Meteorologists use satellites and AI to warn about flash floods sooner

In short: Meteorologists are starting to combine satellite data with machine learning to spot flash flood risk sooner and send earlier warnings.

What's going on

Flash floods can build up fast, sometimes within minutes, which leaves people little time to react. The Verge describes one example from June 9, when Laura Lin was working at home in Lanesville, Indiana, during heavy rain. She noticed wood floating near her barn, shut her laptop, woke her kids, and rushed to safety as floodwater rose.

Floods are a major danger. In the United States, they are the second-deadliest type of weather event, and globally they are the deadliest, according to data cited in the story. Even a small amount of fast-moving water can be dangerous, with as little as 6 inches able to knock an adult off their feet, and about 12 inches able to lift a car.

Researchers and forecasters are working on new warning tools that use satellite information plus machine learning. Machine learning is a type of AI that finds patterns in large amounts of data (like a very fast assistant that sorts through millions of clues). Satellites can provide a broad view of conditions over large areas, including places that may not have many ground sensors.

What to watch

The key question is how quickly these tools can move from research into everyday forecasting, and whether they can reduce false alarms while still giving earlier alerts. If they do, more people could get a few extra minutes or hours to move to higher ground, protect property, or evacuate.

Source: The Verge AI