Estimating the oil flow rate from the BP Macondo well
Contents
Estimating the oil flow rate from the BP Macondo well
Editor's Note: This article is excerpted directly from National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling,"The Amount and Fate of Oil," Draft, Staff Working Paper No. 3. It has been edited only to conform to the Encyclopedia's style guidelines.
The federal government?s estimates of the amount of oil flowing into and later remaining in the Gulf of Mexico in the aftermath of the Macondo well explosion were the source of significant controversy, which undermined public confidence in the federal government?s response to the spill (Oil spill). By initially underestimating the amount of oil flow and then, at the end of the summer, appearing to underestimate the amount of oil remaining in the Gulf, the federal government created the impression that it was either not fully competent to handle the spill or not fully candid with the American people about the scope of the problem.
Federal government responders may be correct in stating that low flow-rate estimates did not negatively affect their operations.1 Even if responders are correct, however, loss of the public?s trust during a disaster is not an incidental public relations problem. The absence of trust fuels public fears, and those fears in turn can cause major harm, whether because the public loses confidence in the federal government?s assurances that beaches or seafood are safe, or because the government?s lack of credibility makes it harder to build relationships with state and local officials, as well as community leaders, that are necessary for effective response actions.
This working paper first tells the story of the government?s struggle to accurately estimate the rate of oil flow from the Macondo well. It next discusses the debate surrounding the government?s report on the fate of the oil.2 More extensive, peer-reviewed government reports, which will allow for greater substantive evaluation of government estimates related to flow rate and fate, are forthcoming. In the meantime, this paper discusses some of the key government estimates with a view towards eventual Commission findings regarding whether flow-rate estimates should have been more accurate from the outset, and whether the government presented information regarding the amount and fate of the oil to the public in an appropriate manner. Commission staff believe that recommendations aimed at improving the quality of information provided to the public are critical to improving public confidence, and thus to the success of future emergency responses.
his Part of the paper describes governmental and non-governmental efforts to accurately estimate the flow of oil from the Macondo well. It attempts to frame the questions of why initial government estimates were inaccurate, and whether the information conveyed to the public was different from operational estimates used by responders or other information known to the government. Section A charts flow-rate estimates created during the spill?s first month. Section B discusses the question of which estimates—the low flow-rate estimates, or worst-case estimates—were the basis of government response operations. Section C considers whether the Flow Rate Technical Group, the government team assembled to address the gulf between official and independent flow-rate estimates, addressed the problem of failing public confidence due to inaccurate estimates. Finally, Section D compares the flow-rate estimates of non-governmental scientists generated since the well was capped with the government?s current flow-rate figures.
The First Month
On the evening of April 20, 2010, the U.S. [Guard] District Eight command center in New Orleans, Louisiana received a report of an explosion and fire aboard the mobile offshore drilling unit Deepwater Horizon.3 On the morning of April 22, 2010, the [Guard] informed the media that the rig was leaking oil at a rate of 8,000 barrels per day (bbls/day), and that responders were preparing for a leak of up to 700,000 gallons of diesel fuel (the total amount of fuel the rig could hold).4 Later that morning, Deepwater Horizon sank, leaving a one mile by five mile sheen on the ocean?s surface.5
How much oil was leaking into the Gulf of Mexico? For responders, politicians, and the public, the leaking well?s “flow rate” quickly became a crucial and controversial question. Throughout the first month of the spill, government responders officially adhered to what we now know were low and inaccurate estimates. Non-governmental scientists, on the other hand, used the small amount of publicly available flow data to generate estimates that have proven to be much more accurate. To make forward-looking recommendations, it is important to understand how this came to pass.
The Government’s Estimates
As a first step in determining whether, or how much, oil was flowing from the Macondo well, BP enlisted remotely operated vehicles (ROVs) to investigate the immediate wellhead area. These ROVs did not uncover any leaks.6 Rear Admiral Mary Landry, the Federal On-Scene Coordinator (and ranking federal official on the spill response team at the time), told CBS News on April 23, 2010, that “at this time there is no crude emanating from that wellhead at the ocean floor . . . . there is not oil emanating from the riser either.”7
But at the time of Admiral Landry?s statement, the riser had not yet been inspected. Over the next 24 hours, BP?s ROVs traced the riser from the wellhead to where the Deepwater Horizon rig had come to rest, approximately 1,500 feet from the blowout preventer (BOP). The ROVs discovered two leaks, one from a kink in the riser above the BOP (“kink leak”) and a primary leak from the end of the riser, where it had broken off from the rig.8
After the discovery of these leaks on April 24, 2010, [Guard] and BP officials put out an estimate: Up to 1,000 bbls/day were flowing from the two leaks in the riser.9 Neither the [Guard] nor BP divulged the data or methodology behind this estimate. Based on the information we have to date, it appears the figure came from BP without supporting documentation.10
In the spill?s second week, the official flow-rate estimate increased from 1,000 bbls/day to [bbls/day] as a result of input from the National Oceanic and Atmospheric Administration (NOAA). On April 28, 2010, Admiral Landry stated that “NOAA experts believe the output could be as much as 5,000 barrels.”11 Although Admiral Landry did not provide further explanation, the media speculated that this latest estimate was derived through a method known as the “Bonn Convention.”12 The method involves using aerial data to measure the extent of a spill, using color to estimate the thickness of various parts of the spill, and then calculating the volume.13
Based on the information currently available to Commission staff, the source of the [bbls/day] estimate appears to have been an unsolicited, one-page document emailed to Admiral Landry?s Scientific Support Coordinator on April 26, 2010, by a NOAA scientist.14 The scientist derived an “estimated present volume release rate” of roughly [bbls/day], based on visual observation of the speed at which oil was leaking from the end of the riser.15 While he also used a method based on satellite imagery, similar to the Bonn Convention, to estimate that 10,000 barrels of oil were on the ocean?s surface, he did not base his flow-rate estimate on that surface volume estimate. (He noted, moreover, that estimating surface volume from the visual appearance of an oil slick was “a highly unreliable process.”)16
The NOAA scientist?s [bbls/day] estimate did not take into account the kink leak, and his methodology for estimating the velocity of the leaking oil was imprecise.17 Further, there is no indication that the scientist had expertise in estimating deep-sea flow velocity from video data or that he used an established or peer-reviewed methodology when doing so. This is not a criticism of the scientist, who made clear his assumptions and that the [bbls/day] figure was a “very rough estimate.”18 His stated intent in disseminating the estimate was to warn government officials that the flow rate was multiple times greater than 1,000 bbls/day.19
Despite the acknowledged inaccuracies of the NOAA scientist?s estimate, and despite the existence of other and potentially better methodologies for visually assessing flow rate (discussed below), [bbls/day] was to remain the government?s official flow-rate estimate for a full month, until May 27, 2010.20
Non-governmental Estimates
From the outset, estimates from non-governmental sources were significantly higher than official government estimates. In at least some instances, the cause of the discrepancy appears to be that non-government scientists relied on more refined or better-established methodologies.
