Here’s How Much Better Hurricane Forecasts Are Today Than Decades Ago

Faster computers, a better satellite, and improved scientific understanding of storm dynamics have brought about a revolution in hurricane forecasting.

As the United States braces for its second major hurricane in less than a month, forecasters are again scrambling to get a better read on its path.

While still battering the Caribbean islands with top winds of 185 mph, Hurricane Irma is headed roughly toward the US mainland. With no landfall expected there until the weekend, the National Hurricane Center warns that “a fair amount of uncertainty” remains about its course. But the center has reams of data that it hasn’t had in decades past, and their estimates have improved markedly in recent decades.

to gather readings. Since the 1960s, satellites have been the primary tool forecasters use to get a read on a brewing storm, with the first of a new generation put into orbit in December. And since the 1970s, advances in computing power have given scientists the ability to run more and more sophisticated projections of tropical cyclone behavior.

“Bit by bit, these incremental improvements — a faster computer, a better satellite, knowing how to use that data, learning a little more about the inner science of the hurricane, and being able to put that into the computer models — all of these little things together have led to a revolution in how we can forecast these storms,” said David Titley, the US Navy’s former chief oceanographer.

In the 1980s, the average difference between forecasters’ projections of landfall and where a storm actually hit was around 400 miles (640 km), said Titley, who now teaches meteorology at Penn State. Today, the average difference on a three-day forecast is about 70 miles — and that’s been cut by roughly half in the last decade, he said.

“That would be, for the state of Florida, like saying we don’t know if a hurricane is going to go to Miami or Jacksonville,” Titley said. “Now we’re wondering about Miami versus Fort Lauderdale.”

Those improvements are a big advantage for emergency managers, mayors, or governors who have to make decisions about who should evacuate ahead of a storm, said Bill Read, a former director of the National Hurricane Center in Miami.

“Forty-eight hours is critical,” Read said. “For a whole lot of cities and populated areas, that’s about when they start making decisions on evacuations. South Florida is even more difficult. They’re having to make a decision today, when it may be three days before the actual impact. It’s so difficult to get people out of harm’s way down there, because there’s so many of them.”

Mandatory evacuations had already been issued Wednesday for the Florida Keys, while Florida authorities were moving people with special needs out of the Miami area and urging others living in low-lying areas to start leaving.

RELATED: A ‘Hot Spot’ for Sea Level Rise Whips Up Trouble in the Southeast

The projections that guide those decisions are based on streams of data that get denser every year. Hurricane hunter flights now carry Doppler radar units that produce a vertical cross-section of a storm and microwave imagers that can calculate wind speeds from the behavior of droplets of ocean spray, Read said. They also drop a parachute-equipped pod of sensors called a dropsonde, which transmits windspeed, pressure, temperature, and humidity readings.

“In fact, last night, a flight managed to successfully drop one into the eye wall, where the strongest winds are. Several hundred to 1,000 feet over the surface, it was measuring winds over 200 miles an hour,” Read said. “It was much less than that right at the surface, but it’s still pretty amazing data.”

Forecasters also run numerous computer models under different scenarios to increase their confidence, Titley said.

“No matter how good our observations are, we can never perfectly capture the atmosphere,” he said. But by running 20 to 50 simulations with slightly different initial conditions, “It gives an idea of the probability of a storm going someplace — and that’s very useful if you’re an emergency manager.”

RELATED: Climate Change Is Projected to Worsen Inequality in the United States

Irma’s unwelcome approach comes less than two weeks after Hurricane Harvey hit Texas, killing at least 60 people. The category 4 storm flattened towns on the Texas coast and dropped more than four feet of rain on parts of Houston, the fourth-largest US city and the heart of its oil industry.

Amid the misery, however, each storm provides experts a chance to fine-tune their models for the next disaster. Among the questions that have yet to be answered are why storms like Harvey intensify rapidly, said Ghassan Alaka Jr., who works on computer models for the Hurricane Research Division of the National Oceanic and Atmospheric Administration.

Harvey formed as a tropical storm on August 17, weakened into a tropical depression, then roared back to life in the Gulf of Mexico before its August 25 landfall. And for the first time, hurricane hunter planes were beaming data back to Alaka’s shop as they nosed into the storm.

“This is a big advancement for us on the research side, because we can almost in real time understand what impact these observations are having on the runs — and then dig in a little bit, starting now and on into the future, why those observations make the model better or worse,” Alaka said.

After a storm passes, he said, “We start to dig into these forecasts and understand, for example —looking at Harvey — why did this rapid intensification happen the way it did, and how did it stall out and create those massive rainfalls.”

RELATED: Here’s How Nearly All Nations Could Run on Renewable Power by 2050

In its latest advisories, the Hurricane Center still warns that forecasts four or five days out can be off by 175 to 225 miles. But Titley said better data and modeling has put an end to one tradition.

“Thirty years ago, there was something called the 1,000-mile club,” he said. “It was pretty easy to miss an individual forecast by over 1,000 miles … and that was a 3-day forecast. And yes, I belonged to the 1,000 mile club.

“Nowadays, the average error on those three-day forecasts is less than 100 miles,” Titley added. “So if you belong to the 1,000 mile club, you’d probably be fired.”

WATCH: What We Learned From Hurricane Katrina