Changes in Extreme Events: from GCM Output to Social, Economic and Ecological Impacts

Extreme events can deeply affect social and natural systems. The current generation of global climate model is producing information that can be directly used to characterize future changes in extreme events, and through a further step their impacts, despite their still relatively coarse resolution. It is important to define extreme indicators consistently with what we expect GCM to be able to represent reliably. We use two examples from our work, heat waves and frost days, that well describe different aspects of the analysis of extremes from GCM output. Frost days are "mild extremes" and their definition and computation is straightforward. GCMs can represent them accurately and display a strong consistent signal of change. The impacts of these changes will be extremely relevant for ecosystems and agriculture. Heat waves do not have a standard definition. On the basis of historical episodes we isolate characteristics that were responsible for the worst effects on human health, for example, and analyze these characteristics in model simulations, validating the model's historical simulations. The changes in these characteristics can then be easily translated in expected differential impacts on public health. Work in progress goes in the direction of better characterization of "heat waves" taking into account jointly a set of variables like maximum and minimum temperatures and humidity, better addressing the biological vulnerabilities of the populations at risk.