that the
Troll δ13C carbonate values are fairly clustered, we hypothesize that
the mix of deeper and older thermogenic gas with the younger and
shallower microbial gas, occurred prior to the main gas pulse that
triggered the formation of the pockmarks. A possible analogy to this
scenario is provided by the Peon shallow gas reservoir located on the
Tampen Spur to the North of Troll. Despite its shallow depth
(165 mbsf), Peon is estimated to hold 15 to 30 billion standard cubic
meters of mostly microbial gas (δ13C ~ − 72.‰). Like for the Troll
pockmarks field, this gas is likely to be the result of deeper thermogenic
and shallower microbial gas that is now trapped in a thin 18.7 m unit in
a 37.8 m think formation of porous sands (NPD, 2005).
6.2. Insights of pockmarks activity from petrography and carbonate
precipitation
The methanogenic origin for the studied blocks is also consistent
with the petrographic observations showing typical carbonate phases
paired with pyrite. This confirms that the carbonates formed by
anaerobic oxidation of methane coupled with seawater sulfate reduction, and that the main cations are seawater-derived. The mineralogy
and the textures reflect the evolution during growth. Thus we can
trace the evolution of the samples from initial precipitation within the
sulfate reduction zone to exhumation and exposure to sea water and
oxidizing conditions. The diagenetic evolution of the different cement
zones is summarized in Fig. 7. The carbonates present in Zone 1 (micritic
Mg-calcite) precipitated in the subsurface, as indicated by the abundance of siliciclastic material. The significant amount of pelletoids
could represent carbonate remnants of microbial mats as also indicated
by several authors (e.g. Peckmann and Thiel, 2004; Riding and Tomas,
2006). During the gradual decrease in seepage and slow exhumation
of the blocks, the cement phase became predominantly micritic
aragonite (Zone 2). As the exhumation of the carbonates continued
and the pockmarks activity declined the last carbonate Zone 3
35
precipitated filling the pores with acicular crystals. Once the carbonates
were fully exposed on the seafloor (see next paragraph on exhumation
of the blocks), a later stage of carbonate oxidation occurred (oxidation
of pyrite to iron hydroxides), as highlighted by the reddish oxidation
front present within the dendritic features.
As also pointed out by other authors (Naehr et al., 2000; Aloisi et al.,
2002; Mazzini et al., 2006; Himmler et al., 2011) the different carbonate
phases reflect the varying geochemistry of the fluids during the various
stages of the pockmarks seepage activity. Like also observed at other
localities, even in the Troll case, the last carbonate phase is represented
by aragonite. This is most probably related to 1) the less vigorous gas
seepage during the last stages of the pockmarks activity resulting in a
sulfate-rich diagenetic environment and 2) to the gradual expose of
the carbonate closer to the sea floor.
The internal structure of the carbonate blocks, consisting of
interconnected voids and cavities, could be ascribed to different factors
and alternative scenarios could be depicted. For example Haas et al.
(2010) documented the recovery of porous carbonates whose shape
was influenced by the presence of Vestimentiferan tube worms
populating the seepage sites. At these sites the carbonate precipitation
occurs in the subsurface around sediments surrounding the posterior
‘root’ tubes of the colonies. As video survey revealed the presence of
tube worms at Troll, they could have strongly affected the shapes of
the resulting carbonate.
Similar shaped carbonates have also been described by Himmler
et al. (2011). The authors suggest a model implying a gradual erosion
of the sediments to expose the carbonates followed by carbonate
corrosion resulting from acidity locally produced by aerobic oxidation
of methane and hydrogen sulfide. The internal structure of the Troll
carbonates does not show the presence of e.g. shells or other easily
identifiable features and this hypothesis cannot be easily validated.
Thin sections do not reveal obvious indication of carbonate dissolution
zones, in addition the presence of methanogenic aragonite filling the
numerous vesicles and conduits might rule out this scenario.
Alternatively the cavernous structure present in the blocks could
represent the carbonate precipitation casting portions of free gas
bubbles or tubular fluids conduits in the subsurface sediments created
during the vigorous methane seepage. With the present data, this
seems to be the most likely explanation. The obtained CT scan images
are strikingly similar to that observed by Mazzini et al. (2006).
