Ants don't read the barometer

Ants swarm when the pressure drops. They can feel the storm coming. Ask around and somebody will tell you this with real confidence, and they'll have the memory to go with it: a close, heavy afternoon, thunder somewhere behind the houses, ants pouring out from between the paving stones.

It's a good story. It has the shape of a true thing. Insects are sensitive to all sorts of stuff we can't feel, pressure is exactly the kind of invisible signal you'd expect an animal to be reading, and the anecdote fits.

Then somebody went and counted.

The Royal Society of Biology ran a public flying-ant survey across 2012, 2013 and 2014, and Hart and colleagues published the analysis in Ecography in 2018. It's still the largest thing of its kind: 13,394 usable records of members of the public reporting flying ants, 88.5% of them the black garden ant, Lasius niger.

With that many records you can start asking which bits of weather actually separate a day when ants fly from a day when they don't. Temperature came back enormous. Wind came back large. Pressure came back as nothing at all.

Flights clustered above about 13 °C, and just about every day with a mean temperature over 25 °C put ants in the air somewhere in the country. Wind cut them off fairly hard past 6.3 m/s, which is 14 mph, and you can see the logic there. A queen who launches into that isn't going to come down anywhere she chose.

Pressure gave a chi-square of 0.2 and a P-value of 0.687. That isn't a weak effect. That's roughly the number you'd get from feeding the model a column of random noise. The rate of change of pressure didn't do anything either.

Because the belief is very nearly right, and it's wrong in a way that's genuinely hard to spot.

The conditions that do set off a flight (warm, humid, not much wind, often a day or two after rain) show up in exactly the sticky, close, thundery weather that runs ahead of a summer storm. Pressure happens to be falling at the same time. It's in the room while the flight happens. It just isn't doing any of the work.

And it's an easy mistake to make, not a stupid one. The barometer is the part of that afternoon you can put a number on, so it collects the credit for everything else in the room.

The way you'd catch it is to look for the days where the two things come apart. Warm, still, humid, with the pressure sitting flat: you get ants anyway. Falling pressure with a cold gusty wind behind it: you get nothing.

It would be easy to put pressure in. Every weather API hands you the number for free. It costs nothing. And users would nod along, because it matches what they already think they know about ants.

We don't use it. A variable that has never been shown to do anything doesn't make a forecast more accurate. What it does is make the forecast sound better researched, which is worse than useless, because that's the sort of wrong nobody catches.

Dealate scores flight odds from temperature, humidity, wind, rainfall and the season, weighted differently for different species, because that's what the literature will actually support. Pressure isn't in there and it isn't going in.

The reasoning is public, and the papers underneath it are all cited, so the argument is there to be checked rather than taken on trust.

Hart, A.G., Hesselberg, T., Nesbit, R. & Goodenough, A.E. (2018). The spatial distribution and environmental triggers of ant mating flights: using citizen-science data to reveal national patterns. Ecography 41(6): 877–888. DOI 10.1111/ecog.03140. Open access.

Boomsma, J.J. & Leusink, A. (1981). Weather conditions during nuptial flights of four European ant species. Oecologia 50(2): 236–241. Where the idea comes from that queens fly when the air temperature catches up with the soil.