Sulfate reduction and anaerobic oxidation of methane in sediments of the South-Western Barents Sea

Anaerobic oxidation of methane (AOM) in marine sediments strongly limits the amount of gas reaching the water column and the atmosphere but its efficiency in counteracting future methane emissions at continental margins remains unclear. Small shifts in methane fluxes due to gas hydrate and submarine permafrost destabilization or enhanced methanogenesis in warming Arctic continental shelves may cause the redox boundary in which AOM occurs, known as Sulfate-Methane Transition 10 Zone (SMTZ), to move closer to seafloor, with potential gas release to bottom waters. Here, we investigated the geochemical composition of pore water (SO4 and DIC concentration, CDIC) and gas (CH4, CCH4) in eight gravity cores collected from Ingøydjupet trough, South-Western Barents Sea. Our results show a remarkable variability in SMTZ depth, ranging from 3.5 m to 29.2 m, and that all methane is efficiently consumed by AOM within the sediment. From linear fitting of the sulfate concentration profiles, we calculated diffusive sulfate fluxes ranging from 1.5 nmol cm d to 12.0 nmol cm d. AOM rates obtained for two 15 cores using mixing models are 6.5 nmol cm d and 6.7 nmol cm d and account for only 64% and 56% of total sulfate reduction at the SMTZ (SRRtot), respectively. The remaining 36% and 44% SRRtot correspond to organoclastic sulfate reduction with rates of 3.7 nmol cm d and 5.3 nmol cm d. The shallowest SMTZs (< 5 m) and largest SRRtot rates are associated with a shallow subsurface accumulation of gas visible in seismic data, highlighting how small changes in sulfate reduction rates linked to subsurface methane gradients resulted in vertical shifts in SMTZ position of > 20 m. This study provides new insights into the 20 dynamic and biogeochemistry of the SMTZ in marine sediments of continental margins and may help predict the response of the microbial methane filter to future increase in methane fluxes due to ocean warming.


Introduction
Despite covering the ~70% of planet's surface, oceans only contribute a small fraction of present-day methane emissions into the atmosphere (3-6%) because most of the gas generated in sedimentary basins becomes oxidized during its upward migration to the 25 seafloor before reaching the water column and atmosphere Weber et al., 2019). Marine sediments host an efficient biological filter for methane, constituted by a microbial consortia that uses sulfate to perform the anaerobic oxidation of methane (AOM) (Boetius et al., 2000). AOM occurs within the Sulfate-Methane Transition Zone (SMTZ), a redox boundary marking the transition between sulfate-rich pore waters and the underlying methanogenic zone. The SMTZ is highly dynamic and rapidly responds (10 2 -10 3 yr) to perturbations in pore water composition due to gas hydrate and permafrost decomposition (Hong 30 et al., 2016;Overduin et al., 2015;Sultan et al., 2016) or emplacement of mass-transport deposits . At a global scale, sulfate-driven AOM consumes a total of ~45-61 Tg CH4 yr -1 , which approximately balances the cumulative CH4 production by methanogenesis in the sediments and corresponds to mineralization of ~ 3-4% of the total organic carbon sinking to the seafloor (Egger et al., 2018). Methane fluxes control the depth of the SMTZ (Borowski et al., 1996(Borowski et al., , 1999Jorgensen and Kasten, 2006;Kasten et al., 2003), and under steady-state conditions and in non-hydrate bearing sediments they accumulation and preservation of organic matter in coastal sediments (Diaz and Rosenberg, 2008;Middelburg and Levin, 2009;Wallenius et al., 2021) due to the expansion of oxygen minimum zones (Li et al., 2020;Rabalais et al., 2010). Changes in organic matter burial can increase methanogenic production (Egger et al., 2016(Egger et al., , 2018, affecting SMTZ stability. Global warming is also responsible for the observed Atlantification of the Arctic Ocean, increasing the influx of warm water masses in eastern Barents 40 Sea (Lind et al., 2018;Smedsrud et al., 2013). The warming trend involves the entire water column (Lind et al., 2018) and a rise of near-bottom water temperatures has been recently reported (Skagseth et al., 2020). As the activity of methanogens in marine sediments is strongly dependent on in-situ temperature (Maltby et al., 2018), a warming Arctic may result in enhanced methane production within the shallow sedimentary column and eventually lead to a shoaling of the SMTZ in vast areas of the continental shelf. Ocean warming-related destabilization of gas hydrates and marine permafrost would have the same overall effect on SMTZ 45 dynamic (Borges et al., 2016;James et al., 2016;Stranne et al., 2019). All these processes would lead to a shallow and "compressed" methane oxidation barrier, increasing the net methane export to the ocean, amplifying ocean acidification (Biastoch et al., 2011).
The SW Barents Sea hosts several areas marked by natural seafloor methane emissions and/or associated with subsurface gas accumulations related to gas hydrates or hydrocarbon leakage from deeper Mesozoic reservoirs (Chand et al., 2009; 50 Vadakkepuliyambatta et al., 2015). This setting offers a unique opportunity to investigate SMTZ dynamic and biogeochemical processes in key areas characterized by spatially variable subsurface methane fluxes, providing insights into future scenarios of climate-driven increasing methane fluxes in coastal environments and Arctic continental shelves affected by ocean warming.
In this paper we investigate the geochemistry of pore water (SO4 2and DIC concentration,  13 CDIC) and gas in gravity cores collected from Ingøydjupet trough, SW Barents Sea. This area was selected as it hosts a subsurface seismic anomaly related 55 to gas accumulation and can be used to explore spatial variations in SMTZ depth, sulfate reduction rates and AOM. We used a two-component mixing model based on dissolved inorganic carbon systematics (DIC,  13 CDIC) and constrained by gas isotope analysis ( 13 CCH4) to determine the contribution of organic matter mineralization (OSR) to the overall sulfate reduction in the SMTZ and obtain accurate AOM rate values. This study integrates global datasets of diffusive sulfate and methane fluxes in continental shelves improving the knowledge on the dynamic and biogeochemistry of the SMTZ in marine sediments.

