Description du projet
Utiliser l’intelligence artificielle pour prévoir les effets du changement climatique sur les phénomènes météorologiques extrêmes
Le changement climatique modifie et renforce les phénomènes météorologiques extrêmes tels que les vagues de chaleur, les incendies de forêt dévastateurs, les cyclones, les inondations et les sécheresses. Le projet XAIDA, financé par l’UE, caractérisera, détectera et attribuera les événements extrêmes à l’aide d’une nouvelle approche fondée sur les données et l’impact. Il emploiera de nouvelles techniques d’intelligence artificielle (IA ou AI pour «artificial intelligence») et réunira des spécialistes de l’attribution des événements extrêmes, de la dynamique atmosphérique, de la modélisation climatique, de l’apprentissage automatique et de l’inférence causale. Les résultats feront la lumière sur l’effet du changement climatique sur les phénomènes atmosphériques tels que les cyclones et les tempêtes convectives, qui ne sont pas bien compris ou quantifiés. Le projet fournira également des outils pour évaluer les voies causales menant à des événements extrêmes.
Objectif
Often, extreme events provide representations of the future climate, but not all extremes are harbingers of the future. Thus, in order to be useful for adaptation in support to future projections, a causal link between events and human influence on climate must be established or refuted. This is why the “Extreme event attribution” field has recently developed. However, extreme event detection, attribution and projections studies currently face major limitations.
XAIDA will fill these gaps. Using new artificial intelligence techniques, and a strong two-way interaction with key stakeholders, it will (i) characterize, detect and attribute extreme events using a novel data-driven, impact-based approach, (ii) assess their underlying causal pathways and physical drivers using causal networks methods, and (iii) simulate high-intensity and as yet unseen events that are physically plausible in present and future climates.
To achieve this, XAIDA brings together teams of specialists in extreme event attribution, atmospheric dynamics, climate modelling, machine learning and causal inference, to:
● Understand the effect of climate change on a variety of impacting atmospheric phenomena currently poorly understood or quantified (cyclones, convective storms, long-lived anomalies, or summer compound events), both for past and future evolutions;
● Develop, in co-design with a community of key stakeholders, a novel, broader and impacts-based attribution and projection framework which extracts causal pathways of extremes;
● Develop storylines of events of unseen intensity, based on machine learning methods;
● Provide new tools for model assessment of causal pathways leading to extreme events and investigate the causes of disagreements between models and observations;
● Develop an interaction and communication platform with stakeholders with the ambition to improve training and education on climate change and impacts and to bring these developments to future operational climate services
Champ scientifique
Programme(s)
Régime de financement
RIA - Research and Innovation actionCoordinateur
75794 Paris
France