Session: Earthquake Characterization Using Fiber-optic Cables

Type: Oral

Date: 10/8/2024

Time: 12:00 PM

Room: Stanley Park Ballroom

After the 2018 M6.4 Hualien earthquake in Taiwan, the Milun fault Drilling and All-inclusive Sensing (MiDAS) project drilled two scientific boreholes on the hanging (Hole A) and foot walls (Hole B) of the Milun fault, which had 0.5-m co-seismic slip during the earthquake and caused serious damage in Hualien City. The project deployed a 7.5-km-long optical-fiber cable perpendicular to the Milun fault, passing through Hole A and B (from top to bottom) as well as the main fault zone on the surface. The downhole triaxial seismometer arrays were installed in the boreholes. Both distributed acoustic sensing (DAS) and traditional seismometer systems operate continuously in earthquake monitoring. In the present study, we manually detect earthquakes from the continuous seismic records in a period of 2023/3/16-5/31 (76 days) and identify two local seismic swarms occurring on 4/7 and 5/2. It is worth noting that the swarm on 5/2, including 75 events within 5 hours, has an extremely small ts-tp arrival time of ~0.15 s in the sensors in Hole B, suggesting that the events were very close to the borehole. Benefited from the combined seismic monitoring system of the DAS and downhole seismometers, we determine (i) the DAS node with the first arrival time of S wave in Hole B to estimate the event depth, as well as (ii) the P-wave polarization and amplitude in ground velocity from the downhole seismometers to obtain the horizontal location and magnitude. In the end, we recognize that these earthquakes occurred at the depths of ~10 m and at the locations ~50 m away from the Hole B wellhead with the magnitudes approximating to Mw -2. The ultra-shallow microearthquakes observed in the present study are perhaps a record value of natural earthquakes, which can further contribute to the seismological studies, such as, the earthquake initial-phase and finite-dimension issues, as well as temporal variation of physical parameters (e.g., rigidity and wave velocity) in the ultra-shallow structures.

Presenting Author: Yen-Yu

Additional Authors

Session: Urban Seismology [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Since August 2023, we have been continuously experimenting with DAS capabilities on the Southern Methodist University (SMU) campus using 2-10 km long commercial dark fiber with an OptaSense QuantX Strain Interrogator Unit (IU). During this period, we have recorded a variety of seismoacoustic signals, including natural events (e.g., earthquakes, thunderstorms) and anthropogenic signals (e.g., rotary subwoofer, vehicular movements, building demolitions, football games). All these records are well recorded in a wide frequency range of 0.01-250 Hz at a spatial spacing of 1-2 m with a gauge length of 4-16 m. The recorded signals are clearly visible in Frequency-Wavenumber (F-K) transform plots that are utilized to calculate apparent velocities. For several events, infrasound and conventional broadband measurements were made alongside the DAS measurements, enabling us to explore coupling effects. This presentation summarizes the variety of signals we have detected and explores some applications of the signals for exploring source and path.

Presenting Author: Stephen

Additional Authors

Session: An Innovative Photonic Vision of Volcanoes and Geothermal Systems

Type: Oral

Date: 10/10/2024

Time: 09:00 AM

Room: Stanley Park Ballroom

Seismic tomography, a concept with nearly five decades of history, has significantly advanced our understanding of subsurface structures. In particular, high-quality tomographic imaging can provide valuable insights into the dynamics of volcanic systems. Recent advances in processing algorithms and computational architectures have enabled a new rise in the applications of tomographic approaches to magmatic systems. Furthermore, the advent of fiber sensing technologies for seismological applications has opened new avenues for exploring volcanic systems at a resolution unattainable by conventional station networks. The high spatial sampling and ease of deploying distributed acoustic sensing (DAS) instruments on telecommunication cables make DAS a powerful tool for characterizing subsurface structures of volcanoes at unprecedented scales and resolutions compared to dedicated nodal deployments.

In this presentation, I showcase successful examples of traveltime tomographic applications to volcanic systems, where DAS data are inverted using novel physics-based and machine-learning processing algorithms. The extensive range of modern DAS instruments enables the creation of high-resolution images of structures in the upper crust, offering a new outlook on volcanic science. We conclude by describing how future applications could leverage the full bandwidth DAS recordings inverted using full-waveform methodologies in which primaries and secondary waves contribute to setting a new standard for tomographic imaging.

Presenting Author: Ettore

Additional Authors

Session: Earthquake Characterization Using Fiber-optic Cables [Poster]

Type: Poster

Date: 10/7/2024

Time: 05:00 PM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing (DAS) offers significant advantages for microseismic monitoring due to its high spatial sampling. Unlike traditional geophones, which are sparsely distributed and sample the seismic field every ten to hundreds of kilometers, DAS provides dense sampling with 1-meter spacing and rates up to several kHz. This high spatial resolution allows for unprecedented mapping of seismic wave propagation but also produces TeraBytes of data daily. For this reason, traditional seismological processing techniques are not designed to handle such volumes and high spatial resolution. Consequently, new processing methods are required.

We propose an innovative waveform-based detection method exploiting the high temporal and spatial sampling of DAS data. Our algorithm is based on the calculation of a coherence matrix through the analysis of waveform coherence along hyperbolas with varying curvature and vertex positions (Porras et al., 2024). We then use a convolutional neural network to classify the coherence matrices, effectively distinguishing between seismic events and noise.