Therefore, considering these macrostructures, the formation model
similar to that suggested by Mazzini et al. (2006) could be applied.
Some general considerations can be obtained from the results of the
EMP analyses. In contrast to the carbonates from the G11 pockmark at
Fig. 7. Summary of the petrographic and textural evolution of the carbonate samples. The two first stages are characterized by micritic Mg-Calcite and aragonite, where the aragonite is
more extensive with time. Aragonite dominance during the last stage can be explained by shallower precipitation relative to the sulfate reduction zone (lowers the stability of Mg-Calcite),
but still within reducing condition (pyrite is present). The final exhumation of the samples led to pervasive pyrite oxidation along the margins of the carbonates.
36
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
Nyegga, dolomite was not identified in any of the studied Troll samples.
The Sr and Mg contents of the carbonates are shown in Fig. A2.B
in molar values relative to the end member structural formulas.
Strontium is confined to aragonite, which contains up to 1.8 wt.% SrO
(corresponding to 0.034 mol per formula unit). The SrO content in
aragonite is always higher than 0.7 wt.%. The chemical formula for
that particular mineral analysis (the first analysis from B612-1 in
Table A3) is accordingly Ca1.964Mn0.002 Sr0.034(CO3)2 compared to
Ca(CO3) for pure aragonite. The Sr concentrations in the Troll aragonites
are very similar to the aragonites from the G11 pockmark at Nyegga
suggesting similar formation conditions. The calcites do not contain
any strontium, and plot along the x-axis in Fig. A2.B. Foraminifera shells
have the lowest Mg concentrations. The absence of strontium in the
calcite is evidence for a precipitation of calcite directly from pore
fluids and not from recrystallization of aragonite which normally
takes place during deeper burial and results in some Sr entering the
calcite structure.
Similarly the 87Sr/86Sr of bulk carbonate from the G11 pockmark
(Mazzini et al., 2006) had a ratio of 0.7092 that overlaps that of the
Troll samples. The Troll samples have a significant clastic component
which likely contributed to increasing the 87Sr/86Sr up to 0.7093 in
one sample. However, the results are still consistent with a seawater
source for the strontium, and thus the major source of cations.
Our conclusion about no current seep activity at Troll is further
supported by seismic interpretations that showed no evidence for
deep roots of the pockmarks in the study area. There is further no
evidence of structural control on the location of the pockmarks.
6.4. Pockmark formation
It is remarkable that numerous pockmarks are surrounded by smaller satellite pockmarks on the outskirts of the crater (Fig. 2). Interestingly
these pockmark complexes seem to have the thickest and most laterally
extensive distribution of authigenic carbonates. A model that could explain this was initially proposed by Mazzini et al. (2006), suggesting
that an initial vigorous burst created the major pockmark crater followed by carbonate precipitation; once the methanogenic carbonates reach
a significant thickness, the rising fluids are deflected laterally thus
contributing to a progressive extension and precipitation of carbonates
(see Fig. 8 G11-C in Mazzini et al., 2006). Once the full surface of the
pockmark crater is covered by thick authigenic carbonates, the fluids
rising in the more permeable vertical conduit will be buffered. The
overpressure gathered in the subsurface of the “parent pockmark”
6.3. The sulfate reduction zone as a proxy for methane seepage
The systematic pore water extraction from reference cores and cores
collected for the centre and the flanks of pockmarks was completed in
order to trace the presence of active seepage at pockmark sites. Of
−
particular relevance are the SO2−
4 and Cl values that commonly reflect
the presence of fluids seepage thus affecting the depth of the sulfate
reduction zone (Fig. 6). More specifically, pore water sulfate concentrations in the cores can be used as a proxy for methane flux (e.g. Treude
et al., 2005; Dale et al., 2008; Knab et al., 2009; Treude et al., 2014).
If the sediment pore water sulfate profiles in the pockmarks are
shallower compared to the background cores, it suggests the presence
of methane seepage (sulfate reduction to sulfide induced by microbial
methane oxidation) (Borowski et al., 1996). Deeper sulfate reduction
levels in the pockmarks would require a different explanation, like
permeability variations.
