Abstract. We present a large seismological dataset (https://geofon.gfz.de/doi/network/ZK/2021) composed by continuous recordings from 200 temporary stations (seismic network ZK, 10.14470/MX7576871994), which operated continuously for approximately one year in the Southern Apennines (Southern Italy) as part of the DETECT experiment. The dataset is compliant with the seismological standards for the archiving of data and metadata, and with standardized tools for disseminating them, e.g., the International Federation of Digital Seismograph Networks (FDSN) web services. The data set was collected during the DETECT experiment in the Irpinia area (southern Italy), which is one of the regions with the highest seismic hazard in Italy. From August 2021 to August 2022, a constellation of 20 seismic arrays, with a total of 200 seismic stations, was installed above the fault segments responsible for the M 6.9, 1980 Irpinia earthquake, the strongest and most destructive seismic event in Italy in the last fifty years. Each seismic array had a maximum aperture of ~2 km and was composed of one broad-band sensor, one short- period sensor with 1 Hz and eight 4.5 Hz natural frequency geophones. Data and metadata were managed in accordance with the GIPP/GEOFON policy using SeisComP (Helmholtz Centre Potsdam – GFZ German Research Centre for Geosciences and GEMPA GmbH, 2008) and the final archive size is approximately 5.2 TB. In this contribution, we provide details on how to download waveform data and station metadata, as well as information on data availability and quality.

Indonesia faces a unique combination of exposure to tsunamis, with dense coastal populations, intense seismic and volcanic activity, and multiple tsunami-generation mechanisms that can threaten shorelines with very little time for reaction.BMKG which is in charge of monitoring tsunamis faces 2 limitations: First, the operational system still relies mainly on land-based seismology, coastal sea-level observations, and precomputed modelling, while the availability and operability of deep-ocean buoy observations have remained an ongoing issue. Second, the existing system was designed for tectonic earthquake-generated tsunamis and is not sufficient on its own for non-tectonic sources such as volcanic or landslide-triggered events.To address these limitations, BPPT (subsequently integrated into BRIN in 2021), initiated the planning, design, and deployment of two prototype Ocean Bottom Unit (OBU) sensors between 2020 and 2022. The deployed units integrate high-sensitivity accelerometers and static pressure sensors to enable real-time monitoring of seismic activity and tsunami wave propagation in the northern waters of Labuan Bajo, Flores.The first OBU was installed approximately 25 km offshore at a depth of 2,110 m, whereas the second unit, positioned at the terminal section of the submarine cable, was deployed approximately 55 km offshore at a depth of 4,122 m. Both installations were strategically designed to monitor seismic activity associated with the Flores Thrust, located roughly 100 km north of Flores Island. Historical earthquakes along this tectonic structure, together with the submarine landslides they triggered, generated destructive tsunamis in 1982 and 1992, resulting in significant casualties and coastal run-up heights exceeding 20 m.The deployed submarine cable system has a total length of 55 km and incorporates 12 optical fibres, providing the necessary infrastructure for integrated sensing and real-time data transmission. At the end of April 2026, we connected a Febus Optics A1 DAS interrogator to one of the fibres for a preliminary study of local and regional seismicity. Our objectives with this initial deployment are to catalogue and characterise local seismicity, locate sources, and investigate improvements in the location of local earthquakes by using a combination of land-based seismic stations and the undersea optical cable.

Detecting, locating, and characterising the dynamics of destabilised volcanic material is critical for assessing the extreme hazards posed by volcanic mass flows, such as pyroclastic density currents. Geophysical measurements of these events may offer information otherwise hardly observable at close range. Here we investigate pyroclastic density current dynamics using a multiparameter approach that combines seismic, distributed acoustic sensing, and infrasound data with thermal and visible imagery, supported by numerical simulations. We focus on two events at Stromboli volcano, Italy, that occurred in October and December 2022. By comparing visible imagery with seismic energy and applying array processing techniques, we identify different flow volumes ( ~23.5 ± 9.5 × 103 m3 and ~80 ± 9 × 103 m3, respectively) and velocities (33-42 m/s and 54-59 m/s, respectively). Simulations reveal that reproducing these velocities requires volume-dependent empirical friction angles ( ~27° and 21°), consistent with dry granular flow behaviour and friction weakening. These findings offer new insights into the use of distributed acoustic sensing for volcanic monitoring and underscore the value of integrating multiparameter data with modeling to better understand complex volcanic processes. Integrating seismo-acoustic monitoring with visual observations provides valuable insights into pyroclastic density currents dynamics and associated pre- and post-collapse processes, according to integration of geophysical observations and modelling approaches.