The residual neural network is trained on a dataset comprising coherence matrices derived from synthetic DAS data, which simulate events with varying locations and focal mechanisms. To validate our methodology, we applied it to data collected from a 90 km telecommunication fiber in the Pyrenees, France, and a borehole fiber in an enhanced geothermal system in Utah (FORGE experiment).

J. Porras, D. Pecci, G. M. Bocchini, S. Gaviano, M. De Solda, K. Tuinstra, F. Lanza, A. Tognarelli, E. Stucchi, F. Grigoli, A semblance-based microseismic event detector for DAS data, Geophysical Journal International, Volume 236, Issue 3, March 2024, Pages 1716–1727, https://doi.org/10.1093/gji/ggae016

UniPisa thanks TotalEnergies (TE) for giving access to the data and Febus Optic for the great support and involvement during the data acquisition phase.

Presenting Author: Sonja

Additional Authors

Session: Earthquake Characterization Using Fiber-optic Cables

Type: Oral

Date: 10/8/2024

Time: 12:30 PM

Room: Stanley Park Ballroom

The reliable estimation of earthquake magnitude and stress drop are key in seismology. The novel technology of distributed acoustic sensing (DAS) holds great promise for source parameter inversion owing to the measurements’ high spatial density. In this study, I demonstrate the robustness of DAS for magnitude and stress drop estimation using the empirical Green’s function deconvolution method. This method is applied to 9 co-located earthquakes recorded in Israel following the 2023 Turkey earthquakes. Spectral ratios were stacked along the fiber, and fitted with a relative Boatwright source spectral model. Excellent fits were obtained even for similar sized earthquakes. Stable seismic moments and stress drops were calculated assuming the moment of one earthquake is known. DAS derived estimates were found to be more stable and reliable than those obtained using a dense accelerometer network. The results demonstrate the great potential of DAS for source studies.

Presenting Author: Itzhak

Additional Authors

Session: Urban Seismology [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Seismic monitoring using conventional broadband seismometers can be challenging in urban environments due to limitations of land access. To evaluate the use of fiber optic sensing in urban environments, we interrogate ~95 km-long pre-existing dark fiber in an area with heavy traffic in Dallas, Texas, using a strain Interrogator Unit. To compare with traditional seismometers, we deployed 10 three-component broadband seismometers at different locations along the fiber cable and tested the Distributed Acoustic Sensing (DAS) on different configurations. We observed some known and unknown challenges with using the commercial dark fiber to retrieve reliable estimates of ground motion. These include the presence of aerial fiber sections, slack fiber in handholes, and varying fiber cable coupling with the ground. We have provided a computational-based solution for the identification of sections of fiber that cannot be used to retrieve reliable ground motions. In addition, we provided a criterion for using the Frequency-Wavenumber (F-K) transform technique to convert the DAS strain to ground velocities with different ground coupling conditions. Using a modified F-K transform technique, we have reliably matched the earthquake ground velocity records of DAS with seismometer recordings.

Presenting Author: Jyoti

Additional Authors

Session: Filling the Data Gap: Ocean-bottom Sensing with Fiber-optic Cables [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

The scientific evidence of climate change has led to new energy policies aiming for a carbon-neutral economy by mid-century. Renewable energies, energy storage systems, and carbon footprint reduction are crucial for these goals. Subsurface fluid injection is common in these projects, including green hydrogen storage in salt caverns and CO₂ storage in porous rock formations. Offshore geothermal energy production also involves deep underground water injection. Induced seismicity is a global concern in fluid injection projects, necessitating risk assessments for future operations. Fluid injection changes crustal stress, generating earthquakes whose frequency and magnitude vary with the injection rate. Long-term monitoring plans are essential for these multi-decadal, capital-intensive projects. Mitigating induced seismicity involves addressing four major issues: studying natural seismicity to evaluate fault activity, preventing injection-triggered faults, ensuring underground storage integrity, and managing moderate induced seismicity to prevent social alarm. Offshore storage sites, though challenging for seismic monitoring, reduce risks by being far from populated areas.

Effective monitoring integrates optical fiber cables with Distributed Acoustic Sensing (DAS) systems and AI algorithms for seismic detection, analysis, and location. This technology provides high-density seismological records, being an unparalleled tool to unravel the active tectonics of a target region. This approach was tested at the CASTOR offshore gas storage site, where significant induced seismicity occurred in 2013 due to gas injection. The facility, now in a phase of hibernation, was monitored for three months in 2023 using DAS on a telecom fiber optic cable connecting the offshore platform with the land facility. We recorded and located low-magnitude earthquakes under the platform, undetected by the national seismic network. This successful implementation highlights the potential of advanced monitoring technologies in ensuring the safety and viability of critical energy projects.

Presenting Author: Arantza

Additional Authors

Session: Earthquake Characterization Using Fiber-optic Cables

Type: Oral

Date: 10/8/2024

Time: 11:45 AM

Room: Stanley Park Ballroom

The emerging distributed acoustic sensing (DAS) technology can convert a pre-existing telecommunication fiber cable of tens of kilometers in length into a dense seismic array of thousands of recording channels. Here, we present a workflow for leveraging DAS together with the conventional seismic network to analyze high-resolution earthquake source properties, including seismic phase picking, earthquake relocation, focal mechanism inversion, and high-resolution rupture imaging. We first introduce PhaseNet-DAS, a deep neural network model, for automated earthquake detection and seismic phase identification from DAS recordings. Following detection, our scalable cross-correlation framework, CCTorch, robustly measures differential phase traveltimes and relative polarities between detected seismic events. Furthermore, we have devised a matrix-free adjoint solver that performs double-difference relocation using data from thousands of DAS channels. We have also developed a data-driven method for reliably inverting absolute first-motion polarities on DAS, which are critical for tightly constraining the nodal plane orientations and accurately inverting high-resolution focal mechanisms. We have successfully applied these methods to several DAS arrays in California. Finally, we present the DAS-based high-resolution back-projection rupture imaging using the Mw6 crustal earthquake that occurred in Antelope Valley, CA in 2021 as an example. These successful research progresses underscore the potential of DAS as a next-generation seismic monitoring tool that can significantly enhance and complement existing seismic networks. With the extensive existing and proposed network of onshore and offshore telecommunication fiber cables, DAS could provide critical datasets for systematically investigating detailed seismic source properties.