Estimates Based on Satellite Imagery
The first independent flow-rate estimate surfaced on April 27, 2010, at the time the official estimate was 1,000 bbls/day. Using publicly available satellite images, John Amos, the founder of SkyTruth.org, estimated the leak size to be at least five times the government estimate—5,000 to 20,000 bbls/day.21 Amos generated the low number in his range by multiplying the surface area of the spill by what he considered the minimum thickness for oil to be visible on the Gulf?s surface (1 micron). He then generated the high number by relying on a BP statement that 3% of the slick was significantly thicker (100 microns). Amos?s estimate was conservative (i.e., low) in assuming that none of the oil had burned with the rig, been collected by response crews, evaporated, dispersed, or was then below the surface.22 Within days, Amos?s estimate appeared in the national press.23
On May 1, 2010, Dr. Ian MacDonald (a Florida State University oceanographer) published a new estimate on SkyTruth.org. Based on a [Guard] map that tracked the spill?s surface size and classified the color of the surface oil throughout, Dr. MacDonald generated a flow estimate of 26,500 bbls/day using the Bonn Convention. Like Amos, he assumed that none of the oil had burned, evaporated, dispersed, been skimmed, or was then below the surface.24
Both independent scientists estimated the spill?s volume from the visual appearance of the surface slick—the same general method used by the NOAA scientist who generated the 10,000 barrel surface volume estimate. Experts note that such methods are not reliable for estimating the volume of large spills, due in part to the difficulty of accurately determining oil thickness from aerial data.25 Dr. MacDonald, however, did at least use an established protocol—the Bonn Convention—for determining surface oil thickness. Similarly, Amos explained the basis for his minimum and maximum assumptions regarding thickness. By contrast, the NOAA
surface volume estimate appears to have been based on an unexplained assumption that 99% of the spill was 0.1 microns thick, while the remainder was 100 microns thick.26 Thus, while estimating volume from surface appearance may be inherently unreliable, the non-government scientists appeared to make greater efforts to be clear and rigorous in their methodologies, possibly leading their estimates to be closer to the actual flow rate (though still far off).
Estimates Based on Video of the Flow
On May 12, 2010, BP released a thirty-second video of oil coming out of the end of the broken riser—a crucial piece of data. As discussed above, the government?s estimate of [bbls/day] appears to have been based on visual observation of flow from the riser. Within 24 hours, at least three scientists had used various methodologies to derive estimates of the flow rate substantially greater than the government?s then-current estimate.27
- Dr. Timothy Crone, a marine geophysicist at Columbia University?s Lamont-Doherty Earth Observatory, estimated that 50,000 to 100,000 bbls/day of total flux were flowing out of the end of the riser.28 To determine the velocity of the flow, Dr. Crone used a technique called Optical Plume Velocimetry, which involves temporal cross-correlation of the visual intensity of two pixels in a video (both in the plume, one downstream from the other). He developed this technique in a 2008 peer-reviewed paper relating to flow rates.29
- Dr. Eugene Chiang, an astrophysicist at the University of California at Berkeley, estimated the total flux from the end of the riser to be between 20,000 and 100,000 bbls/day.30 Dr. Chang is an expert in orders-of-magnitude estimation (i.e., estimating size or scale from small amounts of data). He estimated the velocity of oil coming out of the riser based on the angle of flow and the rate at which oil would naturally rise through sea water. He used this information to estimate the diameter of the riser, which generated the high end of his range; he based the low-end number on information that the plume could be emanating from a smaller pipe within the riser.31
- Dr. Steven Wereley, a mechanical engineer at Purdue University and expert in fluid mechanics, estimated that the total flux from the end of the riser was 72,179 bbls/day (±20%).32 To arrive at this estimate, Dr. Wereley used a method called Particle Image Velocimetry, which uses a computer program to identify and track distinct “flow structures” in the plume exiting the riser (akin to the billows of a cloud). The method analyzes how fast structures move across the screen in terms of pixels, and then factors in scale and volume to determine flow rate.33 Dr. Wereley co-authored a 2007 book on this flow-rate estimation method.34
All of these non-government figures estimated the total flux being released from the end of the riser, which includes both oil and natural gas.35 If we were to assume the then-current understanding that the flux was 50% oil, the Crone, Chiang, and Wereley estimates would be, respectively: 25,000-50,000 bbls/day; 10,000-50,000 bbls/day; and 36,090 bbls/day. The Crone, Chiang, and Wereley estimates did not include flow from the kink leak, for which there was then no public data.
BP attempted to dismiss the Crone, Chiang, and Wereley estimates. It told National Public Radio on May 13, 2010, that “there?s no way to estimate the flow coming out of the pipe accurately.”36 Five days later, BP released the first video of the kink leak and an initial estimate that the flux was about 50% oil. Testifying before Congress the next day, Dr. Wereley estimated that the kink leak was producing a flow of roughly 2[bbls/day] (±20%) of total flux. Adding that figure to his previous estimate of flow from the end of the riser (72,179 bbls/day of flux), he arrived at a total flow rate of approximately 50,000 bbls/day of oil.37
The Crone, Chiang, and Wereley estimates proved to be significantly more accurate than the official estimates. The government?s [bbls/day] figure, derived from the same type of visual observation as the Crone, Chiang, and Wereley estimates, appears to have been based on a cruder methodology than at least Crone?s and Wereley?s. It is possible that the early official flow estimates would have been more accurate if the government had either enlisted greater in-house scientific expertise, or enlisted outside scientific expertise by making available the data on which government estimates were based. The government appears to have taken an overly casual approach to the calculation and release of the [bbls/day] estimate—which, as the only official estimate for most of May, took on great importance.
Suggestions for the Commission’s Consideration:
- The Commission may wish to recommend adoption of policies or procedures to ensure that, in a federal spill response, the federal government dedicates appropriate scientific expertise to initial spill volume estimates, to the extent that it wishes to release such estimates.
- The Commission may wish to recommend that, where possible without compromising confidentiality or operations, the federal government disclose the methodology and/or data on which its spill volume estimates are based either to the public or to outside scientific experts. Such information would allow outside scientists to generate estimates or to offer informed criticism of the government?s work, helping to refine and to increase public confidence in official estimates.
The Impact on Operations
Government responders have repeatedly insisted to Commission staff that low initial flow-rate estimates did not impact the response. Responders have uniformly maintained—and, indeed, publicly stated during the response itself—that they scaled their efforts to the “worst-case” spill scenario rather than to official flow-rate estimates. Commission staff is still in the process of evaluating these assertions. Because the worst-case figures that emerged within days of the spill, although imprecise, ended up being roughly equivalent to the actual flow rate, we cannot at this point conclude that inaccurate official estimates adversely impacted the response and clean-up operations. It may, however, have been better practice for the government to disclose the estimates that drove the Unified Command operational plan—that is, the operational worst-case discharge figures. Moreover, it is possible that inaccurate flow-rate figures may have hindered the subsea efforts to stop and to contain the flow of oil at the wellhead. A later staff working paper will consider the impact, if any, of flow-rate estimates on well control and containment.
Soon after the spill began, frontline [Guard] personnel requested worst-case discharge information from the Minerals Management Service and BP, both of which reported a figure of 162,000 bbls/day (the worst-case estimate from BP?s original drilling permit).38 A high-level official, however, told us that the [Guard] did not believe the figure from the drilling plan was a credible worst-case estimate.39 On April 23, 2010, the [Guard] and NOAA received an updated estimate of 64,000-110,000 bbls/day, which appeared in both an - 9 -
internal [Guard] Situation Report and on a dry-erase board in the NOAA Seattle war room.40 By early May, BP had lowered its worst-case estimate to 60,000 bbls/day.41 BP officials disclosed a similar estimate to Congress on May 4, 2010, stating during a briefing that the “maximum estimated flow would be 60,000 barrels a day, with a mid-range estimate of 40,000 barrels a day . . . .”42
Thus, although there is evidence to suggest that the worst-case discharge figures BP disclosed to the Unified Command and Congress did not conform to its internal worst-case estimates,43 front-line responders may have based their decision-making on estimates roughly reflecting the magnitude of the spill. But despite the fact that the Unified Command had this information, relied on it for operations, and publicly stated that it was operating under a worst-case scenario, the government never disclosed what its operational scenario was. As a confidential NOAA report drafted on April 28, 2010, noted: “There is no official change in the volume being released but the Guard is no longer stating that the release rate is 1,000 barrels a day. Instead they are saying that they are preparing for a worst-case release and bringing all assets to bear.”44 Responders stuck to this blueprint, stating that, while 1,000 or [bbls/day] were the official best flow-rate estimates, the government was scaling the response to an unquantified worst-case scenario.45
The decision to withhold worst-case discharge figures may have been made above the operational level. It is the understanding of the Commission staff that the possibility of releasing the worst-case discharge figures was at least discussed at the Unified Command level.46 The Commission staff has also been advised that, in late April or early May 2010, NOAA wanted to make public some of its long-term, worst-case discharge models for the Deepwater Horizon spill, and requested approval to do so from the White House?s Office of Management and Budget.47 Staff was told that the Office of Management and Budget denied NOAA?s request.48
The Commission may wish to consider recommendations that encourage government responders to disclose information about the scenarios under which they are operating—in this case, the operational worst-case discharge estimates. Putting aside the question of whether the public had a right to know the worst-case discharge figures, disclosure of those estimates, and explanation of their role in guiding the government effort, may have improved public confidence in the response. Instead, government officials attempted to assure the public that they were not basing operations on the official flow-rate estimates, while not stating what they were basing operations on instead. That lack of information may have contributed to public skepticism about whether the government appreciated the size of the Deepwater Horizon spill and was truly bringing all of its resources to bear. Moreover, the national response may have benefited early on from a greater sense of urgency, which public discussion of worst-case discharge figures may have generated.49
Suggestions for the Commission’s Consideration:
- The Commission may wish to consider recommendations that encourage government responders to scale operations to a credible worst-case scenario, as it appears they did here, and to disclose information about their operational scenarios. Such a recommendation would be consistent with current [Guard] policy, which directs responders not to “lose sight of the importance of accurate and timely spill volume quantification based on maximum potential volume during initial response actions.”50
The Flow Rate Technical Group
Although responders stated that accurate flow-rate estimates were not important to their operations, the Unified Command eventually felt a need to assert leadership on the issue, possibly as a result of media attention and public criticism of the low early numbers.51 On May 19, 2010, the National Incident Command spearheaded the creation of an inter-agency Flow Rate Technical Group ([Rate Group]) and charged it with generating (1) a preliminary flow rate as soon as possible and (2) a final flow-rate estimate based on peer reviewed methodologies within two months.52 On May 23, 2010, Dr. Marcia McNutt, Director of the [Geological Survey] and Science Advisor to the Secretary of the Interior, was appointed [Group]?s leader.53
The [Rate Group] was originally comprised of three sub-groups, made up of both governmental and non-governmental scientists: (1) the Plume Modeling Team, which used the Particle Image Velocimetry method to estimate flow velocity from video of the leaks; (2) the Mass Balance Team, which estimated spill size from aerial images taken by NASA?s AVARIS aircraft; and (3) a team that analyzed the flow captured by the Riser Insertion Tube Tool to establish a baseline flow rate.
