“All thy trees and fruit of thy land shall the locust consume.”
July 27, 2010 3:37 PM   Subscribe

Math Is No Match for Locust Swarms. "Mathematicians have now figured out the dynamics that drive locusts across the landscape, devastating everything underfoot — and the math says people will never be able to predict where the little buggers will go. The new analysis, reported in an upcoming issue of Physical Review E, suggests that random factors accumulate and influence how swarming locusts collectively decide to change course. “These swarms are driven by intrinsic dynamics,” says team member Iain Couzin, a biologist at Princeton University. “In all practical terms, predicting when a swarm is going to change direction is going to be impossible." More information here.
posted by Fizz (27 comments total) 3 users marked this as a favorite
 
Between this and the ants the front page is making me very itchy today.
posted by The Whelk at 3:42 PM on July 27, 2010 [2 favorites]


s/locusts/people
posted by swift at 3:42 PM on July 27, 2010


I'd like to run similar experiments on swarms of kittens. Not just because I'm curious if it would be different for mammals with a different social organization, but also because it would be adorable.
posted by Tomorrowful at 3:45 PM on July 27, 2010 [5 favorites]


Kitten-Maru Damaci! Giant kitten ball rolling across the world collecting cuteness as it goes.
posted by Babblesort at 3:47 PM on July 27, 2010


Further analysis showed that a number of random factors influence when the insects decide to change direction. Mathematically, the change in the locusts’ direction is similar to switches in magnetic properties that occur among clumps of magnetic particles at high temperatures, Escudero says. In both cases, random influences accumulate until suddenly the whole system changes its behavior.

Peirce was right.
posted by HP LaserJet P10006 at 3:47 PM on July 27, 2010


Between this and the ants the front page is making me very itchy today.

No joke. Is it Bug Horror Day or sumthin?
posted by angrycat at 3:52 PM on July 27, 2010


Good news, everyone! I've invented a device that will predict locust swarm behavior with 50% accuracy! ::flips coin::
posted by Faint of Butt at 4:01 PM on July 27, 2010


This sounds pretty standard, like the same observations about the position of Hyperion or weather patterns. Basically, the complexity of the interactions ensures that measurements taken can't be used to predict future behavior.
posted by Deathalicious at 4:02 PM on July 27, 2010 [1 favorite]


Is it Bug Horror Day or sumthin?

I admit some degree of culpability in that I posted the ant swarm, article, but locusts are adorable! Look at them hop! Boing! Boing! Boing!
posted by infinitywaltz at 4:04 PM on July 27, 2010


I saw this guy talk last week, and it was awesome. He is a total badass. In addition to being an incredible scientist in the lab, he's willing to take to the field and spend several months in the desert in order to see if his computational models work in real life.
posted by solipsophistocracy at 4:28 PM on July 27, 2010 [1 favorite]


I'd like to run similar experiments on swarms of kittens.

Kittens don't swarm. Thus the comparison of doing impossible things to herding cats.
posted by GuyZero at 4:31 PM on July 27, 2010 [1 favorite]


This kind of story is kind of frustrating for me. The paper isn't out, so I can't read it. But the story is very vague, and sounds like a familiar tune to me. Statistical mechanics has been very good at describing systems without knowing the micro detail of the systems' smallest components. This kind of approach is the study of critical phenomena. Systems that exhibit critical phenomena have phase transformations: gasses and magenetic materials are common examples. What's amazing is that you don't need to know the tiny details (what shape the gas molecule is, the origin of the magnetism, etc). You just need to extract a few system-wide parameters, like a critical temperature, and you can predict everything about the system. It's a tremendously powerful tool. So the temptation is to go looking to see if that tool works elsewhere...

...enter biology. I've personally seen this tool used inappropriately here, in fact it's very common. The warning signs are discussions of the "physics" of inherently large-scale biological phenomena, discussion of "phase transitions", "power laws", or "avalanches". Also weirdly-reduced problems (like this one, in only one dimension). Top-down descriptions (we modeled with a Fokker-Plank equation, we just need to work out the math). And worst of all, the trumpeting of large take-away messages (swarms are inherently unpredictable) when the details are still unknown.

This research could be totally on the up-and-up, but there isn't enough to go on from what's linked.
posted by Humanzee at 4:45 PM on July 27, 2010 [9 favorites]


Kittens don't swarm.

The internet begs to differ.
posted by filthy light thief at 5:01 PM on July 27, 2010


filthy light thief, that is the most terrifying thing I've seen on the internet all day.
posted by Fizz at 5:02 PM on July 27, 2010


Kittens can swarm. They just won't swarm in any direction you particularly want them to go (i.e. herding.)
posted by kyrademon at 5:05 PM on July 27, 2010


When life hands you locusts, declare 'em kosher and eat them!
posted by hermitosis at 5:16 PM on July 27, 2010


LADEEZ AND GENTLEMEN! THIS A-THIS CORNUH wearing the href shorts!

Math Is No Match for Locust Swarms.

While in THIS CORNUH wearing the Blue shorts!

Mathematicians have now figured out the dynamics that drive locusts across the landscape.

Let's get readyyyyy tooooo RUUMMMMMMBLLLLLLE!!!!!!
posted by DU at 5:27 PM on July 27, 2010 [1 favorite]


Random factors? That's what scientists say when they can't figure out how something works.

/Laplace
posted by MarshallPoe at 5:51 PM on July 27, 2010


MetaFilter: Random factors accumulate and influence how swarming locusts collectively decide to change course.
posted by ZenMasterThis at 6:34 PM on July 27, 2010 [1 favorite]


So there's sensitive dependence on initial conditions but strangely there is no strange attractor.
posted by localroger at 8:11 PM on July 27, 2010


"Accumulation of many random factors"

Mmmhmm. Even if the factors _are_ random, there is no reason they have to _remain_ random. Perturbation of random systems is something of a sport from where I'm from.
posted by effugas at 8:20 PM on July 27, 2010


Gaia hypothesis, climate change, randomness, string theory, lack to ability to cogently express myself from lack of sleep, itchy eyes: all variables.

I'll sort it all out tomorrow.
posted by eegphalanges at 8:24 PM on July 27, 2010


Chaos theory, that is. Tomorrow, really, figure it all out. Make cogent argument, reference popular culture, make someone feel small. Sleep now.
posted by eegphalanges at 8:27 PM on July 27, 2010 [1 favorite]


Kittens definitely do swarm, as filthy light thief's link indicates. Can't herd 'em? Right, because they swarm in mathematically unpredictable ways. Like locusts.

Adorable, fuzzy, lovable locusts.
posted by Tomorrowful at 8:57 PM on July 27, 2010 [1 favorite]


Why are those kittens being fed on a white carpet? And why is it still white?

I too find the article unsatisfying. "Locusts swarm and we don't know how to predict where to" is old! What's the newsworthy part?
posted by Omnomnom at 1:44 AM on July 28, 2010


One of my first chaos screen savers (this was in the early 1990's) was a simulation of a flock of starlings darting this way, then that. Perhaps I will post an ask metafilter and see if anybody knows where I can find one if google will not provide it.
posted by bukvich at 4:45 AM on July 28, 2010


There's a difference between "we don't know how to predict it" and "we know that we cannot predict it".
posted by Mwongozi at 5:21 AM on July 29, 2010


« Older Ok, you are. But not you.   |   Little Drifters Newer »


This thread has been archived and is closed to new comments