Presenting Author: Jiaxuan

Additional Authors

Session: Sensing Technologies and their Latest Developments [Poster]

Type: Poster

Date: 10/7/2024

Time: 05:00 PM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing data is increasingly used to characterize near-source and far-field signatures of anthropogenic and natural sources. However, conversion from strain rate to particle velocity can be challenging in environments where near-field elastic properties and fiber coupling affect the recorded DAS signal. Calibrated signal amplitudes are critical for wavefield modeling efforts and quantitative interpretation of DAS data. We compare amplitudes of controlled-source signals, ambient noise, explosions, and earthquakes recorded on co-located fiber optic and nodal instrumentation and discuss optimal approaches for converting strain rate to ground motion. Variations in fiber geometry, emplacement, and signal frequency are analyzed to derive transfer functions between fiber and geophone data for different fiber arrays and source types. We show that while empirical relations for converting strain rate to acceleration are useful as a first order approximation, using a gradiometry-derived strain rate provides a more useful comparison. These results will inform future applications where well-calibrated amplitudes are required. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.

Presenting Author: Christian

Additional Authors

Session: Filling the Data Gap: Ocean-bottom Sensing with Fiber-optic Cables [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Detecting and locating marine mammals such as fin whales is critical to understanding their behavior and protecting their habitats. Traditional methods of monitoring using visual observations and tagging are limited in coverage and can be usefully complemented by acoustic methods. Fin whales are well suited for acoustic monitoring because males produce a stereotypical, 1-second, 20-Hz, down-swept chirp that is incorporated into songs associated with breeding and observed in non-song vocalizations in migratory groups. However, the deployment of dedicated hydrophone arrays is expensive. Distributed Acoustic Sensing (DAS) is an emerging technology that shows promise for passive acoustic monitoring of marine mammals in real time. DAS uses fiber optic cables as sensors by exploiting Rayleigh scattering, allowing continuous monitoring over distances of up to ~100 km with an unprecedented resolution of tens of meters. For 4 days in November 2021, a public domain DAS dataset was collected on the two submarine cables of the Ocean Observatories Initiative Regional Cabled Array that extend offshore Pacific City, Oregon. This experiment took place during the fin whale breeding season, and tens of thousands of 20 Hz calls are present in the data. The dataset includes DAS measurements on three fibers extending 65-95 km with 2 m channel spacing. In this study, we demonstrate a scalable example of detection and localization. Different detection techniques have been explored, from classical methods such as matched filtering and spectrogram cross correlation, to image processing, exploiting the new possibilities of DAS time-space representations. Due to the performance of spectrogram correlation, the final detection combines Gabor filters on the recorded wave field envelope and single channel matched filtering. This provides a set of detection times for automated localization using the time difference of arrival method. Scaling these techniques to the full dataset for tracking the whales and inferring their behavior will also be discussed. [Work supported by ONR].

Presenting Author: Quentin

Additional Authors

Session: Real-Time Monitoring and Warning with Fiber Optic Seismology

Type: Oral

Date: 10/8/2024

Time: 03:00 PM

Room: Stanley Park Ballroom

The southernmost Cascadia subduction zone presents considerable earthquake hazard from interplate and intraplate faults and is the most active region within the ShakeAlert EEW System’s reporting area. Both real-time performance assessments and offline simulations indicate that ShakeAlert would benefit from low-latency offshore seismic data in this region. To build towards this possibility, the USGS, Cal Poly Humboldt University, Luna Inc., and Seismics Unusual have begun a long-term effort to collect DAS data in this region to evaluate data quality and analysis methods. As a result of the Middle-Mile initiative and other partnerships, Cal Poly Humboldt’s campus is expected to be a regional hub for fiber optic cables in the coming years that cover the locked portion of the megathrust and dozens of active crustal faults. Since 2022, we have collected ~1.5 years of DAS data on a 15 km cable between Arcata and Eureka CA, including over 35 nearby earthquakes with magnitude estimates between 3.5 and 5.4, and over 400 earthquakes total. We are evaluating both real-time detection algorithms operating directly on the data acquisition system as well as algorithms for magnitude and location estimation. An automated detection algorithm running on the acquisition system detects M3 earthquakes by the time the P-wave has been recorded on about half the cable with a total latency of about 0.9 to 1.2 s between the first P-arrival on one end of the cable and the issuing of the detection. If the earthquake is within about 30 km from the cable, the detection is often issued before the first ShakeAlert message is published. By comparing with co-located nodal seismometer data, we have established the reliability of peak-strain measurements over four orders of magnitude. Magnitude estimates from peak-strain measurements and existing scaling relations are reliable once site effects are accounted for. We are investigating faster P-wave based magnitude estimates for warning applications, and future experiments will involve multiple cables and a variety of interrogators.