The [Rate Group] enlisted non-governmental scientists with applicable expertise and experience, including Dr. Wereley, a critic of the low early estimates. The Group?s initial estimates, however, proved chronically low, too. Moreover, [Group]?s later, more accurate estimates relied primarily upon data collected by a team led by Secretary of Energy Dr. Steven Chu and a team from the Woods Hole Oceanographic Institution, though [Group] did play a significant role in analyzing and interpreting this data.
The initial success of the [Rate Group] is questionable given that it did not release an arguably accurate flow-rate estimate until mid-June, when Secretary Chu?s team was able to provide [Group] with data. The Group?s estimates may also have suffered from a failure to disclose enough information to enable other experts to assess the group?s methodologies and findings. If more of [Group]?s data had been made public, its estimates may have evolved more rapidly with input from the broader scientific community.54
May 27, 2010 Estimate (12,000-2[bbls/day])
The [Rate Group] published its first estimate on May 27, 2010, noting that “[t]he only range of flow rates that is consistent with all 3 of the methods considered by the Group is 12,000 to 19,000 barrels per day. Higher flow rates [[up to 2[bbls/day]]] are consistent with the data considered by (the Plume Team).”55 The Group?s press release contained little information as to how each of the three [Rate Group] teams calculated those ranges, other than to note that the Plume Team?s range of 12,000-2[bbls/day] was “an initial lower bound estimate.”56
On June 2, 2010, the [Rate Group] released a three page Summary Preliminary Report that explained the May 27, 2010 estimate in more detail.57 That document noted that the Plume Team produced “a range of lower bounds” of 12,000 to 2[bbls/day] (±40%), but did not elaborate on the underlying data or calculations.58 Moreover, the June 2, 2010 report did not include the upper ranges of the Plume Team?s estimates because “[t]he experts concluded that the effect of the unknown unknowns made it more difficult to produce a reliable upper bound on the flow rate.”59 It is the Commission staff?s understanding that the “lower bound” range was simply a collection of the minimum estimates produced by each of the Plume Team members. A few members had also produced maximum estimates, several of which were in excess of 50,000 bbls/day, but this upper bound was not released.60 Further, the Plume Team?s report originally contained appendices that revealed some divergence of opinion within the Team.61 The appendices were not publicly released.
June 10, 2010 Estimate (20,000-40,000 bbls/day)
On June 10, 2010, the [Rate Group] announced a revised flow-rate estimate of 25,000 to 30,000 bbls/day with a lower bound of 20,000 and a higher bound of 40,000 bbls/day.62 The Group produced a three page document called Pooling Expert Assessments to accompany those estimates. That document provided intervals with high and low numbers from each of six members of the Plume Team, but only after a “statistical procedure” was applied to “reconcile” the different members? full ranges.63
The June 10, 2010 press release also noted an estimate by researchers with Woods Hole, led by Dr. Richard Camilli, who were conducting their work outside of the auspices of the [Rate Group] but in coordination with Unified Command. On May 31, 2010, these researchers had used an ROV mounted with sonar and acoustic sensors to determine the volume and velocity of the outflow from the end of the riser and kink leak. Their initial rough estimate was a flow rate for total flux (oil plus gas) of between 65,213 and 124,991 bbls/day (0.12 to 0.23 cubic meters/second).64 Along with the [Rate Group]?s press release, the government released a brief one-page statement by the Woods Hole team explaining the methodology behind its estimate.65 Yet, seemingly because this estimate was given in cubic meters/second rather than bbls/day, it did not attract media attention.
Finally, the June 10, 2010 press release announced that two new teams had been added to the [Rate Group]: the Reservoir Modeling Team, which would help determine the rate at which oil flowed from the reservoir into the well, and the Nodal Analysis Team, which would use that information to determine the rate at which the oil traveled through the well and into the Gulf.66 These two teams did not contribute to the June 10 or June 15 estimates. Their purpose was to arrive at a peer-reviewed, final estimate, while the Plume and Mass Balance Teams focused on generating preliminary estimates for public release and for use during the spill.67
June 15, 2010 Estimate (35,000-60,000 bbls/day)
On June 15, 2010, the [Rate Group] announced that it had generated a new official flow estimate of 35,000 to 60,000 bbls/day in conjunction with Secretary Chu and Secretary of the Interior Ken Salazar.68 According to the accompanying press release, the new estimate was “based on a combination of analyses of high resolution videos taken by ROVs, acoustic technologies, and measurements of oil collected by the oil production ship together with pressure measurements inside the top hat.”69 No additional information on methodology was provided.
