A series of biogeochemical studies were conducted at the southern summit of Hydrate Ridge, offshore Oregon. Using the submersible DSV Alvin, sediment push cores were collected from two distinct seep environments characterized by the presence of clam fields (CF) or microbial mats (MM) at the sediment-water interface; samples were also collected from a nearby reference site characterized by a barren surface at the sediment-water interface. Sediment samples from each setting were analyzed for the depth distributions of total organic carbon (concentrations, δ13C and Δ14C), total sedimentary nitrogen, and microbial abundance. Pore fluids were extracted and analyzed for sulfate, alkalinity, sulfide, organic carbon, and volatile organic acids. These depth distributions clearly indicate the presence of three distinctive biogeochemical settings in the surface sediments of Hydrate Ridge, and provide the basis for a comparative biogeochemical analysis. Both CF and MM sites display properties indicating enhanced microbial activity in the subsurface, compared with the reference site. MM sites display evidence of net biomass production in the subsurface; however, a loss of sediment nitrogen relative to the reference site indicates that mineralization is also enhanced. Calculations based on the removal of nitrogen indicate that greater than 30% of autochthonous organic material is lost to enhanced mineralization in the top 23 cm of one MM site. An isotope mass balance of sediment-bound organic carbon indicates a mixed source, including methane and allochthonous organic carbon dissolved in the seep fluids. The concentrations of organic carbon dissolved in seep fluids reach values of 22 mM and provide a first indication that advective transport of dissolved organic carbon from great depth may supply an important source of energy and carbon to “methane seep” communities.
 Marine sediments host the largest reservoir of methane (CH4) on Earth at the present time [Haq, 1999; Kvenvolden, 2002; Kvenvolden et al., 1993; Milkov, 2004; Milkov et al., 2003; Valentine et al., 2004]. Sediment-bound CH4 may be found as solid hydrate, free gas or dissolved in pore fluids, depending on ambient physical and chemical conditions. Because CH4 is a potent greenhouse gas and an important economic resource, the marine reservoir has become an important topic of scientific attention [National Research Council, 2004]. Two current goals in research on marine CH4 are to accurately assess the quantity and distribution of CH4 in marine sediments, and to determine the fluxes of CH4 from the sediment reservoir to overlying waters, and into the atmosphere. Paramount to quantifying the fluxes of methane from the sediment to the water column are understanding the physical, chemical, and biological controls on CH4 in marine sediments, especially along the continental margins where large amounts of CH4 are known to reside. This paper focuses on the biogeochemical processes associated with one such environment, an area of intense methane seepage atop the southern summit of Hydrate Ridge, offshore Oregon. In particular, this work seeks to understand the microbially mediated transformations of carbon and nitrogen in areas of active seepage, how these processes impact sediment chemistry, and the record these processes leave in the sediment.
 All data reported here are from samples collected between 26 July and 31 July of 2002 during cruise 7–18 of the R/V Atlantis. All samples were collected using the DSV Alvin push core system during a series of three dives (3811–3813) to the southern summit of Hydrate Ridge. During each dive, several sites were sampled and at least five parallel push cores were collected from each site. An overview of sample collection and analyses performed is given in Table 1.
Table 1. Data Presented From the Different Study Sites
TOC, total organic carbon; δ13C-TOC, carbon isotopic composition of total organic carbon; 14C-TOC, natural radiocarbon abundance in total organic carbon; TN, total nitrogen; TIC, total inorganic carbon; MA, prokaryotic abundance; Alk, pore water alkalinity; Ca, pore water concentration of Ca2+; S, pore water concentrations of sulfate and sulfide; DOC, concentration of dissolved organic carbon in the pore water; δ13C-DOC, carbon isotopic composition of dissolved organic carbon; VOAs, pore water concentrations of formate, acetate and lactate; δ13C-DIC, carbon isotopic composition of dissolved inorganic carbon.
 The choice of sampling sites was based on visual characteristics of the sediment surface (Figure 2). Sites were distinguished as follows. Microbial mat (MM) sites harbor dense populations of filamentous microbes at the sediment water interface. These sites have previously been shown to be associated with the net outflow of fluids from the sediments to the water column [Tryon and Brown, 2001; Tryon et al., 2002, 1999], with maximum flux rates of several meters per year [Tryon and Brown, 2001]. The visible microbial mats are presumed to be composed primarily of sulfide-oxidizing bacteria, such as Beggiatoa sp. Clam field (CF) sites are characterized by the presence of abundant clams at the sediment water interface. These environments have previously been shown to exhibit a transient advective flux of interstitial waters. That is, sometimes there is net fluid flux out from the sediment to water column, and sometimes seawater is drawn in to the sediment. Net fluid flux rates at these sites are typically only a few centimeters per year either in or out of the sediment, much lower compared with the MM sites. The clams are presumably vesicomyid clams of the genus Calytogena [Sahling et al., 2002]. The reference site (3812-1) chosen for this study was adjacent to two other sites (∼40 m from 3812-2 and 3812-3), but exhibited no surficial features indicative of fluid flow.
2.3. Analytical Procedures
2.3.1. Sample Processing
 After sampling, the push-cores were placed in the front basket of the DSV Alvin (Figure 2b). After the submarine was brought on deck the cores were immediately removed from the basket, capped, and stored near in-situ temperature (∼4°C) under anoxic conditions until they were processed. Processing generally began within an hour of the DSV Alvin returning to the ship. The parallel cores from each site were processed using various methods depending on the analyses. The core designated for geochemistry was extruded and sectioned inside of an anoxic glove bag in an environmental room maintained at 4°C. The sediment was packed into centrifuge tubes, capped immediately inside the glove bag, and centrifuged at 8000 rpm for a period of 10 min inside a refrigerated centrifuge set at 4°C. The centrifuge tubes were then transferred to a second anoxic glove bag at which time the pore waters were filtered, divided, and fixed for various analyses. The resulting pore fluids were used for analyses of sulfate concentration, alkalinity, δ13C and δ18O of dissolved inorganic carbon (DIC), δ15N of dissolved nitrogen, sulfide concentration, dissolved organic carbon (DOC) concentration, δ13C of DOC, and organic acid concentrations, as described below. The remaining sediment was frozen at −20°C and later used for analyses of total organic carbon (TOC), δ13C and Δ14C of TOC, carbonate content, and total nitrogen (TN). A second core was processed in the anoxic bag for cellular counts. Other cores were processed for biological studies and results are presented elsewhere [Hill et al., 2004].
