Sunday, May 9, 2010

BP / Gulf Oil Spill - How Big Is It?

Quite a lot bigger than the estimate being uncritically quoted throughout the media of 210,000 gallons (5,000 barrels) per day. That was the last "official" estimate made by NOAA and accepted by the Coast Guard back on April 29 (see timeline below). Before that, the Coast Guard estimates ranged from 336,000 gallons (8,000 barrels) per day, to zero, to 42,000 gallons (1,000 barrels) per day. None of these estimates has been publicly explained or substantiated. And on May 1, the Coast Guard and NOAA stopped trying to estimate the spill rate, with Admiral Thad Allen saying, "Any exact estimate is probably impossible at this time."

But the media continues to report that oil is leaking into the Gulf at 5,000 barrels per day. At SkyTruth we estimate the spill rate is closer to 1.1 million gallons (26,500 barrels) per day, based on the size of the slick on satellite images and Coast Guard maps, and thickness estimates derived from visual descriptions of the slick. That puts us at a total spill of 21 million gallons so far.

Considering that there are deepwater oil wells in the Gulf of Mexico capable of producing >1.26 million gallons (30,000 barrels) of oil per day under controlled flow rates; and that the oil slick continues to grow in size even though it's been under attack 24/7 by skimmer vessels, burning, chemical dispersants, and natural processes; our estimate seems quite conservative.

Why is it important to get this number right? This is about more than just liability, or PR. You can bet that our future response capacity is going to be overhauled and retooled based on this spill. If we low-ball the spill amount and rate, we run the risk of designing an inadequate new spill-response system that is doomed to fail the next time something this big occurs.

Here's the timeline of spill estimates:

  • 4/22 - Deepwater Horizon rig sinks; Coast Guard estimates "up to" 8,000 barrels per day (bpd) is leaking - source
  • 4/23 - Coast Guard reports no leaking at all from the damaged well - source
  • 4/24 - Coast Guard reports well is leaking, estimates 1,000 bpd - source
  • 4/25 - BP repeats 1,000 bpd estimate - source
  • 4/27 - 1,000 bpd still the official Coast Guard and BP estimate - source
  • 4/27 - SkyTruth and Dr. Ian MacDonald publish first estimate that spill rate is 20,000 bpd - source
  • 4/28 - NOAA weighs in and raises the official estimate to 5,000 bpd based on aerial surveys "and other factors"; BP disputes this higher estimate - source
  • 4/29 - Coast Guard and NOAA repeat their estimate of 5,000 bpd - source
  • 4/29 - BP's Chief Operating Officer admits new estimate of 5,000 bpd may be correct; "He said there was no way to measure the flow at the seabed and estimates have to come from how much oil makes it to the surface" - source
  • 5/1 - SkyTruth and Dr. Ian MacDonald publish revised estimate of at least 26,500 bpd - source
  • 5/1 - Coast Guard Admiral Thad Allen "acknowledged there was no way really to know the extent of the leak" - source - and stated that "Any exact estimate is probably impossible at this time" - source
  • 5/1 - Coast Guard and NOAA cease estimating the rate of the spill. BP continues to use 5,000 barrels per day as their estimate of the spill rate.

14 comments:

  1. Many people are questioning your conclusions about the number of gallons being spilled.

    It would be helpful if you could post your calculations.

    It would also be helpful if you could give a map of the area you say has oil slick.

    You have said the slick is about 4,000 sq. miles. We would like to see a map of this 4,000 sq. mile area because it is simply not discernible to lay people from satellite photos.

    Showing your work and how you arrive at your conclusions would be of real benefit and perhaps answer the questions of your detractors.

    ReplyDelete
  2. Thanks John. I'm wondering if your estimates of the total spill are based on the extent of the oil slick that you can see on the NASA image? The image does not capture the extent of dispersed oil that does not create a sheen of oil on the surface. On a boat in the Gulf on Wednesday, I documented dispersed oil throughout the area of the Breton National Wildlife Refuge. The exact coordinates we registered were N29 23-34.1 and lat 89-12-27.9. I expect the extent of the oil spill will be much greater if the dispersed oil can be tracked as well.

    ReplyDelete
  3. Thank you for the objective analysis of this oil spill. During the last week I have seen countless news articles cite the NOAA estimate of 5,000 barrels/day. I don't like bad news, I just want to know what is going on. Luckily, SkyTruth offers up-to-date, scientific estimates of the flow of oil from the undersea gusher in the Gulf of Mexico.

    ReplyDelete
  4. David - we've been posting our calculations, and the images and maps they are based on, to this blog for days now. It's our intention to be transparent and let others "kick the tires" on our calculations and assumptions.

