Carbon isotope exchange during anaerobic oxidation of methane (AOM) in sediments of the northeastern South China Sea ScienceDirect

Abstract

The major processes that determine the distribution of methane (CH 4) in anoxic marine sediments are methanogenesis and the anaerobic oxidation of methane (AOM), with organoclastic sulfate reduction exerting an important secondary control. However, the factors leading to the distribution of stable carbon isotopes (d 13 C) of CH 4 are currently poorly understood, in particular the commonly-observed minimum in d 13 C-CH 4 at the sulfate-methane transition (SMT) where AOM rates reach maximum values. Conventional isotope systematics predict 13 C-enrichment of CH 4 in the SMT due to preferential 12 CH 4 consumption by AOM. Two hypotheses put forward to explain this discrepancy are the addition of 12 C-enriched CH 4 to pore-waters by methanogenesis in close proximity to AOM, and enzymatically-mediated carbon isotope equilibrium between forward and backward AOM at low concentrations of sulfate. To examine this in more detail, field data including d 13 C of CH 4 and dissolved inorganic carbon (DIC) from the continental margin offshore southwestern Taiwan were simulated with a reaction-transport model. Model simulations showed that the minima in d 13 C-CH 4 and d 13 C-DIC in the SMT could only be simulated with carbon isotope equilibrium during AOM. The potential for carbon cycling between methanogenesis and AOM in and just below the SMT was insignificant due to very low rates of methanogenesis. Backward AOM also gives rise to a pronounced kink in the d 13 C-DIC profile several meters below the SMT that has been observed in previous studies. We suggest that this kink marks the true base of the SMT where forward and backward AOM are operating at very low rates, possibly sustained by cryptic sulfur cycling or barite dissolution.

Full text

Carbon isotope exchange during anaerobic oxidation of
methane (AOM) in sediments of the northeastern South China Sea
Pei-Chuan Chuang
a,b,⇑
, Tsanyao Frank Yang
a,1
, Klaus Wallmann
c
Ryo Matsumoto
d
, Ching-Yi Hu
a,e
, Hsuan-Wen Chen
a
, Saulwood Lin
f
Chih-Hsien Sun
g,1
, Hong-Chun Li
a
, Yunshuen Wang
h
, Andrew W. Dale
c
a
Department of Geosciences, National Taiwan University, Taipei, Taiwan
b
MARUM – Center for Marine Environmental Sciences, University of Bremen, Leobener Str., 28359 Bremen, Germany
c
GEOMAR Helmholtz Centre for Ocean Research Kiel, Wischhofstr. 1–3, 24148 Kiel, Germany
d
Gas Hydrate Laboratory, Organization for the Strategic Coordination of Research and Intellectual Properties, Meiji University,
1-1 Kanda-Surugadai, Chiyoda-ku, Tokyo 101-8301, Japan
e
Exploration & Development Research Institute, CPC, Taiwan
f
Institute of Oceanography, National Taiwan University, Taipei, Taiwan
g
Exploration and Production Business Division, CPC Corporation, Taiwan
h
Central Geological Survey, MOEA, Taipei, Taiwan
Received 22 January 2018; accepted in revised form 5 November 2018; Available online 15 November 2018
Abstract
The major processes that determine the distribution of methane (CH
4
) in anoxic marine sediments are methanogenesis and
the anaerobic oxidation of methane (AOM), with organoclastic sulfate reduction exerting an important secondary control.
However, the factors leading to the distribution of stable carbon isotopes (d
13
C) of CH
4
are currently poorly understood,
in particular the commonly-observed minimum in d
13
C-CH
4
at the sulfate-methane transition (SMT) where AOM rates reach
maximum values. Conventional isotope systematics predict
13
C-enrichment of CH
4
in the SMT due to preferential
12
CH
4
con-
sumption by AOM. Two hypotheses put forward to explain this discrepancy are the addition of
12
C-enriched CH
4
to pore-
waters by methanogenesis in close proximity to AOM, and enzymatically-mediated carbon isotope equilibrium between
forward and backward AOM at low concentrations of sulfate. To examine this in more detail, field data including d
13
Cof
CH
4
and dissolved inorganic carbon (DIC) from the continental margin offshore southwestern Taiwan were simulated with
a reaction-transport model. Model simulations showed that the minima in d
13
C-CH
4
and d
13
C-DIC in the SMT could only be
simulated with carbon isotope equilibrium during AOM. The potential for carbon cycling between methanogenesis and AOM
in and just below the SMT was insignificant due to very low rates of methanogenesis. Backward AOM also gives rise to a
pronounced kink in the d
13
C-DIC profile several meters below the SMT that has been observed in previous studies. We sug-
gest that this kink marks the true base of the SMT where forward and backward AOM are operating at very low rates, pos-
sibly sustained by cryptic sulfur cycling or barite dissolution.
Ó2018 Elsevier Ltd. All rights reserved.
Keywords: Anaerobic oxidation; Sea-floor; Methane; Carbon isotopes; Model; Sulfate reduction; South China Sea
https://doi.org/10.1016/j.gca.2018.11.003
0016-7037/Ó2018 Elsevier Ltd. All rights reserved.
⇑
Corresponding author at: MARUM – Center for Marine Environmental Sciences, University of Bremen, Leobener Str., 28359 Bremen,
Germany.
E-mail addresses: pchuang@marum.de,pcchuang2@gmail.com (P.-C. Chuang).
1
Deceased.
www.elsevier.com/locate/gca
Available online at www.sciencedirect.com
ScienceDirect
Geochimica et Cosmochimica Acta 246 (2019) 138–155

1. INTRODUCTION
Accurate assessment of the sources of methane (CH
4
)to
the atmosphere is needed to better quantify the global car-
bon cycle and its response to climate change (IPCC, 2013).
Most of the CH
4
on Earth is found in marine sediments in
dissolved, gaseous or solid (hydrated) form (Reeburgh,
2007). Only a small fraction of this sedimentary CH
4
reser-
voir reaches the ocean and atmosphere because most of it is
consumed by microorganisms in a process known as the
anaerobic oxidation of methane (AOM) coupled to the
reduction of sulfate (SO
4
2
)(Barnes and Goldberg, 1976;
Boetius et al., 2000; Reeburgh, 2007). AOM takes place
in a sediment layer known as the sulfate-methane transition
(SMT). The SMT broadly defines the interface separating
the sediment layers where organic matter is predominantly
respired by sulfate reduction (above the SMT) and by
methanogenesis (below the SMT). Its depth may vary from
a few centimeters to hundreds of meters below the seafloor
depending on several factors, including the supply of CH
4
from deeper sediments and the amount and reactivity of
organic material reaching the seafloor (e.g., Meister et al.,
2013). CH
4
cycling in and around the SMT and the effi-
ciency of AOM in preventing CH
4
from escaping to the
atmosphere has been the focus of intense research in recent
decades (e.g., Iversen and Jørgensen, 1985; Hinrichs et al.,
1999; Boetius et al., 2000).
In low temperature sedimentary environments, fermen-
tation and hydrolysis of macromolecular organic com-
pounds into simpler moieties such as H
2
and acetate
provides important precursors for CH
4
production by
anaerobic microbes (e.g., Wellsbury and Parkes, 2000;
Rabus et al., 2006). In marine settings, CO
2
reduction by
H
2
is the major pathway of CH
4
production (Whiticar,
1999; Heuer et al., 2009). The turnover rates of CH
4
in
and below the SMT can be quantified using reaction-
transport models that explicitly include stable C isotopes
(d
13
C) of CH
4
and dissolved inorganic carbon (DIC)
(Sivan et al., 2007; Zeebe, 2007; Ussler and Paull, 2008;
Chatterjee et al., 2011; Malinverno and Pohlman, 2011;
Burdige et al., 2016). These approaches consider that both
CO
2
reduction and AOM are associated with normal
kinetic isotope effects (KIE), that is, preferential uptake
of the lighter isotope by microorganisms (Whiticar, 1999).
Microbial fractionation during the cycling of DIC and
CH
4
commonly leads to minimum d
13
C-DIC and d
13
C-
CH
4
values in the SMT (e.g., Borowski et al., 1997;
Ussler and Paull, 2008; Pohlman et al., 2008; Treude
et al., 2014). However, minimum values of d
13
C-CH
4
are
not consistent with conventional isotope systematics
whereby
13
C-enrichment of CH
4
in the SMT is expected
due to preferential
12
CH
4
consumption by AOM. Addition
of
12
C-enriched CH
4
to SMT porewaters, or just below it,
by methanogenesis has been hypothesized to counteract
the production of
13
C-depleted DIC by AOM (Borowski
et al., 1997). More recently, carbon isotope equilibrium
between DIC and CH
4
during AOM at low SO
4
2
concen-
trations has been proposed as an alternative explanation
for d
13
C-CH
4
minima (Yoshinaga et al., 2014). In vitro
experiments by these workers showed that backwards
AOM allows the residual CH
4
pool to be progressively
depleted in
13
CH
4
whilst enriching the DIC pool in DI
13
C.
Our objective here is to explore whether the ideas out-
lined in the previous paragraph are consistent with
observed d
13
C-DIC and d
13
C-CH
4
minima in the SMT in
sediment cores retrieved from the northeastern South China
Sea, offshore SW Taiwan. We use a reaction-transport
model that includes the stable carbon isotope dynamics of
CH
4
and DIC. To our knowledge, no studies have exam-
ined carbon isotope equilibrium during AOM using an
empirical reaction-transport model tuned to field data.
We conclude that backward AOM is occurring in the South
China Sea sediments. Our findings contribute to ongoing
efforts to understand carbon cycling in and around the
SMT (Burdige et al., 2016; Flury et al., 2016; Komada
et al., 2016; Beulig et al., 2018a, 2018b).
2. METHODS
2.1. Site description
The geology offshore southwestern Taiwan (northeast-
ern South China Sea) is characterized by an active subduc-
tion zone separated from the Chinese passive continental
margin by a deformation front (Liu et al., 2006; Lin
et al., 2008)(Fig. 1). The active margin displays a series
of ridges, troughs, mud diapiric structures and submarine
mud volcanoes (Lin et al., 2008; Lin et al., 2009; Hsu
et al., 2014). On the passive margin, the main structural fea-
tures are marked by normal faults with numerous canyon
incisions. Bottom Simulating Reflectors on both margins
have been reported by Liu et al. (2006). Geothermal gradi-
ents and heat flow data (Chi and Reed, 2008) have been
linked to upward advection of pore fluids at all of the cor-
ing sites studied here (Chen et al., 2012). Further details on
geological setting have been reported elsewhere (Huang
et al. 2006; Liu et al. 2006; Lin et al. 2008).
2.2. Sampling and analytical methods
Five giant piston cores (MD10-3262, MD10-3287,
MD10-3290, MD10-3291 and MD10-3293) were recovered
during R/V Marion Dufresne cruise MD178/MARCO
POLO 1/IMAGES XII in May/June 2010 to investigate
the geochemistry of the active and passive margins (Fig. 1
and Table 1). After core retrieval, sediments were sampled
in 6 or 7 cm-long segments every 50 cm. Sediment
subsamples (15 mL) for hydrocarbon gas composition and
d
13
C–CH
4
analysis were obtained from the 6 cm-long sam-
ples using plastic syringes with the tips cut off. The samples
were immediately extruded into 30 mL glass serum bottles
filled with saturated NaCl solution and sealed with butyl
stoppers and aluminum crimp caps. Following this, 5 ml
of the saturated NaCl solution was replaced by helium
gas to generate a headspace. Porewater CH
4
was fully equi-
librated with the headspace after shaking the sealed serum
bottles for ten minutes, with at least three minutes of stand-
ing equilibration time. The d
13
C and dD isotopic composi-
tion of CH
4
was measured on a Finnigan Delta-Plus
XL
isotope ratio mass spectrometer (IRMS) with a HP 5890
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 139

