BOULDER — As Texas and Louisiana continue to recover from Harvey and as Irma’s remnants move farther into the U.S. mainland, some of the federal government’s top weather experts are already analyzing the once-massive hurricanes to better prepare for those that follow.
Billions of data points from the two high-profile storms — along with many others that struck in previous years — are being fed into burly supercomputers at the National Center for Atmospheric Research in Boulder so that future weather models and meteorological forecasts can be even more granular and precise when new hurricanes barrel ashore, threatening lives and property.
“We’re at the forefront of the science,” said Rebecca Morss, deputy director of the Mesoscale and Microscale Meteorology Lab at NCAR.
It’s a science that has achieved notable advances in just the past decade or so as computing power and expertise have increased, said Morss’ boss and the lab’s director, Christopher Davis. NCAR has approximately 500 scientists, engineers and software engineers — mostly in Boulder — who work to better understand and predict hurricanes and flooding, as well as other weather events.
Andy Cross, The Denver PostChristopher Davis, Director Mesoscale and Microscale Meteorology Laboratory at National Center for Atmospheric Research in Boulder Sept. 12, 2017.
Hurricane Irma was seen as a potential threat to Florida nearly two weeks before it made landfall along the state’s Gulf coast with gusts as high as 135 mph — a far earlier warning than was possible 10 years ago, Davis said. The models and data from the storm’s infancy were passed on to the National Weather Service, which disseminated them to hundreds of local meteorologists across the country.
“What has changed tremendously — and what is not obvious to the public — is the tools we have access to that have revolutionized the process since then,” Davis said. “We’re enabling the next generation of forecasting to improve.”
Not too long ago, computer forecast models displayed a storm’s characteristics in choppy, 400-kilometer blocks, said Bill Mahoney, interim director of NCAR’s Research Applications Laboratory. Now the resolution has improved to 12 kilometers, allowing scientists to see ever-greater detail and nuance.
On the ground, the hope is that better forecasts translate to more orderly and more effective preparation and evacuation protocols and fewer casualties in the wake of a hurricane, Morss said.
“That greater lead time helps with emergency response and allows evacuations to take place before traffic gets too bad,” she said.
Even so, NCAR’s scientists concede that as powerful and comprehensive as their computer models have gotten, Mother Nature can still keep them on their toes.
With Hurricane Harvey last month, few models anticipated just how rapidly the storm would blow up into a Category 4 hurricane and then suddenly slow to a crawl once it hit land, allowing it to dump 40 to 52 inches of rain in the span of a couple of days in and around Houston.
“It was a very challenging prediction problem,” Davis said.
And last weekend, Irma’s last-minute western pivot spelled some relief for Miami residents but caught people by surprise in the Florida communities of Naples, Clearwater and Tampa.
Mahoney said forecasters’ ability to determine Irma’s track as well as they did was no small feat, given that atmospheric and hydrologic phenomena from around the world affect how hurricanes ultimately move and turn.
“You have to predict across the globe every weather event,” he said. “We were off by the width of Florida, which is pretty good.”
Even when the forecast is dead-on, it isn’t always effective in convincing those in harm’s way to leave and seek safety. NCAR has been exploring ways to make the messaging associated with forecasting more compelling.
One important tool that could emerge in coming years would offer three-dimensional, street-level predictions of floodwater levels at specific addresses — whether it’s the front porch, the ground level windows or even the roofline of a home.
The idea for a 3-D representation of potential storm damage came from focus groups that were held in the wake of 2012’s Hurricane Sandy and last year’s Hurricane Matthew.
“People said they’d like to visualize the impacts,” said NCAR’s Olga Wilhelmi, project scientist with the Research Applications Laboratory, as a storm surge inundated a hypothetical town on a computer screen nearby. “With animation, it really drives the message.”
But the best forecasting can only help so much. It doesn’t mean everything will emerge unscathed, said Mari Tye, an engineer with NCAR’s Capacity Center for Climate and Weather Extremes.
“We can’t stop extreme weather,” she said. “What we must learn to do is accept that there will be graceful failures.”...Read more