Overall, the sulfate is lowered in concentration by up to 63% compared to the upper parts of the cores. Even though this is a significant
depletion, it is evident that the cores were not deep enough to penetrate
the level of sulfate reduction. Of relevance is the comparison between
the three sites. Fig. 6 shows the virtual similar sulfate profiles when
comparing the centre with the flank and the reference stations. The
presence of a regional and uniform sulfate reduction profile suggests
that there are no indications that the pockmarks are currently active
conduits for methane bearing fluids. Similar profiles are indeed collected form cores collected at exposed or buried pockmarks. It is reasonable
to conclude that the similar profiles inside and outside the pockmarks
indicate that, after a period of widespread seepage, the region
re-equilibrated the vertical distribution of chemical reactions and
fluids composition. Similar cores from active pockmarks in Nyegga,
reveal instead drastic sulfate decrease in the top 50 cm bsf clearly
in contrast with the results reported herein (Ivanov et al., 2010).
Cl content supports the conclusion that there is no indication of flow
of low salinity pore fluids through the pockmark sediments, which
would have been the case if gas hydrates were actively dissociating.
This conclusion is supported also by the similarities in the Na, Ca,
and K trends.
Insights about pockmarks inactivity also come from the
metagenomics studies to characterize the prokaryotic communities
inhabiting the surface sediments in the Troll area in relation to
geochemical parameters (Havelsrud et al., 2012). The sampled locations
show no obvious evidence of methane seepage.
Fig. 8. Correlation between pore water sulfate concentrations and pockmark activity. With
no significant active seepage, the depth of the sulfate reduction zone will be similar outside
and inside the pockmark (upper case). With active methane seepage, the methane will
reduce the sulfate and thus result in a shallower sulfate reduction zone within the
pockmark (middle case). The lower case represents a potentially deeper sulfate reduction
zone inside the pockmark. Here deep penetration of sea water into the sediments is likely
occurring possibly due to high permeabilities in the pockmark sediments.
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
will then be radially released triggering the formation of “satellite
pockmarks” framing the main crater.
6.5. Pockmarks dynamics
We do not exclude that the diffusive methane seepage at pockmarks
sites might have encountered period of more vigorous or explosive activity as suggested by the presence of isolated and angular large carbonate blocks. Similar isolated blocks have also been observed at Nyegga
(Mazzini et al., 2006; Ivanov et al., 2010). In any case, the fact that
meter sized carbonates remained exposed on the seafloor after an ancient period of seepage suggest that significant sediment erosion must
have occurred after the subsurface carbonate precipitation. Moreover
the deposits inside the pockmarks consist mainly of ice-rafted clasts
and coarse-grained sands. This additional piece of evidence suggests
that two mechanisms are involved in the pockmarks formation and
preservation. Initially the seeping gas transports in suspension preferentially the fine-grained sediments. These particles are constantly washed
away by the currents leaving behind the coarser fraction. The strong N–S
currents swiping the area also explain why the pockmarks have not
been buried by hemipelagic deposition until now. The direction of the
currents matches that of the main axis of the pockmarks that are statistically mostly elongated towards the NNW–SSE. The degraded N slope of
the pockmarks compared to the southern slope also supports influence
from a dominating current direction. Laboratory and numerical simulations support this scenario and reveal that typical bottom currents are
efficient to prevent the sedimentation of particles up to fine sand
(Hammer et al., 2009; Pau et al., 2014) and may be efficient mechanisms
for sediment erosion. Thus the oxidation rims present around the
collected carbonates, are the result of their exposure on the seafloor.
7. Conclusions
• A large pockmark field was mapped and investigated in the Troll region.
More than 7000 pockmarks are found in the Troll East gas field region.
• Large blocks of carbonate deposits were observed and sampled
at pockmarks sites. The presence of these carbonates uniquely
inside the pockmarks indicates widespread carbonate precipitation occurring only at these location.
• The methanogenic origin of these carbonates is confirmed by their 12C
enrichment (i.e. δ13C as low as −59.7‰). The methane oxidation
coupled with seawater sulfate reduction is supported by abundant
framboidal pyrite observed within the carbonate cements.