GC and a GC-combustion III interface at the Exploration
and Development Research Institute, Chinese Petroleum
Corporation (CPC, Taiwan). Analytical details have been
described by Sun et al. (2010). The precision of repeated
analyses (1r) was ±0.5‰for d
13
C and ±3‰for dD. The
isotopic composition for individual C compounds are
reported as d-values (in ‰) relative to the international
standards; Vienna Peedee Belemnite (VPDB) for d
13
C and
Vienna Standard Mean Ocean Water (VSMOW) for dD.
Porewater hydrocarbon gas concentrations were measured
as described in Hu et al. (2017).
Comparison of porewater data from the piston cores
and box cores suggests that up to 10–20 cm of the surface
sediments was lost during piston coring (Horng, 2010).
The data reported here are for piston cores only and sedi-
ment depths are uncorrected for surface sediment loss.
Sediment pore fluids for analysis of d
13
C–DIC and dis-
solved ions were extracted from the 7 cm-long sediment sec-
tion on board under pressure and filtered through a
0.22 mm nylon membrane syringe filter. The porewater sam-
ples were then split into two or three 5 mL polypropylene
vials without headspace. Samples for cation analysis
(barium (Ba
2+
), total dissolved iron (dFe), ammonium
(NH
4
+
), calcium (Ca
2+
), and magnesium (Mg
2+
)) were acid-
ified with 0.1 mL of 8 N nitric acid to prevent oxidation and
precipitation. All porewater samples were preserved at 4 °C
until analysis. Porewater d
13
C-DIC was measured at the
Department of Geosciences (National Taiwan University)
using a total organic carbon (TOC) analyzer (OI Analyti-
cal) combined with a Picarro G1101–i cavity ring down
spectrometer isotopic analyzer (CRDS). A total of
10–15 mL of porewater, without dilution, was treated with
5% H
3
PO
4
in a glass vial at 25 °C. Total dissolved inorganic
carbon (DIC), produced as CO
2
, was stripped from the
sample with N
2
and introduced into the detectors. The ana-
lytical precision for d
13
C-DIC was ±0.5‰. Porewater Ba
2+
Fig. 1. Bathymetric map of offshore southwestern Taiwan showing the five coring sites recovered during MD178/MARCO POLO 1/
IMAGES XII cruise of R/V Marion Dufresne. The deformation front is defined according to Han et al. (2017).
Table 1
Sampling locations, water depths and piston core recovery lengths.
Site/core Latitude (N) Longitude (E) Water depth (m) Core length (m)
MD10–3262 22°06024.100 119°17026.500 1200 29.20
MD10–3287 22°29013.900 119°41045.600 1050 24.65
MD10–3290 21°52059.900 120°20008.500 1272 15.65
MD10–3291 21°41029.400 120°13054.500 2070 38.50
MD10–3293 22°16037.900 120°01032.500 1004 22.72
140 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

and total dFe concentrations were measured by ICP-OES
and NH
4
+
concentrations were determined using ion chro-
matography (882 Compact IC). The analytical errors of
Ba
2+
, dFe, and NH
4
+
were less than 5%, respectively. Pore-
water DIC, SO
4
2
,Ca
2+
, and Mg
2+
concentrations were
measured as described by Hu et al. (2017). POC was deter-
mined following Lin (2010) and porosity as described by
Chen (2011).
2.3. Reaction-transport model
2.3.1. Model set-up
To determine and quantify the processes controlling
porewater d
13
C-CH
4
and d
13
C-DIC distributions in the
upper 30 m of sediments, the data from sites MD10-3287,
MD10-3291 and MD10-3293 were simulated with a
steady-state coupled reaction-transport model. Here we
mainly focus on MD10-3287 where the most comprehensive
data set is available as well as d
13
C-CH
4
data above the
SMT. Model results for the other two sites are provided
in the Supplementary Information.
The model simulates the distributions of POC, SO
4
2
,
total CH
4
(=
12
CH
4
+
13
CH
4
),
13
CH
4
,NH
4
+
, total DIC
(=DI
12
C+DI
13
C), DI
13
C, Ca
2+
and Mg
2+
. Biogeochemi-
cal reactions included are listed in Table 2. The sources
and sinks of the model variables are listed in Table 3, and
Table 4 provides the model parameters. Further model
details of the model are described in the Supplementary
Information.
Vertical depth profiles of the dissolved species and POC
were simulated using 1-D mass conservation equations
(Berner, 1980; Boudreau, 1997):
uxðÞ
@Ciðx;tÞ
@t¼
@uðxÞDSðxÞ@Ciðx;tÞ
@x

@x
@uðxÞvðxÞCiðx;tÞðÞ
@xþuðxÞRRðx;tÞ
ð1Þ
1uðxÞðÞ
@POCðx;tÞ
@t¼@1uðxÞðÞwðxÞPOCðx;tÞðÞ
@x
þð1uðxÞÞ  RRðx;tÞð2Þ
where x(cm) is sediment depth, t(yr) is time, uis porosity,
D
S
(cm
2
yr
1
) is the solute–specific diffusion coefficient in
the sediment, C
i
(mmol cm
3
of porewater) is the concentra-
tion of solute i, POC is the content of POC (dry weight per-
cent, wt.%), v(cm yr
1
) is the net velocity of solutes by
burial and compaction in addition to upward fluid flow
imposed at the lower boundary of the model, w(cm yr
1
)
is the burial velocity of solids and RRis the sum of biogeo-
chemical reactions (Table 3). Constant concentrations for
solutes and a fixed POC content were imposed at the upper
boundary (Table 4). At the lower boundary, a zero-gradient
condition was used for POC. Due to the upward-migrating
fluid, the model is solved with fixed solute concentrations at
the lower boundary.
The rate of POC mineralization (R
POC
, wt.% yr
1
) was
defined using the continuum model proposed by
Middelburg (1989) and modified by Wallmann et al. (2006):
RPOC ¼KC
CDIC þCCH4þKC
0:16 POC0:95
age

POC ð3Þ
where C
DIC
and C
CH4
are the simulated concentrations of
DIC and CH
4
, respectively, and K
C
(mM) is the inhibition
coefficient for POC degradation (Wallmann et al., 2006).
This has the effect of slowing down POC mineralization if
CH
4
and DIC accumulate in the porewater. POC
age
is the
age of the sediment layer where POC is being degraded:
POCage ¼a0þx
wf
þufu0

epx1ðÞ
pwfuf1
 ð4Þ
Table 2
Biogeochemical reactions used in the model.
Process Reaction stoichiometry
Organoclastic sulfate reduction, R
SR
CH2OðNH3ÞrNþ0:5SO2
4þrNCO2þrNH2O!1þrN
ðÞHCO
3þ0:5H2SþrNNHþ
4
Methanogenesis, R
Ma
CH2OðNH3ÞrNþrNHþ!0:5CO2þ0:5CH4þrNNHþ
4
Anaerobic oxidation of methane, R
AOM
CH
4
+SO
4
2
?HS

+ HCO
3

+H
2
O
CaCO
3
precipitation, R
CP
Ca
2+
+CO
3
2
?CaCO
3
MgCO
3
precipitation, R
MP
Mg
2+
+CO
3
2
?MgCO
3
a
The net reaction for methanogenesis summarizes POC mineralization to CO
2
and the reduction of CO
2
to CH
4
(e.g. Burdige et al. 2016):
CH
2
O(NH
3
)r
N
+H
2
O+r
N
H
+
?CO
2
+2H
2
+r
N
NH
4
+
.
0.5CO
2
+2H
2
?0.5CH
4
+H
2
O.
These reactions are considered separately for quantifying the isotope dynamics (see Table 3).
Table 3
Rate expressions describing the biogeochemical reactions (RRin
Eqs. (1) and (2)).
Variable Rates
a
SO2
40:5RSR RAOM
NHþ
4RAMF
CH40:5RMRAOM
13CH40:5URM
DIC RM13RAOM k
DIC
b
fcRPOC 0:5RMþRAOM RCP RMP
DI
13
C
b
fc13RPOC 0:5URM
DIC RM
þ13RAOM kUDIC ðRCP þRMP Þ
Ca2þRCP
Mg2þRMP
POC RPOC
a
fcis defined in Eq. (6),UDIC is defined in Eq. (15), and URM
DIC is
defined in Eq. (19).
b
The DIC mass balance explicitly considers the gross production
of DIC by POC degradation in addition to the loss of DIC by
methanogenesis (see Table 2).
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 141

The parameter a
0
(yr) defines the initial age of organic
matter or the reactivity of organic matter being degraded
(Middelburg, 1989) and other parameters in Eq. (4) are
listed in Table 4. Low a
0
enhances POC degradation close
to the sediment surface whereas high a
0
allows POC to be
buried to greater depths. a
0
can be considered as a fitting
parameter constrained mainly from POC, DIC and NH
4
+
concentrations. A similar approach was used to simulate
CH
4
dynamics in nearby sediments offshore SW Taiwan
(Chuang et al., 2013).
POC was defined chemically as CH
2
O(Table 2) with an
oxidation state of zero (e.g., Chuang et al., 2013; Luo et al.,
Table 4
Imposed and best-fit (in bold) parameters for site MD10-3287.
Parameter Description Value Unit Source
Physical parameters
T Bottom water temperature 278.5 K a
L Length of sediment column 3000 cm a
S Bottom water salinity 34.6 – a
ds Dry sediment density 2.5 g cm
3
b
w
f
Burial velocity of compacted sediment 0.08 cm yr
1
c
MAR Mass accumulation rate (ds (1 u
f
)w
f
) 0.112 g cm
2
yr
1
a
P Pressure at seafloor 107 bar a
v
0
Upward fluid velocity 0.12 cm yr
1
a
u0Sediment porosity at zero depth 0.64 – a
ufSediment porosity at infinite depth 0.44 – a
pDepth attenuation coefficient of porosity 1/1160 cm
1
a
DSO4Diffusion coefficient for SO
4
2
180 cm
2
yr
1
d
DCH4Diffusion coefficient for CH
4
,
13
CH
4
294 cm
2
yr
1
d
DNH4Diffusion coefficient for NH
4
+
345 cm
2
yr
1
d
DDIC Diffusion coefficient for DIC, DI
13
C
*
190 cm
2
yr
1
d
DCa Diffusion coefficient for Ca
2+
133 cm
2
yr
1
d
DMg Diffusion coefficient for Mg
2+
124 cm
2
yr
1
d
Biogeochemical parameters
KSO4SO
4
2
limitation constant 0.2 lmol cm
3
e
k
AOM
Kinetic constant for R
AOM
0.04 lmol
1
cm
3
yr
1
a
k
CP
Kinetic constant for carbonate precipitation 0.05 yr
1
a
k
MP
Kinetic constant for carbonate precipitation 0.05 yr
1
a
r
N
N:C mineralization ratio of POC 16/106 - c
a
0
Initial age of organic matter 1000 kyr a
K
C
Inhibition coefficient for POC degradation 40 lmol cm
3
b
eDIC;CO2red KIE for DIC by CO
2
reduction 40 ‰a
eCH4;AOM fKIE for CH
4
by AOM forward reaction 12 ‰f
eDIC;AOMbKIE for DIC by AOM back reaction 94 ‰f
eDIC;CP KIE for DIC by (Ca,Mg)CO
3
precipitation 0 ‰g
d
13
C–POC d
13
C of POC in surface sediments -24 ‰h
r
1
AOM reversibility ratio for CSO2
4>T1SO2
40.047 – i
r
2
AOM reversibility ratio for T1SO2
4>CSO2
4>T2SO2
40.99 – a
r
3
AOM reversibility ratio for CSO2
4<T2SO2
410
14
/(10
14
+1) – a
T1SO2
4Higher SO
4
2
threshold concentration 0.05 lmol cm
3
a
T2SO2
4Lower SO
4
2
threshold concentration 5 10
6
lmol cm
3
a
b1 Parameter determining shape of the error function 1 10
3
lmol cm
3
a
b2 Parameter determining shape of the error function 1 10
6
lmol cm
3
a
Boundary conditions
CSO2
4x¼0;LUpper/lower boundary concentration for SO
4
2
28.4/0 lmol cm
3
a
CCH4x¼0;LUpper/lower boundary concentration for CH
4
110
20
/18 lmol cm
3
a
CNHþ
4x¼0;LUpper/lower boundary concentration for NH
4
+
0/8 lmol cm
3
a
CDICx¼0;LUpper/lower boundary concentration for DIC 5.5/5.5 lmol cm
3
a
CCa2þx¼0;LUpper/lower boundary concentration for Ca
2+
10/3.5 lmol cm
3
a
CMg2þx¼0;LUpper/lower boundary concentration for Mg
2+
48/35 lmol cm
3
a
d13CCH4x¼0;LUpper/lower boundary value for CH
4
-44/-65 ‰a
d13CDICx¼0;LUpper/lower boundary value for DIC -10/-7 ‰a
CPOCx¼0POC content at upper boundary 0.6 wt.% a
POCx¼LLower boundary condition for POC oPOC/ox= 0 wt.% cm
1
a
Source:
a
This study;
b
Wallmann et al. (2006);
c
Chuang et al. (2013);
d
Boudreau (1997);
e
Habicht et al. (2002);
f
Yoshinaga et al. (2014);
g
Teichert et al. (2005);
h
Kao et al. (2006);
i
Holler et al. (2011).
*
Assumed to be equivalent to bicarbonate ion (HCO
3