Distributed Acoustic Sensing (DAS) is of critical value for the offshore expansion of seismological networks. The work presented here is part of the 5-years ERC ABYSS project, which aims at building a permanent seafloor seismic observatory leveraging offshore telecommunication cables along the central coast of Chile. The ABYSS project near-real time data collection started the 30th of September 2023 using three ASN (Alcatel Submarine Networks) OptoDAS units to sense three segments over two offshore telecommunications cables connecting the cities of Concón to La Serena and La Serena to Caldera. The DAS data covers over 500 km of cable, comprising 26,664 virtual sensors sampled at 62.5 Hz and 100 Hz. These data are synchronized once a day with a storage server located in France, the volume of which is anticipated to reach an estimated 608 TB by the end of the project. We developed an automatic workflow to detect an average of 100 daily local, regional and teleseismic events with magnitudes down to ML = 0.5, over 59 GB of data per day after compression.As a first step, we perform automatic seismic phase arrival picking using PhaseNet pretrained on conventional seismological stations, followed by phase association with GaMMA. We then apply a correction of the phase picks to account for shallow sedimentary layers and invert for the event hypocenters with NonLinLoc software. Finally, we estimate the Richter local magnitude based on peak ground displacements. The results show that DAS data combined with data from the national onland seismic network greatly increases the accuracy of the earthquake hypocenters. Once the earthquake catalogs are built, we can perform a relative relocation of the earthquakes with HypoDD software using cross-correlation and/or the catalog results. With this workflow we show that conventional tools used in seismology can be used on DAS data with few adjustments. Furthermore, the size of our catalog, enriched with numerous undetected offshore events is a significant improvement over the existing regional catalogs, which may aid future studies of the Chilean margin subduction zone seismicity.

Abstract Distributed Acoustic Sensing (DAS) provides dense, wide‐aperture observations potentially ideal for imaging earthquake rupture processes. By combining data from a ∼450 km‐long seafloor DAS array offshore Chile and a conventional seismic network, we image an Mw 6.0 intermediate‐depth earthquake ∼400 km away in Argentina, achieving km‐scale spatial resolution of its rupture process. Back‐projection imaging reveals two sub‐events and northward propagation across 6–8 km in ∼3 s. The results favor rupture on a north–south striking (∼350°), west‐dipping (∼53°) normal fault within the subducting slab, consistent with reactivation of a pre‐existing weakness plane such as an ancient outer‐rise fault or the fractured flank of a subducted seamount. This study demonstrates the capability of DAS to resolve moderate‐size (Mw 6) earthquake ruptures at regional distances with high resolution and wide coverage, greatly expanding the range of events that can illuminate the physics of earthquakes.

Monitoring temporal variations of seismic velocities (dv/v) is a key tool for investigating stress changes and damage processes in active tectonic regions. Traditional dv/v studies rely on dense seismic networks or on repeating earthquakes, which can limit their spatial resolution and applicability. Distributed acoustic sensing (DAS), by providing continuous and densely sampled measurements of the seismic wavefield, offers new opportunities to overcome these limitations and to develop high-resolution velocity monitoring strategies.In this study, we investigate how dv/v can be estimated from DAS data by focusing on the analysis of seismic swarms. We develop a processing workflow and apply it to DAS data acquired from three submarine telecommunication fiber-optic cables of ~150 km each, with ~10,000 sensing points per cable, deployed in the central part of Chile (Abyss network). We first identify seismic swarms and quantify waveform similarity between events using multi-channel cross-correlation analysis. We then select event pairs exhibiting high waveform similarity across multiple DAS channels for further analysis. We analyze coda waves using a cross-spectral approach to estimate coherence and phase delays between events, and we infer relative seismic velocity variations from a linear regression of the measured time delays over selected coda time windows starting a few seconds after the S-wave arrival.Through this work, we present a systematic framework for estimating dv/v from DAS-recorded seismic swarms and assess its sensitivity to event similarity, frequency band, and coda window selection. This work shows that seismic swarms, when recorded by dense DAS arrays, provide a promising basis for developing high-resolution seismic velocity monitoring strategies.