• Three main phases of carbonate growth are recorded in the samples,
including one initial phase of Mg calcite and micritic aragonite, and
intermediate phase of micritic aragonite and a third, final phase of
botryoidal aragonite devoid of siliciclastic sediment. The macroscopic
texture of these carbonates strongly resemble to those collected from
Nyegga pockmarks.
• 18O enrichment in the carbonates suggests that the source of methane
might have originated by the dissociation of gas hydrates ongoing at
the moment of carbonate precipitation.
• The 87Sr/86Sr is consistent with a seawater source for the strontium, and
thus the major source of cations.
• The carbonates have been exposed to seawater for a long time,
as highlighted by the abundant outer surface alteration and pyrite
oxidation. This suggests ancient carbonate precipitation that is not
occurring any longer.
• Further evidence of paleao seepage is provided by water analyses
extracted from cores sampled inside and outside the pockmarks.
The results show no difference between pockmark and background
pore water sulfate concentrations, indicating that the pockmarks are
currently not active fluid flow conduits.
• Seafloor observations showed no evidence of bubbles, fluids seepage,
microbial colonies or other typical chemosymbiotic alive assemblages
present at active seepage sites.
37
Although active pockmarks are reported elsewhere in the North Sea
and in the Nyegga area to the north, the evidences collected reveal that
there is no activity in the Troll area and that the retrieved carbonates are
the result of a palaeo methane seepage.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.margeo.2015.12.012.
Acknowledgments
We thank the editor Gert J. De Lange, Luca Martire and two
anonymous reviewers for their constrictive comments that helped to
improve the manuscript. We are very grateful to Statoil and the Troll
license partners for giving access to data used in this study. The research
leading to these results has received funding from the European
Research Council under the European Union's Seventh Framework
Program (FP7/2007-2013)/ERC Grant agreement n° 308126. The Centre
for Earth Evolution and Dynamics (CEED) is thanked for the support
during this work.
References
Akhmetzhanov, A.M., Kenyon, N.H., Ivanov, M.K., Westbrook, G., Mazzini, A., 2008.
Deep-water depositional systems and cold seeps of the Western Mediterranean, Gulf
of Cadiz and Norwegian continental margins. IOC Technical Series No. 76. UNESCO.
Aloisi, G., Bouloubassi, I., Heijs, S.K., Pancost, R.D., Pierre, C., Sinninghe Damste, J.S.,
Gottschal, J.C., Forney, L.J., Rouchy, J.-M., 2002. CH4-consuming microorganisms
and the formation of carbonate crusts at cold seeps. Earth Planet. Sci. Lett. 203
(1), 195–203.
Aloisi, G., Pierre, C., Rouchy, J.M., Foucher, J.P., Woodside, J., 2000. Methane-related
authigenic carbonates of eastern Mediterranean Sea mud volcanoes and their possible
relation to gas hydrate destabilisation. Earth Planet. Sci. Lett. 184 (1), 321–338.
Andreassen, K., Mienert, J., Bryn, P., Singh, A., 2000. A double gas-hydrate related
bottom simulating reflector at the Norwegian continental margin. Ann. N. Y.
Acad. Sci. 912, 126–135.
Berndt, C., Bunz, S., Clayton, T., Mienert, J., Saunders, M., 2004. Seismic character of bottom
simulating reflectors: examples from the mid-Norwegian margin. Mar. Pet. Geol. 21
(6), 723–733.
Boetius, A., Suess, E., 2004. Hydrate Ridge: a natural laboratory for the study of microbial
life fueled by methane from near-surface gas hydrates. Chem. Geol. 205 (3–4),
291–310.
Boetius, A., Wenzhöfer, F., 2013. Seafloor oxygen consumption fuelled by methane from
cold seeps. Nat. Geosci. 6 (9), 725–734.
Borowski, W.S., Paull, C.K., Ussler, W., 1996. Marine pore-water sulfate profiles indicate in
situ methane flux from underlying gas hydrate. Geology 24 (7), 655–658.