), which constitutes >95% of DIC at the typical pH (8) of sediment porewaters.
142 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

2015). Its degradation is stoichiometrically linked to the
rate of ammonification, R
AMF
(mmol cm
3
yr
1
of NH
4
+
)
that releases NH
4
+
to the porewater at a constant (Redfield)
molar ratio, r
N
(mol N (mol C)
1
) (e.g., Chuang et al., 2013;
Luo et al., 2015):
RAMF ¼rNfCRPOC ð5Þ
where the factor f
C
converts R
POC
from wt.% to mmol cm
3
of porewater:
fC¼ds 1uðxÞðÞ104
12 uðxÞð6Þ
and where ds (g cm
3
) is the dry sediment density.
The rate of POC degradation by organoclastic sulfate
reduction (SR), R
SR
(mmol cm
3
yr
1
) and methanogenesis,
R
M
,(mmol cm
3
yr
1
) are derived from R
POC
:
RSR ¼fCfSO4RPOC ð7Þ
RM¼fC1fSO4
ðÞRPOC ð8Þ
with RPOC fC¼RSR þRMð9Þ
The factor f
SO4
is a rate-limiting term with a half satura-
tion constant, K
SO4
, that determines the extent to which SR
is limited by SO
4
2
and to which methanogenesis is inhibited
by SO
4
2
:
fSO4 ¼CSO4
CSO4þKSO4
ð10Þ
The rate of AOM, R
AOM
(mmol cm
3
yr
1
of CH
4
), was
expressed using bimolecular kinetics (e.g., Regnier et al.,
2011):
RAOM ¼kAOM CCH4CSO2
4ð11Þ
where k
AOM
is the empirical rate constant constrained by
the curvature in the SO
4
2
profile at the base of the SMT
(e.g. Dale et al., 2008). Low k
AOM
values lead to a large
overlap of SO
4
2
and CH
4
, that is, a very broad SMT,
and vice versa.
Finally, the rate of removal of calcium (R
CP
,mmol cm
3
yr
1
of Ca
2+
) and magnesium (R
MP
,mmol cm
3
yr
1
of
Mg
2+
) by carbonate precipitation was defined as a fitting
function that minimizes the difference between the observed
(OBS) and simulated concentrations:
RCP ¼kCP CCa CCa OBSðÞ
 ð12Þ
RMP ¼kMP CMg CMg OBSðÞ
 ð13Þ
The rate constants k
CP
and k
MP
were set to high values
to ensure that the terms in parentheses were close to zero,
meaning that the model simulated the measured concentra-
tions (Wallmann et al., 2006).
2.3.2. Stable carbon isotopes
Carbon isotopes were simulated by defining mass bal-
ance equations for the total concentration of DIC and
CH
4
and the corresponding concentration of
13
C-bearing
species, i(DI
13
C,
13
CH
4
). Carbon isotopes are reported rel-
ative to the PDB scale (
13
C/
12
C)
PDB
= 0.011237):
d13Ci¼
12C=13 CðÞ
i
12C=13 C
ðÞ
PDB
1

1000 ð14Þ
The d
13
C value is related to the mole fraction of
13
Cin
compound i,U
i
:
Ui¼
13Ci
12Ciþ13 Ci
¼
13Ci
Ci
¼d13Ciþ1000
d13Ciþ1000 þ1000=ð13 C=12CÞPDB
ð15Þ
The mole fraction of
13
C
i
produced or consumed by reac-
tion jwas calculated using the fractionation factor (a
i,j
)
(Burdige et al., 2016):
Uj
i¼
13Ci
ai;jCi13Ciðai;j1Þð16Þ
The rate of change of
13
C
i
due to biogeochemical reac-
tions was calculated by multiplying the reaction rate (R
j
)
by the corresponding value of Uj
i(Table 3). The fractiona-
tion factor was defined as the ratio of the rate constants (k)
of the light and heavy carbon isotopes (i.e.
L
k/
H
k). It is
related to the kinetic isotope effect, KIE (e
i,j
) as follows:
ei;j¼ai;j1

1000 ð17Þ
All fractionations considered here display a normal KIE,
that is, where the lighter isotope reacts faster than the heav-
ier one (a>0,e> 0).
Fermentation and hydrolysis of macromolecular organic
carbon to reactive intermediates such as hydrogen (H
2
) and
fatty acids followed by terminal oxidation to DIC is not
associated with significant fractionation of the carbon moi-
eties (Blair et al., 1994; Meyers, 1994). The rate of DI
13
C
produced by the breakdown of POC is thus equal to the
d
13
C of carbon being mineralized:
13RPOC ¼1000 þd13 CPOC
d13CPOC þ1000 þ1000=ð13 C=12CÞPDB
RPOC ð18Þ
where d
13
C
POC
is the d
13
C of sedimentary organic carbon
determined in sediments offshore SW Taiwan (-24.0‰,
Kao et al., 2006).
In marine sediments, reduction of DIC or, strictly, CO
2
,
by H
2
is the principal methane-producing pathway and asso-
ciated with significant fractionation of carbon (Whiticar,
1999). The
13
C mass balance for methanogenesis (R
M
)
implicitly accounts for fractionation during this step
(Table 3):
URM
DIC ¼DI13C
aDIC;CO2red DIC DI13 CðaDIC;CO2red 1Þð19Þ
The KIE for methanogenesis, eDIC;CO2red , was derived by
tuning the model to the data.
Finally, carbonate precipitation was assumed to occur
without a significant KIE (eDIC;CP ¼0; Teichert et al.,
2005). The incorporation of
13
C relative to
12
C into carbon-
ates is thus set by the d
13
C of DIC in the porewater
(Table 3).
2.3.3. AOM isotope dynamics
Carbon and sulfur back flux during AOM has been
shown to be dependent on the ambient SO
4
2
concentration
(Holler et al., 2011; Yoshinaga et al., 2014). The AOM back
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 143

reaction leads to a progressive enrichment and depletion in
12
CofCH
4
and DIC, respectively, when SO
4
2
concentra-
tion is below a certain threshold. This threshold has been
estimated to be 0.5 mM (Yoshinaga et al., 2014). Following
these workers, we introduce the term rto represent the ratio
between the backward and forward reaction rates (f/f+in
Yoshinaga et al., 2014). The net rate of DI
13
C and
13
CH
4
production and consumption by AOM can then be
described as:
13RAOM k¼X
k
13RAOM k¼X
k
fk
TSO2
4kAOM CCH4
CSO2
4URAOMk
CH4
1rk
URAOMk
DIC rk
1rk
!
ð20Þ
The first term in parenthesis defines the rate of the for-
ward reaction (f+) and the second term to the backward
reaction (f). The subscript ‘k’ refers to the AOM rate cor-
responding to the ambient SO
4
2
concentration. We
obtained the best data simulations assuming that backward
AOM depends on two SO
4
2
thresholds (TSO2
4), that is, with
k= 1, 2, 3, rather than the single threshold determined by
Yoshinaga et al. (2014). The dependence of ron TSO2
4is
illustrated schematically in Fig. 2.
The term fk
TSO2
4in Eq. (20) modulates the extent of the
backward reaction at each TSO2
4and was defined using
error functions:
f1
TSO2
4¼0:5þ0:5erf CSO2
4T1SO2
4
b1

ð21Þ
f2
TSO2
4¼0:5erfc CSO2
4T1SO2
4
b1

0:5þ0:5erf CSO2
4T2SO2
4
b2

ð22Þ
f3
TSO2
4¼0:5erfc CSO2
4T2SO2
4
b2

ð23Þ
where parameters b1 and b2 determine the transitional
overlap between R
AOM
at each TSO2
4(Table 4). These equa-
tions are written so that RAOM1corresponds to SO
4
2
concen-
trations greater than T1SO2
4and RAOM3to SO
4
2
concentrations below T2SO2
4.SO
4
2
concentrations between
these thresholds correspond to RAOM2.
Kinetic isotope effects for the forward reaction, eCH 4;AOMf
(12‰) and backward reaction, eDIC;AOM b(94‰), were taken
from the experimental results by Yoshinaga et al. (2014).
Sensitivity of the model to this parameterization is dis-
cussed later. The value of r
1
(0.047) was imposed a priori
based on observations the back flux under SO
4
2
replete con-
ditions (Holler et al., 2011). In other words, 4.7% of the
CH
4
undergoing AOM exchanges
13
C with
12
C from the
DIC pool when SO
4
2
concentrations are above (T1SO2
4Þ.
Values of r
2
,r
3
,T1
SO2
4and T2SO2
4were treated as fitting
parameters constrained using the d
13
C-DIC and d
13
C-CH
4
data.
3. RESULTS
3.1. Sediment geochemistry
Geochemical data are shown for site MD10-3287 in
Fig. 3. Data from all sampling locations are presented in
Fig. S1. We focus our discussion here on MD10-3287 since
the biogeochemical trends are the same at all locations (see
Hu et al., 2017). MD10-3287 also offers the most complete
d
13
C dataset for modeling purposes. The sites mainly differ
in the depth of the SMT that varies by up to 200 cm (Hu
et al., 2017).
SO
4
2
concentration decreased quasi-linearly from the
surface sediment to the SMT which extends over ca. 1 m
between 6.5 and 7.5 m depth. SO
4
2
and CH
4
concentrations
in the SMT were <1 mM due to near quantitative consump-
tion by AOM. In the following discussion, we occasionally
refer to this as the canonical SMT, because we later posit
that the true SMT actually extends over greater depths.
The maximum observed concentration of CH
4
below the
SMT was ca. 1 mM. As is typical for methanic sediments,
degassing of dissolved CH
4
following core retrieval caused
severe scatter in the data. CH
4
was the dominant hydrocar-
bon gas and only minor amounts of C
2+
gases were
detected (C
1
/(C
2
+C
3
)>10
3
;Hu et al., 2017).
Porewater DIC concentrations ranged between 4 and
17 mM with highest concentrations at the SMT. DIC con-
centrations decreased markedly below the SMT to 5mM
at 25 m depth. This contrasts with the more common obser-
vation of an increase in DIC concentrations below the SMT
due to methanogenesis (e.g. Malinverno and Pohlman,
2011). Decreases in DIC below the SMT have been attrib-
[SO
4
2-
]
f-/f+
r
1
r
2
T2
SO
4
2-
T1
SO
4
2-
r
3
0
0
R
AOM
3
R
AOM
2
R
AOM
1
Fig. 2. A schematic plot of the dependence of the ratio of the
AOM backward to forward flux (r
k
=f/f+, Yoshinaga et al.,
2014)onSO
4
2
concentration. Note that the transitions between r
1
and r
2
, and between r
2
and r
3
at the two sulfate threshold values are
defined using error functions (Eqs. (21)–(23)). The SO
4
2
concen-
tration domains corresponding to each AOM rate are shown.
144 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

