Skip to main content
  1. Home
  2. Emerging Tech
  3. News

A decade after Katrina, meteorologists say hurricane models vastly improved

Add as a preferred source on Google

Meteorologists are getting better at predicting hurricanes because our ability to model the atmosphere has improved dramatically. And while you’d think NASA is more concerned with getting astronauts into space, that’s only part of its job. The agency actually has a team at the Goddard Space Flight Center in Greenbelt, Maryland working tirelessly day and night to improve those forecasts, and it has seen much success over the past decade.

Every forecast you see these days is not only the product of a human meteorologist, but also a computer. Modern forecasters use computer models of the atmosphere to help them figure out what the weather will do next. With hurricanes, that’s all the more important because it could be the difference between life and death.

Recommended Videos

“Freshwater floods, often caused by hurricanes, are the number one cause of death by natural disasters in the world, even above earthquakes and volcanoes,” Goddard tropical meteorologist Oreste Reale said. “Seeing how the research we do could have an impact on these things is very rewarding.”

One of these tools is something meteorologists call “reanalysis.” Essentially, programmers will make changes to a weather model, and then put weather data from the past into it. From there, they’ll run the model, and the gauge of success is how well the model forecasts the actual conditions that followed the point in time they picked.

The strategy here is to use past major hurricanes, some of which weren’t well-modeled the first time around. “We are able to use cases like Hurricane Katrina to run tests and show us how we can improve, or how this new change affected the forecast or the analysis of the storm system,” fellow Goddard tropical meteorologist Marangelly Fuentes explained.

Katrina is a great example of model error. Atmospheric pressure is typically a good indicator of storm strength. In 2005, computer models forecasted Katrina’s central pressure would reach 955 millibars (28.20″) before striking the Gulf Coast. Observed data showed its pressure fell to 902 millibars (26.64″), one of the lowest pressures of any Atlantic hurricane ever.

Reale and Fuente’s modeling can now get much closer to that actual pressure in attempting to model the storm, which in turn produces better forecasts. They are also able to model at a much better resolution, further increasing accuracy.

Think of it this way. Cheap cell phone cameras produce images that are sometimes coarse and not sharp. Higher-end cameras will give you a better picture because they have higher resolution, and it’s the same way with weather models. In 2005, the atmosphere was split into blocks of 31 miles by 31 miles: today, that’s down to 8-mile-wide squares.

“The smaller the size of the cube, the more realistic the representation of the atmosphere,” Reale said.

While there are drastic improvements to our modeling capabilities, there is still work to be done, but Reale says this massive job of improving the capabilities of our hurricane models should result in better hurricane forecasts. And if the next Katrina comes along, we’ll be far more ready than we were over a decade ago.

Ed Oswald
For fifteen years, Ed has written about the latest and greatest in gadgets and technology trends. At Digital Trends, he's…
Students who use AI every day may be falling behind, while moderate users pull ahead
New global testing data suggests more AI in the classroom isn't automatically better.
Logo, Text

I assumed more AI access might help students learn faster. However, a new OECD study dispels that misconception for me and everyone else who gave their children a paid version of AI chatbots, thinking it would help them through their educational journey. 

According to PISA 2025 data from 760,000 students in 91 countries, students who use AI daily score meaningfully worse on standardized tests than their peers. The full picture is more complicated than a simple ban would fix.

Read more
XPENG just put its humanoid robot on the production line, and it walked itself out
IRON clocks in for its first shift
XPENG Iron

XPENG’s humanoid robot has officially joined the workforce, and it even got an employee badge. The Chinese automaker announced on September 8 that its next-generation IRON had completed assembly and autonomously walked off its newly commissioned humanoid robot production line in Guangzhou.

CEO He Xiaopeng then pinned a staff badge onto the robot, symbolically welcoming it to the XPENG team. It is a charming little ceremony for a machine that the company hopes will eventually become a regular part of everyday life. XPENG is moving IRON beyond R&D prototypes toward production-line manufacturing, applying the quality-control systems and manufacturing experience it has developed through its electric vehicle business.

Read more
An AI-designed drug might help slow aging, but we’re not there yet
The results hint at a possible anti-aging effect, but there’s still a long way to go before we know what it means.
medical test tubes

Rentosertib, a drug Insilico Medicine designed almost entirely using AI, was created to treat idiopathic pulmonary fibrosis, a disease that scars and stiffens the lungs over time. But fighting lung disease was only half the plan from the start.

According to a new study published in Nature Biotechnology, the drug's target was chosen specifically because it overlaps heavily with the biology of aging, and the whole trial was designed to test whether treating the lungs could also mean turning back the body's biological clock.

Read more