    ReplyDelete
  5. Why does it matter to get the release magnitudes correct? There is growing concern regarding the underwater transport of dispersant-treated oil (which does not form a surface layer, but still tends to be buoyant) and oil that has weathered and begun to sink.

    Due to the unprecedented wide-spread nature of this spill, the prospect is that many ecological systems will receive a low, but potentially harmful dose of hydrocarbons including benzine, ethyl-benzine, toluene, and xylene. Essentially, BP is dumping a large amount of toxin into the northeastern Gulf. Magnitude will directly determine dosage, although there are many complicating factors. Dispersant may not make these compounds less-biologically available. However, transport and toxicity modeling is urgently need to predict the process and mitigate to the extent possible.

    It is simply untrue that measuring the volume of a turbulent plume of buoyant material in a deep-ocean setting is beyond the ability of scientists. Video records have been used to estimate gas and oil releases from natural seeps. Similar challenges have been met at hydrothermal vents, where a combination of acoustic current measurements, down-plume concentration sampling, and visual estimates of plume behavior have provided accurate estimates of discharge.

    If there are no attempts being made, it is because there has been no priority put on this problem by the on-scene coordinator (US Coast Guard) and little attempt to reach out to the scientific community that could assist them.

    ReplyDelete
  6. John:

    I can't find them and I have been very diligent looking for them here.

    Can you post the links and then maybe I can respond to their adequacy or deficiencies as I see them.

    I want this blog to be a record of what is going on in the gulf and to be used as authoritative. But I am reluctant to use it as such when I can't figure out how you arrive at your numbers.

    Believe me, I think there is a good possibility that the numbers are under-reported. So I don't want a myth to develop that everyone takes as gospel. If something is not done in a hurry to dispel the myth of 5,000 barrels per day, the myth will become gospel, it it hasn't become gospel already.

    ReplyDelete
  7. David - the calculations are linked right out of this blog post. Here they are again:

    http://tinyurl.com/22osz2k
    http://tinyurl.com/235x8yd

    ReplyDelete
  8. John,

    Do you have any idea why the Coast Guard has stopped publishing proper overflight maps? The latest overflight map that I can find is dated May 3, and even it is not nearly as detailed as the one used to make the 26 Kbbl estimate. I sent a question about the overflights to the USCG Dist. 8 publicity person, but didn't really get much of a reply. I'm curious if you know anything more...

    Those maps are invaluable; They tell a much better story of what the slick is doing than the MODIS images.

    ReplyDelete
  9. christoph - I don't know, those maps were very helpful. If anyone finds them, please send us the link.

    ReplyDelete
  10. David: It is basic math.. calculating the volume of the oil, vol=Length x Width x Height.

    In these oil calculations the stated surface area is Length x Width. The height is estimated based on the appearance of the oil and then referenced to the table on estimated widths for that type of oil spill.

    Next, the volume as micrometers is converted to bbl/acre using the coast guard conversion chart.

    Divide total volume by the total days of spilling to reach an average spill rate per day.

    The variables open to debate that would influence reported spill rate are total surface area and thickness of the spill.

    ReplyDelete
  11. so...doing the math...it looks like the 5,000 barrel number is based on the following...

    11,355 km/sq * .035 = 397.425 m/3

    397.425/.0078 = 50952 barrels (2,038,076 gallons)

    58230/10 = 5095 barrels per day

    This is a calculation that assumes the ENTIRE spill is sheen, but not the thickest possible sheen because it uses a multiplier of .035. It does use the highest possible conversion number for barrels. STILL, this is a gigantic low-ball estimate by any standard.

    ReplyDelete
  12. If people have seen the video of the pipe leaking that was released yesterday it does provide some independent confirmation of the volume of the oil spill. BP has said that the pipe is 12" in diameter and is approx 85% of the spill. Looking at the rate of discharge it has been estimated (by my and others) to be between 1.5 and 3 ft/sec. If 30% by volume is natural gas then the total leak would be estimated to be between 600,000 and 1.2M+ gallons per day. There are fairly large error bars on this estimate and it should confirmed by experts BUT it does provide some secondary confirmation of Johns flow rate estimate of 1.1M gallons per day.

    ReplyDelete
  13. According to a number of news sites, BP is now saying the inner diameter is 20in. Pumping that into the equation changes it to 1.5M - 3M gallons per day (35-70k barrels) out of the pipe (excluding the other leak points).

    Google search's built in calculator is perfect for this sort of thing since it handles all of the conversions for you. Try pasting this into the Google Search box:
    (pi x (10in) ^2) x (36in / sec) x 70% in oil barrels / day

    ReplyDelete
  14. http://www.nytimes.com/2010/06/11/us/11spill.html

    sadly, it looks like you have been right all along.

    thank you so much for what you have been doing.

    ReplyDelete