Mineralization pathways of organic matter deposited in a river–lake transition of the Rhone River Delta, Lake Geneva

During the éLEMO endeavour (a research project in which the Russian MIR submersibles were used for studying Lake Geneva) four sediment cores were retrieved on a transect from the delta of the Rhone River towards the profundal part of the lake. The degradation pathways of organic material (OM) were investigated considering different electron acceptors. Essentially, OM at the delta sites had a higher fraction of terrestrial material than the lake sites indicated by higher C/N ratios, and higher long-chain nalkane and alcohol concentrations. The concentrations of chlorins were higher at the distant sites indicating more easily degradable OM in the sediments. However, the chlorin index that was used to determine the degradation state of the OM material indicated that pigment derived OM of deltaic sediments was less degraded than that of the profundal sediments. The fluxes of reduced species from the sediments decreased from the delta to the profundal for CH4 (from 2.3 to 0.5 mmol m 2 d ) and NH4 + (from 0.31 to 0.13 mmol m 2 d ). Fluxes of Fe(II) and Mn(II), however, increased although they were generally very low (between 9 10 5 and 7.6 10 3 mmol m 2 d ). Oxygen concentration profiles in the pore waters revealed lower fluxes close to the river inflow with 4.3 and 4.1 mmol m 2 d 1 compared to two times higher fluxes at the profundal sites (8.8 and 8.2 mmol m 2 d ). The rates for totally mineralized OM (Rtotal) at the shallower sites (4.7 mmol C m 2 d ) were only half of those of the deeper sites (9.7 mmol C m 2 d ). Accordingly, not only the rates but also the mineralization pathways differed between the shallow and profundal sites. Whereas only 0–6% of the OM was mineralized aerobically at the shallow sites (since almost all O2 was used to oxidize the large flux of CH4 from below) the situation was reversed at the deeper sites and the fraction of aerobically degraded OM was 72–78%. We found a better efficiency in CH4 production per carbon equivalent deposited at the deeper sites as a result of the higher degradability of the mainly autochthonous OM in spite of the lower deposition rate and the higher degradation state of the OM compared to the delta sites.


Introduction
Whereas countless studies on the transformation of organic material (OM) during degradation have been performed in the marine environment, [1][2][3] studies in lakes on the same topic are rather rare. 4,5The understanding of preservation of OM that has been formed during photosynthesis and subsequently has been moved to and buried into the sediments is crucial for at least two reasons: rst, we need to understand how OM is preserved or degraded on its way to the formation of oil or gas depending on the environment and redox conditions during burial/ diagenesis.Second, and this might be a more important question at this moment of time, how much of the aquatic OM primarily built from atmospheric carbon dioxide (CO 2 ) is nally buried in the sediments.The increase of CO 2 and methane (CH 4 ) in the atmosphere since the last ice age and, even more dramatically, over the last 150 years [6][7][8] is naturally balanced only (i.e., not considering man made possibilities like carbon storage in the ocean or deeper terrestrial environments) by the uptake of CO 2 by oceans/lakes and/or terrestrial biomass.The xation of CO 2 in the biomass and the subsequent burial in soils and sediments diminishes CO 2 concentrations in the atmosphere and hence is crucial to understand the conditions that eventually decide about either mineralization or burial of OM.The burial efficiency of OM in lakes was highlighted recently. 4These authors found high organic carbon (OC) burial efficiency (ratio between OC buried to OC deposited) in lakes in general (0.48) and especially in lakes inuenced by a high allochthonous input.A previous study estimated the OC buried in lake sediments to be 50% of what is stored in the ocean 9 which is a huge number when comparing the ocean area (71%) to the integrated global lake area (2-3%). 10In addition to the fact that sediments receiving a high input of suspended particulate matter have high OC burial efficiencies a closely linked factor was the oxygen (O 2 ) exposure time. 4However, it is not only O 2 that is used as an electron acceptor when OC is degraded but also other oxidants like nitrate, iron, manganese, and sulfate can play a substantial role. 11In a study of the meromictic Lake Zug, sediments in which the input of allochthonous material can be neglected the distinction between aerobic and anaerobic degradation of OM showed that in the shallower part of the lake with oxic bottom water around 50% of the OM was degraded with O 2 . 12Hence, it is important to consider other electron acceptors besides O 2 that might play a role in OM degradation.The study presented here focuses on the diagenetic degradation pathways and the factors controlling the burial efficiency of OM in Lake Geneva sediments.Lipids and pigments were studied in the sediments as indicators of the origin of sediment OM and electron acceptors in the sediment pore waters when moving from the delta of the main tributary, the Rhone River (higher river inow, deltaic sites) to the deeper part of the lake (less river inuence, profundal sites).