uted to carbonate precipitation (Sivan et al., 2007;
Wehrmann et al., 2011) and upward advection of DIC-
poor fluid (Pohlman et al., 2008; Chatterjee et al., 2011;
Ussler and Paull, 2008). The modeling work points towards
the latter explanation for the sites investigated here (see
Section 3.2 Key model findings). NH
4
+
concentrations
increased with depth to a maximum concentration of
7.8 mM. The curvature in NH
4
+
between 10 and 25 m is evi-
dence for a source of NH
4
+
at that depth due to organic
matter mineralization. An apparent local minimum in
NH
4
+
centered at ca. 10 m depth has been observed else-
where and explained as NH
4
+
oxidation with SO
4
2
(Schrum et al., 2009).
POC content was around 0.6% in the top 8 m and then
decreased to ca. 0.4%. The increase in POC content in the
upper section occurred around 10 kyr ago, for a given sed-
imentation rate of 80 cm kyr
1
(Chuang et al., 2013). Sim-
ilar coeval increase in POC in the South China Sea has been
explained as a rise in primary production in the Holocene
(Luo et al., 2015).
The decrease in Ca
2+
and Mg
2+
concentrations down to
the SMT is likely caused by carbonate precipitation. Below
the SMT, Ca
2+
and Mg
2+
concentrations were constant,
implying that carbonate precipitation had ceased at this
point. Ba
2+
concentrations increased from 0.04 lM at the
surface toward a maximum of 2.18 lM at the SMT and
then gradually decreased with depth. Total dissolved iron
concentrations (dFe) were less than 3 lM with a concentra-
tion peak situated below the SMT at 15 m.
The smooth concave-down profiles of d
13
C-CH
4
below
the SMT suggest that d
13
C-CH
4
was not significantly influ-
enced by degassing during core recovery (Wallace et al.,
2000). d
13
C-CH
4
in all cores ranged from 68‰to
109‰with minimum values in the SMT (Figs. 3 and
S1). At site MD10-3287, d
13
C-CH
4
exhibited a minimum
of 101‰at 750 cm. CH
4
was relatively enriched in
13
CH
4
above the SMT due to near-complete consumption
by AOM occurring with a normal KIE. Although the
d
13
C-CH
4
could not be measured above the SMT at
the other sites due to analytical limitations, the trend of
13
C-depletion in CH
4
in the SMT was evident. Minimum
d
13
C-CH
4
values were similar (MD10-3262, 109‰;
MD10-3290, 100‰; MD10-3291, 107‰; MD10-3293,
107‰) with an average for all sites of 105 ± 4‰.
d
13
C-CH
4
distributions below the SMT in all cores ranged
from ca. 70‰to 90‰. The increase in d
13
C-CH
4
below
the SMT is caused by the increasingly heavy porewater DIC
being transformed into CH
4
.dD-CH
4
was >250‰which,
when plotted together with d
13
C-CH
4
, indicates a biogenic
CH
4
source based on CO
2
reduction (Fig. S2;Whiticar,
1999). The difference between d
13
C-CH
4
in the SMT and
d
13
C-CH
4
in sediments at the base of the core that are
Fig. 3. Modeled (curves) and measured (symbols) depth profiles of (a–f) solutes, (g–h) d
13
CofCH
4
and DIC, (i) POC, and (j) porosity at site
MD10-3287. Note the break in the methane concentrations axis. The seawater d
13
C-DIC (0‰) is not evident in the uppermost porewater data
point due to loss of the surface sediment layer by the piston coring procedure. Dd
13
C
CH4
is indicated in the d
13
C-CH
4
plot, and d
13
C-CH
4
data
shallower than 5 m are undefined due to negligible CH
4
concentrations. Canonical SMT depths are represented by the gray bars.
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 145

unaffected by AOM, termed Dd
13
C
CH4
by Yoshinaga et al.
(2014), was 32.8‰at MD10-3287 (Fig. 3g), and between 20
and 40‰at the other sites (Fig. S1b).
d
13
C-DIC displayed similar down-core trends as
d
13
C-CH
4
, yet with higher values (more
13
C-enriched). Min-
imum d
13
C-DIC values in the SMT were MD10-3262,
46‰; MD10-3287, 39‰; MD10-3290, 26‰; MD10-
3291, 25‰; MD10-3293, 41‰, with an average of
35 ± 9‰(Figs. 3 and S1a). The profiles of d
13
C-DIC in
the SMT tended to mirror those of bulk DIC, indicating
a source of
13
C-depleted DIC by AOM. In the SR zone,
d
13
C-DIC values became heavier toward the sediment sur-
face due to mixing with seawater with a d
13
Cof0‰.In
the methanogenic zone below the SMT, DIC became
13
C-enriched at all sites, reaching 10‰at site MD10-
3287 at 13 m depth. At this depth, a second abrupt change
in the isotopic gradient was observed, below which
d
13
C-DIC values remained constant to the bottom of the
core. This kink in the d
13
C-DIC profile was also observed
at MD10-3293 and, to a lesser extent, at MD10-3291. The
invariable d
13
C-DIC of ca. 10‰between 13 and 25 m
depth with attendant decrease in DIC concentration
(Fig. 3c and h) implies only minor rates of an
enzymatically-mediated DIC sink with a strong KIE, e.g.
methanogenesis.
3.2. Key model findings
Baseline model simulations for site MD10-3287, MD10-
3291 and MD10-3293 showed a very good correspondence
with the field data (Figs. 3 and S1). The exception was CH
4
due to the aforementioned degassing. Imposed and best-fit
model parameters for these sites are listed in Table 4 and in
the Supplementary Information. The data were simulated
with a sedimentation velocity of 40–80 cm kyr
1
deter-
mined for a nearby site offshore SW Taiwan (Chuang
et al., 2013). Given the apparent lack of carbonate precipi-
tation below 10 m depth (Fig. 3d), the simulated concave-
down DIC profile below the SMT could only be achieved
(at steady state) with upward fluid advection velocity of
120–150 cm kyr
1
. This is on the same order, but larger,
than the sediment burial velocity and creates a net
upward-directed flux of all solutes at the lower boundary.
Yet, the model does not perfectly capture the DIC gradient
below 10 m at site MD10-3287 (Fig. 3c), which suggests
that further refinement of the model may be necessary.
POC degradation rates were mainly constrained from
the NH
4
+
and DIC data. Changes in bulk POC content
are a poor indicator of POC degradation rate since only a
small fraction is ‘bioavailable’, with the remainder trapped
inside mineral matrices or selectively preserved (Hedges and
Keil, 1995). Depth-integrated rates of POC mineralization
(3.89 mmol m
2
yr
1
of C), SR (0.49 mmol m
2
yr
1
of
SO
4
2
) and methanogenesis (1.46 mmol m
2
yr
1
of CH
4
)
were similar to those obtained for offshore SW Taiwan
(Chuang et al., 2013)(Table 5). Upward advection at the
lower boundary supplied 84% of CH
4
in the model domain
(7.57 mmol m
2
yr
1
of CH
4
) versus only 16% by methano-
genesis (1.46 mmol m
2
yr
1
of CH
4
). All CH
4
was con-
sumed within the sediment column by AOM
(9.03 mmol m
2
yr
1
of CH
4
). AOM was also the major
sink for SO
4
2
, with SR accounting for only 5%. The model
also predicted a loss of DIC into Ca-Mg carbonates of
10 mmol m
2
yr
1
; equivalent to two-thirds of total DIC
input to the sediment column.
Modeled profiles of selected reaction rates for site
MD10-3287 showed the classical spatial segregation of
redox processes around the SMT (Fig. 4); a SR zone above
the SMT with decreasing rates through the SMT, a peak in
AOM in the SMT and the onset of methanogenesis within
and below the SMT (Parkes et al., 2007; Yoshioka et al.,
2010; Treude et al., 2005). The peak AOM rate
(11.8 nmol cm
3
yr
1
;Fig. 4b) was three orders-of-
magnitude higher than the rates of SR and methanogenesis.
This is due to the high input of CH
4
at the lower boundary
plus the focusing of AOM in a narrow sediment layer. Net
rates of AOM for different SO
4
2
thresholds (RAOMk) showed
vertical displacements, with maximum rates at 713 cm
(RAOM1), 813 cm (RAOM 2), and 992 cm RAOM3

depth. RAOM3
was five orders-of-magnitude lower than RAOM1since mod-
eled SO
4
2
concentrations at 10 m depth were in the
nanomolar range. These low rates were constrained mainly
from the d
13
C data (see below). AOM was thus occurring,
albeit at very low rates, 2–3 m below the canonical SMT
depth where SO
4
2
apparently disappeared.
Distributions of d
13
C-DIC and d
13
C-CH
4
were well-
simulated with the model by accounting for both a forward
Table 5
Modeled depth–integrated turnover rates (R) and fluxes (F) at the
top (x= 0) and bottom (x=L) of the sediment core at site MD10-
3287 (in mmol m
2
yr
1
) under steady state conditions. Positive
fluxes are directed downwards into the sediment and vice versa.
Rate or flux Value Unit
R
POC
3.89 mmol m
2
yr
1
of C
R
SR
0.49 mmol m
2
yr
1
of SO
4
2
R
M
1.46 mmol m
2
yr
1
of CH
4
RAMF 0.59 mmol m
2
yr
1
of NH
4
+
RRAOM 9.03 mmol m
2
yr
1
of CH
4
RAOM18.51 mmol m
2
yr
1
of CH
4
RAOM20.52 mmol m
2
yr
1
of CH
4
RAOM37.97 10
5
mmol m
2
yr
1
of CH
4
R
CP
5.44 mmol m
2
yr
1
of Ca
2+
R
MP
4.55 mmol m
2
yr
1
of Mg
2+
FDICðx¼0Þ3.42 mmol m
2
yr
1
of DIC
FSO2
4ðx¼0Þ9.52 mmol m
2
yr
1
of SO
4
2
FCH4ðx¼0Þ4.72 10
5
mmol m
2
yr
1
of CH
4
FCaðx¼0Þ3.99 mmol m
2
yr
1
of Ca
2+
FMgðx¼0Þ10.0 mmol m
2
yr
1
of Mg
2+
FNHþ
4ðx¼0Þ3.98 mmol m
2
yr
1
of NH
4
+
F13CH 4ðx¼0Þ5.01 10
7
mmol m
2
yr
1
of CH
4
FDI13 Cðx¼0Þ0.034 mmol m
2
yr
1
of DIC
FPOCðx¼0Þ560 mmol m
2
yr
1
of C
FDICðx¼LÞ1.95 mmol m
2
yr
1
of DIC
FSO2
4ðx¼LÞ0 mmol m
2
yr
1
of SO
4
2
FCH4ðx¼LÞ7.57 mmol m
2
yr
1
of CH
4
FCaðx¼LÞ1.46 mmol m
2
yr
1
of Ca
2+
FMgðx¼LÞ14.56 mmol m
2
yr
1
of Mg
2+
FNHþ
4ðx¼LÞ3.40 mmol m
2
yr
1
of NH
4
+
F13CH 4ðx¼LÞ0.079 mmol m
2
yr
1
of CH
4
FDI13 Cðx¼LÞ0.022 mmol m
2
yr
1
of DIC
FPOCðx¼LÞ556.1 mmol m
2
yr
1
of C
146 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

