Rapid characterization, physical understanding and knowledge-based assessment of potential extreme event impacts are the focus of this Subtopic. Our main concerns are three different types of extreme events: 1) Events that affect areas of an extreme spatial extent, such as magnitude 8-9 earthquakes, ‘super-volcano’ eruptions, mega-tsunamis, or extreme solar storms. 2) Moderate magnitude eventsthat, if occurring today in proximity to urban areas, would inflict serious economic and societal costs. 3) Multi-hazard cascade scenarios, where coupling between events leads to major cumulative impacts, e.g., the interaction between volcanism, landslides and tsunamis.

Three key themes are central to this Subtopic:

Forensic analysis of past events: High-quality catalogs and geological archives of extreme events are essential for developing and validating the scenarios, and to quantify the impact of future extreme events, as well as describing their future occurrence in a probabilistic way. New forensic analysis methods are developed to better understand cascading phenomena. We focus especially on massive marine landslides triggered by earthquakes, or volcanic eruptions that generate tsunamis.

High resolution and global characterization of extreme events: We monitor multiple and well selected sites (e.g. seismic gaps) simultaneously to potentially capture one of these extremely rare events aided by global observations and open data policies. New sensors and platforms allow rapid deployment of instrumentation and novel multi-parameter source inversion methods are tested at the pilot sites and on global scales. Realistic multi-hazard simulations in pilot regions are developed and used to estimate site effects, infrastructure damage and losses based on exposure and vulnerability models. Crowd-sourced approaches are integrated into the data analysis framework.

Modeling future extremes beyond past events: By combining physics-based simulation tools with probabilistic and data-mining approaches, new opportunities for understanding the natural process chains and tipping points that culminate in extreme events are developed. These simulations are integrated with vulnerability and exposure models to estimate potential damages, losses, and fatalities, including the transparent communication of their uncertainties.