Multi-messenger AI-enhanced earthquake early warning

Earthquakes caused nearly one million fatalities in the last two decades and billions of euros of economic loss. In our current state of knowledge, these hazardous events remain unpredictable.

Early Warning Systems (EWS) exist in some countries at risk, but they only send alerts once the earthquake has started, leveraging information recorded close to the epicenter before the most destructive seismic waves reach more distant populated areas. These systems provide – in the best-case scenarios – only a few seconds of warning before the strongest shakings.

Moreover, for fundamental reasons, they systematically underestimate the magnitude of large events, which results in dramatic underestimation of potential subsequent tsunamis, which typically cause most of the fatalities and damage.

Therefore, making EWS faster and more accurate is crucial to mitigate the hazard associated with these catastrophic events.

In the framework of the ERC StG project EARLI, we developed prototype Artificial Intelligence (AI) algorithms providing faster and more accurate theoretical estimates of the location and magnitude of large earthquakes than state-of-the-art EWS.

We propose to implement these AI algorithms in the operational EWS of Peru, with the objective of transforming the theoretical developments of the ERC StG EARLI (Licciardi et al., Nature, 2022; Lara et al., JGR, 2023) into concrete operational improvements in EWS performance. The algorithms we developed use short records of traditional seismic waves and light-speed gravity signals.

We will complement these two algorithms by a third AI-based one using GNSS, allowing the implemented EWS to leverage complementary real-time data, and making it the first operational multi-messenger AI-based EWS.

The system will rapidly benefit millions of people at high risk from earthquakes in Peru, and serve as a Proof-of-Concept for every region exposed to earthquake and tsunami hazard worldwide.