We now know that the high end of this estimate was accurate because of pressure readings from a sensor that Secretary Chu?s team had BP place in the Top Hat above the blowout preventer on June 13, 2010.70 These sensor readings, along with data on the amount of oil being captured by the Top Hat and an estimate of the percentage of oil still escaping into the Gulf, allowed Secretary Chu?s team to generate a flow rate estimate of approximately 60,000 bbls/day.71
On June 14, 2010, Secretary Chu and his team, Secretary Salazar, and members of the [Rate Group] hosted a conference call.72 On the call, the teams decided that they would jointly announce a flow-rate range of 35,000-60,000 bbls/day. The Chu team?s estimate accounted for the high end of the range, while the [Rate Group]?s work provided the low end.73
The Current Estimate (52,700-62,200 bbls/day)
The June 15, 2010 estimate was finally updated on August 2, 2010. A press release announced that, at the outset of the spill, the flow rate was 62,000 bbls/day (±10%), but that it had declined to 53,000 bbls/day (±10%) by the time the well had been capped on July 14, 2010.74
Another document released on August 4, 2010, the Deepwater Horizon MC252 Gulf Incident Oil Budget, provides some additional details, but none concerning the data upon which the June 15, 2010 estimate was based. It notes only that the “[g]overnment estimate of discharge ranged from 62,200 bbl[/day] on April 22, 2010 to 52,700 bbl[/day] on July 14, 2010.”75
We now understand that Secretary Chu?s team calculated the 52,700 bbls/day figure by taking pressure readings on July 14, 2010, using a sensor inside the capping stack that eventually stopped the flow of oil entirely. Before all of the valves on the stack had been closed, the sole channel for flow into the Gulf was an opening in the capping stack?s kill line. Pressure readings from inside that line, along with some other data points, allowed the government to generate a flow-rate estimate with an uncertainty of plus or minus ten percent.76
Once the capping stack was closed on July 15, 2010, the pressure from the reservoir was about 2,000 psi lower than anticipated, signaling that it had decreased by that amount during the spill. Using this information, and modeling backwards, Secretary Chu?s team and the [Rate Group] together arrived at an estimate of 62,200 bbls/day for the first day of the spill, based on the Woods Hole finding that 43.7% of the total flux was oil.77 Given the new figures, the Deepwater Horizon MC252 Gulf Incident Oil Budget concluded that the total amount of oil discharged during the spill was 4,928,100 barrels (± 10%, which gives a range of 4,435,290 to 5,420,910 total barrels), a number not reduced by the amount of oil captured at the wellhead.78
The [Rate Group] is presently compiling more information on the flow rate and total amount of oil discharged, with the intention of generating a peer-reviewed paper. Release of this paper will allow for a better assessment of [Group]?s work and value. It is worth noting now, however, that the [Rate Group] did not succeed in releasing an accurate high-end estimate until mid-June and that Secretary Chu?s team, rather than any of [Group]?s teams, appears to have been responsible for the accuracy of that June 15 estimate. Moreover, the best currently available data on flow rate were collected by Secretary Chu?s team and the Woods Hole team, although the [Rate Group]?s Reservoir and Nodal Analysis Teams, established in June, played a significant role in interpreting and analyzing the data to arrive at the current government estimate.
Suggestions for the Commission’s Consideration:
The [Rate Group] may be a valuable model for integration of outside scientific expertise, and the Commission may wish to recommend certain methodologies used by [Group] as best practices. However, the Commission may also wish to recommend a review of why [Group]?s estimates were initially inaccurate. Finally, the Commission may want to consider whether scientific working groups such as the [Rate Group] should disclose more of their underlying data and methodologies, allowing for greater input from the rest of the scientific community.
Final Government Estimate Versus Estimates of Independent Scientists
The flow-rate estimates of non-governmental scientists, generated since the well was capped, are useful in assessing the accuracy and durability of the government?s current figures.
In a peer-reviewed paper published on September 23, 2010, Dr. Timothy Crone and Dr. Maya Tolstoy of Columbia University?s Lamont-Doherty Earth Observatory describe their total estimate of the flow from the Macondo well. Using the Optical Plume Velocimetry method referenced above, they conclude that, from April 22, 2010, until the riser was cut on June 3, 2010, the flow rate was 55,900 bbls/day (±21%); and that between June 3 and July 15, 2010, when the well was capped, the flow was 67,500 bbls/day (± 19%).79 Crone and Tolstoy estimate the total release to be 5,174,887 barrels (± 20%).80 Their calculations assume that oil represents 40% of the total flux from the well and do not include oil that was released from the kink leak prior to the riser cut on June 3, 2010.81 If the kink leak were taken into account, and the oil ratio was increased to the 43.7% figure generated by the Woods Hole team, this estimate would be on the high end of the government?s current estimate for the total release.
The Woods Hole team also generated an estimate for the total flow from the well.82 On May 31, 2010, with the aid of the [Guard], Woods Hole took readings from the end of the broken riser and kink leak. The measurements at each site were taken using an ROV-mounted acoustic Doppler current profiler to determine velocity, and imaging multi-beam sonar to determine flow volume.83
Following the [Rate Group]?s press release on June 10, 2010, the Woods Hole team refined its data and factored in the assumption that oil accounted for 43.7% of the total flux.84 With this new assumption, the team concluded that, on May 31, 2010, the riser was leaking oil at 40,700 bbls/day and the kink was leaking at 18,500 bbls/day, for a total oil flow of 59,200 bbls/day.85 Using that figure and the 53,000 bbls/day estimate for July 14, 2010 generated by Secretary Chu?s team, the Woods Hole team calculated the declining flow rate over time, from April 22 to July 14, 2010. The team estimated a total release of approximately five million barrels during the course of the spill.86
The emerging consensus is that roughly five million barrels of oil were released by the Macondo well, with roughly 4.2 million barrels pouring into the waters of the Gulf of Mexico. Using different methods, the government teams and independent scientists arrived at the same approximate figure.
Suggestions for the Commission’s Consideration:
The Commission may wish to recommend the technology and/or methods used by the Woods Hole team as a best practice going forward, if flow rate has to be determined rapidly in the absence of accurate pressure readings.
Notes
1 As explained further in Part I.B below, Commission staff is still gathering information with which to evaluate responders? assertions.
2 DEEPWATER HORIZON MC252 GULF INCIDENT OIL BUDGET OIL BUDGET (Aug. 4, 2010), available at http://www.noaanews.noaa.gov/stories2010/PDFs/DeepwaterHorizonOilBudget20100801.pdf.
3 Press Release, United States [Guard], [Guard] Responding to Oil Drilling Platform Fire (Apr. 21, 2010), http://app.restorethegulf.gov/release/2010/04/22/coast-guard-responding-oil-drilling-platform-fire-0.
4 See CNN Wire, [Guard]: Oil Rig That Exploded Has Sunk, CNN (Apr. 22, 2010), http://news.blogs.cnn.com/2010/04/22/coast-guard-oil-rig-that-exploded-has-sunk/.
5 See CNN Wire, Oil Slick Spreads from Sunken Rig, CNN (Apr. 22, 2010), http://www.cnn.com/2010/US/04/22/oil.rig.explosion/index.html.
6 Press Conference, Admiral Mary Landry and Doug Suttles, New Orleans, LA (Apr. 28, 2010), http://cgvi.uscg.mil/media/main.php?g2_itemId=843309.
7 Television Interview of Mary Landry, [Guard]: Oil Not Leaking from Sunken Rig, CBS NEWS (Apr. 23, 2010), http://www.cbsnews.com/video/watch/?id=6424647n.