2 Study area
The Barents Sea is an epicontinental sea marked by an average depth of 230 m and extending from the coasts of northern Norway and Russia to the south, to 80° N and bordered to the east by the coast of Novaya Zemlya and to the west by the Norwegian Sea.
In the present study, we focus on sediment cores collected from Ingøydjupet, an NW-SE-stretching ice-carved trough located in the southwestern sector of the Barents Sea (Fig. 1a). Ingøydjupet area has water depth less than 500 m and seafloor morphology 65 includes numerous iceberg plough-marks, mega-scale lineations as well as glacial moraines, indicating a dynamic history of ice sheet advance and retreat since the Pliocene (Winsborrow et al., 2010). This area is associated with widespread subsurface fluidflow systems and major Mesozoic hydrocarbon reservoirs (Chand et al., 2009;Vadakkepuliyambatta et al., 2013). Core locations are spread over an area of ~ 3 km 2 in the NE part of Ingøydjupet (Fig. 1b), about 10 km from a gas field (Caurus field; well no.7122/2-1 in Fig. 1b)
Pore water samples were taken every 10-20 cm and extracted by pressure filtration through 0.2 m cellulose acetate filters using 10 cm long rhizon samplers. Each pore water sample was split into 2 aliquots: 1) subsamples for DIC analyses (1 ml) were transferred to 1.5 ml micro tubes with screw caps and we added 10 l of HgCl2 to stop microbial activity. 2) Subsamples for sulfate analysis (> 1 ml) were transferred to 5 ml Eppendorf tubes and stored at -20˚C.

90
We collected bulk sediment samples every 10-20 cm in cores 354-GC, 355-GC and 356-GC and from the bottom of each section for the other cores. A 5 mL syringe without the luer tip was used to collect 5 mL of sediments, that were transferred to a 20 mL serum vial containing two glass beads and to which we added 5 mL of 1M NaOH to stop microbial activity. The vial was immediately closed with a septum and an aluminum crimp seal and stored at 4˚C for onshore analyses. https://doi.org/10.5194/bg-2021-58 Preprint. Discussion started: 9 March 2021 c Author(s) 2021. CC BY 4.0 License.

95
Pore water sulfate (SO4 2-) concentration was measured at the University of Bergen, Department of Earth Science (Bergen, Norway), via inductively coupled plasma -optical emission spectrometry using a Thermo Scientific iCAP7000. Repeated measurements of ERM CA016a certified material were used to check the accuracy and precision during the analyses: measured values agree within the uncertainty of the certified value.

