Sophisticated computer modeling developed during the past 25 years allowed the National Weather Service to accurately predict the paths of Tropical Storm Isidore and Hurricane Lili in the course of a week. The forecasting on Lili's intensity went awry, however, as she approached the Louisiana coast.
One theory reported in Lili's aftermath was that Isidore had sucked the heat out of Gulf of Mexico waters, taking away one of the factors that contributes power to a hurricane. This might be a factor -- but the science is not yet exact enough to accurately predict a storm's intensity much in advance of it approaching shore, says James Franklin, a hurricane specialist with the National Hurricane Center and one of the experts involved in forecasting Lili.
"We understand intensity change a lot less well than we understand motion," says Franklin, who in 2001 was awarded the National Weather Service's highest honor, the Isaac M. Cline Award. "We have a general idea of what kinds of factors are associated with strengthening and weakening. But how they all play together in a given situation is not something that is well known."
Some cooler water near shore along the continental shelf might have contributed to Lili's diminishing -- but Franklin says that shouldn't necessarily be attributed to Isidore. There is a lot of warm water in the Gulf of Mexico and Lili was moving at a fairly rapid pace; cooler water becomes more of a factor when storms are moving slow. Also, the water in the northern Gulf does not have as much overall heat in it as in the central Gulf. "Whether that was the reason, I don't think we can say yes or no," says Franklin. "There are other possibilities."
Among the other possible factors: Lili contained a fair amount of dry air that may have been entrained into the system and contributed to the collapse. Another possibility is that there was some increase in vertical shear over the storm that would have sapped the storm's strength, but it didn't seem like a lot to meteorologists.
"There's a lot about the inner physics of a hurricane that goes on irrespective of what's going on in the environment," Franklin says. "The strongest winds in a hurricane occur in the eyewall. These eyewalls have certain behaviors. They are dynamic, not static. They change."
The eyewall of Lili was observed shrinking very rapidly and, like a figure skater pulling in her arms, it spun faster, increasing its intensity. Then the eyewall became tiny. In one of the last observations made before Lili made landfall, Franklin says, the eye was only about 7 miles wide. "[The eyewall] just disappeared. It's possible it got so small that it just collapsed."
So it might not have had anything to do with the water temperature, the shear or the dry air. "Unfortunately, while we have an idea of what some of these factors are, even after the fact it's very hard to tell which of those was most important," Franklin says. "Maybe it was a combination of two of them that all came together in a particular way."
Franklin, a Miami native and Massachusetts Institute of Technology graduate, has published more than 20 scientific articles on hurricanes. Among his notable accomplishments: the proof that coastal residents living in high rise buildings should flee rather than take shelter in the middle and upper-floors, because winds become more intense and dangerous the higher you go. When asked to rank the four theories for why Lili diminished, he put the internal shrinking process first, the theory that cooler waters close to shore took Lili's strength second, the presence of dry air third, and the shear fourth. "Or maybe, there's some other factor that we have not identified yet," he says.
Franklin points to Hurricane Opal, which hit the Florida panhandle in 1995 the night O.J. Simpson was acquitted. Opal, which has been studied extensively, behaved much like Lili.
"It had a very rapid intensification, and then a very rapid weakening, right before landfall," Franklin says. "We still don't understand why Opal intensified so rapidly."
One school of thought says Opal went over a warm ring of water in the Gulf and intensified. Another camp says the intensity came from influences in the upper atmosphere. "In most of these situations there isn't enough data to really analyze exactly what happened," Franklin says. "You have a lot of hypotheses, but no definite answers."
While the track projection of Hurricane Lili proved very accurate, some public officials in coastal areas may worry that the experience may make the public complacent in the face of the next storm. Franklin says that's the wrong lesson to take from Hurricane Lili.
"If people come away with an appreciation that there are uncertainties in intensity forecasting, there's certainly nothing wrong with that, because that is the state of the science," he says. "It really can work both ways."
Franklin said Lili exemplifies where the science is to date -- and where it falls short. The known factors are the big high and low pressure systems that can be seen easily by satellites in space. The physics are simple, and their effects on the storms are well known. "But the kinds of things that affect intensity are occurring on much smaller scales and are much harder to measure," he says. "I would say that there's very little immediate hope that we would have this problem solved in the next few years."
Franklin says a strong research effort is now needed to show how all the factors come together to form a hurricane's intensity. "Computer models for track have been great. But we don't have the information to put into computer models for intensity. ... We don't necessarily measure the things that would allow us to measure how the intensity is going to change."
For storms far from land, satellite photos are analyzed using what's called the Dvorak technique. The idea is that storms have certain appearances corresponding to different stages of development and different intensities. The appearance of the storm from the image is matched with a set of diagrams on file with the National Oceanic and Atmospheric Administration. "It's kind of like an ink blot test," Franklin says.
Then there are the hurricane tracking planes, which fly at 10,000 feet, crisscrossing through the eye and measuring the winds going through the eyewall, and measuring the pressure once they make it to the center. From that data, with statistical adjustments, the surface wind speed is estimated.
A third tool for measurement is a device that Franklin helped develop, called a "dropwindsonde" for "drop wind sounding," or "dropsonde" for short. Dropped from a plane, the device measures the winds all the way down to the earth's surface. If placed properly, the dropsonde can measure the actual wind speed about every 15 feet. But the devices -- which look like a Federal Express poster tube with a parachute -- cost about $500 and are not recoverable, so they are used very selectively.
"That's not necessarily what we need to know in order to predict how the intensity is going to change," Franklin says. "And I'm not sure anybody knows exactly what we need to measure. It's a very young science. We're not on the verge of an answer."
Given this uncertainty, will critics now accuse the agency of crying wolf, projecting a strong category 4 hurricane with 145 mph winds, only to have it make landfall as a category 2?
"Boy, I hope people would be happy that it weakened," Franklin says. "A category 2 hurricane is still dangerous."
Anytime you see a forecast for a category 3 hurricane, Franklin says, it could make landfall as a 2 or a 4 -- sort of like the margin of error in public opinion research. Indeed, 1992's Hurricane Andrew went from a category 3 to a category 5 in the hours right before it hit. "All hurricanes need to be taken seriously. If Lili had come in 12 hours earlier, it might have been during its peak intensification time. So it could have been worse than what we forecasted.
"We just don't know enough to put all the pieces together to predict that," he said. "That's why we make sure the emergency managers understand, if we tell you it's going to be a 3, it could be a 3, it could be a 2 or a 4."
Approaching shore, the land itself can have a positive or negative impact on a storm. Once it gets over land, it is robbed of fuel. In addition, friction under the storm is increased, slowing it down. If you had wind sloping down from high land into the storm, like the California cliffs, that would tend to dry out the air and kill off thunderstorm activity, making the storm weaker. Barrier islands, cliffs or dunes can decrease the storm surge, but the flat, low, wetlands in coastal Louisiana do not offer much of an impediment to slow storms down.
"The land was not a player here," Franklin says. "I suspect we will never understand why this storm did what it did."
- National Weather Service