AI dataset transforms flood forecasting in data-scarce regions globally
A new AI-driven dataset is helping forecast flash floods in areas with little weather data. Meanwhile, a German tech firm is planning a major renewable-powered AI data centre. Elsewhere, a pilot strike at Lufthansa caused limited disruption, and Iran warned of potential energy market instability amid ongoing conflicts.
Google researchers used artificial intelligence to scan over 5 million news articles. They identified 2.6 million flood events across the globe. These findings were then turned into a geo-tagged dataset called Groundsource, covering 150 countries from 2000 to the present. The dataset improves flood risk predictions, particularly in regions where traditional weather data is scarce. Google's Flood Hub platform now uses this information to highlight potential flash-flood risks in urban areas worldwide.
In Germany, the start-up Polarise announced plans for a 30-megawatt AI data centre in Amberg. The facility will run entirely on renewable energy and could become one of the country's largest domestically operated AI computing hubs.
Lufthansa managed to operate over half of its scheduled flights on the first day of a two-day pilot strike. The airline kept disruptions to a minimum despite the industrial action.
Separately, the International Energy Agency released 400 million barrels of oil from emergency reserves. The move aims to stabilise markets after Iran warned that prolonged conflict with the US and Israel could disrupt global energy supplies. Tehran specifically highlighted risks to oil prices and shipping routes through the Strait of Hormuz.
The Groundsource dataset is now being used to enhance flood forecasting in data-scarce regions. Polarise's renewable-powered data centre, if completed, would mark a significant step in Germany's AI infrastructure. Lufthansa's operations remained largely unaffected by the strike, while the IEA's oil release seeks to prevent market instability amid geopolitical tensions.
