Three cups, one oil spill and a lot of balls

Today, agencies are reporting the wonderful news that 75% of the 4.9 million barrels of spilled oil from the Deepwater Horizon disaster has miraculously just disappeared (mostly by evaporation and dispersal). Phew, that is good news isn’t it? There’s no need to worry folks, these massively destructive oil spills aren’t so bad after all: a week after you cap them it’s all fine again.

You will forgive me if I take my government spin with a hefty dose of finest Gulf-Sea salt won’t you?

The numbers are coming from this document; an unedifying 5-page pdf containing 20% white space, 20% author list, 10% using a font size a little larger than necessary to make it look longer, 45% “estimated” data (most of it duplicated) and 5% unnecessary definitions for words apparently only the author doesn’t understand[1]. The report is based on this document; an equally execrable 10-page pdf with, as far as I can make out, exactly the same “data” presented in a slightly different format and repeated a few times. Apparently there are footnotes that lead you to the models and research on which the estimates are based, but I can’t seem to find them.

Forgive my cynicism, but it is well known that if you write a document, cite the information in it in another document and then release a summary, then no-one will go back and check the validity of the original. This is a very useful tool if you are a government official who wishes your version of events to quickly become the established truth.

Bloggers are filling the hole left by the press and, on this occasion, I tried to go back through the archives and do some fact-checking for you, but there aren’t any actual facts to check. At all. That in itself is a major problem; I have enormous misgivings about these numbers (that I will come to shortly) but a well-ordered presentation of the data going back through the methodology and explaining what assumptions were made and why would have put my mind at rest. Without that no-one can state categorically that this report is flawed, but why not publish that data if you have it? Its absence is, in and of itself, exceedingly troubling. That the mainstream press have apparently swallowed the numbers in this report without even a cursory look at them  is utterly unsurprising, but my lack of surprise doesn’t reduce my anger at their utter incompetence.

Oil spills are (mercifully) quite rare and the rates of various things like evaporation and dispersal will vary from oil field to oil field making the previous values of limited relevance to whatever spill is ongoing. It’s not difficult research (bung a sample in a dish and record how fast it disappears), so it’s probably easier to conduct it anew as and when you need it. That’s all fair enough and, indeed, they state in their report that their numbers are based on something of the sort[2]. But this is supposed to be a report on the disappearance of the oil, if that data exists and these figures are based upon it, why on earth isn’t the data in the report?!

Crude oil is a mixture of different length hydrocarbons, that is, chains of carbon atoms each with hydrogens attached like the legs of a centipede[3]. These hydrocarbons can be straight with the maximum possible number of hydrogens attached, like those pictured below, or they can have missing hydrogens (those are “unsaturated, like your olive oil), or the carbons can wrap around and make rings, sometimes multiple rings joined together.

Image from chemistryland.com

Crude oil is a vastly complicated mixture of many different lengths and shapes, all the way from methane (one carbon) to the various long-chain hydrocarbons found in bitumen, which contain 35 or more carbons. How quickly these molecules evaporate depends, mostly, on how long they are. Methane, of course, is a gas and it comes out of oil deposits as it is formed (this is why natural gas and oil are found together) but as the hydrocarbon chain length increases evaporation is slower. This is the same thing as saying it evaporates at a higher temperature and this property is how petrol and various other products of crude oil are separated; the oil is placed in a still and heat is applied, the still will be very hot at the bottom and will get cooler towards the top. The smaller hydrocarbons are gasses at lower temperatures and so they will raise all the way to the top, where they are collected. As they rise through the column, the longer molecules condense back to a liquid at higher temperatures and can be recovered lower down the still.  Bitumen, which has a boiling temperature of around 900C, sits stubbornly in the bottom of the waiting to be scraped out at the end and sent for road building.