Presenting Author: Jeffrey

Additional Authors

Session: Real-Time Monitoring and Warning with Fiber Optic Seismology [Poster]

Type: Poster

Date: 10/7/2024

Time: 05:00 PM

Room: Stanley Park Ballroom

DAS seismology offers new opportunities for earthquake early warning (EEW) as existing “dark fibers” are used as dense linear arrays of strain meters. Here, we examine the statistical features of DAS measurements of the first few seconds after P-wave arrivals from small-to-moderate local earthquakes for applications to EEW. Our data are from a 15-km long fiber onshore between Arcata and Eureka near the Mendocino Triple Junction in Northern California, where a M6.4 occurred about 30 km away, near Ferndale, on December 20, 2022. We analyze 5 months of local seismicity following the 2022 Ferndale event, including 12 M>4 events and one M5.4 on January 1, 2023. We calculate DAS waveform statistics before and after P-wave arrivals for the purpose of designing a DAS-based “tripwire” alert mechanism for EEW for nearby large earthquakes, thus expanding EEW coverage wherever dark fibers are available. We identify 70 two-minute-long DAS records of local earthquakes using the Comcat event catalog and an automatic phase picker (PhaseNet-DAS; Zhu et al., 2023), and analyze their statistical features in both time and frequency domains in the ten seconds before and four seconds after a P wave is detected across a spatial segment of the fiber. We find that some features correlate with catalog magnitude including continuous wavelet transform coefficients, short-term average/long-term average (STA/LTA) ratios, and strain-rate amplitudes, and are largely independent of distance and azimuth. We suggest that near-field terms which are thought to be characteristic of large, but not small, rupture processes, may help broadly distinguish between large and small earthquakes on DAS fibers, thus improving our ability to characterize potentially hazardous earthquakes in a timely and effective manner.

Presenting Author: Theresa

Additional Authors

Session: Urban Seismology [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing along dark fiber has become a common place way to increase sensor density in urban environments. A common challenge for these surveys is noise characteristics. Fiber coupling along the conduit hosting the fiber unknown and often results in varying sensor quality along a fiber. Moreover, temporal changes in noises sources, either continuous or impulsive, can hinder analysis where specific noise assumptions are needed. Here, we present an initial analysis of different preprocessing techniques to improve data quality for ambient noise DAS survey conducted along a dark fiber located in New Mexico. Seven days of data was collected along a seven-kilometer-long dark fiber adjacent to a heavily used truck route which has distinct and temporally varying noise characteristics. We implement different techniques to identify and improve signal quality for noisy gauges, account for time-varying noise sources, and locate and remove impulsive events. We show how each of these techniques improves the correlation of ambient noise between gauges and improves our ability to detect transient velocity changes between sensors. Even though the survey target is urban environments, since most telecom fiber is installed adjacent to roads, we suggest our findings will also be applicable rural areas as well.

Presenting Author: Nathan

Additional Authors

Session: Exploring the Frontier of Environmental Processes using Fiber-optic Sensing [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing (DAS) represents a leap in seismic monitoring capabilities. Compared to traditional single-seismometer stations, DAS measures seismic strain at meter to sub-meter intervals along fiber-optic cables thus offering unprecedented temporal and spatial resolution. However, the effective use of the wealth of information provided by DAS for natural hazard and in particular mass movement monitoring remains a challenge.

We propose a semi-supervised neural network tailored to screen DAS data related to a major rock collapse on 15 June 2023 and precursory failures near Brienz, Eastern Switzerland. Besides DAS, our dataset from 16 May to 30 June 2023 includes Doppler radar data for ground-truth labeling. The proposed algorithm is capable of distinguishing between rock-slope failures and background noise, including road and train traffic, with a detection precision of over 90%. It identifies hundreds of precursory failures and shows sustained detection hours before and during the major collapse. Moreover, we have identified key performance dependencies: event size and signal-to-noise ratio (SNR). As a critical part of our algorithm operates in an unsupervised way we suggest that it is suitable for general monitoring of gravitational hazards.

The major collapse is characterized by strong low-frequency signals between 0.01 and 0.03 Hz across multiple channels of the cable's trajectory-parallel section. Most of this low-frequency energy arrives after the high-frequency signal. Our initial hypothesis is that the low-frequency signal relates to a bulk mass moving downslope. However, the Doppler radar does not fully capture this movement, possibly due to radar insensitivity immediately following the collapse. Further investigation will focus on force-history inversion to study the seismic source mechanisms during the major collapse.

Presenting Author: Jiahui

Additional Authors

Session: Urban Seismology

Type: Oral

Date: 10/10/2024

Time: 11:45 AM

Room: Stanley Park Ballroom

Distributed acoustic sensing (DAS) has gained significant traction in the oil and gas industry and shows promising potential for broader engineering applications, particularly in digital twins for smart cities. Digital twins create virtual replicas for smart cities, and DAS can provide complementary information for bidirectional communication between the virtual and physical worlds.

To explore the benefits of integrating DAS into digital twins for smart cities, our study designed a multi-story DAS array to recover building mode shapes through ambient noise interferometry. The impacts of weather parameters on building vibration modes were investigated. Our contribution leverages the ambient noise interferometry approach, utilizing a multi-story DAS array to enhance the depth and accuracy of building mode shape analysis. This method marks a significant departure from conventional practices by introducing a 3D perspective to structural health monitoring (SHM), a notable advancement over the 1D or 2D approaches typically employed.

By doing so, our work expands the toolkit available for SHM and opens new pathways for understanding the complex dynamics of buildings under various environmental influences. The inclusion of a DAS array facilitates a more nuanced capture of structural responses, promising improvements in the precision of monitoring and early detection of potential structural issues.

