Strategies and Challenges in Applications of DAS-based Earthquake Early Warning Systems

Distributed Acoustic Sensing (DAS) transforms fiber-optic cables into ultra-dense strainmeter arrays, providing spatially and temporally continuous earthquake recordings. While its potential for offline seismic characterization is increasingly recognized, a key application of this sensing paradigm is real-time monitoring for Earthquake Early Warning (EEW). The use of existing fiber-optic infrastructures allows for sensing cables located close to seismogenic sources, such as offshore subduction zones, potentially extending the lead time of issued alerts. DAS deployments within Near Fault Observatories further provide dense spatial coverage of epicentral areas, favouring the rapid extraction of robust source information.

The application of DAS to EEW – alone or as a complement to standard accelerometers - has been recently explored, specifically focusing on the estimate of earthquake magnitude from the first seconds of recorded data. Existing approaches rely either on conversion strategies to ground-motion proxies or on direct analysis in the strain-rate domain. However, both the robustness of different conversion strategies and the selection of the most informative physical quantity for early magnitude estimation are not yet consolidated. In offshore environments, additional complexity arises from fiber-optic cables deployed on sediments, where strong converted phases often dominate early waveforms and hinder the direct P-wave signal traditionally used for EEW.

In this work, we analyse earthquakes recorded by the ABYSS network, supported by the ERC – starting program, consisting of 450 km of offshore telecommunication cables deployed along the Chilean subduction trench and interrogated by three DAS units. At this high-seismicity testbed, we develop a strategy for fast magnitude estimation with DAS. We show that converted Ps phases preceding S-wave arrivals carry significant information on earthquake magnitude. Furthermore, we investigated whether the use of time and space-integrated observables on DAS recordings can enhance the predictive power of amplitudes from the first seconds of seismic signals.

Finally, we assess the performance of a DAS-based EEW, grounded on the software PRESTo (Satriano et al., 2011). Using moderate-to-large offshore Chilean earthquakes, we highlight potential and limitations of DAS in regions with sparse conventional instrumentation. Complementary analyses using data from the Irpinia Near Fault Observatory demonstrate the benefits of jointly exploiting DAS and traditional seismic stations within dense monitoring networks, confirming the applicability of DAS-based EEW systems across different tectonic settings.