2.3.2. Chemistry of the Bulk Sediment
 Frozen sediment was analyzed for total organic carbon content, δ13C and Δ14C of total organic carbon, total inorganic carbon content, and total nitrogen content. TOC, total inorganic carbon (TIC), and TN were determined for the sediment solid phase using a PE 2400 CHN analyzer. Dried and homogenized sediment samples were first analyzed for total carbon (TC) and TN contents. TOC was analyzed after acidifying the sediments with 10% HCl. The difference between TC and TOC was calculated as TIC. The standard deviations of the measurements were ±3% for TOC and ±4% for TN, respectively, based on replicated standard analyses. For δ13C and Δ14C measurements, approximately 1.0 g dried and homogenized sediment from each depth was first acidified with ultra purity 10% HCl in pre-combusted quartz tubes. After adding CuO and Ag foil (pre-combusted at 550°C for 4 hours), the samples were vacuum dried and the tubes were flame-sealed under vacuum. The sediment samples were combusted at 850°C for 1 hour according to standard techniques [Druffel et al., 1992]. Resultant CO2 was measured and split into subsamples for 13C and 14C analyses. δ13C was first measured using a Finnigan Delta-S isotope ratio mass spectrometer at the Boston University Stable Isotope Laboratory with an overall uncertainty of ±0.2‰. The Δ14C was then measured using accelerator mass spectrometry (AMS) at the National Ocean Sciences AMS (NOSAMS) facility at Woods Hole Oceanographic Institution. 14C contents were calculated as fraction modern relative to NBS Oxalic Acid I standard and reported as Δ14C (‰).
2.3.3. Abundance of Prokaryotic Cells
 Intact cores were extruded in 3-cm sections, and two separate samples were taken from each section. Sediment was placed into eppendorf tubes, and fixed with 10% formalin in sterile seawater. Samples were stored at 4°C until being processed in the lab (2–3 weeks later). Roughly 0.1–0.2 g of sediment were diluted in 7 mL sterile seawater in a 15-mL Falcon™ tube. Tween-80 was added to a final concentration of 0.1 mg/mL, and the samples were individually sonicated using a Fisher Scientific Sonic Dismembrator for 3 min, with 30 s of shaking between each minute of sonication. The samples were centrifuged briefly at low speed to settle the sediment, and the supernatant was removed. For a discussion of the impact of sonication and centrifugation on enumeration of prokaryotic cells, see Lunau et al. . The remaining sediment was then dried at 70°C overnight, and the dry weight for each sediment sample was measured. This normalizes for differences in water content among different sampling sites and different depths. DAPI staining was performed according to Porter and Feig  with adjustments given here. DAPI stain was added to the sonicated supernatants to a final concentration of 5 μg/mL, and samples were placed in the dark for at least 10 min. Samples were further diluted to concentrations optimal for cell counting with sterile seawater and filtered onto a black polycarbonate 0.2-μm filter (Osmonics). Samples were visualized using an Olympus BX51 epifluorescence microscope. At least 10 fields of view were counted on each slide examined. This method quantifies the total number of non-aggregated prokaryotic cells. Similar methods have been shown to not break apart the archaeal/bacterial aggregates involved in anaerobic methane oxidation [Boetius et al., 2000; Knittel et al., 2003; Orphan et al., 2001a; Orphan et al., 2002]; such aggregates have previously been shown to constitute a significant portion of the microbial population in CF and MM sites [Boetius et al., 2000; Knittel et al., 2003].
2.3.4. Chemistry of the Pore Fluids
 Following separation by centrifugation, purified pore fluids were divided into aliquots intended for several geochemical analyses. For DIC measurements, 2 mL of pore water was injected into a Vacutainer™ containing 0.5 ml of 5% W/V mercuric chloride. For alkalinity, 3 mL of fluid was pipetted into a glass vial using a calibrated pipetter. One milliliter of pore fluid was sealed in a glass ampoule to be used for the analysis of major elements: Cl, Mg, Ca, total S, K, and Na. Another 1-mL aliquot was injected into a Vacutainer™ containing 1 mL of a 4 M cadmium nitrate solution for analysis of oxidized sulfur species. The remaining pore fluid was used for quantification of DOC, including organic acids.
 Fresh aliquots of pore fluid were analyzed shipboard for alkalinity [Gieskes et al., 1991]. The pH was measured with an Accumet 925 pH/ion meter, and the samples were titrated with 0.1 N hydrochloric acid (previously standardized to determine its true normality). The Gran method was used to evaluate the second equivalence point of carbonate titration in seawater. Samples that obviously contained large amounts of sulfide were given highest priority for analysis.
 Pore water sealed in glass ampoules was returned to shore and analyzed for Mg, Ca, total sulfur, K, and Na using inductively coupled plasma optical emission spectroscopy (Perkin Elmer Optima 3000DV ICP-OES), as described previously [Wardlaw and Valentine, 2005]. These samples were diluted 250-fold using a combination of gravimetric and (calibrated) volumetric methods. The Cd(NO3)2-treated pore waters in the Vacutainer™ were diluted and analyzed in a similar manner. The precision for these analyses is 3% for sulfur. The sulfur species in the Vacutainer™ solutions were taken to represent the sulfate content of the pore waters, since the cadmium nitrate was in clear excess of any anticipated sulfide content. By subtracting these sulfate values from the total sulfur obtained from analysis of untreated ampoules, minimum sulfide concentrations were obtained. Although the unsealed ampoules were kept under anoxic conditions in the glove bag and the openings covered with parafilm during sample processing, this method would not completely prevent degassing of hydrogen sulfide. In addition, it is necessary to remove the ampoules from the glove bag in order to flame-seal them, and it is possible that during this process, some sulfide could oxidize to elemental sulfur and sulfate.
