Methane flux to the atmosphere from the Deepwater Horizon oil disaster



[1] The sea-to-air flux of methane from the blowout at the Deepwater Horizon was measured with substantial spatial and temporal resolution over the course of seven days in June 2010. Air and water concentrations were analyzed continuously from a flowing air line and a continuously flowing seawater equilibrator using cavity ring-down spectrometers (CRDS) and a gas chromatograph with a flame ionization detector (GC-FID). The results indicate a low flux of methane to the atmosphere (0.024 μmol m−2 d−1) with atmospheric and seawater equilibrium mixing ratios averaging 1.86 ppm and 2.85 ppm, respectively within the survey area. The oil leak, which was estimated to contain 30.2% methane by weight, was not a significant source of methane to the atmosphere during this study. Most of the methane emitted from the wellhead was dissolved in the deep ocean.

1. Introduction

[2] Natural seafloor seeps can release methane and oil into the deep ocean and can potentially contribute to the atmospheric methane burden [MacDonald et al., 2002, 2004; Farwell et al., 2009; Leifer et al., 2004, 2006; Kinnaman et al., 2007; Wardlaw et al., 2008; Solomon et al., 2009; Valentine et al., 2010]. A variety of factors such as oil coatings and clathrate hydrate skins on the bubble surface, bubble size, seep strength and release depth all contribute to enhancing the transport of methane directly to the atmosphere [Leifer and MacDonald, 2003; Leifer and Patro, 2002; McGinnis et al., 2006; Rehder et al., 2009; Solomon et al., 2009; Socolofsky and Adams, 2005]. However, a significant component of the methane still dissolves in seawater before rising to the surface and can be available in the deep water for methane oxidation [Valentine et al., 2001].

[3] One of the big questions when looking at natural methane venting from the seafloor is “Could a destabilization of methane hydrate in deep water influence the atmospheric methane budget?” The unfortunate blowout at the BP Deepwater Horizon oil rig on 20 April 2010 released oil and methane at an average rate of 58,000 barrels per day into the deep ocean, until it was capped on 12 July 2010 resulting in a total of ∼4.4–4.9 million barrels released [Crone and Tolstoy, 2010; see also National Incident Command data,]. The natural gas component of the emission was estimated at 30% by weight based on analyzes published by the Department of Energy from 4 June to 16 July 2010 ( With a methane to natural gas ratio of 0.875 [Valentine et al., 2010], this results in 9.14 × 109 moles of methane released over the duration of this event. This devastating event provides a chance to look at the deep water oily seep/methane hydrate blowout scenario, directly. During a quickly designed research expedition, we were able to intensively map the surface saturation state (deviation from equilibrium) and sea-to-air flux of methane over an area extending approximately 28 × 34 km surrounding the wellhead.

2. Methods

[4] To investigate the emission and fate of methane from the Deepwater Horizon oil leak, the Persistent Localized Underwater Methane Emission Study (PLUMES) was conducted aboard the R/V Cape Hatteras. The ship departed Gulfport, MS on 12 June 2010 and returned to Gulfport, MS on 20 June 2010. During this cruise, continuous measurements were made of atmospheric methane and surface water methane concentrations along with a suite of meteorological measurements including wind speed and direction. These measurements were made intermittently with an Agilent 6850 gas chromatograph flame ionization detector system (GC-FID) and continuously using two different cavity ring-down spectrometers (CRDS). The air samples and recirculating equilibrator headspace gas streams each passed through Nafion dryers before analysis. The GC-FID was equipped with a GS GasPro column (0.32mm, 1m pre-column, 14m main column) and a 50 μL sample loop. One CRDS measured just the methane concentrations and water vapor (Picarro, Sunnyvale, CA), while the other CRDS measured methane, water vapor and isotopes of methane (Picarro, Sunnyvale, CA). The calibration of the GC-FID instrument is tied to the NOAA-04 calibration scale (NOAA/ESRL/GMD). The standard has an uncertainty of <1%. The concentration only CRDS offset from the GC-FID (<0.02 ppm) was used to adjust the CRDS to the observed GC-FID levels, and the offset between the GC-FID and the isotope CRDS (<0.36 ppm) was used to adjust the isotope CRDS to the GC-FID level. The concentration only CRDS was used on the air sampling line for most of the cruise, while the isotope CRDS was used in line to sample the recirculating equilibrator headspace for most of the cruise. CRDS is a non-destructive analytical technique, so gases can be analyzed in the equilibrator headspace and returned in a completely closed system; this, and the rapid nature of a CRDS analysis (ca. 10 sec), enables unprecedented data acquisition rates and high spatial resolution mapping of sea-air fluxes. The use of the Weiss type equilibrator [Johnson, 1999; Butler et al., 2007; Bange et al., 1996] and continuous air side measurements allow us to determine the saturation state of the surface ocean with respect to the atmosphere continuously over the surveyed region.