Presenting Author: Chen

Additional Authors

Session: Sensing Technologies and their Latest Developments [Poster]

Type: Poster

Date: 10/7/2024

Time: 05:00 PM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing (DAS) measures strain along an optical fiber at specific intervals. Although converting DAS strain into ground units seems straightforward, practical implementation can result in coupling and instrument response errors. To address this, we compared DAS strain measurements with total strain from an optical fiber strainmeter (OFS) within the same cable. We also simulated strain using single-point velocity measurements from a nearby broadband STS2 seismometer.
Our study includes both field and lab experiments. In the field, DAS and OFS were deployed together on Whidbey Island, Washington, on land and beneath a shallow waterway. We used a teleseismic event recorded by DAS and converted it to total strain by stacking recordings from each channel. The integrated DAS strain data were then compared with OFS measurements, showing highly similar waveforms, nearly identical amplitudes, and a correlation coefficient above 0.95. Strain simulations from seismic velocity also showed consistency with offshore recordings, though with an amplitude factor of ~2.5, suggesting imperfect coupling between the fiber and the ground.
In the lab, we introduced a new method for evaluating DAS amplitude response using DAS and OFS. This method employs a piezoelectric sensor to generate ground truth strain, avoiding biases from natural earthquakes. Lab findings indicated a -3.2 dB decrease in DAS recorded amplitude at low frequencies (0.01 Hz), emphasizing the need for amplitude calibration in such studies.
Our investigation of underwater signals revealed a strong correlation with tide speed, with signals concentrated around slope turning points on the eastern side. This localization was achieved by calculating the coherence between DAS and OFS recordings, providing insight into the unknown source of these underwater signals.

Presenting Author: Chih-Chieh

Additional Authors

Session: Exploring the Frontier of Environmental Processes using Fiber-optic Sensing [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Infrasonic and seismic signals generated by hypersonic objects transiting Earth’s atmosphere provide critical information to classify the arriving object as natural or anthropogenic. Historically, these sound waves have been recorded by permanent arrays of widely separated seismometers and infasound instruments. During the 24 September 2023 reentry of the OSIRIS-REx space capsule, a total of 12 km of fiber optic cable was deployed at two sites in Nevada, which each sampled the capsule’s acoustic wavefield with high spatiotemporal resolution by utilizing Distributed Acoustic Sensing (DAS) technology. A straight-line cable array and a T-shaped array were deployed during the span of a few days directly onto the surface, exposing the cable to environmental forcings which may have induced signal distortion. Three separate Interrogator Units (IU’s) were used to capture DAS data, and multiple seismometers and infrasound sensors were co-located along the cable to inform our interpretation of DAS data and environmental noise sources. Here, we sample portions of the dataset recorded prior to the capsule arrival to investigate environmental sources affecting signals in the 5-100 Hz range which vary with the time of day, such as temperature, solar exposure, and wind speed. We quantify the signal distortion frequency range, intensity, and triggering factors to identify which noise sources may have impacted the spectral signature of the capsule arrival data. Finally, we compare the variations in noise captured by the three IU’s to determine which systems performed best given the most significant noise sources. Our results have implications for mitigating the impact of environmental noise sources in DAS datasets to better constrain target signals. Characterization of these noise sources will also aid in efforts to advance future design and operation of DAS in remote and extreme environments where large temperature and wind variations affect signal acquisition.

Presenting Author: Elisa

Additional Authors

Session: Urban Seismology [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing (DAS) has been shown to be a reliable method in the detection of various seismic sources from megathrust earthquakes to small-magnitude explosive sources. As Lior et al. (2023) showed, DAS can be possibly used for real-time magnitude estimation and ground motion prediction. Here we report the DAS recordings of mining blast events from the Six Mile and Shawville mines approximately 25 kilometers northwest of Philipsburg, Pennsylvania, and 50 kilometers northwest of the FORESEE DAS array using telecommunication fiber-optic cables beneath the Pennsylvania State University campus in the city of State College, Pennsylvania. Our DAS array detected 325 of 808 recorded blasts, 234 of which lay outside our monitoring period between 2019-2021, for a 57% detection rate and observed 99 more events than The Pennsylvania State Seismic Network (PASEIS). Observed explosive tonnage ranged from 6,156 to 57,780lbs (lowest to highest yield source observed) at 19 and 70-foot hole depths respectively. Using pre-detected events, we use template matching to search for hidden blast events. Our final goal is to assign magnitude estimates to the observed blasts by using the methods described in Yin et al. (2023). If this approach proves to estimate earthquake magnitude of low-yield explosive sources reliably and rapidly, we propose incorporating this method into geohazard early warning on the seismically vulnerable east coast of the United States.

Presenting Author: Joseph

Additional Authors

Session: Urban Seismology [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Dark fibers in the telecommunication network can be useful tools for urban site characterization with Distributed Acoustic Sensing (DAS). In 2022, we collected 5 days of continuous strain-rate data from the dark fibers on the Massachusetts Institute of Technology (MIT) campus in Cambridge, Massachusetts. The cables are buried 2 – 5 feet below ground, lying horizontally in an air-filled conduit. We performed a tap test with a hammer to locate the channels along the cable. The fiber captured environmental signals from human activities such as traffic, passing trains, and construction. We use ambient noise surface waves to investigate the shallow shear-wave velocity structure (Vs). To extract surface waves from ambient noise, we remove large transient signals and calculate cross-coherence between channel pairs using a 1-minute time window. We manage to extract coherent Rayleigh waves at two segments on the cable (40% of the total available length) after stacking in time and bin-stacking coherograms according to the inter-station distance. We use the dispersion relation of Rayleigh waves to invert for Vs structure on the top ~120 m. The resolution is constrained by (1) the length of the target cable segment and (2) twice the gauge length, which place bounds on the longest and shortest wavelength we can reliably resolve, respectively. The velocity profiles show low Vs (0.1 – 0.3 km/s) materials overlying a hard bedrock (1.5 – 1.8 km/s), which results in a strong impedance contrast at 75 m –95 m. Our results agree with previous studies on nearby sites using well logs, active and passive seismic surveys. The transfer functions simulated by the 1D Vs profile show resonance peaks at 0.6 – 1 Hz, which is important to consider for improving the seismic resistance of local buildings.

