SEDIMENT COMMUNITY OXYGEN CONSUMPTION

IN THE DEEP GULF OF MEXICO

by

Gilbert T. Rowe1, 2, John Morse2, Clifton Nunnally2 and Gregory Boland3

1Department of Marine Biology, Texas A&M University, Galveston, TX.

2Department of Oceanography, Texas A&M University, College Station, TX

3Minerals Management Service, U.S. Dept. of the Interior, New Orleans, LA

ABSTRACT

Sediment community oxygen consumption (SCOC) has been measured from the continental shelf out to the Sigsbee Abyssal Plain in the NE Gulf of Mexico (GoM). SCOC rates on the continental shelf were an order of magnitude higher than those on the adjacent continental slope (450 to 2750 m depth) and two orders of magnitude higher than those on the abyssal plain at depths of 3.4 to 3.65 km. Oxygen penetration depth into the sediment was inversely correlated with SCOC measured within incubation chambers, but rates of SCOC calculated from either the gradient of the [O2] profiles or the total oxygen penetration depth were generally lower than those derived from chamber incubations. SCOC rates seaward of the continental shelf were lower than at equivalent depths on most continental margins where similar studies have been conducted, and this is presumed to be related to the relatively low rates of pelagic production in the GoM. The SCOC, however, was considerably higher than the input of organic detritus from the surface-water plankton estimated from surface-water pigment concentrations, suggesting that a significant fraction of the organic matter nourishing the deep GoM biota is imported laterally downslope from the continental margin.

Introduction

The measurement of sediment community oxygen consumption (SCOC) has become a standard approach for estimating the carbon and energy requirements of sediment-dwelling organisms within aquatic sediments. It has been estimated on the continental margins and abyssal plains of the western North Atlantic (Smith 1978; Smith and Hinge 1983), the eastern Pacific(Archer and Devol 1992; Smith 1992; and Jahnke and Jackson 1991); the sub-Arctic Atlantic (Glud et al. 1998); the Arabian Sea (Witte and Pfannkuche 2000); the eastern Mediterranean (Buhring et al. 2006); and the deep south Atlantic (Wenzhofer and Glud 2002). It is generally presumed that SCOC at shallow depths is a direct estimate of the coupling between benthic and pelagic processes (Rowe et al. 1975; Grebmeier and Barry 1991; Graf 1992). Oxygen utilization by deep-ocean sediments has been equated stoichiometrically to the net input of organic matter delivered to the sea floor (Jahnke 1996, Smith et al. 2001). In the Gulf of Mexico, numerous SCOC measurements have been made in the area of seasonal hypoxia associated with the Mississippi River effluent on the continental shelf (Miller-Way et al. 1994; Morse and Rowe 1999, Rowe et al. 2002, Rowe and Chapman 2002). Calcium carbonate shell deposit dissolution rates were shown to be directly proportional to SCOC on the upper continental slope of the northern GoM(Powell et al. 2002). To date, SCOC has been reported at two locations on the Sigsbee Abyssal Plain, one just north of the Yucatan Strait (Hinga et al. 1979) and the other in the west central gulf (Rowe et al. 2003).

The present investigation was designed to complement comprehensive sampling of the biota associated with the benthic boundary layer. Our goal was to determine if community structure, in terms of biomass, diversity and species composition, was related to community function, or rates of biogeochemical processes. Most investigations of biological stocks and processes in surficial sediments focus either on community structure (e. g., Sibuet et al. 1984) or function (Devol and Christensen 1993, Sayles et al. 1994, Wenzhofer and Glud 2002), not both. Previous investigations of this nature have been limited to single sites (Smith 1992, Rowe et al. 2003, Smith et al. 2006) or at best a single transect (Smith 1978). The complete set of investigations in this special issue (“Deep Gulf of Mexico Benthos” or DGoMB) has attempted to address both for the northern GoM.