Bouriak, S., Vanneste, M., Saoutkine, A., 2000. Inferred gas hydrates and clay diapirs near
the Storegga Slide on the southern edge of the Vøring Plateau, offshore Norway. Mar.
Geol. 163 (1–4), 125–148.
Brothers, L.L., Kelley, J.T., Belknap, D.F., Barnhardt, W.A., Andrews, B.D., Legere, C., Hughes
Clarke, J.E., 2012. Shallow stratigraphic control on pockmark distribution in north
temperate estuaries. Mar. Geol. 329–331 (0), 34–45.
Bünz, S., Mienert, J., Berndt, C., 2003. Geological controls on the Storegga gas-hydrate
system of the mid-Norwegian continental margin. Earth Planet. Sci. Lett. 209 (3–4),
291–307.
Chand, S., Thorsnes, T., Rise, L., Brunstad, H., Stoddart, D., Bøe, R., Lågstad, P., Svolsbru, T.,
2012. Multiple episodes of fluid flow in the SW Barents Sea (Loppa High) evidenced
by gas flares, pockmarks and gas hydrate accumulation. Earth Planet. Sci. Lett.
331–332 (0), 305–314.
Christiansson, P., Faleide, I.J., Berge, A.M., 2000. Crustal structure in the northern North
Sea: an integrated geophysical study. Geol. Soc. Lond. Spec. Publ. 167, 15–40.
Dale, A.W., Regnier, P., Knab, N.J., Jørgensen, B.B., Van Cappellen, P., 2008. Anaerobic oxidation of methane (AOM) in marine sediments from the Skagerrak (Denmark): II.
Reaction-transport modeling. Geochim. Cosmochim. Acta 72 (12), 2880–2894.
Dählmann, A., de Lange, G.J., 2003. Fluid–sediment interactions at Eastern Mediterranean
mud volcanoes: a stable isotope study from ODP Leg 160. Earth Planet. Sci. Lett. 212
(3–4), 377–391.
Faleide, J.I., Bjørlykke, K., Gabrielsen, R.H., 2010. Geology of the Norwegian continental
shelf. In: Bjørlykke, K. (Ed.), Petroleum Geoscience: From Sedimentary Environments
to Rock Physics, pp. 467–499.
Forsberg, C.F., Planke, S., Tjelta, T.I., Strout, J.M., Svensen, C., 2007. Formation of pockmarks
in the Norwegian Channel. Proceedings of the 6th International Offshore Site
Investigation and Geotechnics Conference: Confronting New Challenges and Sharing
Knowledge, 11–13 September 2007, London, UK, pp. 221–230.
Gontharet, S., Pierre, C., Blanc-Valleron, M.M., Rouchy, J.M., Fouquet, Y., Bayon, G., Foucher,
J.P., Woodside, J., Mascle, J., 2007. Nature and origin of diagenetic carbonate
crusts and concretions from mud volcanoes and pockmarks of the Nile deep-sea
fan (eastern Mediterranean Sea). Deep-Sea Res. II Top. Stud. Oceanogr. 54 (11–13),
1292–1311.
Greinert, J., Bohrmann, G., Suess, E., 2001. Gas hydrate-associated carbonates and
methane-venting at Hydrate Ridge classification, distribution, and origin of
authigenic lithologies. In: Paull, C.K., Dillon, W.K. (Eds.), Natural Gas Hydrates:
38
A. Mazzini et al. / Marine Geology 373 (2016) 26–38
Occurrence, Distribution and DetectionGeophysical Monograph 124. American
Geophysical Union, pp. 99–113.
Greinert, J., Lewis, K.B., Bialas, J., Pecher, I.A., Rowden, A., Bowden, D.A., De Batist, M., Linke,
P., 2010. Methane seepage along the Hikurangi Margin, New Zealand: overview of
studies in 2006 and 2007 and new evidence from visual, bathymetric and
hydroacoustic investigations. Mar. Geol. 272 (1–4), 6–25.
Haas, A., Peckmann, J., Elvert, M., Sahling, H., Bohrmann, G., 2010. Patterns of carbonate
authigenesis at the Kouilou pockmarks on the Congo deep-sea fan. Mar. Geol. 268
(1–4), 129–136.