 Aliquots of pore water ranging in mass from 1 g to 4 g were separated for DOC analysis. Each sample was forced through a 0.2-μm nitrocellulose syringe filter to remove residual particulate matter, and transferred into a pre-combusted (450°C for 4 hours) borosilicate vial. Samples were then treated with 20 μL of 6 N HCl, and frozen until being analyzed. Because of the high DOC concentrations in the pore waters, samples were diluted 50 to 150 fold into low carbon water (LCW) prior to analysis. DOC samples were analyzed via high temperature combustion using a Shimadzu TOC-V in a shore-based laboratory at the University of California, Santa Barbara. The operating conditions of the Shimadzu TOC-V were slightly modified from the manufacturer's model system. The condensation coil was removed and the head space of an internal water trap was reduced to minimize the system's dead space. The combustion tube contained 0.5-cm Pt pillows placed on top of Pt alumina beads to improve peak shape and to reduce alteration of combustion matrix throughout the run. CO2-free carrier gas was produced with a Whatman® gas generator [Carlson et al., 2004]. Each sample was drawn into a 5-mL injection syringe, acidified with 2 M HCl (1.5%) and sparged for 1.5 min with CO2-free gas. Three to five replicate samples (100 μL) were injected into the combustion tube maintained at 680°C. The resulting gas stream passed though several water and halide traps, the CO2 in the carrier gas was analyzed with a non-dispersive infrared detector, and the resulting peak area was integrated with Shimadzu chromatographic software. Injections continued until at least three injections met the specified range of a SD of 0.1 area counts, CV ≤ 2% or best three of five injections. Extensive conditioning of the combustion tube with repeated injections of LCW and deep seawater was essential to minimize the machine blanks. After conditioning, the system blank was assessed with UV oxidized LCW. The system response was standardized daily with a four-point calibration curve of potassium hydrogen phthalate solution in LCW. All samples were systematically referenced against low carbon water and deep Sargasso Sea reference waters (2600 m) and surface Sargasso Seawater every fourth to sixth analysis [Hansell and Carlson, 2001]. Daily reference waters were calibrated with DOC certified reference material provided by D. Hansell (University of Miami). The term DOC, as applied in this work, excludes volatiles such as methane.
 The δ13C of pore water DOC and δ15N of total dissolved nitrogen was analyzed from the same aliquots as the DOC samples above. Analyses were performed using an automated elemental-analyzer isotope-ratio mass spectrometer at UMass Dartmouth. This system consists of an elemental analyzer coupled to a Finnigan MAT 251 mass spectrometer for isotopic analyses in continuous flow mode. For these analyses, 500–1000 μl acidified pore water (due to limited sample volume, but DOC concentration ranged from 1.3–21 mM) were placed on pre-combusted 25-mm GF/F filters. The filters were then vacuum dried and analyzed for δ13C and δ15N. Blanks were measured and used to correct sample results (based on blank C and N amount and their isotopic values). Blank corrections were generally within ±1‰ of the sample signals. Standardization was ensured both by combustion of solid materials of known isotopic composition and by injections of standard gases into the carrier gas (He). Reproducibility for δ13C and δ15N were better than ±0.15‰ and ±0.2‰, respectively.
 Aliquots of pore water ranging in volume from 1 to 2 mL were used for quantification of short chain (1–5 carbons in length) volatile organic acids (VOA). Each sample was forced through a 0.2-μm nitrocellulose syringe filter to remove residual particulate matter, and transferred into pre-combusted (450°C for 4 hours) borosilicate vials. The method used for quantifying VOAs is a modified version of one described by Albert and Martens . The 2-nitrophenyl hydrazine (NPH) and l-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC, protein sequencing grade) were obtained from Sigma Chemical Company, St. Louis, Missouri. The NPH was recrystallized from hot water and then kept frozen until it was used. The EDC was kept frozen and not further purified. Reagent grade pyridine was obtained from Fisher Scientific and was redistilled twice before use. Trace metal grade hydrochloric acid was obtained from Fisher Scientific. Derivatizations were carried out in 8-mL borosilicate glass vials with Teflon-lined caps. The vials were acid-cleaned, rinsed with nano-pure (or milli-Q) water and then combusted at 450°C for 3–4 hours before use to eliminate contaminants. One-milliliter aqueous pore water samples were derivatized using the following steps: (1) 0.2 mL pyridine buffer was added (equal volumes of redistilled pyridine and trace metal HCl), (2) samples were then bubbled for 4 min with a stream of nitrogen to remove CO2, (3) 0.1 mL 0.1 M NPH in 0.25 M HCl was added to the solution, (4) 0.1 mL 0.3 M EDC in milli-Q water was added and the vials were mixed and allowed to sit at room temperature for 1.5 hours, (5) 0.2 mL 40% KOH (w/v) was added and the samples were mixed and heated in a heating block at 70°C for 10 min. The metal hydroxide flock (from the calcium and magnesium in seawater) was removed from the supernatant by centrifugation. The resulting derivatives were analyzed using a Shimadzu model LC600 liquid chromatograph equipped with a Shimadzu SPD-6AV UV/VIS detector. The detector was fitted with a 10-mm flow cell and was operated at a wavelength of 400 nm. The column used was a 22-cm Brownlee C8 cartridge with a 1.5-cm C8 guard column and a polymeric reversed-phase (PRP) guard cartridge in the sample loop as a pre-concentrating column. The total flow rate for the acetate-buffered mobile phase was 1.5 mL min−1. The chromatographic solvent contained the following: 2.5% n-butanol, 50 mM sodium acetate, 2 mM tetrabutylammonium hydroxide (TBAOH) (Sigma Chemical Co.), 2 mM tetradecyltrimethylammonium bromide (TDTMABr) (Sigma), and phosphoric acid to adjust the pH to 4.5. The mobile phase was presaturated with silica to help preserve the columns by adding ∼0.33 g silicagel to a 2-L batch prior to pH adjustment. After pH adjustment the undissolved silicagel was removed by double vacuum filtration, in series, through disposable 0.45-μm filters. The injected sample volume was 0.5 mL, and the supernatant above the flock was forced through the preconcentrator column in the sample loop. This was followed by 1 mL of milli-Q water to wash the strong KOH from the sample loop prior to injection onto the analytical column. Standards of C1-C5 organic acids were derivatized immediately prior to analyzing environmental samples. The standards prepared and ran on the HPLC ranged from 500 nM to 1 mM. All organic acid analyses were performed shipboard.