This five-year study (DGoMB) was conducted in stages, with the initial phase in 2000 being a survey of benthic community structure of the continental slope of the northern GoM from the Texas/Mexico border to northern Florida. The second phase, in 2001, initated “process” measurements at five contrasting locations selected on the basis of the community structure information available from the survey. The selection was based on extremes in standing stocks or densities of the sediment-dwelling bacteria, meiofauna and macrofauna. Fortunately, the stocks all displayed extremes at the same sites. The original five sites chosen (Fig 1) were MT1, MT3 and MT6, for maximum and minimum standing stocks at upper slope and lower slope depths associated with the Mississippi Canyon in the central GoM; and S42 and S36 for minimum and maximum densities at comparable and intermediate depths in carbonate-rich sediments associated with the Florida continental margin. The following year (2002), two additional sites (S1 and S4) were added in order to extend the sampling out onto the Sigsbee Abyssal Plain (Fig 1), a logical extension to encompass the gulf’s greatest depths.

Methods

SCOC in this study was measured using both in situ and ship-board laboratory incubations of surficial sediments and bottom water within chambers, as described previously(Tengberg et al. 1995, 2005), with modifications indicated below. A free-vehicle bottom lander was utilized to the carry incubation chambers to the sea floor in order to measure SCOC without decompression or alteration of temperature. The “GOMEX” lander, as we call it (Rowe et al. 1994), is composed of an aluminum frame, glass floatation spheres, disposable anchor weights released by an electronic timer, a B&W film camera and strobe, a strobe and radio direction finder to assist recovery, and two incubation chambers. The two chambers themselves are constructed of plexiglass cylinders covering an area of 900 cm2 each, with the top sealed by a flat piece of lexan. They are held above the bottom of the lander on deployment, but then are dropped to the sediments by a timer approximately 30 minutes after the lander has settled to the sea floor. The chambers slide down steel runners with hydraulic dampers to slow their sinking and minimize disturbance. One-way flapper valves in the lexan top release water as the chambers descend into the sediments. Once on the bottom, each chamber contains 7 liters of sea water when fully engaged with the surface sediment. A two-cm rim around the outside of each chamber prevents deeper penetration of the cylinders into the sediments. The thickness of the diffusive boundary layer within incubation chambers is controlled by circulation rates. Circulation in the chambers appears to be adequate to make reasonably accurate oxygen flux measurements with the stirring motors and pumping system utilized (Rowe et al. 1994; Tengberg 2005). Oxygen concentration is monitored with polarographic electrodes, with the output recorded on a data logger (produced by Sea Bird for use in CTDs). The camera is positioned to take photographs of the chambers’ contact with the sea floor. Spring-loaded 50 cc syringe samples are taken at the beginning, middle and end of each incubation in order to estimate nutrient fluxes between the water and the sediments. All in situ sea-floor mechanical operations are controlled by an electronic timer and burn-wire release system.

Small plexiglass chambers measuring 25 cm in diameter were utilized aboard ship to make parallel incubations in the laboratory (Miller-Way et al. 1994). These samples were procured by inserting the chambers into the sediments and water contained in 0.2 m2 box cores (Boland and Rowe 1991). The area of sediment covered is 125 cm2 and sea water volume above the interface varied from 0.8 to 1.3 liters. A polarographic oxygen electrode (YSI) is screwed into the sealed top of each chamber and a small stirring magnet is suspended below the electrode membrane. The mixing appears to be adequate for reasonably accurate flux measurements to be made (Tengberg et al. 2005). These chambers were maintained in the ship’s laboratory in the dark at in situ temperature, during which oxygen concentration was measured continuously over periods of 6 to 36 hours. In general two to three replicate, ship-board incubations were made at each site. Following each incubation, the sediments were sieved using a 300 micrometer mesh sieve to separate out the macrofauna for comparison with standard faunal sampling that had already been conducted at each site.

The flux of oxygen into the sediments in the incubations is calculated using the following formula:

Flux = [Change in concentration] x [Volume of incubation chamber]

[Area covered by chamber] x [Time]

The oxygen flux values (SCOC) are reported (Table 1) as mg carbon in carbon dioxide remineralized by respiration, wherein the flux of oxygen is multiplied by a Respiratory Quotient of 0.85. An RQ of 0.85 assumes that the organic matter consumed is proportioned equally between lipid, protein and carbohydrate.

