Significant empirical and theoretical advancements have revealed the departure of hydrometeorological processes from classical statistical models, highlighting the scaling behavior of their variables, especially extremes, across state, space, and time. These extremes, along with the general statistics of hydrometeorological processes, are crucial inputs for hydrological applications, which have increasing importance in the (re)insurance industry. Among the most common applications, catastrophe models are developed to manage risk accumulation; disaster response is used to prepare (re)insurers financially after major events; Real Disaster Scenarios are built to stress-test (re)insurers exposure both in the present-day and future climate.
For instance, in the context of a flood risk model, estimating design rainfall not only involves determining the absolute rainfall amount for a specific return period but also requires understanding the intra-event rainfall distribution, spatial extension, and rainfall intensities at neighboring stations. When these details are underestimated, it can easily turn into a poor risk assessment and weaker financial protection. Additionally, connections between hydrometeorological extremes and climatic oscillations, such as NAO or ENSO, and their evolution in a changing climate, provide insights for long-term risk management in the re-insurance sector, as required for regulatory purposes.
The integration of supporting information and the application of advanced AI approaches offer as well unprecedented opportunities to enhance these estimates. This session invites submissions, among others, on the following topics:
- Coupling stochastic approaches with deterministic hydrometeorological predictions to better represent predictive uncertainty.
- Developing robust statistics under non-stationary conditions for design purposes.
- Parsimonious models of hydrometeorological extremes across various spatial and temporal scales for risk analysis and hazard prediction.
- Improving the reliable estimation of extremes with high return periods, considering physical constraints.
- Linking underlying physics and hydroclimatic indices with the stochastics of hydrometeorological extremes.
- Exploring supporting data sets for additional stochastic information and utilizing novel AI and machine learning approaches.
- Applications carried out jointly by the (re)insurance industry and research institutions.