Investigating the typicality of the dynamics leading to extreme temperatures in the IPSL-CM6A-LR model - CEA - Université Paris-Saclay Accéder directement au contenu
Pré-Publication, Document De Travail (Preprint/Prepublication) Année : 2023

Investigating the typicality of the dynamics leading to extreme temperatures in the IPSL-CM6A-LR model

Résumé

Determining the underlying mechanisms leading to extreme events in dynamical systems is a challenging task. Under mild hypotheses, large deviations theory predicts that as one increases the threshold defining an extreme, dynamical trajectories which reach the extreme will look more and more like one another: they converge towards a typical, i.e. most probable, one called the instanton. In this paper, we use a 2000-year simulation of the IPSL-CM6A-LR model under a stationary pre-industrial climate to test this prediction on the case of hot extremes. We investigate whether the physical mechanisms leading to extreme temperatures at four locations in Europe are more similar with increasing extreme temperatures. Our results show that most physical variables exhibit the expected convergence towards a most probable trajectory, with some geographical and temporal variations. In particular, we observe the presence of a cut-off low in some trajectories, which suggests the existence of multiple pathways leading to extreme temperatures. These findings confirm the relevance of instanton dynamics in understanding the physical mechanisms driving extreme events in climate models.
Fichier principal
Vignette du fichier
Investigating_the_typicality.pdf (13.84 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
licence : CC BY - Paternité

Dates et versions

hal-04043595 , version 1 (23-03-2023)

Identifiants

Citer

Robin Noyelle, Pascal Yiou, Davide Faranda. Investigating the typicality of the dynamics leading to extreme temperatures in the IPSL-CM6A-LR model. 2023. ⟨hal-04043595⟩
70 Consultations
44 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More