AI Uncovers Hidden Climate Extremes in Europe

A recent study led by Étienne Plésiat from the German Climate Computing Center has demonstrated the transformative power of artificial intelligence (AI) in climate science, Earth.com revealed yesterday.

Utilizing data from over 30,000 weather stations worldwide, the research applied AI to reconstruct European climate extremes, confirming known trends and revealing previously unrecorded extreme events.

Traditional methods of analyzing historical temperature data often struggle due to gaps in records, particularly in regions with sparse weather stations. To address these challenges, Plésiat and his team developed CRAI (Climate Reconstruction AI), which outperformed conventional interpolation methods in reconstructing extreme temperature events.

CRAI was trained on historical simulations and validated against standard metrics, successfully uncovering long-forgotten climate extremes, including a severe cold spell in 1929 and a major heatwave in 1911.

The findings are published in the journal Nature Communications.