ROBERT SIEGEL, HOST:
This past week, tropical storm Erika was a reminder of the difficulty of forecasting hurricanes. Forecasters said they expected the storm would intensify into a hurricane with a possible U.S. landfall. Over the weekend, Erika defied those forecasts and dissipated, still bringing heavy rain and flooding to parts of Florida and South Carolina. NPR's Greg Allen reports on the push to improve hurricane forecasting, a push that began after Hurricane Katrina.
GREG ALLEN, BYLINE: Ten years ago, as Hurricane Katrina approached the Gulf Coast, the biggest question was where would it come ashore?
(SOUNDBITE OF ARCHIVED RECORDING)
UNIDENTIFIED REPORTER #1: Here's the path. This is what we're forecasting.
ALLEN: Two days before Katrina hit land, a Weather Channel meteorologist gave the forecast. A huge area, some 500 miles of the Gulf Coast, was potentially in the storm's path.
(SOUNDBITE OF ARCHIVED RECORDING)
UNIDENTIFIED REPORTER #1: We're talking about from Eastern Louisiana eastward, including you in the New Orleans area, across the coast of Mississippi and Alabama.
ALLEN: Ten years later, new technology has improved the accuracy of tracking by 40 percent, reducing errors shrinking the area put on alert and potentially evacuated because of an approaching hurricane. And now, instead of three-day forecasts, the National Hurricane Center in Miami is issuing five-day forecasts. Louis Uccellini, the head of the National Weather Service, says improved tracking has helped coastal communities and emergency managers prepare for storms.
LOUIS UCCELLINI: If we can give them a smaller area with increased confidence, they can really reduce the cost of that evacuation and make the evacuation more effective. So these track improvements have been extremely important.
ALLEN: After the active hurricane seasons of 2004 and 2005, the National Weather Service decided it needed to improve hurricane forecasting. It launched a project that led to the adoption of new technology like drones, upgraded weather satellites and better data gathering instruments on the aircraft that fly into hurricanes.
Ten years ago, Katrina highlighted another challenge for meteorologists - the difficulty of forecasting a hurricane's strength. Here's a Weather Channel report on Katrina the day before it made landfall.
(SOUNDBITE OF ARCHIVED RECORDING)
UNIDENTIFIED REPORTER #2: Hurricane Katrina is a Category 5 hurricane. That means catastrophic damage to any city in its wake.
ALLEN: By the next day, Katrina surprised forecasters by weakening and making landfall as a Category 3. Even so, because of its massive size and storm surge, it still was the most destructive hurricane ever. One year earlier, Uccellini says, Hurricane Charley was another example of a storm that showed the difficulty of forecasting intensity, building from a Category 2 to a Category 4 storm very quickly just off the Florida coast.
UCCELLINI: People who were expecting to deal with a 2 found themselves engulfed in a landfalling hurricane. That was a wake-up call.
ALLEN: Chris Ruf, a professor of atmospheric science at the University of Michigan, says one of the challenges in forecasting intensity is gathering the data from a hurricane as it develops.
CHRIS RUF: The things that cause intensity to change happen in the inner core of the hurricane, right near the eye and in the surrounding intense part of the storm.
ALLEN: Ruf is working with NASA on a project that he says will improve the ability of forecasters to get that information and predict rapid changes in intensity. By next year, they'll launch a group of microsatellites that will use GPS signals to measure wind speeds in a hurricane's inner core.
Just three years ago, Hurricane Sandy was what Uccellini describes as another watershed moment for the National Weather Service. Several days out, a European computer model correctly protected the storm's East Coast landfall while U.S. models still had it turning out to sea. Since then, the National Weather Service has been playing catch-up. After Sandy, it scrambled to upgrade its models and data crunching.
Better models have now improved intensity forecasts by 20 percent. More progress is expected by the end of the year when the National Weather Service says it will quadruple its current computing capacity. Greg Allen, NPR News, Miami. Transcript provided by NPR, Copyright NPR.