Hammer, Ø., Webb, K., Depreiter, D., 2009. Numerical simulation of upwelling currents in
pockmarks, and data from the Inner Oslofjord, Norway. Geo-Mar. Lett. 29 (4),
269–275.
Havelsrud, O.E., Haverkamp, T.H.A., Kristensen, T., Jakobsen, K.S., Rike, A.G., 2012.
Metagenomic and geochemical characterization of pockmarked sediments
overlaying the Troll petroleum reservoir in the North Sea. BMC Microbiol. 12.
Himmler, T., Brinkmann, F., Bohrmann, G., Peckmann, J., 2011. Corrosion patterns of
seep-carbonates from the eastern Mediterranean Sea. Terra Nova 23 (3), 206–212.
Holgate, N.E., Jackson, C.A.L., Hampson, G.J., Dreyer, T., 2013. Sedimentology and sequence
stratigraphy of the Middle–Upper Jurassic Krossfjord and Fensfjord formations, Troll
Field, northern North Sea. Pet. Geosci. 19, 237–258.
Horstad, I., Larter, S.R., 1997. Petroleum migration, alteration, and remigration within
Troll field, Norwegian North Sea. AAPG Bull. 81 (2), 222–248.
Hovland, M., Judd, A.G., 1988. Seabed Pockmarks and Seepages: Impact on Geology, Biology and the Marine Environment. Graham & Trotman (293 pp.).
Hovland, M., Svensen, H., Forsberg, C.F., Johansen, H., Fichler, C., Fossa, J.H., Jonsson, R.,
Rueslatten, H., 2005. Complex pockmarks with carbonate-ridges off mid-Norway:
products of sediment degassing. Mar. Geol. 218 (1–4), 191–206.
Hovland, M., Talbot, M.R., Qvale, H., Olaussen, S., Aasberg, L., 1987. Methane-related
carbonate cements in pockmarks of the North Sea. J. Sediment. Petrol. 57 (5),
881–892.
Ivanov, M., Blinova, V., Kozlova, E., Westbrook, G., Mazzini, A., Minshull, T., Nouzé, H.,
2007. First sampling of gas hydrate from the Vøring Plateau. Eos 88 (19), 209–210.
Ivanov, M., Mazzini, A., Blinova, V., Kozlova, E., Laberg, J.-S., Matveeva, T., Taviani, M.,
Kaskov, N., 2010. Seep mounds on the Southern Vøring Plateau (offshore Norway).
Mar. Pet. Geol. 27 (6), 1235–1261.
Kastner, M., Elderfield, H., Jenkins, W.J., Gieskes, J.M., Gamo, T., 1993. Geochemical and
isotopic evidence for fluid flow in the Western Nankai subduction zone, Japan.
Proceedings of the Ocean Drilling Program, Scientific Results. vol. 131, pp. 397–413.
Kelley, J.T., Dickson, S.M., Belknap, D.F., Barnhardt, W.A., Henderson, M., 1994. Giant
sea-bed pockmarks — evidence for gas escape from Belfast Bay, Maine. Geology 22
(1), 59–62.
Knab, N.J., Cragg, B.A., Hornibrook, E.R.C., Holmkvist, L., Pancost, R.D., Borowski, C., Parkes,
R.J., Jorgensen, B.B., 2009. Regulation of anaerobic methane oxidation in sediments of
the Black Sea. Biogeosciences 6 (8), 1505–1518.
Kocherla, M., Teichert, B.M.A., Pillai, S., Satyanarayanan, M., Ramamurthy, P.B., Patil, D.J.,
Rao, A.N., 2015. Formation of methane-related authigenic carbonates in a highly dynamic biogeochemical system in the Krishna–Godavari Basin, Bay of Bengal. Mar. Pet.
Geol. 64, 324–333.
Magalhães, V.H., Pinheiro, L.M., Ivanov, M.K., Kozlova, E., Blinova, V., Kolganova, J.,
Vasconcelos, C., McKenzie, J.A., Bernasconi, S.M., Kopf, A.J., 2012. Formation processes
of methane-derived authigenic carbonates from the Gulf of Cadiz. Sediment. Geol.