3. Results and Discussion

[5] Over the entire cruise, surface ocean equilibrium concentrations range from 1.87 to 26.67 ppm with the highest concentrations observed during the transits to and from Gulfport closer to or on the shelf (Figures 1a, 1b, and 2a). The atmospheric concentrations range from 1.79 ppm to 2.11 ppm with the highest concentrations to the northwest of the leaking wellhead (Figures 1c and 2a). It is clear that over the region surrounding the wellhead and extending ∼14 to 17 km from the wellhead, the concentrations observed in the atmosphere are not significantly elevated averaging 1.86 ppm, and the concentrations in the surface water are only slightly elevated averaging 2.85 ppm resulting in supersaturations with saturation anomalies ((pw − pa)/pa X 100%) averaging 53% (Figure 2b) and a small mean net flux to the atmosphere of 0.024 μmol m−2 d−1 over the detailed survey region (Figures 1d and 2).

Figure 1.

Distributions of (a) equilibrium methane concentrations in surface water for the entire cruise, (b) equilibrium methane in surface water for the survey region, (c) methane concentrations in the air for the survey region, and (d) sea-to-air fluxes of methane in the survey region. The Deepwater Horizon wellhead is indicated by a black star, and the inverted triangles in Figures 1a and 1b represent data collected in July 2009. Note that 1 minute of latitude equals 1.8 km, and at this latitude 1 minute of longitude = 1.6 km.

Figure 2.

Time series of (a) methane in air (black square) and surface water (red triangle), (b) methane saturation anomalies, (c) wind speed and (d) sea-to-air methane fluxes. The vertical gray lines indicate the beginning and end if the time spent in the survey region.

[6] The equilibrium surface water concentrations measured in the survey region around the wellhead are consistent with other measurements of methane in surface waters (Table 1). During a cruise going out of Gulfport, MS in July 2009, we measured methane in the surface water and air using the same equilibrator and GC-FID system used for the current study (S. A. Yvon-Lewis, unpublished data, 2009). Results of measurements made during that cruise in close proximity to the transits for the current study show concentrations similar to those high concentrations observed along the transits between Gulfport and the survey region near the wellhead (Figure 1a). The high concentrations observed over the slope and shelf are not unique to the Deepwater Horizon oil leak and are likely from a different near-shore source.

Table 1. Surface Water Methane Concentrations From Other Sites in the Northern Gulf of Mexico for Comparison
Study RegionSurface Concentration (nmol L−1)Reference
Survey Region3.3This study
Mississippi/Alabama Shelf22Brooks [1975]
Northern Gulf of Mexico2.4Kelley and Jeffery [2002]
Northern Gulf of Mexico9.9–343Kelley [2003]
Northern Gulf of Mexico0.8–1609Solomon et al. [2009]

[7] The sea-to-air flux (mol m−2 d−1) of any dissolved gas is a function of the concentration gradient between the surface ocean and atmosphere and the wind speed:

equation image

where KW is the wind speed dependant gas exchange coefficient (m d−1) calculated from Sweeney et al. [2007]; Cw is the seawater dissolved concentration (mol m−3); pa is the atmospheric partial pressure (atm); and H is the solubility (m3 atm mol−1) calculated from Wiesenburg and Guinasso [1979].