Sampling site
Lake Geneva is the largest western European lake with a surface area of 582 km 2 , a volume of 89 km 3 and a maximum depth of 309 m.With a catchment area of 5220 km 2 , the Rhone River makes up 68% of the input to the lake. 13The annual discharge ranged between 127 and 227 m 3 s À1 for the period 1935-2008. 14he annual sediment load transported by the Rhone River into Lake Geneva varies between 2 and 9 Â 10 6 tonnes per year 15 providing large amounts of allochthonous OM supply mainly to the large delta.

Sampling
Sediment cores and peeper samples 16 were collected at 4 locations from the northeastern part of Lake Geneva in front of the inow of the Rhone River with increasing distance from the shore using the MIR submersibles 17 and piston cores from a boat (Fig. 1 ESI; † Table 1).For each site, 3 sediment cores (inner diameter ¼ 59 mm) were collected: one for CH 4 analysis, one for porosity and bulk parameters and one for biomarker analysis.Sediment cores were sliced in 1 cm sections, frozen and freeze dried.Pore water equilibrators (peepers; 18,19 ) were used and consisted of a 50 Â 15 cm solid, plexiglass frame into which cells (1.5 Â 3 cm) were milled at a spacing of either 0.5 or 1.5 cm.The cells were lled with distilled, deionized water, and a 0.2 mm lter (HT 200, Tuffryn polysulfone membrane; Gelman Inc.) was laid over the surface of the frame and held in place by an outer frame of plexiglass. 16To remove O 2 from the water in the cells, the peeper was immersed in a tank of deionized water that was then purged for 24 hours with nitrogen. 20When the peeper is placed in the sediments, solutes from pore water in contact with the lter diffuse into the cells such that, aer equilibrium is reached, concentrations inside each cell equal those at the corresponding depth in the pore water.A one week period was allowed for equilibration which allows e.g. over 90% of sulfate (which has a low diffusion coefficient) to be equilibrated (aer 5.6 days; a more than 99% equilibration would need about 130 days; 19 ).Peepers were mounted on a steel tripod that was positioned on the sediment surface.Exposure of the peeper to lake water following retrieval from the sediments was minimized and lasted no more than 3-5 minutes.Sampling of the cells was done immediately upon retrieval within less than an hour (preferable between 15 and 30 min) aer recovery.Water from within the cells was sampled with pipettes and immediately transferred into vials containing appropriate xative agents (see below for different species).

Sedimentation rates
To determine sedimentation rates (SRs) in the profundal area, core #5 was dated by gamma spectrometry using 137 Cs on freezedried and ground sediment with a HPGe detector (Canberra GCW-3523) using an energy of 661.7 keV.SRs and accumulation rates (ARs) were calculated based on the 137 Cs peaks from the Chernobyl accident and nuclear bomb tests assuming a constant rate of supply (CRS) according to the model of Niessen and Appleby. 21,22The SR of core #4, was estimated by correlation using the TOC prole with the nearby core #5 (see Fig. 2A, ESI †).Since cores #1 and #2 could not be dated through neither 210 Pb nor 137 Cs due to low activities we correlated the cores via the TOC prole to the nearby core LP (ESI Fig. 1, † N 46 24.190 0 ; E 6 50.493) from 103 m water depth which was 14 C and 137 Cs dated (see Fig. 2B, ESI, † c.f. 23 ).

