NASA's impressive new AI can predict when a hurricane intensifies

Meteorologists have gotten pretty damn good at forecasting a hurricane's track. But they still struggle to calculate when it will intensify, as it's seriously hard to understand what's happening inside a tropical cyclone. A new machine learning model developed by NASA could dramatically improve their calculations, and give people in a hurricane's path more time to prepare. Scientists at the space agency's Jet Propulsion Laboratory in Southern California developed the system after searching through years of satellite data. They discovered three strong signals that a hurricane will become more severe: abundant rainfall inside the storm's inner core; the amount of ice water in the clouds within the tropical cyclone; and the temperature of the air flowing away from the eye of the hurricane. The team then used IBM Watson Studio to build a model that analyzes all these factors, as well as those already used by the National Hurricane Center, a US government agency that monitors hazardous tropical weather.