and backward pathway for AOM, although the model
result is biased towards the lower d
13
C-DIC values in the
SMT (Figs. 3 and S1). Fractionation of CH
4
during
AOM (eCH4;AOM f) in the model (5–16‰) is at the lower
end of range of 5‰to 39‰reported for diverse marine
habitats (Whiticar, 1999; Holler et al. 2009).
12
CH
4
is pref-
erentially returned to the porewater CH
4
pool in the back-
ward AOM reaction with a KIE determined by the in vitro
study, eDIC;AOM b, of 94–99‰(Yoshinaga et al., 2014). The
ratio of the backward-to-forward reaction (r
1
) is only
4.7% (Holler et al., 2009) when the modeled SO
4
2
concen-
tration is above the higher threshold concentration, T1SO2
4
(i.e. for RAOM1,Fig. 2). T1SO2
4ranged from 0.05 to 5 mM
across the simulated sites, which compares favorably to
0.5 mM determined in the laboratory study (Yoshinaga
et al., 2014).
The best-fit values of r
2
for the three simulated sites ran-
ged from 0.909 to 0.99, meaning around 91–99% of CH
4
exchanges carbon with DIC when SO
4
2
concentrations
are between T1SO2
4and T2SO2
4. For SO
4
2
concentrations
below 5 nM (T2SO2
4) the forward and backward reaction
rates are essentially in equilibrium at all sites with r
3
1.
The KIE for DIC (CO
2
) reduction to CH
4
,eDIC;CO2red , was
40‰for sites MD10-3287 and MD10-3291, which agrees
with values from culture experiments that cluster around
40–50‰(e.g. Botz et al., 1996; Londry et al., 2008; Penger
et al., 2012). Sensitivity tests of eDIC;CO2red suggest that 40‰
may be a maximum value for these sites (result not shown).
At MD10-3293, a higher eDIC;CO2red of 80‰was derived,
which is quite a substantial difference. Part of the variability
in the model parameters between the three sites may be
caused by uncertainties in fluid velocities and POC degrada-
tion rates. For instance, the POC reactivity parameter a
0
ranges from 10
3
kyr at MD10-3287 to 10
5
kyr at MD10-
3293, possibly due to varying degrees of surface sediment
loss at each site (Wallmann et al., 2006).
4. DISCUSSION
4.1. Environmental factors controlling d
13
C-DIC and
d
13
C-CH
4
distributions
Minimum values in d
13
C-DIC and d
13
C-CH
4
in the
SMT are commonly observed in the field (e.g. Borowski
Fig. 4. Modeled rate profiles (a) sulfate reduction (RSR), (b) net AOM (PRAOM ), (c) methanogenesis (RM), (d) AOM while CSO2
4
PT1SO2
4
(RAOM1), (e) AOM while T1SO2
4>CSO2
4
PT2SO2
4(RAOM2) and (f) AOM while CSO2
4<T2SO2
4(RAOM3). Canonical SMT depths are represented
by the gray bars.
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 147

et al., 1997; Paull et al., 2000; Pohlman et al., 2008; Burdige
et al., 2016). However, conventional isotope systematics
predict
13
C-enrichment of CH
4
in the SMT due to more
rapid oxidation of
12
CH
4
by methanotrophs (Whiticar,
1999). Addition of
12
C-enriched CH
4
to SMT porewaters,
or just below it, by methanogenesis has been put forward
as a hypothesis to counteract the removal of
12
CH
4
by
AOM (Borowski et al., 1997). Carbon cycling between
methanogenesis and AOM seems like a reasonable explana-
tion for low d
13
C-CH
4
because eDIC;CO2red eCH4;AOMf. Our
study objective was to use field data to test this and the
alternative experimentally-derived hypothesis that the
d
13
C-CH
4
minima are instead driven by enzymatically-
mediated carbon isotope equilibrium between forward
and backward AOM (Yoshinaga et al., 2014). Data from
three sites offshore southwestern Taiwan are consistent with
backward AOM when interpreted using a model that
accounts for biogeochemical reactions and physical trans-
port processes of
12
C and
13
C isotopes. In the following,
we discuss the robustness of this finding with regard to
the parameterization of the model. It turns out that the
CH
4
concentration at lower boundary, [CH
4
]
x=L
, the veloc-
ity of ascending pore fluids (v
0
), and the rate of POC degra-
dation (i.e. its reactivity) are key controls on d
13
C
distributions in our 30 m sediment column. We briefly
address each of these factors in turn and later show that
reasonable ranges of these parameters do not invalidate
the need to invoke carbon isotope equilibrium during
AOM in the model. A schematic summary of the direction
that these variables exert on d
13
C distributions and which
forms the basis of this discussion is shown in Fig. 5.
The first point to note is that the basal fluxes of DIC and
CH
4
at the lower boundary are decoupled with a CH
4
: DIC
ratio of 4(Table 5). If methanogenesis were the only pro-
cess controlling DIC and CH
4
dynamics in deeper sedi-
ments, a ratio closer to 1:1 would be expected. The
reason for the decoupling of boundary fluxes is beyond
the scope of the study, but could be connected to deep
hydrate destabilization or upward transport of gaseous
CH
4
(Liu et al., 2006; Chatterjee et al., 2011; Chuang
et al., 2013; Chen et al., 2014; Hsu et al., 2014; Burdige
et al., 2016). Regardless of the mechanism, oxidation of
external CH
4
provides a source of DIC beyond that
expected from a coupled system of organoclastic SR,
AOM and methanogenesis. In sediments without external
CH
4
input, d
13
C-DIC in the SMT should tend toward the
d
13
C of POC because the net sum of methanogenesis and
AOM is stoichiometrically indistinguishable from organ-
oclastic SR (Malinverno and Pohlman, 2011; Komada
et al., 2016). Oxidation of external CH
4
is expected to lower
the d
13
C value of porewater DIC in the SMT relative to the
d
13
C of degradable POC, since biogenic or thermogenic
CH
4
is generally depleted in
13
C relative to sedimentary
POC (Whiticar 1999). At site MD10-3287, d
13
C-DIC in
the SMT (39‰) is significantly more depleted than the
d
13
C of POC being degraded (24.0‰,Kao et al., 2006).
An increase in external CH
4
input with a d
13
C-CH
4
of
65‰, expressed in the model as an increase in [CH
4
]
x=L
,
indeed shows that a decoupling of the basal fluxes lowers
d
13
C-DIC in the SMT (Fig. 6, blue versus green curves).
d
13
C-CH
4
is much less sensitive to [CH
4
]
x=L
because the
d
13
C-CH
4
value at the lower boundary remains unchanged
(see also Fig. S3c).
The advective velocity of rising fluids at the lower model
boundary is another control on basal DIC and CH
4
fluxes,
and on the SMT depth relative to the seafloor. Weaker fluid
advection causes a downward shift in the SMT because the
basal CH
4
flux is lower, leading to lower rates of SO
4
2
con-
sumption by AOM (Fig. 6, blue versus black curves). High
fluid velocities have the opposite effect (orange versus black
curves). However, minimum d
13
C-DIC and d
13
C-CH
4
val-
ues are not greatly affected by the tested changes in v
0
.If
v
0
were a factor of 10 higher, representative of a seep site
for instance, d
13
C-DIC and d
13
C-CH
4
become markedly
13
C-enriched as the SMT becomes shallower (Fig. S3a).
This is opposite to the effect of increasing [CH
4
]
x=L
whereby
d
13
C-DIC becomes more
13
C-depleted as the SMT shallows
(Fig. S3c). We suggest that the increase in d
13
C-DIC with
high v
0
is caused by more intense mixing of seawater with
ad
13
Cof0‰as the SMT closer to the sediment surface
(Zeebe, 2007).
A shift toward more positive d
13
C-DIC and d
13
C-CH
4
values is generated with a decrease in the POC age param-
eter a
0
, that is, with an increase in the freshness or reactivity
Fig. 5. Schematic of the impact of environmental parameters and fractionations on d
13
C values. (a) The velocity of ascending pore fluids (v
0
),
the methane concentration at lower boundary ([CH
4
]
x=L
), the rate of POC degradation (R
POC
, expressed through a
0
), (b) the ratio of AOM
backward to forward flux (r
k
), and (c) kinetic carbon isotope effects (eDIC;CO2red ;eCH 4;AOMfand eDIC;AOM b). The arrows qualitatively indicate the
direction of the shift in d
13
C for an increase in the parameter values. Canonical SMT depths are represented by the gray bars.
148 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

of POC (Fig. 6, blue versus red curves; Fig. S3b). The SMT
becomes shallower with decreasing a
0
because of higher
rates of SO
4
2
consumption by organoclastic SR and
AOM. In this scenario, increased POC degradation adds
DIC with a d
13
C that is higher than d
13
C-CH
4
at the lower
boundary. The overall effect is a marked increase both
d
13
C-DIC and d
13
C-CH
4
, and hence a decrease in the iso-
topic fingerprint of backward AOM.
The term describing the difference between d
13
C-CH
4
in the SMT and d
13
C-CH
4
in sediments at the base of
the core (Dd
13
C
CH4
) has been suggested to reflect the
extent of the backward AOM reaction when plotted
against the diffusive flux of SO
4
2
to the SMT
(Yoshinaga et al., 2014). The idea is that lower SO
4
2
fluxes, that is, lower AOM rates, result in higher Dd
13
-
C
CH4
because the extent of carbon isotope equilibration
for the back reaction is inversely dependent on the rate
(Holler et al., 2011). Our data broadly agree with the cor-
relation between Dd
13
C
CH4
and the SO
4
2
flux based in
the database presented by Yoshinaga et al. (2014)
(Fig. S4). However, it is clear from the previous model
analysis that migrating fluids can also strongly affect
Dd
13
C
CH4
. On the one hand, Dd
13
C
CH4
can be greatly
reduced if fluid advection pushes the SMT close to the
sediment surface, keeping [CH
4
]
x=L
constant (Fig. S3a).
On the other hand, similar shallow SMT depths and,
by implication, similar diffusive SO
4
2
fluxes to the SMT
can be generated by keeping v
0
constant and increasing
Fig. 6. Model sensitivity analysis of upward fluid velocity, POC degradation rates (expressed through a
0
) and methane concentrations at the
lower boundary at site MD10-3287. Baseline values are shown in black curves and measured data as symbols. Canonical SMT depths for the
baseline scenario are represented by the gray bars.
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 149

