Science

AI model beat traditional forecasts on Hurricane Isaias's track

NPR reports Google DeepMind's model correctly tracked Isaias, though it trailed the NHC blend on intensity.

Updated Oct 26, 2026 · 10:14 UTC1 reports

The short version

Hurricane Isaias made landfall near Destin, Florida, as a Category 2 storm in early October.[NPR]
Google DeepMind's AI model predicted the Florida Panhandle track, while conventional models initially pointed toward Louisiana.[NPR]
National Hurricane Center Director Michael Brennan says the model topped others on track but trailed the NHC's blended forecast on intensity.[NPR]
Lowry says AI models rely more on historical patterns than on atmospheric physics equations.[NPR]

Hurricane Isaias made landfall near Destin, Florida, as a Category 2 storm in early October.[NPR]
Google DeepMind's AI model predicted Isaias would hit the western Florida Panhandle, while conventional models initially pointed toward southeastern Louisiana.[NPR]
The National Hurricane Center first incorporated the Google DeepMind model into its official forecasts last year.[NPR]
According to Michael Lowry, a Miami meteorologist, AI models rely more on historical storm patterns than on atmospheric physics equations.[NPR]
Lowry says the Google DeepMind and European Centre for Medium-Range Weather Forecasts models have so far proven the most accurate AI forecasting models.[NPR]
Brennan says the Google DeepMind model topped other models in tracking Isaias but trailed the NHC's blended forecast on intensity.[NPR]

The model comparisons cited are drawn from one storm, Isaias, and the accuracy claims are attributed to officials and researchers rather than independently verified.[NPR]
Lowry says many unproven AI forecasting models exist and that the leading models have pitfalls despite their accuracy.[NPR]

Outlet counts describe coverage, not independent confirmation. Reports may share a wire service or original source.