Presenting Author: Hilary

Additional Authors

Session: Sensing Technologies and their Latest Developments [Poster]

Type: Poster

Date: 10/7/2024

Time: 05:00 PM

Room: Stanley Park Ballroom

The power of distributed acoustic sensing (DAS) lies in its ability to sample deformation signals along an optical fiber at hundreds of locations with one interrogator only. While the interrogator is calibrated to record ‘fiber strain’, the properties of the cable and its coupling to the rock control the ‘strain transfer rate’ and such how much of ‘rock strain’ is represented in the recorded signal. We use DAS recordings in an underground installation (BFO observatory) colocated to an array of strainmeters in order to measure the ‘strain transfer rate’ in situ. A tight-buffered cable and a standard loose-tube telecommunication cable are used, where a section of both cables covered by sand and sandbags is compared to a section, where cables are just unreeled on the floor. The ‘strain transfer rate’ is largely independent of frequency in the band from 0.05 Hz to 1 Hz and varies between 0.15 and 0.55 depending on cable and installtion type. The sandbags show no obvious effect and the tight-buffered cable generally provides a larger ‘strain transfer rate’. The noise background for ‘rock strain’ in the investigated band is found at about an rms-amplitude of 0.1 nstrain in 1/6 decade for the tightbuffered cable. This allows a detection of the marine microseisms at times of high microseism amplitude. We recently cemented a "naked" optic fibre into the concrete floor of the observatory and the collected data will be presenred, which should give new insight into the cable coupling an sensitivity limits of the DAS recordings.

Presenting Author: Andreas

Additional Authors

Session: Sensing Technologies and their Latest Developments

Type: Oral

Date: 10/8/2024

Time: 09:30 AM

Room: Stanley Park Ballroom

Distributed Acoustic Sensing (DAS) can convert optical fibers into dense arrays of strainmeters. Amplitudes in DAS recordings, particularly for pre-existing telecommunication cables with uncertain fiber-ground coupling, have not been fully quantified. By calibrating with co-located seismometers, we systematically assessed the performance of three DAS arrays in California and one DAS array at the South Pole. Our results indicate that the average DAS amplitude of earthquake signals aligns with seismometer data, showing fluctuations within an order of magnitude across channels and frequencies from 0.01 to 10 Hz. The noise floor of DAS, crucial for detecting weak signals, is comparable to strong-motion stations within this frequency range but exceeds broadband stations below 0.1 Hz. The saturation amplitude of DAS, essential for capturing strong signals, is adjustable by modifying pulse repetition rate and gauge length, although this adjustment comes with trade-offs regarding array length and data quality. Our comprehensive evaluation highlights the potential of using both DAS phase and amplitude data in seismic studies.

Presenting Author: Qiushi

Additional Authors

Session: Earthquake Characterization Using Fiber-optic Cables [Poster]

Type: Poster

Date: 10/7/2024

Time: 05:00 PM

Room: Stanley Park Ballroom

DAS (Distributed Acoustic Sensing) technology provides an economical way to deploy thousands of linear strain meters along the fiber cable. To fully utilize the information of the strain field, at least three linear strain meters are needed at the same site to obtain a horizontal strain tensor. Taking the opportunity by the redeployment of the surface fiber cable of the MiDAS (Milun fault Drilling and All-inclusive Sensing project) at the end of 2022, three components of fiber cable oriented at 0-45-90 degrees with cable length of 20 meters each were deployed on the surface. This allows us to compute the full strain tensor from the three arms of the fiber cable, rotating the surface strains into a radial-transverse coordinate system. Together with a collocated 3-axis seismometer and 3-axis rotation sensor, 10 components of the elastic wavefield can be achieved. In this study we merge the strain, rotation, and seismometer measurements at the same station and investigate if we could determine the direction of wave propagation and phase velocity for specific wave type from the combination of these measurements.

Presenting Author: Chin-Jen

Additional Authors

Session: Exploring the Frontier of Environmental Processes using Fiber-optic Sensing [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

DAS recordings close to the coast are influenced by pressure signals from land- and seaward ocean surface gravity waves. The amplitude and period of the signal can be interpreted as a proxy for the sea state. Measurements along the cable at larger water depths show secondary microseisms related to the sea state away from the shore. The significant wave height (SWH) and ocean currents along the cable can be evaluated continuously during the experiment, resembling a dense sampling array of closely spaced buoys. However, in order to provide serviceable results, the measurements have to be calibrated with existing buoy or wave model data.

In October 2023, the GeoLab dark fibre off Madeira Island in the Atlantic was fitted with a DAS interrogator under a project by ARDITI and the Oceanic Observatory of Madeira. As a pilot site, the experiment is linked to the SUBMERSE project that is trying to establish continuous DAS monitoring along fibre-optic cables at multiple locations around Europe.