 Results presented here are from six sites, including three designated as microbial mat sites, two designated as clam field sites and one designated as a reference site. The distribution of these sites between dives and a summary of the analyses performed for each site are provided in Table 1. Throughout the remainder of this section, distinct symbols are used in the figures to differentiate the reference site (circles) from the CF sites (squares) and the MM sites (triangles and diamonds). Images of the seafloor atop hydrate ridge are given in Figure 2 and are representative of the three types of sites investigated.
 The chemical and isotopic results from the solid fraction of Hydrate Ridge sediment provide time-averaged information about biogeochemical processes related to methane seeps, on a timescale of years to centuries. The sediment pore waters provide information about these same processes on shorter timescales, from days to months. Superimposed on these depth distributions is the hydrogeology of Hydrate Ridge, most importantly the advection of pore fluids and gas through the sediment [Deyhle et al., 2001; Haeckel et al., 2004; Teichert et al., 2003; Torres et al., 2001, 2002; Tryon and Brown, 2001; Tryon et al., 2002, 1999]. Previous studies have shown that MM sites are characterized by net outflow of fluids from the sediment to the water column, with flux rates ranging from near zero to 1000 cm yr−1; CF sites are characterized by lower flow rates (typically only a few cm yr−1, with occasional spikes to ∼100 cm yr−1) that exhibit transient directionality, sometimes flowing into the sediment and sometimes flowing out of the sediment. The variability and migration patterns of flow channels on decadal to centennial timescales is not known. The depth distributions of chemicals in the sediment and pore fluid are considered here in relation to these timescales and dynamic flow regimes.
3.1. Reference Site
 The distribution of organic carbon in the sediments of the reference station (Figure 3a) displays some characteristics of other marginal environments [Kaplan et al., 1963]. The gradually decreasing levels of organic carbon from the sediment surface to 13 cm depth, coupled with the invariant δ13C-TOC indicate that subsurface methane has little impact on the organic carbon present in this interval. A slight increase in TOC and depletion in δ13C-TOC below 13 cm depth indicates the possibility of active methane oxidation at greater depth in the sediments, likely driven by upward diffusion of subsurface methane [Borowski et al., 1999; D'Hondt et al., 2002; Valentine, 2002]. The trend of generally decreasing TOC and total nitrogen (TN; Table 2) with depth indicate that normal mineralization processes are active at the reference site. This interpretation is further supported by the higher radiocarbon content of the surface sediment (0–2 cm depth), relative to the deeper sediment (Figure 4), and by the decrease in prokaryotic abundance to depths of 18 cm (Figure 5). An increase in prokaryotic abundance below 18 cm is consistent with a deeper zone of methane oxidation. The isotopic composition of organic carbon at the reference site ranges from −22.8 to −23.7‰, and the carbon to nitrogen ratio (C/N) ranges from 9.3 to 12.5; these values are consistent with a mixed terrestrial and marine source of organic material and the corresponding trends are further consistent with normal diagenetic patterns in marine sediments (Table 2). Overall, the sediments at the reference station display no indications of methane seepage, past or present.
Table 2. Chemical Properties of Seep and Reference Sediments From Dive 3812
 The depth distributions of pore water sulfate concentration and alkalinity (Figure 6a) at the reference site display a homogenous distribution from the sediment surface to 13 cm depth. This same trend is apparent in the calcium concentrations, and dissolved sulfide concentration was below detection (Table 3). Below 13 cm depth the pore water alkalinity begins to increase, while sulfate and calcium concentration decrease (Figure 6a; Table 3). The δ13C-DIC is relatively enriched in 13C from the sediment surface to 8 cm depth, and becomes significantly depleted between 13 and 26 cm depth (Figure 7a). The linear depth distribution of sulfate, alkalinity and calcium in the upper 13 cm of sediment provides a strong indication of bioturbation in this zone. Below the depth of 13 cm, the increase in alkalinity and decrease in sulfate, calcium and δ13C-DIC indicate active methane oxidation at sediment depths below 26 cm. Previous investigators have identified sediment communities with high abundance of the solemyid bivalve, Acharax, at the southern summit of Hydrate Ridge [Sahling et al., 2002]. Acharax communities have been observed at the periphery of CF sites and are believed to “mine” the sediments for sulfide, which is oxidized by endosymbiotic bacteria [Sahling et al., 2002]. Results presented here are consistent with the bioturbation activity of Acharax, or similar infauna, to depths between 13 and 18 cm at the reference site. The results further indicate that methane oxidation in the shallow subsurface provides sulfide to this community, apparently independent of fluid advection.
Table 3. Depth Distributions of Selected Chemical Species in Sediment Pore Watersa
 The sediment pore waters at the reference station display high levels of DOC (Figure 8a). Typical DOC levels in deep ocean waters are below 60 μM [Hansell and Carlson, 2001], compared to 1300 to 4400 μM for the pore waters at the reference site. The depth distribution of DOC displays an increase from 1400 to 4300 μM around 5 cm depth, followed by a gradual decrease down core. The δ13C-DOC displays a decrease of 4‰ between the sediment-water interface and 5 cm depth, followed by a gradual decrease down core. The δ13C-DOC is depleted by about 3‰ relative to sediment TOC between the sediment surface and 5 cm depth, and is depleted by about 7‰ relative to TOC at depths greater than 5 cm, indicating that DOC is not derived solely from the diagenesis of autochthonous TOC. Depth distributions of pore water acetate (Figure 9a), formate (Figure 9b), and lactate (Figure 9c) display more complex profiles including subsurface maxima. Other VOAs including propionate, butyrate, isobutyrate and isovalerate were found to have concentrations below 1 μM. VOAs account for about 4% of all DOC present, averaged over the whole core. VOAs represent a labile component of the DOC pool and are presumably produced in situ by the activity of fermentative prokaryotes and heterotrophic eukaryotes.
