New statistical model could give Atlantic hurricane warnings up to five days early
Scientists at NSF NCAR have built a logistic-regression tool that forecasts rapid hurricane intensification up to five days in advance for Atlantic storms.
A group of atmospheric scientists led by Christopher Rozoff at NSF NCAR created a logistic-regression based prediction tool that evaluates 55 sea-surface, wind-shear and other atmospheric factors to estimate the probability of a tropical cyclone’s rapid intensification. Using ensemble forecasts from NOAA's Global Ensemble Forecast System spanning 2019-2024, the model delivered reliable five-day lead-time forecasts for Atlantic hurricanes and three-day forecasts for eastern-Pacific systems.
It also highlights which projected tracks are most conducive to sudden strengthening, giving emergency managers more time to prepare. The study, published in Weather and Forecasting, cites past failures to predict events like Hurricane Michael and Hurricane Otis as motivation. Future work will expand testing to additional hurricane seasons and integrate the tool with other forecasting models, aiming for routine operational use by the National Hurricane Center.
Why it matters
Earlier warnings of sudden hurricane strengthening can give coastal residents crucial time to protect lives and property.
In this story