8 Press Conference, Admiral Mary Landry and Doug Suttles, New Orleans, LA (Apr. 28, 2010); Campbell Robertson, Oil Leaking Underwater From Well in Rig Blast, N.Y. TIMES (Apr. 24, 2010).
9 See, e.g., Press Release, Unified Command, Unified Command Continues to Respond to Deepwater Horizon (Apr. 25, 2010), http://app.restorethegulf.gov/release/2010/04/25/update-7-unified-command-continues-respond-deepwater-horizon-0; Robertson, Oil Leaking Underwater From Well in Rig Blast.
10 Interviews with government officials.
11 Press Conference, Admiral Mary Landry and Doug Suttles, New Orleans, LA (Apr. 28, 2010); Tim Dickinson, The Spill, the Scandal, and the President, ROLLING STONE (June 8, 2010).
12 See, e.g., Joel Achenbach, How Big is the Gulf Spill, Really?, SEATTLE TIMES (May 13, 2010); Justin Gillis, Size of Spill Is Underestimated, Scientists Say, N.Y. TIMES (May 13, 2010).
13 NOAA, OFFICE OF RESPONSE AND RESTORATION, OPEN WATER OIL IDENTIFICATION JOB AID FOR AERIAL OBSERVATION (Nov. 2007), http://response.restoration.noaa.gov/book_shelf/1462_FINAL%20OWJA%202007.pdf.
14 NOAA Document. The scientist also verbally noted to the Scientific Support Coordinator that the flow rate might be upwards of 10,000 bbls/day. Internal NOAA e-mail.
15 NOAA Document. The scientist generated this number by assuming that the flow came from a hole with a diameter of 40 centimeters, at a velocity of 15 cm/sec, and that 50% of the total flow was oil (as opposed to gas and other material).
16 Id.; Interview with government official. The surface volume calculation was based on an American Society for Testing and Materials standard for determining surface oil thickness. The scientist also assumed that at least half of the oil released evaporated or dispersed in the water column before reaching the surface. Based on those assumptions, the surface volume estimate could have confirmed or yielded a flow-rate estimate of [bbls/day]. E.g., if the oil first began leaking on April 22, 2010, one could then take the 10,000 barrel figure, double it to account for evaporation and dispersion, and then divide by 4 days to arrive at [bbls/day]. The one-page document, however, did not take these steps. Instead, it estimated the “present volume release rate” of [bbls/day] based on visual observation of the velocity of the oil plume leaking from the riser.
17 NOAA Document. For example, the document notes that the velocity could be “between 7 cm/sec and 30 cm/sec” and then, without explanation, uses 15 cm/sec when generating the [bbls/day] estimate (using 30 cm/sec, the flow rate would be over 10,000 bbls/day).
18 Internal NOAA e-mail.
19 Interview with government official. On the day following the release of the [bbls/day] estimate, President Obama made his first public statement about the spill and Secretary of Homeland Security Janet Napolitano declared it a Spill of National Significance, paving the way for the appointment of Admiral Thad Allen as National Incident Commander. BP Oil Spill Timeline, GUARDIAN (July 22, 2010), http://www.guardian.co.uk/environment/2010/jun/29/bp-oil-spill-timeline-deepwater-horizon; Campbell Robertson, White House Takes a Bigger Role in the Oil Spill Cleanup, N.Y. TIMES (Apr. 29, 2010).
20 See Press Release, Unified Command, [Rate Group] Provides Preliminary Best Estimate of Oil Flowing from BP Oil Well (May 27, 2010), http://app.restorethegulf.gov/release/2010/05/27/flow-rate-group-provides-preliminary-best-estimate-oil-flowing-bp-oil-well.
21 See John Amos, Gulf Oil Spill Rate Must Be Much Higher Than Stated—6 Million Gallons So Far?, SKYTRUTH.ORG (Apr. 27, 2010), http://blog.skytruth.org/2010/04/gulf-oil-spill-rate-must-be-much-higher.html; John Amos, Gulf Oil Spill—Bigger Than Exxon Valdez, SKYTRUTH.ORG (Apr. 28, 2010), http://blog.skytruth.org/2010/04/gulf-oil-spill-bigger-than-exxon-valdez.html.
22 Id. Amos assumed that all oil leaking from the well reached the surface to be observed. If that were not the case, the estimated flow rate would be higher.
23 Ian Talley, Experts: Oil May Be Leaking at Rate of 25,000 Barrels a Day in Gulf, WALL ST. J. (Apr. 30, 2010); see also Emily Gertz, Gulf Oil Spill Far Worse Than Officials, BP Admit, Says Independent Analyst, ONE EARTH BLOG (Apr. 29, 2010), http://www.onearth.org/node/2084.
24 John Amos, Gulf Oil Spill—New Spill Calculation—Exxon Valdez Surpassed Today, SKYTRUTH.ORG (May 1, 2010), http://blog.skytruth.org/2010/05/gulf-oil-spill-new-spill-rate.html.
25 Achenbach, How big is Gulf Spill, Really?; Gillis, Size of Oil Spill Underestimated, Scientists Say; NOAA Paper, Visual Observations and the Bonn Agreement.
26 NOAA Document.
27 Richard Harris, Gulf Spill May Far Exceed Official Estimates, NATIONAL PUBLIC RADIO (May 14, 2010), http://www.npr.org/templates/story/story.php?storyId=126809525.
28 Id.; Raymond Gellner, BP Oil Spill Rate in Gulf May Be 3 Million Gallons Per Day, WORLD NEWS EXAMINER (May 14, 2010); Telephone Interview with Dr. Timothy Crone, Lamont-Doherty Earth Observatory (Aug. 18, 2010 and Sept. 3, 2010). Dr. Crone noted that his early estimates could not be more precise because of the low quality and short duration of the video.
29 Crone, McDuff, & Wilcock, Optical Plume Velocimetry: A New Flow Measurement Technique for Use in Seafloor Hydrothermal Systems, EXPERIMENTS IN FLUIDS, vol. 45, no. 4, at 899-915 (2008).