3.2. Clam Field Sites
 The elemental and isotopic composition of sediment at the CF site differs significantly from the sediment at the reference site. The amount of TOC is slightly lower in surface intervals (0–5 cm) of the CF site relative to the reference site, but is generally greater at depths below 8 cm (Figure 3b). The δ13C-TOC ranges from −24.5 to −27.1‰ between the sediment surface and 18 cm depth, but becomes depleted by more than 7‰ below 18 cm depth (Figure 3b). The radiocarbon profile for the CF site displays an inversion with the highest radiocarbon content in the 13–18 cm interval (Figure 4). The abundance of single celled prokaryotes is greater than twice that at the reference site, and shows an increase below depths of 15 cm. The C/N ratio of the sediments between the surface and 18 cm depth ranges from 8 to 8.9, and is slightly lower than at the reference site; however, the C/N ratio increases to 14 in the 23 to 28 cm interval (Table 2). These results are consistent with biomass accumulation from two distinct biogeochemical regimes at this site: methane-based productivity at depths below ∼18 cm, and sulfide based productivity in the surface interval (0–18 cm depth). Below 18 cm depth the increase in TOC, corresponding decrease in δ13C-TOC, increase in the C/N ratio, and increase in abundance of prokaryotes indicate the accumulation of biomass derived in part from methane. Between 13 and 18 cm depth the elevated levels of radiocarbon are consistent with CO2 fixation associated with the sulfide-oxidizing symbionts of Calyptogena species. Between 2 and 18 cm depth the δ13C-TOC values of −27‰ and the low C/N ratios are further consistent with the accumulation of carbon and nitrogen by Calyptogena species or their sulfide-oxidizing symbionts.
 Pore water chemical data is available from two CF sites, 3811-2 and 3812-2. The depth distributions of pore water sulfate concentration and alkalinity (Figure 6b; Table 3) at both sites display a rapid depletion of sulfate and increase in alkalinity between the sediment surface and 13 cm depth. A similar trend is apparent in the depth distributions of calcium (Table 3) and δ13C-DIC (Figure 7b; 3811-2 only). Dissolved sulfide concentration (Table 3) is low at the sediment water interface and increases to a maximum at 8–13 cm depth. A down core decrease in sulfide concentration is apparent below 13 cm depth in the deeper of the two cores. These profiles indicate rapid methane oxidation coupled to sulfate reduction (e.g., AOM; for a review see Valentine and Reeburgh ), with maximum rates between 2 and 13 cm depth, and are in good agreement with previous observation of CF sites at the southern summit of Hydrate Ridge [Boetius et al., 2000; Boetius and Suess, 2004; Treude et al., 2003]. The concavities in the depth distributions of sulfate, alkalinity, and calcium between the sediment water interface and 5 cm depth may be related to the transient fluid flow of the CF environment, or to the oxidation of sulfide near the sediment water interface.
 The concentrations of DOC in the pore waters at the CF site are nearly twice that at the reference site, and exceed typical deep water DOC by 2 orders of magnitude. The DOC concentration increases from the sediment water interface to a maximum at 5–8 cm depth, with lower concentrations deeper in the core. However, the δ13C-DOC is variable throughout the core within a range of 18‰. The presumed zone of anaerobic methane oxidation at 8–13 cm depth displays a relatively enriched δ13C-DOC of −22.8‰, while the surface layer (0 to 2 cm depth) displays a more-depleted δ13C-DOC value of −34.3‰. The variability in δ13C-DOC is consistent with a complex flow regime at this site, including the impact of transient flow on carbon release by prokaryotes and infauna. The depth distribution of acetate and lactate both display maxima in the surface layer (0–2 cm depth), with acetate reaching concentration of 100 μM, and lactate reaching concentrations of 1 to 5 μM (Figures 9d and 9f). Formate displays a subsurface maximum between 2 and 8 cm depth, with maximum concentrations of 20–30 μM (Figure 9e). VOAs account for about 1.7% of all DOC present, averaged from the surface to 18 cm depth at site 3812-2. However, in the surface interval (0–2 cm depth) at this site, VOAs comprise 8% of the DOC.
3.3. Microbial Mat Sites
 Sediments underlying the MM site display characteristics different from both the CF and reference sites. The levels of TOC are much higher, with more than twice the overall quantity of organic carbon present compared to the CF site. The δ13C-TOC ranges between −37.7 and −44.3‰ and clearly indicates the accumulation of carbon derived from isotopically depleted sources. The TOC in the 0–2 cm interval is depleted in radiocarbon relative to the same interval at the CF and control site, further indicating the source of organic carbon is radiocarbon dead. The abundance of (non-aggregated) prokaryotes is an order of magnitude greater than at the reference site, indicating enhanced microbial activity. The TN content of the MM site is lower than either the CF or reference sites leading to C/N ratios as high as 49:1, values not typically found in marine sediments. These results indicate the long-term accumulation of 13C-depleted, radiocarbon-dead, allochthonous carbon, presumably originating from lower in the sediment column. A significant component of this carbon is clearly related to microbial metabolites and biomass, though other processes may also contribute.
 Pore water chemical data is available from three MM sites, 3811-1, 3812-3 and 3813-1. The depth distributions of pore water sulfate concentration and alkalinity (Figure 6c; Table 3) at each site display a rapid depletion of sulfate and increase in alkalinity between the sediment surface and 13 cm depth. At two of the sites, pore water sulfate is nearly depleted by 5 cm depth, indicating strong outward flow and/or rapid rates of sulfate reduction. Similar trends are apparent in the depth distributions of calcium (Table 3). The δ13C-DIC (Figure 7c) is relatively depleted at the sediment-water interface in the two cores analyzed. In one core the δ13C-DIC increases gradually with depth from −42‰ at the sediment-water interface to −28‰ at 18 to 23 cm depth. A second core showed a slight decrease in δ13C-DIC from the sediment-water interface to a depth of 8–13 cm. A significant portion of the DIC present is derived from methane oxidation. Dissolved sulfide concentration (Table 3) is low at the sediment water interface, and increases to a maximum at 8–13 cm depth, followed by a down core decrease. Several of the reported sulfide values seem unusually high, but could not be excluded on the basis of experimental error. These depth distributions clearly indicate rapid sulfate reduction and methane oxidation in the shallow subsurface underlying microbial mat sites, and are in good agreement with previous investigations [Boetius et al., 2000; Boetius and Suess, 2004; Treude et al., 2003].
