Our project partner DKRZ has recently released a publication in Nature Communications on how to bring the use of AI one step further. Recent extreme climate events like heatwaves, heavy rainfall, and droughts have intensified concerns about their impact on ecosystems, agriculture, water supplies, and human health. To fully assess these risks, understanding how these events compare to those in the past is essential. EXPECT researchers from the German Climate Computing Centre (DKRZ) along with Met Office and the Barcelona Supercomputing Center (BSC), have developed a novel method using artificial intelligence to reconstruct historical data on climate extremes across Europe.
Traditional statistical methods used to extract missing climate data have been limited by inaccuracies, particularly in overlooked areas and periods before the mid-20th century. DKRZ’s team overcame these limitations by applying AI, trained on large datasets. Their AI model, run on DKRZ’s high-performance “Levante” system, successfully revealed spatial climate trends that most reanalysis datasets lack.
This advancement provides a valuable historical context for assessing recent climate risks and understanding regional variations in climate extremes across Europe. By revealing patterns not visible with previous methods, AI offers a new perspective on the long-term shifts in climate extremes, supporting improved climate adaptation and risk assessment.
DKRZ’s innovative approach marks a major step forward in climate science, demonstrating that AI has the potential to significantly enhance our understanding of historical climate data, informing future strategies for effective risk management and policy development.
Further information can be found in DKRZ’s the press release.
