Moving Towards Dynamic/Causal Process Models for Spatio-Temporal Extremes

Extreme weather events, such as droughts, floods, heat domes and wildfires, frequently occur in spatial clusters that evolve over time. In the past two decades, spatial extremes modeling has advanced significantly, particularly through block maxima and peaks-over-threshold models. However, these methods often rely on stringent assumptions about the joint tail behavior and exclude the bulk of the data to focus on extremes. Additionally, copula-based approaches often assume spatial stationarity and temporal independence, neglecting the complex dynamic relationships among multiple factors that interact across space and time. This session will showcase four leading experts in dynamic and causal approaches to spatio-temporal extremes, presenting diverse viewpoints to enhancing the practicality and effectiveness of modeling these extremes.