Distributed Acoustic Sensing (DAS) transforms fibre-optic cables into densely spaced arrays of strain sensors, providing metre-scale resolution over distances of up to hundreds of kilometres. By analysing interferometric backscatter from laser pulses, DAS measures deformation rate along the fibre, yielding single-component recordings of longitudinal strain.In seismology, DAS supports both high-resolution imaging and seismic monitoring. Its dense sampling enables regional tomography and detailed characterization of shallow structures, while coherent travel-time data can be integrated into standard processing workflows. Phase picking can be performed using adapted artificial intelligence methods for continuous data, although earthquake location remains challenging due to cable geometry. Nevertheless, DAS observations allow for source characterization, including magnitude and source parameter estimation, and in some cases focal mechanism determination.Beyond offline applications, DAS shows strong potential for real-time monitoring, particularly for Earthquake Early Warning (EEW). Its integration into operational systems requires tailored strategies to exploit dense spatial sampling and to address system-specific features such as directional sensitivity and strain-rate saturation.Here, we evaluate the applicability of EEW methodologies to DAS data using three interrogators from the ABYSS network in central Chile that sense 450km of offshore cables running parallel to the subduction trench. The region, characterized by frequent moderate-to-large earthquakes, provides an ideal testbed for assessing EEW performance and the potential for rapid alerting.We develop a real-time magnitude estimation approach suited to offshore DAS conditions, where direct P-wave signals are often weak and followed in the first seconds by stronger arrivals of secondary phases, and integrate it into the QuakeUp algorithm. Performance is assessed using M≥4 events and 60 days of continuous data to evaluate both source characterization capability and robustness to false alerts and missed detections. Our results demonstrate the feasibility of a prototype DAS-based EEW system and highlight its potential to improve response times in high-seismicity regions.Part of this work has been performed by the Transnational Access to the GeoAzur laboratory MAREA (Magnitude estimAtion in Real TimE using DAS) supported by the EU project Geo-INQUIRE. Geo-INQUIRE is funded by the European Commission under project number 101058518 within the HORIZON-INFRA-2021-SERV-01 call.

The tsunami generated by the 2025 Mw 8.8 Kamchatka earthquake was recorded by a 450-km-long distributed acoustic sensing (DAS) array leveraging telecom cables offshore Chilean coasts. Tsunami waves of ~2 cm amplitude induced measurable strain on the cable despite the low-frequency sensitivity limitations of DAS. From the conversion of the cable distributed strain-rate to water elevations considering compliance and Poisson effects, we evaluate the potential contribution to tsunami warning systems. We estimate coastal tsunami arrival times and amplitudes based on assimilation alone of the DAS data recorded at distances ranging from 10 to 30 km off the coast. We find that warning can potentially provide 3 to 15 minutes of lead time before the tsunami waves reach the coast. We also show through synthetic waveform tests that the accuracy of both arrival-time and amplitude estimates improves as longer portions of the DAS record become available. This DAS-based approach highlights the potential of leveraging existing telecommunication cables as a cost-effective and complementary tsunami warning system in many exposed regions of the world.

Distributed Acoustic Sensing provides a transformative view of seismic wavefields, offering spatially continuous recordings that enable tracking of seismic phases along fibre optic cables. Offshore, fibres often cross sedimentary layers that generate phase conversions dominating early waveforms, increasing signal complexity and hindering rapid magnitude estimation for Earthquake Early Warning. Here, we analyse the peak amplitudes of seismic signals in a wide magnitude range (2.5≤ M ≤7.4), recorded by three interrogators along a 400km-long fibre array offshore Chile. We show that the direct P-phase is quickly dominated by secondary phases, limiting its use for source size estimation of moderate-to-large earthquakes. Conversely, we report that converted P-to-S waves carry a signature of the source similar to that of the S-phase within a few seconds from the first arrivals, offering a valuable proxy for rapid magnitude estimation. This finding reinforces the potential of Distributed Acoustic Sensing for Early Warning when sensing offshore cables. Converted P-to-S seismic waves detected in fibre optic cables offshore Chile carry information on earthquake source properties, like S-waves, but arrive much faster, like P-waves, and could represent a useful tool for rapid magnitude estimation.