30 Harris, Gulf Spill May Far Exceed Official Estimates.
31 Id.; Telephone Interview with Dr. Eugene Chiang, University of California at Berkeley (Aug. 13, 2010).
32 Harris, Gulf Spill May Far Exceed Official Estimates; Sizing up the BP Oil Spill: Science and Engineering Measuring Methods, Briefing Before the Subcomm. on Energy and Environment of the H. Comm. on Energy and Commerce, 111th Cong. (May 19, 2010) (testimony of Dr. Steven Wereley).
33 Id.
34 RAFFEL, WILLERT, WERELEY, & KOMPENHANS, PARTICLE IMAGE VELOCIMETRY: A PRACTICAL GUIDE (2d ed. 2007).
35 Telephone Interview with Dr. Chiang, (Aug. 13, 2010); Telephone Interview with Dr. Crone, (Aug. 18, 2010); Telephone Interview with Dr. Steven Wereley, Purdue University (Aug. 12, 2010). While estimates of the oil-to-gas ratio in the flux varied over the course of the spill, scientists from Woods Hole Oceanographic Institution took measurements at the source and concluded that the flux was 43.7% oil.
36 MORNING EDITION, Transcript: Gulf Spill May Far Exceed Official Estimates, NATIONAL PUBLIC RADIO (May 14, 2010), http://www.npr.org/templates/transcript/transcript.php?storyId=126809525; Suzanne Goldenberg, Marine Scientists Study Ocean-Floor Film of Deepwater oil leak, GUARDIAN (May 13, 2010), http://www.guardian.co.uk/business/2010/may/13/bp-oil-spill-ocean-footage.
37 Sizing up the BP Oil Spill: Science and Engineering Measuring Methods, Briefing Before the Subcomm. on Energy and Environment (May 19, 2010) (testimony of Dr. Steven Wereley).
38 Interviews with [Guard] officials; INITIAL EXPLORATION PLAN, MISSISSIPPI CANYON BLOCK 252, 7-1 (Mar. 10, 2009), available at http://www.gomr.boemre.gov/PI/PDFImages/PLANS/29/29977.pdf.
39 Interview with [Guard] official.
40 U.S. COAST GUARD, DISTRICT EIGHT SITUATION REPORT 18 (April 23, 2010), available at http://s3.documentcloud.org/documents/3176/uscg-logs.pdf; Ben Raines, Video Shows Federal Officials Knew Quickly of Potential for Massive Oil Flow in Gulf Spill, MOBILE PRESS-REGISTER (May 1, 2010), http://blog.al.com/live/2010/05/video_shows_federal_officials.html; Dickinson, The Spill, the Scandal, and the President. The refined worst-case figure apparently came from either the Minerals Management Service or BP, though its origin and the methodology underlying it have not been established. Interviews with government officials.
41 Interviews with government officials.
42 Press Release, Markey: New Flow Rate Shockingly Close to BP?s Initial “Worst Case” Scenario (Aug. 2, 2010), http://markey.house.gov/index.php?option=com_content&task=view&id=4076&Itemid=141.
43 In May 2010, BP turned over a document to congressional investigators that demonstrated that their in-house estimates were as follows: “[e]xpected range of possible flow rates is 5,000 to 40,000 BOPD,” the “[m]aximum theoretical flow rate is 60,000 BOPD,” and, if the BOP and wellhead are removed, “the rate could be as high as ~ 100,000 barrels per day . . . .”; BP WORST CASE SCENARIO DOCUMENT, available at http://globalwarming.house.gov/files/WEB/flowrateBP.pdf; Ernest Scheyder, BP Estimates Oil Spill Up to 100,000 Barrels Per Day in Document, REUTERS (June 20, 2010), http://www.reuters.com/article/idUSN1416392020100620; Bryan Walsh, The Worse Case Scenario Gets Worse for BP as New Documents Come to Light, TIME (June 21, 2010), http://ecocentric.blogs.time.com/2010/06/21/the-worse-case-scenario-gets-worse-for-bp-as-new-documents-come-to-light/.
44 Ben Raines, Leaked Report: Government Fears Deepwater Horizon Well Could Become Unchecked Gusher, MOBILE PRESS-REGISTER (Apr. 30, 2010), http://blog.al.com/live/2010/04/deepwater_horizon_secret_memo.html.
45 See, e.g., Press Briefing, Admiral Thad Allen (May 1, 2010), http://www.marinelog.com/DOCS/NEWSMMIX/2010may00010.html (“At the outset, when we realized that the unit had sunk, we made preparations to stage equipment for a worst-case scenario. The deployment of our equipment was not related to any of the early estimates related to 1,000 barrels a day or 5,000 barrels a day . . . .”); Press Briefing, Admiral Mary Landry (May 14, 2010), http://app.restorethegulf.gov/release/2010/05/19/transcript-press-briefing-may-14-2010 (“Whether the flow is one, five, 10, or 15 thousand barrels per day, the mobilization of resources has been to prepare for a worst-case scenario. Our resources and tactics are not constrained by flow estimates—I have to emphasize that.”); Press Briefing, NOAA Administrator Jane Lubchenco (May 20, 2010), http://app.restorethegulf.gov/release/2010/05/24/teleconference-lubchenco-may-20 (“5,000 was always understood to be a very rough estimate. That number was useful and sort of the best estimate at the time. . . . The response to the spill has never been pegged to that estimate of Bbls/day or any other estimate. We?ve always pegged our response to the worst-case scenario and had much more significant effort than would have been required had it only been five.”).
46 Interviews with government officials.
47 Id.
48 Id.
49 See Draft Staff Working Paper: Decision-Making Within the Unified Command, National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling, at § II(B).
50 See [Guard] “ALCOAST” Bulletin, Subject: [Guard] Marine Environmental Response Doctrine—Incident Specific Preparedness Review (ISPR) Update (Jan. 12, 2009), http://www.uscg.mil/announcements/alcoast/ALCOAST02209.txt.
51 Interview with government official.
52 [Guard] Document; Press Release, Deepwater Horizon Incident Joint Information Center, The Ongoing Administration-Wide Response to the Deepwater BP Oil Spill (May 21, 2010), http://www.restorethegulf.gov/release/2010/05/21/ongoing-administration-wide-response-deepwater-bp-oil-spill.
53 [Guard] Document.