100
Sample preparation for the measurement of concentration and carbon isotopic composition ( 13 C) of dissolved inorganic carbon followed the protocol for marine pore waters described in Spötl (2005) and briefly reported here: 100 l of 85% phosphoric acid (H3PO4) were added to 4.5mL vials and flushed with helium. 400 l or 200 l of sample was injected into the vial using a syringe and left to equilibrate for 24h at 25°C prior to analysis. Measurements of the isotopic composition of released CO2 were done using a Thermo Scientific MAT253 IRMS coupled to a Gasbench II at the Stable Isotope Laboratory at UiT -The Arctic University of 105 Norway in Tromsø. Data are reported in‰ notation relative to the Vienna Pee Dee Belemnite (VPDB) standard and analytical error was better than 0.1‰ (1SD). DIC concentration was calculated according to peak heights from the IRSM results and by comparison with signal strength of a NaHCO3(aq) calibration curve. Headspace gas composition was measured on a ThermoScientific GC Trace 1310 gas chromatograph-mass spectrometer equipped with a Restek Rt-Alumina BOND/Na2SO4 column for hydrocarbon gas separation. The GC system was calibrated using 2 external standards at 100 ppm and one standard at 110 1000 ppm. Analyses were conducted at the the Stable Isotope Laboratory at UiT -The Arctic University of Norway in Tromsø.
The isotopic composition of methane was measured in one sample collected at 305 cm in core 358-GC. Methane in the headspace gas was separated using a Restek MXT Molsieve 5Å gas chromatography column, ionized at 950°C and measured using a Nu Horizon IRMS. Calibration is done with in-house standard calibrated with the international USGS standards NGS-2 and USGS NGS-3. Precision was better than 0.5‰.

Modelling
The pore water sulfate concentration profiles display a linear shape in all the examined cores, which indicates that sulfate reduction at these sites is at a steady state and in a diffusion-controlled system Kasten et al., 2003;Schulz, 2006). Diffusive sulfate fluxes (J, nmol cm -2 d -1 ) for each core were calculated from the sulfate concentration gradients using Fick's First Law (Eq. 1): where φ is the sediment porosity, Ds the whole-sediment diffusion coefficient and dC/dz is the concentration gradient (Boudreau, 1996). Porosity (φ) was determined onshore from the water loss per volume of sediment upon drying (Jørgensen and Parkes, 2010) in sediment samples collected from core 354-GC, and gave an average value φ = 0.64 ± 0.06 (SD; n = 3). Ds was obtained from the free-solution molecular diffusion coefficients of sulfate, Do, corrected for tortuosity using the empirical formula of Iversen and 125 Jørgensen (1993) (Eq. 2): where Do = 0.48 cm 2 d -1 at 4 °C (Schulz, 2006). The concentration gradient dC/dz was obtained by linear least squares fitting of the sulfate concentration profile Graves et al., 2017;Mazumdar et al., 2007;Schulz, 2006). Pore water profiles were extrapolated to the depth of the SMTZ by linear regression (Graves et al., 2017). The diffusive sulfate flux into the SMTZ feeds both oxidation of buried organic matter (OSR) and AOM, and also corresponds to total sulfate reduction rate (SRRtot) . In order to quantify AOM and OSR rates in the SMTZ, we calculated their relative contributions to the pore water DIC pool using a mixing model, based on mass balance equations Eq.

135
where  13 Cadded represents the isotopic composition of bulk DIC released into pore water by biogeochemical reactions (i.e. OSR, AOM) and added to the initial background pore water DIC (buried seawater DIC) Hong et al., 2013;Hu et al., 2017Hu et al., , 2010Ussler and Paull, 2008;Wu et al., 2016). In these equations, fOSR and fAOM are the fractions of DIC inputs (%) from OSR and AOM, respectively, and  13 CSOC (SOC = sedimentary organic carbon) and  13 CAOM indicate their isotopic composition.
 13 Cadded was obtained by linear regression of DIC concentration against the product of DIC concentration by its isotopic 140 composition (DIC ×  13 CDIC) (Hu et al., 2010;Wu et al., 2016); the regression slope corresponds to  13 Cadded. In our model we assumed an average  13 CSOC = -21.8‰ (Knies and Martinez, 2009) and  13 CAOM = -63.1‰, the latter representing the composition of DIC produced by anaerobic oxidation of methane in the study area ( 13 CCH4 = -54.1‰, our study), and assuming ~9‰ fractionation during AOM (Alperin et al., 1988;Beulig et al., 2019;Whiticar, 1999). The uncertainty on fOSR and fAOM was estimated by solving the equations using  13 CSOC of pure marine and pure terrestrial organic matter end-members (Knies and 145 Martinez, 2009) and is less than ± 0.05. We applied this model to cores 358-GC and 359-GC as the collected DIC samples reach the depth of the SMTZ (or close to it) and allow accurate modelling.