 The concentrations of DOC in the pore waters at the MM sites are higher than at the CF sites and as much as 400-fold higher than typical deep water DOC levels. DOC concentrations were typically lower near the sediment water interface and higher at depth in the core, though no clear trend is apparent (Figures 8c and 8d). The range of δ13C-DOC is from −34‰ to −38‰ for 10 samples collected from two different sites (Figures 8c and 8d), displaying much lower variability than the CF site. The tendency for lower δ13C-DOC values with depth in the core is similar to the reference site but shifted more negative by about 10‰. The δ13C-DOC from 3812-3 is more enriched than either sediment TOC or the underlying methane, indicating the DOC is produced deeper in the sediment column and transported to the surface. The δ13C of DOC produced at depth is presumed to be similar to its source: organic material deposited in the distant past. Marine and terrestrial organic matter both typically have δ13C values more enriched than −30‰. Transport of DOC from depth presumably occurs through the rapid advection of pore waters. VOAs account for between 1% (3812-3) and 2.6% (3811-1) of the DOC present between the sediment-water interface and 13 cm depth, compared with 1.9% for the CF site and 4.3% for the reference site averaged over the same depth interval. Acetate concentrations in the pore fluids range from 21 to 143 μM, and are greater than 100 μM between the sediment-water interface and 2 cm depth (Figure 9g). Formate concentrations range from undetectable to 26 μM and are lower than 4 μM between the sediment-water interface and 5 cm depth (Figure 9h). Lactate concentrations range from undetectable to 6.4 μM, and are generally greatest between the sediment-water interface and 5 cm depth (Figure 9i).
 The geochemical and microbiological analyses performed on surface sediments from three distinct environments atop Hydrate Ridge allows for a comparative biogeochemical analysis of these environments. We focus on three primary processes that distinguish the seep environment: (1) seepage-enhanced mineralization of autochthonous organic material, (2) accumulation of carbon associated with seepage, and (3) heterotrophic versus methanotrophic metabolism in “methane” seeps. In developing these ideas we rely on a variety of previous observations from this environment, including fluid flow patterns [Tryon et al., 2002], methane oxidation and sulfate reduction rate measurements [Treude et al., 2003], microbial diversity assessments [Knittel et al., 2003], and ecological investigations of infauna [Sahling et al., 2002; Sommer et al., 2002].
4.1. Enhanced Mineralization Associated With Fluid Seepage
 Here we seek to quantify the extent to which organic matter mineralization occurs in high-flux seeps relative to background levels of mineralization from a reference site. The MM environment shows clear evidence for net carbon accumulation in the surface sediments, associated with active seepage (Figure 3). However, while average TOC levels are more than twice those found at the reference site, the TN levels are significantly lower. We interpret the lower TN levels in the MM site to indicate enhanced mineralization of autochthonous organic material. This concept is supported by the work of Sommer et al. , who observe an ∼50% depletion in chlorophyll a and pheopigments integrated over the top 10 cm in areas of subsurface gas hydrate, relative to a reference site.
 A simple mass balance can be created to determine the extent of seepage-enhanced mineralization by comparing the amount of TN at the MM site to that present at the reference site, assuming that seepage is the only factor differentiating the two sites. Here we define the amount of organic material lost to enhanced mineralization as OMem. This value can be calculated from the total nitrogen initially present in the sediment, TNi, the total nitrogen present in the MM sediment, TNs, and the quantity of organic material initially present in the sediment, OMi, given the following assumptions: (1) Regular mineralization processes have occurred to the same extent in both reference and MM sites, (2) MM sediments originally contained the same amount of nitrogen as the reference sediment, (3) nitrogen was lost in proportion to carbon during seepage-enhanced mineralization, and (4) no nitrogen was added to the sediments after deposition. Given these assumptions (considered in detail below), OMem can be calculated for each depth interval, after correcting for carbonate precipitation, using the following equation:
 Results from application of this mass balance to a MM site (3812-3) are provided in Table 4. These calculations indicate that between 12 and 50% of autochthonous organic material is lost in the MM site relative to the reference site, with an average of 30% integrated between the sediment-water interface and 23 cm depth. Application of this mass balance to a CF site (3812-2) yields ambiguous results (not shown) as the levels of TN are similar to, and sometimes higher than, the reference site.
Table 4. Calculation of Enhanced Mineralization and Carbon Accumulation at a MM Site (3812-3)a
Quantity of organic material lost owing to enhanced mineralization (em) in seep sediments, expressed as percent of sediment dry weight, corrected for carbonate content.
Quantity of organic material formed (f) in situ, expressed as percent of sediment dry weight.
The δ13C of organic material formed (f) in situ, expressed in per mil, relative to the PDB standard.
 The assumptions made in deriving equation (1) deserve further consideration. The implicit assumption needed to solve equation (1) is that seep and non-seep sediments have similar depositional origins, and that seepage and mineralization are the primary processes acting to distinguish the sites. While the depositional histories are not known for any of the sites, this assumption is supported by the water depth and close proximity of the sites, and by the dramatic impact of seepage on sediment geochemistry. The first explicit assumption is that mineralization has occurred to the same extent in both reference and MM sites; that is, we assume depth distribution of mineralization in the seep and reference sediment should be similar, and that any difference is due to the seepage. We consider this assumption to be conservative as the primary exception, enhanced mineralization by infauna at the reference site, would cause calculated values of OMem to be lower than the actual values. The second assumption, that MM sediments originally contained the same amount of nitrogen as the reference sediment, is not certain. Confounding factors, such as a sediment blow-out, could invalidate this assumption. Results also indicate that methane oxidation is occurring at depth at the reference site, which has an apparent effect on TOC (Figure 3a); this process could also have a slight effect on TN and could impact calculation of OMem for depths below 18 cm. The third assumption is that carbon and nitrogen are removed proportionally during mineralization. This assumption is supported by the linear relationship of carbon and nitrogen abundance in sediments where mineralization is occurring [e.g., Emmer and Thunell, 2000; Stein and Rack, 1995]. However, we are unable to eliminate the possibility that organisms selectively remove nitrogen from the sediment in seep environments, possibly for the purpose of storing oxidized nitrogen species for eventual use as a terminal electron acceptor [McHatton et al., 1996]. The fourth assumption, that nitrogen was not added to the sediments after deposition, is likely false. Some nitrogen was likely incorporated into the MM sediment concurrently with carbon accumulation. We consider this to be a conservative assumption as any nitrogen accumulation at the MM site would cause an underestimation of OMem. Overall, we consider these assumptions to be conservative and find that calculated values for enhanced mineralization likely underestimate the true extent of mineralization.