Picture from http://www.elmhurst.edu/

The above matters, because the amount of crude oil that can reasonably be expected to have evaporated over the length of this disaster will be related to the temperature in the gulf (hot by human standards, but darned chilly if you’re bitumen) and the distribution of the various length molecules in this particular brand of oil (each field is unique). The picture above shows a very crude[4] diagram of what an oil distillation tower looks like, that doesn’t matter, what’s important is the length of the hydrocarbons and the temperatures at which they evaporate. You should be able to see that, right at the top, the easiest things to distill is petroleum gas. That evaporates at temperatures up to 40C and only contains hydrocarbons up to 3 carbons long. Anything longer than that and you need to go higher to get evaporation. Let me just say that again for emphasis: anything longer than 3 carbons, and you need temperatures above 40C for evaporation.

The report claims that 16% of the oil has “naturally dispersed” and 25% has “evaporated or dissolved”. Those are some pretty high numbers; I’m no oil-field geologist, but I’d stake quite a large sum of money that considerably less than 25% of Mexican gulf crude oil comprises volatile hydrocarbons of 3Cs or shorter.

There is some margin for error here: longer hydrocarbons will evaporate even at 40C, they just do it very slowly. They also break down in the sea to become shorter more volatile compounds. But still, 25%… REALLY?!

Of course, that 25% isn’t only oil that evaporates, some of it dissolves. “Surely not!” I hear you cry, “oil and water don’t mix, how can crude oil possibly dissolve?”. Well, the short answer is that it doesn’t. Some hydrocarbons do dissolve but the relationship is the same as that for evaporation: some of the shorter ones might manage it, but the longer ones won’t. That means that we don’t have two fractions of the crude with which to get to 25%: our dissolution and evaporation are both coming from the same pool of molecules. A pool that I remain somewhat unconvinced makes up 25% of the total oil that has leaked.

Of course, none of my doubts would matter a jot if they had only published their data, but they didn’t. So I shall continue to doubt.

Then there is the oil that dispersed. They define “dispersed” as “droplets that are less than 100 microns”. Coincidentally, the human eye is capable of resolving objects down to around 100 microns. Translation: “if we can’t see it any more, it isn’t there”. I don’t need to tell you that, even if we can’t see it any more, this oil is still there and it can still enter the food chain and kill countless animals.

Their figure for oil that has dispersed is 16%. Length of the molecules doesn’t effect dispersion so maybe that’s a reasonable number, maybe it isn’t. However, I find it rather difficult to believe anyone can put any sort of meaningful number on it, given that we know there are huge plumes of stringy oily gloop floating below the surface that we can’t even come close to estimating the size of. If we don’t know how much oil there is or where to find it, how on earth can we ever hope to quantify it?

Which brings us to my final question: what exactly was the purpose of this report in the first place? The numbers are estimates, the authors say as much themselves; the figures don’t tell us where the oil is, or how to get rid of it; they don’t quantify the damage already done to this delicate ecosystem or try to predict how much is still to come; and they don’t give any clue as to how long the area will take to recover, or if it ever can. Worse than that, these are estimates based on such little data[5] that they are meaningless.

The only reason I can find for publishing such spurious and dubious data is so that President Obama can stand in front of the US press core in the White House briefing room and tell the world that the fight to contain the spill “is finally close to coming to an end“. I’m sorry Mr President, I don’t believe you.

I’ll leave you with one final staggering quote from the report. The person who wrote this travesty obviously felt that, rather than hard data and figures, what readers really needed was an explanation of the difference between dispersal and dissolution. They explain dissolution like this:

“Dissolution is the process by which individual hydrocarbon molecules from the oil separate and dissolve into the water just as sugar can be dissolved in water.”

No. No they don’t.

If your are interested in finding out some of the unspun real impact of this horrifying disaster, I recommend you check out TEDxOilSpill. I warn you now: it will depress you.

1. These numbers are based on government caca de torro budget best estimates from top scientists and researchers.

2. It is a shame they don7t then publish that information, however, because it would be useful to the environmental scientists and communities that are left trying to clean up the mess.

3. OK, not at all like the legs of a centipede, but the analogy will do for now.

4. Badum tsh.

5. And so much wishful thinking