Sulfate data
Pore water sulfate concentration profiles in all the examined cores have a linear shape and allowed an accurate calculation 150 of sulfate gradients by linear regression. Diffusive fluxes calculated using Fick's First Law range from 1.5 nmol cm -2 d -1 (core 354-GC) to 12.0 nmol cm -2 d -1 (core 358-GC) and correspond to the total sulfate reduction rates in the SMTZ (SRRtot) (Fig. 2, Tab. 1).
The depth of the SMTZ, determined by linear extrapolation of the pore water profile to the depth of complete sulfate depletion, is consistent with sulfate fluxes (the greater the flux, the shallower the SMTZ) and ranges from 3.5 m to 29.2 m (Tab. 1). In all cores except for 358-GC, the depth of the SMTZ was deeper than the length of the core (Fig. 2)

Methane data
Methane concentration in the sediment cores was low (< 50 µM) except for 358-GC (Fig. 2), where the concentration value measured below the SMTZ reaches 5.4 mM. These observations validate our assumptions, i.e. the examined data reflect a diffusive-controlled system with complete methane consumption in the SMTZ by AOM. Hydrocarbon gas diffusing into the SMTZ is dominated by methane, representing 99.8% of the total hydrocarbon gas fraction (core 358-GC). The carbon isotopic composition

Organoclastic sulfate reduction and anaerobic oxidation of methane at the SMTZ
The pore water concentration profiles provided important details about sulfate reduction in the sediment. Specifically, the linear shape of the sulfate profiles found in all the cores is indicative of steady-state conditions and diffusive sulfate transport into 190 the SMTZ Schulz, 2006) (Fig. 2). We quantified the sulfate flux from linear regression of the concentration extrapolated SMTZ depths is negligible (Fig. 2), indicating a complete methane consumption within the SMTZ by AOM microbial consortia. Considering a 1 : 1 ratio between sulfate and methane consumption by AOM in the SMTZ (Boetius et al., 2000), the diffusive sulfate flux into the SMTZ would correspond to the upward diffusive methane flux, and SRRtot = AOM rate (Graves et 200 al., 2017;Hoehler et al., 2000). However, recent studies Jørgensen et al., 2019), demonstrated that OSR may also contribute to the overall sulfate reduction within the SMTZ, resulting in flux discrepancies. Therefore, SRRtot can only provide the upper threshold value for AOM rates (Graves et al., 2017). In agreement with those studies, our results indicate that AOM contributes to 56-64% of total sulfate reduction at the SMTZ (Tab. 1). Conversely, OSR accounts for the 36-44% of total sulfate reduction. These proportions are consistent with published range of values obtained in other marine settings worldwide marked by 205 variable depth of the SMTZ ( < 20 m), e.g. Baltic Sea (OSR = 14-59% SRRtot) , Ulleung Basin (OSR = 10-30% SRRtot) (Hong et al., 2013), Santa Barbara Basin (OSR = 35-45% SRRtot) (Komada et al., 2016), Gulf of Mexico (OSR up to 50% SRRtot) (Hu et al., 2010). Our observations seem confirming that in different diffusive systems at steady-state, OSR contribution to total sulfate reduction is approximatively the same both in shallow (< 5 m) SMTZs (our study) and deeper SMTZs (previous studies) as discussed by Egger et al. (2018).

210
The modeled reaction rates represent the net rates, resulting from all primary (e.g. AOM, OSR) and secondary reactions (e.g.re-oxidation processes), and gross rates are only determined by radiotracers techniques Jørgensen et al., 2019;Kasten et al., 2003). A recent study by Beulig et al. (2019) on cores from the Baltic Sea, showed that up to 60% of organic matter oxidation taking place within the SMTZ first produced methane, the latter becoming subsequently oxidized by AOM, stimulating a cryptic methane cycle . This process results in an additional release of 13 C-depleted DIC into 215 pore water that is not linked to the diffusive methane flux from below the SMTZ, but from in-situ methanogenesis. In order to provide a rough estimation of the potential effect of this process on the overall results of our model, we recalculated OSR and AOM contributions to SRRtot, by adding a third component in the mixing model, represented by AOM related to in-situ microbial methane production and hereafter named AOM2. AOM related to the diffusive methane flux is hereafter named AOM1. As boundary conditions for the new model, we constrained fAOM2 to be ≤ fOSR, and  13 CAOM2 = -80‰ (as an example, we assumed a 220 hypothetical microbial  13 CCH4 ~ -70‰ corrected for AOM-induced fractionation). For the two  13 Cadded values obtained in our study,  13 Cadded = -45.0‰ (core 348-GC) and  13 Cadded = -48.4‰ (core 359-GC), we obtained fAOM2 values of 0.25 and 0.29, respectively, corresponding to 25% and 29% SRRtot. fAOM1 values are 0.21 (core 358-GC) and 0.24 (core 359-GC), corresponding to 21% and 24% SRRtot. We note that fAOM (fAOM1 + fAOM2) values, and consequently also fOSR values, are very close to the values obtained from the two-component model, demonstrating the reliability of the results obtained using that approach (Tab. 1).

