The Doppler radar planes are aloft for at most 12 hours a day. But the result is far more precise, with data collected every few hundred yards - sort of like having many more megapixels in a camera. This information is in turn fed into a sophisticated model that requires massive amounts of computing power.
Penn State meteorology professor Fuqing Zhang cautioned that no model can be evaluated on the basis of one storm.
But for what it's worth, Penn State's Doppler-aided model had Irene pegged pretty closely, correctly predicting that the storm would weaken by the time it reached New Jersey on Sunday morning.
At 8 p.m. Aug. 23 - the Tuesday before the storm - the Penn State model predicted that Irene would slow to about 80 miles per hour by 8 a.m. Sunday - soon after it was expected to hit New Jersey. The actual speed at that point was 75 miles per hour, Zhang said. The official government forecast had the wind speed above 100.
Franklin, the government forecaster, said Zhang's results from dozens of previous storms were "very promising."
"It may be one of our best hopes for trying to find models that can depict these kinds of changing conditions," he said.
Still unclear is just what happened to sap Irene's energy, but there are theories, Franklin said.
The strongest winds in a hurricane are found in a ring around the eye, called the eyewall. Often a second ring will develop outside the eyewall and start to contract, stealing energy from the original eyewall until it dissipates and disappears. The process then repeats.
But in Irene, it seems that multiple outer eyewalls developed, diffusing the available energy, Franklin said.