[CH
4
]
x=L
, but with only minor changes in Dd
13
C
CH4
(Fig. S3c). Therefore, whilst Dd
13
C
CH4
may provide infor-
mation on the extent of backward AOM, the wider phys-
ical characteristics of the environment may also make an
important contribution to Dd
13
C
CH4
. This questions its
use as a proxy for the extent of backward AOM (in each
of the tested model scenarios the ratio of the backward-
to-forward reaction was kept constant). In the next sec-
tion, however, we show that an observable difference
between d
13
C-CH
4
in the SMT and at the base of the
core does require backward AOM.
4.2. Microbial fractionation and carbon equilibration during
AOM
A major finding of this work is that the isotope data can
only be simulated using SO
4
2
dependent backward AOM.
Without the back flux (rk¼0), d
13
C-DIC profiles shift to
more negative values below the SMT due to a lack of
DI
12
C transfer to CH
4
(Fig. 7a). Moreover, d
13
C-CH
4
and Dd
13
C
CH4
appear to be strong diagnostic indicators
that a back reaction is occurring. Without it, Dd
13
C
CH4
goes
to zero and d
13
C-CH
4
rises sharply below, in and above the
SMT in line with the normal KIE by AOM. With r
k
=0,
none of the tested combinations of eCH 4;AOMfand eDIC;CO2red
is able to reproduce the minimum in d
13
C-CH
4
. Without
backward AOM, methanogenesis cannot compensate for
preferential
12
CH
4
oxidation because methanogenesis rates
are a factor of 100 lower than AOM (Fig. 4). Increasing
eDIC;CO2red in the model thus does not help to redress the
balance (Fig. 7a, blue versus red curves). More generally,
conservation of isotopes between DIC and CH
4
in the
SMT prevents these constituents becoming simultaneously
lighter if both are being recycled between one and another
using conventional KIEs. According to the present model
results, the d
13
C-CH
4
minima in the SMT in the South
China Sea sediments cannot be solely driven by intertwined
methanogenesis and AOM (Borowski et al., 1997).
This conclusion is sensitive to the choice of lower
boundary conditions used in the model. For example, in a
steady state system with zero-gradient (no diffusive flux)
boundary conditions at the bottom of the core, no mini-
mum in d
13
C-CH
4
in the SMT is produced in the absence
of backward AOM (Burdige et al., 2016). Minimum values
Fig. 7. Model sensitivity analysis of (a) no backward AOM (rk= 0), eDIC;CO2red and eCH4;AOM f, (b) r1, (c) r2, (d) r3, (e) eCH 4;AOMf, and (f) eDIC;AOM b
on d
13
C-CH
4
and d
13
C-DIC profiles for site MD10-3287. Baseline values are shown in black curves. The blue curve for d
13
C-DIC is almost
identical to the green curve. Blue curves and black curves are in (b), and for d
13
C-DIC in (c) are also overlapping. Canonical SMT depths for
the baseline scenario are represented by the gray bars. (For interpretation of the references to color in this figure legend, the reader is referred
to the web version of this article.)
150 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

of d
13
C-DIC and d
13
C-CH
4
of 24‰and 90‰in Santa
Barbara Basin (SBB) have been simulated without invoking
backward AOM (Burdige et al., 2016). There is no external
fluid advection at this site, and the CH
4
and DIC fluxes at
the lower boundary (5 m) are apparently decoupled with a
CH
4
:DIC ratio of 5. In contrast to the cores studied here,
in SBB there is an upward diffusive flux of CH
4
at the lower
boundary with a d
13
C of -68‰. This flux is enriched in
13
CH
4
compared to the d
13
CofCH
4
produced in the model
domain by methanogenesis, which leads to a minimum in
d
13
C-CH
4
in the SMT even when backward AOM is
excluded. This dynamic may mask backward AOM at this
site. It is therefore imperative to define realistic lower
boundary conditions in order to interpret d
13
C-DIC and
d
13
C-CH
4
distributions in the SMT. This may not always
be straightforward given that degassing of CH
4
often pla-
gues the accurate determination of CH
4
concentrations
below the SMT.
The minimum d
13
C-CH
4
of 100‰observed in our
data is more negative than the d
13
C-CH
4
minimum of -
62‰reported in the experimental study (Yoshinaga et al.,
2014). d
13
C-CH
4
and d
13
C-DIC are insensitive to the
AOM backward flux when SO
4
2
concentrations are above
T1SO2
4(Fig. 7b). Instead, to simulate these extremely nega-
tive d
13
C-CH
4
values using the observed eDIC;AOM bof 94 and
99‰, a high ratio of the backward-to-forward AOM rate of
0.99 (r
2
) at site MD10-3287 was required for SO
4
2
concen-
trations between T1SO2
4and T2SO2
4(Yoshinaga et al., 2014)
(Fig. 7c). This compares to calculated r
2
values of up to 0.78
in the in vitro experiment (Yoshinaga et al., 2014). The dis-
crepancies in r
2
may be explained by (i) reaction rates in the
field being orders of magnitude lower than those in the lab-
oratory, and (ii) differences in the relative concentrations of
CH
4
and DIC. Given that enzymatic fractionation tends to
be more expressed under low reaction rates (Holler et al.,
2011), the slowly-reacting, more severely SO
4
2
limited con-
ditions in the field may allow near-equilibrium conditions
to become established.
The back flux corresponding to r
3
affects both d
13
C-DIC
and d
13
C-CH
4
below the SMT (Fig. 7d). These changes
take place when SO
4
2
concentrations are below T2SO2
4,
determined to be 5 nM. A second threshold for backward
AOM was not reported in the in vitro studies, possibly
because SO
4
2
concentrations did not drop below 40 mM
(Yoshinaga et al., 2014). Interestingly, the peak in
RAOM3(Fig. 4f) is just above the pronounced kink in the
d
13
C-DIC at 13 m depth. This feature is located several
meters below the canonical SMT and was simulated by
varying r
3
assuming fixed KIEs of the backward AOM
reaction. The AOM zone possibly extends the SMT deep
into the methanogenic zone, as hypothesized by
Yoshinaga et al. (2014). Although this model result is spec-
ulative, it is supported by the identification of anaerobic
methanotrophs (ANME-1 groups) below the canonical
SMT in nearby sediments offshore SW Taiwan (Lin et al.,
2014). These microorganisms are known to perform both
methanogenesis and methanotrophy under SO
4
2
limiting
conditions (Hoehler et al., 1994; Treude et al., 2007;
Lloyd et al., 2011; Timmers et al., 2017).
Based on these findings, we tentatively suggest that the
gradient change in d
13
C-DIC at 12–13 m depth marks the
lower vertical extent of the SMT. The kink disappears when
the back reaction is turned off (Fig. 7d). The sharpness of
the kink is affected by eCH4;AOM fand eDIC;AOM b, but not its
depth (Fig. 7e and f). Identical findings were made for site
MD10-3293 displaying a strong kink, but less so at MD10-
3291 where a kink was apparently absent (Fig. S1). A pro-
nounced gradient change in d
13
C-DIC profiles below the
canonical SMT has been observed in some settings (Sivan
et al., 2007; Ussler and Paull, 2008; Chatterjee et al.,
2011; Hiruta et al., 2015; Ijiri et al., 2018), but not in others
(Kim et al., 2011; Hiruta et al., 2015; Komada et al., 2016;
Gepra
¨gs et al., 2016). Our model suggests that this is caused
by a varying importance of backward AOM across sites,
although we cannot currently explain why this may be the
case. The kink is maintained for all the tested ranges of
environmental parameters, which would appear to rule
these out as the key factor (Fig. S3). The availability of
SO
4
2
below the SMT may be relevant, however, which
we discuss next.
4.3. A cryptic source of sulfate?
Backward AOM requires a source of SO
4
2
below the
canonical SMT. SO
4
2
is present there at our study sites
at concentrations of 1–2 mM (Figs. 3 and S1). The origin
of this pool is currently unclear, but may be related to a
cryptic sulfur cycle based on the oxidation of sulfide with
deeply buried iron (Holmkvist et al. 2011; Treude et al.,
2014). This requires reactive iron (Fe) to be present at all
sites displaying similar sharp kinks in d
13
C-DIC profiles,
including those in deep-seated sediments several tens of
meters below the seafloor (e.g. Sivan et al., 2007; Paull
et al., 2000). The relevant Fe pool that survives burial from
the surface to below the SMT on the time-scales applicable
to our sediments is silicate-bound, or poorly-reactive, Fe
with a half-life of 10
3
yr (Raiswell and Canfield, 1998).
For a typical poorly-reactive Fe content on the margins
of 7000 ppm (Raiswell and Canfield, 1998), and for sedi-
ment mass accumulation of 0.112 g cm
2
yr
1
(Table 4),
around 140 mmol m
2
yr
1
of Fe could be available for sul-
fidic dissolution. Considering that 8 moles of Fe are
required to produce 1 mole of SO
4
2
(Holmkvist et al.
2011), cryptic sulfur cycling could support the in situ pro-
duction of 18 mmol m
2
yr
1
SO
4
2
, which is double the
total rate of AOM. Indeed, a maximum in dissolved Fe
occurs close to the kink in d
13
C-DIC (Fig. 3). Yet, the
potential rate of SO
4
2
production is likely to be much lower
than this if sulfide becomes limited by iron sulfide precipita-
tion (Holmkvist et al. 2011). Dissolved hydrogen sulfide
was not detected in the sediment cores analyzed in this
study.
Dissolution of barite below the SMT may be another
important source of SO
4
2
(Torres et al., 1996; Riedinger
et al., 2006; Treude et al., 2014). Free dissolved Ba
2+
within
the SMT in core MD10-3287 apparently signifies barite dis-
solution and mobilization of SO
4
2
(Fig. 3). Multiplying a
typical sedimentary barite content of 500 ppm (Dickens,
2001; Riedinger et al., 2006) with the mass accumulation
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 151