243–244, 155–168.
Mazzini, A., Aloisi, G., Akhmanov, G.G., Parnell, J., Cronin, B., Murphy, P., 2005. Integrated
petrographic and geochemical record of hydrocarbon seepage on the Vøring Plateau.
Geol. Soc. 162, 815–827.
Mazzini, A., Svensen, H., Hovland, M., Planke, S., 2006. Comparison and implications from
strikingly different authigenic carbonates in a Nyegga complex pockmark, G11,
Norwegian Sea. Mar. Geol. 231, 89–102.
Menot, L., Galeron, J., Olu, K., Caprais, J.C., Crassous, P., Khripounoff, A., Sibuet, M., 2010.
Spatial heterogeneity of macrofaunal communities in and near a giant pockmark
area in the deep Gulf of Guinea. Mar. Ecol. Evol. Perspect. 31 (1), 78–93.
Naehr, T.H., Rodriguez, N.M., Bohrmann, G., Paull, C.K., Botz, R., 2000. Methane-derived
authigenic carbonates associated with gas hydrate decomposition and fluid venting
above the Blake Ridge Diapir. In: Paull, C.K., Matsumoto, R., Wallace, P.J., Dillon,
W.P. (Eds.), Proc. ODP Scientific Results vol. 164. Ocean Drilling Program, College
Station, TX, pp. 286–300.
Nickel, J.C., di Primio, R., Kallmeyer, J., Hammer, Ø., Horsfield, B., Stoddart, D., Brunstad, H.,
Mangelsdorf, K., 2013. Tracing the origin of thermogenic hydrocarbon signals in
pockmarks from the southwestern Barents Sea. Org. Geochem. 63, 73–84.
Nickel, J.C., di Primio, R., Mangelsdorf, K., Stoddart, D., Kallmeyer, J., 2012. Characterization
of microbial activity in pockmark fields of the SW-Barents Sea. Mar. Geol. 332–334,
152–162.
NPD, 2005. well 35/2–1. http://factpages.npd.no/ReportServer?/FactPages/PageView/
wellbore_exploration&rs:Command=Render&rc:Toolbar=false&rc:Parameters=
f&NpdId=5135&IpAddress=129.240.240.30&CultureCode=nb-no.
Olu, K., Lance, S., Sibuet, M., Henry, P., Fiala-Medioni, A., Dinet, A., 1997. Cold seep
communities as indicators of fluid expulsion patterns through mud volcanoes
seaward of the Barbados accretionary prism. Deep-Sea Res. 44 (5), 811–842.
Pau, M., Gisler, G., Hammer, Ø., 2014. Experimental investigation of the hydrodynamics in
pockmarks using particle tracking velocimetry. Geo-Mar. Lett. 34 (1), 11–19.
Paull, C., Ussler III, W., Holbrook, W.S., Hill, T., Keaten, R., Mienert, J., Haflidason, H.,
Johnson, J., Winters, W., Lorenson, T., 2008. Origin of pockmarks and chimney
structures on the flanks of the Storegga Slide, offshore Norway. Geo-Mar. Lett. 28
(1), 43–51.
Peckmann, J., Thiel, V., 2004. Carbon cycling at ancient methane-seeps. Chem. Geol. 205
(3–4), 443–467.
Riboulot, V., Cattaneo, A., Sultan, N., Garziglia, S., Ker, S., Imbert, P., Voisset, M., 2013.
Sea-level change and free gas occurrence influencing a submarine landslide and
pockmark formation and distribution in deepwater Nigeria. Earth Planet. Sci. Lett.
375, 78–91.
Riding, R., Tomas, S., 2006. Stromatolite reef crusts, Early Cretaceous, Spain: bacterial
origin of in situ-precipitated peloid microspar? Sedimentology 53 (1), 23–34.
Ritt, B., Pierre, C., Gauthier, O., Wenzhofer, F., Boetius, A., Sarrazin, J., 2011. Diversity
and distribution of cold-seep fauna associated with different geological and
environmental settings at mud volcanoes and pockmarks of the Nile Deep-Sea
Fan. Mar. Biol. 158 (6), 1187–1210.
