Reporters, a little more numeracy, please?

British Petroleum (and the government) apparently provided implausibly estimates of how much oil was spewing into the Gulf. This was obvious from the start. If reporters were more numerate, they would have known.

Mother Jones has a piece today about whether British Petroleum will own up to the amount of oil it spewed into the Gulf. (I have clarified wording here based on reader comments.)

This controversy raises one of my pet peeves as a technically-trained person. Most reporters aren’t numerate, and it shows.

Often this innumeracy reveals itself in small ways, often in economics or budget stories since these are most likely to involve actual numbers: errors with percentages, confusion about basic financial calculations involving interest rates. Sometimes innumeracy shows itself in other ways, such as believing transparently implausible numbers in specific news stories.

The Gulf spill provides an especially depressing example because this was such a huge and well-covered story. British Petroleum and the government stated that 5,000 barrels per day (***See comments below. My commenters are right to call me for sloppiness here. ***) were leaking into the Gulf. This never made any sense. Bear with me, while I perform a sloppy but revealing calculation that demonstrates why.

There are 42 gallons in a barrel. So this 5,000 barrel figure is equivalent to 210,000 gallons per day. That seems like a big number–until you recall that there are 86,400 seconds in a day. So one would have to believe that a gushing 20-inch-wide pipe was releasing oil at a rate of 2.4 gallons per second.

Is this a little or a lot to be flowing out of a 20 inch pipe? As you recall, the area of a circle is pi-r-squared. So the cross-sectional area of that pipe was about 314 square inches. As you may not recall but could google, a gallon is about 231 cubic inches. So a flow rate of 2.4 gallons/second is about 561 cubic inches per second. This implies fluid would have to be moving through the pipe at a corresponding rate of about 561/314=1.79 inches per second. How fast is that? Since there are 12 inches in a foot, 5,280 feet in a mile, and 3,600 seconds in an hour, you can verify that this flow rate is about 1.79*3600/(12*5280), or …. 0.1 miles per hour. If you saw the live shots of oil gushing out of the Deepwater Horizon, you know this couldn’t be right.

Of course my sloppy ballpark calculation could be pretty far off. The diameter of the pipe might have been slightly smaller. Perhaps the flow was partly blocked. Suppose I left out enough to bias the calculation by a factor of 5. That calculation would still indicate that either (a) BP’s estimate was transparently implausible, (b) other junk was flowing out of the pipe alongside the oil, or (c) some combination of (a) and (b) were both true.

When experts challenged these estimates, reporters treated this as an esoteric matter or a murky he-said, she-said dispute. It wasn’t. (Kudos, though, to NPR for this story.)

It’s telling that even on the most extensively-covered pollution story of the decade, these rather obvious points were so rarely communicated.

Postscript As noted, my commenters are right to call me on this. That 5,000 barrel/day estimate was actually an initial figure embraced by the government. Reporters should have been less credulous in questioning both the government and BP on these points.

Author: Harold Pollack

Harold Pollack is Helen Ross Professor of Social Service Administration at the University of Chicago. He has served on three expert committees of the National Academies of Science. His recent research appears in such journals as Addiction, Journal of the American Medical Association, and American Journal of Public Health. He writes regularly on HIV prevention, crime and drug policy, health reform, and disability policy for American Prospect,, and other news outlets. His essay, "Lessons from an Emergency Room Nightmare" was selected for the collection The Best American Medical Writing, 2009. He recently participated, with zero critical acclaim, in the University of Chicago's annual Latke-Hamentaschen debate.

10 thoughts on “Reporters, a little more numeracy, please?”

  1. It is possible that the pipe above the seafloor or the well could initially have been partially or nearly completely blocked. I seem to recall that the removal of part of the pipe above the BOP caused the flow to increase. The erosive force of the oil flow in the well may also have enlarged the width of the well which would allow more oil to flow through the well. The Oil Drum site could give the answer to these questions.

  2. "BP prefers to rely on measurements of oil on the sea surface made by the Coast Guard and the National Oceanic and Atmospheric Administration" (emphasis added).

    That's from the NPR story you linked to. The 5,000 bpd estimate came from the Coast Guard and NOAA, not BP. In fact, the initial Coast Guard/NOAA estimate was 1,000 bpd. All along, BP has acquiesced to the government's estimates as they increased over time. You can beat up on BP for going along with early estimates they knew were too low, but you can't legitimately accuse BP of having made them in the first place.

    The Mother Jones article says, "We know that BP has consistently low-balled the figure, trying for days to convince us that just 1,000 barrels of oil was coming from the well before they finally acknowledged a revised (and now known to be far too low) 5,000 barrel initial government estimate" (emphasis added). Mother Jones makes it explicit that the 5,000 bpd estimate was from the government. (The initial 1,000 bpd estimate, also from the government, was only in effect for a few days, as Mother Jones also acknowledges–while trying to make it sound like a lengthy period and failing to say where it came from.)

  3. Mother Jones has a piece today about whether British Petroleum low-balled its estimate of the amount of oil it spewed into the Gulf.

    Oh, and by the way, that is not what the Mother Jones piece is about. It's about whether BP will fight the latest government estimate in court. Markey's letter was disingenuous; of course BP is going to "remain coy" about whether it will dispute the estimate, as he surely knows, and as Mother Jones acknowledges at the end of the piece.

    Dan, you're correct that there's considerable uncertainty about the rate of flow in the early days, from what I've been reading on The Oil Drum.

  4. My commenters are right to call me for some sloppiness here. I will modify the piece to note Dan's and Swift Loris's points. I do not believe these undermine my argument. Thanks for the attentive read.

  5. Harold, reporters can't grasp orders of magnitude, let alone deal with unit conversions. To most reporters, a million, a billion, and a trillion are all "big numbers". That a trillion dollars of federal spending (or debt, or taxes) is about $10,000 per household, a billion is about $10, and a million is about a penny, would probably strike them as too much math.


  6. Tony that is certainly true of many reporters. We can't accept that as the way things have to be.

  7. Math? what does any of this have to do with math? (well, maybe the area of a circle). This is arithmetic and very basic (not common, sadly; basic) physical intuition. Harold put out a bleg on this and I was all ready to go find my fluid mechanics handbook and do pipe/pressure/flow calculations, possibly showing off with Reynolds number matching, but it would have added nothing. Harold is an academic health economist, somewhat insecure in the presence of a pair of pliers, and I salute him for not being intimidated about grappling with a straightforward physical laugh test using common sense and real numbers. Everyone is authorized to do this kind of analysis and a lot more of us should, especially reporters.

  8. As you can see from this picture, the oil is not leaking from the pipe itself, but from the mechanism on the *top* of the pipe, possibly from a much smaller area than the cross-section of the pipe seen below.

    As well, the escaping oil is about 40% methane, so yes, there's other gunk mixed in. Again, the photo shows that the escaping oil is expanding rapidly, typical of a mixture of liquid and gas, not straight liquid. Depending on the density of methane at that depth, (My CRC is packed away.) actual oil might be a fairly small fraction of the stream's volume at that point.

    Given those circumstances, I'd say you'd be hard put to get a result within a factor of ten of the real flow, without additional data. Which is not to say you're wrong, just that there's little reason for confidence.

  9. One more point: BP didn't release the first (low-res) video of the leak until May 12; the live streaming video from underwater cameras didn't begin until May 21. IOW, reporters didn't have anything to work with visually until two weeks after the government's 5,000 bpd estimate, which was announced on April 28. As soon as the video became available, doubts about the initial estimates were widely reported.

Comments are closed.