METHANE CONSUMPTION IN CARIACO TRENCH WATERS AND SEDIMENTS

Detailed measurements of CH4 in the water column and sediments of the Cariaco Trench show that CH4 is non-conservative in both environments. Concentration differences between the sediments and adjacent overlying water suggest that the sediments are the source of the water column CH4. Co-metabolism of CH4 by sulfate reducers appears to be the CH4 sink in anoxic environments.


Introduction
Methane has been observed In the water columns of a number of anoxlc basins [1][2][3], but no clear conclusions have been reached as to whether it as produced in the water column or whether it IS supplied by the sediments One of the reasons for this uncertainty IS our lnablhty in most Instances to distinguish the effects of mtxlng from reactions in the water column Steadystate vertical advectxon-dlffuston models [4] have been applied in anoxxc basins such as the Black Sea [5] and the Carlaco Trench [6,7] and offer a means of distinguishing the relative effects of mixing and reactions Since some hope of steady-state conditions at the sediment-water interface is possible in these environments, anoxlc basins appear to be reasonable places to attempt combining water column and sediment models Detailed measurements of CH4 In the water column and sediments of the Carlaco Trench are presented and discussed in this paper 2 Previous work The Canaco Trench is a depression in the continental shelf off the coast of Venezuela containing two 1300--1400 m deep basins separated by a 900-m saddle and from the Caribbean by a 150-m sill The waters are anoxlc and contain sulfide below depths of 300 m The environment has been studied extensively for 20 years, Rlchards [8] has summarized and reviewed these studies Methane has been measured in the Carlaco Trench water column on several occasions [1-3] Below 300 m, all of these studies showed approximately linear increases with depth Methane In the sediments has been studied only once During JOIDES leg 15, gas pocket samples were collected at site 147 and analyzed at ten depths ranging from 47 to 175 in [9,10] Steady-state vertical advection-dlffusxon models have been apphed to the Carlaco Trench water column by Broenkow [6] and Fanning and Ptlson [7], who differed in their assumptions regarding vertical transport terms and in their choice of boundaries Broenkow considered an Interval between 300 m and the bottom, set the vertical advectlon velocity equal to zero, and assumed a constant vertical eddy dlffuslvlty of 0 5 cm 2 sec -1 Fanning and Pllson used vertical advectlon in their treatment and considered a hnear T-S region between 300 m and 1000 m They considered the ratio of vertical eddy diffusion (K) to vertical advectlon (w) constant over this interval and assumed w = 0 below 1000 m Silica, sulfide and phosphate were linear functions of potential temperature, Indicating they were conservative, 1 e neither produced nor consumed in the hnear T-S region Fanning and Pllson used silica fluxes to calculate a mlmmal vertical advectlon velocity of 0 75 m yr -I Their treatment, however, did not use least-squares fits and made no distinction between basins Their work indicated that the sediments were the source of silica in this environment The same source was suggested for sulfide Methane, like sulfide, is formed by obligate anaerobes, but controversy surrounds the principal production mechanism in nature, CO2 reduction is regarded as the principal production pathway by some workers [11,12], whtle acetate fermentation is favored by others [1,13] Inhibition of CH4 production by SO4 2-has been suggested [14], but recent inhibition studies [13] and continuous culture work by Cappenberg [15] Indicates commensahsm between sulfate reducers and methane producers The sulfate reducers reqmre lactate and release acetate, which is fermented by the methane producers The commensal relationship is complicated by sulfide inhibition of the methane bacteria at concentratmns above 10 -1° s M Cappenberg's observations are in accord with those of Lawrence and McCarty [16] who observed concurrent sulfate reduction and methane production m a sewage digester Microbial oxidation of CH4 with dissolved oxygen IS reported to occur in lakes at rates up to 1/lmole 1-1 hr -1 [17] An oxidizer other than dissolved oxygen must be responsable for CH4 consumption reactions in anoxlc environments Oxidation by SO42-by the general reaction CH4 + SO42-+ 2 H +-> H2S + CO2 + 2H20,  [24] Samples were stored under refrigeration In glassstoppered bottles with sodium azlde untd analysis at Old Dominion University Precision of the water column measurements is estimated to be 3-4% The data are shown in Fig 1 Methane and total CO2 were determined on board using samples of freshly collected gravity cores No core catcher was used m order to minimize disturbance of the sediment-water interface Interstitial water was separated from samples of the cores [25] and analyzed for gases using techniques previously described by Reeburgh [26] A thermal conductivity detector was used, duplicate measurements were precise to 3 5% Figs 2 and 3 show the depth dtstrabutlons of CH4 and total CO2 in Carlaco Trench sediments We encountered no slgmficant problems with bubble formation and degassmg of the cores on deck as reported previously [9] Bubbles were observed in the VlClmty of the deepest sample of the eastern basra CH4 core (Fig 2) as it was extruded, so we consider this sample unrehable Ambient temperatures often ex- ceeded the maximum use temperature (30°C) of the Aplezon N stopcock grease used on ball joints and stopcocks m the sediment gas analysis SIhcone grease was subsmuted, but non-wetting interior surfaces led to unreliable volume determinaUons for the water column total CO2 measurements When leaks were discovered, a CH4 standard was run under the same flow conditions 4 Discussion

