Groundbreaking research reveals that adjusting the way we model atmospheric friction can significantly improve the accuracy of hurricane intensity and rainfall predictions, potentially saving millions of dollars and countless lives.
Unraveling the Effects of Friction
Hurricanes, along with other natural disasters, can have tsunami-like impacts causing widespread destruction and loss of lives. So far, from 1980 to 2023 in the United States just, weather and climate disasters have cost $2.6 trillion in damages, according to an estimation by NOAA. Last year alone, the U.S. suffered 18 $1 billion disasters in 2022(metrics from NOAA)
Most intriguingly, this research concerns the messy interaction between a ton of sunlight near the surface in low latitudes — which aids in storm invigoration — and massive amounts of atmosphere resistance via friction. But the degree of this friction, and how it affects storm strength, has been largely a mystery. A team of researchers at University Of Houston delved into this crucial aspect and went ahead to evaluate the implications on hurricane forecasting.
Access the Power of Supercomputers
To solve the problem, a team led by graduate student Md Murad Khondaker and his advisor, Mostafa Momen, Assistant Professor of Civil and Environmental Engineering with UH Cullen College turned to supercomputing. Through an allocation from the NSF’s network of supercomputing resources, ACCESS, they accessed the Pittsburgh Supercomputing Center’s Bridges-2 system.
The magnitude and intricacy of these research operations needed significant amounts of computational power to handle the large data sets they worked with. In the end, Bridges-2 was indispensable offering 300k CPU-core hours along with the high-performance of its 16 high-RAM nodes, stuffed with a total of 512 gigabytes – or eight times standard laptop specs.
Once they had the computational tools at hand, the researchers were able to simulate 17 days of Hurricane Irma with eight-kilometer horizontal resolution in just 22 hours on 128 processors. The broader computational power required was necessary because this research could provide the insights that might one day change hurricane forecasting.
Conclusion
The implications of this groundbreaking research are significant, as recent hurricanes like Katrina, Harvey, and Maria have had total adjusted costs exceeding 400 billion dollars according to estimates provided by the NOAA. The researchers achieved a large reduction of the intensity forecast error, as well as a significant improvement in rain predictions—part of social adaptation to prevent risk areas from being flooded. Millions of dollars—and countless lives—might be saved if hurricane winds and floods could be more accurately forecast, allowing for better evacuation planning, as well as the deployment of crews and resources for emergency response. By more perfectly understanding the complex physiology involved in these storms, our ability to predict and prepare for the consequences of hurricanes should improve as climate change exacerbates these systems.