54 On September 16, 2010, Public Employees for Environmental Responsibility (PEER), an environmental whistleblower group, filed a complaint in the U.S. District Court for the District of Columbia under the Freedom of Information Act seeking to compel the [Geological Survey] to produce documents related to the [Rate Group]. PEER requested these records “in order to learn about how the Geological Survey and the Rate Group developed a scientific estimate of the rate of oil leaking from the BP Deepwater Horizon blowout in the Gulf of Mexico.” Complaint at 2, Pub. Emp. for Envtl. Responsibility v. Dep’t of Interior (D.D.C. filed Sept. 16, 2010).
55 Press Release, Deepwater Horizon Incident Joint Information Center, [Rate Group] Provides Preliminary Best Estimate of Oil Flowing From BP Oil Well (May 27, 2010), http://app.restorethegulf.gov/release/2010/05/27/flow-rate-group-provides-preliminary-best-estimate-oil-flowing-bp-oil-well.
56 Id.
57 MARCIA MCNUTT, SUMMARY PRELIMINARY REPORT FROM THE FLOW RATE TECHNICAL GROUP (June 2, 2010), available at http://www.doi.gov/deepwaterhorizon/loader.cfm?csModule=security/getfile&PageID=33972.
58 Id.
59 Id.
60 Interview with non-governmental source.
61 E-mail from Steven Wereley to members of the Flow Rate Technical Group?s Plume Team (May 27, 2010), available at http://www.usgs.gov/foia/budget/05-27-2010...experts%20statements%201.pdf.
62 Press Release, Deepwater Horizon Incident Joint Information Center, Admiral Allen; Dr. McNutt Provide Updates on Progress of Scientific Teams Analyzing Flow Rates From BP?s Well (June 10, 2010), http://app.restorethegulf.gov/release/2010/06/10/admiral-allen-dr-mcnutt-provide-updates-progress-scientific-teams-analyzing-flow-.
63 ANTONIO POSSOLO AND PEDRO ESPINA, POOLING EXPERT ASSESSMENTS (June 8, 2010), available at http://www.doi.gov/deepwaterhorizon/loader.cfm?csModule=security/getfile&PageID=34800. On the same day, the Plume Team issued a statement noting: “On May 27, the Team issued an Interim Report that established an estimated range for the minimum possible spillage rate but did not issue an estimate for a possible maximum value because the quality and length of the video data could not support a reliable calculation.” BILL LEHR, STATEMENT OF THE PLUME TEAM OF THE FLOW RATE TECHNICAL GROUP (June 10, 2010), available at http://www.doi.gov/deepwaterhorizon/loader.cfm?csModule=security/getfile&PageID=34638.
64 Press Release, Deepwater Horizon Incident Joint Information Center, Admiral Allen; Dr. McNutt Provide Updates on Progress of Scientific Teams Analyzing Flow Rates From BP?s Well (June 10, 2010). Commission staff converted the Woods Hole team?s estimate from cubic meters per second to barrels per day.
65 RICHARD CAMILLI, PRELIMINARY REPORT FROM THE WOODS HOLE OCEANOGRAPHIC INSTITUTION FLOW RATE MEASUREMENT GROUP (June 10, 2010), available at http://www.doi.gov/deepwaterhorizon/loader.cfm?csModule=security/getfile&PageID=34799.
66 Press Release, Deepwater Horizon Incident Joint Information Center, Admiral Allen; Dr. McNutt Provide Updates on Progress of Scientific Teams Analyzing Flow Rates From BP?s Well (June 10, 2010).
67 Interviews with government officials.
68 Press Release, Deepwater Horizon Incident Joint Information Center, U.S. Scientific Team Draws on New Data, Multiple Scientific Methodologies to Reach Updated Estimate of Oil Flows from BP?s Well (June 15, 2010), http://app.restorethegulf.gov/release/2010/06/15/us-scientific-team-draws-new-data-multiple-scientific-methodologies-reach-updated.
69 Id. “Top Hat” was the nickname for the loose-fitting cap placed over the top of the blowout preventer, which collected up to approximately 15,000 bbl/day from the Macondo well between June 3 and July 10, 2010.
70 Interviews with government officials; see also Henry Fountain, BP Provides Plan to Speed Up Siphoning, N.Y. TIMES (June 14, 2010).
71 Interviews with government officials.
72 NOAA Document.
73 Interviews with government officials.
74 Press Release, Deepwater Horizon Incident Joint Information Center, U.S. Scientific Teams Refine Estimates of Oil Flow from BP?s Well Prior to Capping (Aug. 2, 2010), http://app.restorethegulf.gov/release/2010/08/02/us-scientific-teams-refine-estimates-oil-flow-bps-well-prior-capping.
75 OIL BUDGET at 1.
76 Interview with government official.
77 Interviews with government officials.
78 OIL BUDGET at 1.
79 Timothy J. Crone, et al., Magnitude of the 2010 Gulf of Mexico Oil Leak, SCIENCE EXPRESS, at 1 (Sept. 23, 2010); Telephone Interview with Dr. Crone (Aug. 18, 2010 and Sept. 3, 2010).
80 Id.
81 Id.
82 Woods Hole was originally contacted by BP on May 1, 2010 to undertake diagnostic work on the failed blowout preventer, which would include measuring the flow rate. BP, however, cancelled the project, citing a need to focus on the containment dome effort. Sizing up the BP Oil Spill: Science and Engineering Measuring Methods, Briefing Before the Subcomm. on Energy and Environment of the H. Comm. on Energy and Commerce, 111th Cong. (May 19, 2010) (testimony of Dr. Richard Camilli).
83 RICHARD CAMILLI, PRELIMINARY REPORT FROM THE WHOI FLOW RATE MEASUREMENT GROUP; Interview with Dr. Richard Camilli and Dr. Christopher Reddy, Woods Hole Oceanographic Institution, Washington, D.C. (Aug. 19, 2010); Telephone Interview with Dr. Camilli. (Sept. 10, 2010).
84 Id. Scientists from Woods Hole took measurements at the source and concluded that 43.7% of the flux was oil.
85 Telephone Interview with Dr. Camilli (Sept. 10, 2010).
86 Id.
