Technical Comments

Comment on “A Persistent Oxygen Anomaly Reveals the Fate of Spilled Methane in the Deep Gulf of Mexico”

See allHide authors and affiliations

Science  27 May 2011:
Vol. 332, Issue 6033, pp. 1033
DOI: 10.1126/science.1203307


Kessler et al. (Reports, 21 January 2011, p. 312) reported that methane released from the 2010 Deepwater Horizon blowout, approximately 40% of the total hydrocarbon discharge, was consumed quantitatively by methanotrophic bacteria in Gulf of Mexico deep waters over a 4-month period. We find the evidence explicitly linking observed oxygen anomalies to methane consumption ambiguous and extension of these observations to hydrate-derived methane climate forcing premature.

Kessler et al. (1) reported that methanotrophic bacteria in Gulf of Mexico deep waters consumed 1 × 1010 moles of methane released from the Macondo oil well between June and September 2010 after the Deepwater Horizon blowout. They linked observed oxygen depletion to methane oxidation based on mass balance considerations, molecular biological data, and modeling. However, we find the methane oxidation hypothesis unconvincing because of large uncertainties in hydrocarbon inputs, challenges in linking oxygen anomalies to methane consumption, problematic interpretation of 16S ribosomal RNA (rRNA) gene libraries, and shortcomings of the one-dimensional (1D) model used to estimate the temporal dynamics of methane oxidation.

The fraction of oil in deep water is poorly constrained (2), and complete oxidation of recent estimates of Macondo hydrocarbon discharge (3) would support up to a 10-fold higher oxygen demand (7 × 1010 to 12 × 1010 moles), assuming the same fraction of oil in the intrusion layer as in (1). Moreover, relief-well drilling and response efforts introduced methanol to Gulf of Mexico deep waters, further confounding the carbon budget and precluding mass balance of oxygen and methane budgets.

Kessler et al. (1) estimated the total oxygen anomaly by 2D areal integration of deepwater oxygen minima off the Texas/Louisiana shelf, west of the wellhead. Although the 2D distribution suggests continuity, the 3D oxygen distribution is dominated by spatially localized features (Fig. 1) and does not suggest a well-mixed, westward-drifting plume. Thick, midwater oxygen depletion anomalies are located south and shelfward of the Mississippi river outflow canyons. Oxygen depletion hot spots also are scattered at depth near sites with prolific natural hydrocarbon seepage [GC600 (4); Fig. 1]. These observations indicate that the oxygen anomalies noted in (1) could result from mineralization of carbon (nepheloid layers) derived from the Mississippi River outflow or hydrocarbons from natural seepage, both of which are known sources of deepwater oxygen depletion in the Gulf of Mexico (5). Notably, the presence of dispersants in oxygen-depleted waters (6) is insufficient evidence of the Macondo plume because dispersants were applied widely at the seafloor and sea surface and are thus likely present throughout northern Gulf of Mexico waters. The location of the sampling stations primarily southwest of the wellhead (1) excludes areas of documented oxygen consumption via biodegradation of Macondo-derived saturates and n-alkanes northeast of the wellhead (7). The limited spatial coverage of the data in (1) thus implies that the oxidant budget is constrained poorly, underscoring the difficulty in definitive attribution of large-scale oxygen deficits solely to methane oxidation.

Fig. 1

Three-dimensional oxygen anomaly southwest of the Macondo wellhead. Colored vertical lines denote the oxygen anomaly (mg/L) between 700- and 1300-m water depth. Anomalies were computed as the difference between observed O2 concentrations and background values, which are determined using polynomial fitting excluding depth regions visually identified as oxygen deficient. Gray-scale circles denote depth-integrated oxygen anomalies ranging from 0 (black) to approximately 3 mol m−2 (white). The magenta star indicates the location of the wellhead, and the magenta square represents the location of the natural seep site GC600. Data used are from Research Vessel Pisces Cruise IV (16).

Direct rate measurements by Valentine and co-workers from June 2010 (8) showed low methane oxidation rates. No other time-series rate measurements were presented in (1) to support the hypothesis of complete methane oxidation. Other biogeochemical data that could have provided substantial support for the hypothesis—such as O2 anomalies associated with strong δ13C enrichment of the residual methane, notable nutrient or trace metal drawdown, increased microbial biomass, and/or a substantial δ13C and ∆14C signature in dissolved inorganic carbon and microbial biomass—were not described.

Kessler et al. offered bacterial 16S rRNA gene clone libraries [figure 2 in (1)] constructed from sites with strong oxygen anomalies relative to control sites with no oxygen or fluorescence anomalies to support their hypothesis. However, these sequence data contain little or no evidence of a significant methanotrophic bacterial bloom, because samples from the control stations (samples 203, 242, and 191) and oxygen anomaly stations (samples 192, 211, 222, and 230) show no significant difference in the relative abundance of methanotrophs (e.g., Methylococcaceae). Although two of the four clone libraries from oxygen anomaly stations showed higher percentages (>20%) of aerobic methylotrophs (samples 222 and 230) than the controls, the other two libraries (samples 192 and 211) were similar to the controls (samples 203, 242, and 191; methylotrophic clones <10%).

These methylotrophs were members of the genus Methylophaga and the family Methylophilaceae, which are known to use methanol and/or methylated amines as carbon substrates and electron donors, but not methane (9). Methylophaga showed the largest increase in relative abundance (samples 222 and 230). Although stable isotope probing experiments have documented methane-derived carbon in the biomass of both Methylophaga and Methylophilaceae (1012), this has been attributed to cross-feeding (10, 11). Defining these microorganisms as methanotrophs, as suggested by (12), requires laboratory verification using pure cultures. Alternatively, the increased relative abundance of Methylophaga may have resulted from their degradation of methylated compounds derived from cooperative microbial breakdown of high molecular weight organic matter or from methanol introduced during response efforts.

Kessler et al. (1) also presented a 1D model that predicts methane oxidation and transport in a plume to support their hypothesis of complete methane oxidation by methanotrophic bacteria after the Deepwater Horizon event. Although qualitatively useful to fill in observational gaps, neglecting important factors that affect methane transport and biodegradation—such as bubble hydrate skins, sudden hydrate flake formation in the intrusion layer (3), or nutrient limitation, and the possibility that the original oil-gas mixture may have separated (so that methane no longer coregisters with fluorescence, which was used to track oxygen anomalies)—requires rigorous model validation. The poor fit of their model to the average oxygen anomaly data [figure 3 in (1)] and the absence of other corroborating data severely limits the utility of these simulations.

Extension of Kessler et al.’s (1) conclusions to natural massive hydrate destabilization events is problematic on physical, chemical, and microbiological grounds. The giant solitary jet of oil-saturated gas at the Macondo site differs dramatically from hydrate-depth seepage plumes (13) and experimental deep sea hydrocarbon releases (14), which featured weaker intrusions near the thermocline (13, 14). Although it remains largely unknown whether methanotrophic microbial communities are efficient biofilters in hydrate dissociation settings, it is notable that δ13C profiles above deep, natural Gulf of Mexico seeps (500 to 1000 m) show considerable methane bypassing the microbial biofilter under low ebullition conditions (13, 14).

Finally, today’s most vulnerable marine methane hydrate deposits underlie shallow Arctic waters (15), where methane is entrapped by submarine permafrost or stabilized as hydrates by year-round cold temperatures. Here, released methane rapidly reaches the atmosphere with minimal microbial oxidation (15). Therefore, extension of Kessler et al.’s conclusions to such contemporary natural hydrate destabilization events would be inappropriate.


View Abstract


Navigate This Article