In the 20 years since Hurricane Andrew struck Florida, building codes have improved dramatically. So has hurricane forecasting--at least in some ways.
Scientists are getting better at predicting where and when storms are heading. Figuring out how strong the storms will be still needs work.
"The errors in one generation--just 20 years--have dropped in half," says Chis Landsea, PhD, Science and Operations Officer at the National Hurricane Center. "So if we make a two-day forecast and we say the hurricane is coming toward Tampa, 20 years ago that would have been all the way up to Cedar Key down to Key West. And now that's only about 75 or 100 miles north and south of Tampa Bay."
A smaller prediction cone means a smaller evacuation area and the potential of saving millions of dollars. All of that savings would be erased though if the prediction tracks are wrong.
"Our errors continue to get smaller and smaller," says NHC Hurricane Specialist John Canglialosi. "And these are big-time improvements. These are not small. For example, ten years ago, our error at three days is what it is today at five days. So these advancements are very significant and that's thanks to the computer models that we use."
What still needs work is predicting how strong a storm will become.
"Part of the difficulty with hurricane intensity or wind forecasting is--a--we don't understand completely how it works," says Chris Landsea. "We don't know all the physics and we don't observe it very well. Satellites don't give us a whole story, nor do the aircraft. And the computer modes are just too simple compared to the real world."
This year the National Hurricane Center is testing new computer models to help with intensity predictions and working to extend its forecasts to six or seven days out.