1 Water column
For a stable non-conservative (SNC) parameter, the steady-state verUcal profile has been shown by Craig [4] to be

KC" wC' +J=O
( 1) where K is the vertical eddy diffusion coefficient (cnfl sec -a), w is the vertical advectlve velocity (length yr-1), Jls a reacnon rate term (/~mole 1-1 yr-1), C as the concentratlon (/lmole l-I)and the primes derivatives with respect to depth, which is posxtwe upward Eq 1 has the soluUon for constant J (2) where f(z)= (eZ/z*-1)/(eZm/z*-1), Co is the con- The CH4 profiles (Fig 1) are very nearly linear with depth Two interpretations are possible for a hnear gradient (1) it may represent a pure diffusion case (w = 0) for a conservative species, or (2) it may be the result of non-conservative behavior (J 4= 0) The temperature and sahmty profiles (Fig 5) agree well with fits to the model assuming vertical advectxon and diffusion Dissolved CH4 cannot diffuse or advect independently, so the observed linear profiles can only result from net consumption of CH4 m the water column Conservative profiles are shown in Fig 1 for  comparison The ratio J/w for CH4 may be obtained analytically from the depth profde Craxg [4,27] shows that for a linear profile The Co shown an Fig 1 is the average CH 4 concentrataon below 1200 m The slope, J/w, was obtained by fitting linear least-squares hne through Co at z = 0 The J/w values for CH4 in the eastern and western basins are -7 60 + 0 31 and -5 64 -+ 0 23/1mole 1-1 km -1 Vertical advective velocities may range between 0 0002 and 0 002 km yr -1 [28], resulting in a possible range of CH4 consumption rates of 1 13 X 10 -3 to 1 52 X 10 -2 /amole 1-1 yr-t The CH4 flux from the homogeneous bottom layer through the lower boundary of the mixing interval can be determined using Cra]g and Clarke's [29] eq 2 Modifying this equation for use with absolute volume concentrations rather than saturation anomalies and mass fractions yields = wCo

Kc; = w[Co -z'C;] (4)
where K, w, z* and the prime are identical to those in eq 1 and Co is the CH4 concentration at the lower boundary of the mixing interval The CH4 gradient is linear, so Co = J/w (eq 3) and Fig 2 shows eastern and western basin sediment CH4 profiles that are concave upward, have a low concentratlon CH4 zone between the se&ment-water interface and 35 cm (eastern basin) to 55 cm (western basin) and increase to much higher CH4 concentrations below these depths The inset shows the low CH4 values from the upper 60 cm plotted on an expanded concentration scale Such profiles can only result from CH4 consumptlon in the surface sediments Slmdar profdes have been reported in other marine environments [31,32,14], where they have been interpreted as resulting from CH4 production at depth and oxidation in a biologically or physically mixed surface zone Unlike the other environments, the waters overlying these sediments are permanently anoxlc, so bioturbatlon and oxidation by molecular oxygen are precluded Processes like in situ bubble formation, bubble formation during core retrieval, hydrate formation and physical mixing can be ehminated on physical grounds as causes of the low CH4 surface zone The CH4 must be consumed by anaerobic biochemical reactions, resulting in an effective barrier to transfer of large quantities of CH4 across the sediment-water interface The total COz distributions (Fig 3) show linear gradients and slope changes at depths smallar to those observed for CH4 Presley [33] has measured SO42depth distributions in eastern and western basin sediments and observed linear gradients and near complete reduction of SO42-at similar depths The similarity of the depths at which the CH4, total CO2 and Presley's SO42-profiles change slope, combined with the directions of the nearly linear gradients for each suggest the presence of a diffusioncontrolled CH4 consuming zone located at the depth of the slope changes Methane diffuses upward into the zone and is consumed, SO42-diffuses downward from the overlying water and is reduced, and Increased amounts of CO2 diffuse upward from the zone We have no way of estimating the vertical extent of this zone, but the sample spacing and abrupt changes In the profiles suggest that it is probably not greater than 10 cm thick The CH4 flux to the consuming zone may be calculated by applying the diffusion equation where q5 is the flux 0amoles cm -z yr-I), D s is an interstitial diffusion coefficient (molecular diffusion corrected for tortuosity) and dC[dz is the concentration gradient Using CH4 concentration gradients from Fig 2 and Fanning and Prison's [34] interstitial diffusion coefficient for silica (3 X 10 .6 cm z sec -1 = 94 5 cm 2 yr-1), CH4 fluxes to the consuming zone in eastern and western basin sediments are estimated to be 15 9 and 5 48 pmole cm -2 yr -1 If consumption is uniform in a 10 cm thick zone, this flux corresponds to CH4 consumption rates of 1 59 and 0 55 mmole 1-1 yr-1 for the sediments This rate is sensitive to the choice of the consuming zone thickness and is 106 times greater than CH4 consumption rates in the water column Eq 1 may be used for steady-state sediment profiles by substituting a sediment diffusion coefficient D s for K and a sedimentation rate co for w This gives an equation equivalent to Berner's [35] eq 6-38 Consldering the CH4 gradient below the consuming zone to be linear, KC" = 0 and J = CO C' A sedimentation rate of 0 05 cm yr -I has been reported for the Carlaco Trench [36], maximum CH4 production rates are 8 45 and 2 92 pmole 1-1 yr-1 for eastern and western basin sediments

