Log in

No account? Create an account
Hurricane SCIENCE! - The Mad Schemes of Dr. Tectonic [entries|archive|friends|userinfo]

[ userinfo | livejournal userinfo ]
[ archive | journal archive ]

Hurricane SCIENCE! [Jan. 20th, 2006|04:26 pm]
I just got back from a really neat seminar about why hurricanes form and what the effects of global warming are.

The short answers:

Will global warming cause more hurricanes? No.

The number of hurricanes seems to be pretty constant -- 2005 was an unusualy active year, but there isn't a trend that correlates with sea surface temperature (SST).

Will global warming cause stronger hurricanes? Yes.

A global study of all the differnt ocean basis shows a statistically significant trend of increasing SST that correlates with an increase in hurricane intensity.

Aside: here we get into the disagreements. There are some prominent people who say that it's just due to natural variability in atmospheric or thermohaline circulation, but the speaker showed results from a couple papers (which are currently in press) that show (1) math (some crazy Bayesian multi-entropy thingy) says SST is the best explanation for most of the variation in hurricane intensity; and (2) the variations in thermohaline circulation actually have the wrong sign to account for it. On top of that, there are good theoretical reasons to expect SST to increase hurricane intensity, and numerical models also show it.
There are some NOAA climate variability patterns that appear to explain the hurricane variability patterns, but it turns out that those are coming from EOFs that were generated from data that wasn't detrended, which creates a spurious long-term oscillation mode. So, as they say, "the proponderance of the evidence" supports the idea that global warming increases hurricane intensity.

So what causes hurricanes to form? A "negative stretching deformation".

That's only part of it; we don't really know why they form because, well, it's a hard problem, but this is an important new piece of the whole hurricane genesis puzzle.

The traditional conditions for hurricane formation seem to be wrong, because the conditions occur far more frequently than hurricanes actually do, plus there's no physics behind several of them. The new criteria he theorizes:

A negative stretching deformation of the zonal flow (dU/dx negative) causes energy from Rossby waves to concentrate as the wave group velocity (which transports energy) drops to zero due to doppler shifting. In other words, you get these easterly waves coming off a continent, and when they hit a region where the winds are reversing, their wavelengths get shorter and the energy they're carrying gets all piled up in that spot. And having a whole bunch of energy in one spot in the atmosphere is what leads to hurricanes. So a "negative stretching deformation" is the atmospheric condition necessary for a hurricane to form, and in a way, that's "what causes it".

And I won't claim to totally understand it -- but I mostly followed, and it was a neat thing to be able to write about!

[User Picture]From: dcseain
2006-01-20 07:29 pm (UTC)
This is fascinating! :)I love weather-geek stuff. And i find the concept of negative streching deformation to be as eloquent as that of gossamer threads, which is still my favourite phrase in the field of DNA/RNA/Genetics.
(Reply) (Thread)
[User Picture]From: nematsakis
2006-01-20 08:12 pm (UTC)
So one thing I'm curious about: is it possible, in principle, to stop a hurricane? Specifically, when it is a young storm could we do something that would dissapate its energy or at least alter its course so it wouldn't hit land?
(Reply) (Thread)
[User Picture]From: backrubbear
2006-01-21 06:50 am (UTC)
I recall reading something recently about past experiments to affect the course of hurricanes. Unfortunately, it was a fly-by blurb on the Net so I don't have the reference.

The other comment I remember from that blurb was that their experiments were successful and a failure: They were able to radically affect the course of the hurricane. The effect of the course adjustment was exactly the wrong thing to do If they were willing to continue experimenting with the possibility of putting hurricanes on unsuspecting communities, they might actually get it right. :-)

If I understood one of the comments above, they might be able to collaspse/disrupt the energy/stress front of the relevant system prior to formation and prevent the hurricane from forming at all. What little I took from the text above is that the formation is somewhat similar (by analogy) to an earthquake forming by rubbing tectonic plates with a single area receiving the majority of the stress. Release some of the stress and you can affect it.