Factors influencing SMTZ dynamic in the study area
The SMTZ represents an important oxidation barrier for methane in marine sediments and limits methane emissions into the hydro-and atmosphere (Weber et al., 2019). Its depth below the seafloor is a valuable qualitative proxy for methane fluxes (e.g. Borowski et al., 1999Borowski et al., , 1996. At a regional-scale, methane fluxes in non-hydrate-bearing sediments are mainly controlled by rates of methanogenesis, which in turn is influenced by tectonic and glacial processes (loading and erosion) affecting sediment 230 deposition and burial history of the area and kerogen evolution (Ferry and Lessner, 2008). The sediment cores in this study were collected within a 3 km 2 area so we rule out the possibility that the different methane gradients are due to differential in-situ methanogenesis. 3D seismic data shows a high-amplitude anomaly in the shallow subsurface of the study area (Fig. 4), providing strong evidence for gas accumulation in the sediments. In particular, the spatial distribution of this anomaly perfectly matches with the location of cores 358-GC and 359-GC (Fig. 4), which are the cores associated with the shallowest SMTZ (< 5 m) in our study.
Therefore, we confidently interpret the variability in modeled sulfate reduction rates and depths of the SMTZ (Tab. 1) to reflect the differential distribution of gas accumulation in the subsurface and associated upward methane fluxes (Fig. 5). The carbon isotopic composition of the gas sample collected at 305 cm from core 358-GC ( 13 C = -54.3‰) is similar to the average isotopic composition of methane in Mesozoic successions from a nearby well (Well no. 7122-2-1, Caurus field, 245 factpage.npd.no). We could not measure the hydrogen isotopic composition for our gas sample, due to limited sample size, and the determination of the gas source was out of the aim of this study. However, it is likely that the gas accumulated in the subsurface, and visible in seismic section, has a similar mixed microbial-thermogenic origin as the gas found in well no. 7122-2-1 located only ~ 10 km away. In conclusions, the data obtained in this study clearly highlight a link between subsurface methane gradients, sulfate reduction rates and depth of the SMTZs (Fig. 5)

260
In this study we investigated the geochemical composition of pore water (SO4 2and DIC concentration,  13 CDIC) and gas (CH4,  13 CCH4) samples collected from 8 gravity cores from the Ingøydjupet trough, SW Barents Sea. Sulfate concentration profiles in all cores show a linear shape indicating steady-state conditions and a diffusion-controlled solute transport system. The SMTZ shows a wide range of depths below the seafloor, from 3.5 m to 29.2 m, and all methane is efficiently consumed by AOM within the sediment. The diffusive sulfate fluxes into the SMTZ feeding organoclastic sulfate reduction and anaerobic oxidation of 265 methane correspond to 1.5 nmol cm -2 d -1 and 12.0 nmol cm -2 d -1 . Mixing modeling on two cores indicated that total sulfate reduction at the SMTZ is related for 36-44% to organoclastic sulfate reduction and for 56-64% to anaerobic oxidation of methane, corresponding to rates of 3.7-5.3 nmol cm -2 d -1 and 6.5-6.7 nmol cm -2 d -1 , respectively. In the study area, the shallowest SMTZs are associated with a spatially-localized subsurface gas accumulation, clearly highlighting how small changes in sulfate reduction rates in the order of ~ 10 nmol cm -2 d -1 linked to subsurface methane gradients result in large vertical shifts in SMTZ position (> 270 20 m). Due to the proximity (~ 10 km) of this area to a nearby gas field, and based on the isotopic composition of methane in the examined sample, we suggest that the gas diffusing toward the seafloor at the Ingøydjupet site has a mixed microbial-thermogenic source and migrated from Mesozoic reservoirs. The results of this study help better understand SMTZ dynamic in marine sediments https://doi.org/10.5194/bg-2021-58 Preprint. Discussion started: 9 March 2021 c Author(s) 2021. CC BY 4.0 License.
of continental margins and are of critical importance to predict the response of the microbial methane filter to future increase in methane fluxes due to climate change.

Data availability
The datasets used for this study are openly available in Dataverse.NO at https://doi.org/10.18710/LI7QNQ.

Author contributions
C.A. and G.P. conceived the study. C.A. and K.A.W. conducted the offshore sampling operations. C.A.
wrote the initial manuscript and C.A. and K.A.W produced figures and tables. C.A., S.B. G.P. and K.A.W. 280 all contributed to the discussion and improvement of the paper.