MELISSA BLOCK, Host:
There were of course two hurricanes that churned through the Gulf of Mexico in 2005, Katrina and Rita. They followed similar paths, but Katrina remained strong as it approached land while Rita faded. As NPR's Jon Hamilton reports, scientists say they're beginning to understand why the two storms behaved so differently.
JON HAMILTON: A hurricane is a bit like a figure skater doing a spin. Skaters revolve slowly when their arms and legs are extended, but pull in those arms and legs and the skater becomes a spinning blur. Robert Houze of the University of Washington says hurricanes rotate, too.
ROBERT HOUZE: The storm is a circular vortex swirling around very, very fast.
HAMILTON: And in a hurricane, he says, the equivalent of a skater's arms and legs is the wall of thunderstorms that surrounds the eye. As this eyewall contracts, the storm spins faster. Houze says that's what was happening with Katrina as it approached Louisiana and Mississippi. Just hours before landfall, it was still a Category 5 storm with winds of more than 160 miles per hour. An amateur storm-chaser managed to record those winds as they blasted through a multi-level parking lot in Biloxi, Mississippi.
(SOUNDBITE OF WIND)
HAMILTON: Rita was a different story. It had become a Category 5 storm out in the Gulf, but by the time it neared the coast forecasters at the National Hurricane Center said the news was getting better.
Unidentified Man: Rita remains a very dangerous hurricane, still Category 3 on the Saffir-Simpson scale. Maximum winds remain at 125 miles per hour.
HAMILTON: Remember that spinning skater extending her arms and legs? Houze says Rita had done something similar. The eyewall had gotten bigger.
HOUZE: A second ring of clouds formed, surrounding the inner ring of clouds.
HAMILTON: It eventually replaced them.
HOUZE: When that happens, the storm de-intensifies. It loses its punch rather quickly. Although it was still a very powerful storm, it was not as strong.
HAMILTON: If Rita had spent another day or two in the Gulf, the new eyewall would have begun to contract, causing the storm to speed up again. Houze says he and other scientists were able to look at both storms in detail because during that summer they were studying hurricanes using planes equipped with high-resolution radar. Their timing was perfect.
HOUZE: Katrina and Rita have given us almost a natural controlled experiment, because the storms followed very similar paths, yet Rita formed a secondary eyewall and Katrina didn't.
HAMILTON: And now they've got to figure out why this so-called eyewall replacement happens. There are some hints. For example, Rita had higher humidity near its center and lower humidity farther away. But even without knowing why a new and bigger eyewall might form, it should now be possible to create a computer model that includes this possibility. Hugh Willoughby is a researcher at Florida International University. He says the limitations of current models mean that forecasters still occasionally underestimate a storm.
HUGH WILLOUGHBY: Yeah, you're nervous about it, but it's not the monster storm of your nightmares. And then wham, 12 hours later or six hours later it's this nasty, devastating storm.
HAMILTON: Hurricane Andrew in 1992 behaved a bit like that. It went through an eyewall-replacement cycle out at sea. At first, the new, larger eyewall caused Andrew to slow down. Then, Willoughby says, the new eyewall began to contract.
WILLOUGHBY: It had completed the cycle, and it was re-intensifying as it came across the Gulf Stream and as it came on shore. That was what made it such a bad hurricane.
HAMILTON: Willoughby says forecasters are getting better at predicting the intensity of storms, and he says investing more of the kind of the technology used to study Rita and Katrina should help with future hurricanes. The new research appears in the journal Science.
Jon Hamilton, NPR News.
BLOCK: You can watch a narrated animation exploring the Atlantic's hurricane nursery at npr.org. You're listening to ALL THINGS CONSIDERED.