Oxygen concentration profiles within the sediments were produced at six process stations (MT3, MT6, S36, S42, S1, and S4) using a microelectrode, according to the method described by Brendel and Luther (1995). Measurements were made at depth intervals of 2 mm. Concentrations in the porewaters of the top 10 to 15 cm of sediment, and 1 cm of overlying bottom water, were made by cathodic-stripping voltammetry, using solid state amperometric microelectrodes and an Analytical Instrument Systems model DLK-100 voltammetric analyzer (Brendel, 1995; Brendel and Luther, 1995; Luther et al., 1998). Measurements were made with an Au/Hg amalgam glass microelectrode. Instrumental parameters for the linear sweep and cyclic voltammetry modes were typically 200 mV s-1 scan rate over the potential range -0.1 to -1.8 V with a 10 s deposition at –0.1 V. Minimum detection limits for O2 were approximately 5 M. Calibration of each electrode was based on the pilot ion method where Mn2+ was the standardized ion (Brendel, 1995).

Sediment oxygen consumption within the sediment was estimated from the oxygen gradient within the sediments and from the oxygen penetration depth at each site, utilizing Fick’s first law, assuming molecular diffusion-limited rates in the sediments (Berner 1980):

Js = -Dsdc/dz

where Js is the flux in moles per unit area over a layer of surface sediment per unit time,  is porosity, Ds is the whole sediment diffusion coefficient and dc/dz is the concentration gradient. The simplified approach, however, of Cai and Sayles (1996) was also used to calculate consumption of oxygen based oxygen penetration depth (OPD), rather than the gradient. In doing so, we used the temperature-dependent O2 diffusion coefficient in water (Jahnke et al. 1987) corrected for tortuosity, according to Berner (1980), yielding a value of Ds = 3.08E-6 cm2 s-1 at 4oC. An average porosity of 0.8 was assumed for each site. The flux was calculated from:

F = 2  Ds [O2]BW/L

where F is the oxygen flux, L is the OPD and [O2]BW is the bottom water concentration.

Results

SCOC was measured at seven sites (Fig 1, Table 1), from depths of 460 m in the Mississippi Canyon (MT1) down to 3,650 m on the Sigsbee Abyssal Plain (Fig 1). The three deep sites (LD97, S1, S4) located on the abyssal plain have been averaged as a single value (Table 1). The variability in the measurements at any given site was high, with the Standard Deviation often half the mean, reflecting meager precision in the methods or small-scale variation on the sea floor. The mean rates at each site decreased from a range of 32 to 37 mg C m-2-d-1 on the upper continental slope down an order of magnitude to the low mean value of 3.9 mg C m-2d-1 on the Sigsbee Abyssal Plain.

Oxygen penetration depth (OPD)and concentration profiles for stations MT3, MT6, S36 and S42, are illustrated in Figure 2. At station MT3, the OPD was close to 0 mm. The S36 and S42 stations displayed rather similar OPD of 45 and 40 mm, respectively. The profiles were parallel at depth, but the shallower site (S42) had a lower initial value, reflecting the oxygen concentration in the water column. The deep station on the central Sigsbee Abyssal Plain, S1, had the deepest OPD (100 mm), as might be expected, but at S4, just north of the Yucatan Strait, it was only 30 mm (Table 2).

Dissolved iron and sulfide were not detectable at any of the sites within the sediment depths sampled. Dissolved Mn2+ was observed only at MT3 (Fig 3), where it exhibited a classic (e.g., Berner 1980) broad subsurface maximum from about 30 to 80 mm, which averaged about 210 M. This is the zone of major manganese reduction. Above this zone diffusive transport of manganese is the primary process and below this zone manganese is precipitated primarily as the carbonate mineral pseudokuntnahorite (MnCa(CO3)2).

Sulfate reduction rates were below detection limits. The deep oxygen penetration depths at these sites, with the exception of MT1 and MT3, indicate no sulfate reduction is likely to occur, at least to the sediment depths sampled. It could however be occurring at deeper depths than those sampled (~20 cm). This was clearly evident in the earlier work of Lin and Morse (1991), where reduced sulfur species were not observed in sediments for water depths greater than about 200 m, until several meters of sediment were penetrated.

Discussion

Comparison of Methods:

The SCOC values measured with incubation chambers (Table 1)were considerably higher than the estimates from the gradient of oxygen into the sediments and the OPD (Table 2), with the exception of MT3 and S4. The estimate at MT3, however, was not based on a profile or penetration, since there was none. As we cannot divide by zero, a penetration of 1 mm was assumed. At S4, the two estimates were very close. This set of relationships is similar to other efforts to compare the two approaches (Wenzhofer and Glud 2002): the chamber fluxes, referred to in previous work as “total oxygen utilization” or TOU, are generally larger than the “diffusive oxygen utilization” or DOU. In the GoM, with the exception of MT3, the DOU rates were lower that the TOU rates. We interpret this to indicate that considerable activity occurs at the sediment – water interface involving motile invertebrates that consume reactive particulate organics as soon as it reaches the sea floor. The subsurface diffusive flux of oxygen (DOU) down into the sediments on the other hand is driven by bacterial activity that is limited by the particulate organic matter that is mixed downward by bioturbation, the diffusion of dissolved organics downward following its remobilization from particulates, and chemolithotrophic bacterial oxidation of reduced metabolic end-products diffusing up toward the sediment – water interface. The SCOC, which is referred to as the TOU, is the sum of these two processes, one at the surface dominated by the metazoans and the other at depth dominated by heterotrophic bacteria and possibly protists. The low rates of bacterial organic carbon utilization measured in these same sediments (Deming and Carpenter, this volume) may reflect this dichotomy. That is, the oxygen profiles and OPD may be more closely related to measured bacterial activity than is the SCOC (TOU) measured by incubation chambers, in which the rates are dominated by the biological processes very near or on the interface.

Patterns within the Gulf of Mexico:

All the SCOC rates presently available for the northern GoM continental margin have been plotted as a log-normal function of depth in order to infer what controls SCOC in the GoM (Figure 4). The rates on the continental slope and abyssal plain were substantially lower than those on the adjacent continental shelf just inshore of MT1 (Figure 4). Mean values dropped by about one order of magnitude for each km increase in depth. The wide range in values on the continental shelf reflects oxygen limitation during hypoxic periods and seasonal variation in temperature (ca. 20o C in winter and just under 30o C in summer, Rowe et al. 2002). The seasonal controls, as far as we know, are limited to the continental shelf. This comparison of the shelf and slope illustrates that the upper slope rates, although high compared to abyssal plain rates, are low compared to the continental shelf, a relationship which is typical of all continental margins. The pattern parallels that of benthic macrofaunal biomass (Figure 5).

Comparison with Other Ocean Margins:

The pattern of SCOC relative to depth in the GoM (Fig. 4) is similar to that described along other ocean basin margins, and reflects, we believe, the variation in offshore deposition patterns of organic matter. The ‘flat’ part of the regresson on the upper slope (0.5 to 2.0 km) is believed to correspond to a depocenter in which fine particulates exported from the continental shelf are accumulating (Rowe et al. 1994). The rates then decline from that plateau down at least one order of magnitude out on the abyssal plain. The values encountered on the GoM slope SCOC plateau were10 to 50% lower than SCOC rates at equivalent depths at suspected depocenters on the western margin of the Atlantic (Anderson et al. 1994; Rowe et al. 1994), as well as rates measured on the eastern boundary of the Atlantic (Wenzhofer and Glud 2002) and Pacific (Archer and Devol 1992) just below the “oxygen minimum zone” (OMZ) characteristic of coastal upwelling ecosystems. Likewise, the rates on the GoM Sigsbee Abyssal Plain at 3.4 to 3.7 km depth appear to be slightly lower than rates at similar depths on other margins (Rowe et al. 1994, Witte and Pfannkuche 2001), although the data are too sparse to confirm this statistically. The GoM SCOC appears to be similar to the eastern Mediterranean (Buhring et al. 2006).

Organic carbon deposition and cycling in the Northern GoM: