Edge-to-end workflows in natural hazards (invited)

Abstract

Several natural disasters (e.g. earthquakes, volcanic eruptions, landslides or tsunamis) are hard to predict and pose a serious threat to human lives and property. Prompt reaction to these scenarios requires of computing infrastructures, complicated data workflows, and engagement with stakeholders formally involved in emergency management. The Center of Excellence for Exascale in Solid Earth (ChEESE; https://cheese-coe.eu) is preparing flagship HPC applications in Earth sciences and natural hazards for the Exascale era.

The Center has a strong commitment to building services for industry and public governance bodies (e.g. civil protection), including urgent computing edge-to-end workflows of data coming from networks of sensors to HPC resources for earthquakes and tsunamis. A series of Pilot Demonstrators (PDs) on Urgent Seismic Simulations, Faster Than Real-Time Tsunami Simulations, and High-Resolution Volcanic Ash Dispersal Forecast are being developed to test future Exascale Urgent Computing workflows supporting contingency plans for seismic, volcanic and tsunami events. For earthquakes, PDs are being designed in collaboration with the Italian and Icelandic civil protections on generation of physics-based ground shaking (impact) maps shortly after the occurrence of an earthquake or on real-time tsunami inundation forecasts.

In the case of tsunamis, the aim is to increase the accuracy of tsunami forecast by assimilating the largest possible amount of data in quasi real time, and performing simulations in a few minutes wall-clock time, possibly including the coastal inundation stage. This strategy of direct real time computation, that could seem unfeasible a decade ago, it is now foreseeable thanks to the astonishingly recent increase in the computational power and bandwidth evolution of modern GPUs.

These examples can also pave the road for urgent computing in other domains, from fire propagation to atmospheric or marine toxic dispersals.