Bulk parameters
Total carbon (TC) and total nitrogen (TN) were measured in freeze-dried samples with a Thermo Quest CE Instrument NC 2500.Total organic carbon (TOC) was measured on decalcied samples. 24In brief, sediment samples were acid treated with 10% HCl until no further CO 2 evolved and dried at 40 C before weighing for elemental analysis.Total inorganic carbon (TIC) was calculated as the difference between TC and TOC.The precision for TC and TOC was AE0.1 wt% and AE0.2 wt% for TN.d 13 C and d 15 N of the OM were obtained with a GV Instruments IsoPrime isotope ratio mass spectrometer (IRMS).The precision of d 13 C in & vs. Vienna Peedee Belemnite (VPDB) and d 15 N in & vs. air was lower than AE0.3&.The chlorin index (CI) and total chlorin concentration were determined with an analytical precision of AE5% according to Schubert et al. 25

Accumulation rates
Whereas the SR only includes linear sedimentation in a specic core, bulk accumulation rates (AR bulk ) also take compaction into account and hence sediment accumulation per time and area could be estimated.We used the following formulae: 21 AR bulk ¼ SR Â dry density Â (1 À porosity/100) (1) where AR bulk and AR TOC (g cm À2 per year) are the accumulation rates of the bulk sediment and total organic carbon, respectively.Dry density (g cm À3 ), SR (cm per year), porosity (%), and TOC (total organic carbon in %).

Lipid biomarker analysis
The sediment samples were extracted in methanol/dichloromethane using a microwave system (MLS) to obtain the total lipid extracts.Aliquots of the extracts were saponied aer addition of internal standards and separated into fatty acid and neutral (non-saponiable) lipid fractions.The neutral fractions were silylated (BSTFA, Sigma) and analyzed by gas chromatography and gas chromatography-mass spectrometry for the quantication and identication of biomarkers, respectively.For a detailed description of lipid extraction see ref. 26.

Methane concentrations
Sediment samples (2 cm 3 ) were taken from the sediment core by sub-sampling the sediment from the side of the plastic tube through pre-drilled holes that were covered by tape during sampling.The sediment was lled in a 20 mL glass vial with 4 mL of 1 M NaOH.Subsequently the glass was covered with a butyl stopper and sealed with an aluminum crimp.Dissolved CH 4 concentrations were measured in the headspace under controlled temperature conditions with an automated system (Joint Analytical Systems, Germany).In short, samples were introduced via an automated pumping system into a gas chromatograph (GC, Agilent) at constant temperature.The GC was equipped with a Carboxen 1010 column (30 m, Supelco) and a ame ionization detector.The oven temperature was 40 C. CH 4 standards were prepared by dilution of pure CH 4 (99.9%) and calibrated against commercial references of 100 ppm, 1000 ppm and 1% (Scott, Supelco).

pH, alkalinity, and DIC
pH was measured on the boat within minutes following the retrieval of the peepers using a portable pH meter (model 704, Metrohm) and a glass electrode (Minitrode, Hamilton) calibrated with 2 buffer reference solutions (Merck) of pH 6 and 8. Alkalinity was determined by end-point titration using a titration device (Titrino, Metrohm) and hydrochloric acid (HCl, 0.01 M).Dissolved inorganic carbon (DIC) concentrations were estimated using the measurements of pH and alkalinity, and equilibrium constants at 20 C. 27 2.9 Stable isotopic composition of carbon in DIC (d 13 C DIC ) An aliquot of pore water was transferred without air bubbles onsite as fast as possible into a 2 mL glass vial, sealed with a Teon cap using a crimper and kept at 4 C. Before analysis, 1.5 mL of the water sample was transferred into an exetainer® and the headspace was ushed with helium.Concentrated phosphoric acid was added to convert all DIC to CO 2 .100 mL of headspace was withdrawn from the vials with a GasTight Hamilton syringe and injected into a GC (Hewlet Packard 5890) interfaced to a ConFlo IV and a Delta V Plus IRMS (Thermo Fisher Scientic, Bremen, Germany).The GC was equipped with a GS-Carboplot column (Agilent, 30 m Â 0.32 mm Â 1.5 micron) to separate CO 2 from the interfering gases.

Oxygen concentrations
Oxygen at the sediment-water interface was measured within an hour aer the sampling.An oxygen micro-optode operated with a temperature compensated system (Microx TX3, PreSens) and a custom made step motor with a vertical resolution of 0.05 mm was used.The micro-optode was calibrated with oxygenfree and air saturated solutions.The oxygen-free calibration solution was prepared by dissolving 1 g of sodium sulte (Na 2 SO 3 ) in 100 mL of nanopure water and the air saturated calibration solution was prepared by purging with air for 20 minutes using a glass frit following the recommendations of the manufacturer.The micro-optode was positioned in the water overlying the sediments, and slowly brought close to the sediment-water interface until the oxygen values (% air saturation) decreased.From this point on, the oxygen values (% air saturation) were recorded with a vertical resolution of 0.05 to 0.1 mm with an equilibration time between measurements of 20 s.Oxygen was measured until values of 0% air saturation were obtained two times in a row.Up to nine oxygen proles per site were recorded.For Fe(II) and Mn(II) concentrations, pore water aliquots (2 mL) were acidied onsite with 45 mL ultrapure concentrated nitric acid (65%, VWR) within minutes aer sampling and then kept frozen until analysis using inductively coupled plasma optical emission spectroscopy (ICP-OES, Spectro Analytical Instruments).Samples were diluted 25 and 125 times (pore water samples from the center of the lake and from the delta region, respectively).

Data analysis and calculation of mineralization rates
Areal uxes across the sediment-water interface from sediment pore water data were determined with a model considering diffusion and reaction of dissolved compounds. 28The model was tted to measured pore water concentration proles by adjustment of reaction rates.The concentration change, dC, of a compound at depth z below the sediment-water interface with time dt is: and depends on the diffusion coefficient of the compound in the sediment, D s , and the sum of process rates affecting the compound, SR(z).We introduced two sediment layers with different reaction rates that allow accounting for easily degradable (R 1 ) and more refractory (R 2 ) organic matter, hence for the upper layer SR(z) ¼ À(R 1 + R 2 ) and for the lower layer The mathematics of the model are detailed in Epping and Helder (1997). 29Diffusion coefficients at 4 C were taken from Li and Gregory (1974). 30The porosity of the sediment (f) was measured and the resulting formation factor (F ¼ 1.02 Â f À1.81 ¼ 1.23) was used as suggested by Maerki et al. (2004). 31

Results and discussion
The two regions chosen for our investigation showed clear quantitative differences in the measured parameters.Sedimentation rates of the two cores close to the river mouth were much higher (SR ¼ 2.3 cm per year) than the ones in the deeper part of the lake (SR ¼ 0.4-0.6 cm per year).This already showed that sites on the Rhone River delta received more particulate material than the ones further out in the lake.The ARs amount to 26.8 and 28.4 kg m À2 per year at sites #1 and #2, and to 4.1 and 2.4 kg m À2 per year at sites #4 and #5 (Table 1).The values from the delta are in the range of published values in front of the Rhone river of 45.9 kg m À2 per year. 32The lowest OC contents (0.3-0.6%, Fig. 1) as a result of dilution by mineral particles delivered by the Rhone River were found at site #2, which together with site #1 is located closest to the river entrance.High OC values of $1% and up to 1.35% in the deeper sediment were detected at site #5, the core furthest away from the delta.This resulted in an AR TOC of 131.7 and 129.8 gC m À2 per year (sites #1 and #2) and 31.2 and 19.8 gC m À2 per year (sites #4 and #5), respectively for the delta and profundal sites (Table 1).
C/N ratios are a way to characterize the OM as autochthonous, i.e. locally produced in the lake, and allochthonous, i.e. originating from outside the aquatic system.Autochthonous OM consists of algae and bacteria and hence is rich in N (proteins), whereas terrestrial OM (leaves, grass, etc.) is depleted in N and enriched in O-containing compounds (e.g.lignins). 33herefore, C/N values for aquatic plants lie between 4 and 10, whereas terrestrial vascular plants can have values >20. 34In this study C/N ratios decreased continuously from 10.1 close to the river mouth to 9.1 and 7.9, and nally 7.4 at the most profundal site #5 (Table 2).Although TOC contents increased towards the open lake due to decreasing dilution by mineral particles transported by the river, the increase of N was even higher leading to lower C/N ratios.C/N ratios (Table 2) indicated a higher terrestrial inuence at the delta sites contrary to the deeper sites where autochthonous OM was prevalent.
The carbon isotopic composition of the OM (d 13 C org values) is used frequently in the marine environment to distinguish marine from terrestrial OM. 35 However, in lacustrine systems the d 13 C org values of those two sources are rather indistinguishable since autochthonously produced OM is isotopically very similar to terrestrial OM.The reason is that the d 13 C DIC in lakes in general and in the water column of Lake Geneva in particular is already very depleted (around À10 to À15&, see the uppermost peeper values in Fig. 2), which leads to isotopically very light autochthonous OM compared to marine plankton.Carbon isotope values of OM in sediments of Lake Geneva were between À24.8 and À29.0& (Fig. 1) and hence were difficult to interpret with respect to differentiation of the