Volcanic activity encompasses a wide range of seismogenic phenomena occurring from the deep magmatic conduit to the surface of volcanic flanks. Volcanic tremor, long period (LP), and very long period (VLP) seismic signals are commonly associated with magma and fluid movement within the conduit, whereas sliding mass and density currents along the flanks typically produce minute-long, cigar-shaped seismic traces.Characterising these phenomena through seismic analysis requires high measurement accuracy over a broad frequency range, together with high spatial and temporal resolution. Meeting these requirements in complex volcanic environments can be particularly challenging because the deployment and the maintenance of dense seismic networks involves considerable logistical effort. Distributed Acoustic Sensing (DAS) offers the opportunity to bridge the gap between sparse seismic networks and denser arrays, enabling continuous strain measurements along fibre-optic cables at comparatively low operational costs.Here, we investigate several different volcanic processes at Stromboli volcano (Italy) through DAS observations acquired along a 6 km fibre-optic cable integrated within a permanent multiparameter monitoring network comprising broadband seismometers, thermal and visible cameras, and infrasonic pressure sensors. The fibre was deployed on the volcanic flanks between 2020 and 2023 and interrogated during several month-long campaigns using a Febus A1-R. The dataset includes signals generated by ordinary Strombolian explosions, major explosions, lava flows, partial crater collapses and pyroclastic density currents (PDCs), which were analysed using different analytical approaches.Array-processing techniques in the 1–5 Hz frequency range were used to track the source of volcanic tremor, explosions, and PDCs with DAS strain-rate signals. Tremor and explosion signals are consistently located near the crater area, whereas PDCs propagate along the volcanic flanks. Moreover, by combining visible imagery with seismic energy recorded by DAS and inertial seismometers, we estimate the flow velocities and volumes of the PDCs and derive empirical, volume-dependent friction angles that provide insight into flow dynamics.Additionally, we exploit the distributed nature of DAS measurements to reconstruct the axisymmetric principal strain axes of VLP strain signals (between 0.04–0.2 Hz) associated with each explosion. The VLP strain signals recorded along the fibre nicely fit a deformation point-source (Mogi) located beneath the active craters, with an estimated volumetric change of ~30 m³.Our results demonstrate the capability of DAS measurements to characterise the dynamics of volcanic processes and to resolve the VLP strain distribution with enhanced spatial resolution. Overall, these findings highlight the significant potential of DAS as an innovative tool for analysing and monitoring a wide range of volcanic phenomena across different spatial and temporal scales.

Distributed acoustic sensing transforms fiber-optic cables into giant and very dense seismic networks. Although less sensitive to ground motion than traditional networks, they offer new possibilities for passive imaging and temporal monitoring, especially in hardly accessible locations such as the seafloor. From two case studies - in South of France, on a 42km long cable off-shore Toulon and in Central Chile on the northern leg of the Concón landing site of the GTD telecom cable - we explore the capability to perform passive imagery using ambient seismic noise and coda waves.Despite a higher instrumental noise level and uneven ground coupling, underwater telecom cables can record the microseismic noise. This may be strong microseismic noise generated locally, or microseismic noise amplified by the resonance of the water column. The recorded microseismic noise at the seafloor allows a better understanding of its generation and provides high resolution images of shallow crustal structures.From the observation of ocean gravity waves and microseismic noise, we highlight the strong localization of seismic noise sources near the coast, which can be highly variable over short time scales. Due to the localized nature of the noise sources, and because it is not always possible to average the noise recorded over long periods of time (months, years), conventional methods for ambient noise imagery show significant discrepancies in velocity estimates, up to 30%, especially at greater depths. We present here a method that minimizes the errors due to highly localized sources by carefully correcting the apparent velocities from the azimuth of the sources.In seismic areas, in addition to microseismic noise, it is possible to expand the frequency content toward higher frequencies using seismic coda. Coda waves are dominated by multi-diffracted surface waves on local heterogeneities. The spatial distribution of their energy is more isotropic. Using dispersion curves stacked over the coda of several earthquakes, we image the shallow crustal structure of the sediments. This innovative approach opens up new horizons for structural imaging and monitoring.In coastal environments, the distribution of noise sources must be systematically studied in order to obtain reliable results.

Distributed acoustic sensing (DAS) provides an attractive solution for ocean-bottom seismological instrumentation by providing a dense and long-distance measurement of the deformation of the ground along offshore submarine fiber-optic cables. This study reports analyses made on records acquired with a network located along the Chilean margin. We focus onto the analysis of the structure of the shallow crust, in particular, the sedimentary layer of the overlying crust, whose lateral variations suggest strong contrasts of the sedimentary recharge of the slab. The POST experiment was carried out from October 27 to December 3, 2021 on a fiber-optic cable connecting the city of Concón (100km northwest of Santiago) to La Serena. Using strain-rate recordings for twenty local and regional earthquakes, we estimated both the thickness and shear wave velocity of sediments. We used jointly (1) travel time delays between the direct P-wave and converted Ps at the bedrock/sediment interface that were estimated from manual picks and (2) coda wave interferometry. This later was done by identifying the phase velocities of the fundamental Rayleigh wave mode on frequency-wavenumber (FK) diagrams over 2km linear arrays along the fiber in the 0.3 to 7Hz frequency band. Each dispersive curve and travel time delays between the direct and converted wave were then jointly inverted to create a 2D S-wave velocity (Vs) structure of the sedimentary layer under the fiber. Our results show significant differences in thickness and in Vs along the cable. Two basins are observed, including the Valparaiso Forearc Basin separated by the Punta Salinas Ridge and another basin limited by a thin sedimentary layer with Vs of a few hundred m/s. In the extreme northern part of the cable, a thin layer of unconsolidated Quaternary sediments is on top of a deeper compacted sediments with faster Vs. The developed methodology comforts the potential of DAS for subsurface imaging purposes. Moreover, accurate modeling of the subsurface could be used to correct the location of earthquakes on the \iber sensors.

Distributed Acoustic Sensing (DAS) is establishing as a promising technique in Seismology. This novel system turns a fibre optic cable into a continuous single component array with very dense spatial sampling. Simplicity of installation and availability of telecommunication cables (dark fibres) make the technique very advantageous for investigating harsh environments like seafloors, sensing up to hundreds of kilometres of fibre with fine spatial resolution. Given the high potential, the technology has been successfully tested in recent years for several earthquake monitoring tasks, such as location, subsurface characterization, focal mechanism determination, tomography, or source back projection. The transferability of standard seismological tools to DAS data is straightforward when working with time picking, while analysis of the amplitude content of the signal demands further research. This is the case of earthquake source characterization, where standard approaches require conversion of strain rate data into more classical kinematic quantities (i.e. acceleration or velocity). In this work we develop a new formulation that allows to estimate source parameters without the need for conversion. We start from the description of the far-field strain radiation emitted from a circular seismic rupture, showing that the time integral of the strain is related to the Source Time Function. Using this quantity, we develop the spectral modelling allowing for frequency domain inversion of DAS data for estimation of moment magnitude and corner frequency. The formulation accounts for the unique azimuthal sensitivity of the cable in the radiation pattern average, and explicitly shows DAS enhanced sensitivity to slow scattered waves propagating beneath the fibre.We validated the proposed approach on two case-studies, for events in local magnitude range 0.4 - 4.3, comparing the results with estimates from standard seismic instruments. Earthquakes recorded on a 150km long cable offshore the coast of central Chile during a 1-month DAS survey exhibit scale invariant stress drops, with an average of Δσ=(0.8±0.6)MPa. Also, moment magnitude estimates agree with results from standard seismic instruments. The analysis of small magnitude events (ML