[8] During this cruise the wind speeds were exceptionally low with 10m wind speeds ranging from 0.13 m s−1 to 1.03 m s−1 over the survey area (Figure 2c). The presence of a significant slick or surfactant layer should also significantly reduce the flux between the ocean and the atmosphere [Asher and Pankow, 1986; Frew et al., 1990; Asher, 1997; Frew, 1997; Frew et al., 2004; Wanninkhof et al., 2009]. The calculated mean net flux of 0.022 μmol m−2 d−1 for the survey region (Figure 2d) does not include any reduction in the gas exchange coefficient due to the presence of the slick or surfactant layer and is therefore an upper limit.

[9] The integrated mass flux of methane to the atmosphere over the survey region ranges from 0.1 to 250 mol d−1 when applying the lowest and highest fluxes observed in the survey region to the entire 950 km2 survey area. Extrapolating for 83 days, this accounts for 8.3 to 20750 moles (0.14–332 kg) of methane escaping to the atmosphere over the 83 days of active leaking from the wellhead. Even if higher wind speeds (e.g. storms) prevailed over the site, the sea-to-air flux would only be enhanced enough to account for ca. 0.01% of the methane released from the reservoir (Table 2).

Table 2. Calculated Fluxes Over the Survey Region (Not Including Transits) for This Study and Extrapolated to Higher Wind Speeds
10 m Wind Speed (m s−1)Flux (μmol m−2d−1)Flux (mol d−1)83 Day Total Flux (mol)
0.43 (avg. obs.)0.02422.61876

4. Conclusions

[10] As discussed above, the majority of the methane emitted from the wellhead was not reaching the surface and venting to the atmosphere in the survey region. This is not entirely surprising, as prior studies have shown that methane released near the bottom in deep water rarely reaches the atmosphere directly but instead remains in the deep water [Johansen, 2000; Daling et al., 2003; Johansen, 2003; Chen and Yapa, 2004; Simecek-Beatty and Lehr, 2007; Dasanayaka and Yapa, 2009]. The methane appears to stay dissolved and suspended as intrusion layers resulting from entrainment of seawater into the oil and gas buoyant jet as described by Socolofsky and Adams [2005].

[11] Depth profiles of C1-C3 hydrocarbons measured during this study (using the same GC-FID described above and by Valentine et al. [2010]) show large concentrations of dissolved methane, ethane and propane at approximately 1050m with a few very high concentrations seen at ∼850m in some profiles (Figure 3). While the concentrations of methane in the water column above these depths and below the mixed layer are higher than background, the majority of the methane emitted appears to be dissolved or suspended in the deep water at ∼1050m. Empirical relationships derived from the experiments of Socolofsky and Adams [2005] are consistent with intrusions at these depths for the scales of the Deepwater Horizon event. The presence of this high C1-C3 concentration layer and its biogeochemical cycling as well as impact on dissolved oxygen is described elsewhere [Adcroft et al., 2010; Valentine et al., 2010; J. D. Kessler et al., Death of a methane intrusion, submitted to Science, 2010].

Figure 3.

Selected depth profiles of methane (top) collected from the stations (yellow labels) shown on the map (bottom) of the survey region. The black star indicates the location of the wellhead. Note that 1 minute of latitude equals 1.8 km, and at this latitude 1 minute of longitude = 1.6 km.


[12] This research was supported by National Science Foundation (NSF) grants OCE-1042650 and OCE-0849246. We wish to thank the captain and crew of the R/V Cape Hatteras for all of their hard work in helping us field this program so quickly and under such unusual circumstances. We also thank Fennix Garcia Tigreros, Eric Chan, Mengran Du, and David L. Valentine for sample collection and support at sea.