Paper
Environmental Science: Processes & Impacts OM as mentioned above.Nevertheless, our d 13 C org results showed that OM isotopically was increasingly heavier towards the bottom of the cores indicating the more labile character of the isotopically light aquatic OM.Based on this nding one could argue that values above À28& are indicative of a more autochthonous material since these values are unusual for an allochthonous material and hence the profundal cores have a higher contribution of autochthonous, i.e., algal material.
One parameter used quite frequently in sediments is the chlorin index (CI) together with chlorin concentrations. 25,36hereas the chlorin concentration gives a measure of how much pigment related OM is present in a sediment sample, the CI indicates how refractory is the OM.Relatively fresh OM has CI values of $0.6 (fresh chlorophyll has a value of 0.2), whereas more degraded OM shows values between 0.8 and 1.0.The sediments of Lake Geneva showed CI values between 0.5 and 0.9

View Article Online
with lower values on the delta and higher values at the distant sites #4 and #5 (Fig. 1).Hence, the material closer to the river mouth was fresher than the material deposited far away from the river mouth.Chlorin concentrations normalized to TOC, however, were higher for the distant sites #4 and #5 with 1226 and 1358 mg g À1 TOC compared to the delta sites #1 and #2 with 1032 and 822 mg g À1 TOC.Combined we see higher amounts of degraded pigment material in the distant sites.8][39] Especially long chain n-alkanes and alcohols have been used as indicators for terrestrial OM.These compounds are constituents of surface waxes covering the leaves of higher plants. 40The sum of the long chain n-alkanes (C 27 , C 29 , C 31 ) and long-chain alcohols (C 24 , C 26 , C 28 ) was calculated.The n-alkanes showed contents of 277 to 429 mg per gC (#1 and #2) and 126 to 175 mg per gC (#4 and #5) with around 2 times higher concentrations closer to the river mouth.Similar results were obtained for the long-chain alcohols with 804 to 1128 mg per gC and 368 to 538 mg per gC, respectively.Calculating the AR for long chain nalkanes and alcohols revealed an even more drastic difference and showed that the deltaic sites received 6 to 25 (n-alkanes) and 8-16 (alcohols) times more terrestrial OC than the profundal sites.In conclusion, a higher terrestrial input prevailed towards the river mouth supporting the above interpretation based on C/N ratios.
The nal products of OM degradation are CH 4 and CO 2 (quantied as DIC dissolved in the pore water, neglecting the DIC produced by calcite dissolution in the sediment).Both constituents were released from Lake Geneva sediments into the water column, as suggested by the shape of their concentration proles (Fig. 2), which decreased towards the sediment surface.CH 4 uxes varied spatially and were more intense in the river delta (1.96 and 2.34 at sites #2 and #1) compared to the deeper part of the lake (0.69 and 0.50 at sites #4 and #5, Table 2).The DIC uxes were 2.0 and 1.4 mmol m À2 d À1 at sites #1 and #2 and 1.0 and 0.9 mmol m À2 d À1 at sites #4 and #5, respectively.Although DIC uxes at the deltaic sites were not 4 to 5 times higher than those at the profundal sites, as was observed for CH 4 , they were still 1.5 to 2 times higher hinting to more pronounced uxes of the degradation products of OM at the deltaic sites.
The main focus of this study was to investigate the difference in electron acceptors used for the degradation of OM depending on its concentration and origin, i.e. lacustrine or terrestrial, in a lacustrine system.Hence, uxes of O 2 , SO 4 , and NO 3 were calculated from the concentration proles determined from peepers and sediment cores (Fig. 3, O 2 concentration proles not shown but up to 9 proles per site were combined to estimate uxes).
The O 2 consumption of the sediment was similar for the profundal cores with 8.8 (site #4) and 8.2 mmol m À2 d À1 (site #5), whereas only half of this ux was measured at sites #1 and #2 close to the river inow with 4.3 and 4.1 mmol m À2 d À1 , respectively (Table 2).O 2 concentrations in the deep waters were around 4 mg L À1 during the time of sampling (The Commission Internationale pour la Protection des Eaux du Léman, CIPEL).As described before for lacustrine sediments 12  The uxes of reduced species from the sediments (Table 2) decreased from the delta to the sites in the open lake for CH 4 (from 2.3 to 0.5 mmol m À2 d À1 ) and NH 4 + (from 0.31 to 0.13 mmol m À2 d À1 ).Fluxes of Fe(II) and Mn(II), however, increased although they were generally very low (between 9 Â 10 À5 and 7.6  View Article Online CH 4 uxes, whereas sites #4 and #5 comprised higher OC contents but lower CH 4 uxes.However, one has to take into account that the AR of OC was about ve times higher at the delta sites than that at the deep sites and the low content is a 'dilution' by non organic, i.e. minerals, delivered by the river. If the production rate of CH 4 is related to the deposition rate of OC we receive an efficiency for the methanogenesis from total OC (in mmol CH 4 per gC) that is now clearly higher at the deeper sites than in the delta (Table 1).Hence, it seems that the AR OC is not solely responsible for the degree of CH 4 production, but that beside the accumulation of OC additional parameters like freshness/degradability and composition of the OM play a role.
The set of available pore water data allowed estimating the mineralization rates of OM R total at the four sites, and the fraction of OM degraded aerobically (R aerobic ) and anaerobically (R anaerobic ).As exemplied by Maerki et al. (2009) 12 oxic mineralization was calculated from the ux of O 2 to the sediment