We use a one-week recording in late 2023 to show changes in the DAS data close to the shore where the water depth is small, and temporal variations of the secondary microseisms further along the cable. Spectrograms of individual channels and f-k spectra of different time intervals show effects of varying sea states, such as currents on the dispersion curves between land- and seaward waves.

To calibrate the measurements, the data are compared with measurements from a buoy located close to the coast and the DAS cable. In addition, we use model data from the Atlantic-Iberian Biscay Irish-Ocean Wave Analysis and Forecast model (0.05 x 0.05 deg; Toledano et al., 2022), which shows good correlation with the buoy data.

This work was funded by the Portuguese Fundação para a Ciência e a Tecnologia (FCT) I.P./MCTES through national funds (PIDDAC) – UIDB/50019/2020 (DOI: 10.54499/UIDB/50019/2020), UIDP/50019/2020 (DOI: 10.54499/UIDP/50019/2020) and LA/P/0068/2020 (DOI: 10.54499/LA/P/0068/2020), and by EC project SUBMERSE, HORIZON-INFRA-2022-TECH-01-101095055.

Presenting Author: David

Additional Authors

Session: How to Scale Up [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

This work presents an analysis of potential areas in Brazil for the application of Distributed Acoustic Sensing (DAS) technology with optical fiber, highlighting the difficulties of support from foundations and the private sector, thus proposing solutions to promote effective collaborations. In the last 10 years, Brazil has not yet explored the use of DAS with optical fiber. This study aims to identify promising regions for implementation, characterize the main difficulties encountered and suggest strategies to overcome these obstacles, facilitating the advancement of technology in the country.

Presenting Author: George

Additional Authors

Session: Filling the Data Gap: Ocean-bottom Sensing with Fiber-optic Cables - I

Type: Oral

Date: 10/9/2024

Time: 12:45 PM

Room: Stanley Park Ballroom

Various types of wavefield have been captured by submarine fiber optic cables, including earthquake and ocean waves. However, tsunami signals have been less captured because whether their observations depend on experimental periods and locations for distributed acoustic sensing (DAS). In southern Japan on October 8, 2023 (UT), sea level changes due to tsunamis were observed by tide gauges, and the tsunami generation is likely to be related to activities of submarine volcanoes. We observed the tsunami signals in the continuous records of DAS off Muroto in southern Japan. In this study, to investigate the characteristics of the tsunami signals, we calculate the frequency-dependent phase velocities, and compare the waveforms with those observed at nearby absolute pressure gauges (APGs).

In the DAS records, coherent signals propagating landwards are observed in the northern part of the submarine cable. The dominant frequency is 6–30 mHz. The frequency-dependent propagation velocities of the observed signals correspond to those of infragravity waves (i.e., deep water waves, ocean surface gravity waves), and the infragravity waves are categorized into high-frequency tsunamis.

We also estimated time-series of the tsunami generation at the source location. Relatively high frequency components within high-frequency tsunamis are excited in early stage of the volcanic activities, and their frequency components are shifted to middle and low ones in later stages. This indicates that the volcanic activities tend to be large at the later stage. The obtained features of the time-series were consistent with those using records of the APGs. Our results indicate that DAS records are useful for detecting tsunami propagations and also elucidating excitation mechanisms of tsunamis.

Data citation

DONET: doi:10.17598/NIED.0008

Acknowledgement

This work was supported by JSPS KAKENHI Grant No. JP21H05202, JP21H05204 in Scientific Research on Transformative Research Areas “Science of Slow-to-Fast earthquakes”.

Presenting Author: Takashi

Additional Authors

Session: Exploring the Frontier of Environmental Processes using Fiber-optic Sensing [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Ice wedge polygons are among the most distinctive landforms in permafrost regions. Conceptually, they form when freezing ground undergoes thermal contraction cracking during winter, followed by crack filling during the spring snowmelt. However, there is a lack of field observations hindering the understanding of cracking dynamics. In this study, we report over two thousand strong seismic signals throughout the winter months identified in the continuous recordings of yearly-deployed distributed acoustic sensing (DAS) system. These signals are characterized as short-duration events, with center frequencies ranging from 5 to 30 Hz. The events present body wave arrivals closely followed by surface waves, indicating a source near the installed fiber. We found a strong correlation between the daily occurrence of these events and periods of extremely low ground temperatures. Consequently, we interpret these events as cryoseisms originating from the permafrost. During periods of intense cold, thermal stress within the frozen ground can induce tensional fracturing, resulting in the sudden release of stress in the form of seismic waves, known as frost quakes. Through analysis of each event's moveout, we were able to distinguish the relative source locations. Our findings reveal distinct seismicity patterns between wet tundra and dry tundra environments. Finally, we modeled elastic wave propagation to examine the source mechanisms and compared several potential mechanisms for generating local seismicity within the permafrost with the observed waveforms. This research presents an innovative method by using DAS for monitoring and analyzing the dynamics of thermally fracturing frozen ground in the changing Arctic environment.

Presenting Author: Gabriel

Additional Authors

Session: Urban Seismology

Type: Oral

Date: 10/10/2024

Time: 12:00 PM

Room: Stanley Park Ballroom

In addition to regional earthquakes, Mexico City is also affected by local seismic events. To better understand these local phenomena, it is important to monitor local earthquakes and create a comprehensive catalog that includes events with very low signal-to-noise ratios as well as those with higher amplitudes.