 Enhanced mineralization, as quantified here, is likely related to the action of heterotrophic prokaryotes in the seep environment. The high abundance of non-aggregated prokaryotes present in the sediments of the MM site (Figure 5), coupled with the diversity of heterotrophic bacteria in these environments [Knittel et al., 2003], provides supporting evidence for this assertion. We propose that seepage of reduced fluids through the sediment provides energy and carbon for the microbial community (see section 4.3) and that this community further facilitates the mobilization and degradation of autochthonous organic material (both TOC and TN). The lack of enhanced mineralization in the CF site (3812-2) is notable, and may be related to the mode of fluid flux or to the action of macrofauna.
4.2. Carbon Accumulation Associated With Fluid Seepage
 Here we seek to quantify the impact of seepage on the accumulation of organic carbon in seep sediments and to determine the ultimate sources of this carbon. Assuming mineralization is the sole process responsible for loss of organic carbon, the following mass balance can be generated for any sediment:
where OMs and Rs represent the quantity and δ13C of organic material currently in the sediment, respectively, OMi and Ri represent the quantity and δ13C of organic material initially deposited in the sediment, OMf and Rf represent the quantity and δ13C of organic material formed in-situ, OMm represents the quantity of initial carbon that has subsequently been lost owing to mineralization, and αm is the fractionation factor associated with mineralization of organic material.
 If we further assume that the reference site and MM sites initially had the same OMi and Ri, and that diagenetic processes have occurred to the same extent in both reference and MM sites (as in section 4.1), then we can compare each depth interval directly between the MM and reference site. In this case the mineralization term, OMm, becomes OMem, the amount of organic carbon lost to enhanced mineralization,
Furthermore, we assume that 13C fractionation associated with mineralization is unity, as isotopic discrimination in large molecules is typically minimal. Equation (3) can then be rearranged and simplified to solve for Rf,
The sole remaining unknown on the right side of the equation, OMf, can be determined from a simple mass balance,
leading to the following equation:
 Values for Rf at different depths in the microbial mat sediment are given in Table 4, along with calculated values of OMem and OMf. Calculations indicate the δ13C of organic material formed in seep sediments ranges from −48.5 to −85‰, with a mass-weighted average value of −54‰. This value is significantly depleted relative to carbon at the reference site, but is significantly enriched relative to the underlying methane, especially considering the isotopic composition of methane-oxidizing prokaryotes [Orphan et al., 2001b] and their lipid biomarkers [Elvert et al., 2003; Hinrichs et al., 1999]. Figure 10 shows the δ13C values for the various carbon pools at this and other seep sites on the southern summit of Hydrate Ridge, both in the solid sediment and in the pore fluid. We conclude that methane is unlikely to be the sole source of carbon contributing to organic carbon accumulation.
 On the basis of the isotopic composition of the various carbon pools, we propose that four primary processes are responsible for the carbon accumulation in seep sediments: (1) methanotrophy, (2) CO2-based autotrophy, (3) DOC-based heterotrophy, and (4) DOC accumulation by sediment particles. A mass balance approach cannot be used to determine the exact partitioning of carbon sources because there are too many unknowns, including the fractionation factor from methane to biomass. However, if we assume CO2-based autotrophy is not important, a fractionation of 15‰ from methane to biomass [Alperin et al., 1988], an unlimited methane supply with an isotopic composition of −65‰, and an unlimited DOC supply with an isotopic composition of −36.5‰, we calculate that 40% of carbon originates from methane, and 60% originates from DOC. Further differentiating the importance of biological versus non-biological DOC accumulation is not feasible with the current data. Despite the various assumptions, this calculation indicates that DOC is an important source of carbon to seep environments, and may play an important role in the ecology of these environments. The assertion of DOC feeding “methane” seeps is in direct contrast to the commonly accepted notion that only sedimentation and methane provide carbon to seep environments.
4.3. DOC Feeds Heterotrophy in Seep Sediments
 Here we consider evidence for the importance of DOC-based heterotrophy in the seep environment. The geochemical evidence presented above indicates an important role for DOC in cold seep sediments, particularly high-flux seeps overlain by microbial mats. Carbon accumulation by heterotrophic prokaryotes and adsorption of DOC onto sediment particles, possibly the interlayers of smectite clays [Kennedy et al., 2002; Sposito et al., 1999], are the most likely mechanisms by which DOC would accumulate in surface sediments. Here we focus on microbial activity, and consider two non-exclusive mechanisms by which prokaryotes convert DOC to TOC. The first mechanism is through the net mineralization of DOC. In this case heterotrophic prokaryotes catabolize DOC, either oxidizing or fermenting the DOC to more simple products. A portion of the energy conserved by the organism is then used to convert DOC into biomass. This mechanism presumably applies to the fermentative and sulfate-reducing bacteria abundant in MM settings [Knittel et al., 2003]. A second mechanism involves DOC uptake by methane-oxidizing consortia. Though these consortia appear to use only methane and sulfate for catabolism [Nauhaus et al., 2002, 2005], it is possible they may fix DOC for biomass, behavior which has been observed previously in chemoautotrophic methanogens [e.g., Fuchs et al., 1978]. Differential uptake of DOC by methane-oxidizing consortia could explain the observed variability in δ13C of the consortia [Orphan et al., 2001b, 2002].
