Proposal Description

According to the Aon Annual Climate and Catastrophe Insight report, the number of large-loss natural hazard events reached record levels in 2023, with 66 billion-dollar economic loss events, and 37 billion-dollar insured loss events. Earthquakes caused the most economic losses, while severe convective storms were most costly to insurers. 95,000 people globally lost their lives due to natural hazards in 2023 – the highest number since 2010 – resulting largely from earthquakes and heatwaves. With insurance covering only $118 billion, or 31 percent of total losses, the 'protection gap' stood at 69 percent, which highlights the urgency to expand insurance coverage.
The speakers are addressing prediction in real-time of extreme events whose dynamics are complex. These predictions, accompanied by uncertainties, form the basis of early warnings systems. These include novel statistical approaches to forecast tropical cyclones that form over the ocean (Pierre Tandeo, IMT-Atlantique, France), novel data assimilation at large scale and high time frequency using High Performance computing with the Fugaku supercomputer in Japan for heavy rain events (Takemasa Miyoshi, RIKEN-CCS, Japan). Also the session will have a specific focus on multihazards for near-term of long term assessments, either triggered by one another or as a diverse outcome. Elaine Spiller (Marquette University, USA) combines understanding of the fire conditions that create the conditions for subsequent catastrophic debris flow due to rain following the fire (as in deadly evens in California recently). Mark Bebbington (Massey University, New Zealand). will focus on probabilistic forecasting the various hazards that can result from a volcanic eruption. The use of surrogate models, also known as statistical emulation, will be key to some of the discussions. Indeed, the need to quickly and fully assess the range of outcomes, and their distributions, known as Uncertainty Quantification, is essential in the field of hazard assessments.