Distributed Acoustic Sensing (DAS) is becoming a standard solution for ocean-bottom seismological acquisition by providing a dense and long-distance measurement of ground deformation along offshore submarine fiber-optic cables. In the context of an offshore deployment in Central Chile, fiber-optic cables provide real-time seismic data dominated by scattered and converted phases. In a previous work, we have developed a methodology to determine both the velocity and thickness of the shallow sedimentary layer under the fiber using surface waves and split P-waves. Our current objective is to enhance the crustal imaging by identifying fault zones characterized by strong wavefront scattering and sharp lateral velocity contrasts, and sedimentary basins geometry at sub-kilometer scales in the same area, using scattered surface waves.We focus on seismic events recorded along a 150-km-long fiber in Central Chile. After partitioning the wavefield to separate direct waves from surface waves, we compute local backprojections of the scattered wavefield. By analyzing multiple seismic events across different frequencies, we investigate variations in wave propagation at multiple scales. The resulting energy profiles reveal spatially resolved fault zone structures and sharp lateral contrasts that align with topographic and structural features. Additionally, using standard seismic noise processing procedures, we compute time-domain cross-correlation functions, autocorrelations, and spectral densities. These analyses provide further insights into the behavior of surface waves near reflector features. For instance, we identify lateral discontinuities associated with basin edges by measuring their frequency-dependent resonance.Finally, to assess the seismogenic potential of the imaged structures, we will compare the geographical distribution and extent of the detected structures with the shallow seismicity automatically detected in the area using DAS data.

Applying Distributed Acoustic Sensing (DAS) to sense offshore telecom cables offers a unique opportunity to expand permanent seismological networks far offshore with an unprecedented instrumentation density but presents challenges in managing and processing data. This study introduces a workflow for automating earthquake catalog generation by combining DAS arrays with regional on-land seismic networks. We apply it to the ABYSS offshore permanent observatory in central Chile, which began continuous recording in October 2023 on three 150-km segments of a regional dark-fiber telecommunication network. Our workflow includes state-of-the-art components originally developed for conventional seismic sensor data. We first show that PhaseNet, pretrained on conventional seismic stations, can perform automatic phase picking on DAS strain-rate data converted to velocity. In a second step, we demonstrate that the GaMMA associator, tuned with parameters adapted to our dense data set, can cluster picked arrivals of small and moderate-sized earthquakes. We then adapt NonLinLoc to compute hypocenter locations using numerous phase picks, combining DAS and onland seismic data for improved accuracy. Finally, we show that reliable local magnitudes can be estimated on DAS recordings using the Richter scale. The earthquake catalog is stored in an SQL database using QuakeML formalism and is accessible via a user-friendly webpage. By applying the workflow to onshore station recordings, we evaluate the benefits of adding permanent seafloor DAS instrumentation, demonstrating that offshore DAS significantly improves the detection of small-magnitude earthquakes and reduces the offshore magnitude of completeness to 1.3.

Distributed Acoustic Sensing (DAS) technology facilitates the instrumentation of areas that are challenging to access with conventional instruments. In Chile, the presence of offshore submarine telecommunication cables offers a unique opportunity to instrument a major subduction zone close to the trench. Here we report an analysis of DAS data collected during a one–month campaign, sensing a commercial telecom cable connecting Concón to La Serena positioned several dozen kilometers off the coast. The earthquake recordings displayed P and S arrivals along with an additional Ps arrival, which is the result of the conversion of the P-wave at the bedrock/sediment interface. These three phase arrivals were identified and manually picked taking advantage of the spatial continuity of DAS measurements. To correctly account for the presence of the sediment layer in the localization procedure we introduced sedimentary corrections, which are a modification of the conventional station corrections. Instead of introducing an arbitrary constant time delay for each station and each phase, the corrections are derived from a physical first order modeling of the wave propagation in the sediments. The estimation of sedimentary parameters relies on: (i) the observed delay between the transmitted P-phase and the converted Ps-phase that give an indication of the sediment thickness; (ii) an inversion of the P- and S-wave speed in the sediments which is made possible thanks to the high sensor spatial density. We show that sedimentary corrections: (i) can represent most of the observed pick residual bias while only requiring the inversion of two global parameters (compared to station correction that requires three parameters per station); (ii) allow one to retrieve the sediment thickness and wave speed values that are consistent with common values for sediments; (iii) reduces the residuals of the earthquake hypocenter localization. The proposed correction method should improve the hypocenter estimation quality, facilitating the analysis of geological structures, and will contribute to a more detailed view of seismic activity in the studied area.

