Google researchers have found a unique means of predicting flash flooding: asking artificial intelligence to process the news.
The technology giant has been forecasting riverine and coastal flooding caused by overflowing bodies of water on its online platform, Flood Hub, for years using deep learning computer programs. But flash flooding — caused by excessive rainfall — has been more difficult to predict because the amount of flooding depends on local factors like drainage and topography.
Flash floods happen fast and are often short-lived. But those same factors that make them among the deadliest types of extreme weather also mean there’s a lack of data to be used by forecasting models.
Advertisement
Until now.