Using data collected during a DAS experiment conducted from May 2022 to June 2023, we have detected nearly a hundred local seismic events using classical algorithms such as STA/LTA. However, manual inspection of some records revealed that STA/LTA missed many low-amplitude events, likely due to the noise produced by the city.

To address the limitations of classical detection algorithms, we propose the implementation of a convolutional neural network (CNN), which has been proven effective in similar situations. The challenge, however, is that a labeled dataset of only a hundred samples is insufficient to properly train this type of neural network. To overcome this limitation, we performed data augmentation on the original dataset to generate several thousand labeled samples. The advantage of this approach is that it eliminates the need for simulations or modeling, which would be time-consuming and computationally intensive given the complexity of Mexico City's geology.

Once trained, the proposed CNN can process the entire 120 terabytes of data acquired during the DAS experiment in just a couple of days, enabling the detection of hundreds of seismic events each month.

Presenting Author: Alfonso

Additional Authors

Session: Urban Seismology [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Localizing earthquakes is not always easy. In particular, we show that most of the localizations of local earthquakes in Mexico City can be seriously questioned. Although the seismic network has been densified for the 20 past years, the highly heterogeneous media in the city and the rapidly evolving ratios Vp/Vs through depth (about a factor of 100 in the first 80m of depth and then about 1.7 at 1km depth) make obsolete the general tools for localization. Worst, the first observed shear arrival does not always follow a direct path, and phase identification with a scattered seismic network in such a geological context is uncertain.Then, the high spatial and temporal resolution measurements in Mexico City, such as the ones associated with Distributed Acoustic Sensing (DAS) technology, are of significant interest. We experimented from 2022 to 2023 and detected a hundredth of local earthquakes. We show that it is easy to follow several phases through a large number of channels. However, DAS measurements conducted with dark fibers in urban environments don’t allow for observation of compressional wave.

We will also show that localizations can be improved by considering more than the direct P and S arrivals. However, identifying each phase is out of the question. Instead, we would apply convolutional neural networks (CNN), which are well-suited for this task due to their ability to capture spatial and local patterns within the data. The CNN architecture was designed to include multiple convolutional layers with appropriate kernel sizes to extract features from data, followed by fully connected layers to map these features to the 3D coordinates. Data preprocessing involved denoising, normalization, and segmentation of DAS signals to focus on relevant temporal windows associated with microseismic events. The model was trained using a supervised learning approach, where the known positions of microseismic events served as ground truth. A comprehensive evaluation using a separate validation set was conducted to assess the model’s performance and generalization capability.

Presenting Author: Kevin

Additional Authors

Session: Filling the Data Gap: Ocean-bottom Sensing with Fiber-optic Cables [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

The EU-INFRATECH funded SUBMERSE project will establish continuous monitoring of several oceanic telecom cables for landing sites in Portugal, Greece, and Svalbard. We develop tools for the automated analysis of these upcoming data sets as well as other submarine DAS data. Employing DeepLab v3, a cutting-edge deep neural network architecture renowned for its exemplary performance in semantic segmentation, this project aims to develop a specialized machine learning model for the detection of earthquakes and the identification of P and S waves using submarine DAS data. The inherently two-dimensional nature of our input data mandates the adoption of DeepLab v3.

Confronted with the distinctive challenges posed by submarine DAS data, which encompass varied oceanic noise environments and a spectrum of operational parameters including cable length, configuration, channel spacing, deployment conditions, and geographic diversity, we have chosen to implement a more expansive model to enhance detection capabilities. This project exclusively utilizes submarine DAS seismic records, now encompassing nearly three million earthquake records from multiple international locales. This approach ensures that our model is more effective to the unique characteristics of the submarine environment.

Our findings corroborate the model’s good capacity to detect seismic events and accurately categorize P and S waves using single-component multi-channel DAS data—a notable feature considering the absence of the three components typically essential for wave discrimination in conventional seismology. The model proficiently differentiates between P and S waves, even in scenarios where the order of these waves is reversed or their amplitudes are markedly altered, indicating that the model discerns complex patterns and features inherent in seismic data beyond mere amplitude analysis.

Presenting Author: Han

Additional Authors

Session: Filling the Data Gap: Ocean-bottom Sensing with Fiber-optic Cables [Poster]

Type: Poster

Date: 10/9/2024

Time: 07:00 AM

Room: Stanley Park Ballroom

Offshore Distributed Acoustic Sensing (DAS) has emerged as a powerful technology for seismic monitoring, expanding the capacities of cable networks and coastal seismic networks to monitor offshore seismicity. However, DAS data often combine signals unfamiliar to seismologists, including new types of instrumental noise, fiber cable coupling issues, and ocean signals that overprint those from tectonic sources, all of which hinder seismological research. We develop a self-supervised deep learning algorithm, a Masked Auto-Encoder (MAE), to denoise DAS data for seismological purposes. The model is trained on randomly masked DAS channel recordings of local earthquakes in the Cook Inlet, offshore Alaska. To demonstrate the benefits of denoising for seismological research, we conduct the most fundamental steps to any earthquake catalog building: seismic phase picking, signal-to-noise ratio estimates, and event association. We leverage the generalizability of ensemble deep learning models with cross-correlation to predict phase picks with sufficient precision for post-processing (e.g., earthquake location). The signal-to-noise ratio (SNR) of the denoised testing DAS data increased by 2. The MAE denoised DAS data allows manyfold more S picks than the original noisy data for smaller regional earthquakes. The results demonstrate that our self-supervised MAE holds significant potential for enhancing seismic monitoring with rapid earthquake characterization.

Presenting Author: Qibin

Additional Authors