Flash floods are a growing concern due to climate change, making accurate predictions vital for community safety. Google’s recent research highlights how artificial intelligence, specifically its Gemini AI, can enhance flash flood forecasts. By analyzing over 9 million global news articles since 2000, Google created Groundsource—the most extensive dataset of historical flood events, containing 2.6 million records across 150 countries. This dataset trains predictive models to provide urban flash flood forecasts in 24-hour intervals via Google’s Flood Hub, although currently, it indicates only broad risk levels for cities. Early results show promising application, such as aiding Southern African disaster response efforts. While the data’s initial accuracy is about 82%, its potential for improving flood predictions underscores the role of AI in disaster readiness. Future applications may extend beyond flooding to other extreme weather events, showcasing AI’s capability to harness publicly sourced data for enhanced crisis management.
Source link
