The National Institute of Standards and Technology created a new digital simulation method for hurricanes using artificial intelligence and the National Hurricane Center’s Atlantic Hurricane Database.
The experimental technique shows promise in helping to improve building design guidelines particularly in storm-prone areas, NIST said Wednesday.
Researchers implemented machine learning algorithms on hurricane datasets from the HURDAT2 database, which provides information on storm paths, windspeeds and other factors recorded from the past 100 years. The algorithms were able to devise models that can mimic specific storm behaviors in a matter of seconds.
NIST clarified that the model cannot currently simulate the effects of climate change on future hurricanes. The wind maps developed from the ML-based model paints a more realistic pattern. They are intended to be used by the researchers to propose new guidance in building standards and codes.