Thanks To Technology, We're Getting Better At Predicting Hurricanes
In 2012, Hurricane Sandy was roaring up the East Coast of the U.S. when it suddenly made a detour, heading straight west to make a cataclysmic surprise landfall in New Jersey. In 2017, Hurricane Harvey was predicted to touch down in Texas, but what experts thought it would do next was anybody's guess. With so much death and destruction left in their wake, why don't we know more about the behavior of hurricanes? It's because nature is very hard to predict — but new technology is helping.
Hey Mister, Where You Headed?
Hurricane Andrew had reached Category 5 when it made landfall in Florida in 1992. It killed 15 people, leveled 100,000 houses in Miami-Dade County alone, and racked up $26 billion in damages, making it the costliest storm to hit the U.S. until Hurricane Katrina. At the time, scientists only had two ways of measuring a hurricane's strength: analyzing wind speed with an aircraft, and comparing satellite images of the hurricane to those of similar hurricanes in the past. To predict what the storm might do next, scientists used statistical models that relied on finding patterns in past storms in the hopes that the current one would follow them. All that technology gave them the ability to make a rough three-day forecast of where hurricanes like Andrew were headed, with an average error of about 300 miles (480 km).
But after Hurricane Andrew, scientists and government officials realized that in order to save lives in future hurricanes, we had to do a lot better. That led to a boom in research and regulation changes that have improved our ability to warn people in hurricane-prone areas of impending danger.
Hurricane Tracking Today
The National Weather Service unveiled new computer models, including HMON and HWRF, that rely on supercomputers and dynamical approaches that predict things based on current events rather than past storms. These new models can extend the possible hurricane forecast from three days to five days. They narrow the range of uncertainty, too — whereas the average error in 1992 was 300 miles (480 km) at three days out, in 2017, the error is about the same at five days out. At three days out, it's now less than 100 miles (160 km). The hope is to extend the forecast out even further and give people a full seven days' warning.
One of the most costly elements of a hurricane isn't the wind or rain, but the ocean swells known as storm surges. They can send flooding far inland, often in places well outside the hurricane's warning area, and have been responsible for about half the deaths associated with tropical cyclones in the last half century. But most people have barely heard the term, much less how dangerous storm surges can be. To fix that, the NHC began issuing storm surge warnings alongside maps that show all of the areas that could be reached by potential flooding. They're also issuing warnings of high winds further in advance.
But there's still a lot we don't know. Hurricanes can change strength without warning — the winds of Hurricane Andrew sped up from 75 mph to 165 mph (120 km/h to 265 km/h) in just 36 hours — and their paths can shift in confusing ways. Any change in water temperature or surrounding winds can disrupt a hurricane's pattern, and that complex mix of influences makes a hurricane's intensity hard to predict. "Our ability to predict intensification is not quite there yet," Frank Marks, Director of NOAA's Hurricane Research Division, told National Geographic. "We're getting there."