 A first constraint on the potential importance of DOC is to quantify the flux through the surface sediment. To our knowledge, this flux has not previously been measured or calculated for MM sites. We observed an average pore water DOC concentration of 12 mM. Assuming a fluid flux of 100 cm yr−1 [Tryon and Brown, 2001; Tryon et al., 2002], we calculate a purely advective DOC flux of 12 moles C m−2 yr−1. This flux is an order of magnitude greater than the downward POC flux for other ocean margin environments at the depth of Hydrate Ridge [Boetius and Suess, 2004; Sommer et al., 2002], and about half the diffusive sulfide flux calculated by Sahling et al.  for MM sites. The magnitude of the estimated DOC flux is sufficient to support an active microbial community. This flux is also sufficient to drive carbon accumulation in the sediments on geologic timescales. For example, if DOC were converted anaerobically to microbial biomass with 1.0% net efficiency, the result would be accumulation of 1.4 g C m−2 yr−1. At this rate of accumulation, several thousand years would be required to accumulate the levels of carbon observed at site 3812-3.
 Volatile organic acids constitute about 1% of the calculated DOC flux at site 3812-3. These compounds are known to be key substrates for sulfate reduction in marine sediments, and likely contribute to heterotrophic metabolism in the seep environment. Assuming a fluid flux of 100 cm yr−1 we estimate a flux of 120 mmoles C m−2 yr−1 in the form of VOAs. This advective VOA flux alone could drive sulfate reduction in the MM sediments at a rate of 60 mmoles SO42− m−2 yr−1, assuming a VOA: SO42− reaction stoichiometry of 2:1. The chemical composition of the remaining 99% of the DOC at site 3812-3 is not known. Much of the DOC is presumed to be more recalcitrant than the VOAs, though there are likely significant amounts of labile carbohydrates and functionalized organic acids. The labile fractions are presumably available to the subsurface microbial community, while the more refractory components may penetrate into the aerobic microbial mats and possibly into the bottom water. The impact of seepage on marine DOC has not been determined, but could play an important role through the introduction of 14C-depleted DOC into the deep ocean [Bauer and Druffel, 1998; Wang et al., 2001].
 Evidence supporting an important role for heterotrophy in MM sediments can also be inferred from previous studies conducted at Hydrate Ridge. Here we consider three cases. First, Knittel et al.  analyzed the microbial community structure in MM sediments using a combination of 16SrDNA sequencing and fluorescence in situ hybridization (FISH) microscopy. They found large numbers of apparently heterotrophic bacteria, especially sulfate-reducing bacteria, in MM sediments. We suggest the advective DOC flux through the sediments supplies carbon substrates to this heterotrophic microbial community. Second, Treude et al.  quantified the rates of sulfate reduction and methane oxidation in CF, MM and reference sites at Hydrate Ridge. They observed a mismatch between these rates at MM sites, with methane oxidation rates accounting for only about 25% of the sulfate reduction rates. While some of the mismatch may arise from environmental variability, we suggest that sulfate reduction coupled to oxidation of DOC may be responsible for much of the observed mismatch. Last, several investigators have analyzed the δ13C of lipid biomarkers from seep environments, including those from Hydrate Ridge [Elvert et al., 2003, 2001, 2002; Hinrichs et al., 1999, 2000; Orphan et al., 2001a; Pancost et al., 2001, 2000; Thiel et al., 2001a, 2001b, 1999]. In general the δ13C of lipid biomarkers spans the range from typical marine and terrestrial organic material to more depleted than −100‰. Bacterial biomarkers in particular often have a δ13C between −30 and −70‰. We suggest that lipid δ13C values in this range partially reflect the uptake and utilization of DOC by heterotrophic bacteria.
 The ultimate source of DOC observed in the present study is not clear. We assume most of the observed DOC is formed at great depth in the sediment during diagenesis and migrates upward with the pore fluids. The DOC is presumed to be somewhat recalcitrant; otherwise it would be utilized to feed methanogenesis in the deep subsurface. However, the DOC must be somewhat labile; otherwise it would not be available to sulfate-reducing bacteria. The δ13C of the DOC is depleted slightly relative to most marine and terrestrial organic carbon. We interpret this to indicate the deeply sourced DOC is mixing or exchanging with isotopically depleted carbon in the surface sediments, though we cannot exclude the possibility that deeply sourced DOC is already depleted in 13C.
5. Concluding Remarks
 Results presented here are consistent with previous observations of distinctive hydrogeologic regimes in clam field and microbial mat environments. The accumulation of carbon in anaerobic environments is a slow process, and our results point to long-term stability of and distinction between microbial mat and clam field environments. Some individual seeps may be stable for hundreds to thousands of years.
 The depletion of nitrogen in sediments underlying microbial mats is interpreted as a signature of enhanced mineralization associated with fluid flux. That is, the mineralization of organic material appears to be stimulated in areas of sustained seepage. We favor the concept that carbon and nitrogen are mineralized in stoichiometric proportion to their source, though we are unable to rule out the selective mineralization of nitrogenous compounds.
 Geochemical and microbiological evidence indicates that seep communities atop Hydrate Ridge harbor a significant heterotrophic component, distinct from anaerobic methane oxidizers. These studies indicate an import role for allochthonous dissolved organic carbon in driving heterotrophy in seep environments typically considered to be driven by methane.
 The authors thank all members of the scientific party from cruise 7–18 of the R/V Atlantis, the captain and crew of the R/V Atlantis, as well as the crew and pilots of the DSV Alvin. We also acknowledge Gretchen Robertson, super-tech, who performed several of the chemical analyses presented, Wiebke Ziebis, whose unpublished observations aided us in developing our sampling scheme, Tina Truede who provided data from her investigations, Stefan Sommer who provided the map shown in Figure 1, Craig Carlson for assistance with DOC measurements, Mike Tryon for useful insights into fluid flow, Bill Reeburgh for providing us an HPLC unit, and Mark Altabet for assistance with δ13C and δ15N measurements. The associate editor and two reviewers provided valuable comments. This work was funded by the National Science Foundation (NSF OCE 0096475 to M. K. and D. B., and NSF LExEn-0085607/BioGeo-0311894 to D. V.).