View Article Online
subtracting reduced compounds diffusing up from the deeper sediments thus contributing to the consumption of O 2 : The stoichiometric factors dene the equivalents of O 2 consumed per equivalent of reduced compounds.The areal mineralization of OM is expressed in mmol C m À2 d À1 (Table 2) assuming the mineralization of one equivalent of OM per equivalent of O 2 , thus neglecting nitrication.The fraction of OM mineralized with electron acceptors other than O 2 (R anaerobic ) is calculated as: Since we could not account for compounds that were potentially removed from the pore water and buried such as precipitates of Fe(II)/Mn(II) suldes, vivianite, siderite, etc., those sinks were neglected.
The rates for totally mineralized OM (R total ¼ R aerobic + R anaerobic ) presented in Table 2 at the deeper sites #4 and #5 (average of 9.7 mmol C m À2 d À1 ) were two times higher than those at the shallower sites #1 and #2 (4.7 mmol C m À2 d À1 ).Not only the rates were different but also the mineralization pathway differed between the shallow sites #1 and #2 in the Rhone River delta, and the profundal sites #4 and #5.Only 0-6% was mineralized aerobically at the shallow sites, since almost all O 2 was used to oxidize the large ux of CH 4 from below (Fig. 4).At the deeper sites, the situation was reversed and the fraction of aerobically degraded OM was between 72 and 78% (Fig. 4).Also, the contribution of SO 4 2À to R anaerobic was very distinct between both areas with 44-49% at the deeper sites and only 6-14% at the shallower sites.The contributions of NO 3 À (1-5.5%), and of Fe and Mn (<0.2%) to the mineralization of OM in the sediments were negligible.The increased generation of alkalinity observed in the sediment pore water of the shallow sites (Fig. 2) lessens up to 10% by the dissolution of calcite (determined from Ca 2+ pore water proles, data not shown), and even less at the deeper sites.However, the reduction of SO 4 2À with both, OM or CH 4 produces one equivalent of alkalinity, which for sites 1 and 2 amounts to 11-16% of the HCO 3À ux but to $60% for sites 4 and 5. Similarly, alkalinity generated by denitrication is $3% at site 2, but 6-11% at sites #4 and #5.Contributions from Fe and Mn reduction are negligible.The remaining HCO 3À originates from the production of CO 2 by mineralization and the subsequent decrease of pH.
In conclusion, a consistent picture emerged that the mineralization pathways and rates in the sedimentary depositions in the Rhone delta area were signicantly different from those out in the profundal parts of the lake.The much higher CH 4 uxes in the delta compared to the open lake sites are explained by the high deposition rate of OC in this area where OC accumulation was approximately 5 times higher in the delta at sites #1 and #2 (132 and 130 gC m À2 per year) than at sites #4 and #5 (31.2 and 19.8 gC m À2 per year, Table 1).Bulk deposition rates differed even by a factor of 8 (Table 1) demonstrating the high load of mineral particles transported by the river and diluting the OC.These "diluting" conditions led to almost two times higher content of OC (in percent %) at the profundal than at the delta sites.Estimating the efficiency of CH 4 formation per equivalent of OC (the annual ux of CH 4 per annually deposited C) therefore resulted in 5.5 to 6.4 mmol CH 4 per gC and of 8.0 to 9.1 mmol CH 4 per gC for the delta sites and open lake sites, respectively (Table 1).This supports our ndings that a significant part of the OC deposited in the delta contained more refractory and thus microbially less degradable, terrestrial material.The higher CH 4 ux on the delta must be recognized as caused by the high AR and the small fraction of aerobic mineralization of OC, which was qualitatively different from the more autochthonous material at the deeper sites and thus could be converted to CH 4 only with a lower efficiency.This study supports earlier ndings that have claimed that burial in lakes is high 9 and it is especially high when high amounts of allochthonous material are delivered. 4It also points out that it is necessary not only to measure the OM concentration but to characterize the origin and the degradation state of OM to estimate how OM is transformed and which share is emitted as CO 2 and/or CH 4 .
Similar to river deltas in the coastal ocean, river deltas of large lakes are active players in the carbon cycle as OM is buried, transformed or emitted from bottom sediments.However, an accurate estimate of the amount of carbon cycled in these active zones remains relatively unconstrained, largely because the fractions of various carbon sources are unresolved.Our results suggest that carbon sources and degradation pathways can differ widely in some freshwater systems, further complicating the situation.While similar analyses as those presented here would be necessary to accurately identify OC sources and degradation pathways, the clear trends in our data indicate that river deltas are regions of substantial carbon cycling, and that any system resembling our study site, particularly if it is an obvious hot spot for CH 4 production (i.e., presence of ebullition), will exhibit the same redox conditions and OM degradation patterns we observed in the Rhone delta.