rate (0.112 g cm
2
yr
1
) gives a potential mineral SO
4
2
sup-
ply of 2.4 mmol m
2
yr
1
. This is around five times the rate
of RAOM2þRAOM 3(Table 5). Whilst speculative, these simple
calculations illustrate the potential for low rates of AOM
far below the canonical SMT and a possible explanation
for the presence or absence of a secondary kink in d
13
C-
DIC profiles from other sites around the world.
5. CONCLUSIONS
Minimum values of d
13
C-CH
4
and d
13
C-DIC were mea-
sured in the SMT in anaerobic sediments at five sites off-
shore Taiwan; features that have been observed at several
sites around the world. The data were simulated using a
diagenetic model that accounts for diffusive and advective
processes as well as biogeochemical reactions including
organoclastic sulfate reduction, methanogenesis, AOM
and carbonate precipitation. The data were used with the
model to test the hypothesis that d
13
C-CH
4
minima in the
SMT are driven by carbon isotope equilibrium between
AOM forward and back reactions at low SO
4
2
concentra-
tions, as demonstrated in vitro using natural sediments
(Yoshinaga et al. 2014). Simulations with an accompanying
sensitivity analysis showed that the d
13
C-CH
4
data could
only be simulated by including backward AOM. This also
leads to a second pronounced kink in the d
13
C-DIC profile
several meters below the canonical SMT where SO
4
2
con-
centrations are diminished. We suggest that this kink marks
the true base of the SMT where forward and backward
AOM are operating at very low rates, possibly sustained
by cryptic sulfur cycling. This work highlights the need to
carefully constrain POC degradation rates, fluid flow veloc-
ities and CH
4
concentrations at the lower model boundary
to properly interpret isotope distributions in the SMT using
models, and the relevance of backward AOM in other set-
tings. Future work should consider extending this model to
include the carbonate system and its isotopes in order to
evaluate the impact of backward AOM on d
13
C signals of
carbonate in the paleorecord.
ACKNOWLEDGEMENTS
We thank the captain and crew of the R/V Marion Dufresne
and all of the MD178 cruise participants for their help. Specially
thank to the Yang research group members provided invaluable
assistance with sample collection and analysis. We thank AE John
W. Moreau and an anonymous reviewer for their detailed com-
ments on our manuscript. This work was funded by the Central
Geological Survey MOEA, Taipei, Taiwan, (Grant No. 101-
5226904000-06-01) to T.F.Y.. P.C.C. was benefited from the schol-
arship from NSC-DAAD Sandwich Program for research visits to
GEOMAR, Germany (NSC99-2911-I-002-053-2) and funded
through DFG-Research Center/Cluster of Excellence ‘‘The Ocean
in the Earth System”(Sediment Geochemistry).
APPENDIX A. SUPPLEMENTARY MATERIAL
Supplementary data associated with this article can be
found, in the online version, at https://doi.org/10.1016/j.
gca.2018.11.003.
REFERENCES
Barnes R. O. and Goldberg E. D. (1976) Methane production and
consumption in anoxic marine sediments. Geology 4, 297–300.
Berner R. A. (1980) Early Diagenesis: A Theoretical Approach.
Princeton University Press, Princeton, p. 241.
Beulig F., Røy H., Glombitza C. and Jørgensen B. B. (2018a)
Control on rate and pathway of anaerobic organic carbon
degradation in the seabed. Proc. Natl. Acad. Sci. 115, 367–
372.
Beulig F., Røy H., McGlynn S. E. and Jørgensen B. B. (2018b)
Cryptic CH4 cycling in the sulfate–methane transition of
marine sediments apparently mediated by ANME-1 archaea.
ISME J..
Blair N. E., Plaia G. R., Boehme S. E., DeMaster D. J. and Levin
L. A. (1994) The remineralization of organic carbon on the
North Carolina continental slope. Deep Sea Res. Part II 41,
755–766.
Boetius A., Ravenschlag K., Schubert C. J., Rickert D., Widdel F.,
Gieseke A., Amann R., Jorgensen B. B., Witte U. and
Pfannkuche O. (2000) A marine microbial consortium appar-
ently mediating anaerobic oxidation of methane. Nature 407,
623–626.
Borowski W. S., Paull C. K. and Ussler W. (1997) Carbon cycling
within the upper methanogenic zone of continental rise
sediments; An example from the methane-rich sediments
overlying the Blake Ridge gas hydrate deposits. Mar. Chem.
57, 299–311.
Botz R., Pokojski H.-D., Schmitt M. and Thomm M. (1996)
Carbon isotope fractionation during bacterial methanogenesis
by CO2 reduction. Org Geochem. 25, 255–262.
Boudreau B. P. (1997) Diagenetic Models and Their Implementa-
tion: Modelling Transport and Reactions in Aquatic Sediments.
SpringerVerlag, Berlin, 414p.
Burdige D. J., Komada T., Magen C. and Chanton J. P. (2016)
Methane dynamics in Santa Barbara Basin (USA) sediments as
examined with a reaction-transport model. J. Mar. Res. 74,
277–313.
Chatterjee S., Dickens G. R., Bhatnagar G., Chapman W. G.,
Dugan B., Snyder G. T. and Hirasaki G. J. (2011) Pore water
sulfate, alkalinity, and carbon isotope profiles in shallow
sediment above marine gas hydrate systems: a numerical
modeling perspective. J. Geophys. Res.: Solid Earth 116, n/a–
n/a.
Chen J. C. (2011) The relationship between physical properties of
cored sediments and gas hydrate accumulations in the gas
hydrate potential area offshore of southwestern Taiwan (4/4).
Rep. Central Geol. Surv. 2011 100-25-F, pp. 122 (in Chinese
with English abstract).
Chen L., Chi W.-C., Liu C.-S., Shyu C.-T., Wang Y. and Lu C.-Y.
(2012) Deriving regional vertical fluid migration rates offshore
southwestern Taiwan using bottom-simulating reflectors. Mar.
Geophys. Res. 33, 379–388.
Chen S.-C., Hsu S.-K., Wang Y., Chung S.-H., Chen P.-C., Tsai
C.-H., Liu C.-S., Lin H.-S. and Lee Y.-W. (2014) Distribution
and characters of the mud diapirs and mud volcanoes off
southwest Taiwan. J. Asian Earth Sci. 92, 201–214.
Chi W.-C. and Reed D. L. (2008) Evolution of shallow, crustal
thermal structure from subduction to collision: an example
from Taiwan. Geol. Soc. Am. Bull. 120(5–6), 679–690.
Chuang P.-C., Dale A. W., Wallmann K., Haeckel M., Yang T. F.,
Chen N.-C., Chen H.-C., Chen H.-W., Lin S., Sun C.-H., You
C.-F., Horng C.-S., Wang Y. and Chung S.-H. (2013) Relating
sulfate and methane dynamics to geology: accretionary prism
offshore SW Taiwan. Geochem. Geophys. Geosyst. 14, 2523–
2545.
152 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

Dale A. W., Regnier P., Knab N. J., Jørgensen B. B. and Van
Cappellen P. (2008) Anaerobic oxidation of methane (AOM) in
marine sediments from the Skagerrak (Denmark): II. Reaction-
transport modeling. Geochim. Cosmochim. Acta 72, 2880–2894.
Dickens G. R. (2001) Sulfate profiles and barium fronts in sediment
on the Blake Ridge: present and past methane fluxes through a
large gas hydrate reservoir. Geochim. Cosmochim. Acta 65, 529–
543.
Flury S., Røy H., Dale A. W., Fossing H., To
´th Z., Spiess V.,
Jensen J. B. and Jørgensen B. B. (2016) Controls on subsurface
methane fluxes and shallow gas formation in Baltic Sea
sediment (Aarhus Bay, Denmark). Geochim. Cosmochim. Acta
188, 297–309.
Gepra
¨gs P., Torres M. E., Mau S., Kasten S., Ro
¨mer M. and
Bohrmann G. (2016) Carbon cycling fed by methane seepage at
the shallow Cumberland Bay, South Georgia, sub-Antarctic.
Geochem. Geophys. Geosyst. 17, 1401–1418.
Habicht K. S., Gade M., Thamdrup B., Berg P. and Canfield D. E.
(2002) Calibration of Sulfate Levels in the Archean Ocean.
Science 298, 2372.
Han W.-C., Liu C.-S., Chi W.-C., Chen L., Lin C.-C. and Chen S.-
C. (2017) Westward advance of the deformation front and
evolution of submarine canyons offshore of southwestern
Taiwan. J. Asian Earth Sci. 149, 6–19.
Hedges J. I. and Keil R. G. (1995) Sedimentary organic matter
preservation: an assessment and speculative synthesis. Mar.
Chem. 49, 81–115.
Heuer V. B., Pohlman J. W., Torres M. E., Elvert M. and Hinrichs
K.-U. (2009) The stable carbon isotope biogeochemistry of
acetate and other dissolved carbon species in deep subseafloor
sediments at the northern Cascadia Margin. Geochim. Cos-
mochim. Acta 73, 3323–3336.
Hinrichs K.-U., Hayes J. M., Sylva S. P., Brewer P. G. and
DeLong E. F. (1999) Methane-consuming archaebacteria in
marine sediments. Nature 398, 802–805.
Hiruta A., Ishizaki O., Tuchinaga K., Snyder G. T. and
Matsumoto R. (2015) Influence of the carbon isotopic compo-
sition of methane and the proportion of methane-derived
bicarbonate on the 13C/12C ratio of dissolved inorganic carbon
at the sulfate–methane transition in the Joetsu Basin area,
eastern margin of the Sea of Japan. Mar. Pet. Geol. 67, 468–
480.
Hoehler T. M., Alperin M. J., Albert D. B. and Martens C. S.
(1994) Field and laboratory studies of methane oxidation in an
anoxic marine sediment: Evidence for a methanogen-sulfate
reducer consortium. Global Biogeochem. Cycl. 8, 451–463.
Holler T., Wegener G., Knittel K., Boetius A., Brunner B.,
Kuypers M. M. M. and 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.
Holler T., Wegener G., Niemann H., Deusner C., Ferdelman T. G.,
Boetius A., Brunner B. and Widdel F. (2011) Carbon and sulfur
back flux during anaerobic microbial oxidation of methane and
coupled sulfate reduction. Proc. Natl. Acad. Sci. 108, E1484–
E1490.
Holmkvist L., Ferdelman T. G. and Jørgensen B. B. (2011) A
cryptic sulfur cycle driven by iron in the methane zone of
marine sediment (Aarhus Bay, Denmark). Geochim. Cos-
mochim. Acta 75, 3581–3599.
Horng C. S. (2010) Palaeomagnetism, dating and magnetic
property studies in the sediments of the Gas Hydrate Potential
Area Offshore of Southwestern Taiwan (3/4). Rep. Central Geol.
Surv. 2010 99-26-E, pp. 54 . (in Chinese with English abstract).
Hsu S.-K., Chiang C.-W., Evans R. L., Chen C.-S., Chiu S.-D., Ma
Y.-F., Chen S.-C., Tsai C.-H., Lin S.-S. and Wang Y. (2014)
Marine controlled source electromagnetic method used for the
gas hydrate investigation in the offshore area of SW Taiwan. J.
Asian Earth Sci. 92, 224–232.
Hu C.-Y., Yang T. F., Burr G. S., Chuang P.-C., Chen H.-W.,
Walia M., Chen N.-C., Huang Y.-C., Lin S., Wang Y., Chung
S.-H., Huang C.-D. and Chen C.-H. (2017) Biogeochemical
cycles at the sulfate-methane transition zone (SMTZ) and
geochemical characteristics of the pore fluids offshore south-
western Taiwan. J. Asian Earth Sci. 149, 172–183.
Huang C.-Y., Chien C.-W., Zhao M.-X., Li H.-C. and Iizuka Y.
(2006) Geological study of active cold seeps in the syn-collision
accretionary prism Kaoping slope off SW Taiwan. Terrestr.
Atmosp. Ocean. Sci. 17, 679–702.
Intergovernmental Panel on Climate Change (IPCC) (2013)
Climate Change 2013: The Physical Science Basis. Contribution
of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge Univ.
Press, Cambridge, U.K., and New York.
Iversen N. and Jorgensen B. B. (1985) Anaerobic methane
oxidation rates at the sulfate-methane transition in marine
sediments from Kattegat and Skagerrak (Denmark)1. Limnol.
Oceanogr. 30, 944–955.
Ijiri A., Inagaki F., Kubo Y., Adhikari R. R., Hattori S., Hoshino
T., Imachi H., Kawagucci S., Morono Y., Ohtomo Y., Ono S.,
Sakai S., Takai K., Toki T., Wang D. T., Yoshinaga M. Y.,
Arnold G. L., Ashi J., Case D. H., Feseker T., Hinrichs K.-U.,
Ikegawa Y., Ikehara M., Kallmeyer J., Kumagai H., Lever M.
A., Morita S., Nakamura K.-I., Nakamura Y., Nishizawa M.,
Orphan V. J., Røy H., Schmidt F., Tani A., Tanikawa W.,
Terada T., Tomaru H., Tsuji T., Tsunogai U., Yamaguchi Y. T.
and Yoshida N. (2018) Deep-biosphere methane production
stimulated by geofluids in the Nankai accretionary complex.
Sci. Adv. 4.
Kao S.-J., Shiah F.-K., Wang C.-H. and Liu K.-K. (2006) Efficient
trapping of organic carbon in sediments on the continental
margin with high fluvial sediment input off southwestern
Taiwan. Cont. Shelf Res. 26, 2520–2537.
Kim J.-H., Park M.-H., Chun J.-H. and Lee J. Y. (2011) Molecular
and isotopic signatures in sediments and gas hydrate of the
central/southwestern Ulleung Basin: high alkalinity escape
fuelled by biogenically sourced methane. Geo-Mar. Lett. 31,
37–49.
Komada T., Burdige D. J., Li H.-L., Magen C., Chanton J. P. and
Cada A. K. (2016) Organic matter cycling across the sulfate-
methane transition zone of the Santa Barbara Basin, California
Borderland. Geochim. Cosmochim. Acta 176, 259–278.
Lin A. T., Liu C.-S., Lin C.-C., Schnurle P., Chen G.-Y., Liao W.-
Z., Teng L. S., Chuang H.-J. and Wu M.-S. (2008) Tectonic
features associated with the overriding of an accretionary wedge
on top of a rifted continental margin: an example from Taiwan.
Mar. Geol. 255, 186–203.
Lin C.-C., Tien-Shun Lin A., Liu C.-S., Chen G.-Y., Liao W.-Z.
and Schnurle P. (2009) Geological controls on BSR occurrences
in the incipient arc-continent collision zone off southwest
Taiwan. Mar. Pet. Geol. 26, 1118–1131.
Lin L.-H., Wu L.-W., Cheng T.-W., Tu W.-X., Lin J.-R., Yang T.
F., Chen P.-C., Wang Y. and Wang P.-L. (2014) Distributions
and assemblages of microbial communities along a sediment
core retrieved from a potential hydrate-bearing region offshore
southwestern Taiwan. J. Asian Earth Sci. 92, 276–292.
Liu C.-S., Schnu
¨rle P., Wang Y.-S., Chung S.-H., Chen S.-C. and
Hsiuan T.-H. (2006) Distribution and characters of gas hydrate
offshore of Southwestern Taiwan. Terrestr. Atmosp. Ocean. Sci.
17, 615–644.
Lin S. (2010) Authigenic carbonates and variations of sulfur species
in relations to the gas hydrate accumulations in the sediments
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 153

