Today’s exercise reveals a case where no particular knowledge of science is needed to assess the story – mere basic logic and grade-school math are the only tools required.
And now, in a new study in Nature Climate Change, Princeton’s Ning Lin and MIT’s Kerry Emanuel demonstrate that when it comes to three global cities in particular — Tampa, Fla., Cairns, Australia, and Dubai, United Arab Emirates — there could come a storm that is much worse than anything in recent memory (or in any memory).
Granted, these theoretical storms are also highly unlikely to occur — in some cases, they are 1-in-10,000-year events, or even rarer. The researchers refer to these possible storms as “gray swans,” riffing on the concept of a “black swan” event, an unpredictable catastrophe, or highly impactful event. A “gray swan,” by contrast, can indeed be predicted, even if it is extremely rare.
On a logical level, saying something could happen isn’t saying much – it’s not a logical impossibility. A square circle is a logical impossibility, but it could be true that Genghis Khan built a gigantic base on the dark side of the moon from which his decedents will launch an attack on earth marking the 1,000 anniversary of the sacking of Baghdad. It’s absurd, but not logically impossible.
The devil, as they say, is in the details. No sane person, if he gave it any thought, would think it impossible that there might occur a storm greater than any they have personally seen or even hear of. So what we’re doing here is trying to layer a little math on this commonsense idea – it could be worse – and lay down some odds. So, how do we come up with those odds?
The purpose of the study is “to raise awareness of what a very low probability, very high impact hurricane event might look like,” said Emanuel. The gray swan storms were generated by a computer model that “coupled” together, in the researchers’ parlance, a very high-resolution hurricane model with a global climate model. That allowed the researchers to populate the simulated world with oodles of different storms.
OK, then. Sounds like something a couple grad assistants might fiddle with for kicks and giggles: Take model A, which incorporates a bazillion assumptions and presumes a gazillion dependencies, and use it to feed modal B, which may or may not use the same, or may even use mutually exclusive, assumptions and dependencies, and see what comes out. Good times, on the level of a purely intellectual exercise. Just don’t try to make predictions in the real world from it.
Of course, if one or both of the two models have been shown to consistently fail to match reality, we would add yet another level of caution. No sane person would dream of announcing the results of such an exercise as if it were a scientific finding of some sort, right?
Must be a really slow news day. Or maybe science isn’t the point?
One other note: the time frames they are talking about are large – these mega storms are something like 1 in 200,000 year probabilities. We’d expect one or more glaciations – ice ages – in that time frame. The return of miles-thick ice sheets covering all of Canada, some of the Midwest and much of Northern Europe and England, with a 300’+ drop in sea levels, for millennia at a stretch, would be bad, way worse than any storm. So, I’d say: let’s not sweat the storms too much.