2. 11
Determination of ammonium, sulfate, nitrate, iron and manganese NH 4 + concentrations were determined within hours following the sampling, using a standard photometric method following the protocol of GSM, 2002 (DIN 38 406-E5).Pore water aliquots were kept frozen until analysis of SO 4 2À and NO 3 À concentrations by ion chromatography (ICS 1000 Ion Chromatography System, Dionex).
O 2 penetrated only the uppermost mm of the sediment with the deepest penetration of 1.8 mm at site #1.The other two electron acceptors NO 3 À and SO 4 2À showed very small uxes compared to O 2 .The NO 3 À ux could not be determined at site #1 due the large scatter of the data.At site #2 the NO 3 À ux was lower (0.04 mmol m À2 d À1 ) than that at the deeper sites #4 and #5 where 0.06 and 0.1 mmol m À2 d À1 were measured.Similarly, the SO 4 2À ux into the sediment was lower at the shallower sites #1 and #2 (0.15 and 0.32 mmol m À2 d À1 ) than that at the deeper sites #4 and #5 (0.62 and 0.54 mmol m À2 d À1 ).

Fig. 2
Fig. 2 Methane (CH 4 ) and dissolved inorganic carbon (DIC) concentrations in the pore waters of the four sediment cores, the carbon isotopic composition of the DIC ( 13 C DIC ), and the alkalinity (ALK) at the four sites.

Fig. 4
Fig.4Aerobic and anaerobic degradation rates of the river delta sites and of the profundal sites.Whereas sites #1 and #2 show a degradation that is mainly triggered by anaerobic degradation (94 and 100%), sites #4 and #5 show a dominance of aerobic degradation with an anaerobic fraction of only 28 and 22%.

Table 1
Dates when cores were taken.Coordinates, water depths, and sediment depths of the studied cores.TOC concentration, sedimentation rates (SRs), porosity, and accumulation rates (ARs) of the analysed sediment cores.TOC concentrations and porosity were averaged below 3 cm and over the uppermost 9 cm, respectively.Efficiency describes the CH 4 produced per g deposited carbon This journal is © The Royal Society of Chemistry 2015 Environmental Science: Processes & Impacts Paper Open Access Article.Published on 01 December 2014.Downloaded on 1/1/2019 9:20:40 PM.This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Online

Table 2
Areal fluxes in [mmol m À2 d À1 ] across the sediment-water interface in the Rhone River delta estimated from sediment pore water measurements.The mineralization rates (R) are in [mmol m À2 d À1 ] Open Access Article.Published on 01 December 2014.Downloaded on 1/1/2019 9:20:40 PM.This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.