Owing to its deployment and sensing characteristics, Distributed Acoustic Sensing (DAS) has been touted as a promising technology for low-cost and low-latency Earthquake Early Warning (EEW). While preliminary experiments conducted by several research groups have yielded encouraging results, it must be acknowledged that these EEW feasibility studies were performed only on low-magnitude events. When exposed to the wavefield of a large magnitude earthquake (being the prime subject of EEW), the DAS strain rate recordings are likely to become highly distorted ("saturated") due to cycle skipping of the optical phase measurements, to an extent that the recorded data start to degrade to uniform random noise. This clearly poses a major challenge to EEW, as neither amplitude nor phase information can be readily extracted from saturated DAS data. In this study, we perform a detailed analysis of the dynamic range of DAS, both from theoretical and practical perspectives. We offer a set of criteria that need to be met for matching the DAS dynamic range with EEW targets, and we propose a computationally convenient method to quantify the information content of saturated recordings. We apply these methods to DAS data recorded offshore Chile, and identify several avenues for future research to improve the feasibility of DAS for EEW.

While several studies have shown the possibility to estimate earthquake magnitude from analysis of the S-phases recorded along fiber optic cables using Distributed Acoustic Sensing (DAS), understanding the source information hidden in the first seconds of these seismic recordings is still an open question requiring further investigation. In fact, in the case of submarine cables, with the fibers located closer to the epicenters, DAS could also be an asset for Early Warning of offshore earthquakes. In this study, we explore the possibility of measuring the size of the earthquake from the first few seconds of the signal received by the DAS by relating measurements of peak amplitudes to earthquake magnitude. We analyze a dataset of over 100 events (2.5 < M < 7.4) recorded along three submarine dark fibers running parallel to the Chilean margin, between Concon and La Serena, forming an approximately 450-km-long linear array. Unfortunately, this sensing technique suffers several limitations that complicate the recording of the first direct seismic arrivals. These include a lower signal-to-noise ratio (SNR) compared to traditional seismometers, and a high sensitivity of the measured parameters to local medium heterogeneities. Additionally, the longitudinal sensitivity of DAS makes it challenging to detect P-waves when using horizontally deployed cables, which are common when telecom fibers are utilized. Finally, modern DAS interrogators struggle to record strong ground motions due to phase wrapping of the backscattered light within the fixed interval [-π, π], resulting in the saturation of DAS recordings during intense shaking. Despite these challenges, we show that the P wave is poorly informative about the seismic source due to the influence of a shallow sedimentary layer, which generates a dominant PS-converted phase in the early DAS data. However, we demonstrate that this converted phase can be effectively used to robustly estimate earthquake sizes up to magnitude 7 within few seconds from the recording of an event. Furthermore, we derive amplitude attenuation laws as a function of distance and magnitude, overcoming the limitations of saturation by leveraging records from large events occurring hundreds of kilometers away from the array. Overall, this work highlights the continued potential of DAS-based seismic monitoring infrastructures while providing valuable insights for the development of a new generation of DAS-based Earthquake Early Warning Systems.

Xdas is a Python library designed to manipulate distributed acoustic sensing (DAS) data. It provides a unified abstraction for reading any DAS file format into a standardized Python object, streamlining data handling across different acquisition systems. To address the challenge of massive, multifile data sets, Xdas aggregates data chunks into virtually contiguous arrays organized by instrument and acquisition. This structure allows for efficient spatial and temporal slicing while minimizing overhead. To enable scalable offline processing of massive DAS data sets, Xdas processes data in manageable chunks. To ensure processing continuity, Xdas uses a stateful pipes-and-filters architecture. Most Xdas operations are multithreaded by default to take full advantage of multicore systems. This approach also enables real-time data processing. Its built-in network streaming capabilities allow Xdas to be deployed on DAS instruments for custom, real-time workflows at the point of data generation. At its core, Xdas uses a labeled ND (N-dimensional) array structure that encapsulates both data values and coordinate metadata and can be used to handle any kind of data set (not just time–space DAS records). This data model adheres to the established standards provided by the NetCDF4/HDF5 formats and the Climate and Forecast conventions. Designed to mirror the application programming interfaces (APIs) of popular libraries such as NumPy, SciPy, and Xarray, Xdas minimizes the learning curve for new users. Its modular and extensible design means that adding support for a new file format or integrating a processing function typically requires less than 10 lines of code.