By Alexander Castro
Predicting the future has never been precise. Sheep or bird entrails were the ancient Romans’ preferred methods. Saint Augustine, who despised all things magical, cataloged water and the dead as other oracular mediums. Aeromancy, or divination by clouds and wind currents, was another arcane specialty.
And today’s clairvoyants? They’re arguably the scientists who have swapped superstitious logic for algorithmic finesse, like Mansur Ali Jisan, a Ph.D. candidate and research assistant at GSO. He’s developed the Hurricane Boundary Layer (HBL) model, an improved way to forecast wind during a hurricane.
“The basic idea,” Jisan says, “is when a hurricane makes landfall, when it transitions from the ocean toward the land, there’s a change in structure and intensity…When a storm is over water, there’s less drag. When it makes landfall, it interacts with different types of land.”
Those interactions make for surface friction, which the HBL model uses to simulate a hurricane’s vortex. Using data from the National Hurricane Center about the storm’s position, speed, and radius, the vortex structure can even be modified throughout the forecast’s duration.
“Previous teams have worked on hurricanes as they move over the ocean,” says David Smith, GSO’s associate dean. “Now they’re moving onto shore with Mansur. What he’s really adding is that all the land is not the same.”
Our research is not just academic, where a few people read your papers and that’s it.Prof. Isaac Ginis
Jisan offers the Miami area as an example: a hurricane will encounter high friction in the city, but nearby wetlands will offer less resistance. The HBL model accommodates these differences in surface and roughness, all of which affect a hurricane’s primary form of destruction: flooding. “Water is the main impact that a storm leaves,” Jisan says. “If you don’t have good wind in your storm-surge simulation, your water-level prediction could be underestimated or overestimated.”
Another advantage of the HBL model is its simplicity compared to numerical weather prediction (NWP). Numerical models are widely used but they require inputting all sorts of specialist physics. The HBL’s biggest extravagance, meanwhile, is “incorporating surface friction,” Jisan explains. “[This] makes our model quite simple.”
That doesn’t mean less accurate. So far, tested against historical storms like hurricanes Florence (1918) and Michael (2018), Jisan says the model has generated “quite reasonable comparisons” against the observation data.
Accuracy and usability are indeed the goal, as the HBL model has been designed to “be used by anyone,” Jisan says. “I really hope that once we publish this code in open source, that emergency managers can use this for decision making.”
Jisan got a real-life chance to validate his model as part of his work with the Hurricane Research Group at GSO, led by Isaac Ginis, who also serves as Jisan’s doctoral advisor. Ginis’ group is affiliated with the Coastal Resilience Center, a Department of Homeland Security-funded effort that focuses on mitigating coastal hazards.
“Our research is not just academic, where a few people read your papers and that’s it,” Ginis says. “We are not only developing this code, we’re making it relevant to the decision makers.” During Hurricane Henri in 2021, Ginis’ hurricane crew made predictions for the storm, with Jisan handling the wind calculations, and sent their results to emergency management in Rhode Island. Ginis describes Jisan as a highly enthusiastic collaborator, one who’s made “significant contributions” to the Hurricane Research Group.
“Not many people venture to develop models,” Ginis says.
Maybe that’s because modeling a hurricane involves writing code for enormous supercomputers. The myriad calculations that comprise simulations need a massive amount of energy to run. Spewing carbon emissions while fighting hurricanes is a patently contemporary paradox, and it’s something Jisan has considered. The HBL model is scalable, meaning it “can be run using any number of [CPUs] depending on the demand of the simulation.”
Jisan’s typical setup uses 256 processors to spit out a five-day forecast. That’s still an improvement over NWP models Jisan has tested, one of which takes around three hours and 480 CPUs to produce a five day forecast. An HBL-based forecast, meanwhile, takes about an hour.
All that power doesn’t come cheap, but thanks to the Coastal Resilience funding, Jisan can achieve such gargantuan number crunching. That’s one perk of living in the US:“It really allows you to do the research you want to do,” Jisan says.
For him, that research is generally grounded in practicality and social benefit: “I always try to think of how [my research] is going to impact our society,” Jisan says. “How will it make people more aware of climate change or sea-level rise?”
Thoughts Not Far from Home
His native Bangladesh is more than aware. The southeast asian country endures a marathon of intense weather yearly. Outside of the monsoon season’s unending rain, Jisan says hurricanes aren’t uncommon. Yes, the nation is geographically poised to be stormy, but rising seas and temperatures don’t help.
“People are unequivocal that it’s causing hazard already in their lives, especially people living near the coast,” Jisan says. His career began in orbit of these very issues, as a civil engineer at the Institute of Water and Flood Management in Dhaka, the Bangladeshi capital.
According to 2021 stats from Germanwatch, a Berlin nonprofit, Bangladesh placed seventh in the Climate Risk Index, an aggregate ranking of adverse climate events from 2000-2019. Yet a 2021 study in the journal Energy Strategy Reviews notes that Bangladesh contributes only 0.56 percent of global CO2 emissions.
That’s one example of the interaction between climate change and social inequity. No wonder Jisan praises GSO’s Justice, Equity, Diversity and Inclusion initiatives as “a really essential part of having a good research environment.”
To that same end, he believes public outreach is crucial to the researcher’s approach: “You make a connection with the public and see what they’re thinking…and whether the research I’m doing is helping them. What are the things people are actually worried about? What’s on their minds?”
Jisan had a lot on his mind after finishing his Ph.D. comprehensive exams, so he went in search of a relaxing hobby. First stop: Best Buy, where he bought a basic drone, the DJI Mini 2. Next stop: New Hampshire and the Adirondacks, where Jisan’s drone snapped many vivid sceneries from above. Jisan’s optimism is almost tangible in these photos: Sweeps of red foliage into bright green. Clusters of yellow. Ripples sculpted into water. A smoldering sunset or two. And locally, the bridges, lighthouses and wee lil’ islands that populate Rhode Island’s coastline and bays.
“I really like living in this area, right near the coast,” Jisan says. He has a one-minute commute to GSO from his home, where he lives with wife Nadia and their pet bunny Oreo (who has an Instagram, naturally). Jisan even appreciates “the cold weather system” of New England. Yes, he’s a fan of the snow—the scourge of many a grumpy Rhode Islander.
But the occasional blizzard seems a fair tradeoff for Rhode Island’s historical luck when it comes to avoiding the worst of hurricanes. Other coastal communities aren’t so lucky, so there’s plenty of incentive to craft more accurate models for extreme weather. After all, a prophecy that can’t be acted upon isn’t very useful.