of Gas Hydrate Potential Area Offshore of Southwestern
Taiwan (3/4). Rep. Central Geol. Surv. 2010 99-26-C, pp. 87.
(in Chinese with English abstract).
Lloyd K. G., Alperin M. J. and Teske A. (2011) Environmental
evidence for net methane production and oxidation in putative
ANaerobic MEthanotrophic (ANME) archaea. Environ. Micro-
biol. 13, 2548–2564.
Londry K. L., Dawson K. G., Grover H. D., Summons R. E. and
Bradley A. S. (2008) Stable carbon isotope fractionation
between substrates and products of Methanosarcina barkeri.
Org Geochem. 39, 608–621.
Luo M., Dale A. W., Wallmann K., Hensen C., Gieskes J., Yan W.
and Chen D. (2015) Estimating the time of pockmark forma-
tion in the SW Xisha Uplift (South China Sea) using reaction-
transport modeling. Mar. Geol. 364, 21–31.
Malinverno A. and Pohlman J. W. (2011) Modeling sulfate
reduction in methane hydrate-bearing continental margin
sediments: does a sulfate-methane transition require anaerobic
oxidation of methane? Geochem. Geophys. Geosyst. 12, n/a–n/a.
Meister P., Liu B., Ferdelman T. G., Jørgensen B. B. and Khalili A.
(2013) Control of sulphate and methane distributions in marine
sediments by organic matter reactivity. Geochim. Cosmochim.
Acta 104, 183–193.
Meyers P. A. (1994) Preservation of elemental and isotopic source
identification of sedimentary organic matter. Chem. Geol. 114,
289–302.
Middelburg J. J. (1989) A simple rate model for organic matter
decomposition in marine sediments. Geochim. Cosmochim. Acta
53, 1577–1581.
Parkes R. J., Cragg B. A., Banning N., Brock F., Webster G., Fry
J. C., Hornibrook E., Pancost R. D., Kelly S., Knab N.,
Jørgensen B. B., Rinna J. and Weightman A. J. (2007)
Biogeochemistry and biodiversity of methane cycling in sub-
surface marine sediments (Skagerrak, Denmark). Environ.
Microbiol. 9, 1146–1161.
Paull C. K., Lorenson T. D., Borowski W. S., Ussler, III, W., Olsen
K. and Rodriguez N. M. (2000) Isotopic composition of CH
4
,
CO
2
species, and sedimentary organic matter within samples
from the Blake Ridge: Gas source implications. Proc. ODP Sci.
Res. 164, 229–236.
Penger J., Conrad R. and Blaser M. (2012) Stable carbon isotope
fractionation by methylotrophic methanogenic archaea. Appl.
Environ. Microbiol. 78, 7596–7602.
Pohlman J. W., Ruppel C., Hutchinson D. R., Downer R. and
Coffin R. B. (2008) Assessing sulfate reduction and methane
cycling in a high salinity pore water system in the northern Gulf
of Mexico. Mar. Pet. Geol. 25, 942–951.
Rabus R., Hansen T. A. and Widdel F. (2006) Dissimilatory
sulfate- and sulfur-reducing prokaryotes. In The Prokaryotes,
vol. 2 (eds. M. Dworkin, S. Falkow, E. Rosenberg, K.-. H.
Schleifer and E. Stackebrandt), pp. 659–768. Ecophysiology
and Biochemistry. Springer, New York, NY.
Raiswell R. and Canfield D. E. (1998) Sources of iron for pyrite
formation in marine sediments. Am. J. Sci. 298, 219–245.
Reeburgh W. S. (2007) Oceanic methane biogeochemistry. Chem.
Rev. 107, 486–513.
Regnier P., Dale A. W., Arndt S., LaRowe D. E., Mogollo
´n J. and
Van Cappellen P. (2011) Quantitative analysis of anaerobic
oxidation of methane (AOM) in marine sediments: a modeling
perspective. Earth Sci. Rev. 106, 105–130.
Riedinger N., Kasten S., Gro
¨ger J., Franke C. and Pfeifer K. (2006)
Active and buried authigenic barite fronts in sediments from the
Eastern Cape Basin. Earth Planet. Sci. Lett. 241, 876–887.
Schrum H. N., Spivack A. J., Kastner M. and D’Hondt S. (2009)
Sulfate-reducing ammonium oxidation: a thermodynamically
feasible metabolic pathway in subseafloor sediment. Geology
37, 939–942.
Sivan O., Schrag D. P. and Murray R. W. (2007) Rates of
methanogenesis and methanotrophy in deep-sea sediments.
Geobiology 5(2), 141–151.
Sun C.-H., Chang S.-C., Kuo C.-L., Wu J.-C., Shao P.-H. and
Oung J.-N. (2010) Origins of Taiwan’s mud volcanoes: evidence
from geochemistry. J. Asian Earth Sci. 37, 105–116.
Teichert B. M. A., Gussone N., Eisenhauer A. and Bohrmann G.
(2005) Clathrites: archives of near-seafloor pore-fluid evolution
(d44/40Ca, d13C, d18O) in gas hydrate environments. Geology
33(3), 213–216.
Timmers P. H. A., Welte C. U., Koehorst J. J., Plugge C. M., Jetten
M. S. M. and Stams A. J. M. (2017) Reverse methanogenesis
and respiration in methanotrophic archaea. Archaea 2017, 22.
Torres M. E., Brumsack H. J., Bohrmann G. and Emeis K. C.
(1996) Barite fronts in continental margin sediments: a new
look at barium remobilization in the zone of sulfate reduction
and formation of heavy barites in diagenetic fronts. Chem. Geol.
127, 125–139.
Treude T., Krause S., Maltby J., Dale A. W., Coffin R. and
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., Orphan V., Knittel K., Gieseke A., House C. H. and
Boetius A. (2007) Consumption of methane and CO
2
by
methanotrophic microbial mats from gas seeps of the Anoxic
Black Sea. Appl. Environ. Microbiol. 73, 2271–2283.
Treude T., Kru
¨ger M., Boetius A. and Jørgensen B. B. (2005)
Environmental control on anaerobic oxidation of methane in
the gassy sediments of Eckernfo
¨rde Bay (German Baltic).
Limnol. Oceanogr. 50, 1771–1786.
Ussler Iii W. and Paull C. K. (2008) Rates of anaerobic oxidation
of methane and authigenic carbonate mineralization in
methane-rich deep-sea sediments inferred from models and
geochemical profiles. Earth Planet. Sci. Lett. 266, 271–287.
Wallace P. J., Dickens G. R., Paull C. K. and Ussler, III, W. (2000)
Effects of core retrieval and degassing on the carbon isotope
composition of methane in gas hydrate-and free gas-bearing
sediments from the Blake Ridge. In Gas Hydrate Sampling on
the Blake Ridge and Carolina Rise. Proceedings of the Ocean
Drilling Program. Scientific Results (College Station, TX), vol.
164 (eds. C. K. Paull, R. Matsumoto, P. J. Wallace and W. P.
Dillon), pp. 67–78.
Wallmann K., Aloisi G., HaeckelM.,ObzhirovA.,PavlovaG.and
Tishchenko P. (2006) Kinetics of organic matter degradation,
microbial methane generation, and gas hydrate formation in anoxic
marine sediments. Geochim. Cosmochim. Acta 70, 3905–3927.
Wehrmann L. M., Risgaard-Petersen N., Schrum H. N., Walsh E.
A., Huh Y., Ikehara M., Pierre C., D’Hondt S., Ferdelman T.
G., Ravelo A. C., Takahashi K. and Zarikian C. A. (2011)
Coupled organic and inorganic carbon cycling in the deep
subseafloor sediment of the northeastern Bering Sea Slope
(IODP Exp. 323). Chem. Geol. 284, 251–261.
Wellsbury P. and Parkes R. J. (2000) Deep biosphere: sources of
methane for oceanic hydrates. In Natural Gas Hydrate in
Oceanic and Permafrost Environments (ed. M. D. Max). Kluwer
Acad., Norwell, Mass, pp. 91–104.
Whiticar M. J. (1999) Carbon and hydrogen isotope systematics of
bacterial formation and oxidation of methane. Chem. Geol. 161,
291–314.
Yoshinaga M. Y., Holler T., Goldhammer T., Wegener G.,
Pohlman J. W., Brunner B., Kuypers M. M. M., Hinrichs K.-
U. and Elvert M. (2014) Carbon isotope equilibration during
154 P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155

sulphate-limited anaerobic oxidation of methane. Nature
Geosci. 7, 190–194.
Yoshioka H., Maruyama A., Nakamura T., Higashi Y., Fuse H.,
Sakata S. and Bartlett D. H. (2010) Activities and distribution
of methanogenic and methane-oxidizing microbes in marine
sediments from the Cascadia Margin. Geobiology 8, 223–233.
Zeebe R. E. (2007) Modeling CO
2
chemistry, d
13
C, and oxidation
of organic carbon and methane in sediment porewater: impli-
cations for paleo-proxies in benthic foraminifera. Geochim.
Cosmochim. Acta 71, 3238–3256.
Associate editor: John W. Moreau
P.-C. Chuang et al. / Geochimica et Cosmochimica Acta 246 (2019) 138–155 155