3 Anaerobw CH oxtdatton
The reaction responsible for anaerobic CH4 consumption is of interest not only because of the apparent conflict in geochemical studies of anaerobic CH4 oxidation [ 19,20] but also because anaerobic organisms capable of using CH 4 as the sole carbon source are unknown [37] Davis and Yarbrough [191 and Sorokln [20] worked with pure cultures of sulfate reducers, but used very different media, the first used a lactate medium and observed oxidation of labeled hydrocarbons supplied as a minor carbon source while the latter used a mineral medium with CH4 as the sole carbon source and observed no reaction Mechalas [38] demonstrated that sulfate reducers may derive energy from the oxidation of recalcitrant molecules by cometabolism, using lsobutanol as an example He further pointed out this group's co-metabolic ablhty significantly broadens the spectrum of organic compounds available to them The geochemical studies [19,20] are consistent with the notion of co-metabolism of CH4, they show that CH4 does not serve as the principal carbon source for sulfate reducers, but that It may be used as a secondary carbon source Since sulfate reducers are active in the Cariaco Trench and 8042-IS the only likely oxidant, co-metabohsm by sulfate reducers must be responsible for the anaerobic consumption of CH4 observed The CH4 sediment profdes are maintained by dlf-fusion of S042and CH4 into a subsurface zone where co-metabohsm of CH4 occurs at a much higher rate than the CH4 production rate The sedimentation rate and orgamc carbon flux to the sediments appear to control the depth of the CH4 consuming 70ne fluxes for CH4 and CO2 estimated front tills study with organic carbon and SO42-fluxes estimated by Presley [33] In the following schematic summary the numbers in parentheses are fluxes in/anlole cm -2 yr -1 The se&ment CH4 and CO2 fluxes are from eastern basin sediments The water column discussion (subsection 4 1) showed clearly that net consumptmn of CH4 takes place in the nnxlng zone of the Carlaco Trench and that the CH4 source for this zone lies below 1200 m The sediment discussion (subsection 4 2) also showed consumption of CH4 in a subsurface consuming zone Mixing in the lower 200 m of the water column cannot be assessed with our present data, but anything other than consumptlon of CH4 here seems unlikely Methane concentratlons in the surface 45 cm of the sediments are between 3-and 10-fold higher than those in the adjacent overlying water, suggesting that the sediments are the water column CH4 source However, a distinct concentration gradmnt is not evident With a hnear concentration gradient, CH 4 concentrations between 60 and 800/~mole 1 -I are needed at depths of 45 cm to maintain the range of water column fluxes These concentrations are easily observed with the sednnent gas measurements The levels to which CH4 may be co-metabohsed by sulfate reducers are unknown at present This knowledge, as well as more detailed and sensitive CH4 measurements an the surface se&ments, is needed to fully understand the transfer of CH 4 across the sediment-water Interface

Summary
A summary of CH4 behavior in Carlaco Trench waters and sediments may be obtained by combining The upward fluxes of CH4 and CO2 below the CH4 consuming 7one are near equal, which is expected for formation of CH4 by acetate fermentation With carbohydrate as substrate, sulfate reduction releases 2 moles of CO2 for each mole of SO42-reduced, if CH4 is oxidized, the Stolchlometry is 1 1 Co-metabolism of CH4 by sulfate reducers accounts for about half of the downward SO42-flux and for 25% of the upward CO2 flux m the surface sediments This CO2 may be an Important source of lSOtOplcally light CO2 to the deep waters of the Canaco Trench [39] It appears that about 40% of the organic carbon flux is preserved in the sediments The upward flux of CH4 tO the water column, although not clearly identified, may be maintamed by 1-10% of the CHa flux to the sediment consumlng zone Consumption reactions in the mixing zone of the water column require 85% of the water column flux This anaerobic metabohsm of CH4 by sulfate reducers suggests that H2S is probably not conservative in the water column Am Geophys Union 55 (1974)