(Reply) (Parent) (Thread)
[User Picture]From: ocschwar
2006-01-21 07:05 pm (UTC)
I keep reading how serious windmill coverage actually does slow down the winds significantly. I wonder if anyone simulated the effect on a hurricane landfall if every mile of coast was covered with trifoil windmills.
(Reply) (Parent) (Thread)
[User Picture]From: dr_tectonic
2006-01-22 12:05 am (UTC)
In principle, probably. In practice, I doubt it.

Whatever you did to alter the distribution of water and air flow structures that create the hurricane is almost certain to have heinous side-effects that would likely (in the long run, maybe in the short run, too) be vastly worse than the hurricanes themselves. It's just the nature of dealing with anything at that macro-scale.

But I might be totally wrong. Maybe there's a very clever tweak you could do to nudge them just a tiny bit and make things nicer...
(Reply) (Parent) (Thread)
From: tolkiencub
2006-01-20 08:45 pm (UTC)

Mmm...near-chaotic science

I love this stuff.

And, as a geeky aside: "Lear's Daughters" by M. Bradley Kellogg has weather science as the central fact of the story. It's hideously sci-fi, but very, very fun to read. And, well - it informed a memorable year-long scifi campaign I'm running (the characters stumbled across the world of the novel, wound through the same mystery as the novel's characters, and moved on).

"The Wave and the Flame" and "Reign of Fire" are the 2 books.
(Reply) (Thread)
From: (Anonymous)
2006-01-20 09:17 pm (UTC)

stronger, faster, better

One of the things contributing to Katrina's incredible intensification off the LA/MS Gulf Coast was a deep pocket of very warm water. Hurricanes churn the ocean surface and so the upwelling of cooler subsurface water tends to stabilize the rate of intensification, but in Katrina's case that balancing aspect was minimized. Was there any discussion of causes & effects re: higher subsurface sea temperatures?

(Reply) (Thread)
[User Picture]From: psyclonic
2006-01-20 09:17 pm (UTC)

Re: stronger, faster, better

Dagnabit, I was so logged in!
(Reply) (Parent) (Thread)
[User Picture]From: dr_tectonic
2006-01-22 12:07 am (UTC)

Re: stronger, faster, better

Yeah, everybody's cookies got expired for a security issue.

He did indeed talk about the effect of subsurface sea temperatures, with several pretty color graphs explaining it all. The gist of it was mostly, "warm deep water causes stronger hurricanes". =)
(Reply) (Parent) (Thread)
[User Picture]From: bats22
2006-01-21 07:46 am (UTC)
it turns out that those are coming from EOFs that were generated from data that wasn't detrended

Does EOF=end of file in this case? Or something else?

So a "negative stretching deformation" is the atmospheric condition necessary for a hurricane to form

Um, is 'negative stretching' equivalent to 'squishing'? Or am I reading that paragraph wrong?

Cool stuff. As my coworker put it, this summer seemed like a season that somebody set the hurricane machine on 'automatic' and let it go...
(Reply) (Thread)
[User Picture]From: dr_tectonic
2006-01-22 12:17 am (UTC)
EOF = Empirical Orthogonal Functions. Also known as Principal Component Analysis, Karhunen-Loeve Transformation, Singular Value Decomposition, the Hotelling transform, and about fifty other things, depending on which field you're working in. Everyone discovers it independently and names it something new.

It's basically just finding the eigenvectors of the covariance matrix for your data. That gives you the "modes" of your system.

I think negative stretching is equivalent to squishing. It's a region where the sign of dU/dx is negative, where U is velocity in the x-direction. Except I think it's actually the flow anomaly, because we're working in some reference frame attached to the atmosphere, not the earth, or something like that, so the signs might be doing something weird.
(Reply) (Parent) (Thread)