Rogers, J.N., Kelley, J.T., Belknap, D.F., Gontz, A., Barnhardt, W.A., 2006. Shallow-water
pockmark formation in temperate estuaries: a consideration of origins in the western
gulf of Maine with special focus on Belfast Bay. Mar. Geol. 225 (1–4), 45–62.
Savin, S.M., Epstein, S., 1970. The oxygen and hydrogen isotope geochemistry of clay
minerals. Geochim. Cosmochim. Acta 34 (1), 25–42.
Sejrup, H.P., Larsen, E., Haflidason, H., Berstad, I.M., Hjelstuen, B.O., Jonsdottir, H.E., King,
E.L., Andvik, J., Longva, L., Nygård, A., Ottesen, D., Raunholm, S., Rise, L., Stalsberg, K.,
2003. Configuration, history and impact of the Norwegian Channel Ice Stream. Boreas
32, 18–36.
Sibuet, M., Olu, K., 1998. Biogeography, biodiversity and fluid dependence of deepsea cold-seep communities at active and passive margins. In: Smith, C.R.,
Mullineaux, L.S., Levin, L.A. (Eds.), Deep-Sea Biodiversity: A Compilation of
Recent Advances in Honor of Robert R. Hessler, Volume 45; Number 1/3.
Elsevier, pp. 517–567.
Smith, A.J., Mienert, J., Bunz, S., Greinert, J., 2014. Thermogenic methane injection via
bubble transport into the upper Arctic Ocean from the hydrate-charged Vestnesa
Ridge, Svalbard. Geochem. Geophy. Geosyst. 15 (5), 1945–1959.
Thomas, B.M., Møller-Pedersen, P., Whitaker, M.F., Shaw, N.D., 1985. Organic facies and
hydrocarbon distributions in the Norwegian North Sea. In: Thomas, B.M., Thomas,
M.T., Doré, A.G., Home, P.C., Larsen, R.M. (Eds.), Petroleum Geochemistry in
Exploration of the Norwegian Shelf. Norwegian Petroleum Society, pp. 3–26.
Tjelta, T.I., Svanø, G., Strout, J.M., Forsberg, C.F., 2007. Shallow gas and its multiple
impact on a North Sea production platform. Proceedings of the 6th International
Offshore Site Investigation and Geotechnics Conference: Confronting New
Challenges and Sharing Knowledge, 11–13 September 2007, London, UK,
pp. 205–220.
Treude, T., Krause, S., Maltby, J., Dale, A.W., Coffin, R., Hamdan, L.J., 2014. Sulfate reduction
and methane oxidation activity below the sulfate-methane transition zone in Alaskan
Beaufort Sea continental margin sediments: implications for deep sulfur cycling.
Geochim. Cosmochim. Acta 144, 217–237.
Treude, T., Niggemann, J., Kallmeyer, J., Wintersteller, P., Schubert, C.J., Boetius, A.,
Jørgensen, B.B., 2005. Anaerobic oxidation of methane and sulfate reduction along
the Chilean continental margin. Geochim. Cosmochim. Acta 69 (11), 2767–2779.
Valentine, D.L., Reeburgh, W.S., 2000. New perspectives on anaerobic methane oxidation.
Environ. Microbiol. 2 (5), 477–484.
Vaular, E.N., Barth, T., Haflidason, H., 2010. The geochemical characteristics of the hydratebound gases from the Nyegga pockmark field, Norwegian Sea. Org. Geochem. 41 (5),
437–444.
Westbrook, G.K., Thatcher, K.E., Rohling, E.J., Piotrowski, A.M., Pälike, H., Osborne, A.H.,
Nisbet, E.G., Minshull, T.A., Lanoisellé, M., James, R.H., Hühnerbach, V., Green, D.,
Fisher, R.E., Crocker, A.J., Chabert, A., Bolton, C., Beszczynska-Möller, A., Berndt, C.,
Aquilina, A., 2009. Escape of methane gas from the seabed along the West Spitsbergen continental margin. Geophys. Res. Lett. 36 (15), L15608.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising