Artificial Intelligence for Enhanced Representation of Processes and Extremes in Earth System Models - Hop On

The proposed “Hop On” facility will contribute to accomplishing and enhancing several aspects of the AI4PEX project and, at the

same time, develop new competencies and skills through joining this transnational, multi-disciplinary research initiative. The primary goal is to further improve the reliability of climate projections at global and regional scale, by enhancing their ability to represent extreme hydroclimatic phenomena that are not inherently captured in the model conceptualization.

In particular, we aim at (i) adopting a model evaluation philosophy that seeks and identifies new system properties, such as compound extreme events, emerging from the interaction of its components, (ii) reducing uncertainties related to observational disagreement, (iii) improving the representation and understanding of recent trends in compound extreme evaporation/precipitation events, (iv) identifying hotspots of critical future change in the frequency, intensity, and duration of compound extreme evaporation/precipitation events across Europe and globally, and (v) investigating the ability of ESMs to generate extremely rare events with severe socioeconomic impacts, such as catastrophic floods. This approach is unique in its investigation of how new system properties emerge in model simulations by analysing compound extreme events, with a particular focus on the interactions between extreme evaporation and precipitation. The Czech University of Life Sciences Prague, located in Eastern Europe in the broader hot-spot region of the Danube River basin, will also contribute to maximizing the impact of AI4PEX by bringing new channels and tools for project dissemination and communication.