Thursday, April 28, 2011

Tornado tragedies: science advances, but prediction of Nature still found wanting

Today's sobering headlines involve the recent spate of deadly tornadoes that have cut swaths of destruction across the South.  The death toll alone, not counting injuries and economic damage, is nearing 200.  Here we are in our technological age, with AccuWeather (located here at Penn State,  one of the Meccas of meteorology).  But people are still swept up to Oz.

I was for a few years a weather forecaster in the US Air Force.  That was some time ago, when technology was much less advanced than it is today.  We studied severe storms such as the tornadoes so typical of spring in the US.  They're due to the collision between warm, muggy air from the Gulf of Mexico, meeting cooler, drier air sweeping down from Canada (the first map below, from AccuWeather).  The denser air mass undercuts the warmer one, pushing it up, where it cools, condenses, and releases tremendous amounts of latent energy.  The whole system is steered rapidly along to the north and east by the jet stream, a strong wind high in the atmosphere, like corks floating along a river.

For a number of known reasons, these mass collisions also typically involve zones where air converges on a line or point.  In these places, swirling convergence leads to thunderstorms of majestic power, and they can spawn the trailing intense vortices that are the tornadoes.

All of this can nowadays be predicted in general--to an uncanny accuracy relative to what we could do in the bad old days.  But where the most intense energy will be released is still only probabilistically predictable.  Tornadoes are such local phenomena relative to, say, a line of T-storms hundreds of miles long (map below, from AccuWeather), and they are so brief (touching down usually only for a few miles), that the complexity of stormy turbulence does not make their specific occurrences predictable.  As can be seen, we in Pennsylvania also got some nasty T-storms during the night, and there are ominous skies outside my window right now, but so far we've been spared the tornadoes. 

Whether the desired precision can ever be achieved is one of the many questions in science related to understanding complexities in Nature.  Until it can, we may invest heavily and in general successfully in science technologies, but Nature's gremlins may still be able to hide between the bits and bytes even of the fastest computers.

And so long as that is the case, awful tragedies such as what has occurred in Alabama will still happen.  And meteorologists will struggle to do better against what may be theoretically prohibitive impossibilities.

1 comment:

Ken Weiss said...

Here is a Times story that discusses the prediction issues that we raised yesterday: